Citation
Direct and Nearby Lightning Strike Interaction with Test Power Distribution Lines

Material Information

Title:
Direct and Nearby Lightning Strike Interaction with Test Power Distribution Lines
Creator:
SCHOENE, JENS DANIEL
Copyright Date:
2008

Subjects

Subjects / Keywords:
Data lines ( jstor )
Electric current ( jstor )
Electric potential ( jstor )
Flashover ( jstor )
Launching bases ( jstor )
Lightning ( jstor )
Lightning arresters ( jstor )
Power lines ( jstor )
Resistors ( jstor )
Transmission lines ( jstor )
Camp Blanding ( local )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Jens Daniel Schoene. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
7/12/2007
Resource Identifier:
659815032 ( OCLC )

Downloads

This item has the following downloads:

schoene_j ( .pdf )

schoene_j_Page_003.txt

schoene_j_Page_139.txt

schoene_j_Page_338.txt

schoene_j_Page_036.txt

schoene_j_Page_048.txt

schoene_j_Page_375.txt

schoene_j_Page_143.txt

schoene_j_Page_184.txt

schoene_j_Page_031.txt

schoene_j_Page_264.txt

schoene_j_Page_281.txt

schoene_j_Page_129.txt

schoene_j_Page_391.txt

schoene_j_Page_042.txt

schoene_j_Page_384.txt

schoene_j_Page_013.txt

schoene_j_Page_266.txt

schoene_j_Page_292.txt

schoene_j_Page_240.txt

schoene_j_Page_296.txt

schoene_j_Page_050.txt

schoene_j_Page_295.txt

schoene_j_Page_369.txt

schoene_j_Page_176.txt

schoene_j_Page_236.txt

schoene_j_Page_152.txt

schoene_j_Page_075.txt

schoene_j_Page_168.txt

schoene_j_Page_106.txt

schoene_j_Page_365.txt

schoene_j_Page_092.txt

schoene_j_Page_024.txt

schoene_j_Page_318.txt

schoene_j_Page_227.txt

schoene_j_Page_039.txt

schoene_j_Page_136.txt

schoene_j_Page_340.txt

schoene_j_Page_010.txt

schoene_j_Page_317.txt

schoene_j_Page_305.txt

schoene_j_Page_335.txt

schoene_j_Page_350.txt

schoene_j_Page_113.txt

schoene_j_Page_242.txt

schoene_j_Page_186.txt

schoene_j_Page_118.txt

schoene_j_Page_020.txt

schoene_j_Page_237.txt

schoene_j_pdf.txt

schoene_j_Page_279.txt

schoene_j_Page_019.txt

schoene_j_Page_307.txt

schoene_j_Page_099.txt

schoene_j_Page_283.txt

schoene_j_Page_060.txt

schoene_j_Page_395.txt

schoene_j_Page_096.txt

schoene_j_Page_304.txt

schoene_j_Page_140.txt

schoene_j_Page_257.txt

schoene_j_Page_155.txt

schoene_j_Page_213.txt

schoene_j_Page_203.txt

schoene_j_Page_018.txt

schoene_j_Page_386.txt

schoene_j_Page_144.txt

schoene_j_Page_352.txt

schoene_j_Page_145.txt

schoene_j_Page_341.txt

schoene_j_Page_407.txt

schoene_j_Page_120.txt

schoene_j_Page_052.txt

schoene_j_Page_226.txt

schoene_j_Page_399.txt

schoene_j_Page_151.txt

schoene_j_Page_043.txt

schoene_j_Page_009.txt

schoene_j_Page_051.txt

schoene_j_Page_228.txt

schoene_j_Page_332.txt

schoene_j_Page_196.txt

schoene_j_Page_402.txt

schoene_j_Page_301.txt

schoene_j_Page_259.txt

schoene_j_Page_171.txt

schoene_j_Page_165.txt

schoene_j_Page_379.txt

schoene_j_Page_100.txt

schoene_j_Page_252.txt

schoene_j_Page_300.txt

schoene_j_Page_094.txt

schoene_j_Page_329.txt

schoene_j_Page_086.txt

schoene_j_Page_410.txt

schoene_j_Page_204.txt

schoene_j_Page_104.txt

schoene_j_Page_166.txt

schoene_j_Page_097.txt

schoene_j_Page_378.txt

schoene_j_Page_331.txt

schoene_j_Page_008.txt

schoene_j_Page_127.txt

schoene_j_Page_289.txt

schoene_j_Page_053.txt

schoene_j_Page_232.txt

schoene_j_Page_111.txt

schoene_j_Page_249.txt

schoene_j_Page_247.txt

schoene_j_Page_006.txt

schoene_j_Page_254.txt

schoene_j_Page_416.txt

schoene_j_Page_046.txt

schoene_j_Page_061.txt

schoene_j_Page_251.txt

schoene_j_Page_420.txt

schoene_j_Page_234.txt

schoene_j_Page_109.txt

schoene_j_Page_282.txt

schoene_j_Page_397.txt

schoene_j_Page_084.txt

schoene_j_Page_275.txt

schoene_j_Page_208.txt

schoene_j_Page_403.txt

schoene_j_Page_343.txt

schoene_j_Page_310.txt

schoene_j_Page_326.txt

schoene_j_Page_190.txt

schoene_j_Page_200.txt

schoene_j_Page_012.txt

schoene_j_Page_214.txt

schoene_j_Page_280.txt

schoene_j_Page_398.txt

schoene_j_Page_142.txt

schoene_j_Page_074.txt

schoene_j_Page_022.txt

schoene_j_Page_270.txt

schoene_j_Page_055.txt

schoene_j_Page_392.txt

schoene_j_Page_374.txt

schoene_j_Page_179.txt

schoene_j_Page_357.txt

schoene_j_Page_021.txt

schoene_j_Page_014.txt

schoene_j_Page_037.txt

schoene_j_Page_346.txt

schoene_j_Page_291.txt

schoene_j_Page_056.txt

schoene_j_Page_161.txt

schoene_j_Page_250.txt

schoene_j_Page_297.txt

schoene_j_Page_302.txt

schoene_j_Page_077.txt

schoene_j_Page_126.txt

schoene_j_Page_224.txt

schoene_j_Page_103.txt

schoene_j_Page_054.txt

schoene_j_Page_258.txt

schoene_j_Page_334.txt

schoene_j_Page_349.txt

schoene_j_Page_382.txt

schoene_j_Page_358.txt

schoene_j_Page_112.txt

schoene_j_Page_038.txt

schoene_j_Page_206.txt

schoene_j_Page_212.txt

schoene_j_Page_298.txt

schoene_j_Page_210.txt

schoene_j_Page_169.txt

schoene_j_Page_325.txt

schoene_j_Page_389.txt

schoene_j_Page_324.txt

schoene_j_Page_033.txt

schoene_j_Page_342.txt

schoene_j_Page_231.txt

schoene_j_Page_133.txt

schoene_j_Page_122.txt

schoene_j_Page_089.txt

schoene_j_Page_026.txt

schoene_j_Page_263.txt

schoene_j_Page_029.txt

schoene_j_Page_183.txt

schoene_j_Page_002.txt

schoene_j_Page_163.txt

schoene_j_Page_066.txt

schoene_j_Page_269.txt

schoene_j_Page_034.txt

schoene_j_Page_417.txt

schoene_j_Page_191.txt

schoene_j_Page_078.txt

schoene_j_Page_195.txt

schoene_j_Page_361.txt

schoene_j_Page_105.txt

schoene_j_Page_339.txt

schoene_j_Page_199.txt

schoene_j_Page_262.txt

schoene_j_Page_080.txt

schoene_j_Page_359.txt

schoene_j_Page_347.txt

schoene_j_Page_355.txt

schoene_j_Page_063.txt

schoene_j_Page_243.txt

schoene_j_Page_162.txt

schoene_j_Page_098.txt

schoene_j_Page_057.txt

schoene_j_Page_287.txt

schoene_j_Page_194.txt

schoene_j_Page_364.txt

schoene_j_Page_368.txt

schoene_j_Page_413.txt

schoene_j_Page_032.txt

schoene_j_Page_239.txt

schoene_j_Page_185.txt

schoene_j_Page_272.txt

schoene_j_Page_217.txt

schoene_j_Page_314.txt

schoene_j_Page_160.txt

schoene_j_Page_245.txt

schoene_j_Page_076.txt

schoene_j_Page_319.txt

schoene_j_Page_180.txt

schoene_j_Page_406.txt

schoene_j_Page_412.txt

schoene_j_Page_153.txt

schoene_j_Page_135.txt

schoene_j_Page_158.txt

schoene_j_Page_128.txt

schoene_j_Page_058.txt

schoene_j_Page_070.txt

schoene_j_Page_209.txt

schoene_j_Page_246.txt

schoene_j_Page_225.txt

schoene_j_Page_381.txt

schoene_j_Page_333.txt

schoene_j_Page_293.txt

schoene_j_Page_371.txt

schoene_j_Page_028.txt

schoene_j_Page_294.txt

schoene_j_Page_284.txt

schoene_j_Page_030.txt

schoene_j_Page_323.txt

schoene_j_Page_216.txt

schoene_j_Page_223.txt

schoene_j_Page_004.txt

schoene_j_Page_110.txt

schoene_j_Page_065.txt

schoene_j_Page_073.txt

schoene_j_Page_286.txt

schoene_j_Page_222.txt

schoene_j_Page_035.txt

schoene_j_Page_253.txt

schoene_j_Page_238.txt

schoene_j_Page_148.txt

schoene_j_Page_229.txt

schoene_j_Page_027.txt

schoene_j_Page_170.txt

schoene_j_Page_322.txt

schoene_j_Page_198.txt

schoene_j_Page_182.txt

schoene_j_Page_313.txt

schoene_j_Page_108.txt

schoene_j_Page_255.txt

schoene_j_Page_219.txt

schoene_j_Page_385.txt

schoene_j_Page_146.txt

schoene_j_Page_354.txt

schoene_j_Page_137.txt

schoene_j_Page_134.txt

schoene_j_Page_351.txt

schoene_j_Page_316.txt

schoene_j_Page_025.txt

schoene_j_Page_016.txt

schoene_j_Page_248.txt

schoene_j_Page_261.txt

schoene_j_Page_367.txt

schoene_j_Page_220.txt

schoene_j_Page_405.txt

schoene_j_Page_372.txt

schoene_j_Page_419.txt

schoene_j_Page_189.txt

schoene_j_Page_387.txt

schoene_j_Page_083.txt

schoene_j_Page_276.txt

schoene_j_Page_101.txt

schoene_j_Page_321.txt

schoene_j_Page_244.txt

schoene_j_Page_130.txt

schoene_j_Page_309.txt

schoene_j_Page_147.txt

schoene_j_Page_177.txt

schoene_j_Page_187.txt

schoene_j_Page_288.txt

schoene_j_Page_005.txt

schoene_j_Page_085.txt

schoene_j_Page_373.txt

schoene_j_Page_370.txt

schoene_j_Page_007.txt

schoene_j_Page_362.txt

schoene_j_Page_041.txt

schoene_j_Page_091.txt

schoene_j_Page_047.txt

schoene_j_Page_123.txt

schoene_j_Page_388.txt

schoene_j_Page_017.txt

schoene_j_Page_233.txt

schoene_j_Page_044.txt

schoene_j_Page_087.txt

schoene_j_Page_023.txt

schoene_j_Page_353.txt

schoene_j_Page_049.txt

schoene_j_Page_202.txt

schoene_j_Page_068.txt

schoene_j_Page_344.txt

schoene_j_Page_067.txt

schoene_j_Page_218.txt

schoene_j_Page_117.txt

schoene_j_Page_093.txt

schoene_j_Page_285.txt

schoene_j_Page_315.txt

schoene_j_Page_114.txt

schoene_j_Page_193.txt

schoene_j_Page_173.txt

schoene_j_Page_181.txt

schoene_j_Page_312.txt

schoene_j_Page_174.txt

schoene_j_Page_320.txt

schoene_j_Page_215.txt

schoene_j_Page_265.txt

schoene_j_Page_081.txt

schoene_j_Page_150.txt

schoene_j_Page_230.txt

schoene_j_Page_274.txt

schoene_j_Page_197.txt

schoene_j_Page_360.txt

schoene_j_Page_390.txt

schoene_j_Page_414.txt

schoene_j_Page_404.txt

schoene_j_Page_211.txt

schoene_j_Page_308.txt

schoene_j_Page_040.txt

schoene_j_Page_090.txt

schoene_j_Page_271.txt

schoene_j_Page_401.txt

schoene_j_Page_115.txt

schoene_j_Page_059.txt

schoene_j_Page_095.txt

schoene_j_Page_396.txt

schoene_j_Page_188.txt

schoene_j_Page_393.txt

schoene_j_Page_141.txt

schoene_j_Page_154.txt

schoene_j_Page_138.txt

schoene_j_Page_088.txt

schoene_j_Page_172.txt

schoene_j_Page_377.txt

schoene_j_Page_409.txt

schoene_j_Page_330.txt

schoene_j_Page_260.txt

schoene_j_Page_102.txt

schoene_j_Page_125.txt

schoene_j_Page_079.txt

schoene_j_Page_411.txt

schoene_j_Page_119.txt

schoene_j_Page_178.txt

schoene_j_Page_121.txt

schoene_j_Page_072.txt

schoene_j_Page_337.txt

schoene_j_Page_164.txt

schoene_j_Page_167.txt

schoene_j_Page_241.txt

schoene_j_Page_156.txt

schoene_j_Page_157.txt

schoene_j_Page_421.txt

schoene_j_Page_107.txt

schoene_j_Page_348.txt

schoene_j_Page_069.txt

schoene_j_Page_015.txt

schoene_j_Page_064.txt

schoene_j_Page_306.txt

schoene_j_Page_336.txt

schoene_j_Page_311.txt

schoene_j_Page_201.txt

schoene_j_Page_001.txt

schoene_j_Page_131.txt

schoene_j_Page_124.txt

schoene_j_Page_205.txt

schoene_j_Page_235.txt

schoene_j_Page_207.txt

schoene_j_Page_045.txt

schoene_j_Page_192.txt

schoene_j_Page_175.txt

schoene_j_Page_383.txt

schoene_j_Page_071.txt

schoene_j_Page_356.txt

schoene_j_Page_082.txt

schoene_j_Page_408.txt

schoene_j_Page_299.txt

schoene_j_Page_159.txt

schoene_j_Page_116.txt

schoene_j_Page_363.txt

schoene_j_Page_268.txt

schoene_j_Page_327.txt

schoene_j_Page_376.txt

schoene_j_Page_273.txt

schoene_j_Page_415.txt

schoene_j_Page_149.txt

schoene_j_Page_267.txt

schoene_j_Page_328.txt

schoene_j_Page_380.txt

schoene_j_Page_278.txt

schoene_j_Page_221.txt

schoene_j_Page_290.txt

schoene_j_Page_062.txt

schoene_j_Page_400.txt

schoene_j_Page_256.txt

schoene_j_Page_303.txt

schoene_j_Page_345.txt

schoene_j_Page_011.txt

schoene_j_Page_418.txt

schoene_j_Page_366.txt

schoene_j_Page_132.txt

schoene_j_Page_394.txt

schoene_j_Page_277.txt


Full Text











DIRECT AND NEARBY LIGHTNING STRIKE INTERACTION
WITH TEST POWER DISTRIBUTION LINES














By

JENS DANIEL SCHOENE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2007

































Copyright 2007

by

Jens Daniel Schoene
















ACKNOWLEDGMENTS

I would like to thank Dr. M.A. Uman and Dr. V.A. Rakov for their guidance,

advice, and support. Furthermore, I would like to thank Dr. C.T. Mata and A.G. Mata for

their work on the FPL experiment in 1999-2003, results of which were used in this

dissertation, and K.J. Rambo who was always there to "get things done." The

experimental data presented in this dissertation were obtained in a team effort. The team

players involved were M.V. Stapleton, J.E. Jerauld, Dr. D.M. Jordan, G.H. Schnetzer,

B.D. Hanley, J.Howard, B.D. DeCarlo, R. Sutil, A. Guarisma, G. Bronsted, and A. Mata.

I also wish to thank Dr. C.A. Nucci, Dr. F. Rachidi, Dr. M. Paolone, and Dr. E. Petrache

for their invaluable help with the modeling of results from the nearby strike experiments

and for the pleasure of working with them at Camp Blanding during the summers of 2002

and 2003. Last but not least, I would like to thank my wife Gisele and my son Gabriel for

their patience and good behavior, respectively.

The research reported in this thesis was funded in part by the Florida Power and

Light Company, the Lawrence Livermore National Laboratory, and NSF.











TABLE OF CONTENTS


page

ACKNOWLEDGMENT S ............ ...... ._._ .............._ iii..

TABLE OF CONTENT S............... ............... iv

LIST OF TABLES ........._.._ ..... ._._ .............._ ix...

LIST OF FIGURES ........._.._ ..... ._._ ..............xiii...

AB STRAC T ................ .............. xxix

CHAPTER

1 INTRODUCTION .............. ............... ..........1

1.1 1999 FPL Experiment, Direct Lightning Strike Interaction with a
Horizontally-configured Distribution Line ................ .... ......... .................3
1.2 2000 FPL Experiment, Direct Lightning Strike Interaction with a
Horizontally-configured Distribution Line ................ .... ......... .................5
1.3 2001 FPL Experiment, Direct Lightning Strike Interaction with a
Vertically-conf gured Distribution Line ................ ....... ..... ..........8
1.4 2002 FPL Experiment, Direct Lightning Strike Interaction with a
Vertically-conf gured Distribution Line .................. ... ........ .........9
1.5 2002 FPL Experiment, Nearby Lightning Strike Interaction with a
Vertically-conf gured Distribution Line .................. .... ......... ................. 11
1.6 2003 FPL Experiment, Direct Lightning Strike Interaction with a
Vertically-conf gured Distribution Line ................... ........... .... .............. .12
1.7 2003 FPL Experiment, Nearby Lightning Strike Interaction with a
Vertically-conf gured Distribution Line .................. .... ......... ................. 13
1.8 2004 FPL Experiment, Direct Lightning Strike Interaction with a
Vertically-configured Distribution Line with Overhead Ground Wire ...........15
1.9 2005 Lawrence Livermore Experiment, Induced Currents ........._.._.................16
1.10 Summary of Original Contributions .............. .....................17

2 LITERATURE REVIEW .............. .................... 20

2.1 Cloud Formation and Electrifieation .............. ...............20....
2.1.1 Formation of a Cumulonimbus ................. ............_._......... ..21
2. 1.2 Electrical Structure of a Cumulonimbus............... ..............2
2.1.3 Electrifieation of a Cumulonimbus ................. ....._._. ........._....25
2.2 Natural and Rocket-triggered Lightning ................. ............ .. ............... .27
2.2.1 Lightning Discharges between Cloud and Ground ................... ..........28
2. 2.2 Rocket-tri gge red Li ghtning................ ..............3
2.2.3 Return Stroke Current........................ ............3
2.3 Transmission-line Type Return Stroke Models ................ ............ .........39











2.4 Distribution Line Design Parameters ................. ...............................42
2.4.1 Insulation Strength of Distribution Lines ................. ............... .....43
2.4.2 Overhead Ground Wires ................ ...............44................
2.4.3 Metal-Oxide Arresters .................... ..... ..............4
2.5 Modeling Direct Strikes to Power Distribution Line ................. ................. .49
2.5.1 History of the Electromagnetic Transient Program ................... .........50
2.5.2 EMTP Current Sources ................. ...............51................
2.5.3 EMTP Arrester Model ................ ...............51...
2.5.4 EMTP Transmission Line Models................ ......... ............... 5
2.5.5 Leads Connecting the Neutral Conductor to Ground Rods ................53
2.5.6 Ground Rod M odel ........................... .... ...........5
2.6 Modeling Nearby Strikes to Power Distribution Lines................ .................5
2.6.1 Calculation of Lightning-induced Overvoltages. .............. .... ........._..56
2.6.2 Testing of Lightning-induced Overvoltage Models.............._.._.. ........59
2.6.3 LIOV-EMTP96 Code ................. .... .. .....................6
2.7 Experimental Studies of Lightning Strike Interaction with Power Lines.........61
2.7.1 Japanese Study of Nearby Rocket-Triggered Lightning Strike
Interaction with a Test Distribution Line (1977-1985) ....................61
2.7.2 South African Study of Direct and Nearby Natural Lightning Strike
Interaction with a Test Distribution Line (1978, 1979) ......................62
2.7.3 DoE Study of Direct Natural Lightning Strike Interaction with
Distribution Lines (1978)................. ........... .. .. ........6
2.7.4 DoE Study of Nearby Natural Lightning Strike Interaction with
Distribution Lines (1979) ................... .... .._.._..... ..... _.._...........6
2.7.5 Japanese Study of Nearby Natural Lightning Strike Interaction with
a Distribution Line (1980-1988) ................. ... .................... 6
2.7.6 DoE Study of Nearby Natural and Rocket-triggered Lightning
Strike Interaction with Distribution Lines (1985, 1986) ....................65
2.7.7 EPRI Study of Direct and Nearby Natural Lightning Strike
Interaction with Distribution Lines (1987-1990) ........._..... ..............66
2.7.8 Japanese Study of Direct Natural and Rocket-triggered Lightning
Strike Interaction with a Test Transmission Line (1987-1996) .........67
2.7.9 EPRI Study of Direct and Nearby Rocket-triggered and Natural
Lightning Strike Interaction with Distribution Lines (1993-1996)...69
2.7.10 Japanese Study of Direct Natural Lightning Strike Interaction with
a Test Distribution Line (1999) .................. .. ............ ................. .72
2.7.11 FPL Study of Direct and Nearby Rocket-triggered Strike
Interaction with Distribution Lines (1999-2004) .............. ..... ........._.75

3 EX PE RIMENT AL F AC LITY ............... ...............7

3.1 Rocket Launchers. ........... ......__ ...............79...
3.1.1 Stationary Launcher .....__.....___ ..........._ ............7
3.1.2 Mobile Launcher................ ...............8
3.2 Test Distribution Lines .............. .. .......... ...................8
3.2.1 Horizontally-configured Line, 1999 and 2000 .............. ..............82
3.2.2 Vertically-configured Line, 2001 through 2003 .............. .... ............_84











3.2.3 Vertically-configured Line with Overhead Ground Wire, 2004.........88
3.3 Grounding of the Test Distribution Lines ....._____ .........__ ............... 90
3.3.1 1999 Experiment, Grounding ............_...... .__ .................91
3.3.2 2000 Experiment, Grounding .............. ...............91....
3.3.3 2001 Experiments, Grounding ....._____ ..... ... .__ ...........__....92
3.3.4 2002 and 2003 Experiments, Grounding ............___ .........__ ......93
3.3.5 2004 Experiment, Grounding .............. ...............93....
3.4 Arresters on the Test Distribution Lines ....._____ .........__ .............. .94
3.4.1 1999 Experiment, Arresters ............_...... .__ ........._........94
3.4.2 2000 Experiment, Arresters .............. ...............94....
3.4.3 2001 Experiment, Arresters .............. ...............95....
3.4.4 2002 Experiment, Arresters .............. ...............95....
3.4.5 2003 Experiment, Arresters .............. ...............96....
3.4.6 2004 Experiment, Arresters .............. ...............96....
3.5 Line Terminators on the Test Distribution Lines................ ...............97
3.5.1 1999 Experiment, Line Terminators ....._____ .........__ ..............98
3.5.2 2000 Experiment, Line Terminators ....._____ .........__ ..............98
3.5.3 2001-2004 Experiments, Line Terminators .............. .............. .98
3.6 Instrumentation .............. ...............99....
3.6.1 Current Sensors............... ...............99
3.6.2 Voltage Dividers ......___ ..... ... .__ ....___ ...........10
3.6.3 Fiber Optic Link. ....__ ......_____ .......___ ............0
3.6.4 Anti-aliasing Filter ................. ...............107................
3.6.5 Data Recording Equipment ................. .............. ......... .....107
3.6.6 Wireless Control System ................. ...............110...............

4 EXPERIMENTAL CONFIGURATIONS AND RESULTS ................. ................1 12


4.1 Description and Terminology of Measured Parameters .............. .... ............1 12
4.2 1999 Experiment, Horizontally-configured Line ................. .........._ .....113
4.3 2000 Experiment, Horizontally-configured Line ................. .........._..__.....115
4.4 2001 Experiment, Vertically-configured Line ................. ........._._.........117
4.5 2002 Experiment, Vertically-configured Line ....._____ ..... .....___..........119
4.6 2003 Experiment, Vertically-configured Line ............... .. .. .... ......__ .........121
4.7 2004 Experiment, Vertically-configured Line with Overhead Ground Wire 124
4.8 2005 Experiment, Induced Currents ................. .............. ......... .....125
4.9 Triggering Results ................. ...............129................

5 DATA PRESENTATION ................. ...............133................


5.1 2003 FPL Direct Strike Experiment: Phase-to-neutral and Ground
Currents During Stroke FPLO312-5 ................ ........ ..... .... ................... ....134
5.2 2004 FPL Direct Strike Experiment: Complete Data Set for Stroke
FPLO403-2 .................... ... ... .. ... ...................13
5.3 2003 FPL Nearby Strike Experiment: Lightning Currents Traversing Soil
And Entering Line Grounding ................. ...............141...............











5.4 2005 Lawrence Livermore Experiment: Induced Currents on a Buried
Counterpoise and Vertical Wire............... ...............150.

6 DATA ANALYSIS, MODELING, AND DISCUSSION ................. ................ ..157

6.1 Current Consistency Check. .............. ... .. ............. ..... .. ....... .............6
6.1.1 Consistency of Inj ected Current and Total Ground Current .............161
6. 1.2 Consistency of Inj ected Current and Total Struck-phase-to-neutral
C current ....................... .. ... .. ..... ..................16
6.1.3 Consistency of "High" and "Low" Lightning Currents ................... .165
6. 1.4 Consistency of Lightning Currents from the 2000 Experiment........ 169
6.1.5 Consistency of Current from the 2004 Experiment ........................172
6.2 Characterization of the Lightning Return Stroke Current. ...........................175
6.2.1 Statistical Data of Lightning Return Stroke Current Parameters......176
6.2.2 Statistical Distributions of Lightning Return Stroke Current
Param eters ................. ....... ._._ ....... ..._.. .... .........7
6.2.3 Discussion of Lightning Return Stroke Current Parameters ............181
6.2.3.1 Return stroke current peaks ................. ............ .........181
6.2.3.2 Return stroke current 10-90% risetime .............................184
6.2.3.3 Return stroke current half-peak width ............... .... ...........186
6.2.3.4 Return stroke charge transfer ................. ........._.._.......187
6.3 Arrester Disconnector Operation and Flashovers ........._.._.. .. ......._.._.. ....190
6.3.1 Discussion of Arrester Disconnector Operation ............... .... ...........193
6.3.2 Discussion of Flashovers .................. ... ................ ... ............... 196....
6.4 Measured Lightning Current Division on the Test Distribution Lines..........198
6.4.1 Measured Phase-to-neutral and Ground Current Division ...............198
6.4.2 Discussion of the Measured Lightning Current Division ................. 205
6.5 Modeled Lightning Current Divisions on the Test Distribution Lines..........210
6.5.1 Modeled Lightning Current Division on the Horizontal Line..........211
6.5.2 Modeled Lightning Current Division on the Vertical Line .............220
6.5.3 Simple Model of the Lightning Current Division. ..........................23 5
6.5.4 Discussion of the Modeled Lightning Current Division .................. .241
6.6 Estimation of the Arrester-absorbed Energy..................... ..... .... ...._...........245
6.7 Modeling of Induced Currents on the Test Line due to Nearby Lightning ...248
6.8 Lightning Currents Traversing Soil and Entering Line Grounding ...............252
6.8.1 Analy si s of Lightning Currents Traversing S oil ............... .... ...........25 2
6.8.2 Discussion of Lightning Currents Traversing Soil ................... ........255
6.9 Induced Currents on a Buried Counterpoise............... .............26

7 SUMMARY OF ORIGINAL RESULTS .............. ...............263....

7.1 Data Consistency ............... .... ........ .... .. ........6
7.1.1 Data Consistency during 2000 Horizontal Line Experiment............264
7. 1.2 Data Consistency during 2002, 2003, and 2004 Vertical Line
Experim ents ............... ........ .. .......... ..........6
7.2 Characterization of the Lightning Return Stroke Current. ...........................266
7.3 Arrester Di sconnector Operation and Arrester Energy Ab sorption ...............267











7.4 Flashover Occurrence .............. ...............268....
7.5 Phase-to-neutral Current Division .............. ...............269....
7.6 Neutral-to-ground Current Division .............. ........... ................7
7.7 Lightning Currents Traversing Soil and Entering Vertical Line Grounding .271
7.8 Induced Currents on Distribution Lines........................ ...... ........7
7.9 Induced Currents on a Buried Counterpoise and Vertical Wire ....................272

8 RECOMMENDATIONS FOR FUTURE RESEARCH............... ...............27


APPENDIX


A MEASUREMENT SETTINGS .............. ...............277....

B DATA PRESENTATION ................. ...............296................


C RETURN STROKE CURRENT PARAMETERS .............. .....................6

D FLASHOVERS AND DISCONNECTOR OPERATION................ ...............37


LIST OF REFERENCES ............ ..... ._ ...............379...

BIOGRAPHICAL SKETCH .............. ...............391....










LIST OF TABLES


Table pg
3-1: Experimental configurations used in 1999 through 2004 experiments. Pole
numbers are identified in Figure 3-1. ............. ...............78.....

3-2: Conductor placement and specifications for the vertically-configured test
distribution line. ............. ...............84.....

3-3: Conductor placement and specifications for the vertically-configured test
distribution line with overhead ground wire. ............. ...............88.....

3-4: Grounding resistances in ohms for the horizontally-configured line tested
during the 1999 experiment. ....___.................. ...............91.....

3-5: Grounding resistances in ohms for the horizontally-configured line tested
during the 2000 experiment. ....___.................. ...............91. ....

3-6: Grounding resistances in ohms of the vertically-configured line tested during
the 2001 experiment. The number of ground rods is given in parenthesis. .............93

3-7: Measured and theoretically-derived grounding resistances in ohms for the
vertically-configured line. The number of ground rods is given in parenthesis. .....93

3-8: Grounding resistances in ohms measured on 7/12/2004 for the vertically-
configured line tested during the 2004 experiment. The number of ground
rods is given in parenthesis. ............. ...............93.....

3-9: VI-characteristics of the Cooper Power Systems Ultra SIL Housed VariSTAR
Heavy Duty and Ohi o-B ras s PDV- 1 00 arre sters ......___ ... .....__ ..............94

3-10: 1999 experiment, arresters used on the horizontally-configured test line. ..............94

3-11: 2000 experiment, arresters used on the horizontally-configured test line. ...............95

3-12: 2001 experiment, arresters used on the vertically-configured test line. .................. .95

3-13: 2002 experiment, arresters used on the vertically-configured test line. ...................96

3-14: 2003 experiment, arresters used on the vertically-configured test line. ...................96

3-15: 2004 experiment, arresters used on the vertically-configured test line with
overhead ground wire ................. ...............97........... ....

3-16: Termination resistors during the 2000 experiment (horizontally-configured
di stribution line). ............. ...............98.....

3-17: Termination resistors during the 2001 through 2004 experiments ................... ........99










3-18: Specifieations of the Pearson Electronics, Inc. current transformers. ................... .101

3-19: Specifieation of the T&M Research Products, Inc. current viewing resistors. ......102

3-20: Specifieations for the capacitive-compensated voltage dividers used in the
1999 and 2000 experiments............... ..............10

3-21: Specifieations for each of the four loops of the magnetic-flux-compensated
voltage dividers used in the 1999 and 2000 experiments ................. ................. 105

3-22: Recording devices used during the 1999 through 2004 experiments. The
number of recording devices used for a particular year is given. .................. .........108

4-1: 1999 experiment, references to the measurement settings for all rocket-
triggered lightning strikes. All strikes triggered to the test line in 1999
contained return strokes. ................ ......... ...............115 .....

4-2: 2000 experiment, references to the measurement settings for all rocket-
triggered lightning events. Event IDs in bold printing denote flashes with
return strokes and event IDs in italic printing denote flashes without return
stroke s ................ ...............117................

4-3: 2001 experiment, references to the measurement settings for all rocket-
triggered lightning events. Event IDs in bold printing denote flashes with
return strokes and IDs in italic printing denote flashes without return strokes......119

4-4: 2002 experiment, references to the measurement settings for all rocket-
triggered lightning events. Event IDs in bold printing denote flashes with
return strokes and IDs in italic printing denote flashes without return strokes......121

4-5: 2003 experiment, references to the measurement settings for all rocket-
triggered lightning events. Event IDs in bold printing denote flashes with
return strokes and event IDs in italic printing denote flashes without return
stroke s ................ ...............123................

4-6: 2004 experiment, references to the measurement settings for all rocket-
triggered lightning events. The event ID in bold printing denotes a flash with
return strokes and the event ID in italic printing denotes a flash without return
stroke s ................ ...............125................

4-7: Successful triggering events during the 1999 through 2004 experiments. ............130

4-8: Successful triggering events during the 2005 experiment ................. .................1 32

6-1: Retumn stroke current statistics for the 1999 through 2004 experiments. ...............177

6-2: Comparison of return stroke statistics ................ .......................... ......182










6-3: Disconnector operations and flashovers during the 2000-2004 experiments. .......191

6-4: Statistical information on the arrester and ground current equilibration times......200

6-5: Comparison of peak values and charge transfers of lightning currents and
currents entering the line through the pole 15 grounding. The charge transfers
were obtained by numerically integrating the measured currents over 1 ms
(return stroke current) and 10 ms (initial continuous current) time intervals. .......253

6-6: Comparison of 10-90% risetimes and half-peak widths of lightning currents
and currents entering the line through the pole 15 grounding............... ................25

A-1: 2002 measurement settings for flash FPLO236 (8/18/2002). ............. ..... ..........._279

A-2: 2002 measurement settings for flash FPLO240 (8/27/2002). ............. .................280

A-3: 2002 measurement settings for flashes FPLO244, FPLO245, and FPLO246
(9/13/2002). ............. ...............28 1....

A-4: 2003 measurement settings for flashes FPLO301, FPLO302, and FPLO303
(6/30/2003). ............. ...............283....

A-5: 2003 measurement settings for flashes FPLO305 and FPLO306 (7/6/2003). ........284

A-6: 2003 measurement settings for flashes FPLO312 and FPLO314 (7/13/2003). ......285

A-7: 2003 measurement settings for flashes FPLO315 and FPLO317 (7/14/2003). ......286

A-8: 2003 measurement settings for flash FPLO321 (7/18/2003). ............. ..... ........._..288

A-9: 2003 measurement settings for flashes FPLO329 and FPLO331 (7/22/2003). ......289

A-10: 2003 measurement settings for flash FPLO336 (8/2/2003). ............. ..................290

A-11: 2003 measurement settings for flash FPLO341 (8/7/2003). ................ ...............291

A-12: 2003 measurement settings for flashes FPLO342 and FPLO345 (8/11/2003). ......292

A-13: 2003 measurement settings for flashes FPLO347, FPLO348, and FPLO350
(8/15/2003). ................ ............... ......... ........ ......... ....... ..293

A-14: 2004 Measurement settings for flashes FPLO402 and FPLO403 (7/24/2004). ......295

C-1: Retumn stroke statistics for the 1999 experiment. ............. ......................6

C-2: Retumn stroke statistics for the 2000 experiment. ............. ......................6

C-3: Retumn stroke statistics for the 2001 experiment. ............. ......................6











C-4: Return stroke statistics for the 2002 direct strike experiment. ............. ................368

C-5: Return stroke statistics for the 2002 nearby strike experiment. ................... ..........3 69

C-6: Return stroke statistics for the 2003 direct strike experiment. ............. ................370

C-7: Return stroke statistics for the 2003 nearby strike experiment. ................... ..........371

C-8: Return stroke statistics for the 2004 experiment. ............. ....... .............37

D-1: Flashovers and disconnector operation during the 2000 horizontal line
experim ent. .............. ...............375....

D-2: Flashovers and disconnector operation during the 2001 vertical line
experim ent. .............. ...............376....

D-3: Flashovers and disconnector operation during the 2002 vertical line
experim ent. .............. ...............377....

D-4: Flashovers during the 2003 vertical line experiment. No disconnector
operated during this experiment. .............. ...............378....










LIST OF FIGURES


Figure page
2-1: Small thundercloud over a train and a power line. The picture was taken in
Finland at the Russian border. The thundercloud produced only a few
lightning strikes. ............. ...............21.....

2-2: Formation of sea/land-breeze thunderstorms. ............. ...............22.....

2-3: Cold front moves under warm front resulting in the formation of
cumulonimbus (frontal storm)............... ...............23.

2-4: Electrical structure inside a cumulonimbus............... ..............2

2-5: Precipitation theory (left) and convection theory (right) ................. ............... ....26

2-6: Discharge types for a thundercloud. .............. ...............27....

2-7: Simplified drawing of four discharges between cloud and ground. ................... ......28

2-8: Drawings illustrating some of the various processes comprising a negative
cloud-to-ground lightning flash ................. ...............29........... ....

2-9: Rocket-triggered lightning in Camp Blanding, Florida (Flash U9910). ................. .33

2-10: Sequence of events in classical rocket-triggered lightning. ............. ...................34

2-11: Typical triggered lightning return stroke current waveform measured at the
channel base (stroke FPLO315-2). The following return stroke current
parameters are illustrated: a) peak value, b) 10-90% risetime, c) half-peak
width, and d) charge transfer. .............. ...............38....

2-12: Current versus time waveforms specified by TL model at ground (z'= 0) and
at two heights zl' and z2' ................. ...............41..............

2-13: Current versus height z' above ground at time t = tl for the TL model ................... .42

2-14: Typical varistor VI-characteristic plotted on a log-log scale. ............. ..................46

2-15: Structure of a metal-oxide varistor. .............. ...............47....

2-16: Illustrations of gapless metal-oxide arresters. ............. ...............48.....

2-17: Lumped parameters representation of ground rods. The model consists of n
RLC-sections. .............. ...............55....

2-18: Measured pole 13 arrester current and voltage during a natural flash that
connected to pole 6............... ...............73...










2-19: Pole 13 arrester current during a natural lightning strike to pole 13. .......................74

3-1: Overview of the ICLRT. ........... ........... ...............77...

3-2: Stationary tower launcher employed from 1999 through 2004 primarily used
to simulate direct lightning strikes to the test line. The intercepting structure
(PVC poles and intercepting conductor) were only used for the 2002, 2003,
and 2004 experiments. ........... _............ ...............80...

3-3: Mobile bucket-truck rocket launcher placed a few meters from one end of the
vertically-configured test distribution line. ............. ...............81.....

3-4: Sketch of the horizontally-configured line ................. .............. ......... .....83

3-5: Sketch of the vertically-configured line tested from 2001 through 2003 .................85

3-6: Vertically-configured distribution line in 2003, arrester station at pole 10. ............86

3-7: Vertically-configured distribution line in 2003, arrester station with
transformer at pole 2. ............. ...............87.....

3-8: Vertically-configured distribution line with overhead ground wire in 2004,
arrester station at pole 10. ............. ...............89.....

3-9: Vertically-configured distribution line with overhead ground wire in 2004,
arrester station at pole 10, close-up view. .............. ...............89....

3-10: Grounding scheme for the vertically-configured distribution line. ..........................90

3-11: Vertically-configured distribution line in 2003, line termination at pole 15...........97

3-12: Measurement scheme used during the 1999 through 2004 experiments ................100

3-13: Magnetic-flux compensated voltage divider. ............. ...............104....

3-14: Diagram of the wireless control system topology used during the 2002
through 2004 experiments ................. ...............111................

4-1: 1999 test distribution line having a horizontal framing configuration with
measurement points and the two different lightning strike locations identified. ...114

4-2: 2000 test distribution line having a horizontal framing configuration with
measurement points and the two different lightning strike locations identified. ...116

4-3: 2001 test distribution line having a vertical framing configuration with
measurement points and the three different lightning strike locations
i denti fi e d ............... .............118............ ...










4-4: 2002 test distribution line having a vertical framing configuration with
measurement points and the four different lightning strike locations
i denti fie d ................ ...............120........... ...

4-5: 2003 test distribution line having a vertical framing configuration with
measurement points and the four different lightning strike locations
i denti fie d ................ ...............122........... ...

4-6: 2004 test distribution line having a vertical framing configuration and an
overhead ground wire with measurement points and the lightning strike
location identified ................. ...............125................

4-7: Satellite image of the International Center for Lightning Research and
Testing. The location of the rocket launch facilities, test house, and induced
current measurements are indicated. Additionally, the location of the electric
field derivative measurements (Stations 1, 4, 8, and 9) are shown including
their distances to the vertical wire measurement in meters ................. ................1 26

4-8: Satellite image of the experimental site of the induced currents experiments.
The obj ects relevant to this experiment and their locations relative to the
north-west corner of the counterpoises are indicated ................. .....................127

4-9: Vertical wire shown with peak current sensor cards. The length of the wire is
7 m ................. ...............129................

5-1: Measured currents for stroke FPLO312-5, 2 ms and 100 Cls time scales. ...............135

5-2: Stroke FPLO312-5, injected current and phase A-to-neutral currents. ................... 137

5-3: Measured currents for stroke FPLO403-1, 2 ms and 100 Cls time scales. ...............140

5-4: Individual ground currents and sum of ground currents leaving the system for
stroke FPLO341-1. Displayed on a) 1 ms and b) 50 Cls time scales. ......................143

5-5: Normalized lightning current injected into ground 11 m from pole 15 (blue)
and current inj ected into the line through the pole 15 ground (black) for
stroke FPLO347-1. The plateau in the pole 15 ground current may be related
to ground arcing at the lightning current inj section point. ................. ................ 144

5-6: Individual ground currents and sum of ground currents leaving the system for
stroke FPLO3 47- 1 ................ ...............144...............

5-7: Normalized lightning current injected into ground 11 m from pole 15 (blue)
and current inj ected into the line through the pole 15 ground (black) for
stroke FPLO347-2 ................. ...............145................

5-8: Individual ground currents and sum of ground currents leaving the system for
stroke FPLO347-2 ................. ...............145................










5-9: Normalized lightning current injected into ground 11 m from pole 15 (blue)
and current inj ected into the line through the pole 15 ground (black) for
stroke FPLO3 50-1 ......_. ................ ........._.._.......14

5-10: Individual ground currents and sum of ground currents leaving the system for
stroke FPLO3 50-1 ......_. ................ ........._.._.......14

5-1 1: Normalized lightning current inj ected into ground 11 m from pole 15 (blue)
and current inj ected into the line through the pole 15 ground (black) for the
ICC of flash FPLO34 7. ............. ...............147....

5-12: Individual ground currents and sum of ground currents leaving the system for
the ICC of flash FPLO347. ............. .....................147

5-13: Normalized lightning current injected into ground 11 m from pole 15 (blue)
and current inj ected into the line through the pole 15 ground (black) for the
ICC of flash FPLO348. ............. ...............148....

5-14: Individual ground currents and sum of ground currents leaving the system for
the ICC of flash FPLO348. ............. .....................148

5-15: Normalized lightning current injected into ground 11 m from pole 15 (blue)
and current inj ected into the line through the pole 15 ground (black) for the
ICC of flash FPLO3 50. ............. ...............149....

5-16: Individual ground currents and sum of ground currents leaving the system for
the ICC of flash FPLO3 50. ........._.._.. ...............149........_....

5-17: Measured channel base current and induced currents on the runway
counterpoise (290 m from the lightning) and the vertical wire (300 m from
the lightning) for stroke 0503-2, initiated from the mobile launcher. ................... .151

5-18: Measured channel base current and induced currents on the runway
counterpoise (50 m from the lightning) and the vertical wire (100 m from the
lightning) for stroke 0517-2, initiated from the tower launcher. ................... .........151

5-19: Counterpoise and vertical wire currents during natural lightning flash
MSE0504 striking ground approximately 300 m from the vertical wire. The
spikes in the vertical wire current occurring before the return stroke initiation
at t = are labeled. ............. ...............152....

5-20: Spikes 1 through 5 in the vertical wire current overlaid with unfiltered and
filtered dE/dt records measured at Stations 1, 4, 8, and 9. All data are
normalized and the normalization factors are given in the legends. ......................154

6-1: Stroke FPLO315-1, ground current consistency check..........._.._.._ ......._.._.. ....162

6-2: Stroke FPLO315-1, ground current consistency check..........._.._.._ ......._.._.. ....162










6-3: Stroke FPLO315-1, phase A-to-neutral current consistency check. ............._._. ....164

6-4: Stroke FPLO315-1, phase A-to-neutral current consistency check. ............._._. ....165

6-5: Stroke FPLO305-2, "high"/"low" lightning current comparison. ................... ........167

6-6: Stroke FPLO312-8, "high"/"low" lightning current comparison. ................... ........167

6-7: Stroke FPLO336-7, "high"/"low" lightning current comparison. ................... ........168

6-8: Stroke FPLO403-2, (1) "high" and (2) "low" channel base currents, (3) sum
of poles 6 through 10 ground currents and poles 6 and 10 overhead ground
wire currents, and (4) pole 8 ground current ................. ............................173

6-9: Spikes during the FPL and test-house experiments. The spikes in the pole 8
ground current and the point A current were isolated by subtracting the low-
frequency components from the total data. ............. ...............174....

6-10: Histogram of return stroke current peaks. An adjustment factor of 0.75 has
been applied to the current peaks from the 2000 experiment. ............. ................178

6-11:. Histograms of return stroke current 10-90% risetimes ................. ............... .... 179

6-12: Histogram of return stroke current half-peak widths. ................ .....................180

6-13: Histogram of return stroke charge transfers within 1 ms. An adjustment factor
of 0.75 has been applied to the charge from the 2000 experiment. ................... .....180

6-14: Current peaks as a function of charge transfer within 1 ms for rocket-
triggered lightning return strokes (143 individual values and regression line
are shown). The regression power equation and R2 value are given. The
regression line for 89 negative first return strokes in natural lightning found
by Berger is also displayed. The shaded area represents an envelope that
encompasses all Berger data points (only the outside values that confine the
shaded area are shown). ............. ...............189....

6-15: Struck-phase arrester currents with the transient mode (dark shaded area) and
steady-state mode (light shaded area) indicated ................. .........................199

6-16: Stroke FPLO312-5, the individual currents flowing from phase A to neutral
divided by the sum of all phase A-to-neutral currents on a 100 Cls time scale.
The pole 15 terminator current was not measured and was assumed to be
equal to the pole 1 terminator current. ............. ...............203....

6-17: Stroke FPLO312-5, the individual currents flowing to ground divided by the
sum of all ground currents on a 100 Cls time scale. The low-frequency, low-
current ground resistance of each of the pole grounds is given in the










parentheses. The percentages of the individual currents at 100 Cls are
displayed on the right side ................. ...............204........... ...

6-18: Stroke FPLO312-5, the individual charges flowing from phase A to neutral
divided by the sum of all phase A-to-neutral charges, on a 2 ms time scale.
The pole 15 terminator charge was not measured and was assumed to be
equal to the pole 1 terminator charge. ............. ...............204....

6-19: Stroke FPLO312-5, the individual charges flowing to ground divided by the
sum of all ground charges, on a 2 ms time scale. The dc ground resistance of
each of the pole grounds is given in the parentheses. The percentages of the
individual charges at 2 ms are displayed on the right side .................. ...............205

6-20: Stroke FPLOO36-1, Unfiltered and filtered channel base currents and its
piecewise-linear approximation. The piecewise linear approximation of the
filtered current was used as input to the model. .................. ...............21

6-21: Stroke FPLOO36-1, comparison of the overall phase C-to-neutral current
divisions (left) and ground current divisions (right) of measured currents
(top), model 1 currents (center), and model 2 currents (bottom) on a 500 Cls
time scale ................. ...............215................

6-22: Stroke FPLOO36-1, individual comparison of all successfully measured phase
C-to-neutral currents with model 1- and model 2-predicted results on a 50 Cls
tim e scale. ................. ...............216......... .....

6-23: Stroke FPLOO36-1, individual comparison of all successfully measured phase
C-to-neutral currents with model 1- and model 2-predicted results on a 500
Cls tim e scale. .............. ...............217....

6-24: Stroke FPLOO36-1, individual comparison of all measured ground currents
with model 1 and model 2 predicted results on a 50 Cls time scale. .......................218

6-25: Stroke FPLOO36-1, individual comparison of all measured ground currents
with model 1- and model 2-predicted results on a 500 Cls time scale. ...................219

6-26: Stroke FPLO312-5, unfiltered and filtered channel base currents and its
piecewise-linear approximation. The piecewise linear approximation of the
filtered current was used as input to the model. .................. ...............21

6-27: Stroke FPLO315-1, measured channel base current and its piecewise-linear
approximation. ...._.._................. ........_.._.........22

6-28: Stroke FPLO3 12-5, comparison of the overall measured and modeled phase
A-to-neutral current divisions at (left) poles 1, 2, and 6 and (right) poles 10
and 14 on a 100 Cls time scale. ............. ...............223....


XV111










6-29: Stroke FPLO3 15-1, comparison of the overall measured and modeled phase
A-to-neutral current divisions at (left) poles 1, 2, and 6 and (right) poles 10
and 14 on a 100 Cls time scale. ............. ...............224....

6-30: Stroke FPLO3 12-5, comparison of the overall measured and modeled ground
current divisions at (left) poles 1, 2, and 6 and (right) poles 10, 14, and 15 on
a 100 Cls time scale. ............. ...............225....

6-3 1: Stroke FPLO3 15-1, comparison of the overall measured and modeled ground
current divisions at (left) poles 1, 2, and 6 and (right) poles 10, 14, and 15 on
a 100 Cls time scale. ............. ...............226....

6-32: Stroke FPLO312-5, individual comparison of all measured phase A-to-neutral
currents with model-predicted results on a 50 Cls time scale............... .................22

6-33: Stroke FPLO315-1, individual comparison of all measured phase A-to-neutral
currents with model-predicted results on a 50 Cls time scale............... .................22

6-34: Stroke FPLO312-5, individual comparison of all measured phase A-to-neutral
currents with model-predicted results on a 500 Cls time scale. Note that the
pole 15 terminator current was not measured. ............. ...............229....

6-35: Stroke FPLO315-1, individual comparison of all measured phase A-to-neutral
currents with model-predicted results on an 800 Cls time scale. Note that the
pole 15 terminator current was not measured. ............. ...............230....

6-36: Stroke FPLO312-5, individual comparison of all measured ground currents
with model-predicted results on a 50 Cls time scale. Note that the pole 15
terminator current was not measured. .........__..... .._.._ ............... 231....

6-37: Stroke FPLO315-1, individual comparison of all measured ground currents
with model-predicted results on a 50 Cls time scale. Note that the pole 15
terminator current was not measured. ............. ...............232....

6-38: Stroke FPLO312-5, individual comparison of all measured ground currents
with model-predicted results on a 500 Cls time scale. ............. ......................3

6-40: Idealized distribution line and its circuit representation............... ..............3

6-41: Currents through the close arrester (iR) and the far arrester (iRL) calculated
using the simple model. The yellow-shaded area enclosed by the two currents
represents the charge transfer AQ through the close arrester due to the
impulsive current. ........._._ .. ...._.. ...............239....

6-42: Currents through the close and far arresters and AQ calculated with the
simple model and the EMTP model for 3, 4, and 5 span lengths between the
two arrester stations. The time constants used in the simple model are given.......240










6-43: EMTP-calculated absorbed energy in one of the two closest arresters during a
typical natural lightning first return stroke current inj ected into the phase
conductor at midspan. The vertical line contained (a) 4, (b) 8, or (c) 16
arrester stations. The transient mode (dark shaded area) and steady-state
mode (light shaded area) determined from the arrester currents for case (a)
are indicated. ............. ...............247....

6-44: Comparison of measured data with model-predicted results for stroke
FPLO336-6................. .............25

6-45: Peak value of current inj ected into the line through the pole 15 grounding as a
function of peak value of lightning current inj ected into ground 11 m from
pole 15. The linear regression equations and R2 ValUeS are given. ........................254

6-46: Charge inj ected into the line through the pole 15 grounding as a function of
lightning charge injected into ground 11 m from pole 15. The linear
regression equation and R2 value are given. The integration time used to
obtain the charge transfers from the return stroke currents and initial
continuous currents was 1 ms and 10 ms, respectively ................. ................ ...254

6-47: Magnitudes of counterpoise current peaks vs. channel base current peaks for
the mobile launcher experiment. The linear regression equation and R2 ValUe
are given. Distance from the lightning to the north-west corner of the
counterpoise is 290 m............... ...............261...

6-48: Magnitudes of counterpoise current peaks vs. channel base current peaks for
the tower launcher experiment. The linear regression equation and R2 ValUe
are given. Distance from the lightning to the north-west corner of the
counterpoise is 50 m............... ...............261...

B-1: Stroke FPL0011-1, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................9

B-2: Stroke FPL0011-2, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................9

B-3: Stroke FPL00 11-3, phase C to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................9

B-4: Stroke FPL0011-4, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................9










B-5: Stroke FPL0011-5, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................9

B-6: Stroke FPL0014-1, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-7: Stroke FPL0014-2, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-8: Stroke FPL0014-3, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-9: Stroke FPL0018-1, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-10: Stroke FPL0018-2, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-11: Stroke FPL0018-3, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-12: Stroke FPL0018-4, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-13: Stroke FPL0018-5, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-14: Stroke FPL0018-6, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-15: Stroke FPLOO32-1, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-16: Stroke FPLOO32-2, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0










B-17: Stroke FPLOO32-3, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-18: Stroke FPLOO32-4, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-19: Stroke FPLOO32-5, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-20: Stroke FPLOO32-6, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-21: Stroke FPLOO32-7, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-22: Stroke FPLOO33-1, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-23: Stroke FPLOO34-1, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-24: Stroke FPLOO34-2, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-25: Stroke FPLOO34-3, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................0

B-26: Stroke FPLOO34-4, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-27: Stroke FPLOO34-5, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-28: Stroke FPLOO36-1, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ................ ................ ...._.311










B-29: Stroke FPLOO36-2, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ................ ................ ...._.311

B-30: Stroke FPLOO36-3, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-31: Stroke FPLOO36-4, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-32: Stroke FPLOO36-5, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-33: Stroke FPLOO37-1, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-34: Stroke FPLOO37-2, phase C-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-35: Stroke FPLO208-1, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-36: Stroke FPLO210-1, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-37: Stroke FPLO218-1, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-38: Stroke FPLO219-2, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-39: Stroke FPLO219-3, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-40: Stroke FPLO220-1, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1


XX111










B-41: Stroke FPLO220-2, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-42: Stroke FPLO220-3, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-43: Stroke FPLO220-4, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-44: Stroke FPLO220-5, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................1

B-45: Stroke FPLO221-2, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-46: Stroke FPLO221-3, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-47: Stroke FPLO221-4, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-48: Stroke FPLO221-5, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-49: Stroke FPLO221-6, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-50: Stroke FPLO226-1, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-51: Stroke FPLO226-2, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-52: Stroke FPLO226-3, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2


XX1V










B-53: Stroke FPLO226-4, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-54: Stroke FPLO226-5, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-55: Stroke FPLO226-6, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right), 1 ms time scale.........................325

B-56: Stroke FPLO228-1, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-57: Stroke FPLO228-2, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-58: Stroke FPLO228-3, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-59: Stroke FPLO228-4, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-60: Stroke FPLO228-5, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-61: Stroke FPLO228-6, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-62: Stroke FPLO229-1, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-63: Stroke FPLO229-2, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2

B-64: Stroke FPLO229-3, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................2










B-65: Stroke FPLO229-4, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................3

B-66: Stroke FPLO229-5, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................3

B-67: Stroke FPLO229-6, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................3

B-68: Stroke FPLO229-7, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................3

B-69: Stroke FPLO229-8, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................3

B-70: Stroke FPLO229-9, phase A-to-neutral currents (top, left), ground currents
(top, right), and the respective sum of the individual currents overlaid with
the channel base current (bottom, left and right). ............. ......................3

B-71: Data matrix for stroke FPLO3 01-1 (direct strike), 100 Cls time windows ..............33 4

B-72: Data matrix for stroke FPLO3 01-2 (direct strike), 100 Cls time windows ..............33 5

B-73: Data matrix for stroke FPLO3 01-3 (direct strike), 100 Cls time windows ..............33 6

B-74: Data matrix for stroke FPLO3 05 -1 (direct strike), 100 Cls time windows ..............33 7

B-75: Data matrix for stroke FPLO3 05 -2 (direct strike), 100 Cls time windows ..............33 8

B-76: Data matrix for stroke FPLO3 05 -3 (direct strike), 100 Cls time windows ..............33 9

B-77: Data matrix for stroke FPLO305-4 (direct strike), 100 Cls time windows............_..340

B-78: Data matrix for stroke FPLO312-1 (direct strike), 100 Cls time windows............_.341

B-79: Data matrix for stroke FPLO312-2 (direct strike), 100 Cls time windows ..............3 42

B-80: Data matrix for stroke FPLO312-3 (direct strike), 100 Cls time windows. .............343

B-81: Data matrix for stroke FPLO312-4 (direct strike), 100 Cls time windows............_.3 44

B-82: Data matrix for stroke FPLO312-5 (direct strike), 100 Cls time windows............_..345


XXV1










B-83: Data matrix for stroke FPLO312-6 (direct strike), 100 Cls time windows............_..346

B-84: Data matrix for stroke FPLO312-7 (direct strike), 100 Cls time windows............_..347

B-85: Data matrix for stroke FPLO312-8 (direct strike), 100 Cls time windows............_..348

B-86: Data matrix for stroke FPLO312-9 (direct strike), 100 Cls time windows............_..349

B-87: Data matrix for stroke FPLO312-10 (direct strike), 100 Cls time windows............ 350

B-88: Data matrix for stroke FPLO315-1 (direct strike), 100 Cls time windows............_.351

B-89: Data matrix for stroke FPLO315-2 (direct strike), 100 Cls time windows ..............3 52

B-90: Data matrix for stroke FPLO317-1 (direct strike), 100 Cls time windows ..............353

B-91: Data for wireburn FPLO402, 600 Cls time scale. .................. .................5

B-92: Data for stroke FPLO403-1, 2 ms and 100 Cls time scales. ............. ...................356

B-93: Data for stroke FPLO403-2, 2 ms and 100 Cls time scales. ............. ...................357

B-94: Measured channel base current and induced currents on the runway
counterpoise (290 m from the lightning) and the vertical wire (300 m from
the lightning) for flash 0501, stroke 1 during the mobile launcher experiment.....3 58

B-95: Measured channel base current and induced currents on the runway
counterpoise (290 m from the lightning) and the vertical wire (300 m from
the lightning) for flash 0503, stroke 1 during the mobile launcher experiment.....3 58

B-96: Measured channel base current and induced currents on the runway
counterpoise (290 m from the lightning) and the vertical wire (300 m from
the lightning) for flash 0503, stroke 2 during the mobile launcher experiment.....3 59

B-97: Measured channel base current and induced currents on the runway
counterpoise (290 m from the lightning) and the vertical wire (300 m from
the lightning) for flash 0503, stroke 3 during the mobile launcher experiment.....3 59

B-98: Measured channel base current and induced currents on the runway
counterpoise (290 m from the lightning) and the vertical wire (300 m from
the lightning) for flash 0503, stroke 4 during the mobile launcher experiment.....3 60

B-99: Measured channel base current and induced current on the runway
counterpoise (50 m from the lightning) for flash 0510, stroke 1 during the
tower launcher experiment. .............. ...............360....


XXV11










B-100: Measured channel base current and induced current on the runway
counterpoise (50 m from the lightning) for flash 0512, stroke 1 during the
tower launcher experiment. .............. ...............361....

B-101: Measured channel base current and induced current on the runway
counterpoise (50 m from the lightning) for flash 0512, stroke 2 during the
tower launcher experiment. .............. ...............361....

B-102: Measured channel base current and induced currents on the runway
counterpoise (50 m from the lightning) and the vertical wire (100 m from the
lightning) for flash 0517, stroke 1 during the tower launcher experiment. ............3 62

B-103: Measured channel base current and induced currents on the runway
counterpoise (50 m from the lightning) and the vertical wire (100 m from the
lightning) for flash 0517, stroke 2 during the tower launcher experiment. ............3 62

B-104: Measured channel base current and induced currents on the runway
counterpoise (50 m from the lightning) and the vertical wire (100 m from the
lightning) for flash 0520, stroke 1 during the tower launcher experiment. ............3 63

B-105: Measured channel base current and induced currents on the runway
counterpoise (50 m from the lightning) and the vertical wire (100 m from the
lightning) for flash 0521, stroke 1 during the tower launcher experiment. ............3 63

B-106: Currents during natural lightning flash MSE0504, stroke 1................ ...............364

B-107: Currents during natural lightning flash MSE0504, stroke 2. ............... ...............364


XXV111
















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

DIRECT AND NEARBY LIGHTNING STRIKE INTERACTION
WITH TEST POWER DISTRIBUTION LINES

By

Jens Daniel Schoene

May 2007

Chair: M.A. Uman
Cochair: V.A. Rakov
Maj or Department: Electrical and Computer Engineering

The interaction of direct and nearby rocket-triggered lightning with two

unenergized three-phase power distribution lines of about 800 m length was studied at the

International Center for Lightning Research and Testing in Florida. A horizontally-

configured line was tested in 1999 and 2000, a vertically-configured line in 2001, 2002,

and 2003, and a vertically-configured line with overhead ground wire in 2004. All lines

were equipped with arresters and, additionally, in 2003, the vertical line had a pole-

mounted transformer. During the 2000, 2001, and 2002 direct strike experiments,

arresters frequently failed, but there was no arrester failures either during the 2003 direct

strike experiment when the transformer was on the line or during the 2004 direct strike

experiment when the lightning current was injected into the overhead ground wire. All

line configurations except the one tested in 2004 commonly exhibited flashovers. The

division of return stroke currents for the vertically-configured line was initially similar to

the division on the horizontally-configured line (that is, the arresters closest to the strike


XX1X










point conducted the bulk of the impulsive current). After some tens of microseconds the

currents in all arresters on the vertically-configured line equalized, while the close

arrester currents on the horizontally-configured line still conducted significantly more

current than the remote arresters. The lightning current division for direct strikes to a

phase conductor is successfully modeled with the Electromagnetic Transient Program

(EMTP) for the vertically-configured line and, if the residual voltage of the close

arresters is reduced by 20%, successfully modeled for the horizontally-configured line.

Currents on the vertically-configured line induced by nearby lightning strikes were

measured and compared to results calculated using the LIOV-EMTP96 code. It was

found that during a lightning strike 11 m from a grounded line pole, a significant fraction

of the lightning currents entered the neutral conductor through the line grounding.

Additional topics include an investigation of the response of metallic structures (a buried

counterpoise and a vertical wire of 7 m height) to nearby rocket-triggered and natural

lightning strikes and the characterization of rocket-triggered lightning return stroke

currents.


xxx















CHAPTER 1
INTTRODUCTION

Lightning commonly strikes to and near power distribution lines causing flashovers

and/or damage to equipment connected to the line. Lightning-caused service interruptions

are particularly frequent in areas exhibiting a high ground flash density, such as the

southeastern United States.

The purpose of this dissertation is to investigate the behavior of distribution lines

subj ected to currents from rocket-triggered lightning that were (1) directly inj ected into

one of the line conductors and (2) directed to ground near the line causing (a) induced

currents on the line and (b) some fraction of the lightning current to enter the line through

the line groundings. The data presented and discussed are needed in order to develop

realistic models for designing lightning protection for distribution lines.

During six summers (1999 through 2004) the Lightning Research Laboratory of the

University of Florida has been studying, under Florida Power and Light (FPL) support,

the interaction of triggered lightning with three distribution-line framing configurations

that were constructed by FPL at the International Center for Lightning Research and

Testing (ICLRT) at Camp Blanding, Florida. A horizontally-configured line was the

primary subj ect of the 1999 and 2000 experiments, a modified vertically-configured line

was studied during summers 2001, 2002, and 2003, and a modified vertically-configured

line with an overhead ground wire was studied during summer 2004. Triggered lightning

current was directly inj ected into distribution lines during all years of the six year study.

The effects of nearby triggered lightning, at distances from the line ranging from 7 to









100 m, were examined in 2002 and 2003 for the case of the modified vertical framing

configuration.

Additionally, results from a proj ect funded by Lawrence Livermore National

Laboratories in 2005 are presented in this dissertation. This proj ect examined the

response of grounded structures to nearby strikes. Nearby rocket-triggered or natural

lightning was the source of induced currents in a buried test-runway counterpoise and in a

grounded vertical wire of 7 m height. The measured currents in these two systems are

presented in this dissertation since they have implications regarding the issue of induced

currents on power lines and power line grounding systems.

Some of the data obtained, with the help of many participants, during the 7 years of

experiments has been previously discussed in

* 2 dissertations (Mata, 2000; Mata, 2003),

* 3 journal publications (Mata et al., 2003; Schoene et al., 2006a; Schoene et al.,
2006b),

* 6 conference proceedings (Rakov, 1999a; Chrzan and Rakov, 2000; C.T. Mata et
al. 2002; Rakov, 2003; Rakov et al., 2003a; Rakov et al., 2003b), and

* 7 technical reports (Mata et al., 2000b; Mata et al., 2001; A.G. Mata et al., 2002;
Schoene et al., 2003a; Schoene et al., 2004a; Schoene et al., 2004b; Hanley et al.,
2006).

The following sections give a review by year of the 7 years of experiments

including a list of the participants and their contributions/responsibilities, a list of

previous publications discussing the experimental data, and results presented in previous

publications. The original contributions of the author in obtaining the new experimental

data and in analyzing both previously-recorded data and new data are delineated. A

description of the 9 separate experiments listed below that took place during the 7 years,

1999-2005 follows:









1. 1999 FPL Experiment, Direct Lightning Strike Interaction with a Horizontally-
configured Distribution Line

2. 2000 FPL Experiment, Direct Lightning Strike Interaction with a Horizontally-
configured Distribution Line

3. 2001 FPL Experiment, Direct Lightning Strike Interaction with a Vertically-
configured Distribution Line

4. 2002 FPL Experiment, Direct Lightning Strike Interaction with a Vertically-
configured Distribution Line

5. 2002 FPL Experiment, Nearby Lightning Strike Interaction with a Vertically-
configured Distribution Line

6. 2003 FPL Experiment, Direct Lightning Strike Interaction with a Vertically-
configured Distribution Line

7. 2003 FPL Experiment, Nearby Lightning Strike Interaction with a Vertically-
configured Distribution Line

8. 2004 FPL Experiment, Direct Lightning Strike Interaction with a Vertically-
configured Distribution Line with Overhead Ground Wire

9. 2005 Lawrence Livermore Experiment, Induced Currents


1.1 1999 FPL Experiment, Direct Lightning Strike Interaction with a Horizontally-
configured Distribution Line

The 245 m long horizontally-configured three-phase test distribution line studied in

1999 consisted of six wooden poles and two arrester stations. For part of the experiment

the phase conductors were terminated in their characteristic impedances to simulate to

some extent an infinitely long line. The experience gained during the 1999 FPL

experiment has laid the groundwork for the larger-scale 2000 FPL experiment. However,

the relatively short length of the line (only 6 poles), the small number of arrester stations

(only 2 arrester stations), and problems with the termination resistors (Section 3.5.1)

made the 1999 experiment of limited use for the purpose of investigating the effects of

direct lightning strikes to power lines. A description of the 1999 experiment, previously










published in Mata et al. (1999) and Mata (2000), is included in this dissertation for

completeness. A statistical analysis of the 1999 return stroke current data was performed

for this dissertation since it had not been done in previous publications.

1999 Experiment description found in this dissertation:
* Experimental facility (Chapter 3)
* Experimental configuration and references to measurement settings (Section 4.2)
* Triggering results (Section 4.9)


New results from 1999 found in this dissertation:
* Statistical analysis of return stroke currents (Section 6.2, Appendix C)

1999 Results presented in previous publications:
PhD Dissertation: Mata (2000); Technical Report: Mata et al. (1999)
* Experiment description
* Presentation of all data
* Consistency check of return stroke currents and currents to ground: The return
stroke currents were found to be 20% to 40% smaller than the sum of ground
currents. "This might be due to the fact that in 1999 no end-to-end calibration
factors were experimentally obtained for current sensors,..." (Mata, 2000, p.91)
* Flashover analysis: At least 4 of the 7 flashes that contained return strokes
produced flashovers
* Disconnector operation analysis: One arrester and several termination resistors
were destroyed. "Arresters did no fail as frequently as they did in the 2000
experiments due to several factors including the presence of inductors in parallel
with terminating resistors in some of the 1999 tests, relatively small return strokes,
arcing between conductors (including that facilitated by the presence of nylon cords
between the line conductors and residual triggering wires), and failures of
terminating resistors and voltage dividers" (Mata, 2000, p.90). No arrester failure
was observed for the last 3 of the 7 flashes when terminating resistors in parallel
with inductors were installed on the line.

Participants and their responsibilities during the 1999 Experiment:
* Principal investigators (proposal, experimental design, data analysis, general
advice): M.A. Uman and V.A. Rakov
* Experimental design and implementation, data analysis, and report writing:
C.T. Mata
* ICLRT manager: K.J. Rambo
* General assistance: M.V. Stapleton, A.G. Mata









1.2 2000 FPL Experiment, Direct Lightning Strike Interaction with a Horizontally-
configured Distribution Line

The 8 56 m long hori zontal ly -configure d three-phase te st di stributi on li ne, a l onger

version of the line studied in 1999, consisted of 18 wooden poles and six arrester stations.

The phase conductors were terminated in their characteristic impedances to simulate to

some extent an infinitely long line. The description of the 2000 experiment, previously

published in Mata (2000), Mata et al. (2000b), and Mata et al. (2003), is also included in

this dissertation for completeness. The new results gained from the 2000 data presented

in this dissertation as well as results from previous publications are summarized below.

2000 Experiment description found in this dissertation:
* Experimental facility (Chapter 3)
* Experimental configuration and references to measurement settings (Section 4.3)
* Triggering results (Section 4.9)

New results from 2000 found in this dissertation:
* Comparison of struck-phase-to-neutral current divisions during the 2000
experiment with the divisions during the 2002 and 2003 experiments (Section 6.4):
The division of return stroke currents among multiple arresters for the line tested in
2000 was initially similar to the division on the line tested in 2002 and 2003.
However, the time during which the return stroke current flowed primarily through
the closest arresters to the neutral conductor was significantly longer on the line
tested in 2000. Possible explanations for the different current division are given.
* Additional modeling results (Section 6.5): The current division on the line during
stroke FPLOO36-1 is modeled using EMTP96. The arrester current division
modeled here does not match the measured division if the arresters in the model are
represented by the manufacturer-provided VI-characteristics. The arrester currents
modeled here match the measured arrester currents well if the VI-characteristics of
the two closest arresters are modified (the residual voltage is reduced by
approximately 20%). The current division during stroke FPLOO36-1 has been
previously modeled in Mata (2000) and A.G. Mata (2002) using ATP. Previously
published modeling results of the closest arrester current exhibited ringing if the
lightning in the model is represented as an ideal current source (that is, the
lightning channel's characteristic impedance is infinity). Ringing is neither present
in the measured arrester currents nor in the modeling results presented here.
* Additional data presentation (Section 6.4.1, Appendix B): Additional data from
eight strokes in two flashes (FPLOO32 and FPLOO33) showing currents measured
simultaneously through both closest struck-phase arresters (at pole 8 and pole 11)
are presented and discussed. The existence of currents simultaneously measured at









both struck-phase arresters has been overlooked in previous publications and only
currents through one of the two closest struck-phase arresters have been presented.
The newly discovered data support the common assumption in the data analysis
performed previously that the two closest struck-phase arrester currents are equal
due to symmetry.
* Complete consistency check of return stroke currents ("high" and "low"
measurements), sum of ground currents, and sum of struck-phase-to-neutral
currents (Section 6.1, Appendix B): The return stroke currents presented in Mata
(2000), A.G. Mata et al. (2002), and Mata et al. (2003) have likely been
overestimated by 25%.
* Revised statistical analysis of return stroke currents (Section 6.2, Appendix C): The
overestimated return stroke current peaks and charge transfers (see previous bullet)
have been adjusted. The risetimes determined from currents that exhibited ringing
during the rising edge have been excluded from the statistics.
* Revised arrester damage analysis (Section 6.3.1): The average number of
disconnector operation per triggering day was determined as opposed to arrester
damage per flash in Mata et al. (2003), which the author of this dissertation
believes to be neither a realistic measure (arresters could only be examined for
damage at the end of the triggering day and not after each flash) nor an appropriate
measure (arrester damage was determined by disconnector operation even though
disconnectors could have operated on healthy arresters).
* Revised and expanded flashover analysis (Section 6.3.2, Appendix D): Flashovers
during the 2000 experiment and their approximate location have been determined
based on the charge transfers in non-struck-phase-to-neutral connections and non-
struck phases. Using this approach 5 out of 8 flashes (8 out of 34 strokes) were
determined to have caused flashovers on the line as opposed to 6 out of 8 flashes
determined to have caused flashovers in Mata (2000). The exact number of strokes
causing flashovers was not determined previously. A comparison of flashovers
during the 2000 experiment with flashovers during the 2001 through 2003
experiments shows that flashovers on the vertical line were much more common.
Possible explanations for the different flashover behavior are given.

2000 Results presented in previous publications:
PhD Dissertation: Mata (2000); Technical Reports: Mata et al. (2000b) and A.G. Mata et
al. (2002); Journal Publication: Mata et al. (2003)
* Experiment description
* Presentation of selected data
* Statistical analysis of return stroke currents
* Consistency check of the return stroke currents and currents to ground. The return
stroke charge (the integrated current) were found to be 25% to 30% larger than the
sum of ground currents. Corona on line conductors and other elements of the
system near the strike point were considered a possible cause of the missing charge
(Mata, 2000). Flashovers that would account for the missing charge were not ruled
out (Mata et al., 2003).
* Arrester damage analysis. "Six of the eight triggered lightning flashes caused
damage to one of the two closest arresters" (Mata et al., 2003, p.1i). The more










frequent arrester damage in the 2000 experiment compared to the 1999 experiment
was attributed to the absence of failed line terminating resistors and failed voltage
dividers in the 2000 experiment, the presence of inductors in parallel with
terminating resistors in some of the 1999 tests, the relatively small return strokes in
the 1999 experiment, and arcing between phase conductors (some facilitated by the
presence of nylon cords between line conductors and residual triggering wires) in
the 1999 experiment. (Mata, 2000) The arrester damage analysis is revised in this
dissertation.
* Flashover analysis: At least 6 of the 8 flashes that contained return strokes
produced flashovers (Mata, 2000). The flashover analysis is revised in this
dissertation.
* Struck-phase-to-neutral current divisions: "about 40% of the return stroke peak
current and about 25% or more of the return stroke charge transferred in the first
millisecond passed to the neutral conductor through each of the two closest
arresters" (Mata et al., 2003, p. 1). Based on this result Mata et al. (2003) estimated
that over half of all first strokes in natural lightning would result in an arrester
failure on the 2000 test distribution line.
* Ground current division: "An important finding in this study is that the distribution
of peak currents to ground is more or less symmetrical with respect to the strike
point and appears to be strongly dependent on the distance from the strike point,
regardless of the low-frequency, low-current resistances, whereas the charges
transferred to ground are distributed according to the low-frequency, low-current
grounding resistances" (Mata, 2000, p. 136).
* Modeling of measured currents on the line: According to Mata (2000) model-
predicted currents through the two arresters closest to the strike point resemble
measured current waveforms if oscillations in the modeled currents are eliminated
by the implementation of a crude lumped "corona" model (a shunt resistor in
parallel with a capacitor at the strike point) and/or the implementation of the
lightning channel's characteristic impedance in the model. However, no modeling
results that support this statement are presented in Mata (2000). The modeling
results (including the closest arrester currents) presented on a 100 Cls time scale
assuming different values for the lightning channel impedance effects presented in
A.G. Mata et al. (2002) match the measured currents reasonably well. Modeled and
measured results after 100 Cls are not compared.

Participants and their responsibilities during the 2000 Experiment:
* Principal investigators (proposal, experimental design, data analysis, general
advice): M.A. Uman and V.A. Rakov
* ICLRT manager: K.J. Rambo
* Experimental design and implementation, data analysis, and report writing:
C.T. Mata
* General assistance: M.V. Stapleton, A.G. Mata, G.H. Schnetzer, R. Sutil, A.
Guarisma, A. Mata, and G. Bronsted









1.3 2001 FPL Experiment, Direct Lightning Strike Interaction with a Vertically-
configured Distribution Line

The 812 m long vertically-configured three-phase test distribution line studied in

2001 consisted of 15 wooden poles and 4 arrester stations. The phase conductors were

terminated in their characteristic impedances to simulate to some extent an infinitely long

line. Lightning currents were inj ected into the phase A conductor at pole 8 and between

poles 7 and 8 at midspan. Similar to the 1999 FPL experiment, which laid the

groundwork for the 2000 FPL experiment, the experience gained during the 2001 FPL

experiment was invaluable for the successful completion of the experiments on the same

line in 2002 and 2003. Also similar to the 1999 experiment, the data collected in 2001

were not useable for the investigation of the current division on the line due to numerous

instrumentation problems. The description of 2001 experiment, previously published in

Mata et al. (2001) and Mata (2003), is included in this dissertation for completeness. The

statistical analysis of the 2001 return stroke current data included in this dissertation has

not been performed in previous publications.

2001 Experiment description found in this dissertation:
* Experimental facility (Chapter 3)
* Experimental configuration and references to measurement settings (Section 4.4)
* Triggering results (Section 4.9)

New results from 2001 found in this dissertation:
* Statistical analysis of return stroke currents (Section 6.2, Appendix C).
* Disconnector operation analysis (Section 6.3.1): The average number of operated
disconnectors per triggering day was determined based on the information provided
in Mata (2003).
* Revised and expanded flashover analysis (Section 6.3.2, Appendix D): Flashovers
during the 2001 experiment and their approximate location have been determined
based on the charge transfers in non-struck-phase-to-neutral connections and non-
struck phases. Using this approach flashovers were determined in all 4 flashes that
contained return strokes and were examined for flashovers (12 out of 14 strokes
were determined to have caused flashovers) as opposed to only 2 flashes









determined to have caused flashovers in Mata (2003). The exact number of strokes
causing flashovers was not determined previously.

2001 Results presented in previous publications:
Master Thesis: Mata (2003); Technical Report: Mata et al. (2001)
* Experiment description
* Presentation of selected data
* Arrester damage analysis: "Of the total of nine flashes triggered during the 2001
experiments, only flashes FPL0101 and FPL0 102 (both wireburns) showed no
evidence of failed arresters during the direct strike tests" (Mata, 2003, p. 84).
* Flashover analysis: "During summer 2001, no evidence of trailing wires on the line
were found after any event and possible flashovers (probably between phase A and
phase B conductors at pole 8) were observed during two flashes (FPL0107 and
FPL0 108)." (Mata, 2003, p. 84). The number of flashovers during the 2001
experiment has been revised in this dissertation (see "New results from 2001 found
in this dissertation" above).

Participants and their responsibilities during the 2001 Experiment:
* Principal investigators (proposal, experimental design, data analysis, general
advice): M.A. Uman and V.A. Rakov
* ICLRT manager: K.J. Rambo
* Experiment implementation: A.G. Mata, M.V. Stapleton, J.E. Jerauld, and D.M.
Jordan
* Data analysis and report writing: A.G. Mata
* General assistance: C.T. Mata, G.H. Schnetzer, and A. Guarisma

1.4 2002 FPL Experiment, Direct Lightning Strike Interaction with a Vertically-
configured Distribution Line

The maj or differences between the line tested in 2002 and the line tested in 2001

were that (1) two arresters in parallel were installed on the struck phase (2001: only

single arresters) and (2) the initial continuous current was diverted from the line.

Lightning currents were inj ected into the phase A conductor between poles 7 and 8 at

midspan. The description of the 2002 experiment, previously published in A.G. Mata et

al. (2002) and Mata (2003), is included in this dissertation for completeness. The author

of this dissertation was actively involved in the 2002 experiment by checking the validity

of the acquired data after each triggering day. The statistical analysis of the 2002 return

stroke current data included in this dissertation has not been performed previously.









2002 Direct Strike Experiment description found in this dissertation:
* Experimental facility (Chapter 3)
* Experimental configuration and references to measurement settings (Section 4.5)
* Triggering results (Section 4.9)

New results from 2002 Direct Strike Experiment found in this dissertation:
* Statistical analysis of return stroke currents (Section 6.2, Appendix C)
* Disconnector operation analysis (Section 6.3.1): The average number of operated
disconnectors per triggering day was determined based on the information provided
in Mata (2003).
* Revised and expanded flashover analysis (Section 6.3.2, Appendix D): Flashovers
during the 2002 experiment and their approximate location have been determined
based on the charge transfers in non-struck-phase-to-neutral connections and non-
struck phases. Using this approach flashovers were determined in all 7 flashes that
contained return strokes and were examined for flashovers (39 out of 43 strokes
were determined to have caused flashovers) as opposed to only 2 flashes
determined to have caused flashovers in Mata (2003). The exact number of strokes
causing flashovers was not determined previously.

2002 Direct Strike Experiment results presented in previous publications:
Master Thesis: Mata (2003); Technical Report: A.G. Mata et al. (2002)
* Experiment description
* Presentation of selected data
* Flashover analysis: Flashovers were determined from video records and
photographs. Using this approach 6 out of the 9 examined flashes were determined
to have possibly caused flashovers (Mata, 2003). The number of flashovers during
the 2002 experiment has been revised in this dissertation (see "New results from
2002 Direct Strike Experiment found in this dissertation" above). "For flashes
FPLO208 to FPLO226 an instrumentation device (at pole 7) might possibly have
helped drain some current from phase A (struck phase) to phase B (closest phase to
the struck one), most likely via a flashover" (Mata, 2003, p. 89).

Responsibilities of the dissertation author during the 2002 Direct Strike Experiment:
* Assessment of validity of measured currents after each triggering day

Other participants and their responsibilities during the 2002 Direct Strike Experiment:
* Principal investigators (proposal, experimental design, data analysis, general
advice): V.A. Rakov and M.A. Uman
* ICLRT manager: K.J. Rambo
* Data analysis and report writing: A.G. Mata
* Experiment implementation: A.G. Mata, M.V. Stapleton, J.E. Jerauld, and D.M.
Jordan
* General assistance: C.T. Mata, G.H. Schnetzer, and A. Guarisma









1.5 2002 FPL Experiment, Nearby Lightning Strike Interaction with a Vertically-
configured Distribution Line

The 2002 nearby strike experiment was performed after the 2002 direct strike

experiment. The differences between the 2002 nearby strike and direct strike experiments

were (1) the attenuation settings of the current measurements on the line were adjusted to

the lower magnitudes expected for the induced currents, (2) the wire used for the direct

strike experiment to inj ect the lightning current into the line was disconnected from the

line, and (3) lightning was triggered from the mobile launcher located 30 m or 100 m

north of pole 7. The author of this dissertation was involved in the design of the

experiment, checking the validity of the data after each triggering day, and the

analysis/modeling of the experimental data. He is co-author in A.G. Mata et al. (2002)

and wrote the sections involving the presentation, modeling, and discussion of the

"nearby strike" data in A.G. Mata et al. (2002).

2002 Nearby Strike Experiment description found in this dissertation:
* Experimental facility (Chapter 3)
* Experimental configuration and measurement settings (Section 4.5, Appendix A)
* Triggering results (Section 4.9)

Results from 2002 Nearby Strike Experiment presented in previous publications:
Technical Report: A.G. Mata et al. (2002)
* Experiment description
* Presentation of selected data
* Modeling results: The measured and model-predicted ground and neutral currents
generally match reasonably well. The measured and model-predicted arrester and
phase currents show little or no resemblance. The former mismatch was possibly
due to an inaccurate VI-characteristic.

Responsibilities of the dissertation author during the 2002 Nearby Strike Experiment:
* Experiment design
* Analysis and modeling of experimental data and report writing
* Assessment of validity of measured currents after each triggering day

Other participants and their responsibilities during the 2002 Nearby Strike Experiment:










* Principal investigators (proposal, experimental design, data analysis, general
advice): M.A. Uman and V.A. Rakov
* ICLRT manager: K.J. Rambo
* Modeling of experimental data: M. Paolone, C.A. Nucci, and F. Rachidi
* Experiment implementation: A.G. Mata, M.V. Stapleton, J.E. Jerauld, and D.M.
Jordan
* General assistance: C.T. Mata, G.H. Schnetzer, and A. Guarisma

1.6 2003 FPL Experiment, Direct Lightning Strike Interaction with a Vertically-
configured Distribution Line

The main differences between the 2003 direct strike experiment and the 2002 direct

strike experiment are (1) only single arresters were used in 2003 and (2) a transformer

was connected to the struck phase at pole 2. The author of this dissertation was involved

in the design of the experiment, checking the validity of the data after each triggering

day, and the analysis/modeling of the experimental data. He is the main-author of

Schoene et al. (2003a).

2003 Direct Strike Experiment description found in this dissertation:
* Experimental facility (Chapter 3)
* Experimental configuration and measurement settings (Section 4.6, Appendix A)
* Triggering results (Section 4.9)

Results from 2003 Direct Strike Experiment presented in this dissertation and in Schoene
et al. (2003a):
* Statistical analysis of return stroke currents (Section 6.2, Appendix C)
* Data presentation: All measurements of the 2003 experiment are presented in
Schoene et al. (2003a). Selected measurements are presented in this dissertation
(Section 5.1, Appendix B).
* Struck-phase-to-neutral current divisions: "The primary path of the return stroke
current for the first 50 CLs or so is through the two arresters that are closest to the
inj section point of the return stroke current, that is, the arresters at poles 6 and 10,
and for the first 20 CLs or so from those arresters to the grounds of the closest
arrester poles. At the time of the peak value of the arrester currents (about 2 CLs after
the return stroke initiation) the two closest arresters pass about 80% of the total
inj ected current and after 100 CLs the current is evenly distributed among all
arresters." (Schoene et al., 2003a, p. 14). Additional analysis is included in this
dissertation (Section 6.4).
* Modeling results: In this dissertation the model-predicted current division during
strokes FPLO312-5 and FPLO315-1 are compared to measured results (Section 6.5).
C.T. Mata modeled the current division on the line during stroke FPLO312-3. The










author of this dissertation discussed C.T. Mata' s modeling results in Schoene et al.
(2003a). Model-predicted currents in the two closest struck-phase arresters matched
the measured currents well while currents in the other two struck-phase arresters
were poorly modeled. Ground currents for times before 20 Cls, or so, are poorly
modeled. Ground currents for times after 20 Cls, or so, are reasonably well modeled.
The discrepancies in the struck-phase-to-neutral currents are likely due to a
flashover on the line which the model does not take into account.
* Disconnector operation analysis (Section 6.3): No disconnector operated during the
2003 experiment. Possible reasons for the absence of disconnector operation are
discussed in this dissertation.
* Flashover analysis (Section 6.3): Flashovers during the 2003 experiment and their
approximate location have been determined based on the charge transfers in non-
struck-phase-to-neutral connections and non-struck phases. Using this approach
flashovers were determined in all 5 flashes that contained return strokes and were
examined for flashovers (22 out of 26 strokes were determined to have caused
flashovers). The percentage of strokes causing flashovers on the line tested in 2003
is similar to the flashover percentage determined on the line tested in 2002.

Responsibilities of the dissertation author during the 2003 Direct Strike Experiment:
* Experiment design
* Analysis of the experimental data
* Writing of a report
* Verification of the experimental data during the experiment

Other participants and their responsibilities during the 2003 Direct Strike Experiment:
* Principal investigators (proposal, experimental design, data analysis, general
advice): M.A. Uman and V.A. Rakov
* ICLRT manager: K.J. Rambo
* Experiment implementation: J.E. Jerauld, M.V. Stapleton, and D.M. Jordan
* Modeling of the experimental data: C.T. Mata
* General assistance: G.H. Schnetzer

1.7 2003 FPL Experiment, Nearby Lightning Strike Interaction with a Vertically-
configured Distribution Line

The 2003 nearby strike experiment was performed after the 2003 direct strike

experiment and continued the study of lightning-induced current conducted in 2002. The

differences between the 2003 nearby strike and direct strike experiments were (1) the

attenuation settings of the current measurements on the line were adjusted to the lower

magnitudes expected for the induced currents, (2) the wire used for the direct strike

experiment to inject the lightning current into the line and the transformer was









disconnected from the line, and (3) lightning was triggered from the mobile launcher

located 7 m or 15 m south of pole 4, or 11 m south-east of pole 15. The author of this

dissertation was involved in the design of the experiment, checking the validity of the

data after each triggering day, and the analysis/modeling of the experimental data. He is

the main-author of Schoene et al. (2003a) and co-author of Paolone et al. (2004b).

2003 Nearby Strike Experiment description found in this dissertation:
* Experimental facility (Chapter 3)
* Experimental configuration and measurement settings (Section 4.6, Appendix A)
* Triggering results (Section 4.9)

Results from 2003 Nearby Strike Experiment presented in this dissertation, Schoene et al.
(2003a), and Paolone et al. (2004b):
* Statistical analysis of return stroke currents (Section 6.2, Appendix C)
* Data presentation: All measurements of the 2003 experiment are presented in
Schoene et al. (2003a). Selected measurements are presented in this dissertation
(Section 5.3).
* Data analysis (Section 6.8): The issue of currents from nearby lightning strikes
entering the line through one of the line groundings is investigated.
* Modeling results: "The measured and modeled ground and neutral currents
generally match reasonably well. The measured and modeled arrester and phase
currents generally match reasonably well up to the first peak. The modeled arrester
and phase currents are typically larger than the measured currents for times after
the first peak. In general, the model works reasonably well for the first microsecond
after the beginning of the return stroke and not so well for later times" (Schoene et
al., 2003a, p. 38). The comparison of modeling results with selected data from the
2003 experiment is also included in this dissertation (Section 6.7) and in Paolone et
al. (2004b).

Responsibilities of the dissertation author during the 2003 Nearby Strike Experiment:
* Experiment design
* Analysis and modeling of the experimental data and report writing
* Assessment of the validity of the measured currents after each triggering day

Other participants and their responsibilities during the 2003 Nearby Strike Experiment:
* Principal investigators (proposal, experimental design, data analysis, general
advice): M.A. Uman and V.A. Rakov
* ICLRT manager: K.J. Rambo
* Experiment implementation: J.E. Jerauld, M.V. Stapleton, and D.M. Jordan
* Modeling of the experimental data: M. Paolone, C.A. Nucci, and F. Rachidi
* General assistance: G.H. Schnetzer









1.8 2004 FPL Experiment, Direct Lightning Strike Interaction with a Vertically-
configured Distribution Line with Overhead Ground Wire

A vertically-configured distribution line with overhead ground wire and 4 arrester

stations was studied during the summer of 2004. The three phase conductors of the

vertically-configured line tested from 2001 through 2003 were lowered and the

underneath neutral conductor was moved above the three phase conductors to function as

an overhead ground wire. The overhead ground wire was grounded at each of the line' s

15 poles. Lightning currents were inj ected into the overhead ground wire between poles 7

and 8 at midspan. The number of measurements was reduced compared to the number of

measurements during previous years-currents on the line were only measured between

poles 6 and 10.

The author of this dissertation was involved in the design and implementation of

the experiment, checking the validity of the data after each triggering day, and the

analysis of the experimental data. He is the main-author of Schoene et al. (2004b).

2004 Experiment description:
* Experimental facility (Chapter 3)
* Experimental configuration and measurement settings (Section 4.7, Appendix A)
* Triggering results (Section 4.9)

Results from the 2004 Experiment presented in this dissertation and in Schoene et al.
(2004b):
* Statistical analysis of return stroke currents (Section 6.2, Appendix C).
* Data presentation: All measurements of the 2004 experiment are presented in this
dissertation (Section 5.2, Appendix B) and in Schoene et al. (2004b).
* Data analysis: Schoene et al. (2004b) estimated the peak voltage across one of the
two arresters closest to the current inj section points to be 25 kV. This relatively low
voltage is well under the critical flashover voltage of a typical "real world"
distribution line, indicating that the arresters contributed successfully to preventing
backfl ashovers.
* Ground current division: At the time of the inj ected return stroke current peak
roughly 50% of the inj ected currents flow through one of the two closest grounds at
pole 8, 25% of the inj ected currents flow through the other closest ground at pole 7,
and the remaining 2% of the inj ected currents flow through the remaining grounds.










The pole 8 ground initially carries most of the currents, although the pole 7 ground,
which has a smaller low-frequency, low-current ground resistance, is at the same
distance from the inj section point. Apparently, the transient ground resistance at
pole 8 is smaller than the one at pole 7-maybe due to the two ground rods installed
at pole 8 (only one ground rod is installed at pole 7). For later times, the
magnitudes of the ground currents are determined by the value of the measured
low-frequency, low-current grounding resistances. (Schoene et al. 2004b)
* Disconnector operation analysis and flashover analysis: No disconnector operated
and no flashovers were observed during the 2004 experiment.

Responsibilities of the dissertation author during the 2004 Experiment:
* Experiment design and implementation
* Analysis of experimental data and report writing
* Verification of experimental data during experiment

Other participants and their responsibilities during the 2004 experiment:
* Principal investigators (proposal, experimental design, data analysis, general
advice): M.A. Uman and V.A. Rakov
* ICLRT manager: K.J. Rambo
* Experiment implementation: J.E. Jerauld and D.M. Jordan
* General assistance: M.V. Stapleton, J. Howard, and G.H. Schnetzer

1.9 2005 Lawrence Livermore Experiment, Induced Currents

Measurements of peak current in the lightning protection systems of munitions

storage structures in the UK made using an OBO Bettermann measurement system that

incorporates peak current sensor cards (plastic cards containing a magnetic stripe much

like credit cards) have yielded anomalously high peak current values, in excess of 120

kA, that are not associated with direct lightning flash attachment to the structures but

potentially could be associated with nearby lightning. To examine the issue of nearby

lightning inducing high currents on grounded structures Lawrence Livermore National

Laboratories has funded a proj ect to measure currents in a test runway counterpoise and

in a grounded vertical wire of 7 m height induced by rocket-triggered and natural

lightning currents during the 2005 lightning campaign. The results of this study also have

implications on the issue of induced currents on power lines and power line groundings.










The experimental configurations are described in Section 4.8 and an overview of

the rocket-triggered and natural lightning events that induced currents measured on the

test structures is given in Section 4.9. The experimental data are presented in Section 5.4

and analyzed in Section 6.9. The OBO Bettermann measurement system has been tested

by Newton (2004), Schoene et al. (2004a), and Hanley et al. (2006).

Responsibilities of the dissertation author during the 2005 Lawrence Livermore
Experiment:
* Design of experiments and installation of vertical wire
* Implementation of all measurements
* Calibration and maintenance of all measurements
* Analysis of experimental data
* Supervising writing of the final report (Hanley et al., 2006)

Other participants and their responsibilities during the 2005 Lawrence Livermore
Experiment:
* Principal investigator (proposal, experimental design, data analysis, general
advice): Dr. M.A. Uman
* Co-principal investigator: K.J. Rambo
* Measurement of lightning currents and/or lightning electric fields: J.E. Jerauld,
J. Howard, and B.D. DeCarlo
* Triggering: J.E. Jerauld, J. Howard, and B.D. DeCarlo
* Maintenance of measurements after end of the 2005 triggered lightning campaign:
J. Howard and B.D. DeCarlo
* Assistance in building of vertical wire, measurement
instrumentation/calibration/maintenance, and writing of final report: B. Hanley
* General advice: J.E. Jerauld, V.A. Rakov, and G.H. Schnetzer

1.10 Summary of Original Contributions

The author' s original contributions in taking new data and analyzing already-taken

and new data are outlined in Sections 1.1 through 1.9. The following paragraphs

summarize the author' s work and the maj or new findings from the data analysis.

Additional original contributions are outlined in Chapter 7.

New and additional statistical analysis of return stroke currents measured during the 1999
through 2004 FPL experiments (Section 6.2, Appendix C):
* The statistical distribution of the 10-90% risetimes of return stroke currents directly
inj ected into the line is different than the distribution of the 10-90% risetimes of









return stroke currents during the nearby strike experiment. This discrepancy is
likely due to the different strike obj ect and grounding system.
* It was determined that the charge transferred during the first millisecond after the
return stroke initiation and the return stroke current peaks are correlated by a power
regression equation (y 12.3x0.54, where y is the peak current and x is the charge
transfer, R2 0.76).

Analysis of the arrester disconnector behavior during the 2000 through 2004 FPL direct
strike experiments (Section 6.3):
* The disconnector operation common during the 2001/2002 vertical line experiment
was absent during the 2003 vertical line experiment, probably due to a transformer
on the line which protected the arresters by shunting the low-frequency current
components to ground.
* Disconnector operation during the 2000 horizontal line experiment was
considerably less frequent than during the 2001/2002 vertical line experiment,
which was possibly due to the larger number of arrester stations on the 2000
horizontal line reducing the long-duration current through each individual arrester.

New, additional, and revised analyses of the flashover occurrences during the 2000
through 2004 FPL direct strike experiments (Section 6.3):
* The number of flashovers on the horizontal line was smaller than the number of
flashovers on the vertical line which is likely related to the different number of
spans between arrester stations (horizontal line: 3 spans, vertical line: 4 spans).

Comparison of struck-phase-to-neutral and ground current divisions during the 2000
experiment with the divisions during the 2002 and 2003 experiments (Section 6.4):
* The division of return stroke currents among multiple arresters for the line tested in
2000 was initially similar to the division on the line tested in 2002 and 2003.
However, the time during which the return stroke current flowed primarily through
the closest arresters to the neutral conductor was significantly longer on the line
tested in 2000. Possible explanations for the different current division are given.

Additional modeling results of the currents on the distribution lines tested during the
2000, 2002, and 2003 direct strike experiments (Section 6.5):
* The model-predicted arrester currents match well all arrester currents measured
during the 2002/2003 vertical line experiment if the arresters are modeled with the
manufacturer-provided VI-characteristic.
* The model-predicted arrester currents do not match well any arrester currents
measured during the 2000 horizontal line experiment if the published VI-
characteristic is used for all arresters in the model. The model-predicted arrester
currents match well all arrester currents measured on the horizontal line if a
modified VI-characteristic is used for the two arresters closest to the lightning
current inj section point (that is, the residual voltages of the two closest arresters are
reduced by 20%).









Modeling the arrester-absorbed energy during a natural lightning strike to distribution
lines of various lengths (Section 6.6):
* The minimum arrester-absorbed energy is defined in this dissertation as the energy
absorbed in the arrester during the transient mode. The minimum arrester-absorbed
energy during a typical natural lightning first strokes to a distribution lines with 4,
8, and 16 arrester stations separated by 4 spans was estimated to be between 40 and
45 kJ.

Modeling results of currents induced on the vertical distribution line during the 2003
nearby strike experiment (Section 6.7):
* The measured and modeled ground and neutral currents generally match reasonably
well.
* The measured and modeled arrester and phase currents generally match reasonably
well for the duration of the first peak. The modeled arrester and phase currents are
typically larger than the measured currents for times after the first peak.

Analysis of lightning currents traversing soil and entering the neutral conductor through a
pole grounding during the 2003 nearby strike experiment (Section 6.8):
* During the 2003 nearby strike experiment a considerable fraction of the lightning
current inj ected into ground 11 m from the closest line grounding traverses the
ground and flows in the neutral conductor through the line grounding.

Presentation and analysis of currents induced on a vertical wire and buried counterpoise
during rocket-triggered and natural lightning (Section 6.9):
* The peak values of lightning return stroke currents and the peak values of induced
currents in the buried counterpoise at distances of 50 m and 290 m from the
lightning strike point are strongly linear correlated.
* The vertical wire functions as a dE/dt antenna during the rocket-triggered lightning
strikes.
* The largest induced current in the grounded vertical wire associated with natural
lightning striking ground about 300 m away was 140 A. This current was likely
associated with an upward-directed unconnected leader generated in response to an
overhead downward-propagating stepped leader step.















CHAPTER 2
LITERATURE REVIEW

The first part of Chapter 2 provides general information about pertinent physical

processes related to the lightning discharge: the physics that leads to the development of

thunderclouds (Section 2.1), the processes comprising a natural and a rocket-triggered

lightning discharge, and the characteristics of the lightning return stroke in both (Section

2.2). The first part concludes with a review of return stroke models (Section 2.3).

The second part of Chapter 2 is concerned with the effects of the lightning

discharge on power lines: distribution line design parameters are introduced (Section

2.4), the literature on modeling of overvoltages on power distribution lines induced by

nearby lightning is reviewed (Section 2.6.1), and previous work on the interaction of

direct and nearby lightning with power distribution and transmission lines is outlined

(Section 2.7).

2.1 Cloud Formation and Electrification

Lightning (also referred to as a lightning discharge or a lightning flash) attempts to

equalize regions of opposite electrical charge. At least one of these regions is provided by

an electrically charged cloud, a so-called thundercloud or cumulonimbus, which is the

energy source for most lightning strikes. The process of charge generation/separation is

called electrification. Electrification processes take place in a number of cloud types. The

anvil shaped cumulonimbus shown in Figure 2-1 (used with permission of Ari Laakso) is

the most common type. Lightning can also originate from other sources that involve

different electrification processes (e.g., clouds produced by forest fires, volcanic










eruptions, and atmospheric charge separation in nuclear weapons' blasts). The electrical

structure of a cumulonimbus and the electrification process inside this cloud type are

discussed in the following section.


























Figure 2-1: Small thundercloud over a train and a power line. The picture was taken in
Finland at the Russian border. The thundercloud produced only a few
lightning strikes.

2.1.1 Formation of a Cumulonimbus

A thundercloud is a lightning-producing deep convective cloud. Thunderstorms

form in an "unstable" atmosphere. According to Henry et al. (1994), eight types of

thunderstorms are known. The five thunderstorms common to Florida are: (1) air-mass,

(2) sea/land-breeze, (3) oceanic, (4) squall line, and (5) frontal. The first three

thunderstorm types are also known as convective or local thunderstorms. In Florida the

mechanisms that lead to the development of cumulonimbi are strongly seasonal depend.

Air-mass and sea/land-breeze thunderstorms constitute the maj ority of Florida










thunderstorms and occur most frequently during the warm summer months. Frontal

storms are a typical wintertime phenomenon, although they also occur, but less

frequently, during summertime. Air-mass thunderstorms develop if landmasses are

heated by the sun. The landmasses radiate heat to the moist layers of air near ground,

causing the air to rise. The air forms an updraft if the air mass is sufficiently large.

Sea/land-breezes are caused by the temperature difference between water and land.

Sea breezes (from sea to land near the surface) occur typically during early afternoon on

sunny days; land breezes (from land to sea near the surface) occur typically during

nighttime (Figure 2-2). Sea/land-breeze storms are primarily induced by the convergence

of air over land that accompanies sea breezes. Less common are sea/land-breeze storms

due to land breezes.


T iea breeze atter-noon


0 1 994 Ency clopaedia Britannica. Inc .
Figure 2-2: Formation of sea/land-breeze thunderstorms. Adapted from the
"Encyclopedia Britannica."










Frontal storms are usually formed if a cold front advances toward warm air,

undercutting the warm air and forcing it to rise (Figure 2-3).


Figure 2-3: Cold front moves under warm front resulting in the formation of
cumulonimbus (frontal storm). Adapted from Kuehr (1996).

When parcels of warm, moist air rise in an updraft, which is caused by one of the

mechanisms listed above, several other effects take place (Rakov and Uman, 2003):

* The air pressure decreases with height causing the parcels of moist air to expand.

* The rising air cools and condenses on small particles in the atmosphere
(condensation nuclei) once the relative humidity in the parcel exceeds saturation.
The resulting small water particles form the visible cloud. The height of the
condensation level that is the bottom of the visible cloud increases with decreasing
relative humidity at ground.

* During the condensation process the condensation heat (the energy absorbed as
water changes from liquid to vapor) is released. This heat supports the continued
upward movement of the air masses and water particles.

* Some of the water particles freeze once they reach a height where the temperature
is below 00C. At -400C, or so, all water particles freeze. The freezing process
releases the heat of freezing (the energy absorbed as water changes from solid to
liquid), which supports the further upward movement of the particle.

* A cumulonimbus develops if the decrease in atmospheric temperature with height,
the so-called lapse rate, exceeds a certain specific value which depends on the
humidity of the air, the so-called moist-adiabatic lapse rate.


C ulonumbu










2.1.2 Electrical Structure of a Cumulonimbus

The charge distribution inside a cumulonimbus is complex and changes

continuously as the cloud evolves. Most charge inside the cloud resides on hydrometeors

(liquid or frozen water particles), but also some free ions are present. Probably, charged

particles and ions of positive and negative polarity coexist in the same regions inside the

cloud, but in some areas particles of one polarity are dominant, forming regions of

positive or negative net charge.





10 -n 10

E ~ ~ ++ +
8- +
4+_ -+ + +
r+ +


4 -- +
I I- + /-+
I 4 + 000 -

++-


+FTosTn"E = : : = "'NEGATIVE
re 4 RAIN := = : : RAIN

Figure 2-4: Electrical structure inside a cumulonimbus. Adapted from Simpson and
Scrase (1937).

Early ground-based measurements of the cloud charge via its electric field by

Wilson (1916, 1920) revealed a vertical, positive dipole structure (regions of positive net

charge located above regions of negative net charge) for the primary charge regions.

Later in-cloud measurements (Simpson and Scrase, 1937) confirmed this result and

additionally identified a localized lower region of positive net charge. This positive

charge region was not always present (or could not always be detected). Figure 2-4 shows










a typical electrical structure inside a thundercloud. The top two charge regions are usually

referred to as main charge regions. In Florida the height of the main negative charge

center is typically 7-8 km above ground level, the main positive charge center is typically

located 10-12 km above ground level, and the positive charge center at the bottom of the

cloud is located 1-2 km above ground level (Rakov, 2001).

2.1.3 Electrification of a Cumulonimbus

The detailed physical processes that lead to the generation and separation of charge

and the formation of the charged regions inside the thundercloud are poorly understood.

Several hypotheses that try to explain this phenomenon have been proposed. Two of the

most important hypotheses are labeled the "precipitation theory" and the "convection

theory."

In the precipitation theory relatively heavy and large hydrometeors (precipitation in

the form of soft hail, called graupel) with a high fall speed (> 0.3 m/s) collide with

lighter, smaller hydrometeors (cloud particles in the form of ice and water) carried

upwards by updrafts (left half of Figure 2-5). Charge is transferred during the interaction

between the heavy and light particles. In relatively cold regions (T < -150C, or so) the

heavy particles will become negatively charged and the lighter particles positively

charged. In warmer regions (T > -150C, or so) at the lower part of the cloud, the process

reverses such that the heavy particles will become positively and the lighter particles

negatively charged. Gravity and updrafts separate the lighter particles from the more

heavy ones, and a main positive dipole with an additional separate localized positive

region at the bottom of the cloud forms (MacGorman and Rust, 1998).









In the convection theory electric charges are supplied by external sources-fair-

weather space charge, corona discharges on the ground and cosmic rays (right half of

Figure 2-5). Updrafts of warm air carry positive fair-weather space charge to the top of

the developing cumulonimbus. Negative charge above the cloud produced by cosmic rays

is attracted to the cloud' s surface by the positive charge within the cloud. Most of the

negative charge resides on cloud particles. Cloud particles can carry more charge per unit

volume of cloudy air than precipitation. The negatively charged cloud particles are

carried downward by downdrafts, causing corona at the surface. The corona generates

positive charge below the cloud that is carried to the upper cloud by the updrafts. This

hypothetical mechanism results in the formation of a positive dipole.











Figure~: 2-:Peiiainter lf)adcovcinter rgt.Aatdfo
Williams (199)
Alhog iti osbeta oh rcptto n ov ctinmcaim r
impotan forclod eectrfiction th prcipiatin mchansm s viwedin he ltertur
as~ th oe infiat









2.2 Natural and Rocket-triggered Lightning

Lightning discharges that neutralize charge inside a thundercloud can be

categorized into the following discharge types (Figure 2-6):

* Intracloud discharges (discharge within the cloud)

* Cloud-to-ground discharges

* Intercloud discharges (discharge from cloud to cloud)

* Cloud-to-air discharges


S+


S3+++


T I
+ J + "1"
+ + +
II++
-1U OC +
++ by ~~. **
+++++ 1

w
-15 OC -~e"-




r +r


~

+++
t +
+ ~-







++


Figure 2-6: Discharge types for a thundercloud. Adapted from the "Encyclopaedia
Britannica."

Most research has been conducted on cloud-to-ground discharges since this

discharge type is the cause of most lightning damage, injury, and death. Intercloud and

cloud-to-air discharges are thought to be relatively rare compared to intracloud and

cloud-to-ground discharges.










2.2.1 Lightning Discharges between Cloud and Ground

Berger (1978) classifies lightning discharges between cloud and ground into four

categories, based on the direction of propagation of the initial leader and the polarity of

the charge transferred from the cloud to earth (Figure 2-7):

* Type 1: Downward lightning discharge, lowering negative charge to earth

* Type 2: Upward lightning discharge, lowering negative charge to earth

* Type 3: Downward lightning discharge, lowering positive charge to earth

* Type 4: Upward lightning discharge, lowering positive charge to earth


~~%v~~ "
1


4


3 7


Figure 2-7: Simplified drawing of four discharges between cloud and ground. Adapted
from Uman (1987).

Upward directed flashes (type 2 and 4) are typically initiated from tall structures on

flat ground or structures of moderate heights on mountains. The highest recorded

lightning currents (up to 300 kA) and the largest recorded transferred charge (hundreds of










coulombs or more) are thought to be associated with lightning lowering positive charge

to ground (type 3 or 4). Positive lightning contains typically one stroke while most

negative lightning contains more than one stroke. Negative downward lightning (type 1)

constitutes approximately 90% of all cloud-to-ground flashes. The remaining 10% of all

cloud-to-ground flashes are covered by the other three categories; with positive

downward lightning (type 3) being the most frequent of the three.

A rough outline of the physical processes involved in negative downward lightning

is described and illustrated in Figure 2-8.


*, yia rpl tutueo ihnn rdcn
cumulonimbus-upper positive main charg~eregion; lower
OLOUD CHARGE negative main charge region; localized positive charge
olspermours region at the cloud base. If the electric field at the bottom
of the negative charge region reaches a critical value, a
e preliminary breakdown starts (t=0).
t= 0




*An in-cloud breakdown starts from the negative charge
PRELIMINARY TegiOn toward ground carrying negative charge and
neutralizing the positive charge region at the cloud base (1
ms).

1.00 ms




*A stepped leader, consisting of a thin, highly ionized core
surrounded by a wider corona sheath, leaves the cloud (1.1
STEPPED mS>.
LEADER



1.10 ms

Figure 2-8: Drawings illustrating some of the various processes comprising a negative
cloud-to-ground lightning flash. Adapted from Uman (1987).





















1.20 ms


1.15 ms
Figure 2-8: continued


*The stepped leader moves with a typical average speed of about 2*105 m/s toward
ground. Negative charge from the cloud flows more or less continuously into the
leader channel (1.15 ms-19 ms).


20.10 ms


20,00 ms
Figure 2-8: continued


*When the stepped leader approaches ground, the electric field strength at certain
points on ground (notably at sharp and elevated obj ects) exceeds the breakdown
value of air. At these points one or more positive upward-going leaders develop in
the direction of the negative downward going leader from the cloud (20 ms).

An upward-going leader connects with a downward-going leader branch. A current wave
with a current peak value of typically 30 kA, the first return stroke, starts propagating
upward along the ionized channel prepared by the leader (20.10 ms).


19.00 ms


ATTACHMENT
PROCESS

























Figure 2-8: continued

*The first return stroke neutralizes the negative charge deposited in the leader
channel and in the process lowers negative charge to ground. The return stroke
travels upward with a speed in the order of 10s m/s (20.10 ms-20.2 ms).


*Following the first return stroke, the cloud
region, where the leader has started, is near
ground potential. Discharges between this
region and negatively charged regions in the
cloud can occur, so-called K and J processes
(40 ms).


SK AND J
i' PROCESSES


20.15 ms


20.20 ms


40.00 ms
Figure 2-8: continued























60.00 ms 61.00 ms 62.00 ms
Figure 2-8: continued

*A dart leader may form if the channel of the first return stroke has not yet
dissipated. The dart leader usually follows the already existing channel prepared by
the return stroke; therefore it is typically not branched. The speed of the dart leader
is typically 107 m/s (Uman, 1987) and it lowers negative charge onto the defunct
channel of the previous stroke (60 ms-62 ms).







Once the dart leader approaches ground, a
second return stroke develops in similar
SECONDmanner to the first return stroke (62.05 ms).
RETURN Additional subsequent returns strokes can
STROKE occur if this process repeats itself.


DART
LEADER


Figure 2-8: continued


62.05 ms









2.2.2 Rocket-triggered Lightning

The probability of a lightning strike to a structure, even in areas of high lightning

activity, is very low. Thus, experiments involving close natural lightning are difficult to

conduct (or take an inordinately long time). A more practical approach to conducting

such experiments is to artificially initiate a lightning strike using the rocket-and-wire

technique (Figure 2-9).
























Figure 2-9: Rocket-triggered lightning in Camp Blanding, Florida (Flash U9910).

The lightning type initiated by using the rocket-and-wire technique is termed

rocket-triggered lightning. In rocket-triggered lightning, a small rocket with an attached

conducting wire is used to artificially initiated lightning. Under favorable conditions, i.e.,

measured static electric Hield on ground < -5 kV/m, a rocket trailing a conducting wire is

launched with a speed of 200 m/s towards a thundercloud. In the classical rocket-

triggered lightning technique, illustrated in Figure 2-10 (Rakov, 1999b), the triggering

wire is a continuous conductor that is connected to ground.



















































Figure 2-10: Sequence of events in classical rocket-triggered lightning.


* As the rocket ascends with a speed of approximately 200 m/s, the electric field at
the tip of the rocket is distorted.

* When the rocket reaches an altitude of about 200 m to 300 m, the field
enhancement at the rocket tip can result into the development of a positive leader
(provided that a sufficient ambient negative field is present) ascending with a speed
of the order of 105 m/s from the rocket tip towards the thundercloud.





















































Figure 2-10: continued

* The upward-going positive leader vaporizes the wire and establishes an initial
continuous current (ICC), which flows for typically some hundreds of milliseconds
through the channel.

* After the cessation of the ICC a no current interval having a typical duration of tens
of milliseconds occurs.










,,,~ .;.*r


Figure 2-10: continued

*The ICC may be followed by one or more downward leader/upward return stroke
sequences. These leader/return stroke sequences are believed to be very similar, if
not identical, to the subsequent leader/return stroke sequences occurring in natural
lightning.









2.2.3 Return Stroke Current

In this section a typical rocket-triggered lightning return stroke current that lowered

negative charge to ground is shown and parameters that characterize this current are

introduced (Figure 2-11). The return stroke current measured at the channel base is

characterized by a sharp rising edge followed by a much slower decaying part.

The return stroke current peak Ipeak (Figure 2-11a) determines the maximum

overvoltage Upeak CaUSed by a direct lightning strike (17peak =peak Z Z, where Z is the

impedance of the struck obj ect). Berger et al. (1975) found that 50% of natural lightning

first/subsequent return stroke peak values exceed 30/12 kA and 5% exceed 80/30 kA.

The 10-90% risetime tl0-90o/o (Figure 2-11Ib) characterizes the rising edge of the

return stroke current. Berger et al. (1975) measured the 2 kA to peak value (a parameter

similar to the 10-90% risetime) of natural lightning return strokes and found that the

risetime of 50% of first/subsequent stroke currents exceed 5.5/1.1 Cls and 5% of the

risetimes exceed 18/4.5 Cls.

The maximum time rate of current change max(AI/At), an important parameter for

calculating induced effects of lightning, can be estimated using the 10-90% risetime and


the ~ ~ ~ = retun srok pek mx -peak Berger et al. (1975) found that 50% of


natural lightning first/subsequent stroke dl/dt peak values (a parameter that gives the

exact maximum time rate of current change) exceed 12/40 kA/Cls and 5% of the peak

values exceed 32/120 kA/Cls.



SRetumn stroke currents in rocket-triggered lightning resemble subsequent return stroke currents in natural
lightning (see Section 2.2).




































0 5 10 15 20 25 3(


-* -4 Charge Transfe


-12 -~ls~- d)

05 10 1~5~ 201 25 31
Figure 2-11: Typical triggered lightning return stroke current waveform measured at the
channel base (stroke FPLO315-2). The following return stroke current
parameters are illustrated: a) peak value, b) 10-90% risetime, c) half-peak
width, and d) charge transfer.









The half-peak width (Figure 2-11Ic) characterizes the tail of the return stroke

current. The half-peak width is related to the charge Q (Q! = Ildt, the area under the

current waveform) transferred during the return stroke (Figure 2-11d). The charge

transfer is related to the electrical energy E a lightning strike delivers to a struck obj ect

(E = U-I Idt, where U is the voltage across the struck object) and which can cause

melting damage. Berger et al. (1975) measured the 2 kA to half-peak value (a parameter

similar to the half-peak width) of natural lightning return strokes and found that the

widths of 50% of first/subsequent strokes exceed 75/32 Cls and 5% of the widths exceed

200/140 Cis.

2.3 Transmission-line Type Return Stroke Models

Return stroke models are used to calculate lightning-induced overvoltages on

power lines (Section 2.6). A transmission-line type return stroke model is a so-called

"engineering" model, that is, a model which relates the longitudinal return stroke current

for every time and position along the lightning channel I(z',t) to the current at the channel

base I(0,t). An equivalent expression specifying the line charge density along the channel

can be obtained using the continuity equation (Thottappillil et al., 1997). These

expressions can be used to calculate the lightning electric and magnetic fields at specified

locations.

"Engineering" models have been discussed in the literature by, for instance, Bruce

and Golde (1941), Uman and McLain (1969), Rakov and Dulzon (1987, 1991), Willett et

al. (1988), Willett et al. (1989), Diendorfer and Uman (1990), Nucci et al. (1990),

Thottappillil et al. (1991), Thottappillil and Uman (1993), Thottappillil and Uman (1994),

Thottappillil et al. (1997), and Rakov and Uman (1998, 2003).









Rakov (1997) proposed a generalized equation for a channel current that represents

many "engineering" models:

I(z', t) = u(t z'/vf) P(z') I(0, t z'/v) 2-1

where vf is the upward propagating front speed (return stroke speed), v is the current

wave propagation speed, z' is the height of the channel section, u is the Heaviside unit

step function (u(t z'/ve) = 1 for t < z'/ve, otherwise u(t- z'/ve) = 0), and P(z') is the height

dependent current attenuation factor introduced by Rakov and Dulzon (1991).

In transmission line type models the lightning channel behaves similar to a

transmission line, that is, a current wave inj ected at the channel origin travels upward.

The transmission line type return stroke models discussed in this section are (1) the

transmission line model (TL model), (2) the modified transmission line model with linear

current decay with height (MTLL model), and (3) the modified transmission line model

with exponential current decay with height (MTLE model).

Figure 2-12 illustrates the current division specified by the TL model. The current

wave is shown for height z' = 0, z'= zl', and z' = z2'. The current wave travels upward

without distortion or attenuation and with constant speed vt (illustrated by the dotted

line). Current wave and current front propagate in the same direction with the same speed

(v = yr). The current at for instance height z2' is the same as the current at ground (z2l

0) at time t = z2'/vf earlier. Therefore the equation for the TL model is:

I(z', t) = u(t z'/vf)- I(0, t z'/vf) 2-2

In the TL model no charge is deposited or removed from the channel; that is, the

total net charge of the channel remains unchanged after the current wave has passed.









MTLL and MTLE model are similar to the TL model except that the current decays

with height. The current decay is linear for the MTLL model and exponential for the

MTLE model. Therefore the equation for the MTLL model is

I(z', t) = u(t z'/ vf)- (1- z'/H) I(0, t z'/vf) 2-3

where H is the total height of the channel, and the equation for the MTLE model is

I(z',tI) = u1(t z-'/ ,)- e": I(0,t z-'/ ,,,) 2-4

where h is the current decay constant. Figure 2-13 shows the height dependent current

wave for the TL model at an arbitrary fixed moment in time (t = tl). The current wave

moves from left to right, that is, upward along the channel. The upward direction of

propagation of the current wave is a characteristic feature of transmission line type

models.























0 t

Figure 2-12: Current versus time waveforms specified by TL model at ground (z'= 0) and
at two heights zl' and z2 -









I (z', t)















Figure 2-13: Current versus height z' above ground at time t = tl for the TL model.

2.4 Distribution Line Design Parameters

Distribution lines are relatively low-voltage lines (below 50 kV) that deliver power

from the substation to the consumer. The goal of lightning protection of distribution lines

is to prevent (1) flashovers and (2) system damage. A direct lightning strike to a phase

conductor may cause flashovers from the struck-phase conductor to other conductors,

obj ects, or to ground. A direct lightning strike to the shield wire of a distribution line may

cause flashovers to the phase conductor-so-called back-flashovers. Flashovers may also

be caused by induced overvoltages due to a lightning strike near the line. If a flashovers

occurs circuit breakers on the line operate to prevent that the flashover is sustained by the

power-frequency current (Rakov and Uman, 2003). The circuit breaker operation reduces

the power quality by causing unwanted transients due to the momentary power

interruption. A direct lightning strike to a phase conductor may also cause damage to

expensive equipment on the line, such as distribution transformers, or damage to the

protection devices on the line, such as arresters.









This section discusses design aspects pertinent to improving the performance of

overhead distribution lines in the lightning environment, that is, reducing flashovers and

system damage due to lightning caused overvoltages.

2.4.1 Insulation Strength of Distribution Lines

The insulation strength of distribution lines is commonly given in terms of the

critical flashover voltage (CFO). The CFO is defined as the peak value of a specified

voltage impulse applied to the line insulation that has a 50% probability of flashover. The

CFO of a 1.2/50 Cls voltage waveshape is typically used for the lightning insulation

coordination. Barker et al. (1996) show that measured induced voltages typically exhibit

a slower risetime and a faster decay than the 1.2/50 Cls test waveform and therefore

produce less stress on insulation. They estimate that the actual CFO for insulation is up to

50% larger than the CFO determined with the 1.2/50 Cls test waveform, primarily because

of the narrow width of the voltage pulse. In the IEEE standard 1410-1997, a flashover is

assumed to occur if the voltage across the insulation exceeds 1.5 times the CFO of the

insulation. The CFO of a distribution line can be found experimentally in laboratory

impulse tests or estimated with "rule of thumb" formulas based on laboratory

experiments. The insulation of a distribution line is composed of different components,

such as insulators, wooden crossarms (insulation strength 360 kV/m), wooden poles

(insulation strength 330 kV/m), fiberglass standoffs (insulation strength 500 kV/m), and

air gaps (insulation strength 600 kV/m) It is difficult to estimate the CFO of a line since

experiments show that the CFO of a line is generally considerably smaller than the sum

of the CFO of the individual insulation components (Jacob et al., 1991). Guidelines for

SAll insulation strength values are given for wet conditions. The dry condition CFO is approximately 10-
30% larger.









estimating the CFO of a distribution line can be found in IEEE standard 1410-1997 and

in Jacob et al. (1991). Note that equipment on the poles or system elements such as guy

wires, fuse cutouts, conducting supports, and other conductors in the system can reduce

the CFO (Rakov and Uman, 2003).

2.4.2 Overhead Ground Wires

An overhead ground wire (OHGW) is a conductor that is placed above the phase

conductors and is typically grounded via a pole ground lead at every pole. An OHGW

can protect line equipment from direct lightning strike damage and can reduce the

number of flashovers due to direct lightning strikes by intercepting and conducting to

ground the lightning current that would otherwise enter a phase conductor. It is important

that the OHGW be well grounded and that the CFO between the system of OHGW and

ground leads and the phase conductors is sufficiently large to avoid back flashovers, that

is, flashovers from the OHGW or from the ground leads to a phase conductor. An

OHGW can reduce the number of flashovers due to nearby lightning strikes by reducing

the induced voltage on the phase conductors through capacitive coupling (shielding). The

shielding effect of the OHGW improves by placing it closer to the phase conductors but

this increases the probability of backflashovers in direct strikes. According to theory, an

OHGW reduces the magnitude of the induced voltages on the phase wires by about 15-

45% (Rachidi et al., 1997a; Yokoyama et al., 1984).

2.4.3 Metal-Oxide Arresters

Surge protective devices (SPD) limit transient overvoltages by diverting the

transients to ground (transient overvoltages on power lines can be caused by, for instance,

lightning, normal utility operations such as capacitor bank switching or tap changing on

transformers, or the turning on/off of inductive loads such as motors or transformers.).









SPDs that divert transients behave as open circuits or high impedance during normal

operation conditions and as short circuit or low impedance during transient overvoltages.

They can be categorized into two types-(1) crowbar devices and (2) voltage clamping

devices (Rakov and Uman, 2003).

Crowbar devices are triggered by the electric breakdown of a gas (e.g., air gaps) or

an insulating layer (e.g., thyristors). When triggered, they become very low impedance

which allows the shunting of large transients with little energy absorbed in the device and

are consequently relatively safe from getting damaged during the transient discharge. The

response time of gas breakdown crowbar devices is relatively large (microsecond time

scale for air gaps) due to the time it takes to establish a breakdown path.

Voltage clamping devices arresters (also know as varistors) have a non-linear,

continuous impedance that becomes low during transient overvoltages and reduces

(clamps) the transients to safe levels. In general, voltage clamping devices have the

advantage of a fast (nanosecond) response time over crowbar devices thus diverting

overvoltages before they cause a flashover or can do damage to the protected equipment.

The following section will focus on Metal-Oxide Varistors (MOVs), which are voltage

clamping devices commonly installed on power lines to protect line equipment and to

prevent flashovers.

MOVs are the most widely used arresters for the protection of power lines. A

typical VI-characteristic of a low-voltage MOV is shown in Figure 2-14. For low

voltages (voltages not larger than the normal operation voltage of the power line) MOVs

have a large resistance and conduct very little current (leakage region). The small current

is called "leakage current." For large voltages (e.g., transient overvoltages on power





























)00


SLOPE 1,





200


lines) MOVs show highly non-linear behavior, that is, a large change of the MOV current

causes only a small change of the MOV voltage-the voltage is clamped to a nearly

constant value ("normal varistor operation"). For very large voltages the MOV has a

small resistance ("upturn region"). The VI-characteristic for the "normal varistor

operation" can be expressed as

I =k-V" 2-5

where k and oc are device constants.

LEAKAGE NORMAL VARISTOR UPTURN
4 ~REGION OPERATION -REGION


10


10 I I I I1[1 1
10g 106 -10- 10-2 100 102 104
CURRENT (A)
Figure 2-14: Typical varistor VI-characteristic plotted on a log-log scale.

MOVs are bipolar ceramic semiconductor devices that are made of a ceramic

compound consisting primarily of a metal oxide, most commonly zinc oxide (ZnO),

doped with additive compounds such as oxides of bismuth, barium, cobalt, manganese,

chromium, or tin (Hillman, 2005). ZnO MOVs consist of microcrystalline zinc oxide

grains that are isolated from each other by a thin intergranular phase consisting of the

additive metal oxides (Figure 2-15).










WIRE LEAD

ELECTRODE


INTER %
GRANULAR
PHASE


GRAINS





EPOXY
ENCAPSU LANT I ~- WIRE LEAD

Figure 2-15: Structure of a metal-oxide varistor. Adapted from Harnden et al. (1972).

The non-linear properties of the arresters impedance are established at the interface

between the ZnO grains and the intergranular phase throughout the whole MOV

(Hillman, 2005). Electrically, each of these boundaries act like a diode, that is, they block

current flow at voltages below the turn-on voltage (this state corresponds to the "leakage

region" of the MOVs VI-characteristic in Figure 2-14) and conduct current at voltages

above the turn-on voltage (this state corresponds to the "normal varistor operation" in

Figure 2-14). The MOV as a whole acts like parallel strings of diodes connected in series.

For very large currents the resistance of the intergranular phase becomes less than that of

the ZnO grains resulting in a linear behavior of the VI-characteristic (this state

corresponds to the "upturn region" in Figure 2-14).

MOVs designed for power line protection (Figure 2-16) contain disks of a few

centimeters thickness that are composed of the metal-oxide compound described above

and are coated with a highly conducting material such as a silver containing compound.











a)b) uino ~-I Stanl..ss ...
Terminal Top Cap


Aluminum
Line Electrode
Silicone
Rubber
Housing


MOV
/ Disks



Interface
Glass
Reinforced
Epoxy Collar




6 & Aluminum
Ground
It f ~ ~Electrode nuae
Ground
Hanger
Terminal

Figure 2-16: Illustrations of gapless metal-oxide arresters. a) Cooper Ultrasil housed
Varistar 10 kV distribution arrester. b) Cutaway illustration of a different
Cooper distribution arrester. Adapted from the 2000 Cooper arrester catalogue
235-35.

The disks are stacked to achieve the desired voltage rating and energy capability and

enclosed inside a housing that provides electrical insulation and mechanical stability.

MOVs designed for power line protection have a large, although limited, capability

to absorb energy. The rated energy capability of an MOV can be significantly reduced if

the MOV is repeatedly subj ected to overvoltages. If the energy capability of an MOV is

exceeded, the MOV can be damaged due to excessive heating and can fail in three

modes, that is, the failed MOV can act as a (1) short circuit, (2) open circuit, and (3)









linear resistor (Brown, 2004). A short circuit failure can be caused by sustained

overvoltages or excessive leakage current causing a puncture site within the disk. An

open circuit failure can be caused by large failure currents causing melting of wire lead

and electrode solder junctions or by cracking/shattering of the MOV disk. Some MOV

employ a safety device that disconnects the leads during the 50/60 Hz current following

short circuit failures of MOVs. The disconnector ensures that the line remains operational

by opening the short circuit caused by the failed arrester. Additionally, the operated

disconnector makes a damaged arrester easily detectable (Lenk, 2004).

2.5 Modeling Direct Strikes to Power Distribution Line

Direct strikes to power distributions lines are modeled in this dissertation using the

Electromagnetic Transient Program (EMTP), version EMTP96 3.2d. The EMTP was

designed to model transient processes in power systems, although it is also capable of

treating steady-state problems such as the load-flow analysis of a power system. The

EMTP provides a discrete solution of ordinary differential equations using the trapezoidal

rule of integration together with traveling wave solution methods based on Bergeron' s

method (Dommel and Meyer, 1974). EMTP96 features EMTP View, a Graphical User

Interface (GUI) that can be used to graphically design a model. The graphically designed

model is converted to a Fortran code that is used as the input to the EMTP. Alternatively,

the user can omit the use of the GUI and create the model by directly writing the Fortran

code. In the following sections a brief history of the development of the EMTP is given

and the various components implemented in the EMTP and used in the EMTP model in

this dissertation are described.









2.5.1 History of the Electromagnetic Transient Program

The EMTP was developed in the late 1960's by Dr. Hermann Dommel as a digital

counterpart to the analog Transient Network Analyzer (TNA), which in turn was

developed in the late 1930's to model transients in power systems. The capability of the

EMTP has been expanded over the years by Dr. Scott Meyer, Dr. Tsu-huei Liu, and

others. The EMTP Development Coordination Group (DCG), founded in 1982,

commercialized the EMTP in 1984. As a consequence of their concerns with the

commercialization of the EMTP, Dr. Tsu-huei Liu resigned as DCG chairman and Dr.

Scott Meyer left DCG and developed a free version of the EMTP (although licensing is

required) called the Advanced Transient Program (ATP). The first commercial DCG

EMTP version was released in 1987 and version 2.0 was released in 1989. Version 3 of

the DCG EMTP was released in 1996 (EMTP96), which is used in this dissertation to

model direct lightning strike effects on power distribution lines (Section 6.5) and is

integrated in the LIOV-EMTP96 code used in this dissertation to model nearby lightning

strike effects on power distribution lines (Section 2.6.3 and Section 6.7). The EMTP96

represents the final version of the EMTP based on the original code. EMTP-RV was

released in 2003, developed by Hydro-Quebec, and is a completely restructured version

of the EMTP. EMTP-RV features better simulation performance in terms of speed and

stability, and greater design convenience, largely accomplished by a new graphical user

interface for the model design (EMTPWorks) and the implementation of a new data

vi sualization/analysi s tool (ScopeView). Note that the older version of the EMTP (that i s,

EMTP96) was used for the simulations presented in this dissertation as the latest version

(EMTP-RV) was not purchased. (sources: Donanel and Meyer, 1974; EM~TP Rulebook,

2001; www. entp. org; www. entp. cons)









2.5.2 EMTP Current Sources

Type-1 through Type 10 current sources in EMTP96 allow the user to specify the

source in a point-by-point fashion (that is, the source function f(t) is defined empirically

at every time step). Type-11 through Type-15 current sources in EMTP96 specify the

source by using analytical functions of time. Currents with a complex shape can be

generated by a parallel combination of Type-1 1 through Type-15 sources and by

choosing appropriate turn-on and turn-off times for the sources. In the EMTP all sources

are connected between a node and local ground. The source functions are evaluated at

discrete time steps only. Linear interpolation between discrete points is assumed by the

program. (source: EM~TP Rule Book, 2001)

2.5.3 EMTP Arrester Model

Models that simulate the non-linear behavior of various arrester types (Section

2.4.3) are implemented in the EMTP. The Type-92, 5555 component implemented in the

EMTP allows the representation of gapless ZnO surge arresters in the model. An arrester

model can be implemented by (1) creating a data card that contains the desired VI-

characteristic and the rated voltage of the arrester, (2) converting the data card to a punch

file using the EMTP support program AUX, and (3) using the punch file in the EMTP

model. The non-linear resistance is represented by a power function of the form



V,-


where v and i are the arrester voltage and current, respectively, and p, V~fand q are

constants. V~fis typically twice the rated voltage of the arrester and is used to normalize

the equation and prevent numerical overflow. In the EMTP each segment of the VI-

characteristic is defined by a separate power function, except for voltages substantially










below Y, efor which a linear representation is used to avoid exponential underflow and to

speed the solution.

The compensation method is used to solve circuits with non-linearities in the

EMTP model. If this method is applied to a circuit with a single non-linear element, two

equations need to be compensated-(1) an equation obtained by removing the non-linear

element and representing the linear portion of the circuit as a Thevenin equivalent circuit

and (2) an equation that exactly describes the non-linear characteristic of the non-linear

element (Equation 2-6 for the case of an arrester being the non-linear element). The

Newton-Raphson iterative method is used in the EMTP to solve the two equations.

Matrix algebra can be applied in the compensation method to solve circuits with multiple

non-linearities. (sources: EM~TP Rule Book, 2001; Dommel, 1986)

2.5.4 EMTP Transmission Line Models

Various models that simulate the electric characteristics of transmission lines are

implemented in the EMTP-(1) lumped parameter models suitable for steady-state studies

and (2) distributed parameter models suitable for steady-state and transient studies. In

lumped parameter models the electrical properties of the line are modeled with a circuit

consisting of a combination of discrete R, L, and C parameters while in distributed

parameter models the line parameters are uniformly distributed along the length of the

line and traveling wave methods are implemented. The two distributed parameter models

implemented in the EMTP are (1) the constant parameter model, which assumes the line

parameters to be constant at a given frequency and (2) the frequency-dependent

parameter model (also known as the JMARTI model) described in Marti et al. (1993), in

which the frequency dependence of the line parameters is taken into account.









2.5.5 Leads Connecting the Neutral Conductor to Ground Rods

The leads connecting the neutral conductor to ground rods can be represented by a

series connection of n sections with each section consisting of a grounded capacitor and

an inductor (Mata, 2000). The values of the capacitor and the inductor in each section can

be calculated by multiplying the capacitance C and inductance L, each per unit length, of

a vertical wire above ground by the length of the section (Bazelyan et al., 1978):

2zie,
C= [F/m] 2-7
In(2h/r)


L = I-n(2h/r) [H/m] 2-8
2xi

where h is the height above ground of the midpoint of the section, r is the conductor

radius, so is the permittivity of free space (8.8541878176-10-12 F/m) and Clo is the

permeability of free space (Clo=4n: 10-7 N/A2)

Mata et al. (2000a) modeled the current division on a two-conductor test power

distribution line exposed to rocket-triggered lightning currents in 1996 (Section 2.7.9) by

representing the 5.5 m long ground leads with 11 sections. The number of sections was

found by incrementally increasing the number of sections until no significant difference

in the calculated currents and voltages was observed by adding an additional section.

Mata (2000) used 20 sections (the number of sections was also determined by trial-and-

error) in his EMTP model that calculated the current division on the horizontally-

configured line tested in 2000.

2.5.6 Ground Rod Model

A lumped parameter representation of grounding rods is commonly adapted to

model the electric behavior of ground rods in response to lightning transients (e.g.,









Verma and Mukhedkar, 1980; Meliopoulos and Moharam, 1983; Mata et al., 2000a;

Paolone et al., 2004b). In this representation the grounding system is modeled as a series

connection of n RLC-sections as shown in Figure 2-17. The capacitance and inductance

in each section is given by


C = el 10-9 [F] 2-9
181n(41/d)n

and
L=211n(41/ d)10 []20


where sr is the relative permittivity of the soil, I is the length of the ground rod, d is the

diameter of the ground rod, and n is the total number of sections (Mata et al., 2000a). The

resistance in each section is given by

R = RDC 77 2-11

where RDC is the measured low-frequency, low-current grounding resistance of the

ground rod and n is the total number of sections. More sophisticated models that take the

non-linearity of the ground rod resistance due to the ionization of soil into account are

available (Imece et al., 1996). However, it has been shown in an example by Mata et al.

(2000a) that for low enough values of RDC (below 56 02), large enough values of the

ground resistivity (above 4000 OZm), and small enough ground rod currents (less than 8

kA) the non-linear resistance approximates RDc. Note that all these conditions are met in

the 2000 and 2003 experiments modeled here and that, consequently, RDC is used for the

ground rod resistance in the EMTP modeling included in this dissertation. Note also that

surface arcs that develop radially from the ground rod along the ground surface thereby

reducing the ground resistance with respect to RDc are HOt taken into account in the






ground rod model presented above. Mata (2000) used 50 sections (the number was
determined by trial-and-error) in his EMTP model that calculated the current division on
the hori zontal ly -confi gure d li ne te sted i n 200 O0.

Ground Lead
Connection



O L C





Cr)I



O L C





Figure 2-17: Lumped parameters representation of ground rods. The model consists of n
RLC-sections.









2.6 Modeling Nearby Strikes to Power Distribution Lines

Information on models that calculate overvoltages on power distribution lines due

to nearby lightning and testing of these models is provided in this section.

2.6.1 Calculation of Lightning-induced Overvoltages

Overvoltages on power lines caused by nearby lightning can be calculated in two

steps:

1. Determining the electric field by (a) adopting a return stroke model to specify the
temporal and spatial current distribution in the lightning channel from which the
electric field is calculated (Rakov and Uman, 1998) or (b) measuring electric fields
and extrapolating to fields at locations where the field was not measured.

2. A coupling model is adopted that calculates the voltages on the distribution line
induced by the external electric fields determined in step (1). Various coupling
models have been developed to estimate the voltage induced on distribution lines
by nearby lightning. The most important models are based on transmission line
theory extended to include appropriate distributed sources and are described in
Rusck (1958, 1977), Chowdhuri and Gross (1967), and Agrawal et al. (1980). The
LIOV-EMTP96 code uses the Agrawal model. Note that in the Rusck model and
the Chowdhuri and Gross model some source terms have been omitted (Nucci et
al., 1995; Cooray, 1994; Cooray and Scuka, 1998).

Rusck gives a simplified formula to calculate the maximum value of induced

overvoltages Vmax inferred from his general coupling model (Rusck, 1977). The electric

fields used in this simplified coupling model are calculated by assuming an infinitely

long single conductor line above a perfectly conducting ground and a step lightning

current impulse in a vertical channel:

I, -h!' 2v 2
Vma = Z 1 21





i Transmission line theory refers to the theory to model a transverse electromagnetic (TEM) waveguide
such as the parallel conductors of a transmission line as a distributed circuit and is not to be confused with
the transmission line lightning return stroke models introduced in Section 2.3.










Zo 1 0- 3002 2-13


where I, is the maximum lightning current at the channel base, h is the height of the

conductor, d is the distance of the stroke location from the line, and v is the ratio between

the return stroke speed and the speed of light.

Rachidi et al. (1997a) discuss the shielding effects due to the presence of other

conductors and a ground wire and give simple analytical formulas to calculate the

shielding coefficients. The conductor shielding coefficient SCi is the ratio between the

peak voltage on conductor i with other conductors present and the calculated peak

voltage on a single conductor with inductance Li and capacitance Ci at the same height.

Rachidi et al. (1997a) derived a formula to approximate SCi:

2-Z
SC, 2-14
Z, +Z*

where Zi is the characteristic impedance of the single-wire line i with radius ri at height



hi.,



Z,' is the characteristic impedance of an equivalent single wire representation of

conductor i in a line with N conductors. The conductor i has inductance/capacitance per

unit length I' / C', self inductance/capacitance per unit length Lii/Cii, mutual

inductance/capacitance per unit length L,/C,, and is located at distance d, from

conductor j. L/C is the inductance/capacitance matrix per unit length.


Z* = 2-16










L' = hf h3 L~ 2-17
1 1=1


C, = hl h3-C, 2-18
1 ]=1


L 00 In '2 h 2-19



L 00 In 1 4 -h h, 2-20
4-x d


C = Clo80L 2-21

Paul (1994) derived Equations 2-19-2-21 for multi-conductors over a perfectly

conducting ground. Note that if the coupling model of Agrawal et al. (1980) is applied to

an infinitely long, lossless line the induced voltages on the conductors are decoupled, i.e.,

the induced voltage on a given conductor is not affected by the presence of other

conductors (Rusck, 1958; Yokoyama, 1984; Rachidi et al., 1997a). A shielding effect due

to the presence of other conductors exists for lossy lines, lines of finite length, and for

lines with a line conductor held at a fixed potential (Rachidi et al., 1997a).

The ground wire shielding coefficient SGi is the ratio between the peak voltage on

conductor i without and with ground wire. Rusck (1958) gives a simplified formula to

estimate SGi:

h, c-L,
SG, = 1 2-22
h, c-L, +R

where the index g denotes ground wire parameters, R is the grounding resistance of the

ground wire, and c is the speed of light. Equation 2-22 assumes that the ground wire is at

ground potential at all times, which is an unrealistic assumption for large spacing









between adj acent groundings (Paolone et al., 2004a), and that the ground is perfectly

conducting.

2.6.2 Testing of Lightning-induced Overvoltage Models

Barker et al. (1996) measured voltages induced by rocket-triggered lightning 145 m

away from a one-phase test distribution line with an underneath neutral conductor

(Section 2.7.9 for more details on that experiment). The measured induced voltage peaks

were 63% larger than the voltage peaks predicted by the simplified Rusck model

(Equation 2-12).

Rachidi et al. (1997a) show that the shielding coefficient SCi due to other

conductors, obtained by solving the simplified equation (Equation 2-14) is practically

equal to the value obtained by solving the Agrawal coupling equations. They also

compared ground wire shielding coefficients SGi calculated with Rusck' s simplified

formula (Equation 2-22) with results calculated using the Agrawal coupling equation and

found that the shielding coefficients from the Rusck formula are 6 to 7% smaller.

Paolone et al. (2004a) found that the modeled induced currents calculated using

their LIOV-EMTP96 code, which implements a transmission line type return stroke

model (Section 2.3) and the Agrawal coupling model, are in very good agreement with

induced currents measured in a reduced scale model.

Paolone et al. (2000, 2004a) compare voltage peaks calculated (a) with the LIOV-

EMTP96 code for a multi-conductor line and (b) with the LIOV-EMTP96 code for a

single conductor line and multiplied by the shielding coefficient obtained from the

simplified Rusck formula (Equation 2-22). They found that the results are in good

agreement if the simplified Rusck model assumption of a zero potential shielding wire is

approximated by periodically grounding the shielding wire at least every 200 m. For less









frequent grounding of the shielding wire (every 500 m or more) the simplified Rusck

model underestimates the induced voltages significantly. According to their simulation,

the shielding effect of the ground wire depends strongly on the number of periodical

groundings of the shielding wire and not so much on the value of the grounding

resistance if the grounding resistance is below 100 0Z and if the strike location is not

facing one of the groundings. If the strike location is located in front of one of the

groundings, the shielding effect is strongly dependent on the value of the grounding

resistance.

2.6.3 LIOV-EMTP96 Code

The LIOV-EMTP96 code (Paolone et al., 2001) has been employed to model the

lightning-induced currents on the UF/FPL test power distribution line (Section 6.7). The

model was provided by Dr. Mario Paolone, Dr. Carlo Alberto Nucci, and Dr. Farhad

Rachidi as part of a j oint University of Florida/University of Bologna/Swiss Federal

Institute of Technology proj ect. Dr. Mario Paolone visited the ICLRT during the

Summers 2002 and 2003 to participate in the validation of the LIOV-EMTP96 code with

UF/FPL experimental data from the test distribution line.

The LIOV-EMTP96 code links the LIOV code (Nucci et al., 1993; Rachidi et al.

1997a) with EMTP96 (Section 2.5). The LIOV code is based on the coupling model of

Agrawal et al. (1980). The Agrawal model has been implemented into the LIOV

computer code to calculate the response of a distribution line to a LEMP (Lightning

Electromagnetic Pulse). The EMTP96 implements the boundary conditions and provides

a convenient way to include electrical components in the overall model. The lightning

vertical electric and horizontal magnetic fields are calculated in the LIOV code by










employing either the Transmission Line Model (Uman and McLain, 1975), the Modified

Transmission Line Model with exponential current decay (Nucci et al., 1990), or the

Modified Transmission Line Model with linear current decay (Rakov and Dulzon, 1987).

The Cooray-Rubinstein formula (Cooray, 1992; Rubinstein, 1996) is implemented in the

LIOV code to calculate the horizontal electric field from the vertical electric field derived

from the models and to take into account the propagation effects of the LEMP over lossy

ground. The calculated electric and magnetic fields are then coupled to the line including

appropriate distributed sources in the transmission line theory.

The LIOV-EMTP96 code has been successfully tested with experimental data at

the Swiss Federal Institute of Technology in Lausanne using a NEMP (Nuclear

Electromagnetic Pulse) simulator and reduced-scale models of single and multi-

conductor lines (Paolone et al., 2000).

2.7 Experimental Studies of Lightning Strike Interaction with Power
Lines

This section reviews some important experimental studies of natural and rocket-

triggered lightning interaction with power distribution and transmission lines. Studies of

both direct lightning strike interaction (lightning current is inj ected into one of the line

conductors) and nearby lightning strike interaction (lightning current is inj ected into

ground at a certain distance from the power line) are considered. The studies are listed in

chronological order.

2.7.1 Japanese Study of Nearby Rocket-Triggered Lightning Strike Interaction
with a Test Distribution Line (1977-1985)

Literature: Horii (1982), Horii and Nakano (1995)

The nearby lightning strike interaction with a test distribution line was studied at

the Kahokugata site in Japan from 1977 to 1985. Currents from rocket-triggered lightning









strikes were injected into ground as close as 77 m from a 9 m high wire that simulated a

phase conductor of a distribution line. Horii (1982) found a linear correlation between the

induced voltage peak values and the corresponding lightning return stroke current peak

values. Adding a grounded wire 1 m above the phase conductor resulted in the reduction

of the induced voltage by about 40%.

2.7.2 South African Study of Direct and Nearby Natural Lightning Strike
Interaction with a Test Distribution Line (1978, 1979)

Literature: Eriksson et al. (1982)

The interaction of nearby and direct natural lightning with an 11l-kV, three phase

test distribution lines was studied in South Africa in 1978 and 1979. The experiment is

discussed in Eriksson et al. (1982). The study was part of a j oint proj ect between the

Electricity Supply Commission (Johannesburg, South Africa) and the National Electrical

Engineering Research Institute (Pretoria, South Africa). The test line was 9.9 km long,

with the western end of the line grounded to a buried counterpoise and the eastern end

open-circuited. Silicon carbide gapped arresters or gapless MOV arresters were installed

on all three phase at 1 km intervals.

Data from 12 direct lightning strikes and 269 nearby lightning strikes were

collected. The largest arrester currents induced by nearby lightning were measured at the

east end of the line and were approximately twice the amplitude of the currents through

the arresters located remote from the ends. The induced arrester current never exceeded 1

kA. The maj ority of the voltages measured at the midpoint were induced by nearby

lightning lowering negative charge to ground. The induced voltages were unipolar with a

positive polarity.









2.7.3 DoE Study of Direct Natural Lightning Strike Interaction with Distribution
Lines (1978)

Literature: Schneider and' Stillwell (1979)

Surges on distribution lines due to natural lightning strikes to the line were

investigated in 1978 as part of a proj ect funded by the U. S. Department of Energy (DoE).

The experiment is discussed in Schneider and Stillwell (1979). Fourteen battery-operated

lightning surge recorders with current sensors (Rogowski coils) were installed on

distribution lines in the St. Petersburg, Florida, area to measure arrester currents. Data

from two lightning strikes that attached to a 7.62-kV, single-phase, overhead distribution

line were recorded. The exact strike locations on the line are not known. One stroke that

lowered negative charge to ground was recorded during the first lightning strike. The

arrester discharge current had a peak value of 15 kA, a rise-time of about 2 Cls, and half-

peak width of about 36 Cls. Three strokes that lowered positive charge to ground were

recorded during the second lightning strike. The peak value/risetime/half-peak width of

the arrester discharge current was 42 kA/5.6 Cls/60 Cls for the first stroke, 32 kA/1 Cls/9 Cls

for the second stroke, and 40 kA/1 Cls/5 Cls for the third stroke.

2.7.4 DoE Study of Nearby Natural Lightning Strike Interaction with Distribution
Lines (1979)

Literature: M~aster et al. (1984, 1986)

The interaction of nearby natural lightning with an unenergized test distribution

line located near Wimauma, Florida, southeast of Tampa, was studied in 1979. The test

distribution line was a single phase 7.62 kV line of 460 m length and consisted of 6 poles.

The neutral conductor was located under the phase conductor and grounded at both line

ends. The phase conductor was open-circuited at both ends.









Lightning-induced voltages at one line end and the vertical electric field of the

lightning (the electric field sensor being located 160 m from the induced voltage sensor)

were measured. Data from over 100 first return strokes and over 200 subsequent strokes

were collected. The lightning strike locations were determined by triangulation using a

network of television cameras and thunder ranging. The majority of the strike locations

were at distances ranging from 4 to 12 km from the line.

Master et al. (1984) observed induced voltages of negative or positive polarity-

the polarity apparently depending on the lightning strike location. Based on this

observation Master et al. argue that the induced voltage cannot be due only to coupling of

the vertical component of the lightning' s electric field and that the contribution from

coupling of the horizontal electric field component to the line must be significant.

2.7.5 Japanese Study of Nearby Natural Lightning Strike Interaction with a
Distribution Line (1980-1988)

Literature: Yokoyamna et al. (1983, 1986, 1989)

The interaction of nearby natural lightning with an unenergized test distribution

line located at the Fukui steam power station in Japan was studied from 1980 to 1988.

The test distribution line was 820 m long and consisted of 17 poles (numbered No. 1

through No. 17) with a distance of about 50 m between adj acent poles. Natural lightning

attached to the 200 m tall chimney which was located at 200 m perpendicular distance

away from the line. The lightning current at the chimney and induced voltages at poles

No. 5, No. 8, No. 11, and No. 14 were measured.

A total of 90 voltage waveforms were recorded from 1980 through 1988 with the

causative lightning stroke current being available for 32 strokes. The polarity of the

induced voltages is discussed in Yokoyama et al. (1989). They concluded that except for









unusual conditions lightning-induced voltages are unipolar. The 26 return strokes that

lowered positive charge to ground induced a negative voltage and the 6 return strokes that

lowered negative charge to ground induced a positive voltage. Furthermore, Yokoyama et

al. observed that the induced voltage is largest at pole No. 5 which is closest to the

lightning strike location. The induced voltage peaks at poles No. 8, No. 11, and No. 14

are about 77%, 52%, and 33%, respectively, of the voltage peaks measured at pole No. 5.

2.7.6 DoE Study of Nearby Natural and Rocket-triggered Lightning Strike
Interaction with Distribution Lines (1985, 1986)

Literature: Georgiadis et al. (1992), Rubinstein et al. (1994), Rachidi et al. (1997b)

The University of Florida lightning research group studied the interaction of nearby

rocket-triggered lightning with an unenergized three-phase test distribution line at the

NASA Kennedy Space Center (KSC). The top phase conductor (10 m above ground) of

the 448 m long line was terminated at both ends of the line in the line' s characteristic

impedance of about 600 02. The other two phases were open-circuited at the ends.

Induced voltages on the top phase conductor at each line end and the electric/magnetic

field at 500/580 m from the lightning were measured. Data from 3 triggered lightning

flashes containing 11 strokes lowering negative charge to ground and triggered 20 m

from one line end were collected. Two types of induced voltages with almost an equal

number of occurrences were measured: (1) Oscillatory voltages with peak values ranging

from tens of kilovolts to about 100 kV and (2) impulsive voltages that were nearly an

order of magnitude larger than the oscillatory voltages (Rubinstein et al., 1994). The

oscillations in the former voltage type were attributed to multiple reflections at the line

ends. Both types were observed to occur for different strokes within a single lightning

discharge. Attempts to model the oscillatory voltages using the time-domain coupling









model of Agrawal et al. (1980) and the measured return stroke electric fields were

moderately successful (Rubinstein et al., 1994). The modeling results were improved by

including the electric Hield of the dart leader into the model (Rachidi et al., 1997b).

Attempts to reproduce the impulsive voltages failed (Rubinstein et al., 1994). Rubinstein

et al. suggested that the impulsive voltages were attributable to the presence of electrical

breakdown in the measuring system.

In a separate experiment on the same line at KSC induced voltages at both line ends

and electric/magnetic Hields due to natural lightning at distances beyond about 5 km were

measured. Georgiadis et al. (1992) found good agreement of the waveshapes of measured

induced voltages and induced voltages calculated using the coupling model of Agrawal et

al. (1980). The calculated voltages were generally larger than the measured waveforms.

This result was attributed to the fact that the measured Hields were accurate whereas the

Shields coupling to the lines were shielded by trees.

2.7.7 EPRI Study of Direct and Nearby Natural Lightning Strike Interaction with
Distribution Lines (1987-1990)

Literature: Barker et al. (1993), Fernand'ez et al. (1999)

The interaction of nearby and direct natural lightning with operating 13 kV

distribution lines was studied from 1987 to 1990. Up to 71 lightning transient recorders

were installed on distribution lines in Florida in 1987 and 1988 and 75 transient recorders

were installed on lines in Tennessee, Colorado, and New York in 1989 and 1990. The

transient recorders measured either arrester voltages or currents.

Voltage peaks of less than 17 kV were recorded for 95% of the 953 lightning-

caused arrester voltages. The largest recorded arrester voltage peak was 28 kV. Barker et

al. reported that only one of the measured voltages was large enough to be positively









attributed to a direct strike to the line within one pole span of the transient recorder. The

other voltages were apparently either induced by nearby lightning strikes or due to direct

lightning strikes at several spans from the transient recorder.

Current peaks of less than 2 kA were recorded for 95% of the 3 57 lightning caused

arrester currents. Three unusual large arrester currents (2 of them occurring in the same

flash) with current peaks above 10 kA were recorded. The largest arrester current peak

recorded without saturation was 28 kA. Barker et al. measured one long-duration slowly

decaying current following the main stroke current. This event saturated the transient

recorder at 6 kA for the first 75 Cls. The non-saturated tail of the current had an average

value of about 2 kA for 2 ms. Fernandez et al. (1999) argue in response to a reviewer' s

comment that the long-duration current reported in Barker et al. shows that a single

arrester can take an appreciable fraction of the low-frequency components of lightning

currents and therefore the low-frequency components do present a threat to arresters .

2.7.8 Japanese Study of Direct Natural and Rocket-triggered Lightning Strike
Interaction with a Test Transmission Line (1987-1996)

Literature: Horii and Nakano (1995), Matsumoto et al. (1996), M~otoyama et al. (1998),
Kobaya~shi et al. (1998)

The interaction of direct natural and rocket-triggered lightning with an unenergized

275 kV test transmission line located on the top of the Okushishiku mountain (930 m

above sea level) in Japan was studied from 1987 to 1996. The Okushishiku test

transmission line had 6 phase conductors of 1.6 km length that were supported by 6 steel

towers (numbered No. 29 through No. 34) of 60 m height and one overhead ground wire

of 2.1 km length that was supported by the same 6 steel towers and one additional tower

SThe opposing view is that the low-frequency components are shared uniformly among many line arresters
and therefore the energy absorption of a single arrester due to these components is negligible.










(No. 28 located 505 m from No. 29) of 85 m height. From 1987 to 1993 154-kV MOV

arresters were installed on each phase. The arresters were not installed for the 1994 to

1996 experiments. The line was terminated in its characteristic impedance of 500 0Z at the

No. 29 tower, which was located at 182 m from the struck No. 30 tower. At the other end

the phase conductors were connected to the grounded metallic crossarm.

From 1993 to 1996 currents from 8 rocket-triggered and 2 natural lightning strikes

were injected into the 4-m long lightning rod installed on the top of the No. 30 tower. The

No. 30 tower was instrumented to measure the lightning currents, overhead ground wire

currents, currents to the 4 grounded tower legs, arrester currents, and insulator string

voltages.

The positive lightning peak currents measured for the 8 triggered strikes ranged

from 27 kA to 102 kA and were 132 kA and 159 kA for the two natural strikes. The

maximum arrester peak current and voltage measured were 3 kA and 293 kV,

respectively (Matsumoto et al., 1996). In 1994, a back flashover was observed on the

line, and a peak voltage of approximately 2.5 MV was measured across the string

insulator where the back flashover occurred (Motoyama et al., 1998).

It was found that the currents to the 4 tower legs were different, which presumably

can be attributed to the different ground impedance of the individual legs. Furthermore, it

was observed that the high-frequency components of the inj ected lightning currents

primarily flow to the ground at the struck tower legs while the low-frequency

components tend to flow to the other grounds through the overhead ground wire (Horii

and Nakano, 1995) .









2.7.9 EPRI Study of Direct and Nearby Rocket-triggered and Natural Lightning
Strike Interaction with Distribution Lines (1993-1996)

Literature: Uman and Rakov (1995), Barker et al. (1996), Fernandez (1997), Fernandez
et al. (1997a, 1997b, 1998a, 1998b, 1998c, 1999), Uman et al. (1997), Mata et al. (1998)

The interaction of direct and nearby rocket-triggered lightning and nearby natural

lightning with an unenergized test distribution line was studied at a research facility at

Camp Blanding, Florida by Power Technologies, Inc. (PTI) from 1993 to August 1994

and by the University of Florida research group from September 1994 to 1996. The

research facility, which is the same research facility where the data presented and

analyzed in this dissertation were collected, was constructed by PTI in 1993 under the

funding and direction of EPRI and named the "International Center for Lightning

Research and Testing" (ICLRT) in 1995. UF took responsibility for operating the facility

in 1994. The 730 m long line had 15 poles, a single phase conductor at a height of about

7.5 m, which was terminated at both line ends in the lines characteristic impedance of

about 500 02, and a neutral conductor located 1.8 m under the phase conductor. The

critical flashover overvoltage (CFO) of the line was about 500 kV.

During the 1993 experiment reported in Barker et al. (1996) the neutral conductor

was grounded at poles 1, 9, and 15, and lightning was triggered at distance of 145 m from

the line. The lightning currents of 30 flashes containing 63 strokes and the resulting

voltages induced on the line were recorded. The largest voltage peak values Vpeak WeTO

measured at pole 9 and ranged from 8 to 100 kV. These voltage peaks were typically

twice the value of the voltage peaks measured at the end poles, and showed strong linear

correlation with the return stroke current peaks Ipeak which ranged from 4 to 44 kA

(correlation coefficient: 0.938, regression equation: Vpeak=2.24*Ipeak). Barker et al. found










a weaker correlation between return stroke current rate of rise and peak voltages

(correlation coefficient: 0.75) and little or no correlation between return stroke current

rise time and peak voltages (correlation coefficient: 0.28).

During the 1994 experiment reported in Uman and Rakov (1995) and Uman et al.

(1997) the line was connected to an underground distribution system through a

transformer at pole 9. Arresters were installed at poles 8, 9, and 10 and the neutral

conductor was grounded at pole 1, 9, and 15. Two flashes were triggered and inj ected

into the phase conductor between poles 9 and 10 although lightning currents were only

measured for the first flash (a flash containing 4 strokes followed by many M-

components). Currents on the line were obtained at the ground connection at pole 9, and

arrester voltages on the line were obtained at pole 9.

During the 1995 and 1996 experiments reported by Fernandez (1997), Fernandez et

al. (1997a, 1997b, 1998a, 1998b, 1998c, 1999), and Uman et al. (1997) the interaction of

direct and nearby natural and rocket-triggered lightning with the test distribution line was

studied. During the two years 38 lightning flashes that contained return strokes were

triggered-28 of these were classical triggers and 10 were altitude triggers Thirteen of

these flashes were inj ected into either the phase conductor or the neutral conductor of the

test line. Eighteen flashes were inj ected into ground at distances ranging from 20 m to

155 m and 7 flashes struck ground at various locations, some determined and other

undetermined locations. Data from 6 natural lightning events on or very near the facility


SAn altitude trigger is caused by a trailing wire that is not continuous but either deliberately interrupted or
breaks during the ascend of the rocket, resulting in a grounded wire section close to ground and an
ungrounded wire section above the grounded wire. The lightning current path is established by a bi-
directional leader (an upward positive and downward negative leader) from the ungrounded wire section as
opposed to a unidirectional leader (an upward positive leader) from the tip of the wire in classical rocket-
triggered lightning.




Full Text

PAGE 1

DIRECT AND NEARBY LIGHT NING STRIKE INTERACTION WITH TEST POWER DISTRIBUTION LINES By JENS DANIEL SCHOENE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007

PAGE 2

Copyright 2007 by Jens Daniel Schoene

PAGE 3

iii ACKNOWLEDGMENTS I would like to thank Dr. M.A. Uman and Dr. V.A. Rakov for their guidance, advice, and support. Furthermore, I would like to thank Dr. C.T. Mata and A.G. Mata for their work on the FPL experiment in 1999, results of which were used in this dissertation, and K.J. Rambo who was alwa ys there to get things done . The experimental data presented in this dissertat ion were obtained in a team effort. The team players involved were M.V. Stapleton, J.E. Jerauld, Dr. D.M. Jordan, G.H. Schnetzer, B.D. Hanley, J.Howard, B.D. DeCarlo, R. Sutil, A. Guarisma, G. Bronsted, and A. Mata. I also wish to thank Dr. C.A. Nucci, Dr. F. Rachidi, Dr. M. Paolone, and Dr. E. Petrache for their invaluable help with the modeling of results from the nearby strike experiments and for the pleasure of working with them at Camp Blanding duri ng the summers of 2002 and 2003. Last but not least, I w ould like to thank my wife Gi sele and my son Gabriel for their patience and good behavior, respectively. The research reported in this thesis wa s funded in part by the Florida Power and Light Company, the Lawrence Liverm ore National Laboratory, and NSF.

PAGE 4

iv TABLE OF CONTENTS page ACKNOWLEDGMENTS................................................................................................. iii TABLE OF CONTENTS................................................................................................... iv LIST OF TABLES............................................................................................................. ix LIST OF FIGURES......................................................................................................... xiii ABSTRACT................................................................................................................... xxix CHAPTER 1 INTRODUCTION............................................................................................. .......... 1 1.1 1999 FPL Experiment, Direct L ightning Strike Interaction w ith a Horizontally-configured Distribution Line........................................................ 3 1.2 2000 FPL Experiment, Direct Lightning Strike Interaction w ith a Horizontally-configured Distribution Line........................................................ 5 1.3 2001 FPL Experiment, Direct Lightning Strike Interaction w ith a Vertically-configured Distribution Line............................................................8 1.4 2002 FPL Experiment, Direct Lightning Strike Interaction w ith a Vertically-configured Distribution Line............................................................9 1.5 2002 FPL Experiment, Nearby Lightni ng Strike Interaction with a Vertically-configured Distribution Line.......................................................... 11 1.6 2003 FPL Experiment, Direct Lightning Strike Interaction w ith a Vertically-configured Distribution Line..........................................................12 1.7 2003 FPL Experiment, Nearby Lightni ng Strike Interaction with a Vertically-configured Distribution Line.......................................................... 13 1.8 2004 FPL Experiment, Direct Lightning Strike Interaction w ith a Vertically-configured Distribution Line with Overhead Ground Wire........... 15 1.9 2005 Lawrence Livermore Experi ment, Induced Currents.............................. 16 1.10 Summ ary of Original Contributions................................................................17 2 LITERATURE REVIEW.......................................................................................... 20 2.1 Cloud Formation and Electrification............................................................... 20 2.1.1 Formation of a Cumulonimbus........................................................... 21 2.1.2 Electrical Structure of a Cumulonimbus............................................. 24 2.1.3 Electrification of a Cumulonimbus..................................................... 25 2.2 Natural and Rocket-triggered Lightning.......................................................... 27 2.2.1 Lightning Discharges between Cloud and Ground.............................28 2.2.2 Rocket-triggered Lightning.................................................................33 2.2.3 Return Stroke Current......................................................................... 37 2.3 Transmission-line Type Return Stroke Models............................................... 39

PAGE 5

v 2.4 Distribution Line De sign Parameters............................................................... 42 2.4.1 Insulation Strength of Distribution Lines ........................................... 43 2.4.2 Overhead Ground Wires..................................................................... 44 2.4.3 Metal-Oxide Arresters........................................................................44 2.5 Modeling Direct Strikes to Power Distribution Line....................................... 49 2.5.1 History of the Electromagne tic Transient Program............................ 50 2.5.2 EMTP Current Sources....................................................................... 51 2.5.3 EMTP Arrester Model........................................................................51 2.5.4 EMTP Transmission Line Models...................................................... 52 2.5.5 Leads Connecting the Neutra l Conductor to Ground Rods................ 53 2.5.6 Ground Rod Model.............................................................................53 2.6 Modeling Nearby Strikes to Power Distribution Lines.................................... 56 2.6.1 Calculation of Lightning -induced Overvoltages................................. 56 2.6.2 Testing of Lightning-indu ced Overvoltage Models............................ 59 2.6.3 LIOV-EMTP96 Code......................................................................... 60 2.7 Experimental Studies of Lightning Stri ke Interaction with Power Lines........ 61 2.7.1 Japanese Study of Nearby Rocket-Triggered Lightning Strike Interaction with a Test Distribution Line (1977)......................61 2.7.2 South African Study of Direct and Nearby Natural Lightning Strike Interaction with a Test Distribution Line (1978, 1979)......................62 2.7.3 DoE Study of Direct Natural Lightning Strike Interaction with Distribution Lines (1978).................................................................... 63 2.7.4 DoE Study of Nearby Natural Lightning Strike Interaction with Distribution Lines (1979).................................................................... 63 2.7.5 Japanese Study of Nearby Natural Lightning Strike Interaction with a Distribution Line (1980)........................................................ 64 2.7.6 DoE Study of Nearby Natural and Rocket-triggered Lightning Strike Interaction with Distribution Lines (1985, 1986).................... 65 2.7.7 EPRI Study of Direct and N earby Natural Lightning Strike Interaction with Distribution Lines (1987)............................... 66 2.7.8 Japanese Study of Direct Natura l and Rocket-triggered Lightning Strike Interaction with a Te st Transmission Line (1987)......... 67 2.7.9 EPRI Study of Direct and Nearby Rocket-triggered and Natural Lightning Strike Interaction with Distribution Lines (1993).... 69 2.7.10 Japanese Study of Direct Natural Li ghtning Strike Inte raction with a Test Distribution Line (1999).......................................................... 72 2.7.11 FPL Study of Direct and Nearby Rocket-triggered Strike Interaction with Distribution Lines (1999).............. .................75 3 EXPERIMENTAL FACILITY...................................................................................76 3.1 Rocket Launchers.............................................................................................79 3.1.1 Stationary Launcher............................................................................ 79 3.1.2 Mobile Launcher................................................................................. 81 3.2 Test Distribution Lines....................................................................................82 3.2.1 Horizontally-configured Line, 1999 and 2000 ................................... 82 3.2.2 Vertically-configured Line, 2001 through 2003................................. 84

PAGE 6

vi 3.2.3 Vertically-configured Line w ith Overhead Ground Wire, 2004......... 88 3.3 Grounding of the Test Distribution Lines........................................................ 90 3.3.1 1999 Experim ent, Grounding............................................................. 91 3.3.2 2000 Experiment, Grounding............................................................. 91 3.3.3 2001 Experiments, Grounding............................................................92 3.3.4 2002 and 2003 Experiments, Grounding............................................ 93 3.3.5 2004 Experiment, Grounding............................................................. 93 3.4 Arresters on the Test Distribution Lines..........................................................94 3.4.1 1999 Experime nt, Arresters................................................................94 3.4.2 2000 Experiment, Arresters................................................................94 3.4.3 2001 Experiment, Arresters................................................................95 3.4.4 2002 Experiment, Arresters................................................................95 3.4.5 2003 Experiment, Arresters................................................................96 3.4.6 2004 Experiment, Arresters................................................................96 3.5 Line Terminators on the Test Distribution Lines............................................. 97 3.5.1 1999 Experiment, Line Terminators................................................... 98 3.5.2 2000 Experiment, Line Terminators................................................... 98 3.5.3 2001 Experiments, Line Terminators....................................... 98 3.6 Instrumentation................................................................................................ 99 3.6.1 Current Sensors ...................................................................................99 3.6.2 Voltage Dividers................................................................................ 102 3.6.3 Fiber Optic Link................................................................................ 105 3.6.4 Anti-aliasing Filter............................................................................ 107 3.6.5 Data Recording Equipment...............................................................107 3.6.6 Wireless Control System.................................................................. 110 4 EXPERIMENTAL CONFIGUR ATIONS AND RESULTS.................................... 112 4.1 Description and Terminology of Measured Parameters ................................ 112 4.2 1999 Experim ent, Horizontally-configured Line........................................... 113 4.3 2000 Experim ent, Horizontally-configured Line........................................... 115 4.4 2001 Experim ent, Vertically-configured Line............................................... 117 4.5 2002 Experime nt, Vertically-configured Line............................................... 119 4.6 2003 Experime nt, Vertically-configured Line............................................... 121 4.7 2004 Experiment, Vertically-configured Line with Overhead Ground W ire 124 4.8 2005 Experim ent, Induced Currents..............................................................125 4.9 Triggering Results .......................................................................................... 129 5 DATA PRESENTATION......................................................................................... 133 5.1 2003 FPL Direct Strike Experiment: Phase-to-neutral and Ground Currents During Stroke FPL0312-5...............................................................134 5.2 2004 FPL Direct Strike Experiment: Com plete Data Set for Stroke FPL0403-2 .....................................................................................................138 5.3 2003 FPL Nearby Strike Experiment: Lightning Currents Traversing Soil And Entering Line Grounding.......................................................................141

PAGE 7

vii 5.4 2005 Lawrence Livermore Experimen t: Induced Currents on a Buried Counterpoise and Vertical Wire ..................................................................... 150 6 DATA ANALYSIS, MODELING, AND DISCUSSION........................................ 157 6.1 Current Consistency Check............................................................................160 6.1.1 Consistency of Injected Curre nt and Total Ground Current............. 161 6.1.2 Consistency of Injected Current and Total Struck-phasetoneutral Current..............................................................................................163 6.1.3 Consistency of High and Low Lightning Currents.................... 165 6.1.4 Consistency of Lightning Cu rrents from the 2000 Experiment........ 169 6.1.5 Consistency of Current from the 2004 Experiment.......................... 172 6.2 Characterization of the Light ning Return Stroke C urrent.............................. 175 6.2.1 Statistical Data of Lightning Retu rn Stroke Current Parameters......176 6.2.2 Statistical Distributions of Li ghtning Return Stroke Current Parameters......................................................................................... 178 6.2.3 Discussion of Lightning Return Stroke Current Parameters............ 181 6.2.3.1 Return stroke current peaks...............................................181 6.2.3.2 Return stroke current 10-90% risetime.............................184 6.2.3.3 Return stroke current half-peak width...............................186 6.2.3.4 Return stroke charge transfer............................................. 187 6.3 Arrester Disconnector Op eration and Flashovers.......................................... 190 6.3.1 Discussion of Arrester Disconnector Operation............................... 193 6.3.2 Discussion of Flashovers.................................................................. 196 6.4 Measured Lightning Current Division on the Test Distribution Lines.......... 198 6.4.1 Measured Phase-to-neutral and Ground Current Division............... 198 6.4.2 Discussion of the Measured Lightning Current Division................. 205 6.5 Modeled Lightning Current Divisions on the Test Distribution Lines.......... 210 6.5.1 Modeled Lightning Current Divisi on on the Horizontal Line..........211 6.5.2 Modeled Lightning Current Divi sion on the Vertical Line.............. 220 6.5.3 Simple Model of the Lightning Current Division............................. 235 6.5.4 Discussion of the Modeled Li ghtning Current Division...................241 6.6 Estimation of the Arrester-absorbed Energy.................................................. 245 6.7 Modeling of Induced Currents on the Test Line due to Nearby Lightning ...248 6.8 Lightning C urrents Traversing Soil and Entering Line Grounding............... 252 6.8.1 Analysis of Lightning C urrents Traversing Soil............................... 252 6.8.2 Discussion of Lightning Currents Traversing Soil...........................255 6.9 Induced Currents on a Buried Counterpoise.................................................. 260 7 SUMMARY OF ORI GINAL RESULTS.................................................................263 7.1 Data Consistency...........................................................................................263 7.1.1 Data Consistency during 2000 Horizontal Line Experim ent............264 7.1.2 Data Consistency during 2002, 2003, and 2004 Vertical Line Experiments...................................................................................... 265 7.2 Characterization of the Light ning Return Stroke C urrent.............................. 266 7.3 Arrester Disconnector Operation and Arrester Energy Absorption............... 267

PAGE 8

viii 7.4 Flashover Occurrence.................................................................................... 268 7.5 Phase-to-neutral Current Division................................................................. 269 7.6 Neutral-to-ground Current Division..............................................................270 7.7 Lightning Currents Traversing Soil a nd Entering Vertical Line Grounding 271 7.8 Induced Currents on Distribution Lines......................................................... 272 7.9 Induced Currents on a Buried C ounterpoise and Vertical Wire.................... 272 8 RECOMMENDATIONS FOR FUTURE RESEARCH........................................... 274 APPENDIX A MEASUREMENT SETTINGS................................................................................ 277 B DATA PRESENTATION......................................................................................... 296 C RETURN STROKE CURRENT PARAMETERS................................................... 365 D FLASHOVERS AND DI SCONNECTOR OPERATION........................................373 LIST OF REFERENCES .................................................................................................379 BIOGRAPHICAL SKETCH...........................................................................................391

PAGE 9

ix LIST OF TABLES Table page 3-1: Experimental configurations used in 1999 through 2004 experiments. Pole numbers are identified in Figure 3-1........................................................................ 78 3-2: Conductor p lacement and specificati ons for the vertically-configured test distribution line........................................................................................................ 84 3-3: Conductor placement and specifications for the ve rtically-configured test distribution line with ov erhead ground wire............................................................88 3-4: Grounding resistances in ohms fo r the horizontally-configured line tested during the 1999 experim ent...................................................................................... 91 3-5: Grounding resistances in ohm s fo r the horizontally-configured line tested during the 2000 experiment...................................................................................... 91 3-6: Grounding resistances in ohm s of the vertically-configured line tested during the 2001 experiment. The number of ground rods is given in parenthesis.............. 93 3-7: Measured and theoretically-deriv ed grounding resistances in ohm s for the vertically-configured line. The number of ground rods is given in parenthesis......93 3-8: Grounding resistances in ohm s measured on 7/12/2004 for the verticallyconfigured line tested during the 2004 experiment. The number of ground rods is given in parenthesis...................................................................................... 93 3-9: VI-characteristics of the Cooper Power Systems Ultra SIL Housed VariSTAR Heavy Duty and Ohio-Brass PDV-100 arresters......................................................94 3-10: 1999 experime nt, arresters used on th e horizontally-configured test line................ 94 3-11: 2000 experiment, arresters used on th e horizontally-conf igured test line................ 95 3-12: 2001 experiment, arresters used on th e vertically-configured test line. ...................95 3-13: 2002 experiment, arresters used on th e vertically-configured test line. ...................96 3-14: 2003 experiment, arresters used on th e vertically-configured test line. ...................96 3-15: 2004 experiment, arresters used on the vertically-configured test line with overhead ground wire. ..............................................................................................97 3-16: Termination resistors during the 2000 experiment (horizontally-configured distribution line)....................................................................................................... 98 3-17: Termination resistors duri ng the 2001 through 2004 experim ents........................... 99

PAGE 10

x 3-18: Specifications of the Pearson El ectronics, Inc. current transformers..................... 101 3-19: Specification of the T& M Research Products, Inc. current viewing resistors....... 102 3-20: Specifications f or the capacitive-compensated voltage dividers used in the 1999 and 2000 experiments....................................................................................103 3-21: Specifications for each of the four loops of the m agnetic-flux-compensated voltage dividers used in the 1999 and 2000 experim ents....................................... 105 3-22: Recording devices used durin g the 1999 through 2004 experiments. The number of recording devices used f or a particular year is given............................ 108 4-1: 1999 experiment, references to the measurement settings for all rockettriggered lightning strikes. All strikes triggered to the test line in 1999 contained return strokes......................................................................................... 115 4-2: 2000 experiment, references to the measurem ent settings for all rockettriggered lightning events. Event IDs in bold printing denote flashes with return strokes and event IDs in italic printing denote flashes without return strokes..................................................................................................................... 117 4-3: 2001 experiment, references to the measurem ent settings for all rockettriggered lightning events. Event IDs in bold printing denote flashes with return strokes and IDs in italic printing denote flashes without return strokes...... 119 4-4: 2002 experiment, references to the measurem ent settings for all rockettriggered lightning events. Event IDs in bold printing denote flashes with return strokes and IDs in italic printing denote flashes without return strokes...... 121 4-5: 2003 experiment, references to the measurem ent settings for all rockettriggered lightning events. Event IDs in bold printing denote flashes with return strokes and event IDs in italic printing denote flashes without return strokes..................................................................................................................... 123 4-6: 2004 experiment, references to the measurem ent settings for all rockettriggered lightning events. The event ID in bold printing denotes a flash with return strokes and the event ID in italic printing denot es a flash without return strokes..................................................................................................................... 125 4-7: Successful triggering events during the 1999 through 2004 experiments.............. 130 4-8: Successful triggering even ts during the 2005 experiment...................................... 132 6-1: Return stroke current sta tistics for the 1999 through 2004 experim ents................ 177 6-2: Com parison of return stroke statistics.................................................................... 182

PAGE 11

xi 6-3: Disconnector op erations and flashovers during the 2000 experiments........ 191 6-4: Statistical inform a tion on the arrester and ground current equilibration times...... 200 6-5: Comp arison of peak values and charge transfers of lightning currents and currents entering the line through the pole 15 grounding. The charge transfers were obtained by numerically integra ting the m easured currents over 1 ms (return stroke current) an d 10 ms (initial continuous current) time intervals........ 253 6-6: Comparison of 10-90% risetimes a nd half-peak widths of lightning currents and currents entering the line through the pole 15 grounding................................ 255 A-1: 2002 measurem ent settings for flash FPL0236 (8/18/2002)................................. 279 A-2: 2002 measurem ent settings for flash FPL0240 (8/27/2002)................................. 280 A-3: 2002 measurem ent settings for flashes FPL0244, FPL0245, and FPL0246 (9/13/2002).............................................................................................................281 A-4: 2003 measure ment settings for flashes FPL0301, FPL0302, and FPL0303 (6/30/2003).............................................................................................................283 A-5: 2003 measure ment settings fo r flashes FPL0305 and FPL0306 (7/6/2003)......... 284 A-6: 2003 measurem ent settings fo r flashes FPL0312 and FPL0314 (7/13/2003)....... 285 A-7: 2003 measurem ent settings fo r flashes FPL0315 and FPL0317 (7/14/2003)....... 286 A-8: 2003 measurem ent settings for flash FPL0321 (7/18/2003)................................. 288 A-9: 2003 measurem ent settings fo r flashes FPL0329 and FPL0331 (7/22/2003)....... 289 A-10: 2003 measurem ent settings for flash FPL0336 (8/2/2003)................................... 290 A-11: 2003 measurement settings for flash FPL0341 (8/7/2003)................................... 291 A-12: 2003 measurement settings for flashes FPL0342 and FPL0345 (8/11/2003).......292 A-13: 2003 measure ment settings for flashes FPL0347, FPL0348, and FPL0350 (8/15/2003).............................................................................................................293 A-14: 2004 Measurement settings f or flashes FPL0402 and FPL0403 (7/24/2004).......295 C-1: Return stroke stat istics for the 1999 experiment................................................... 365 C-2: Return stroke stat istics for the 2000 experim ent................................................... 366 C-3: Return stroke stat istics for the 2001 experim ent................................................... 367

PAGE 12

xii C-4: Return stroke statistics for the 2002 direct strike experiment............................... 368 C-5: Return stroke statistics fo r the 2002 nearby strike experiment.............................. 369 C-6: Return stroke statistics fo r the 2003 direct strike experiment............................... 370 C-7: Return stroke statistics fo r the 2003 nearby strike experiment.............................. 371 C-8: Return stroke stat istics for the 2004 experim ent................................................... 372 D-1: Flashovers and disconnector op eration during the 2000 horizontal line experim ent.............................................................................................................. 375 D-2: Flashovers and disconnector operation during the 2001 vertical line experim ent.............................................................................................................. 376 D-3: Flashovers and disconnector operation during the 2002 vertical line experim ent.............................................................................................................. 377 D-4: Flashovers during the 2003 vert ical line experim ent. No disconnector operated during this experiment. ............................................................................ 378

PAGE 13

xiii LIST OF FIGURES Figure page 2-1: Small thundercloud over a train a nd a power line. The picture was taken in Finland at the Russian border. Th e thundercloud produced only a few lightning strikes........................................................................................................21 2-2: Formation of sea/land-breeze thundersto rms........................................................... 22 2-3: Cold front moves under warm front resulting in the formation of cumulonimb us (frontal storm).................................................................................. 23 2-4: Electrical structure inside a cumulonimbus.............................................................. 24 2-5: Precipitation th eory (left) and convection theory (right). ......................................... 26 2-6: Discharge types for a thundercloud.......................................................................... 27 2-7: Sim plified drawing of four discharges between cloud and ground.......................... 28 2-8: Drawings illustrating som e of the various processes comprising a negative cloud-to-ground lightning flash................................................................................29 2-9: Rocket-triggered lightning in Camp Blanding, Florida (Flash U9910)................... 33 2-10: Sequence of events in classi cal rocket-triggered lightning......................................34 2-11: Typical triggered lightning return stroke current waveform measured at the channel base (stroke FPL0315-2). The following return stroke current param eters are illustrated: a) peak va lue, b) 10-90% risetime, c) half-peak width, and d) charge transfer....................................................................................38 2-12: Current versus time wavefo rms specified by TL model at ground (z = 0) and at two heights z1 and z2 ...........................................................................................41 2-13: Current versus height z above ground at time t = t1 for the TL model.................... 42 2-14: Typical varistor VI-character istic plotted on a log-log scale...................................46 2-15: Structure of a me tal-oxide varistor........................................................................... 47 2-16: Illustrations of gapl ess metal-oxide arresters........................................................... 48 2-17: Lum ped parameters representation of ground rods. The model consists of n RLC-sections............................................................................................................55 2-18: Measured pole 13 arrest er current and voltage dur ing a natural flash that connected to pole 6................................................................................................... 73

PAGE 14

xiv 2-19: Pole 13 arrester current during a natural lightning strike to pole 13........................ 74 3-1: Overview of the ICLRT............................................................................................ 77 3-2: Stationary tower launcher em pl oyed from 1999 through 2004 primarily used to simulate direct lightni ng strikes to the test line The intercepting structure (PVC poles and intercepting conducto r) were only used for the 2002, 2003, and 2004 experiments............................................................................................... 80 3-3: Mobile bucket-truck rocket launche r placed a few meters from one end of the vertically-configured te st distribution line. .............................................................. 81 3-4: Sketch of the hor izontally-configured line............................................................... 83 3-5: Sketch of the ve rtically-configured line te sted from 2001 through 2003. ................ 85 3-6: Vertically-configured distribution line in 2003, arrester station at pole 10. ............86 3-7: Vertically-configured distribut ion line in 2003, arre ster station with transforme r at pole 2................................................................................................ 87 3-8: Vertically-configured distribut ion line with overhead ground wire in 2004, arrester station at pole 10. ........................................................................................89 3-9: Vertically-configured distribut ion line with overhead ground wire in 2004, arrester station at pol e 10, close-up view................................................................. 89 3-10: Grounding scheme for the verti cally-configured distribution line. ..........................90 3-11: Vertically-configured distribution line in 2003, line termination at pole 15. ........... 97 3-12: Measurement schem e used during the 1999 through 2004 experiments................ 100 3-13: Magnetic-flux compensated voltage divider.......................................................... 104 3-14: Diagram of the wireless contro l system topology used during the 2002 through 2004 experiments...................................................................................... 111 4-1: 1999 test distribution line having a horizontal framing configuration with measurement points and the tw o different lightni ng strike locations identified. ... 114 4-2: 2000 test distribution line having a horizontal framing configuration with measurement points and the tw o different lightni ng strike locations identified. ... 116 4-3: 2001 test distribution line having a vertical framing configuration with measure ment points and the three di fferent lightning strike locations identified.................................................................................................................118

PAGE 15

xv 4-4: 2002 test distribution line having a vertical framing configuration with measure ment points and the four different lightning strike locations identified.................................................................................................................120 4-5: 2003 test distribution line having a vertical framing configuration with measure ment points and the four different lightning strike locations identified.................................................................................................................122 4-6: 2004 test distribution line having a vertical framing configuration and an overhead ground wire with measurement points and the lightning strike location identified. .................................................................................................. 125 4-7: Satellite image of the Internat ional Center for Lightning Research and Testing. The location of the rocket launc h facilities, test house, and induced current measurements are indicated. Add itiona lly, the location of the electric field derivative measurements (Stati ons 1, 4, 8, and 9) are shown including their distances to the vertical wire measurement in meters.................................... 126 4-8: Satellite image of the experimental site of the induced currents experiments. The objects relevant to this experiment and their locations relative to the north-west corner of the c ounterpoises are indicated.............................................127 4-9: Vertical wire shown with peak curre nt sensor cards. The length of the wire is 7 m..........................................................................................................................129 5-1: Measured currents fo r stroke F PL0312-5, 2 ms and 100 s time scales. ............... 135 5-2: Stroke FPL0312-5, injected curr ent and phase A-to-neutral currents.................... 137 5-3: Measured currents fo r stroke FPL0403-1, 2 m s and 100 s time scales. ............... 140 5-4: Individual ground currents and sum of ground currents leaving the system for stroke FPL0341-1. Displayed on a) 1 ms and b) 50 s tim e scales....................... 143 5-5: Normalized lightning current inj ected into ground 11 m from pole 15 (blue) and current injected into the line through the pole 15 ground (black) for stroke FPL0347-1. The plateau in the pole 15 ground current may be related to ground arcing at the lightni ng current injection point........................................ 144 5-6: Individual ground currents and sum of ground currents leaving the system for stroke FPL0347-1...................................................................................................144 5-7: Normalized lightning current inj ected into ground 11 m from pole 15 (blue) and current injected into the line through the pole 15 ground (black) for stroke FPL0347-2. ..................................................................................................145 5-8: Individual ground currents and sum of ground currents leaving the system for stroke FPL0347-2...................................................................................................145

PAGE 16

xvi 5-9: Normalized lightning current inj ected into ground 11 m from pole 15 (blue) and current injected into the line through the pole 15 ground (black) for stroke FPL0350-1...................................................................................................146 5-10: Individual ground currents and sum of ground currents leaving the system for stroke FPL0350-1...................................................................................................146 5-11: Normalized lightning current inject ed into ground 11 m from pole 15 (blue) and current injected into the line through the pole 15 ground (black) for the ICC of flash FPL0347. ........................................................................................... 147 5-12: Individual ground currents and sum of ground currents leaving the system for the ICC of flash FPL0347......................................................................................147 5-13: Normalized lightning current inject ed into ground 11 m from pole 15 (blue) and current injected into the line through the pole 15 ground (black) for the ICC of flash FPL0348............................................................................................ 148 5-14: Individual ground currents and sum of ground currents leaving the system for the ICC of flash FPL0348......................................................................................148 5-15: Normalized lightning current inject ed into ground 11 m from pole 15 (blue) and current injected into the line through the pole 15 ground (black) for the ICC of flash FPL0350............................................................................................ 149 5-16: Individual ground currents and sum of ground currents leaving the system for the ICC of flash FPL0350......................................................................................149 5-17: Measured channel base curren t and induced currents on the runway counterpoise (290 m from th e lightning) and the vert ical wire (300 m from the lightning) for stroke 0503-2, init iated from the m obile launcher..................... 151 5-18: Measured channel base curren t and induced currents on the runway counterpoise (50 m from the lightning) a nd the vertical wire (100 m from t he lightning) for stroke 0517-2, ini tiated from the tower launcher............................. 151 5-19: Counterpoise and vertical wire currents duri ng natural lightning flash MSE0504 striking ground approximately 300 m from the vertical wire. The spikes in the vertical wire current occu rring b efore the return stroke initiation at t = 0 are labeled.................................................................................................. 152 5-20: Spikes 1 through 5 in the vertical wire current overlaid with unfiltered and filtered dE/dt records m easured at Stations 1, 4, 8, and 9. All data are normalized and the normalization fact ors are given in the legends....................... 154 6-1: Stroke FPL0315-1, ground current consistency check........................................... 162 6-2: Stroke FPL0315-1, ground current consistency check........................................... 162

PAGE 17

xvii 6-3: Stroke FPL0315-1, phase A-to-n eutral current consistency check........................ 164 6-4: Stroke FPL0315-1, phase A-to-n eutral current consistency check........................ 165 6-5: Stroke FPL0305-2, high/lo w lightning current comparison............................ 167 6-6: Stroke FPL0312-8, high/low lightning current comparison............................ 167 6-7: Stroke FPL0336-7, high/low lightning current comparison............................ 168 6-8: Stroke FPL0403-2, (1) high and (2 ) low channel base currents, (3) sum of poles 6 through 10 ground currents and poles 6 and 10 overhead ground wire currents, and (4) pole 8 ground current.......................................................... 173 6-9: Spikes during the FPL and test-hous e experim ents. The spikes in the pole 8 ground current and the point A current we re isolated by subtracting the lowfrequency components from the total data. ............................................................ 174 6-10: Histogram of return stroke current peaks. An adjustment factor of 0.75 has been applied to the current peaks from the 2000 experime nt................................ 178 6-11: Histograms of return st roke current 10-90% risetimes...........................................179 6-12: Histogram of return str oke current half-peak widths............................................. 180 6-13: Histogram of return stroke charge tr ansfers within 1 m s. An adjustment factor of 0.75 has been applied to the charge from the 2000 experiment......................... 180 6-14: Current peaks as a function of char ge transfer within 1 ms fo r rockettriggered lightning return strokes (143 individual values and regression line are shown). The regression power equation and R2 value are given. The regression line for 89 negative first retu rn strokes in natural lightning found by Berger is also displayed. The shaded area represents an envelope that encompasses all Berger data points (onl y the outside values that confine the shaded area are shown).......................................................................................... 189 6-15: Struck-phase arrester currents with the transient mode (dark shaded area) and steady-state mode (light shaded area) indicated..................................................... 199 6-16: Stroke FPL0312-5, the individual curr ents flowing from phase A to neutral divided by the sum of all phase A-to-neutral cu rrents on a 100 s time scale. The pole 15 terminator current was no t measured and was assumed to be equal to the pole 1 terminator current.................................................................... 203 6-17: Stroke FPL0312-5, the individual curr ents flowing to ground divided by the sum of all ground currents on a 100 s time scale. The low-frequency, lowcurrent ground resistance of each of the pole grounds is given in the

PAGE 18

xviii parentheses. The percentages of the individual currents at 100 s are displayed on the right side......................................................................................204 6-18: Stroke FPL0312-5, the individual char ges flowing from phase A to neutral divided by the sum of all phase A-to-neu tral charges, on a 2 ms tim e scale. The pole 15 terminator charge was no t measured and was assumed to be equal to the pole 1 te rminator charge..................................................................... 204 6-19: Stroke FPL0312-5, the individual ch arges flowing to ground divided by the sum of all ground charges, on a 2 ms time scale. The dc ground resistance of each of the pole grounds is given in the parentheses. The percentages of the individual charges at 2 ms are displayed on the right side.....................................205 6-20: Stroke FPL0036-1, Unfiltered and filtered channel base currents and its piecewise-linear approximation. The piecew ise linear app roximation of the filtered current was used as input to the model...................................................... 212 6-21: Stroke FPL0036-1, comparison of th e overall phase C-to -neutral current divisions (left) and ground current divi sions (right) of measured currents (top), model 1 currents (center), and model 2 currents (bottom) on a 500 s time scale................................................................................................................ 215 6-22: Stroke FPL0036-1, individual comparison of all successfully measured phase C-to-neutral currents with model 1and model 2-predicte d results on a 50 s tim e scale................................................................................................................ 216 6-23: Stroke FPL0036-1, individual comparison of all successfully measured phase C-to-neutral currents with model 1and model 2-predicte d results on a 500 s tim e scale........................................................................................................... 217 6-24: Stroke FPL0036-1, indi vidual comparison of all measured ground currents with model 1 and model 2 predicted results on a 50 s time scale........................ 218 6-25: Stroke FPL0036-1, indi vidual comparison of all measured ground currents with model 1and mode l 2-predicted results on a 500 s time scale.................... 219 6-26: Stroke FPL0312-5, unfiltered and filtered channel base currents and its piecewise-linear approximation. The piecew ise linear app roximation of the filtered current was used as input to the model...................................................... 221 6-27: Stroke FPL0315-1, measured channel base current and its piecewise-linear approximation.........................................................................................................222 6-28: Stroke FPL0312-5, comparison of th e overall measured and modeled phase A-to-neutral current divisions at (left) poles 1, 2, and 6 and (right) poles 10 and 14 on a 100 s tim e scale................................................................................ 223

PAGE 19

xix 6-29: Stroke FPL0315-1, comparison of th e overall measured and modeled phase A-to-neutral current divisions at (left) poles 1, 2, and 6 and (right) poles 10 and 14 on a 100 s tim e scale................................................................................ 224 6-30: Stroke FPL0312-5, comparison of th e overall measured and modeled ground current divisions at (lef t) poles 1, 2, and 6 and (right) poles 10, 14, and 15 on a 100 s time scale. ................................................................................................ 225 6-31: Stroke FPL0315-1, comparison of th e overall measured and modeled ground current divisions at (lef t) poles 1, 2, and 6 and (right) poles 10, 14, and 15 on a 100 s time scale. ................................................................................................ 226 6-32: Stroke FPL0312-5, individu al comparison of all measured phase A-to-neutral currents with model-predicted results on a 50 s tim e scale.................................. 227 6-33: Stroke FPL0315-1, individu al comparison of all measured phase A-to-neutral currents with model-predicted results on a 50 s tim e scale.................................. 228 6-34: Stroke FPL0312-5, individu al comparison of all measured phase A-to-neutral currents with model-predicted results on a 500 s tim e scale. Note that the pole 15 terminator current was not measured........................................................ 229 6-35: Stroke FPL0315-1, individu al comparison of all measured phase A-to-neutral currents with model-predicted results on an 800 s tim e scale. Note that the pole 15 terminator current was not measured........................................................ 230 6-36: Stroke FPL0312-5, indi vidual comparison of all measured ground currents with model-predicted results on a 50 s time scale. Note that the pole 15 term inator current was not measured..................................................................... 231 6-37: Stroke FPL0315-1, indi vidual comparison of all measured ground currents with model-predicted results on a 50 s time scale. Note that the pole 15 term inator current was not measured..................................................................... 232 6-38: Stroke FPL0312-5, indi vidual comparison of all measured ground currents with model-predicted results on a 500 s tim e scale............................................. 233 6-40: Idealized distribution line and its circuit representation......................................... 236 6-41: Currents through th e close arrester (iR) and the far arrester (iRL) calculated using the simple model. The yellow-sh aded area enclosed by the two currents represents the charge transfer Q through the close arrester due to the impulsive current.................................................................................................... 239 6-42: Currents through the close and far arresters and Q calculated with the simple mo del and the EMTP model for 3, 4, and 5 span lengths between the two arrester stations. The time constants used in the simple model are given....... 240

PAGE 20

xx 6-43: EMTP-calculated absorbed energy in one of the two closest arresters during a typical natural lightning first return stroke current injected into the phase conductor at midspan. The vertical lin e contained (a) 4, (b) 8, or (c) 16 arrester stations. The transient mode (dark shaded area) and s teady-state mode (light shaded area) determined fr om the arrester currents for case (a) are indicated...........................................................................................................247 6-44: Comparison of measured data with model-predicted results for stroke FPL0336-6..............................................................................................................250 6-45: Peak value of current injected into the line through the pole 15 grounding as a function of peak value of lightning cu rrent injected into ground 11 m from pole 15. The linear regression equations and R2 values are given......................... 254 6-46: Charge injected into the line through the pole 15 grounding as a function of lightning charge injected into ground 11 m from pole 15. The linear regression equation and R2 value are given. The integration time used to obtain the charge transfers from the return stroke currents and initial continuous currents was 1 ms and 10 ms, respectively.......................................... 254 6-47: Magnitudes of counterpoise current p eaks vs. channel base current peaks for the mobile launcher experiment. The linear regression equation and R2 value are given. Distance from the lightning to the north-west corner of the counterpoise is 290 m.............................................................................................261 6-48: Magnitudes of counterpoise current p eaks vs. channel base current peaks for the tower launcher experiment. Th e linear regression equation and R2 value are given. Distance from the lightning to the north-west corner of the counterpoise is 50 m...............................................................................................261 B-1: Stroke FPL0011-1, phase C-to-neutral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 297 B-2: Stroke FPL0011-2, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 298 B-3: Stroke FPL0011-3, phase C to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 298 B-4: Stroke FPL0011-4, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 299

PAGE 21

xxi B-5: Stroke FPL0011-5, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 299 B-6: Stroke FPL0014-1, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 300 B-7: Stroke FPL0014-2, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 300 B-8: Stroke FPL0014-3, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 301 B-9: Stroke FPL0018-1, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 301 B-10: Stroke FPL0018-2, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 302 B-11: Stroke FPL0018-3, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 302 B-12: Stroke FPL0018-4, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 303 B-13: Stroke FPL0018-5, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 303 B-14: Stroke FPL0018-6, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 304 B-15: Stroke FPL0032-1, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 304 B-16: Stroke FPL0032-2, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 305

PAGE 22

xxii B-17: Stroke FPL0032-3, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 305 B-18: Stroke FPL0032-4, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 306 B-19: Stroke FPL0032-5, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 306 B-20: Stroke FPL0032-6, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 307 B-21: Stroke FPL0032-7, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 307 B-22: Stroke FPL0033-1, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 308 B-23: Stroke FPL0034-1, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 308 B-24: Stroke FPL0034-2, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 309 B-25: Stroke FPL0034-3, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 309 B-26: Stroke FPL0034-4, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 310 B-27: Stroke FPL0034-5, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 310 B-28: Stroke FPL0036-1, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 311

PAGE 23

xxiii B-29: Stroke FPL0036-2, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 311 B-30: Stroke FPL0036-3, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 312 B-31: Stroke FPL0036-4, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 312 B-32: Stroke FPL0036-5, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 313 B-33: Stroke FPL0037-1, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 313 B-34: Stroke FPL0037-2, phase C-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 314 B-35: Stroke FPL0208-1, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 315 B-36: Stroke FPL0210-1, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 315 B-37: Stroke FPL0218-1, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 316 B-38: Stroke FPL0219-2, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 316 B-39: Stroke FPL0219-3, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 317 B-40: Stroke FPL0220-1, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 317

PAGE 24

xxiv B-41: Stroke FPL0220-2, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 318 B-42: Stroke FPL0220-3, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 318 B-43: Stroke FPL0220-4, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 319 B-44: Stroke FPL0220-5, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 319 B-45: Stroke FPL0221-2, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 320 B-46: Stroke FPL0221-3, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 320 B-47: Stroke FPL0221-4, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 321 B-48: Stroke FPL0221-5, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 321 B-49: Stroke FPL0221-6, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 322 B-50: Stroke FPL0226-1, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 322 B-51: Stroke FPL0226-2, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 323 B-52: Stroke FPL0226-3, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 323

PAGE 25

xxv B-53: Stroke FPL0226-4, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 324 B-54: Stroke FPL0226-5, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 324 B-55: Stroke FPL0226-6, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right), 1 ms tim e scale......................... 325 B-56: Stroke FPL0228-1, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 325 B-57: Stroke FPL0228-2, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 326 B-58: Stroke FPL0228-3, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 326 B-59: Stroke FPL0228-4, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 327 B-60: Stroke FPL0228-5, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 327 B-61: Stroke FPL0228-6, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 328 B-62: Stroke FPL0229-1, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 328 B-63: Stroke FPL0229-2, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 329 B-64: Stroke FPL0229-3, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 329

PAGE 26

xxvi B-65: Stroke FPL0229-4, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 330 B-66: Stroke FPL0229-5, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 330 B-67: Stroke FPL0229-6, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 331 B-68: Stroke FPL0229-7, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 331 B-69: Stroke FPL0229-8, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 332 B-70: Stroke FPL0229-9, phase A-to-neutr al currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current ( bottom, left and right)................................................... 332 B-71: Data m atrix for stroke FPL0301-1 (direct strike), 100 s time windows.............. 334 B-72: Data ma trix for stroke FPL0301-2 (direct strike), 100 s time windows.............. 335 B-73: Data ma trix for stroke FPL0301-3 (direct strike), 100 s time windows.............. 336 B-74: Data ma trix for stroke FPL0305-1 (direct strike), 100 s time windows.............. 337 B-75: Data ma trix for stroke FPL0305-2 (direct strike), 100 s time windows.............. 338 B-76: Data ma trix for stroke FPL0305-3 (direct strike), 100 s time windows.............. 339 B-77: Data ma trix for stroke FPL0305-4 (direct strike), 100 s time windows.............. 340 B-78: Data ma trix for stroke FPL0312-1 (direct strike), 100 s time windows.............. 341 B-79: Data ma trix for stroke FPL0312-2 (direct strike), 100 s time windows.............. 342 B-80: Data ma trix for stroke FPL0312-3 (direct strike), 100 s time windows.............. 343 B-81: Data ma trix for stroke FPL0312-4 (direct strike), 100 s time windows.............. 344 B-82: Data ma trix for stroke FPL0312-5 (direct strike), 100 s time windows.............. 345

PAGE 27

xxvii B-83: Data matrix for stroke FPL0312-6 (direct strike), 100 s time windows.............. 346 B-84: Data ma trix for stroke FPL0312-7 (direct strike), 100 s time windows.............. 347 B-85: Data ma trix for stroke FPL0312-8 (direct strike), 100 s time windows.............. 348 B-86: Data ma trix for stroke FPL0312-9 (direct strike), 100 s time windows.............. 349 B-87: Data ma trix for stroke FPL0312-10 (direct strike), 100 s time windows............ 350 B-88: Data ma trix for stroke FPL0315-1 (direct strike), 100 s time windows.............. 351 B-89: Data ma trix for stroke FPL0315-2 (direct strike), 100 s time windows.............. 352 B-90: Data ma trix for stroke FPL0317-1 (direct strike), 100 s time windows.............. 353 B-91: Data for wireburn FPL0402, 600 s time scale. .................................................... 355 B-92: Data for stroke FPL0403-1, 2 ms and 100 s time scales..................................... 356 B-93: Data for stroke FPL0403-2, 2 ms and 100 s time scales..................................... 357 B-94: Measured channel base current and induced currents on the runway counterpoise (290 m from the lightning) and the vert ical wire (300 m from the lightning) for flash 0501, stroke 1 duri ng the mobile launch er experime nt..... 358 B-95: Measured channel base current and induced currents on the runway counterpoise (290 m from th e lightning) and the vert ical wire (300 m from the lightning) for flash 0503, stroke 1 duri ng the mobile launch er experime nt..... 358 B-96: Measured channel base current and induced currents on the runway counterpoise (290 m from th e lightning) and the vert ical wire (300 m from the lightning) for flash 0503, stroke 2 duri ng the mobile launch er experime nt..... 359 B-97: Measured channel base current and induced currents on the runway counterpoise (290 m from th e lightning) and the vert ical wire (300 m from the lightning) for flash 0503, stroke 3 duri ng the mobile launch er experime nt..... 359 B-98: Measured channel base current and induced currents on the runway counterpoise (290 m from th e lightning) and the vert ical wire (300 m from the lightning) for flash 0503, stroke 4 duri ng the mobile launch er experime nt..... 360 B-99: Measured channel base current and induced current on the runway counterpoise (50 m from the lightning ) for flash 0510, stroke 1 during the tower launcher experiment.....................................................................................360

PAGE 28

xxviii B-100: Measured channel base current and induced current on the runway counterpoise (50 m from the lightning ) for flash 0512, stroke 1 during the tower launcher experiment.....................................................................................361 B-101: Measured channel base current and induced current on the runway counterpoise (50 m from the lightning ) for flash 0512, stroke 2 during the tower launcher experiment.....................................................................................361 B-102: Measured channel base current and induced currents on the runway counterpoise (50 m from the lightning) a nd the vertical wire (100 m from the lightning) for flash 0517, stroke 1 du ring the tower launcher experiment............. 362 B-103: Measured channel base current and induced currents on the runway counterpoise (50 m from the lightning) a nd the vertical wire (100 m from the lightning) for flash 0517, stroke 2 du ring the tower launcher experiment............. 362 B-104: Measured channel base current and induced currents on the runway counterpoise (50 m from the lightning) a nd the vertical wire (100 m from the lightning) for flash 0520, stroke 1 du ring the tower launcher experiment............. 363 B-105: Measured channel base current and induced currents on the runway counterpoise (50 m from the lightning) a nd the vertical wire (100 m from the lightning) for flash 0521, stroke 1 du ring the tower launcher experiment............. 363 B-106: Currents during natural li ghtning flash MSE0504, stroke 1................................ 364 B-107: Currents during natural li ghtning flash MSE0504, stroke 2................................ 364

PAGE 29

xxix Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DIRECT AND NEARBY LIGHT NING STRIKE INTERACTION WITH TEST POWER DISTRIBUTION LINES By Jens Daniel Schoene May 2007 Chair: M.A. Uman Cochair: V.A. Rakov Major Department: Electrical and Computer Engineering The interaction of direct and nearby rocket-triggered lightning with two unenergized three-phase power distribution li nes of about 800 m length was studied at the International Center for Lightning Research and Testing in Flor ida. A horizontallyconfigured line was tested in 1999 and 2000, a vertically-configured line in 2001, 2002, and 2003, and a vertically-configured line with overhead ground wire in 2004. All lines were equipped with arresters and, additionally, in 2003, the vertical line had a polemounted transformer. During the 2000, 2001, and 2002 direct strike experiments, arresters frequently failed, but there was no arrester failures either during the 2003 direct strike experiment when the transformer was on the line or during th e 2004 direct strike experiment when the lightning current was in jected into the overhead ground wire. All line configurations except the one tested in 2004 commonly exhibited flashovers. The division of return stroke curre nts for the vertically-configured line was initially similar to the division on the horizontally-configured line (t hat is, the arresters closest to the strike

PAGE 30

xxx point conducted the bulk of th e impulsive current). After so me tens of microseconds the currents in all arresters on the vertically -configured line equalized, while the close arrester currents on the horizontally-configur ed line still conducted significantly more current than the remote arresters. The lightni ng current division for direct strikes to a phase conductor is successfully modeled with the Electromagnetic Transient Program (EMTP) for the vertically-configured line and, if the residual voltage of the close arresters is reduced by 20%, successfully m odeled for the horizontally-configured line. Currents on the vertically-configured line induced by nearby lightning strikes were measured and compared to results calculated using the LIOV-EMTP96 code. It was found that during a lightning strike 11 m from a grounded line pole, a significant fraction of the lightning currents entered the ne utral conductor thr ough the line grounding. Additional topics include an investigation of the response of metallic structures (a buried counterpoise and a vertical wire of 7 m hei ght) to nearby rocket-t riggered and natural lightning strikes and the char acterization of rocket-triggered lightning return stroke currents.

PAGE 31

1 CHAPTER 1 CHAPTER 1: INTRODUCTION Lightning commonly strikes to and near power distribu tion lines causing flashovers and/or damage to equipment connected to th e line. Lightning-caused service interruptions are particularly frequent in areas exhibiting a high ground flash density, such as the southeastern United States. The purpose of this dissertation is to inve stigate the behavior of distribution lines subjected to currents from rocket-triggered lightning that were (1) directly injected into one of the line conductors a nd (2) directed to ground near the line causing (a) induced currents on the line and (b) some fraction of th e lightning current to enter the line through the line groundings. The data presented and discussed are needed in order to develop realistic models for designing lightning protection for distribution lines. During six summers (1999 through 2004) the Lightning Research Laboratory of the University of Florida has been studying, under Florida Power and Light (FPL) support, the interaction of triggered lightning with three distributio n-line framing configurations that were constructed by FPL at the Inte rnational Center for Lightning Research and Testing (ICLRT) at Camp Blanding, Florida. A horizontally-configured line was the primary subject of the 1999 and 2000 experiments, a modified vertically-configured line was studied during summers 2001, 2002, and 2003, a nd a modified vertically-configured line with an overhead ground wire was st udied during summer 2004. Triggered lightning current was directly injected into distribution lines during al l years of the six year study. The effects of nearby triggered lightning, at distances from the line ranging from 7 to

PAGE 32

2 100 m, were examined in 2002 and 2003 for the case of the modified vertical framing configuration. Additionally, results from a project funded by Lawrence Livermore National Laboratories in 2005 are presented in this dissertation. This project examined the response of grounded structures to nearby stri kes. Nearby rocket-triggered or natural lightning was the source of induced currents in a buried test-runway counterpoise and in a grounded vertical wire of 7 m height. The m easured currents in these two systems are presented in this dissertation since they have implications regarding the issue of induced currents on power lines and pow er line grounding systems. Some of the data obtained, with the help of many participants, during the 7 years of experiments has been previously discussed in 2 dissertations (Mata, 2000; Mata, 2003), 3 journal publications (Mata et al., 2003; Schoene et al., 2006a ; Schoene et al., 2006b), 6 conference proceedings (Rakov, 1999a ; Chrzan and Rakov, 2000; C.T. Mata et al. 2002; Rakov, 2003; Rakov et al., 2003a ; Rakov et al., 2003b ), and 7 technical reports (Mata et al., 2000b; Mata et al., 2001; A.G. Mata et al., 2002; Schoene et al., 2003a ; Schoene et al., 2004a ; Schoene et al., 2004b; Hanley et al., 2006). The following sections give a review by year of the 7 years of experiments including a list of the pa rticipants and their contributions/respons ibilities, a list of previous publications discussing the experiment al data, and results presented in previous publications. The original contri butions of the author in obtaining the new experimental data and in analyzing both previously-reco rded data and new data are delineated. A description of the 9 separate experiments listed below that took place during the 7 years, 1999 follows:

PAGE 33

3 1. 1999 FPL Experiment, Direct Lightning Stri ke Interaction with a Horizontallyconfigured Distribution Line 2. 2000 FPL Experiment, Direct L ightning Strike Interaction with a Horizontallyconfigured Distribution Line 3. 2001 FPL Experiment, Direct L ightning Strike Interaction with a Verticallyconfigured Distribution Line 4. 2002 FPL Experiment, Direct L ightning Strike Interaction with a Verticallyconfigured Distribution Line 5. 2002 FPL Experiment, Nearby Lightning Stri ke Interaction w ith a Verticallyconfigured Distribution Line 6. 2003 FPL Experiment, Direct Lightning Stri ke Interaction with a Verticallyconfigured Distribution Line 7. 2003 FPL Experiment, Nearby Lightning Stri ke Interaction w ith a Verticallyconfigured Distribution Line 8. 2004 FPL Experiment, Direct Lightning Stri ke Interaction with a Verticallyconfigured Distribution Line with Overhead Ground Wire 9. 2005 Lawrence Livermore Experiment, Induced Currents 1.1 1999 FPL Experiment, Direct Lightning Stri ke Interaction with a Horizo ntallyconfigured Distribution Line The 245 m long horizontally-configured threephase test distribution line studied in 1999 consisted of six wooden poles and two arrest er stations. For part of the experiment the phase conductors were terminated in thei r characteristic impedances to simulate to some extent an infinitely long line. Th e experience gained during the 1999 FPL experiment has laid the groundwork for the la rger-scale 2000 FPL experiment. However, the relatively short length of th e line (only 6 poles), the small number of arrester stations (only 2 arrester stations), and problems with the termination resistors (Section 3.5.1) made the 1999 experiment of limited use for th e purpose of investiga ting the effects of direct lightning strikes to pow er lines. A description of the 1999 experime nt, previously

PAGE 34

4 published in Mata et al. ( 1999) and Mata ( 2000), is included in this dissertation for completeness. A statistical analysis of the 1999 return stroke current data was performed for this dissertation since it had not been done in previous publications. 1999 Experiment description found in this dissertation: Experimental facility ( Chapter 3 ) Experimental configuration and referen ces to measurement settings (Section 4.2 ) Triggering results (Section 4.9) New results from 1999 found in this dissertation: Statistical an alysis of retu rn stroke currents (Section 6.2, Appendix C ) 1999 Results presented in pr evious publications: PhD Dissertation: Mata ( 2000); Technical Report: Mata et al. ( 1999) Experiment description Presentation of all data Consistency check of return stroke cu rrents and currents to ground: The return stroke currents were found to be 20% to 40% smaller than the sum of ground currents. This might be due to the fact that in 1999 no end-to-end calibration factors were experimentally obtain ed for current sensors, (Mata, 2000, p.91) Flashover analysis: At least 4 of the 7 flashes that contained return strokes produced flashovers Disconnector operation analysis: One arrest er and several te rmination resistors were destroyed. Arresters did no fail as frequently as they did in the 2000 experiments due to several factors including the presence of inductors in parallel with terminating resistors in some of the 1999 tests, rela tively small return strokes, arcing between conductors (including that facilitated by the presence of nylon cords between the line conductors and residual triggering wires), and failures of terminating resistors and voltage dividers (Mata, 2000, p.90). No arrester failure was observed for the last 3 of the 7 flashe s when terminating resistors in parallel with inductors were installed on the line. Participants and their responsibilities during the 1999 Experiment: Principal investigators (proposal, experi mental design, data analysis, general advice): M.A. Uman and V.A. Rakov Experimental design and implemen tation, data analysis, and report writing: C.T. Mata ICLRT manager: K.J. Rambo General assistance: M.V. Stapleton, A.G. Mata

PAGE 35

5 1.2 2000 FPL Experiment, Direct Lightning Stri ke Interaction with a Horizo ntallyconfigured Distribution Line The 856 m long horizontally-configured threephase test distribution line, a longer version of the line studied in 1999, consisted of 18 wooden poles and six arrester stations. The phase conductors were terminated in their characteristic impedances to simulate to some extent an infinitely long line. The de scription of the 2000 experiment, previously published in Mata ( 2000), Mata et al. ( 2000b), and Mata et al. ( 2003), is also included in this dissertation for completeness. T he new results gained from the 2000 data presented in this dissertation as well as results from previous publications are summarized below. 2000 Experiment description found in this dissertation: Experimental facility ( Chapter 3 ) Experimental configuration and referen ces to measurement settings (Section 4.3 ) Triggering results (Section 4.9) New results from 2000 found in this dissertation: Comparison of struck-phasetoneutral current divisions during the 2000 experim ent with the divisions durin g the 2002 and 2003 experiments (Section 6.4): The division of return stroke currents am ong multiple arresters for the line tested in 2000 was initially sim ilar to the division on the line tested in 2002 and 2003. However, the time during which the return stroke current flowed primarily through the closest arresters to the neutral co nductor was significantly longer on the line tested in 2000. Possible explanations for the different current division are given. Additional modeling results (Section 6.5): The current division on the line during stroke FPL0036-1 is modeled using EM TP96. The arrester current division modeled here does not match the m easured division if the arresters in the model are represented by the manufacturer-provided VI-characteristics. The arrester currents modeled here match the measured arrester currents well if the VI-characteristics of the two closest arresters are modified (the residual voltage is reduced by approximately 20%). The current divi sion during stroke FPL0036-1 has been previously modeled in Mata ( 2000) and A.G. Mata ( 2002) using ATP. Previously published modeling results of the closest a rrester current exhibited ringing if the lightning in the model is represented as an ideal current source (that is, the lightning ch annels characteristic impedance is infinity). Ringing is neither present in the measured arrester currents nor in the modeling results presented here. Additional data presentation (Section 6.4.1, Appendix B ): Additional data fr om eight strokes in two flashes (FPL0032 and FPL0033) showing currents measured simultaneously through both closest struckphase arresters (at pole 8 and pole 11) are presented and discussed. The existence of currents simultaneously measured at

PAGE 36

6 both struck-phase arresters has been overlooked in prev ious publications and only currents through one of the two closest stru ck-phase arresters have been presented. The newly discovered data support the common assumption in the data analysis performed previously that th e two closest struck-phase ar rester currents are equal due to symmetry. Complete consistency check of return stroke currents (high and low measurements), sum of ground currents, and sum of struck-phasetoneutral currents (Section 6.1, Appendix B ): The return stroke currents presented in Mata ( 2000), A.G. Mata et al. ( 2002), and Mata et al. ( 2003) have likely been overestimated by 25%. Revised statistical analysis of retu rn stroke currents (Section 6.2, Appendix C ): The overestimated return stroke current peaks and charge transfers (see previous bullet) have been adjusted. The risetimes determ ined from currents that exhibited ringing during the rising edge have been exc luded from the statistics. Revised arrester damage analysis (Section 6.3.1): The average number of disconnector operation per triggering day was determined as opposed to arrester damage per flash in Mata et al. ( 2003 ), which the author of this dissertation believes to be neither a realistic measur e (arresters could only be examined for damage at the end of the triggering day and not after each flash) nor an appropriate measure (arrester damage was determin ed by disconnector operation even though disconnectors could have opera ted on healthy arresters). Revised and expanded flashover analysis (Section 6.3.2, Appendix D ): Flashovers during the 2000 experiment and their appr oximate location have been determ ined based on the charge transfers in non-st ruck-phasetoneutral connections and nonstruck phases. Using this approach 5 out of 8 flashes (8 out of 34 strokes) were determined to have caused flashovers on the line as opposed to 6 out of 8 flashes determined to have caused flashovers in Mata ( 2000). The exact number of strokes causing flashovers was not determined previously. A comparison of flashovers during the 2000 experime nt with flashovers during the 2001 through 2003 experiments shows that flashovers on th e vertical line were much more common. Possible explanations for the diffe rent flashover behavior are given. 2000 Results presented in pr evious publications: PhD Dissertation: Mata ( 2000); Technical Report s: Mata et al. ( 2000b) and A.G. Mata et al. ( 2002); Journal Publicatio n: Mata et al. ( 2003) Experiment description Presentation of selected data Statistical analysis of return stroke currents Consistency check of the re turn stroke currents and currents to ground. The return stroke charge (the integrat ed current) were found to be 25% to 30% larger than the sum of ground currents. Corona on line conductors and other elements of the system near the strike point were consid ered a possible cause of the missing charge (Mata, 2000). Flashovers that would account for the missing charge were not ruled out (Mata et al., 2003). Arrester damage analysis. Six of the eight triggered lightni ng flashes caused dama ge to one of the two closest arresters (Mata et al., 2003, p.1). The more

PAGE 37

7 frequent arrester damage in the 2000 e xperiment compared to the 1999 experiment was attributed to the absence of failed lin e terminating resistors and failed voltage dividers in the 2000 experiment, the presence of inductors in parallel with terminating resistors in some of the 1999 test s, the relatively small return strokes in the 1999 experiment, and arcing between pha se conductors (some facilitated by the presence of nylon cords between line conduc tors and residual triggering wires) in the 1999 experiment. (Mata, 2000) The arrester damage analysis is revised in this dissertation. Flashover analysis: At least 6 of the 8 flashes that contained return strokes produced flashovers (Mata, 2000). The flashover analysis is revised in this dissertation. Struck-phasetoneutral current divisions: about 40% of the return stroke peak current and about 25% or more of the return stroke charge tran sferred in the first millisecond passed to th e neutral conduc tor through each of the two closest arresters (Mata et al., 2003, p. 1). Based on this result Mata et al. ( 2003) estimated that over half of all first strokes in na tural lightning would re sult in an arrester failure on the 2000 test distribution line. Ground current division: An important finding in this study is that the distribution of peak currents to ground is more or le ss symmetrical with respect to the strike point and appears to be strongly dependent on the distance from the strike point, regardless of the low-frequency, low-cu rrent resistances, whereas the charges transferred to ground are distributed accord ing to the low-frequency, low-current grounding resistances (Mata, 2000, p. 136). Modeling of measured currents on the line: According to Mata (2000) modelpredicted currents through th e two arresters closest to the strike point resem ble measured current waveforms if oscillations in the modeled currents are eliminated by the implementation of a crude lumped corona model (a shunt resistor in parallel with a capacitor at the strike point) and/or the implementation of the lightning channels characteristic impedan ce in the model. However, no modeling results that support this statement are presented in Mata ( 2000). The modeling results (including the closest arre ster currents) presented on a 100 s time scale assumi ng different values for the lightni ng channel impedance effects presented in A.G. Mata et al. ( 2002) match the measured currents reasonably well. Modeled and measured results after 100 s are no t compared. Participants and their responsibilities during the 2000 Experiment: Principal investigators (proposal, experi mental design, data analysis, general advice): M.A. Uman and V.A. Rakov ICLRT manager: K.J. Rambo Experimental design and implementation, data an alysis, and report writing: C.T. Mata General assistance: M.V. Stapleton, A.G. Mata, G.H. Schnetzer, R. Sutil, A. Guarisma, A. Mata, and G. Bronsted

PAGE 38

8 1.3 2001 FPL Experiment, Direct Lightning Stri ke Interaction with a Verticallyconfigured Distribution Line The 812 m l ong vertically-configured three-pha se test distribution line studied in 2001 consisted of 15 wooden poles and 4 arrest er stations. The phase conductors were terminated in their characteristic impedances to simulate to some extent an infinitely long line. Lightning currents were injected into the phase A conductor at pole 8 and between poles 7 and 8 at midspan. Similar to the 1999 FPL experiment, which laid the groundwork for the 2000 FPL experiment, the experience gained during the 2001 FPL experiment was invaluable for the successful completion of the experiments on the same line in 2002 and 2003. Also similar to the 1999 experiment, the data collected in 2001 were not useable for the investigation of the current division on the line due to numerous instrumentation problems. The description of 2001 experiment, previously published in Mata et al. ( 2001) and Mata ( 2003), is included in this dissertation for completeness. The statistical analysis of the 2001 return stroke current data included in this dissertation has not been performed in previous publications. 2001 Experime nt description found in this dissertation: Experimental facility ( Chapter 3 ) Experimental configuration and referen ces to measurement settings (Section 4.4 ) Triggering results (Section 4.9) New results from 2001 found in this dissertation: Statistical an alysis of retu rn stroke currents (Section 6.2, Appendix C ). Disconnector operation analysis (Section 6.3.1): The average number of operated disconnectors per triggering day was determined based on the information provided in Mata ( 2003). Revised and expanded flashover analysis (Section 6.3.2, Appendix D ): Flashovers during the 2001 experiment and their appr oximate location have been determ ined based on the charge transfers in non-st ruck-phasetoneutral connections and nonstruck phases. Using this approach flashovers were determined in all 4 flashes that contained return strokes and were examin ed for flashovers (12 out of 14 strokes were determined to have caused flashovers) as opposed to only 2 flashes

PAGE 39

9 determined to have caused flashovers in Mata ( 2003). The exact number of strokes causing flashovers was not determined previously. 2001 Results presented in pr evious publications: Master Thesis: Mata ( 2003); Technical Report: Mata et al. ( 2001) Experiment description Presentation of selected data Arrester damage analysis: Of the total of nine flashes triggered during the 2001 experiments, only flashes FPL0101 and FPL0102 (both wireburns) showed no evidence of failed arresters during th e direct strike tests (Mata, 2003, p. 84). Flashover analysis: During summer 2001, no evidence of trailing wires o n the line were found after any event and possible fl ashovers (probably between phase A and phase B conductors at pole 8) were observed during two flashes (FPL0107 and FPL0108). (Mata, 2003, p. 84). The number of flashovers during the 2001 experiment has been revised in this di ssertation (see New results from 2001 found in this dissertation above). Participants and thei r responsibilities during the 2001 Experiment: Principal investigators (proposal, experi mental design, data analysis, general advice): M.A. Uman and V.A. Rakov ICLRT manager: K.J. Rambo Experiment implementation: A.G. Mata, M.V. Stapleton, J.E. Jerauld, and D.M. Jordan Data analysis and report writing: A.G. Mata General assistance: C.T. Mata, G.H. Schnetzer, and A. Guarisma 1.4 2002 FPL Experiment, Direct Lightning Stri ke Interaction with a Verticallyconfigured Distribution Line The ma jor differences between the line tested in 2002 and the line tested in 2001 were that (1) two arresters in parallel we re installed on the struck phase (2001: only single arresters) and (2) the initial continuous current wa s diverted from the line. Lightning currents were injected into the phase A conductor between poles 7 and 8 at midspan. The description of the 2002 experiment previously published in A.G. Mata et al. ( 2002) and Mata ( 2003), is included in this dissertation for completeness. The author of this dissertation was actively involved in the 2002 experiment by checking the validity of the acquired data after each triggering day. The statistical analysis of th e 2002 return stroke current data included in this disse rtation has not been performed previously.

PAGE 40

10 2002 Direct Strike Experiment descri ption found in this dissertation: Experimental facility ( Chapter 3 ) Experimental configuration and referen ces to measurement settings (Section 4.5 ) Triggering results (Section 4.9) New results from 2002 Direct Strike E xperime nt found in this dissertation: Statistical analysis of retu rn stroke currents (Section 6.2, Appendix C ) Disconnector operation analysis (Section 6.3.1): The average number of operated disconnectors per triggering day was determined based on the information provided in Mata ( 2003). Revised and expanded flashover analysis (Section 6.3.2, Appendix D ): Flashovers during the 2002 experiment and their appr oximate location have been determ ined based on the charge transfers in non-st ruck-phasetoneutral connections and nonstruck phases. Using this approach flashovers were determined in all 7 flashes that contained return strokes and were examin ed for flashovers (39 out of 43 strokes were determined to have caused flashovers) as opposed to only 2 flashes determined to have caused flashovers in Mata ( 2003). The exact number of strokes causing flashovers was not determined previously. 2002 Direct Strike Experime nt results pr esented in previous publications: Master Thesis: Mata ( 2003); Technical Report: A.G. Mata et al. ( 2002) Experiment description Presentation of selected data Flashover analysis: Flashovers were de termined from video records and photographs. Using this approach 6 out of the 9 examined flashes were determined to have possibly caused flashovers (Mata, 2003). The number of flashovers during the 2002 experiment has been revised in th is dissertation (see New results from 2002 Direct Strike Experiment found in th is dissertation above). For flashes FPL0208 to FPL0226 an instrumentation devi ce (at pole 7) might possibly have helped drain some current from phase A (s truck phase) to phase B (closest phase to the struck one), most likely via a flashover (Mata, 2003, p. 89). Responsibilities of the dissertation author during the 2002 Direct Strike Experiment: Assessm ent of validity of measured currents after each triggering day Other participants and thei r responsibilities during the 2002 Direct Strike Experiment: Principal investigators (proposal, experi mental design, data analysis, general advice): V.A. Rakov and M.A. Uman ICLRT manager: K.J. Rambo Data analysis and report writing: A.G. Mata Experiment implementation: A.G. Mata, M.V. Stapleton, J.E. Jerauld, and D.M. Jordan General assistance: C.T. Mata, G.H. Schnetzer, and A. Guarisma

PAGE 41

11 1.5 2002 FPL Experiment, Nearby Lightning St rike Interaction wi th a Verticallyconfigured Distribution Line The 2002 nearby strike experiment was pe rforme d after the 2002 direct strike experiment. The differences between the 2002 ne arby strike and direct strike experiments were (1) the attenuation settings of the current measurements on the line were adjusted to the lower magnitudes expected for the induced currents, (2) the wire used for the direct strike experiment to inject the lightning current into the line was disconnected from the line, and (3) lightning was triggered from the mobile launcher located 30 m or 100 m north of pole 7. The author of this disse rtation was involved in the design of the experiment, checking the validity of the data after each triggering day, and the analysis/modeling of the experimental data. He is co-author in A.G. Mata et al. ( 2002) and wrote the sections invol ving the presentation, mode ling, and discussion of the nearby strike data in A.G. Mata et al. ( 2002). 2002 Nearby Strike Experim ent descri ption found in this dissertation: Experimental facility ( Chapter 3 ) Experimental configuration and measurement settings (Sectio n 4.5, Appendix A ) Triggering results (Section 4.9) Results from 2002 Nearby Strike Experime nt presented in previous publications: Technical Report: A.G. Mata et al. ( 2002) Experiment description Presentation of selected data Modeling results: The measur ed and model-predicted ground and neutral currents generally match reasonably well. The meas ured and model-predicted arrester and phase currents show little or no resemb lance. The former mismatch was possibly due to an inaccurate VI-characteristic. Responsibilities of the dissertation author during the 2002 Nearby Strike Experiment: Experiment design Analysis and modeling of experi mental data and report writing Assessment of validity of measured currents after each triggering day Other participants and thei r responsibilities during the 2002 Nearby Strike Experiment:

PAGE 42

12 Principal investigators (proposal, experi mental design, data analysis, general advice): M.A. Uman and V.A. Rakov ICLRT manager: K.J. Rambo Modeling of experimental data: M. Paolone, C.A. Nucci, and F. Rachidi Experiment implementation: A.G. Mata, M.V. Stapleton, J.E. Jerauld, and D.M. Jordan General assistance: C.T. Mata, G.H. Schnetzer, and A. Guarisma 1.6 2003 FPL Experiment, Direct Lightning Stri ke Interaction with a Verticallyconfigured Distribution Line The ma in differences between the 2003 direct strike experiment and the 2002 direct strike experiment are (1) only single arresters were used in 2003 and (2) a transformer was connected to the struck phase at pole 2. The author of this dissertation was involved in the design of the experiment, checking th e validity of the data after each triggering day, and the analysis/modeling of the experi mental data. He is the main-author of Schoene et al. ( 2003a ). 2003 Direct Strike Experiment descri ption found in this dissertation: Experimental facility ( Chapter 3 ) Experimental configuration and measurement settings (Sectio n 4.6, Appendix A ) Triggering results (Section 4.9) Results from 2003 Direct Strike Experiment presented in this dissertation and in Schoene et al. ( 2003a ): Statistical a nalysis of re tu rn stroke currents (Section 6.2, Appendix C ) Data presentation: All measurements of the 2003 experiment are presented in Schoene et al. ( 2003a ). Selected m easurements are presented in this dissertation (Section 5.1, Appendix B ). Struck-phasetoneutral current divisions: T he primary path of the return stroke current for the first 50 s or so is through the two arre sters that are closest to the injection point of the return stroke current that is, the arresters at poles 6 and 10, and for the first 20 s or so from those arresters to the grounds of the closest arrester poles. At the time of the peak value of the arrest er currents (about 2 s after the return stroke initiation) the two clos est arresters pass about 80% of the total injected current and after 100 s the current is evenly distributed among all arresters. (Schoene et al., 2003a p. 14). Additional analysis is incl uded in this dissertation (Section 6.4). Modeling results: In this di ssertation the model-predic ted current division during strokes FPL0312-5 and FPL0315-1 are compar ed to m easured results (Section 6.5). C.T. Mata modeled the current division on the line during stroke FPL0312-3. The

PAGE 43

13 author of this dissertation discussed C.T. Matas modeli ng results in Schoene et al. ( 2003a ). Model-predicted currents in the two closest struck-phase arres ters matched the measured currents well while currents in the other two struck-phase arresters were poorly modeled. Ground currents for times before 20 s, or so, are poorly modeled. Ground currents for times after 20 s, or so, are reasonably well modeled. The discrepancies in the struck-phase toneutral currents are likely due to a flashover on the line which the m odel does not take into account. Disconnector operation analysis (Section 6.3): No disconnector operated during the 2003 experiment. Possible reasons for the absence of disconnector operation are discussed in this dissertation. Flashover analysis (Section 6.3): F lashovers during the 2003 experiment and their approximate location have been determined based on the charge transfers in nonstruck-phasetoneutral connections and nonstruck phases. U sing this approach flashovers were determined in all 5 flashe s that contained retu rn strokes and were examined for flashovers (22 out of 26 str okes were determined to have caused flashovers). The percentage of strokes causing flashovers on the line tested in 2003 is similar to the flashover percentage determined on the line tested in 2002. Responsibilities of the dissertation author during the 2003 Direct Strike Experiment: Experiment design Analysis of the experimental data Writing of a report Verification of the experimental data during the experiment Other participants and thei r responsibilities during the 2003 Direct Strike Experiment: Principal investigators (proposal, experi mental design, data analysis, general advice): M.A. Uman and V.A. Rakov ICLRT manager: K.J. Rambo Experiment implementation: J.E. Jerauld, M.V. Stapleton, and D.M. Jordan Modeling of the experimental data: C.T. Mata General assistance: G.H. Schnetzer 1.7 2003 FPL Experiment, Nearby Lightning St rike Interaction wi th a Verticallyconfigured Distribution Line The 2003 nearby strike experiment was pe rforme d after the 2003 direct strike experiment and continued the study of li ghtning-induced curren t conducted in 2002. The differences between the 2003 nearby strike and direct strike experiments were (1) the attenuation settings of the current measuremen ts on the line were adjusted to the lower magnitudes expected for the induced currents, (2) the wire used for the direct strike experiment to inject the lightning curren t into the line and the transformer was

PAGE 44

14 disconnected from the line, and (3) lightni ng was triggered from the mobile launcher located 7 m or 15 m south of pole 4, or 11 m south-east of pole 15. The author of this dissertation was involved in the design of the experiment, checking the validity of the data after each triggering day, and the analys is/modeling of the experimental data. He is the main-author of Schoene et al. ( 2003a ) and co-author of Paolone et al. ( 2004b). 2003 Nearby Strike Experiment descri ption found in this dissertation: Experimental facility ( Chapter 3 ) Experimental configuration and measurement settings (Sectio n 4.6, Appendix A ) Triggering results (Section 4.9) Results from 2003 Nearby Strike Experime nt pres ented in this dissertation, Schoene et al. ( 2003a ), and Paolone et al. ( 2004b): Statistical analysis of re tu rn stroke currents (Section 6.2, Appendix C ) Data presentation: All measurements of the 2003 experiment are presented in Schoene et al. ( 2003a ). Selected m easurements are presented in this dissertation (Section 5.3). Data analysis (Section 6.8): The issue of currents from nearby lightning strikes entering the line through one of the line groundings is investigated. Modeling results: The m easured and modeled ground and neutral currents generally match reasonably well. The m eas ured and modeled arrester and phase currents generally match reasonably well up to the first peak. The modeled arrester and phase currents are typical ly larger than the measur ed currents for times after the first peak. In general, the model work s reasonably well for the first microsecond after the beginning of the return stroke a nd not so well for later times (Schoene et al., 2003a p. 38). The comparison of modeling resu lts with selected data from the 2003 experim ent is also included in this dissertation (Section 6.7) and in Paolone et al. ( 2004b ). Responsibilities of the dissertation author during the 2003 Nearby Strike Experim ent: Experim ent design Analysis and modeling of the experimental data and report writing Assessment of the validity of the measur ed currents after each triggering day Other participants and thei r responsibilities during the 2003 Nearby Strike Experiment: Principal investigators (proposal, experi mental design, data analysis, general advice): M.A. Uman and V.A. Rakov ICLRT manager: K.J. Rambo Experiment implementation: J.E. Jerauld, M.V. Stapleton, and D.M. Jordan Modeling of the experimental data: M. Paolone, C.A. Nucci, and F. Rachidi General assistance: G.H. Schnetzer

PAGE 45

15 1.8 2004 FPL Experiment, Direct Lightning Stri ke Interaction with a Verticallyconfigured Distribution Line wi th Overhead Ground Wire A vertically-configured distribution line w ith overhead ground wire and 4 arrester stations was studied during the summer of 2004. The three phase conductors of the vertically-configured line tested from 2001 through 2003 were lowered and the underneath neutral conductor wa s moved above the three phase conductors to function as an overhead ground wire. The overhead ground wi re was grounded at each of the lines 15 poles. Lightning currents were injected into the overhead ground wire between poles 7 and 8 at midspan. The number of measuremen ts was reduced compared to the number of measurements during previous years-current s on the line were only measured between poles 6 and 10. The author of this dissert ation was involved in the de sign and implementation of the experiment, checking the validity of th e data after each triggering day, and the analysis of the experimental data. He is the main-author of Schoene et al. (2004b). 2004 Experiment description: Experimental facility ( Chapter 3 ) Experimental configuration and measurement settings (Sectio n 4.7, Appendix A ) Triggering results (Section 4.9) Results from the 2004 Experiment presented in this dissertation a nd in Schoene et al. ( 2004b ): Statistical analysis of re tu rn stroke currents (Section 6.2, Appendix C ). Data presentation: All measurements of the 2004 experiment are presented in this dissertation (Section 5.2 Appendix B ) and in Schoene et al. ( 2004b). Data analysis: Schoene et al. ( 2004b) estimated the peak voltage across one of the two arrest ers closest to the current injection points to be 25 kV. This relatively low voltage is well under the critical flashove r voltage of a typical real world distribution line, indicating th at the arresters contributed successfully to preventing backflashovers. Ground current division: At the time of th e injected return st roke current peak roughly 50% of the injected currents flow through one of the two closest grounds at pole 8, 25% of the injected currents flow through the other closest ground at pole 7, and the remaining 2% of the injected currents flow through the remaining grounds.

PAGE 46

16 The pole 8 ground initially carries most of the currents, although the pole 7 ground, which has a smaller low-frequency, low-cu rrent ground resistance, is at the same distance from the injection point. Appare ntly, the transient ground resistance at pole 8 is smaller than the one at pole 7maybe due to the two ground rods installed at pole 8 (only one ground rod is inst alled at pole 7). For later times, the magnitudes of the ground currents are determined by the value of the measured low-frequency, low-current grounding resistances. (Schoene et al. 2004b) Disconnector operation analys is and flashover analysis: No disconnector operated and no flashovers were observed during the 2004 experiment. Responsibilities of the dissertati on author during the 2004 Experiment: Experiment design and implementation Analysis of experimental data and report writing Verification of experimental data during experiment Other participants and their res ponsibilities during the 2004 experiment: Principal investigators (proposal, experi mental design, data analysis, general advice): M.A. Uman and V.A. Rakov ICLRT manager: K.J. Rambo Experiment implementation: J.E. Jerauld and D.M. Jordan General assistance: M.V. Staplet on, J. Howard, and G.H. Schnetzer 1.9 2005 Lawrence Livermore Experiment, Induced Currents Measurements of peak current in the lightning protection systems of munitions storage structures in the UK made using an OBO Bettermann measurement system that incorporates peak current sensor cards (plast ic cards containing a magnetic stripe much like credit cards) have yielded anomalously high peak current values, in excess of 120 kA, that are not associated w ith direct lightning flash attach ment to the structures but potentially could be associat ed with nearby lightning. To examine the issue of nearby lightning inducing high current s on grounded structures Lawrence Livermore National Laboratories has funded a project to measure currents in a test r unway counterpoise and in a grounded vertical wire of 7 m height induced by rocket-t riggered and natural lightning currents during the 2005 lightning campai gn. The results of this study also have implications on the issue of induced current s on power lines and power line groundings.

PAGE 47

17 The experimental configurations are described in Section 4.8 and an overview of the rocket-triggered and natura l lightning events that induced currents measured on the test structures is given in Section 4.9. The experimental data are presented in Section 5.4 and analyzed in Section 6.9. The OBO Betterm ann measurement system has been test ed by Newton ( 2004), Schoene et al. ( 2004a ), and Hanley et al. ( 2006). Responsibilities of the dissertation au thor during the 2005 Lawrence Livermore Experiment: Design of experime nts and inst allation of vertical wire Implementation of all measurements Calibration and maintenance of all measurements Analysis of experimental data Supervising writing of the fi nal report (Hanley et al., 2006) Other participants and their responsib ilities during the 2005 Lawrence Livermore Experiment: Principal investigator (proposal, experi mental design, data analysis, general advice): Dr. M.A. Uman Co-principal investigator: K.J. Rambo Measurement of lightning current s and/or lightning electric fi elds: J.E. Jerauld, J. Howard, and B.D. DeCarlo Triggering: J.E. Jerauld, J. Howard, and B.D. DeCarlo Maintenance of measurements after end of the 2005 triggered lightning campaign: J. Howard and B.D. DeCarlo Assistance in building of vertical wire, measurement instrumentation/calibration/maintenance, a nd writing of final report: B. Hanley General advice: J.E. Jerauld, V.A. Rakov, and G.H. Schnetzer 1.10 Summary of Original Contributions The authors original contri butions in taking new data and analyzing already-taken and new data are ou tlined in Sections 1.1 through 1.9. The following paragraphs summarize the authors work and the majo r new findings from the data analysis. Additional o riginal contributions are outlined in Chapter 7 New and additional statistical analys is of return stroke currents measured during the 1999 through 2004 FPL experiments (Section 6.2, Appendix C ): The statistical distribution of the 10-90% rise times of return stroke currents directly injected into the line is d ifferent than the distribution of the 10-90% risetimes of

PAGE 48

18 return stroke currents during the nearby st rike experiment. This discrepancy is likely due to the different stri ke object and grounding system. It was determined that the charge transferred during the first millisecond after the return stroke initiation and the return stroke current peaks are correlated by a power regression equation (y = 12.3x0.54, where y is the peak current and x is the charge transfer, R2 = 0.76). Analysis of the arrester di sconnector behavior during th e 2000 through 2004 FPL direct strike experiments (Section 6.3): The disconnector operation common during the 2001/2002 vertical line experiment was absent during the 2003 ve rtical line experiment, pr obably due to a transfor mer on the line which protected the arrester s by shunting the low-frequency current components to ground. Disconnector operation during the 20 00 horizontal line experiment was considerably less frequent than during the 2001/2002 vertical line experiment, which was possibly due to the larger nu mber of arrester stations on the 2000 horizontal line reducing the long-duration current through each individu al arrester. New, additional, and revised analyses of the flashover occurrences during the 2000 through 2004 FPL direct strike experiments (Section 6.3): The number of flashovers on the horizontal line was smaller than the number of flashovers on the vertical line which is likely related to the different num ber of spans between arrester stati ons (horizontal line: 3 spans, vertical line: 4 spans). Comparison of struck-phasetoneutral and ground current divisions during the 2000 experiment with the divisions during the 2002 and 2003 experiments (Section 6.4): The division of return stroke currents am ong multiple arresters for the line tested in 2000 was initially sim ilar to the division on the line tested in 2002 and 2003. However, the time during which the return stroke current flowed primarily through the closest arresters to the neutral co nductor was significantly longer on the line tested in 2000. Possible explanations for the different current division are given. Additional modeling results of the currents on the distribu tion lines tested during the 2000, 2002, and 2003 direct strike experiments (Section 6.5): The model-predicted arrester currents ma tch well all arreste r currents measured during the 2002/2003 verti cal line experiment if the ar resters are modeled with the manufacturer-provided VI-characteristic. The model-predicted arrester currents do not match well any arrester currents measured during the 2000 horizontal line experiment if the published VIcharacteristic is used for all arresters in the model. The model-predicted arrester currents match well all arrester currents measured on the horizontal line if a modified VI-characteristic is used for the two arresters closest to the lightning current injection point (that is, the residual volta ges of the two closest arresters are reduced by 20%).

PAGE 49

19 Modeling the arrester-absorbed energy during a natural light ning strike to distribution lines of various lengths (Section 6.6): The minimu m arrester-absorbed energy is defined in this dissertation as the energy absorbed in the arrester during the tran sient mode. The minimum arrester-absorbed energy during a typical natural lightning firs t strokes to a distribution lines with 4, 8, and 16 arrester stations separated by 4 spans was estim ated to be between 40 and 45 kJ. Modeling results of currents induced on the vertical distribution line during the 2003 nearby strike experiment (Section 6.7): The measured and modeled ground and neutral currents generally ma tch reasonably well. The measured and modeled arrester and phase currents generally match reasonably well for the duration of the fi rst peak. The modeled arrester and phase currents are typically larger than the measured cu rrents for times after the first peak. Analysis of lightning currents traversing soil and entering th e neutral conduc tor through a pole grounding during the 2003 nearby strike experiment (Section 6.8): During the 2003 nearby strike experiment a considerable fraction of the lightning current injected into gr ound 11 m from the closest line grounding traverses the ground and flows in the neutral c onductor through the line grounding. Presentation and analysis of currents induced on a vertical wire and buried counterpoise during rocket-triggered and natural lightning (Section 6.9): The peak values of lightning return stroke currents and the peak values of induced currents in the buried counterpoise at distances of 50 m and 290 m from th e lightning strike point are strongly linear correlated. The vertical wire functions as a dE/dt antenna during th e rocket-triggered lightning strikes. The largest induced current in the grounded vertical wire associated with natural lightning striking ground a bout 300 m away was 140 A. This current was likely associated with an upward-directed unconnected leader generated in response to an overhead downward-propagati ng stepped leader step.

PAGE 50

20 CHAPTER 2 CHAPTER 2: LITERATURE REVIEW The first part of Chapter 2 provides general information about pertinent physical processes related to the lightni ng discharge: the physics that leads to the development of thunderclouds (Section 2.1), the processes com prising a natural and a rocket-triggered lightning discharge, and the char acteristics of the lightning re turn stroke in both (Section 2.2). The first part concludes with a revi ew of return stroke models (Section 2.3). The second part of Chapter 2 is concerned with the effects of the lightning discharge on power lines: di stribution line design param eters are introduced (Section 2.4), the literature on mo deling of overvoltages on power distribution lines induced by nearby lightning is reviewed (Section 2.6.1 ), and previous work on the interaction of direct and nearby lightning with power dist ribution and transmission lines is outlined (Section 2.7). 2.1 Cloud Formation and Electrification Lightning (also referred to as a lightning discharge or a li ghtning flash) attempts to equalize regions of opposite electr ical charge. At least one of these regions is provided by an electrically charged cloud, a so-called thundercloud or cumulonimbus, which is the energy source for most lightning strikes. The pr ocess of charge generation/separation is called electrification. Electrifi cation processes take place in a num ber of cloud types. The anvil shaped cumulonimbus shown in Figure 2-1 (used with permission of Ari Laakso) is the most common type. Lightning can also or iginate from other sources that involve different electrification processes (e.g., clouds produced by forest fires, volcanic

PAGE 51

21 eruptions, and atmospheric charge separation in nuclear weapons blasts). The electrical structure of a cumulonimbus and the electrif ication process inside this cloud type are discussed in the following section. Figure 2-1: Small thundercloud ov er a train and a power line. The picture was taken in Finland at the Russian border. Th e thundercloud produced only a few lightning strikes. 2.1.1 Formation of a Cumulonimbus A thundercloud is a lightning-producing deep convective cloud. Thunderstorms form in an unstable atmosphere According to Henry et al. ( 1994), eight types of thunderstorm s are known. The five thunderstorm s common to Florida are: (1) air-mass, (2) sea/land-breeze, (3) oceanic, (4) s quall line, and (5) frontal. The first three thunderstorm types are also known as convectiv e or local thunderstorms. In Florida the mechanisms that lead to the development of cumulonimbi are st rongly seasonal depend. Air-mass and sea/land-breeze thunderstorms constitute the majority of Florida

PAGE 52

22 thunderstorms and occur most frequently during the warm summer months. Frontal storms are a typical wintertime phenom enon, although they also occur, but less frequently, during summertime. Air-mass thunderstorms develop if landmasses are heated by the sun. The landmasse s radiate heat to the mois t layers of air near ground, causing the air to rise. The air forms an updr aft if the air mass is sufficiently large. Sea/land-breezes are caused by the temperature difference between water and land. Sea breezes (from sea to land near the surf ace) occur typically during early afternoon on sunny days; land breezes (from land to sea near the surface) occur typically during nighttime (Figure 2-2). Sea/land-breeze storms are primarily induced by the convergence of air over land that accompani es sea breezes. Less common are sea/land-bree ze storms due to land breezes. Figure 2-2: Formation of sea/land-b reeze thunderstorms. Adapted from the Encyclopedia Britannica.

PAGE 53

23 Frontal storms are usually formed if a cold front advances toward warm air, undercutting the warm air and forcing it to rise ( Figure 2-3). Figure 2-3: Cold front moves under warm front resulting in the formation of cumuloni mbus (frontal storm). Adapted from Kuehr ( 1996). When parcels of warm, moist air rise in an updraft, which is caused by one of the mechanism s listed above, several othe r effects take pl ace (Rakov and Uman, 2003): The air pressure decreases with height causing the parc els of moist air to expand. The rising air cools and condenses on small particles in the atmosphere (condensation nuclei) once the relative humidity in the parcel exceeds saturation. The resulting small water particles form the visible cloud. The height of the condensation level that is the bottom of the visible cloud increases with decreasing relative hum idity at ground. During the condensation process the condensation heat (the energy absorbed as water changes from liquid to vapor) is rele ased. This heat supports the continued upward movement of the air ma sses and water particles. Some of the water particles freeze once they reach a height where the temperature is below 0 C. At -40C, or so, all water particles freeze. The freezing process releases the heat of freezi ng (the energy absorbed as water changes from solid to liquid), which supports the further up ward movement of the particle. A cumulonimbus develops if the decrease in atmospheric temperature with height, the so-called lapse rate, exceeds a cert ain specific value which depends on the humidity of the air, the so-cal led moist-adiabatic lapse rate.

PAGE 54

24 2.1.2 Electrical Structure of a Cumulonimbus The charge distribution inside a cu mulonimbus is complex and changes continuously as the cloud evolves. Most ch arge inside the cloud resides on hydrometeors (liquid or frozen water particles), but also so m e free ions are present. Probably, charged particles and ions of positive and negative pola rity coexist in the same regions inside the cloud, but in some areas particles of one polarity are dominant, forming regions of positive or negative net charge. Figure 2-4: Electrical structure inside a cumulonimbus. Adapted from Simpson and Scrase ( 1937). Early ground-based measurements of the cloud charge via its electric field by Wilson ( 1916, 1920) revealed a vertical, positive dipole structure (regions of positive net charge located above regions of negative ne t charge) for the primary charge regions. Later in-cloud measuremen ts (Simpson and Scrase, 1937) confirmed this result and additionally identified a localized low er region of positive net charge. This positive charge region was not always present (or could not always be detected). Figure 2-4 shows

PAGE 55

25 a typical electrical structure inside a thundercloud. The top tw o charge regions are usually referred to as main charge regions. In Florid a the height of the main negative charge center is typically 7 km above ground level, th e main positive charge center is typically located 1012 km above ground level, and the positive charge center at the bottom of the cloud is located 1 km a bove ground level (Rakov, 2001). 2.1.3 Electrification of a Cumulonimbus The detailed physical processes that lead to the generation and separation of charge and the formation of the charged regions inside the thundercloud are poorly understood. Several hypotheses that try to explain this phenomenon have been proposed. Two of the most important hypotheses are labeled the precipitation theory and the convection theory. In the precipitation theory relatively heavy and large hydrometeors (precipitation in the form of soft hail, called graupel) with a high fall speed (> 0.3 m/s) collide with lighter, smaller hydrometeors (cloud particles in the form of ice and water) carried upwards by updrafts (left half of Figure 2-5). Charge is transf erred during the interaction between the heavy and light particles. In relatively cold reg ions (T < -15 C, or so) the heavy particles will become negatively charged and the lighter particles positively charged. In warmer regions (T > -15 C, or so) at the lower part of the cloud, the process reverses such that the heavy particles will become positively and the lighter particles negatively charged. Gravity and updrafts separa te the lighter particles from the more heavy ones, and a main positive dipole with an additional separate localized positive region at the bottom of the cloud forms (MacGorman and Rust, 1998).

PAGE 56

26 In the convection theory electric charge s are supplied by external sourcesfairweather space charge, corona discharges on the ground and cosmic ra ys (right half of Figure 2-5). Updrafts of warm air carry posit ive fair-weather space charge to the top of the developing cum ulonimbus. Negative charge above the cloud produced by cosmic rays is attracted to the clouds surface by the pos itive charge within the cloud. Most of the negative charge resides on cloud particles. Cl oud particles can carry more charge per unit volume of cloudy air than precipitation. Th e negatively charged cloud particles are carried downward by downdrafts, causing corona at the surface. The corona generates positive charge below the cloud that is carri ed to the upper cloud by the updrafts. This hypothetical mechanism results in th e formation of a positive dipole. Figure 2-5: Precipitation theory (left) and convection theory (right). Adapted from Williams ( 1989). Although it is possible that both precip itation and convection mechanisms are im portant for cloud electrificati on, the precipitation mechanism is viewed in the literature as the more significant.

PAGE 57

27 2.2 Natural and Rocket-triggered Lightning Lightning discharges that neutralize charge inside a thundercloud can be categorized into the following discharge types ( Figure 2-6): Intracloud discharges (dis charge within the cloud) Cloud-to-ground discharges Intercloud discharges (discharge from cloud to cloud) Cloud-to-air discharges Figure 2-6: Discharge types for a thunde rcloud. Adapted from the Encyclopaedia Britannica. Most research has been conducted on cloud-to-ground discharges since this discharge type is th e cause of most lightning damage, injury, and death. Intercloud and cloud-to-air discharges are t hought to be relatively rare com pared to intracloud and cloud-to-ground discharges.

PAGE 58

28 2.2.1 Lightning Discharges between Cloud and Ground Berger ( 1978) classifies lightning discharges between cloud and ground into four categories, based on the directi on of propagation of the initia l leader and the polarity of the charge transferred from the cloud to earth ( Figure 2-7): Type 1: Downward lightning discharge, lowering negative charge to earth Type 2: Upward lightning discharge, lowering negative charge to earth Type 3: Downward lightning discharge, lower ing positive charge to earth Type 4: Upward lightning discharge, lowering positive charge to earth Figure 2-7: Simplified drawi ng of four discharges between cloud and ground. Adapted from Uman ( 1987). Upward directed flashes (type 2 and 4) are typically init iated from tall structures on flat ground or structures of moderate heights on mountai ns. The highest recorded lightning currents (up to 300 kA) and the largest recorded tran sferred charge (hundreds of

PAGE 59

29 coulombs or more) are thought to be associ ated with lightning lowering positive charge to ground (type 3 or 4). Positive lightning co ntains typically one stroke while most negative lightning contains more than one st roke. Negative downward lightning (type 1) constitutes approximately 90% of all cloud-to -ground flashes. The remaining 10% of all cloud-to-ground flashes are covered by the other three categories; with positive downward lightning (type 3) being th e most frequent of the three. A rough outline of the physical processes involved in negative downward lightning is described and illustrated in Figure 2-8. Typical tripole structur e of a lightning producing cumulonimb usupper positive main charge region; lower negative main charge region; localized positive charge region at the cloud base. If the electric field at the bottom of the negative charge region reaches a critical value, a preliminary breakdown starts (t=0). An in-cloud breakdown starts from the negative charge region toward ground carrying negative charge and neutralizing the positive charge region at the cloud base (1 ms). A stepped leader, consisting of a thin, highly ionized core surrounded by a wider corona sh eath, leaves the cloud (1.1 ms). Figure 2-8: Drawings illustra ting some of the various pro cesses comprising a negative cloud-to-ground lightning flash. Adapted from Uman ( 1987).

PAGE 60

30 Figure 2-8: continued The stepped leader moves with a typical average speed of about 2*105 m/s toward ground. Negative charge from the cloud flow s more or less continuously into the leader channel (1.15 ms ms). Figure 2-8: continued When the stepped leader approaches ground, the electric field st rength at certain points on ground (notably at sharp and el evated objects) exceeds the breakdown value of air. At these points one or mo re positive upward-going leaders develop in the direction of the negative downwar d going leader from the cloud (20 ms). An upward-going leader connects with a downw ard-going leader branch. A current wa ve with a current peak value of typically 30 kA, the first return stroke starts propagating upward along the ionized channel pr epared by the leader (20.10 ms).

PAGE 61

31 Figure 2-8: continued The first return stroke neutralizes th e ne gative charge deposited in the leader channel and in the process lowers nega tive charge to ground. The return stroke travels upward with a sp eed in the order of 108 m/s (20.10 ms.2 ms). Following the first return stroke, the cloud region, where the leader has started, is near ground potential. Discharges between this region and negatively charged regions in the cloud can occur, so-called K and J processes (40 ms). Figure 2-8: continued

PAGE 62

32 Figure 2-8: continued A dart leader may form if the channel of the first return stroke has not yet dissipated. The dart leader usually follow s the already existing channel prepared by the return stroke; therefore it is typically not branched. The speed of the dart leader is typically 107 m/s (Uman, 1987) and it lowers negative charge onto the defunct channel of the previous stroke (60 ms ms). Once the dart leader approaches ground, a second return stroke develops in sim ilar manner to the first retu rn stroke (62.05 ms). Additional subsequent returns strokes can occur if this pro cess repeats itself. Figure 2-8: continued

PAGE 63

33 2.2.2 Rocket-triggered Lightning The probability of a lightning strike to a struc ture, even in areas of high lightning activity, is very low. Thus, experiments invol ving close natural light ning are difficult to conduct (or take an inordinately long time) A more practical a pproach to conducting such experiments is to artificially initiate a lightning strike usi ng the rocket-and-wire technique ( Figure 2-9). Figure 2-9: Rocket-triggered lightning in Camp Bla nding, Florida (Flash U9910). The lightning type initiated by using th e rocket-and-wire technique is term ed rocket-triggered lightning. In rocket-triggered lightning, a small rocket with an attached conducting wire is used to ar tificially initiated lightning. U nder favorable conditions, i.e., measured static electric field on ground < -5 kV/m, a rocket trailing a conducting wire is launched with a speed of 200 m/s towards a thundercloud. In th e classical rockettriggered lightning technique, illustrated in Figure 2-10 (Rakov, 1999b), the triggering wire is a continuous conducto r that is connected to ground.

PAGE 64

34 Figure 2-10: Sequence of events in cl assical rocket-tri ggered lightning. As the rocket ascends with a speed of approximately 200 m/s, the electric field at the tip of the rocket is distorted. When the rocket reaches an altitude of about 200 m to 300 m, the field enhancement at the rocket tip can result into the development of a positive leader (provided that a sufficient ambient negative field is present) ascending with a speed of the order of 105 m/s from the rocket ti p towards the thundercloud.

PAGE 65

35 Figure 2-10: continued The upward-going positive leader vaporizes the wire and establishes an initial continuous current (ICC), which flows for typically some hundreds of m illiseconds through the channel. After the cessation of the ICC a no current in terval having a typical duration of tens of milliseconds occurs.

PAGE 66

36 Figure 2-10: continued The ICC may be followed by one or more downward leader/upward return stroke sequences. These leader/return stroke sequenc es are believed to be very similar, if not identical, to the subsequent leader/ret urn stroke sequences occurring in natural lightning.

PAGE 67

37 2.2.3 Return Stroke Current In this section a typical rocket-triggered li ghtning return stroke current that lowered negative charge to ground1 is shown and parameters that characterize this current are introduced ( Figure 2-11). The return stroke current measured at the channel base is characterized by a sharp rising edge follo wed by a mu ch slower decaying part. The return stroke current peak Ipeak ( Figure 2-11a) determines the maximum overvoltage Upeak caused by a direct lightning strike (ZIUpeak peak where Z is the impedance of the struck object). Berger et al. ( 1975) found that 50% of natural lightning first/subsequent return stroke peak valu es exceed 3 0/12 kA and 5% exceed 80/30 kA. The 10-90% risetime t10-90% ( Figure 2-11b) characterizes the rising edge of the return stroke current. Berger et al. ( 1975 ) measured the 2 kA to peak value (a parameter similar to the 10-90% risetime) of natural lightning return strokes and found that the risetime of 50% of first/subsequent stroke currents exceed 5.5 /1.1 s and 5% of the risetimes exceed 18/4.5 s. The maximum time rate of current change max( I/ t), an important parameter for calculating induced effects of lightning, can be estimated using the 10-90% risetime and the return stroke peak %90108.0 max t I t Ipeak. Berger et al. ( 1975) found that 50% of natural lightning first/subsequent stroke dI/d t peak values (a para meter that gives the exact ma ximum time rate of current change) exceed 12/40 kA/s and 5% of the peak values exceed 32/120 kA/s. 1 Return stroke currents in rocket-t riggered lightning resemble subsequent return stroke currents in natural lightning (see Section 2.2).

PAGE 68

38 Figure 2-11: Typical triggered lightning return stroke curren t waveform measured at the channel base (stroke FPL0315-2). The following return stroke current parameters are illustrated: a) peak va lue, b) 10-90% risetime, c) half-peak width, and d) charge transfer.

PAGE 69

39 The half-peak width ( Figure 2-11c) characterizes the tail of the return stro ke current. The half-peak width is related to the charge Q ( dtIQ, the area under the current waveform) transferred during the return stroke ( Figure 2-11d). The charge transfer is related to the elec trical energy E a lig htning strike delivers to a struck object ( dtIUE, where U is the voltage across the struck object) and which can cause melting damage. Berger et al. (1975) measured the 2 kA to half-peak value (a parameter similar to the half-peak width) of natura l lightning return strokes and found that the widths of 50% of first/subsequent st rokes exceed 75/32 s and 5% of the widths exceed 200/140 s. 2.3 Transmission-line Type Return Stroke Models Return stroke models are used to ca lculate lightning-induced overvoltages on power lines (Section 2.6). A transmission-line ty pe return stroke model is a so-called engineering model, that is, a model which re lates the longitudinal re turn stroke current for every time and position along the lightning channel I(z ,t) to the current at the channel base I(0,t). An equivalent e xpression specifying the line charge density along the channel can be obtained using the continu ity equation (Thottappillil et al., 1997). These expressions can be used to calculate the lightning electric and m agnetic fields at specified locations. Engineering models have been discussed in the literature by, for instance, Bruce and Golde ( 1941), Uman and McLain (1969), Rakov and Dulzon ( 1987, 1991 ), W illett et al. ( 1988), Willett et al. ( 1989 ), Diendorfer and Uman ( 1990), Nucci et al. ( 1990 ), Thottappillil et al. ( 1991), Thottappillil and Um an ( 1993), Thottappillil and Uman ( 1994), Thottappillil et al. ( 1997), and Rakov and Uman (1998 2003 ).

PAGE 70

40 Rakov ( 1997) proposed a generalized equation for a channel current that represents many engineering mo dels: )/,0()()/(),(vztIzPvztutzIf 2-1 where vf is the upward propagating front speed (r eturn stroke speed), v is the current wave propagation speed, z is the height of the channel section, u is the Heaviside unit step function (u(t z /vf) = 1 for t < z /vf otherwise u(tz /vf) = 0), and P(z ) is the height dependent current attenuation fact or introduced by Rakov and Dulzon ( 1991). In transmission line type m odels the lightning channel behaves similar to a transmission line, that is, a current wave in jected at the channel origin travels upward. The transmission line type return stroke m odels discussed in this section are (1) the transmission line model (TL model), (2) the modified transmission line model with linear current decay with height (MTLL model), and (3) the modified transmission line model with exponential current decay with height (MTLE model). Figure 2-12 illustrates the current division specified by the TL model. The current wave is shown for height z = 0, z = z1 and z = z2 The current wave travels upward without distortion or attenuati on and with constant speed vf (illustrated by the dotted line). Current wave and current front propagate in the same di rection with the same speed (v = vf ). The current at for instance height z2 is the same as the current at ground (z2 = 0) at time t = z2 /vf earlier. Therefore the equation for the TL model is: ) /,0()/(),'(f fvztIvztutzI 2-2 In the TL model no charge is deposited or removed from the channel; that is, the total net charge of the channel remains unc hanged after the current wave has passed.

PAGE 71

41 MTLL and MTLE model are similar to the TL model except that the current decays with height. The current decay is linear for the MTLL model and exponential for the MTLE model. Therefore the equation for the MTLL m odel is ) / ,0( ) /1()/(),'(f fvztIHzvztutzI 2-3 where H is the total height of the channel, and the equation for the MTLE model is )/,0()/(),'(/ f z fvztIevztutzI 2-4 where is the current decay constant. Figure 2-13 shows the height dependent current wave for the TL model at an arbitr ary fixed moment in tim e (t = t1). The current wave moves from left to right, th at is, upward along the channe l. The upward direction of propagation of the current wave is a characte ristic feature of transmission line type models. Figure 2-12: Current versus time wave forms specified by TL model at ground (z = 0) and at two heights z1 and z2 t t t z' 0 z' z' z' /v v = v f f 2 2 1 0

PAGE 72

42 Figure 2-13: Current versus height z above ground at time t = t1 for the TL model. 2.4 Distribution Line Design Parameters Distribution lines are relatively low-voltage lines (below 50 kV) that deliver power from the substation to the consumer. The goal of lightning protection of distribution lines is to prevent (1) flashovers and (2) system damage. A direct lightning strike to a phase conductor may cause flashovers from the st ruck-phase conductor to other conductors, objects, or to ground. A direct li ghtning strike to the shield wi re of a distribution line may cause flashovers to the phase conductorsocalled back-flashovers. Flashovers may also be caused by induced overvoltages due to a ligh tning strike near the line. If a flashovers occurs circuit breakers on the line operate to prevent that the flashover is sustained by the power-frequency current (Rakov and Uman, 2003 ). The circuit breaker o peration reduces the power quality by causing unwanted tr ansients due to the momentary power interruption. A direct lightning strike to a phase conductor may also cause damage to expensive equipment on the line, such as distribution transformers, or damage to the protection devices on the line, such as arresters. 0 z' I(0,t ) { 1 I(z',t) v f f v t 1

PAGE 73

43 This section discusses design aspects pertinent to improving the performance of overhead distribution lines in the lightning environment, that is, reducing flashovers and system damage due to lightning caused overvoltages. 2.4.1 Insulation Strength of Distribution Lines The insulation strength of distribution lines is commonl y given in terms of the critical flashover voltage (CFO). The CFO is defined as the peak value of a specified voltage impulse applied to the line insulation that has a 50% probability of flashover. The CFO of a 1.2/50 s voltage waveshape is typically used for the lightning insulation coordination. Barker et al. ( 1996) show that measured induced voltages typically exhibit a slower risetime and a faster d ecay than the 1.2/50 s test waveform and therefore produce less stress on insulation. They estimate that the actual CFO for insulation is up to 50% larger than the CFO determined with the 1.2/50 s test waveform, primarily because of the narrow width of the voltage pulse. In the IEEE standard 1410-1997, a flashover is assumed to occur if the voltage across the insulation exceeds 1.5 times the CFO of the insulation. The CFO of a distribution line can be found experimentally in laboratory impulse tests or estimated with rule of thumb formulas based on laboratory experiments. The insulation of a distribution line is composed of different components, such as insulators, wooden crossarms (insulation streng th 360 kV/m), wooden poles (insulation strength 330 kV/m), fiberglass sta ndoffs (insulation stre ngth 500 kV/m), and air gaps (insulation strength 600 kV/m)1. It is difficult to estimate the CFO of a line since experiments show that the CFO of a line is generally considerably smaller than the sum of the CFO of the individual in sulation components (Jacob et al., 1991). Guideline s for 1 All insulation strength values are given for wet conditions. The dry condition CFO is approximately 1030% larger.

PAGE 74

44 estimating the CFO of a distribution line ca n be found in IEEE standard 1410-1997 and in Jacob et al. (1991). Note that equipment on the poles or system elem ents such as guy wires, fuse cutouts, conducting supports, a nd other conductors in the system can reduce the CFO (Rakov and Uman, 2003). 2.4.2 Overhead Ground Wires An overhead ground wire (OHGW) is a conduc tor that is placed above the phase conductors and is typically grounded via a pole ground lead at every pole. An OHGW can protect line equipment from direct lightning strike damage and can reduce the number of flashovers due to direct light ning strikes by intercepting and conducting to ground the lightning current that would otherwise enter a phase conductor. It is important that the OHGW be well grounded and that th e CFO between the system of OHGW and ground leads and the phase conductors is suffici ently large to avoid back flashovers, that is, flashovers from the OHGW or from th e ground leads to a phase conductor. An OHGW can reduce the number of flashovers due to nearby lightni ng strikes by reducing the induced voltage on the phase conductors through capacitive coupling (shielding). The shielding effect of the OHGW improves by pl acing it closer to th e phase conductors but this increases the probability of backflashovers in direct stri kes. According to theory, an OHGW reduces the magnitude of the induced voltages on the phase wires by about 15 45% (Rachidi et al., 1997a ; Yokoyam a et al., 1984). 2.4.3 Metal-Oxide Arresters Surge protective devices (SPD) limit tran sient overvoltages by diverting the transients to ground (transient overvoltages on power lines can be caused by, for instance, lightning, normal utility operations such as capacitor bank switching or tap changing on transformers, or the turning on/off of inductive loads such as motors or transformers.).

PAGE 75

45 SPDs that divert transients behave as open circuits or high impedance during normal operation conditions and as shor t circuit or low impedance dur ing transient overvoltages. They can be categorized into two types(1 ) crowbar devices and (2) voltage clamping devices (Rakov and Uman, 2003). Crowbar devices are triggered by the electr ic breakdown of a gas (e.g., air gaps) or an insulatin g layer (e.g., thyristors). W h en triggered, they become very low impedance which allows the shunting of la rge transients with little ener gy absorbed in the device and are consequently relatively safe from getting damaged during the transient discharge. The response time of gas breakdown crowbar devi ces is relatively large (microsecond time scale for air gaps) due to the time it takes to establish a breakdown path. Voltage clamping devices arresters (also know as varistors) have a non-linear, continuous impedance that becomes low dur ing transient overvoltages and reduces (clamps) the transients to safe levels. In general, voltage clamping devices have the advantage of a fast (nanosecond) response time over crowbar devices thus diverting overvoltages before they cause a flashover or can do damage to the protected equipment. The following section will focus on Metal-Oxide Varistors (MOVs), which are voltage clamping devices commonly installed on power lines to protect line equipment and to prevent flashovers. MOVs are the most widely used arrest ers for the protection of power lines. A typical VI-characteristic of a low-voltage MOV is shown in Figure 2-14. For low voltages (voltages not larger than the norm al operation voltage of the power line) MOVs have a large resistance and c onduct very little curre nt (leakage region). The sm all current is called leakage current. For large vo ltages (e.g., transient overvoltages on power

PAGE 76

46 lines) MOVs show highly non-linear behavior, that is, a large change of the MOV current causes only a small change of the MOV volta gethe voltage is clamped to a nearly constant value (normal varistor operation ). For very large voltages the MOV has a small resistance (upturn region). The VI-characteristic for the normal varistor operation can be expressed as VkI 2-5 where k and are device constants. Figure 2-14: Typical varistor VI-characteristic pl otted on a log-log scale. MOVs are bipolar ceramic semiconductor devices that are made of a ceramic compound consisting primarily of a metal oxide, most commonly zinc oxide (ZnO), doped with additive compounds such as oxide s of bismuth, barium, cobalt, manganese, chromium, or tin (Hillman, 2005). ZnO MOVs co nsist of mi crocrystalline zinc oxide grains that are isol ated from each other by a thin intergranular phase consisting of the additive metal oxides ( Figure 2-15).

PAGE 77

47 Figure 2-15: Structure of a metal-oxide va ristor. Adapted from Harnden et al. ( 1972). The non-linear properties of the arresters impedance are establis hed at the interface between the ZnO grains and the interg ranular phase throughout the whole MOV (Hillm an, 2005). Electrically, each of these boundaries act like a di ode, that is, they block current flow at voltages below the turn-on vol tage (this state corresponds to the leakage region of the MOVs VI-characteristic in Figure 2-14) and conduct current at voltages above the turn-on voltage (this state correspond s to the normal varistor operation in Figure 2-14). The MOV as a whole acts like pa rallel strings of diode s conn ected in series. For very large currents the resi stance of the inte rgranular phase becomes less than that of the ZnO grains resulting in a linear beha vior of the VI-characteristic (this state corresponds to the upturn region in Figure 2-14). MOVs designed for power line protection ( Figure 2-16) contain disks of a few centim eters thickness that are composed of the m e tal-oxide compound described above and are coated with a highly conducting materi al such as a silver containing compound.

PAGE 78

48 Figure 2-16: Illustrations of gapless metal-oxide arrester s. a) Cooper Ultrasil housed Varistar 10 kV distribution arrester. b) Cutaway illustration of a different Cooper distribution arrest er. Adapted from the 2000 Cooper arrester catalogue 235-35. The disks are stacked to achieve the desire d voltage rating and en ergy capability and enclosed inside a housing that provides electrical insulati on and mechanical stability. MOVs designed for power line protection ha ve a large, although limited, capability to absorb energy. The rated energy capability of an MOV can be signi ficantly reduced if the MOV is repeatedly subjected to overvoltages. If the energy capability of an MOV is exceeded, the MOV can be damaged due to excessive heating and can fail in three modes, that is, the failed MOV can act as a (1) short circuit, (2) open circuit, and (3)

PAGE 79

49 linear resistor (Brown, 2004). A short circuit failure can be caused by sustained overvoltages or excessive leakage current causing a puncture site within the disk. An open cir cuit failure can b e caused by large f a ilure currents causing melting of wire lead and electrode solder junctions or by cracking/shattering of the MOV disk. Some MOV employ a safety device that disconnects the l eads during the 50/60 Hz current following short circuit failures of MOVs The disconnector ensures that the line remains operational by opening the short circuit caused by the failed arrester. Additionally, the operated disconnector makes a damaged arrester easily detectable (Lenk, 2004). 2.5 Modeling Direct Strikes to Power Distribution Line Direct strikes to power distributions lines are modeled in this dissertation using the Electromagnetic Transient Program (EMTP), version EMTP96 3.2d. The EMTP was designed to model transient pr ocesses in power systems, a lthough it is also capable of treating steady-state problems such as the load-flow analysis of a power system. The EMTP provides a discrete solution of ordinary differential equations using the trapezoidal rule of integration together with traveling wave soluti on methods based on Bergerons method (Dommel and Meyer, 1974). EMTP96 features EMT P View, a Graphical User Interface (G UI) that can be us ed to graphically design a m odel. The graphically designed model is converted to a Fortran code that is used as the input to the EMTP. Alternatively, the user can omit the use of the GUI and create the model by directly writing the Fortran code. In the following sections a brief history of the develo pment of the EMTP is given and the various components implemented in the EMTP and used in the EMTP model in this dissertation are described.

PAGE 80

50 2.5.1 History of the Electromagn etic Transient Program The EMTP was developed in the late 1960 s by Dr. Hermann Dommel as a digital counterpart to the analog Transient Netw ork Analyzer (TNA), which in turn was developed in the late 1930s to model transien ts in power systems. The capability of the EMTP has been expanded over the years by Dr. Scott Meyer, Dr. Tsu-huei Liu, and others. The EMTP Development Coordination Group (DCG), founded in 1982, commercialized the EMTP in 1984. As a c onsequence of their concerns with the commercialization of the EMTP, Dr. Tsu-hue i Liu resigned as DCG chairman and Dr. Scott Meyer left DCG and developed a free version of the EMTP (although licensing is required) called the Advanced Transient Program (ATP). The first commercial DCG EMTP version was released in 1987 and versi on 2.0 was released in 1989. Version 3 of the DCG EMTP was released in 1996 (EMTP96) which is used in this dissertation to model direct lightning strike effects on power distribution lines (Section 6.5) and is integ rated in the LIOV-EMTP96 code used in this disse rta tion to model nearby lightning strike effects on power di stribution lines (Section 2.6.3 and Section 6.7). The EMTP96 represents the final version of the EMTP based on the original code. EMTP-RV was released in 2003, developed by Hydro-Quebec, and is a com pletely restructured version of the EMTP. EMTP-RV features better simu lation performance in terms of speed and stability, an d greater design convenience, largely accomplished by a new graphical user interface for the model design (EMTPWorks) and the implementation of a new data visualization/analysis tool (ScopeView). Note that the older version of the EMTP (that is, EMTP96) was used for the simulations presented in this dissertation as the latest version (EMTP-RV) was not purchased. (sources: Dommel and Meyer, 1974; EMTP Rulebook, 2001; www.emtp.org ; www.emtp.com )

PAGE 81

51 2.5.2 EMTP Current Sources Type-1 through Type 10 current sources in EMTP96 allow the user to specify the source in a point-by-point fashion (that is, th e source function f(t) is defined empirically at every time step). Type-11 through Type -15 current sources in EMTP96 specify the source by using analytical func tions of time. Currents with a complex shape can be generated by a parallel combination of Type-11 through Type-15 sources and by choosing appropriate turn-on and turn-off times for the sources. In the EMTP all sources are connected between a node and local ground. The source functions are evaluated at discrete time steps only. Linear interpolati on between discrete points is assumed by the program. (source: EMTP Rule Book, 2001 ) 2.5.3 EMTP Arrester Model Models that simulate the non-linear behavi or of various arre ster types (Section 2.4.3) are implemented in the EMTP. The Type -92, 5555 component implemented in the EMTP allows the representation of gapless ZnO surge arres t ers in the model. An arrester model can be implemented by (1) creating a data card that contains the desired VIcharacteristic and the rated voltage of the a rrester, (2) converting th e data card to a punch file using the EMTP support program AUX, and (3) using the punch file in the EMTP model. The non-linear resistance is represented by a power function of the form q refV v pi 2-6 where v and i are the arrester voltage and current, respectively, and p, Vref and q are constants. Vref is typically twice the rated voltage of the arrester and is used to normalize the equation and prevent numerical overflow. In the EMTP each segment of the VIcharacteristic is defined by a separate power function, except for voltages substantially

PAGE 82

52 below Vref for which a linear representation is us ed to avoid exponential underflow and to speed the solution. The compensation method is used to solv e circuits with non-linearities in the EMTP model. If this method is applied to a circuit with a single non-linear element, two equations need to be compensated(1) an equation obtained by removing the non-linear element and representing the linear portion of the circuit as a Thevenin equivalent circuit and (2) an equation that exact ly describes the no n-linear characteristic of the non-linear element (Equation 2-6 for the case o f an arrester being the no n-lin ear element). The Newton-Raphson iterative method is used in the EMTP to solve the two equations. Matrix algebra can be applied in the compensa tion method to solve circuits with multiple non-linearities. (sources: EMTP Rule Book, 2001; Dommel, 1986 ) 2.5.4 EMTP Transmission Line Models Various models that simulate the electric characteristics of transmission lines are implemented in the EMTP(1) lumped parameter models suitable for steady-state studies and (2) distributed parameter models suitabl e for steady-state and transient studies. In lumped parameter models the electrical propert ies of the line are modeled with a circuit consisting of a combination of discrete R, L, and C parameters while in distributed parameter models the line parameters are uni formly distributed al ong the length of the line and traveling wave methods are implemen ted. The two distributed parameter models implemented in the EMTP are (1) the constant parameter model, which assumes the line parameters to be constant at a given frequency and (2) the frequency-dependent parameter model (also known as the JMARTI model) desc ribed in Marti et al. ( 1993), in which the frequency dependence of the line param eters is taken into account.

PAGE 83

53 2.5.5 Leads Connecting the Neutral Conductor to Ground Rods The leads connecting the neutral conductor to ground rods can be represented by a series connection of n sections with each section consisting of a grounded capacitor and an inductor (Mata, 2000). The values of the capacitor and the inductor in e ach section can be calcu lated by multiplying the capacitance C and inductan ce L, each p er unit length, of a vertical wire above ground by the le ngth of the section (Bazelyan et al., 1978): ]/[ /2ln 20mF rh C 2-7 ]/[/2ln 20mHrhL 2-8 where h is the height above ground of the mi dpoint of the section, r is the conductor radius, is the permittivity of free space (8.8541878176-12 F/m) and 0 is the permeability of free space ( 0=4 10-7 N/A2). Mata et al. ( 2000a) m odeled the current division on a two-conductor test power distribution line exposed to rocket-tri ggered lightning currents in 1996 (Section 2.7.9 ) by representing the 5.5 m long ground leads with 11 sections. The num ber of sections was found by incrementally increasing the num ber of sections until no si gnificant difference in the calcu lated currents and voltages was observed by adding an additional section. Mata ( 2000) used 20 sections (the number of sect ions was also determ ined by trial-anderror) in h i s EMTP model that calculat ed the current division on the horizontallyconfigured line tested in 2000. 2.5.6 Ground Rod Model A lumped parameter representation of grounding rods is commonly adapted to model the electric behavior of ground rods in response to lightning transients (e.g.,

PAGE 84

54 Verma and Mukhedkar, 1980; Meliopoulos and Moharam, 1983; Mata et al., 2000a ; Paolone et al., 2004b ). In this representation the groundi ng system is modeled as a series connection of n RLC-s ections as shown in Figure 2-17. The capacitance and inductance in each s ection is given by ][10 /4ln189F ndl l Cr 2-9 and ][10 )/4ln(27H n dll L 2-10 where r is the relative permittivity of the soil, l is the length of the ground rod, d is the diameter of the ground rod, and n is the total number of sect ions (Mata et al., 2000a ). The resistance in each sectio n is given b y nRRDC 2-11 where RDC is the measured low-frequency, lo w-current grounding resistance of the ground rod and n is the total number of sections More sophisticated mo dels that take the non-linearity of the ground rod resistance due to the ionization of soil into account are available (Imece et al., 1996). However, it has b een shown in an example by Mata et al. ( 2000a ) that for low enough values of RDC (below 56 ), large enough values of the ground resistivity (above 4000 m), and small enough ground rod currents (less than 8 kA) the non-linear resistance approximates RDC. Note that all these conditions are met in the 2000 and 2003 experiments modeled here and that, consequently, RDC is used for the ground rod resistance in the EMTP modeling included in this di ssertation. Note also that surface arcs that develop radially from the ground rod along the ground surface thereby reducing the ground resistance with respect to RDC are not taken into account in the

PAGE 85

55 ground rod model presented above. Mata ( 2000) used 50 sections (the number was determ ined by trial-and-error ) in his EMTP model that calculated the current division on the horizontally-configured line tested in 2000. Figure 2-17: Lumped parameters representati on of ground rods. The model consists of n RLC-sections.

PAGE 86

56 2.6 Modeling Nearby Strikes to Power Distribution Lines Information on models that calculate ove rvoltages on power di stribution lines due to nearby lightning and testing of these models is provided in this section. 2.6.1 Calculation of Lightning-induced Overvoltages Overvoltages on power lines caused by near by lightning can be calculated in two steps: 1. Determining the electric fiel d by (a) adopting a return st roke model to specify the temporal and spatial current distribution in the lightning channel from which the electric field is calculated (Rakov and Uman, 1998) or (b) measuring electric fields and extrapolating to field s at locatio ns where the field was not m easured. 2. A coupling model is adopted that calculates the voltages on the distribution line induced by the external electric fields de termined in step (1). Various coupling models have been develope d to estimate the voltage i nduced on distribution lines by nearby lightning. The most important models are based on transmission line theory1 extended to include appropriate dist ributed sources and are described in Rusck ( 1958, 1977), Chowdhuri and Gross ( 1967), and Agrawal et al. ( 1980 ). The LIOV-EMTP96 code uses the Agrawal m odel. Note that in the Rusck model and the Chowdhuri and Gross model some source term s have been omitted (Nucci et al., 1995; Cooray, 1994; Cooray and Scuka, 1998). Rusck gives a sim plified formula to calculate the m aximum value of induced overvoltages Vmax inferred from his general coupling model (Rusck, 1977). The electric fields used in this sim plified coupling model are calculated by assuming an infinitely long single conductor line above a perfectly conducting gr ound and a step lightning current impulse in a vertical channel: 2 p 0 maxv2 v 1 d hI Z V 2-12 1 Transmission line theory refers to the theory to model a transverse electromagnetic (TEM) waveguide such as the parallel conductors of a transmission line as a distributed circuit and is not to be confused with the transmission line lightning return stroke models introduced in Section 2.3.

PAGE 87

57 30 4 10 0 0 Z 2-13 where Ip is the maximum lightning cu rrent at the channel base, h is the height of the conductor, d is the distance of the stroke loca tion from the line, and v is the ratio between the return stroke speed and the speed of light. Rachidi et al. ( 1997a ) discuss the shielding effects due to the presence of other conductors and a ground wire and give simple analytical formulas to calculate the shielding coefficients. The conductor shielding coefficient S Ci is the ratio between the peak voltage on conductor i with other conduc tors present and the calculated peak voltage on a single conduc tor with inductance Li and capacitance Ci at the same height. Rachidi et al. ( 1997a ) derived a formula to appro x imate SCi: ii i iZZ Z SC 2 2-14 where Zi is the characteristic impedance of the single-wire line i with radius ri at height hi. i i i i ir h C L Z2 ln60 2-15 iZ is the characteristic impedance of an equivalent single wire representation of conductor i in a line with N conductors. The conductor i has inductance/capacitance per unit length iL/iC, self inductance/capaci tance per unit length Lii/Cii, mutual inductance/capacitance per unit length Lij/Cij and is located at distance dij from conductor j. L/C is the inductance/capacitance matrix per unit length. i i iC L Z 2-16

PAGE 88

58 N 1j ijj i iLh h 1 L 2-17 N 1j ijj i iCh h 1 C 2-18 i i 0 iir h2 ln 2 L 2-19 ij ji 0 ijd hh4 1ln 4 L 2-20 1 00 LC 2-21 Paul ( 1994) derived Equations 2-19 2-21 for multi-conductors over a perfectly conducting ground. Note that if the c oupling model of Agrawal et al. ( 1980) is applied to an infinitely long, lossless line the induced voltages on the conductors are decoupled, i.e., the induced voltage on a given conductor is not affected by the presence of other conductors (Rusck, 1958; Yokoyam a 1984 ; Rachidi et al., 1997a ). A shielding effect due to the presence of other conductors exists for lossy lines, lines of fi nite length, and for lines with a line conductor held at a fixed potential (Rachidi et al., 1997a ). The ground wire shielding coefficient SGi is the ratio between the peak voltage on conductor i without and with ground wire. Rusck (1958) gives a simplified formula to estimate SGi: RLc Lc h h 1SGgg ig i g i 2-22 where the index g denotes ground wire parame ters, R is the grounding resistance of the ground wire, and c is the speed of light. Equation 2-22 assumes that the ground wire is at ground potential at all times, which is an unrealistic assumption for large spacing

PAGE 89

59 between adjacent groundings (Paolone et al., 2004a ), and that the ground is perfectly conducting. 2.6.2 Testing of Lightning-indu ced Overvoltage Models Barker et al. ( 1996) measured voltages induced by rocket-triggered lightning 145 m away from a one-phase test distribution line with an underneath neutral conductor (Section 2.7.9 for m o re details on that experiment). The measured induc ed voltage peaks were 63% larger than the voltage p eak s predicted by the si mplified Rusck model (Equation 2-12). Rachidi et al. ( 1997a ) sh ow that the shielding coefficient SCi due to other conductors, obtained by solving th e simplified equation (Equation 2-14) is practically equal to the value obtained by solving th e Agrawal coupling equations. They also compared ground wire shielding coefficients SGi calculated with Ruscks simplified formula (Equation 2-22) with results calculated using the Agrawal coupling equation and found that the shielding coefficients from the Rusck for m ula are 6 to 7% smaller. Paolone et al. (2004a ) found that the modeled induced currents calculated using their LIOV-EMTP96 code, which implem ents a transmission line type r eturn s troke model (Section 2.3) and the Agrawal coupling m odel, are in very good agreement with induced currents measured in a redu ced scale model. Paolone et al. ( 2000, 2004a ) com pare voltage peaks calculated (a) with the LIOVEMTP96 code for a multi-conductor line and (b) with the LIOV-EMTP96 code for a single conductor line and m ultiplied by the shielding coefficient obtained from the simplified Rusck formula (Equation 2-22). They found that the results are in good agreem ent if the simplif ied Rusck m odel assumption of a zero potential shielding wire is approximated by periodically grounding the shielding wire at least every 200 m. For less

PAGE 90

60 frequent grounding of the shielding wire (every 500 m or more) the simplified Rusck model underestimates the induced voltages significantly. According to their simulation, the shielding effect of the ground wire de pends strongly on the number of periodical groundings of the shielding wire and not so much on the value of the grounding resistance if the grounding resistance is below 100 and if the strike location is not facing one of the groundings. If the strike lo cation is located in front of one of the groundings, the shielding effect is strongl y dependent on the value of the grounding resistance. 2.6.3 LIOV-EMTP96 Code The LIOV-EMTP96 code (Paolone et al., 2001) has been employed to model the lightning-induced currents on the UF/FPL test power distribution line (S ection 6.7). The model was provided by Dr. Mario Paolone, Dr Carlo Alberto Nucci, and Dr. Farhad Rachidi as part of a joint University of Florida/U n iver sity of Bologna/Swiss Federal Institute of Technology project. Dr. Mari o Paolone visited the ICLRT during the Summers 2002 and 2003 to participate in the validation of the LIOV-EMTP96 code with UF/FPL experimental data from the test distribution line. The LIOV-EMTP96 code links th e LIOV code (Nucci et al., 1993; Rachidi et al. 1997a ) with EMTP96 (Section 2.5). The LIOV code is based on the coupling m odel of Agrawal et al. ( 1980). The Agrawal m odel has b een implemented into the LIOV computer code to calculate the response of a distribution line to a LEMP (Lightning Electromagnetic Pulse). The EMTP96 implem ents the boundary conditions and provides a convenient way to include electrical com ponents in the overall model. The lightning vertical electric and horizont al magnetic fields are calcu lated in the LIOV code by

PAGE 91

61 employing either the Transmission Line Model (Uman and McLain, 1975), the Modified Transmission Line Model with expone ntial current decay (Nucci et al., 1990 ), or the Modified Transm ission Line Model with lin ear current decay (Rakov and Dulzon, 1987 ). The Cooray-Rubinstein formula (Cooray, 1992 ; Rubinste in, 1996 ) is implem ented in the LIO V code to calculate the horizon tal electric field from the vertical electric field derived from the models and to take into account th e propagation effects of the LEMP over lossy ground. The calculated electric a nd magnetic fields are then c oupled to the line including appropriate distributed sources in the transmission line theory. The LIOV-EMTP96 code has been successfully tested with experimental data at the Swiss Federal Institute of Technology in Lausanne using a NEMP (Nuclear Electromagnetic Pulse) simulator and re duced-scale models of single and multiconductor lines (Paolone et al., 2000). 2.7 Experimental Studies of Lightnin g Strike Interaction with Power Lines This section reviews some important expe rimental studies of natural and rockettriggered lightning interaction with power dist ribution and transmissi on lines. Studies of both direct lightning strike inte raction (lightning current is injected into one of the line conductors) and nearby lightning st rike interaction (lightning current is injected into ground at a certain distance from the power line) are considered. The st udies are listed in chronological order. 2.7.1 Japanese Study of Nearby Rocket-Tri ggered Lightning Strike Interaction with a Test Distri bution Line (1977) Literature: Horii ( 1982 ), Horii and Nakano ( 1995 ) The nearby lightning strike in te raction with a te st distribution line was studied at the Kahokugata site in Japan from 1977 to 1985. Currents from rocket-triggered lightning

PAGE 92

62 strikes were injected into gr ound as close as 77 m from a 9 m high wire that simulated a phase conductor of a distribution line. Horii ( 1982) found a linear corre lation between the induced voltage peak values and the corres ponding lightning return stroke current peak values. Adding a grounded wire 1 m above th e phase conductor resulte d in the reduction of the induced voltage by about 40%. 2.7.2 South African Study of Direct and Nearby Natural Lightning Strike Interaction with a Test Di stribution Line (1978, 1979) Literature: Eriksson et al. ( 1982 ) The interaction of nearby and direct natu ral lightning with an 11-kV, three phase test distribution lines was st udied in South Africa in 1978 and 1979. The experiment is discussed in Eriksson et al. ( 1982). The study was part of a joint project between the Electricity S upply Commission (Johannesburg, South Africa) and the National Electrical Engineer ing Research Institute (Pretoria, S outh Africa). The test line was 9.9 km long, with the western end of the line grounded to a buried counterpoise and the eastern end open-circuited. Silicon carbide gapped arresters or gapless MOV arre sters were installed on all three phase at 1 km intervals. Data from 12 direct lightning strike s and 269 nearby lightni ng strikes were collected. The largest arrester currents induced by nearby ligh tning were measured at the east end of the line and were approximately twice the amplitude of the currents through the arresters located remote from the ends. Th e induced arrester current never exceeded 1 kA. The majority of the voltages measured at the midpoint we re induced by nearby lightning lowering negative char ge to ground. The induced vo ltages were unipolar with a positive polarity.

PAGE 93

63 2.7.3 DoE Study of Direct Natural Lightning Strike Interaction with Distribution Lines (1978) Literature: Schneider and Stillwell ( 1979 ) Surges on distribution lines due to natu ral lightning strikes to the line were investigated in 1978 as part of a project funde d by the U.S. Department of Energy (DoE). The experiment is discussed in Schne ider and Stillwell ( 1979). Fourteen battery-operated lightning surge recorders with current se nsors (Rogowski coils) were installed on distribution lines in the St. Pe tersbu rg, Florida, area to measure arrest er currents. Data from two lightning strikes that attached to a 7.62-kV, singl e-phase, overhead distribution line were recorded. The exact strike locations on the line are not known. One stroke that lowered negative charge to ground was recorded during the first lightning strike. The arrester discharge current had a peak value of 15 kA, a rise-time of about 2 s, and halfpeak width of about 36 s. Three strokes that lowered positive charge to ground were recorded during the second lightning strike. The peak value/ risetime/half-peak width of the arrester discharge current was 42 kA/5.6 s/60 s for the first stroke, 32 kA/1 s/9 s for the second stroke, and 40 kA/1 s/5 s for the third stroke. 2.7.4 DoE Study of Nearby Natural Lightning Strike Interaction with Distribution Lines (1979) Literature: Master et al. ( 1984, 1986 ) The interaction of nearby natural lightning with an unenergized test distribution line located near Wimauma, Florida, southeas t of Tampa, was studied in 1979. The test distribution line was a single phase 7.62 kV lin e of 460 m length and c onsisted of 6 poles. The neutral conductor was loca ted under the phase conductor and grounded at both line ends. The phase conductor was ope n-circuited at both ends.

PAGE 94

64 Lightning-induced voltages at one line end and the vert ical electric field of the lightning (the electric field sensor being located 160 m from the induced voltage sensor) were measured. Data from over 100 first retu rn strokes and over 200 subsequent strokes were collected. The lightning strike locations were determined by triangulation using a network of television cameras and thunder rangin g. The majority of the strike locations were at distances ranging from 4 to 12 km from the line. Master et al. ( 1984) obse rved induced voltages of negative or positive polarity the polarity apparently depending on the li ghtning strike location. Based on this observation Master et al argue that the induced voltage cannot be due only to coupling of the vertical component of th e lightnings electric field a nd that the contribution from coupling of the horizontal electric field co m ponent to the line mu st be significant. 2.7.5 Japanese Study of Nearby Natural Li ghtning Strike Interaction with a Distribution Line (1980) Literature: Yokoyama et al. ( 1983, 1986, 1989 ) The interaction of nearby natural lightning with an unenergized test distribution line located at the F ukui steam power station in Japa n was studied from 1980 to 1988. The test distribution line was 820 m long and consisted of 17 poles (numbered No. 1 through No. 17) with a distance of about 50 m between adjacent poles. Natural lightning attached to the 200 m tall chimney which wa s located at 200 m perpendicular distance away from the line. The lightning current at the chimney and induced voltages at poles No. 5, No. 8, No. 11, and No. 14 were measured. A total of 90 voltage waveforms were recorded from 1980 through 1988 with the causative lightning stroke current being av ailable for 32 strokes. The polarity of the induced voltages is discussed in Yokoyama et al. ( 1989). They concluded that except f or

PAGE 95

65 unusual conditions lightning-induced voltages are unipolar. The 26 return strokes that lowered positive charge to ground induced a negative voltage and the 6 return strokes that lowered negative charge to ground induced a positive voltage. Furthermore, Yokoyama et al. observed that the induced vol tage is largest at pole No. 5 which is closest to the lightning strike location. The induced voltage peaks at poles No. 8, No. 11, and No. 14 are about 77%, 52%, and 33%, respectively, of the voltage peaks measured at pole No. 5. 2.7.6 DoE Study of Nearby Natural and Rocket-triggered Lightning Strike Interaction with Distri bution Lines (1985, 1986) Literature: Georgiadis et al. ( 1992 ), Rubinstein et al. ( 1994 ), Rachidi et al. ( 1997b ) The Univers ity of Florida lightning research group studied the in teraction of nearby rocket-trigg ered lightning with an unenergized three-phase test distribution line at the NASA Kennedy Space Center (KSC). The top phase conductor (10 m above ground) of the 448 m long line was terminated at both ends of the line in the lines characteristic impedance of about 600 The other two phases were open-circuited at the ends. Induced voltages on the top phase conductor at each line end and the electric/magnetic field at 500/580 m from the lightning were measured. Data from 3 triggered lightning flashes containing 11 strokes lowering nega tive charge to ground and triggered 20 m from one line end were collected. Two types of induced voltages with almost an equal number of occurrences were measured: (1) Os cillatory voltages with peak values ranging from tens of kilovolts to about 100 kV and (2) impulsive voltages that were nearly an order of magnitude larger than the os cillatory voltages (R ubinstein et al., 1994). The oscillations in the form er voltage type were attributed to m u ltiple reflections at the line ends. Both types were observed to occur for different strokes within a single lightning discharge. Attempts to model the oscilla tory voltages using the time-domain coupling

PAGE 96

66 model of Agrawal et al. ( 1980) and the measured return stroke electric field s were moderately successful (Rubinstein et al., 1994). The modeling results were improved by including th e elec tric field of the dart leader into the model (Rachidi et al., 1997b). Attempts to reproduce the impulsive voltages failed (Rubinstein et al., 1994 ). Rubinstein et al. suggested that the impul sive voltages were attributable to the presen ce of electrical breakdown in the m easuring system. In a separate experiment on the sam e line at KSC induced voltages at both line ends and electric/m agnetic fields due to natural lightning at distances beyond about 5 km were measured. Georgiadis et al. ( 1992) found good agreem ent of the waveshapes of measured induced voltages and induced voltages calculate d using the coupling model of Agrawal et al. ( 1980 ). The calcu lated voltages were generally larger than the measured wavefor m s. This result was attributed to the fact that the measured fields were accurate whereas the fields coupling to the lines were shielded by trees. 2.7.7 EPRI Study of Direct and Nearby Natura l Lightning Strike Interaction with Distribution Lines (1987) Literature: Barker et al. ( 1993 ), Fernandez et al. ( 1999 ) The interaction of nearby and direct natural lightning with operating 13 kV distribution lines was studied from 1987 to 1990. Up to 71 lightning transient recorders were installed on distribution lines in Florida in 1987 and 1988 and 75 transient recorders were installed on lines in Tennessee, Colorado, and New York in 1989 and 1990. The transi ent recorders measured either arrester voltages or currents. Voltage peaks of less than 17 kV were recorded for 95% of the 953 lightningcaused arrester voltages. The largest recorded arrester voltage peak was 28 kV. Barker et al. reported that only one of the measured voltages was large enough to be positively

PAGE 97

67 attributed to a direct strike to the line within one pole span of the transient recorder. The other voltages were apparently either induced by nearby lightning strike s or due to direct lightning strikes at several spans from the transient recorder. Current peaks of less than 2 kA were r ecorded for 95% of the 357 lightning caused arrester currents. Three unusual large arrester currents (2 of them occurring in the same flash) with current peaks above 10 kA were recorded. The largest arrester current peak recorded without saturation was 28 kA. Barker et al. measured one long-duration slowly decaying current following the main stroke cu rrent. This event saturated the transient recorder at 6 kA for the first 75 s. The non-saturated tail of the current had an average value of about 2 kA for 2 ms. Fernandez et al. ( 1999) argue in response to a reviewers comment that the long-duration current reported in Barker et al. shows that a single arrester can take an appreci able fraction of the low-fre quency com ponents of lightning currents and therefore th e low-frequency com ponents do present a threat to arresters1. 2.7.8 Japanese Study of Direct Natural and Rocket-triggered Lightning Strike Interaction with a Test Transmission Line (1987) Literature: Horii and Nakano ( 1995), Matsumoto et al. ( 1996 ), Motoyama et al. ( 1998 ), Kobayashi et al. ( 1998) The interaction of direct natural and rock et-triggered lightning with an unenergized 275 kV test transmission lin e located on the top of the Okushishiku mountain (930 m above sea level) in Japa n was studied from 1987 to 1996. The Okushishiku test transm ission line had 6 phase conductors of 1. 6 km length that were supported by 6 steel towers (numbered No. 29 through No. 34) of 60 m height and one overhead ground wire of 2.1 km length that was supported by the same 6 steel towers and one additional tower 1 The opposing view is that the low-frequency components are shared uniformly among many line arresters and therefore the energy absorption of a single arrester due to these components is negligible.

PAGE 98

68 (No. 28 located 505 m from No. 29) of 85 m height. From 1987 to 1993 154-kV MOV arresters were installed on each phase. The ar resters were not installed for the 1994 to 1996 experiments. The line was terminated in its characteristic impedance of 500 at the No. 29 tower, which was located at 182 m from the struck No. 30 tower. At the other end the phase conductors were connected to the grounded metallic crossarm. From 1993 to 1996 currents from 8 rocket-triggered and 2 natural lightning strikes were injected into the 4-m long lightning rod installed on the top of the No. 30 tower. The No. 30 tower was instrumented to measure the lightning currents, overhead ground wire currents, currents to the 4 grounded tower le gs, arrester currents and insulator string voltages. The positive lightning peak currents meas ured for the 8 triggered strikes ranged from 27 kA to 102 kA and were 132 kA and 159 kA for the two natural strikes. The maximum arrester peak cu rrent and voltage measured were 3 kA and 293 kV, respectively (Matsumoto et al., 1996). In 1994, a back flashover was observed on the line, and a peak voltage of approximately 2.5 M V was measured across the string insulator where the back flas hover occurred (Motoyam a et al., 1998). It was found that the currents to the 4 towe r legs w ere different, which presumably can be attributed to the d i fferent ground impedance of the individual legs. Furthermore, it was observed that the high-frequency compone nts of the injected lightning currents primarily flow to the ground at the stru ck tower legs while the low-frequency components tend to flow to the other grounds through the overhead ground wire ( Horii and Nakano, 1995 )

PAGE 99

69 2.7.9 EPRI Study of Direct a nd Nearby Rocket-triggered and Natural Lightning Strike Interaction with Distribution Lines (1993) Literature: Uman and Rakov ( 1995 ), Barker et al. ( 1996 ), Fernandez ( 1997 ), Fernandez et al. ( 1997a, 1997b, 1998a, 1998b, 1998c 1999), Uman et al. ( 1997 ), Mata et al. ( 1998 ) The interaction of direct and nearby rock et-triggered lightning and nearby natural lightning with an unenergized test distribution line was studied at a research facility at Cam p Blanding, Florida by Power Technol ogies, Inc. (PTI) from 1993 to August 1994 and by the University of Florida resear ch group from September 1994 to 1996. The resea rch facility, which is the same research facility where the data presented and analyzed in this dissertation were collect ed, was constructed by PTI in 1993 under the funding and direction of EPRI and named the International Center for Lightning Research and Testing (ICLRT) in 1995. UF took responsibility for operating the facility in 1994. The 730 m long line had 15 poles, a single phase conductor at a height of about 7.5 m, which was terminated at both line ends in the lines characteristic impedance of about 500 and a neutral conductor located 1. 8 m under the phase conductor. The critical flashover overvoltage (CFO ) of the line was about 500 kV. During the 1993 experiment repo rted in Barker et al. ( 1996) the neutral conductor was grounded at poles 1, 9, a nd 15, and lightning was triggere d at distance of 145 m from the line. The lightning currents of 30 flashe s containing 63 strokes and the resulting voltages ind u ced on the line were recorded The largest voltage peak values Vpeak were measured at pole 9 and ranged from 8 to 100 kV. These voltage peaks were typically twice the value of the voltage peaks measured at the end poles, and showed strong linear correlation with the return stroke current peaks Ipeak which ranged from 4 to 44 kA (correlation coefficient: 0. 938, regression equation: Vpeak=2.24*Ipeak). Barker et al. found

PAGE 100

70 a weaker correlation between return stroke current rate of rise and peak voltages (correlation coefficient: 0.75) a nd little or no correlation be tween return stroke current rise time and peak voltages (correlation coefficient: 0.28). During the 1994 experiment reported in Uman and Rakov ( 1995) and Uman et al. ( 1997) the line was connected to an u nderground distribution system through a transformer at pole 9. Arresters were insta lled at poles 8, 9, and 10 and the neutral conductor wa s grounded at pole 1, 9, and 15. Two flashes were triggered and injected into the phase conductor be tween poles 9 and 10 although lightning currents were only m easured for the first flash (a flash containing 4 strokes followed by many Mcomponents). Currents on the line were obtai ned at the ground connection at pole 9, and arrester voltages on the line were obtained at pole 9. During the 1995 and 1996 experiments reported by Fernandez ( 1997), Fernandez et al. ( 1997a 1997b, 1998a 1998b, 1998c 1999), and Um an et al. ( 1997) the interac tion of direct and nearby natural and ro cket-triggered lightni ng with the test di stribution line was studied. During the two years 38 lightning flas hes that contained return strokes were triggered of these were classical triggers and 10 were altitude triggers1. Thirteen of these flashes were injected into either the phase conductor or the ne utral conductor of the test line. Eighteen flashes were injected into ground at distan ces ranging from 20 m to 155 m and 7 flashes struck ground at various locations, some determined and other undetermined locations. Data from 6 natural li ghtning events on or very near the facility 1 An altitude trigger is caused by a trailing wire that is not continuous but either deliberately interrupted or breaks during the ascend of the rocket, resulting in a grounded wire section close to ground and an ungrounded wire section above the grounded wire. The lightning current path is established by a bidirectional leader (an upward positive and downward ne gative leader) from the ungrounded wire section as opposed to a unidirectional leader (an upward positive l eader) from the tip of the wire in classical rockettriggered lightning.

PAGE 101

71 were obtainedone of the natural flash term inated on the phase conductor and another forced current into the pole 1 ground. The te st system was arranged in several different configurations during these tr iggered and natural flashesM OV arresters were either not present or were installed at 2, 3, or 4 poles, the neutral wa s grounded at 3, 4, or 5 poles, and in some configurations the underground di stribution system was connected to the line and for others it was not conn ected. The arrester voltage a nd current were simultaneously measured for 9 triggered flas hes and arrester voltages were measured during 11 triggered flashes and one natural flash. The arre ster voltage waveforms often exhibited characteristics that are not observed in th e corresponding current waveforms and continue long after the arrester cu rrent is no longer detect able (Fernandez et al., 1999). The energy absorptions of the pole 9 arreste r (the arrest er closest to the lightning current injection point) were calculated for 5 strokes injected into the line and ranged from 2 to 6 kJ (Fernandez, 1997). Thes e energy values were well below the rated energy capability of the arresters (around 40 kJ) and consequen tly no arrester was damaged during the experim e nts, although the energy absorptions of the arresters during th e initial continuous currents and M-components were not reported an d might have been significant. Note that a 50 cable with arresters at the far end was atta ched to the line, which diverted some of the lightning energy from the line arresters. It was observed that for a lightning stri ke at 20 m from the line (and 40 m from pole 9, the pole closest to the strike point) about 5% of the total lightning current (a 17 kA stroke) entered the neutral conduct or through the pole 9 ground (Fernandez, 1997; Fernandez et al., 1998b). During this lightning strike no ar res ter was inst alled on the line and the neutral conductor was grounded at poles 1, 9, and 15. Currents from nearby

PAGE 102

72 rocket-triggered lightning entering grounding systems other than the line grounds were also measured% of a 20 kA stroke entered the ground of a simulated house (19 m from the strike point) and 10% of the currents from a two-str oke flash (return stroke peak values of the two strokes: 5 kA and 11 kA ) entered the ground of a transformer 60 m away from the strike point (Fernandez, 1997). 2.7.10 Japanese Study of Direct Natural Lightn ing Strike Interaction with a Test Distribution Line (1999) Literature: Nakada et al. ( 2003 ) The direct strike interaction of natu ral lightning with an unenergized test distribution line constructed in 1999 at the Mikuni Testing Center of Hokuriku Electric Power Company in Japan is di scussed in Nakada et al. ( 2003). The 800 m long test distribution line consis ted of 21 evenly spaced co ncrete poles (span length: 40 m) that supported an 11 m high overhead ground wire and three 10 m high, horizontally arranged, high-voltage phase conductors. A dditionally, poles 11 through 15 supported three vertically arranged low-voltage phase conductors located below the high-voltage conductors. The overhead ground wire and th e three high-voltage conductors were terminated with 400 resistors at both line ends. Ga pped MOV arresters having residual voltages of 29 kV and 20 kV for a discharge current of 2.5 kA were installed at poles 1, 5, 9, 13, 17, and 21. Pole transformers were installed at poles 5 and 13. Natural lightning struck the 59 m tall tower built 50 m from the test line. The lightning current was injected into the test distribution line through a conductor connecting the lightning rod on the tower and the top of pole 13. Currents and voltages were measured at up to 23 points along the line. The experimental data is comp ared to model-predicted results calculated using the EMTP. The model employed to repres ent the transmission line sections in the

PAGE 103

73 EMTP calculation was the same used in the modeling efforts presented in this dissertation (the frequency-dependent model described in Section 2.5.4). During the 2000 experiment currents from th ree flashes that struck the tower were injec t ed into the line at pole 13 (a -27 kA st roke, a +33 kA stroke, and a -4.8 kA stroke). Additionally, one flash, for which the lightning current was not measur ed, struck pole 6. The currents and voltages along the line during the 33 kA stroke (risetime: 5 s, halfpeak width: 250 s) and the pole 13 arrester current and voltage during the flash that struck pole 6 are discus sed in Nakada et al. ( 2003). The pole 13 arrester current and voltage during the flash that s truck po le 6 are sho wn in Figure 2-18. Figure 2-18: Measured pole 13 arrester curr ent and voltage during a natural fl ash tha t connected to pole 6. Adapte d from Nakada et al. ( 2003). Nakada et al. attributed the observed a rrester voltage decrease to approximately half-peak value after 50 s ( Figure 2-18b) to a change of the VI-characteristic of the pole 13 arrester caused by energy absorption of the arrester. Based on this observation, Nakada et al. introdu ced a simple arrester model in which the residual voltage of the

PAGE 104

74 arrester is reduced to half the manufacturer-provided value and em ployed this modified arrester model to calculate pole 13 arrester cu rrents during the 33 kA stroke that attached to the tower and which current was inject ed into pole 13. It can be seen from Figure 2-19 that the m easured pole 13 arres t er current during this stroke ( Figure 2-19a) matches the calculated arrester current poor ly if the unmodified VI-chara cteristic is em ployed in the model ( Figure 2-19b). T he match is significan tly im proved if the modified VIcharacteristic is used in the model ( Figure 2-19c). Figure 2-19: Pole 13 arrester current during a natural lightni ng stri ke to pole 13. a) Measured pole 13 arrester current. b) Calculated pole 13 arre ster current using the manufacturer-provided VI-characteris tic. c) Calculated pole 13 arrester current using the modified VI-character istic. Adapted from Nakada et al. ( 2003).

PAGE 105

75 2.7.11 FPL Study of Direct and Nearby Rocke t-triggered Strike Interaction with Distribution Lines (1999) Literature: C.T. Mata et al. ( 1999, 2000a, 2000b, 2002, 2003), Mata ( 2000 ), A.G. Mata et al. ( 2001, 2002 ), Mata ( 2003 ) Schoene et al. ( 2003a, 2004b ) The Lightning Research Laborat ory of the University of Florida has been studying, under Florida Power and Light (FPL) support, th e interaction of tri ggered lightning with three distribution-line fram i ng configurations that w ere constructed by FPL at the International Center for Lightning Research and Testing (ICLRT) at Camp Blanding, Florida. This dissertation is concerned w ith those experiments and their analysis. A horizontally-configured line was the primary subject of the 1999 and 2000 experiments, a modified vertically-configured line was studied during summers 2001, 2002, and 2003, and a modified vertically-configured line with an overh ead ground wire was studied during the summer 2004. Triggere d lightning current was direc tly injected into one or both distribution lines during all years of the six year study. The effects of nearby triggered lightning, at distances from the line ranging from 7 m to 100 m, were examined in 2002 and 2003 for the case of the modifi ed vertical framing configuration. The 1999 and 2000 experiments were discussed in C.T. Mata et al. (1999, 2000a 2000b, 2002, 2003) and Mata ( 2000). A.G. Mata et al. ( 2001, 2002) and Mata ( 2003) discuss the 2001 and 2002 direct strike experim ents and A.G. Mata et al. ( 2002) discuss the 2002 nearby strike experim ent in less detail. Th e 2003 and 2004 experim ents were presented in Schoene et al. ( 2003a ) and Schoene et al. ( 2004b). This dissertation includes additional presentation, analysis, and disc ussion of the data from all experiments conducted from 1999 through 2004.

PAGE 106

76 CHAPTER 3 CHAPTER 3: EXPERIMENTAL FACILITY The International Center for Resear ch and Testing (ICLRT) shown in Figure 3-1 is an outdoor facility occupying about 1 km2 at the Camp Blanding Army National Guard Base, located in north-central Florida approximately midway between Gainesville and Jacksonville. At the ICLRT lightning is routin ely triggered (artific ially initiated) using the rocket-and-wire technique (Section 2.2.2). The inte raction of triggered lightning with three different 3-phase test distribution lines has been studied at the ICLRT from 1999 through 2004: (1) a cross-arm horizontal configuration, tests on that line being conducted in 1999 and 2000, (2) a vertical configurati on, studied in 2001, 2002, and 2003, and (3) a vertical configuration with an overhead ground wire, studied in 2004. The experiments in 1999 through 2004 have been described and discussed in the following publications: Mata et al. ( 1999), Mata et al. ( 2000b), Mata ( 2000), Mata et al. ( 2001), A.G. Mata et al. ( 2002), Ma ta ( 2003), Mata et al. ( 2003), Rakov et al. ( 2003a ), Schoene et al. ( 2003a ), and Schoene et al. ( 2004b ). Table 3-1 provides an overview of the 6 years of distribution-line experim ents including inform ation about the us e of a diversion tec hnique for the initial continuous current (ICC) of the triggered lightning, the triggered lightning current injection point (current was in jected into one of the line c onductors or into ground at a various distances from the line), the presence of a transformer on the line, the location of the line groundings, the loca tion and type of arresters on the line, and the line termination. More detailed information about each aspect can be found in the section referenced to in the corresponding row in the first column of Table 3-1.

PAGE 107

77 Figure 3-1: Overview of the ICLRT.

PAGE 108

78 Vertical Configuration with OHGW 2004 Three phases and overhead ground wire, 15 poles Overhead Ground Wired No Transformer 2003 Phase A, 7 m, 11 m, or 15 m from the line Transformer on struck phase at pole 2 Ohio-Brass at poles 2, 6, 10, and 14 2002 ICC interception structure Phase A, 20 m, 30 m, or 100 m from the line Poles 1, 2, 6, 10, 14, and 15 Ohio-Brass only or Cooper only at poles 2, 6, 10, and 14, two arresters in parallel on struck p hase Vertical Configuration 2001 Three phases and neutral, 15 poles Phase A Poles 1, 2, 6, 10, 14, and 15 Poles 5 and 11 for part of the ex p erimen t Ohio-Brass /Cooper or Cooper only at poles 2, 6, 10, and 14 2000 Three phases and neutral, 18 poles Phase C Poles 2, 5, 14, 17, and 18 Ohio-Brass at poles 8 and 11, Cooper at poles 2, 5, 8, 11, 14, and 17 Struck phase Two 1.75 MJ resistors in parallel Non-struck phase: 1.2 MJ resistors Horizontal Configuration 1999 No Three phases and neutral, 6 poles Phase C or B No transformer Poles 7, 8, 11, and 12 Ohio-Brass at Poles 8 and 11 6.2 kJ resistors on each phase Table 3-1: Experimental configurations used in 1999 through 2004 experi ments. Pole numbers are identified in Figure 3-1. ICC Diversion (Section 3.1) Test Line (Section 3.2) Current Injection Point (Section 3.2) Transformer (Section 3.2.2) Grounding (Section 3.3) Arresters (all three phases) (Section 3.4) Line Termination (Section 3.5)

PAGE 109

79 3.1 Rocket Launchers Two different types of rocket launchers were employed to artificially initiate lightning using the rocket-a nd-wire technique (Section 2.2.2)(1) a stationary laun ch er primarily used to inject lightning current into the power distribution line to simulate a direct lightning strike and also used to simulate nearby lightning strike s at a fixed distance of 20 m and (2) a mobile launcher in tended to trigger li ghtning at a specified distance from the line to simula te nearby lightning strikes at distances fr om 7 m to 100 m. 3.1.1 Stationary Launcher From 1999 through 2004 triggered lightning curr ent was injected into one of the distribution lines conductors from the stationary rocket launcher shown in Figure 3-2. For part of the 2001 experiment the me tallic c onnection between the launcher and the power line was removed and the launcher was used to si mulate indu ced effects on the line due to nearby lightning strikes. The launcher was mounted on an 11 m wood tower located about 20 m north of and near th e midpoint of the test line ( Figure 3-1). The 2002, 2003, and 2004 experim ents differed from t h e e xperiments conducted from 1999, 2000, and 2001 in that a separate path to ground, othe r than via the test distribution line, was provided for the initial continuous current (ICC, see Section 2.2.2) prec eding the return strokes in a triggered flash. The ICC was diverted f rom the line so that the line arresters would not be subjected to the current and char ge transfer of the IC C, only to the return stroke currents. The ICC flowed to the rock et launcher and then to ground via a wire connection, while return str okes following the ICC and any continuing current after those strokes generally attached to a U shaped intercepting structure mounted above the launcher ( Figure 3-2) and from there were directed to the lin e. In the begi nning of the 2002 experiment the wire connection that grounded the rocket launcher was a ground

PAGE 110

80 lead. It was replaced by a fu se wire during the 2002 experi ment, which was used for the remainder of the 2002 experiment and for th e 2003 experiment. The ground lead was not destroyed during the rocket-tri ggered lightning while the fu se wire exploded during the ICC and electrically isolated the tower th ereby increasing the lik elihood for the return strokes and any continuous current followi ng the ICC to attach to the intercepting structure. Figure 3-2: Stationary tower launcher em ployed from 1999 through 2004 primarily used to simulate direct lightni ng strikes to the test line. The intercepting structure (PVC poles and intercepting conductor) were only used for the 2002, 2003, and 2004 experiments.

PAGE 111

81 3.1.2 Mobile Launcher In 2002 and 2003 lightning was triggered from a mobile launcher to simulate induced effects on the vertically-configured power distributio n line due to nearby lightning. The mobile launcher is a utility serv ice vehicle with a rocket launcher installed on the bucket ( Figure 3-3). The mobile launcher could be m oved relatively easy to different loc ations to sim ulate nearby light ning strikes at variou s distances. Triggered lightning connected to the launcher was injected into the gro und through a ground wire, which is connected to the rocket launcher on one end and to multiple ground rods on the other end. A current viewing resistor (CVR or shunt) is mounted on the bucket to Figure 3-3: Mobile bucket-truck rocket launch er placed a few meters from one end of the vertically-configured te st distribution line.

PAGE 112

82 measure the channel base current of the tri ggered lightning. The he ight of the rocket launcher was approximately 8 m above ground. The primary reasons for elevating the launcher was to create a grounde d structure that is higher th an the power line so very close lightning would attach to the launcher and not to the line as it would do in the absence of the mobile launcher. A very close lightning strike near a real world power distribution line occurs if a large structure (e.g., a tree) is present very close to the line and lightning strikes this structure. 3.2 Test Distri bution Lines A horizontally-configured line was th e primary subject of the 1999 and 2000 experiments, a vertically-configured line was studied during summers 2001, 2002, and 2003, and a vertically-configured line with an overhead ground wire was studied during the summer 2004. 3.2.1 Horizontally-configured Line, 1999 and 2000 A horizontally-configured distribution lin e was studied during the summers of 1999 and 2000. This distribution line is described in detail in Mata ( 2000) and in Mata et al. ( 2003). A sk etch of the horizontally -configured line is given in Figure 3-4. The line had a total length of approximately 245 m (1999) or 812 m (2000) and consisted of three horizontally arranged phase c onductors with an underneath ne utral. The phase conductors were of type 5681 19-strand2 conductors with an equivalent diameter of 2.23 cm and a dc-20 resistance of 0.0994 /km. The neutral was a seven-strand AWG 3/0 conductor with an equivalent diameter of 1. 18 cm and a dc-20 resistance of 0.3380 /km. 1 Concentric-lay-stranded aluminum conductors, aluminum reinforced. 2 15 type 1350-H19 wires and 4 type 6201-TB1 wires.

PAGE 113

83 Figure 3-4: Sketch of the horizontally-c onfigured line. Adapted from Mata ( 2000). Horizontally-configured line, 1999 experiment: The approximately 245 m long horizontally-configured threephase test distribution line st udied in 1999 consisted of six wooden poles and two arrester stations. The tw o arrester stations were located at the second pole and the second to last pole (three spans between arrester stations). OhioBrass PDV-100 MOV arresters (Section 3.4) were installed o n each phase. For part of the experime nt the phase conductors were term in ated in their characteristic impedances (Section 3.5) to sim ulate to some extent an infinitely long line. The height of each phase conductor above ground level was approximately 8 m Th e horizontal distance from phase A to phase B was 1.8 m, and the distan ce from phase B to phase C was 0.7 m. The neutral conductor was attached to the line poles at a height of 6.2 m and grounded at each arrester station (Section 3.3). The lightning current was in jected into phas e C between poles 9 and 10 at m i dspan (from 08/16/1999 to 08/24/1999) or into phase B between poles 9 and 10 at midspan (f rom 09/06/1999 to 09/10/1999). Horizontally-configured line, 2000 experiment: The distribution line studied during the Summer of 2000 differed from the line studied in 1999 in that it consisted of 18 poles (vs. 6 poles in 1999) and six threephase arrester stations (vs. 2 arrester stations in 1999) located at poles 2, 5, 8, 11, 14, and 17 (three spans between arrester stations). The 829 m long line (average span length: 50 m) was term inated in its characteristic impedance at both line-terminating poles (Section 3.5). Ohio-Brass PDV-100 MOV arresters and/or Cooper UltraSIL Housed VarisSTAR Heavy Duty were installed on the line (Section 3.4). The neutral conductor was grounded at each arrester station (Section 3.3). The lightn ing current was injected into p h ase C between poles 9 and 10 at midspan (from 07/16/2000 to 08/03/2000) or into phase C at pole 9 (08/06/2000).

PAGE 114

84 3.2.2 Vertically-configured Line, 2001 through 2003 A vertically-configured di stribution line was built in 2000 and 2001 and studied during the summers of 2001, 2002, and 2003. This distribution line is described in detail in Mata ( 2000), Mata ( 2003), and Schoene et al. (2003a ). Th e line had a total length of approximately 812 m (average span length: 58 m) and consisted of three verticallyarranged phase conductors with and underneath neutral. Th e phase conductors were of type 5681 19-strand2 conductors with an equivalent di ameter of 2.23 cm and a dc-20 resistance of 0.0994 /km. The neutral was seven-stra nd AWG 3/0 conductor conductors with an equivalent diameter of 1. 18 cm and a dc-20 resistance of 0.3380 /km. Information on the conductor placement and type is given in Table 3-2. In this table the sign conven tions for the horizontal d i splacement of the conductors are: a) positive sign the conductor is north of the neutral, b) ne gative signthe conduc tor is south of the neutral Table 3-2: Conductor placement and specifica tions for the vertically-configured test distribution line. Height (m) Horizontal Displacement Relative to Neutral (m) Conductors Phase A 11.5 -0.18 Phase B 10.8 0.24 Phase C 10.2 0.24 Type 568 19-strand Neutral 8.7 0 AWG 3/0 7-strand Figure 3-5 shows a sketch of the verti cally-configured line including the span lengths and the location of a rrester stations, groundings, and line terminators. The four thr e e-phase arrester stations were located at poles 2, 6, 10, 14, and 17 (4 spans between arrester stations). 1 Concentric-lay-stranded aluminum conductors, aluminum reinforced. 2 15 type 1350-H19 wires and 4 type 6201-TB1 wires.

PAGE 115

85 Figure 3-5: Sketch of the vertically-c onfigured line tested from 2001 through 2003. Figure 3-6 shows the arrest er station at pole 10. Th e photograph was taken during the 2003 experiment. The line was term inated in its characteristic im pedance at both lineterminating poles (Section 3.5).

PAGE 116

86 Figure 3-6: Vertically-configur ed distribution line in 2003, arrester station at pole 10. Vertically-configured line, 2001 experiment: Ohio-Brass PDV-100 MOV arresters and Cooper UltraSIL Housed VarisSTAR Heavy Duty were installed on the line (Section 3.4). The neutral conductor was in itially grounde d at poles 5 and 11 and at each arrester station. The groundings at pol es 5 and 11 were removed fo r the later part of the 2001 experim ent and for the experim ents conducted after 2001 (Section 3.3) The lightning current w as injec ted into phase A of the line. Vertically-configured line, 2002 experiment: The 2002 line was similar to the vertically-configured distribut ion line studied in 2001 with the following modifications: Lightning current was not only injected into phase A of the line, but also into ground at 20 m, 30 m, and 100 m distance from the line to investigate induced effects on the line due to nearby lightning. Two arresters in parallel were used on phase A (Section 3.4). Only Ohio-Brass arresters or only Cooper arresters were used on the line (Section 3.4). A U-shaped interception structure was used during the direct strike experiment to divert the initial continuous current from the line (Section 3.1.1).

PAGE 117

87 Vertically-configured line, 2003 experiment: Figure 3-7 shows pole 2 of the distribution line tested in 2003. The 2002 experiment was similar to the 2003 experiment with the f ollowing modifications: Lightning current was injected into phase A of the line and into ground at 7 m, 11 m, and 15 m distance from the line. One arrester was used on phase A in 2003 instead of two parallel arresters that were used in 2002 (Section 3.4). Only Ohio-Brass arresters were used on the line (Section 3.3.3 ). A pole-mount transformer was installed a nd instrumented on phase A at pole 2 ( Figure 3-7). The secondary of the transf ormer was center-tapp ed, and term inated with resistive loads of 4 and 6 ohms. Current was measured at the transformer primary. The transformer was of type ABB 50 kVA. Higher insulation-strength phase A insulators were installed. Figure 3-7: Vertically-confi gured distribution line in 2003, arrester station with transformer at pole 2.

PAGE 118

88 3.2.3 Vertically-configured Line with Overhead Ground Wire, 2004 A vertically-configured di stribution line with overhead ground wire was studied during the summer of 2004. Pole 10 of this distribution line is shown in Figure 3-8 and Figure 3-9. The three phase conductors of the vertically-conf igured line tested from 2001 through 2003 were lowered and the underneath neutral cond uctor was moved above the three phase conductors to function as an overhead ground wire. Inform ation on the placem ent and specifications of the conductors can be found in Table 3-31. Table 3-3: Conductor placement and specifica tions for the vertically-configured test distribution line with overhead ground wire. Height (m) Horizontal Displacement Relative to Overhead Ground Wire (m) Conductors Phase A 9.7 0.3 Phase B 8.9 0.3 Phase C 8.1 0.3 Type: 568, 587.2 MCM, 19-strand Equivalent diameter: 2.23 cm Resistance (dc, 20): 0.099 /km Overhead Ground Wire 11.4 0 Type: AWG 3/0 7-strand Equivalent diameter: 1.18 cm Resistance (dc, 20): 0.34 /km The overhead ground wire was grounded at every pole with a ground lead (a copper wire of 5.1 mm diameter). Stand-offs of 30 cm length provided clearance between the ground lead and the phases (Figure 3-9). One arrester was installed on each of the three phases at each of the four arrester stations at poles 2, 6, 10, and 14 (four spans between arrester stations). The phase arresters were connected to th e overhead ground wire at the nearest point on the grounding down-lead wh ere the down-lead was looped outward to clear the arresters and phase wires. The transform er installed dur ing the 2003 experime nt was rem oved. 1 The conductor heights and horizontal displacements in the table were measured on 06/28/2005 at pole 8.

PAGE 119

89 Figure 3-8: Vertically-configur ed distribution line with overhead ground wire in 2004, arrester station at pole 10. Figure 3-9: Vertically-configur ed distribution line with overhead ground wire in 2004, arrester station at pole 10, close-up view.

PAGE 120

90 3.3 Grounding of the Test Distribution Lines For the 1999 through 2003 experiments the neutral of the horizontally/verticallyconfigured distribution line was grounded at all poles with arre ster stations and at the end poles. For the 2004 experiment the overhead ground wire was grounded at every pole. The grounding at each pole was done by means of vertical ground rods of unknown length. Some grounding consisted of multiple ground rods that were connected with a wire as it can be seen in Figure 3-10. This figure shows a sketch of the vertical distribution line with the ne utral conductor grounded with n ground rods and a shunt (current m easurement resistor) measuring the ground current. The arrangem ent of the multiple g round rods used in the other line configurations was similar. Grounding resistances were determined with the fall-of-potential method (Mata, 2000) or with clamp-on meters. The grounding during the 1999 and 2000 experime nts is discussed in Mata ( 2000 ) and Mata et al. ( 2003). The grounding during the 2001 and 2002 experim ents is discussed in Mata ( 2003). Figure 3-10: Grounding schem e for the vertica lly-configured distribution line. Adapted from Mat a ( 2003).

PAGE 121

91 3.3.1 1999 Experiment, Grounding Four of the six poles (poles 7, 8, 11, and 12) of the horizontally-configured distribution line were grounde d. The grounding resistance was determined by UF with the fall-of-potential method or by FPL with a clamp-on meter. The grounding resistances for the 1999 experiment as reported by Mata ( 2000) are sum marized in Table 3-4. The table also includes the date the grounding resistance was me a sured and the m ethod used for the measurement (fall-of-potential or clamp-on). Table 3-4: Grounding resistances in ohms for the horizonta lly-configured line tested during the 1999 experiment. Date Method Pole 7 Pole 8 Pole 11Pole 12 06/15/99 Clamp-on 41.4 41.6 51 42.6 06/19/99 44.4 35.6 06/21/99 Fall-of-potential 24.6 31.0 3.3.2 2000 Experiment, Grounding Eight of the 18 poles (poles 1, 2, 5, 8, 11, 14, 17, and 18) of the horizontallyconfigured distribution line were grounded. All grounding resistances were determined by UF with the fall-of-potential met hod. The grounding resistances for the 2000 experiment as reported by Mata ( 2000) are sum marized in Table 3-5. The table also includes the date the groundi ng resistance was measured. Gr ounding resistances in bold and ita lic p rints are the v alues used for modeling the current division on the horizontallyconfigured distribution line in Mata ( 2000) and Mata et al. ( 2003). Table 3-5: Grounding resistances in ohm s for the horizonta lly-configured line tested during the 2000 experiment. Date Pole 1 Pole 2 Pole 5Pole 8Pole 11Pole 14Pole 17 Pole 18 07/03/00 46.4 10.6 27.3 07/04/00 40.7 46.8 27.7 51.6 54.9 09/18/00 37.1 22.1

PAGE 122

92 3.3.3 2001 Experiments, Grounding Eight of the 15 poles of the vertically -configured distribut ion line were grounded for the early part (until 7/31/01) of the 2001 expe riment (poles 1, 2, 5, 6, 10, 11, 14, and 15) and 6 of the 15 poles (poles 1, 2, 6, 10, 14, and 15) were grounded for the later part of the experiment (the groundings of poles 5 and 11 were unintentional and were removed on 7/31/01 after they were discovered. During 2001 two flashes with return stroke were triggered with groundings present at poles 5 and 11 and two flashes were triggered after the two groundings were removed.). Single gr ound rods were used when the line was build by FPL during 1999 and 2000. On 06/ 21/01, 06/26/01, and 07/04/01 FPL added additional ground rods to the already existing ones. The lengths of the ground rods are unknown. The multiple ground rod scheme is illustrated in Figure 3-10. Note that all flashes during the 2001 experim ent were tr iggered after the additional grounding rods were installed. The grounding resistances for the 2001 expe rim e nt as reported by Mata et al. ( 2001) are s ummarized in Table 3-6. All grounding resistan ces were measured with a clamp-on meter. The measurem ent on 7/30/01 was done with the gr oundings at poles 5 and 11 in place and the measurement on 7/ 31/01 was done after these two groundings were disconnected. The number given in parenthesis is the nu mber of ground rods used at each pole ground. The numbers of ground rods used for the pole 5 and pole 11 grounds were not reported, but presumably single ground rods were used at these two poles. Grounding resistances given in bold and italic prints are the resistances used by A.G. Mata et al. ( 2002) for modeling the current division on th e vertical distribution line with data obtained during the 2002 experim ent.

PAGE 123

93 Table 3-6: Grounding resistances in ohms of the vertically-c onfigured line tested during the 2001 experiment. The number of ground rods is given in parenthesis. Date Pole 1 Pole 2 Pole 5Pole 6 Pole 10Pole 11 Pole 14 Pole 15 07/30/01 24 (4) 19 (1)25 19.7 (5)17.5 (2)63 29 (5) 24 (5) 07/31/01 24 (4) 20 (1) 18 (5) 17.8 (2) 28 (5) 24 (5) 3.3.4 2002 and 2003 Experiments, Grounding The pole grounds used during the 2002 and 2003 experiments were the same as the pole grounds used during the 2001 expe riment. The grounding resistances were measured again in 2002 ( Table 3-7) with a clamp-on meter and were found to be very sim ilar to the values measured in 20 01. A .G. Mata et al. (2002) suggested a set of theoretically-der ived grounding resistances (Table 3-7) that were chosen to improve the m atch between the ground currents measured during the direct in jection of lightning current into one of the phases and their modeled results. Table 3-7: Measured and th eoretically-derived grounding re sistances in ohm s for the vertically-configured line. The number of ground rods is given in parenthesis. Method Pole 1 Pole 2 Pole 6 Pole 10 Pole 14 Pole 15 Measured 23 (4) 20 (1) 19 (5) 18 (2) 29 (5) 24 (5) Theoretical 46 9 14 12 33 21 3.3.5 2004 Experiment, Grounding FPL re-configured the vertical distri bution line and added an overhead ground wire and additional grounds so that the overhead ground wire was grounded at each pole. The grounding resistances were measured on 7/12/2004 with a clamp-on ground meter. The grounding resistances of each pole and the number of ground rods are given in Table 3-8. The first row in the table gives th e index of the pole (e.g., p1 for pole 1). Table 3-8: Grounding resistances in ohm s measured on 7/12/2004 for the verticallyconfigured line tested during the 200 4 experiment. The num ber of ground rods is given in parenthesis. p1 p2 p3 p4 p5 p6 p7p8p9p10 p11p12p13 p14 p15 22 (4) 19 (2) 28 (2) 28 (3) 27 (2) 18.3 (3) 20 (1) 28 (2) 44 (2) 16.4 (2) 34 (2) 36 (2) 42 (2) 29 (2) 22 (5)

PAGE 124

94 3.4 Arresters on the Test Distribution Lines MOV arresters manufactured by Cooper Power Systems or by Ohio-Brass were installed on the test power distribution lines during the 1999 through 2004 experiments. The Cooper Power Systems arresters were th e UltraSIL Housed VariSTAR Heavy Duty with a rated voltage of 18 kV. The Ohio-Brass arresters we re the PDV-100 with the same rated voltage. The VI-characteristics of the two arrester types in response to an 8/20 s current pulse are found in Table 3-9. Table 3-9: VI-characteristics of the C ooper Power Systems Ultra SIL Housed VariSTAR Heavy Duty and Ohio-Brass PDV-100 arresters. Ohio-Brass Cooper Power Systems Voltage (kV) Current (kA) Voltage (kV) Current (kA) 49 1.5 48.5 1.5 52 3.0 51.6 3.0 55 5.0 53.9 5.0 60 10.0 58.8 10.0 70 20.0 65.0 20.0 82 40.0 73.2 40.0 3.4.1 1999 Experiment, Arresters Single Ohio-Brass arresters were installed at poles 8 and 11 (the only two arrester stations) of the horizontally-configured dist ribution line ( Table 3-10). The arresters were replaced after each storm day with lightning triggered to the line. Table 3-10: 1999 experiment, arresters used on the horizontally-conf igured test line. Pole 8 Pole 11 Phase A Phase B Phase C Ohio-Brass 3.4.2 2000 Experiment, Arresters Single Ohio-Brass arresters were installed at poles 8 and 11 (the two poles closest to the lightning injection point) of the horizontally-configured distribution line and single Cooper Power Systems arresters were installed on all other poles ( Table 3-11). The

PAGE 125

95 arresters installed on phase C (the struck phase) at poles 8 and 11 were replaced after each day during which lightning was triggered to the line. Table 3-11: 2000 experiment, arresters used on the horizontally-conf igured test line. Pole 2 Pole 5 Pole 8 Pole 11 Pole 14 Pole 17 Phase A Phase B Phase C Cooper Power Systems Ohio-Brass Cooper Power Systems 3.4.3 2001 Experiment, Arresters Table 3-12 shows the arresters installed on the vertically-configured distribution line during the 2001 experiment. Single Ohio -Brass and single Cooper Power Systems arresters were installed on the line during the triggering days 7/26/2001 (FPL0101 and FPL0102) and 7/27/2001 (FPL0105, FPL0107, and FPL0108), and only Cooper Power Systems arrester were inst alled during the triggering day 8/18/2001 (FPL0110, FPL0111, and FPL0112). Table 3-12: 2001 experiment, arresters used on the vertically-conf igured test line. Pole 2 Pole 6 Pole 10 Pole 14 Phase A Ohio-Brass Phase B FPL0101-FPL0108 (7/26/2001-7/27/2001) Phase C Cooper Power Systems Ohio-Brass Cooper Power Systems Phase A Phase B FPL0110-FPL0112 (8/18/2001) Phase C Cooper Power Systems 3.4.4 2002 Experiment, Arresters Table 3-13 shows the arresters installed on the vertically-configured distribution line during the 2002 experiment. Cooper Power Systems arresters were installed on the line during the triggering days 7/9/2002 (FPL0205, FPL0206, FPL0208, and FPL0210), 7/19/2002 (FPL0213), 7/20/2002 (FPL0218, FPL0219, FPL0220, and FPL0221), and 7/25/2002 (FPL0226) and Ohio-Brass arrester were installed during the triggering days 8/2/2002 (FPL0228, FPL0229, and FPL0230), 8/18/2002 (FPL0236), 8/27/2002

PAGE 126

96 (FPL0240), 8/28/2002 (FPL0241), and 9/13/200 2 (FPL0244, FPL0245, and FPL0246). During the 2002 experiment 2 arresters in para llel were installed on phase A (the struck phase) and single arresters were installed on phase B and phase C. All phase A arresters were replaced with new arresters of the same type (Cooper Power Systems or OhioBrass) after each storm day with lightning triggered to the line. Table 3-13: 2002 experiment, arresters used on the vertically-conf igured test line. Pole 2 Pole 6 Pole 10 Pole 14 Phase A Cooper Power Systems (2 arresters in parallel) Phase B FPL0205FPL0226 (7/9/2002/25/2002) Phase C Cooper Power Systems (single arresters) Phase A Ohio-Brass (2 arresters in parallel) Phase B FPL0228FPL0246 (8/2/2002/13/2002) Phase C Ohio-Brass (single arresters) 3.4.5 2003 Experiment, Arresters Single Ohio-Brass arresters and one transf ormer-mounted arrester were installed on the line during the 2003 experiment ( Table 3-14). The specifications of the transformermounted arrester are not known. No arrester was replaced during the 2003 experiment. Table 3-14: 2003 experime nt, arresters used on the vertically-conf igured test line. Pole 2 Pole 6 Pole 10 Pole 14 Phase A transformermounted Phase B Phase C Ohio-Brass Ohio-Brass 3.4.6 2004 Experiment, Arresters Single Ohio-Brass arresters were installed on the line during the 2004 experiment ( Table 3-15). The phase arrest ers were connected to the overhead ground wire at the nearest point on the grounding down-lead, wher e the down-lead was looped outward to clear the arresters and phase wires. No arrester was replaced dur ing the 2004 experim ent.

PAGE 127

97 Table 3-15: 2004 experiment, arresters used on the vertically-configur ed test line with overhead ground wire. Pole 2 Pole 6 Pole 10 Pole 14 Phase A Phase B Phase C Ohio-Brass 3.5 Line Terminators on the Test Distribution Lines The test distribution lines were terminated in their characteristic impedances (500 ) to eliminate reflections at the line ends and thereby to simulate to some extent infinitely long lines. Terminating impedances were installed at both line ends between each phase and neutral conductor. Figure 3-11 shows the lin e termination of the vertically-configured distribut ion line at pole 15. The pict ure was taken during the 2003 experiment. The sam e term ination resistors were used for all e xperiments during 2000 through 2004 but not for the 1999 experiment. Figure 3-11: Vertically-c onfigured distribution line in 2003, line termination at pole 15.

PAGE 128

98 3.5.1 1999 Experiment, Line Terminators From 08/08/99 to 08/16/99: Singl e-impulse 6.2 kJ rated 433 termination resistors were installed on each phase at each termin ation pole. The resistors were composed of 13 serial sections of six 200 k, 80 kJ Ohmite resistors connected in parallel. The phase B and C terminators were destroyed during a triggered flash. From 08/22/99 to 08/24/99: No termination im pedances were installed on the line, that is, the line end was open-circuited. From 09/06/99 to 09/10/99: The terminating impedances consisted of wirewound Power Technologies, Inc. (PTI) 25 kJ, 500 resistors connected in parallel with 16 mH inductors. The purpose of the indu ctors was to protect the resistors by bypassing the low-frequency current components. 3.5.2 2000 Experiment, Line Terminators The terminating resistors used for this experiment were high energy absorption resistors manufactured by High Power T echnologies. The specifications of the termination resistors used during the 2000 horizontally-conf igured distribution line experiment can be found in Table 3-16. Table 3-16. Term ination resistors during th e 2000 experiment (horizontally-configured distribution line). Phase A Phase B Phase C Pole 1 single 524 1.25 MJ resistor enclosed in PVC housing single 510 1.25 MJ resistor enclosed in PVC housing 1027 and 964 resistors (each 1.75 MJ) connected in parallel enclosed in PVC housing and immersed in oil Pole 18 single 506 1.25 MJ resistor enclosed in PVC housing single 480 1.25 MJ resistor enclosed in PVC housing 1051 and 911 resistors (each 1.75 MJ) connected in parallel enclosed in PVC housing and immersed in oil 3.5.3 2001 Experiments, Line Terminators The same high energy absorption resistors th at were used in 2000 were used for the 2001 through 2003 experiments (v ertically-configured distri bution line) and for the 2004 experiment (vertically-configured distributi on line with overhead ground wire). For the 2004 experiment the terminating resistors we re connected between the phase and the overhead ground wire or between the phase and the ground wire that connects to the

PAGE 129

99 overhead ground wire. The specifications of th e termination resistors used during the 2001 through 2004 experiments can be found in Table 3-17. Table 3-17: Termination resistors du ring the 2001 through 2004 experiments. Phase A Phase B Phase C Pole 1 1027 and 964 resistors (each 1.75 MJ) connected in parallel enclosed in PVC housing and immersed in oil single 510 1.25 MJ resistor enclosed in PVC housing Single 524 1.25 MJ resistor enclosed in PVC housing Pole 18 1051 and 911 resistors (each 1.75 MJ) connected in parallel enclosed in PVC housing and immersed in oil single 480 1.25 MJ resistor enclosed in PVC housing single 506 1.25 MJ resistor enclosed in PVC housing 3.6 Instrumentation The instrumentation used during the 1999 th rough 2004 experiments is described in this section. References to additional instrumentation information are provided. Figure 3-12 shows the measurement scheme. Th e signals measured by current or voltage sensors (Section 3.6.1 and Section 3.6.2 ) w ere transmitted to the Launch Control trailer ( Figure 3-1) via fiber optic links (Section 3.6.3 ) where th ey w ere low-pass filtered (Section 3.6.4) and stored on oscilloscope s or magnetic tape (Section 3.6.5 ) tha t were connected to the fiber optic receiver with coaxial cable s. In 2002, 2003, and 2004 the transmitters w ere controlled remotely from the Launch Control trailer by a wireless control system (Section 3.6.6), allowing rem ote turning on and off, signal calibration, and setting of attenuation levels for each transmitte r. 3.6.1 Current Sensors Currents were sensed with current viewing resistors (CVRs) or current transformers (CTs). The CVRs have the advantage that th ey have a dc lower frequency response and the disadvantage that they are typically more difficult to install (the current-carrying wire has to be cut to include a CVR in the meas ured current path). The CTs on the other hand

PAGE 130

100 Figure 3-12: Measurement scheme used during the 1999 through 2004 experiments. Adapted from Mata ( 2000).

PAGE 131

101 are easy to install (they are lighter and one type of CTs can be clamped on the currentcarrying wire) but have the disadvantage of a finite lower frequency response. A detailed description of the current sensors used during the 1999 through 2004 experiments is found in Mata ( 2000). Current transformers: Current transformers (CTs ) manufactured by Pearson Electronics, Inc. were used in the experiment to measure currents on the test distribution line other than the ground currents. Thr ee different models (110A, 3025C, and 3525) were employed. Their specifications including the V/A value, the maximum peak current (the largest current value that can be meas ured), the maximum I-t product (transferred charge that exceeds the I-t produc t saturates the transformer co re resulting in a distorted signal), the minimum usable rise time of the input signal (smaller rise times result in overshoot or ringing in the output signal), and the frequency response are found in Table 3-18. Each CT has a 50 output impedance. The CTs used for measuring phase currents and phase-to-neutral currents (that is, arrester, transformer, and terminator currents) produce accurate results for only a few milliseconds of current variation because of magnetic core saturation. Table 3-18: Specifications of the Pearson Electronics, Inc. current transformers. Model Output [V/A] Peak Current [kA] I-t Product [C] 10-90% Rise Time [ns] Frequency Response 110A 0.1 10 0.49 20 1 Hz MHz 3025 0.025 20 3.2 100 7 Hz MHz 3525 0.1 10 0.5 25 5 Hz MHz Current viewing resistors: Current Viewing Resistors (CVRs) manufactured by T&M Research Products, Inc. were used in the experiment to m easure lightning channel currents and current flowing to ground from the test distribution lines. Two different

PAGE 132

102 models (R-5600-8 and R-7000-10) were employe d. Their specifications including the V/A value, the energy and power rating, the minimum usable rise time of the input, the output impedance, and the frequency response are found in Table 3-19. Table 3-19: Specificatio n of the T&M Research Products, In c. current viewing resistors. 3.6.2 Voltage Dividers Both magnetic-flux compensated voltage dividers and capacitive compensated resistive dividers were designed and build by the University of Florida to measure phaseto-phase and phase-to-neutra l voltages in the 1999 and 2000 experiments. A detailed description of these voltage divide rs is found below and in Mata ( 2000). Capacitive-compensated voltage dividers: The high-voltage arm of the capacitive-compensated resistive dividers wa s a high-voltage resistor connected in parallel with a high-voltage capacitor. The high-voltage re sistor was composed of a number of resistors connected in series (47 k, 80 J, 3 W, 20% rated Ohmite resistors or 1 MW, 35 kV, 9 W, 10% rated Ohmite resi stors). The high-voltage capacitor was composed of a number of capacitors connect ed in series (500 pF 30 kV DC, 20% rated Cera-Mite capacitors or 100 pF, 15 V DC, 20% rated Cera-Mite capacitors). The lowvoltage arm was a single 50 low-voltage resistor (the ch aracteristic impedance of a coaxial cable terminated into 50 ) connected in parallel with a single low-voltage capacitor. The function of the highand lowvoltage capacitors was to improve the upper frequency response of the voltage dividers by minimizing the effects of parasitic Model Output [V/A] Energy Rating [kJ] Power Rating [W] 10% Rise Time [ns] Output Impedance [] Frequency Response [MHz] R-5600-8 0.00125 5.2 225 30 0.00125 0 R-7000-8 0.001 7.0 225 45 0.001 0 R-2800-4 0.0025 2.8 225 8 0.0025 0

PAGE 133

103 capacitances across the indivi dual resistors. The specifi cations of the capacitivecompensated voltage dividers including their rating, high voltage (H V) resistance (total resistance = nR individual resist or value, where nR is the number of serially connected resistors), HV capacitance (total capaci tance = individual capacitor value/nC, where nC is the number of serially connected capacito rs), low voltage (LV) resistance, LV capacitance, and the nominal VD ratio are given in Table 3-20. Table 3-20 : Specifications for the capacitive-c ompensated voltage dividers used in the 1999 and 2000 experime nts. VD Rating HV Resistan ce (k) HV Capacitance (pF) LV Resistance () LV Capacitance (nF) Nominal VD Ratio 300 kV 10 47 500/10 50 470 1.0637 10-4 350 kV 5 1000 100/20 50 1000 1.0 10-6 1.41 kV 47 47 500/47 50 470 2.2634 10-5 Magnetic-flux-compensated voltage dividers: The magnetic-flux-compensated voltage divider was designed to minimize ma gnetic field induction effects due to high rate of rise current impulses. The basic principle of the magn etic-flux compensated voltage divider, as shown in Figure 3-13, is to cancel out the induced voltage in the two m easuring loops (upper part) with the indu ced voltage in the two compensating loops (low er pa rt). Ideally, the electromotive forces emf1 and emf2 generated by the magnetic fluxes B1 and B2, respectively, are cancelled by the electromotive forces emf3 and emf4 generated by the magnetic fluxes B3 and B4, respectively. The f our-loop design of this voltage divider minimizes the effect of a possible non-symmetric cu rrent division. The individual voltage divider in each loop is capacitive-compensated (see previous paragraph). The specifications for each of the four loops of the capacitive-compensated voltage dividers including their rating, high voltage (HV) resistance (total resistance = nR

PAGE 134

104 individual resistor value, where nR is the number of serially connected resistors), HV capacitance (total capacitance = individual capacitor value/nC, where nC is the number of serially connected capacitors), low voltage (LV) resistance, LV capacitance, and the nominal VD ratio are given in Table 3-21. Note the two meas uring loops are connected in parallel and consequently the actual HV resistan ce is half of the value given in the table and the actu al capacitan ce is twice of the valu e given in the table. The same type of Figure 3-13: Magnetic-flux compensated vol tage divider. Adapted from Mata ( 2000).

PAGE 135

105 resistors (47 k and 1 M) and capacitors (100 pF) us ed to build the capacitivecompensated voltage dividers described above were used to build the magnetic-fluxcompensated voltage dividers. Table 3-21: Specifications for each of the f our loops of the magnetic-flux-compensated voltage dividers used in the 1999 and 2000 experiments. VD Rating HV Resistance (k) HV Capacitance (pF) LV Resistance () LV Capacitance (nF) Nominal VD Ratio 75 kV 5 47 100/5 50 94 3.2816 10-5 175 kV 5 1000 100/10 50 1000 1.9011 10-6 3.6.3 Fiber Optic Link The analog signals from the current sensors and voltage dividers were transmitted via fiber optic links to the Launch Control trailer where th ey were digitized. The fiber optic links are composed of Nicolet Isobe 3000 receiver-transmitter pairs and connecting fiber optic cables and are desc ribed in detail in Mata ( 2000) and Jerauld ( 2003). Nicolet Isobe 3000 receiver-transmitter: The Nicolet Isobe 3000 receivertransmitter pair has an input impedance of 1 M, an output impedance of 50 and a -3 dB bandwidth of 15 MHz. The input range of the transmitter is selectable from .1 V, V, and V and the output range of the receiver is fixed at V ( V with a 50 termination). The transmitters were battery operated and mounted in shielded containers at the sensor locations. In 2002 and 2003 the transmitters were controlled remotely from the Launch Control trailer by a wire less communication system (Section 3.6.6), allowing their rem ote turning on and off, signal calibration, and setting attenuation levels for each transmitte r. The rece ivers were housed in th e Launch Control trailer where they were powered by batteries or an inde pendent generator. E ach receiver-transmitter pair requires two optical fibers.

PAGE 136

106 Fiber optic cables: The fiber optic cables used during the 1999 through 2001 experiments were 200 m Kevlar reinforced duplex two-fiber cables manufactured by OFS Fitel Corporation. These cables were be ing chewed and damaged by animals. For the 2002 through 2004 experiments 200 m six-fiber armored cables manufactured by OFS Fitel Corporation that provided better protection from animals were used. The six fibers are color coded and tw isted around a strength member. Both types of fiber optic cables had SMA connectors. Fiber optic link delay: The fiber optic link delay is th e sum of the fiber delay (the transmission time of the signal through the fibe r optic cable) and the electronic delay (the time it takes for the sensor signal to be processed by the fiber optic transmitter-receiver pair). The fiber delay primarily depends on th e lengths of the cable and is between a few hundreds of nanoseconds and a few microseconds for the cables used in our experiments (the cable lengths for different measurem ents ranged from 100 m to over 500 m). The electronic delay is much smaller (tens of nanosecond) and consequently the major contributor toward the total delay of the fiber optic link is the fiber delay. The fiber optic link delay can be determined from theory by summing 70 ns for the electronic delay combined with the fiber delay of one meter of fiber optic cable and 5 ns per meter of fiber optic cable beyond the first meter for the additio nal fiber delay. Alte rnatively, the fiber optic link delay can be determined experimental ly by injecting a test signal into one end of the link and measuring the ti me delay of the reflection from the other end of the link as it was done for the delays in the 1999 and 2000 experiments in Mata ( 2000) and for the delays in the 2001 and 2002 experim ents in Mata ( 2003). N o te that some of the experimentally-determ i ned delays in Mata ( 2000) and Mata ( 2003) are apparently

PAGE 137

107 erroneous (the fiber optic links with shorter fiber optic cabl es have a much larger delay than the links with longer cables, which is unr easonable since the fiber delays present the largest contribution to the fibe r optic link delay as noted above ). The erroneous delays are all delays for the 2000 experiment except th e delays for events FPL0011 and FPL0014 in Mata ( 2000) and all delays for the 2002 experim ent in Mata ( 2003). 3.6.4 Anti-aliasing Filter Anti-aliasing filters1 designed and built by Carlos Mata were used for all measurements during the 1999 experiment. The filters were designed to have a -3 dB attenuation of 5 MHz and were connected to the Ni colet Isobe 3000 receivers ( Figure 312). The filters are describe d in de tail in Mata ( 2000). 3.6.5 Data Recording Equipment One of the difficulties of recording lightning parameters is that the duration of some lightning features differs by several orders of magnitudes. For instance, the lightning return stroke occurs on a tens of microseconds time scale while the continuous current portion in a rocket-triggered or natu ral lightning flash and the time between return strokes lasts for tens to hundreds of millis econds. This difficulty makes it necessary to adopt a data recording scheme that includes obtaining a fast (high-bandwidth) and short record of the features associated with the re turn stroke and a slow (low-bandwidth) fullflash record. All 2000 through 2004 fast records were acquired with LeCroy digitizing oscilloscopes that were set to sample between 10 MHz and 50 MHz in segmented 1 Aliasing is an effect that can occur if a continuou s time signal is sampled (converted into a discrete time signal) with an insufficient samp ling frequency. An aliased discrete time signal contains high-frequency components present in the continuous time signal that are indistinguishable from genuine low-frequency components making the reconstruction of the continuous time signal fro m the digitized signal impossible. The Nyquist criterion states that in order to avoid aliasing the sampling frequency has to be larger than twice the highest frequency of the signal. Anti-aliasing filters avoid aliasing by reducing the bandwidth of the continuous time signals before they are sampled.

PAGE 138

108 memory mode. In segmented memory mode each channel is divided into many segments and each segment requires a separate trigger si gnal. This mode is ideal to record the parameters associated with the short-duration return strokes (each return stroke consecutively triggers one of the segments). The disadvantage of this recording method is that lightning currents that do not produce a large enough si gnal to trigger a segment (such as small return strokes, most initial continuous currents, and most M-components) are not recorded. Also, the number of return strokes in a flash can exceed the number of segments in a channel so that return strokes that may occur after the last segment was triggered are not recorded. A continuous r ecord of some important parameters was obtained with an analog tape recorder (in 1999 and 2000) or a digitiz ing oscilloscope (in 2001 through 2004) with a sample setting between 500 kHz and 2 MHz. These slow records were not suitable to capture the high-f requency features in the flash (for instance, currents associated with the fast rising edge of the return stroke) but provided added redundancy and a full flash record that incl uded the low-magnitude continuous currents and all return strokes. Table 3-22: Recording devices used dur ing the 1999 through 2004 experiments. The number of recording devices used for a particular year is given. 1999 2000 2001 2002 2003 2004 Nicolet Pro 90 1 Nicolet Multipro 1 LeCroy 9354 2 1 1 LeCroy Waverunner LT344L 7 7 6 7 4 LeCroy LC574AL 1 LeCroy 9384AL 1 Yokogawa DL716 1 2 2 1 Honeywell 101 1 1 The digitizing oscilloscopes and the magnetic tape r ecorder used during the 1999 through 2004 experiments are described in detail in Mata ( 2000) and Jerauld ( 2003).

PAGE 139

109 Table 3-22 lists the model and quantity of recording devices used during the 1999 through 2004 experiments. Some spe cifications of the recording devices are provided in the following paragraphs. Nicolet Pro 90: This digitizing oscilloscope is a f our-channel recorder with a data capacity of 259 kilosamples per channel. Channe ls 1 and 2 have 8 bit vertical resolution and a maximum sampling rate of 200 MHz. Channels 3 and 4 have 12 bit vertical resolution with a maximum sampling rate of 10 MHz. Some parameters were recorded with two channel sequentially connected to increase the record length. Nicolet Multipro: This digitizing oscill oscope is a four-card recorder, each card having four channels with 12 bit vertical re solution, a data capacity of 517 kilosamples per channel, and a maximum sampling rate of 100 MHz. Some parameters were recorded with two channel sequentially connected to increase the record length. LeCroy 9354: This digitizing oscilloscope is a four-channel recorder with a data capacity of 400 kilosamples per channel. The segmentable channels have 8 bit vertical resolution, a maximum input bandwidth of 500 MHz, and a sampling rate of up to 2 GHz. The oscilloscope has a 500 MB PCMCIA ha rd drive or a floppy disk drive for data storage. LeCroy Waverunner LT344L: This digitizing oscilloscope is a four-channel recorder with a data capacity of 1 megasa mples per channel. The segmentable channels have 8 bit vertical resolution and a maximu m input bandwidth and sampling rate of 500 MHz. An internal low-pass filter of values 25 MHz or 200 MHz (-3 dB) can be used on each channel. The oscilloscope has a 500 MB PCMCIA hard drive or a 128 MB PCMCIA compact flash card fo r data storage and is capab le to move data through a

PAGE 140

110 10Base-T Ethernet connection us ing a proprietary protocol for file transfer. The LT344L can be remotely configured over an IEEE 488.2 (GPIB) bus. LeCroy LC574AL: This digitizing oscilloscope is a four-channel recorder with a data capacity of 2,000,000 samples per channe l. The segmentable ch annels have 8 bit vertical resolution, a maximu m input bandwidth of 1 GHz, and a sampling rate of up to 4 GHz. The oscilloscope has a 500 MB PCMC IA hard drive for data storage. Yokogawa DL716: This digitizing oscilloscope is a sixteen-channel recorder with a data capacity of 16 megasamples per channel when all 16 channels are used simultaneously. Each channel has 12 bit vert ical resolution, a maximum input bandwidth of 4 MHz (-3 dB), and a sampling rate of up to 10 MHz. An internal low-pass filter of values 500 Hz, 5 kHz, 50 kHz, or 500 kHz (-3 dB) can be used on each channel. The oscilloscope has a 9.2 GB internal hard drive for data storage and is capable to move data through a 10Base-T Ethernet connection using the File Transfer Protocol (FTP). The DL716 can be remotely configured over an IEEE 488.2 (GPIB) bus. Honeywell 10: This magnetic tape recorder/rep roducer has 16 analog channelsdirect record channels or FM record channels. The input range of the direct record channels is 0.1.0 V or 1.0.1 V with a bandwidth of 100 Hz to 2 MHz and a source impedance ranging from 75 The input range of the FM record channels is V with 20 k input impedance or .24 V with 75 input impedance. The bandwidth of the FM record channels is DC to 500 kHz. 3.6.6 Wireless Control System The wireless control system employe d during the 2002 through 2004 experiments is described in detail in Jerauld ( 2003). This system is composed of a PIC controller

PAGE 141

111 (named after the PIC 16F873-207SP microproces sor it contains), a control PC, and a wireless link ( Figure 3-14). The PIC controllers were designed and developed by Michael Stapleton and Keith Ram bo of the University of Florida. A PIC controller is located at each measurement and can activate and d eactivate the measurement, does provide attenuation to the sensor, can check the stat us of the measurements battery, and can send a calibration signal for the fiber optic link a ssociated with the measurement. The PIC controller can be controlled remotely from the control PC located in launch control through a fiber optic link. The fiber optic link is a wireless 900 MHz RF unit (aka PIC RF unit) equipped with an Agilent HFBR-1523/ 2523 fiber-optic trans ceiver pair and two short lengths of plastic fibe rs that connect the PIC RF unit and the PIC controller. Figure 3-14: Diagram of the wireless control system topology used during the 2002 through 2004 experiments. Adapted from Jerauld ( 2003).

PAGE 142

112 CHAPTER 4 CHAPTER 4: EXPERIMENTAL CONFIG UR ATIONS AND RESULTS The experimental configurations for the 1999 through 2004 experiments are described in this chapter including a list of the measured parameters and references to the measurement settings for each flash. 4.1 Description and Terminology of Measured Parameters The lightning current at the lightning ch annel base was measured close to the rocket launcher (either the stationary la uncher located on the to wer or the mobile launcher both being described in Section 3.1) and was recorded with a high and a low current m easurement (except in 1999 where only a high current measurem e nt was taken). The high current measurement is supposed to capture an unsaturated record of the lightning current. The l ow current record is redundant in the sense that it records the signal from the same current sensor as the high current r ecord, but with lower attenuation to improve the dynamic range of th e current record. The low current record is useful to identify and analyze low magnit ude lightning current features, such as the initial continuous current (ICC) or M-compone nts, but is usually not suitable for the analyses of return stroke current features since return stroke currents are typically saturated in the low current record. A current intercepti on device to divert the ICC from the line was employed from 2002 through 2004 (Section 3.1.1). Both the current through the interception device a nd the current flowing to the line were measured. L ine currents (currents on phase and neutral conduc tors), line-to-line currents (currents flowing from one line conductor to another one, such as arrester and termination resistor

PAGE 143

113 currents), and ground currents were measured on the test distributi on line in each year. Additionally, voltages between phase conductors and between phase and neutral conductors were measured during the 1999, 2000, and 2001 experiments. The following labeling scheme for the measured parameters is adopted: In the figures, line current labels and pha se-to-phase voltage labels are green. In the figures, phase-to-neutral curr ent and voltage labels are red. In the figures, ground curre nt labels are brown. The first letter indicates the type of m easured parameter-either I for current or V for voltage. An i in the subscript stands for incident and refers to the ch annel base current. A low in the subscript indicate s a low current measurement. A high in the subscript indicates a high current measurement. A dc in the subscript indicates that the oscilloscope recorded in DC-coupling mode (this is the default setting). An ac in the subscript indicates that th e oscilloscope recorded in AC-coupling mode. A sr in the subscript stands for strike ring and refers to the current through the current interception structure. A tower in the subscript refers to the current flowing to the tower ground when the current interception structure was inst alled on the line. This current was not injected into the line. A bucket in the subscript refers to th e current flowing to ground from lightning triggered to the bucket truck. A number in the subscript indicates the number of the pole at which the measurement is located. A G in the subscript stands for Ground and refers to the current flowing from the neutral conductor or the overhead ground wire to the pole grounding. An A, B, C, N, and O in th e subscript stands for phase A conductor, phase B conductor, phase C conductor, neutral conductor, and overhead ground wire, respectively. A single letter in th e subscript refers to a line current measurement (e.g., IO10 is the current through the overhead ground wire measured at pole 10). Two letters in the subscript refer to a line-to-line current or voltage measurement (e.g., in the vertical line experiment IAN6 is the current from the phase A conductor to the neutral conductor through the phase A arrester at pole 6). 4.2 1999 Experiment, Horizontally-configured Line A total of 23 different parameters were measured in the 1999 experiment currents and 7 voltages. All measured parame ters and the two different lightning current injection points (between poles 9 and 10 at midspan to phase C for flashes FPL9904

PAGE 144

114 Figure 4-1: 1999 test di stribution line having a horizontal framing configuration with measurement points and the tw o different lightning stri ke locations identified. a) Line termination resist ors, b) no line termination, and c) line termination resistors with inductors connected in parallel.

PAGE 145

115 FPL9912 and to phase B for flashes FPL9914FPL9917) are identified in Figure 4-1. All parameters were sampled at 10 MHz w ith two Le Croy and two Nicolet digital oscilloscopes and some selected parameters we re also recorded with the Honeywell tape recorder (Section 3.6.5 for a desc ription of the re cordi ng equipm ent). Three different line terminations were used duri ng the 1999 experiment (Section 3.5)line termination with resistors shown in Figure 4-1a, no line term inations shown in Figure 4-1b, and line term inations with resistors and inducto rs connected in parallel shown in Figure 4-1c. Note that for UF9914UF9917 additional current measure ments (IA9, IB9, and IC9) and for events UF9914 and UF9915 additional voltage measurements (VBA9, VBN9, and VCN9) were taken as shown in Figure 4-1c. Table 4-1 lists the date and event IDs of all rockettriggered lightning strikes that were injected into the test line in 1999 and a reference to the m easurement settings for each ev ent. Table 4-1: 1999 experim ent, references to the measurement settings for all rockettriggered lightning strikes. All strikes triggered to the test line in 1999 contained return strokes. Date Event ID Reference to Measurement Settings 8/16/1999 FPL9904 Mata ( 2000) Table C1, pages 175 FPL9911 8/24/1999 FPL9912 Mata ( 2000) Table C2, pages 177 FPL9914 9/6/1999 FPL9915 Mata ( 2000) Table C3, pages 179 FPL9916 9/10/1999 FPL9917 Mata ( 2000) Table C4, pages 182 4.3 2000 Experiment, Horizontally-configured Line A total of 31 different parameters were measured in the 2000 experiment currents and 5 voltages. All measured parameters and th e two different lightni ng current injection points (to phase C between poles 9 and 10 at midspan for flashes FPL0011FPL0036 and to phase C at pole 9 for flash FPL0037) are identified in Figure 4-2. All parameters were

PAGE 146

116 Figure 4-2: 2000 test dist ribution line having a horizontal framing configuration with m easurement points and the two different li g htnin g strike locations identified.

PAGE 147

117 sampled at 20 MHz or 25 MHz with ten LeCr oy digital oscilloscopes and some selected parameters were also recorded with the Honeywell tape recorder (Section 3.6.5 for a description of the recordi ng equipment). Low and high current records of the lightning channel base current (Ii) were taken in this experiment and in all later experiments. Table 4-2 lists the date and event ID s of all rocket-triggered lightning strikes that were injecte d into the test lin e in 2000 and a reference to the measurement settings for each event. Note that flash FPL 0025 was a flash without return strokes that did not trigger the oscilloscopes. Consequent ly, data and measurement settings for this flash are not given. Table 4-2: 2000 experiment, references to the measurement settings for all rockettriggered lightning events. Event IDs in bold printing denote flashes with return strokes and event ID s in italic printing denote flashes without return strokes. Date Event ID Reference to Measurement Settings FPL0011 7/16/2000 FPL0014 Mata ( 2000) Tables C5 and C6, pages 185 7/20/2000 FPL0018 Mata ( 2000) Tables C7 and C8, pages 188 7/28/2000 FPL0025 N/A FPL0032 FPL0033 8/2/2000 FPL0034 FPL0035 8/3/2000 FPL0036 Mata ( 2000) Tables C9 and C10, pages 191 8/6/2000 FPL0037 Mata ( 2000) Tables C11 and C12, pages 195 4.4 2001 Experiment, Vertically-configured Line A total of 31 different parameters were measured in the 2000 experiment currents and 6 voltages. All measured paramete rs and the three diffe rent lightning current injection points (to phase A at pole 8 for flashes FPL0101FPL0108, to phase A between

PAGE 148

118 Figure 4-3: 2001 test di stribution line having a vertical framing configuration with measurement points and the three di fferent lightning strike locations identified. poles 7 and 8 at midspan for flashes FPL0110FPL0112, and to ground 20 m north of pole 8 for flash FPL0115) are identified in Figure 4-3. Arrester voltage measurements (VAN2, VAN6, VBN6, and VCN6) were present during fl ashes FPL0101FPL0108 only. All parameters were sampled in segments at 20 MHz with seven LeCroy digital oscilloscopes and some selected parameters were also continuously sampled at 1 MHz with one Yokogawa digital oscilloscope (Section 3.6.5 for a description of the recording

PAGE 149

119 equipment). Table 4-3 lists the date and event IDs of all rocket-triggered lightning events that were injected into th e test line or into ground in 2001 and a reference to the m easurement settings for each event. Table 4-3: 2001 experim ent, references to the measurement settings for all rockettriggered lightning events. Event IDs in bold printing denote flashes with return strokes and IDs in italic printing denote flashes without return strokes. Date Event ID Reference to Measurement Settings FPL0101 7/26/2001 FPL0102 FPL0105 FPL0107 7/27/2001 FPL0108 Mata ( 2003) Table B-1, pages 121 FPL0110 FPL0111 8/18/2001 FPL0112 Mata ( 2003) Table B-2, pages 124 9/5/2001 FPL0115 Mata ( 2003) Table B-3, pages 127 4.5 2002 Experiment, Vertically-configured Line A total of 26 different current paramete rs were measured in the 2002 experiment. No voltages were measured in this experime nt or in experiments performed after 2002. The lightning current was intended to be either directly injected into the phase A of the test line to investigate direct strike effects on the line (7/9 /2002/2/2002) or injected into ground to investigate induced effects due to nearby strikes on the line (8/18/2003 9/13/2003). The current measuremen t stations on the test line were identical to the ones used in 2001. One additional current measur ement station was implemented on the tower launcher to measure the current through the current inte rception structure (Section 3.1.1). All m easured parameters and the four diffe rent lightning current injection points (to phase A between poles 7 and 8 at mids pan for flashes FPL0205FPL0230 except flash FPL0213 and stroke FP L0219-1, to ground 20 m north of pole 8 for flash FPL0213 and stroke FPL0219-1, to ground 100 m north of pole 7 for flash FPL0236, and to ground

PAGE 150

120 Figure 4-4: 2002 test di stribution line having a vertical framing configuration with measurement points and the four different lightning strike locations identified. 30 m north of pole 7 for flashes FPL0240FPL0246) are identified in Figure 4-4. Note that the return strokes in fl ash FPL0213 and stroke FPL0219 were intended to be injected into the line and instead were injected into the tower gr ound. These strokes can be used to investigate induced effects on the line. Also note that flash FPL0241 was a flash without return strok es that di d no t trigger the oscill oscopes. Consequently, data and measurement settings for this flash are not given. All pa rameters were sampled in segments at 20 MHz with seven LeCroy digital oscilloscopes a nd sampled continuously at 1 MHz with two

PAGE 151

121 Yokogawa digital oscilloscopes (Section 3.6.5 for a description of the recording equipment). Table 4-4 lists the date a nd event IDs of all rocket-triggered lightning strikes that were injected into th e test line or into ground in 2002 and a reference to the m easurement settings for each event. Table 4-4: 2002 experim ent, references to the measurement settings for all rockettriggered lightning events. Event IDs in bold printing denote flashes with return strokes and IDs in italic printing denote flashes without return strokes. Date Event ID Reference to Measurement Settings FPL0205l FPL0206 FPL0208 7/9/2002 FPL0210 Mata ( 2003) Table B-4, pages 130 Table B-5, page 132 7/19/2002 FPL0213 FPL0218 FPL0219 FPL0220 7/20/2002 FPL0221 Mata ( 2003) Table B-6, pages 134 Table B-7, page 136 7/25/2002 FPL0226 Mata ( 2003) Table B-8, pages 137 Table B-9, page 139 FPL0228 FPL0229 8/2/2002 FPL0230 Mata ( 2003) Table B-10, pages 140 Table B-11, page 143 8/18/2002 FPL0236 Appendix A : Table A-1 8/27/2002 FPL0240 Appendix A : Table A-2 8/28/2002 FPL0241 N/A FPL0244 FPL0245 9/13/2002 FPL0246 Appendix A : Table A-3 4.6 2003 Experiment, Vertically-configured Line A total of 27 different current paramete rs were measured in the 2003 experiment. The lightning current was either directly injected into the phase A of the test line to investigate direct strike effects on the line (6/30/2003/14/2003) or in jected into ground l Altitude trigger flash, no return strokes injected into the line.

PAGE 152

122 Figure 4-5: 2003 test di stribution line having a vertical framing configuration with measurement points and the four different lightning strike locations identified.

PAGE 153

123 to investigate induced effects on the lin e due to nearby stri kes (7/18/2003/15/2003). The current measurement stations were iden tically to the ones in 2002 except for one additional measurement during the direct strike experiment that measured either the sum of the pole 2, phase A arrest er and transformer currents ( FPL0301) or the pole 2, phase A transformer current (FPL0305FPL0317). The transformer was removed for the nearby strike experiment. All measured parameters and the four differe nt lightning current injection points (to phase A between poles 7 and 8 at midspan for flashes FPL0301 FPL0317, to ground 7 m south of pole 4 for flashes FPL0321FPL0331, to ground 15 m Table 4-5: 2003 experiment, references to the measurement settings for all rockettriggered lightning events. Event IDs in bold printing denote flashes with return strokes and event ID s in italic printing denote flashes without return strokes. Date Event ID Reference to Measurement Settings FPL0301 FPL0302 6/30/2003 FPL0303 Appendix A : Table A-4 FPL0305 7/6/2003 FPL0306 Appendix A : Table A-5 FPL0310 7/11/2003 FPL0311 Appendix A : N/A FPL0312 7/13/2003 FPL0314 Appendix A : Table A-6 FPL0315 7/14/2003 FPL0317 Appendix A : Table A-7 7/18/2003 FPL0321 Appendix A : Table A-8 FPL0326 7/21/2003 FPL0327 N/A FPL0329 7/22/2003 FPL0331 Appendix A : Table A-9 7/26/2003 FPL0335 N/A 8/2/2003 FPL0336 Appendix A : Table A-10 8/7/2003 FPL0341 Appendix A : Table A-11 FPL0342 8/11/2003 FPL0345 Appendix A : Table A-12 FPL0347 FPL0348 8/15/2003 FPL0350 Appendix A : Table A-13

PAGE 154

124 south of pole 4 for flashes FPL0335 and FPL 0336, and to ground 11 m south-east of pole 15 for flashes FPL0341FPL0350) are identified in Figure 4-5. Note that flashes FPL0310, FPL0311, FPL0326, FPL0327, and FPL0335 were flashes without return strokes that did not trigger the oscilloscopes. Conseque ntly, data and m easurement settings for these flashes are not given. All parameters w e re sampled in segments at 20 MHz with seven LeCroy digital oscilloscopes and continuously sampled at 2 MHz with two Yokogawa digital os cilloscopes (Section 3.6.5 for a description of the recording equipm ent). Table 4-5 lists the date a nd event ID s of all rocket-triggered lightning events that were injected into the test line or into ground in 2003 and re ferences to the m easurement settings for the events. 4.7 2004 Experiment, Vertically-configured Line with Overhead Ground Wire A total of 13 different current paramete rs were measured in the 2004 experiment. The vertically-configured test line tested in 2001 through 2004 was modified to include an overhead ground wire which wa s grounded at every pole (Section 3.3) The lightning current w as direc tly injected from the tower launcher (Section 3.1.1) into the overhead ground wire at midspan, between poles 7 and 8, to investigate direct strike effects. The lightning current m easurem ent stations on and ne ar the tower were identical to the ones in 2003. The number of current measurement sta tions on the line was reduced in 2004. The measurement stations on the line and the lightning current injection point are identified in Figure 4-6. All parameters were sampled in segm ents at 20 MHz with four LeCroy digital oscilloscopes and sampled continuously at 2 MHz with one Yokogawa digital oscilloscopes (Section 3.6.5 for a description of the recording equipm ent).

PAGE 155

125 Figure 4-6: 2004 test di stribution line having a vertical framing configuration and an overhead ground wire with measuremen t points and the lightning strike location identified. Table 4-6 lists the date and event IDs of all rocket-triggered lightning events in 2004 and a reference to the m easurement settings for each ev ent. Table 4-6: 2004 experim ent, references to the measurement settings for all rockettriggered lightning events. The event ID in bold printing denotes a flash with return strokes and the event ID in italic printing denot es a flash without return strokes. Date Event ID Reference to Measurement Settings FPL0402 7/24/2004 FPL0403 Appendix A : Table A-14 4.8 2005 Experiment, Induced Currents During the 2005 lightning campaign currents in a test-runway c ounterpoise (Bejleri et al., 2004) and in a grounded vertical wire of 7 m height induced by rocket-triggered and natural lightning currents were measured to examine the issue of nearby lightni ng inducing high currents on grounde d structures (also Section 1.9 ). Electric field tim e derivatives (dE/dt) during a natural lightning st epped leader and first strok e measured as part of a different experiment (Jerauld, 2003; Jerauld, 2005) are a lso presented in this

PAGE 156

126 dissertation for comparison w ith the induced currents in the vertical wire and counterpoise. The experimental configurations are described in this section. An overview of the rocket-triggered and natural lightning ev ents that induced currents measured on the test structures is given in Section 4.9. The experim ental data is presented in Section 5.4 and analyzed in Section 6.9 Figure 4-7: Satellite image of the Internat ional C e nter for Lightning Research and Testing. The location of the rocket launch facilities, test house, and induced current measurements are indicated. Additionally, the location of the electric field derivative measurements (Stations 1, 4, 8, and 9) are shown including their distances to the vertical wire measurement in meters. Figure 4-7 shows a Google Earth satellite im age of the Intern ational Center for Lightning Research and Testing with the instal lations re levan t to the induced currents

PAGE 157

127 experiment indicated. The location of the two different launch facilities are shown in the figurea mobile launcher described in Section 3.1.2 and a stationary tower launcher described in Section 3.1.1. The lightning currents triggered f rom the m o bile launcher were injected into ground and the lightni ng currents triggered fr om the tower launcher were injected into the te st house (DeCarlo et al., 2006b) loca ted north-west of the tower ( Figure 4-7). The locations of the vertical wire and counterpoise induced current measurements are also indicated in Figure 4-7. The tim e derivatives of the electric fields Figure 4-8 : Satellite image of the ex perim ental site of the induced currents experiments. The objects relevant to this experiment and their locations relative to the north-west corner of the c ounterpoises are indicated.

PAGE 158

128 were measured at Stations 1, 4, 8, and 9 (Jerauld, 2003; Jerauld, 2005). The location of these measurements including their d istan ces to the vertical wire measurement in meters is given in Figure 4-7. Figure 4-8 shows a Google Earth satellite im age of the experim ental site of the induced currents experime nt s. The dim ensions of the runway counterpoise and the locations of the two induced current measurem ents (vertical wire and counterpoise) and the two rocket launcher (mobile and tower launcher) from which triggered lightning was initiated are given relative to the north-west corner of the counterpoise. Figure 4-9 shows a photograph of the 7 m long vertical wire. The wire was grounded with three closely spaced 10-foot ground rods. The grounding resistance of the wire was too high to be measured with th e clamp-on ground m e ter. A current shunt was installed at the bottom to measure the current in the wire. The locations and IDs of peak current sensor cards installed on the vertical wire are also give n in the figure (Schoene et al., 2004a ; Hanley et al. 2006). The current s on the vertical wire and counterpoise were sensed with T&M R-7000-8 1.033 m and T&M R-5600-8 1.27 m shunts, respectively, ( Table 3-19) and transmitted to launch control through Nicolet Isobe 3000 fiber optic links (Section 3.6.3), w here the signals were filtered with 5 MHz custom made low-pass filters (Section 3.6.4 ) and sampled at 50 MHz w ith 8-bit LeCroy Waverunner LT344L oscilloscopes (Section 3.6.5 ). The coun terpo ise and vertical wire curren ts were each measured with two different attenuation settingsa current measurement with high attenuation that can measure currents of up to a few thousand amperes and a current measurement with low attenuation that can measure currents of up to a few hundred amperes.

PAGE 159

129 Figure 4-9: Vertical wire shown with peak current sensor cards. The length of the wire is 7 m. 4.9 Triggering Results All successful triggering events dur ing the 1999 through 2005 experiments are summarized in Table 4-7 and Table 4-8. A successful triggering event is either a classical

PAGE 160

130 trigger (Section 2.2.2), an altitude trigger (bottom of wire not attached to ground), or a wireburn (initial stage only, no return strokes). The experiment (year, test line, nearby or direct strike experiment), the date the li ghtning was triggered, the event ID, the number of return strokes (if any), and the lightning ty p e (classical trigger, altitude trigger, or wireburn; see Section 4.9), are g iven in the tables. Additionally, Table 4-7 contains information about the location of the in jection point of th e lightning current. Table 4-7: Successful triggering events during the 1999 through 2004 experim e nts. Experiment Date Event ID Number of Strokes Lightning Type Injection Point 8/16 UF9904 6 Classical Trigger UF9911 2 Classical Trigger 8/24 UF9912 2 Classical Trigger Phase C, between pole 9 and pole 10 UF9914 2 Classical Trigger 9/6 UF9915 3 Classical Trigger UF9916 7 Classical Trigger 1999, Horizontal Line 9/10 UF9917 4 Classical Trigger Phase B, between pole 9 and pole 10 FPL0011 5 Classical Trigger 7/16 FPL0014 3 Classical Trigger 7/20 FPL0018 6 Classical Trigger 7/28 FPL0025 Wireburn FPL0032 7 Classical Trigger FPL0033 1 Classical Trigger 8/2 FPL0034 5 Classical Trigger FPL0035 Wireburn 8/3 FPL0036 8 Classical Trigger Phase C, between pole 9 and pole 10 2000, Horizontal Line 8/6 FPL0037 2 Classical Trigger Phase C, pole 9 FPL0039 ? Classical Trigger 2000, Vertical Line 8/25 FPL0040 Wireburn Phase A, pole 8 FPL0101 Wireburn 7/26 FPL0102 Wireburn FPL0105 Wireburn FPL0107 2 Classical Trigger 7/27 FPL0108 5 Classical Trigger Phase A, Pole 8 FPL0110 1 Classical Trigger FPL0111 Wireburn 2001, Vertical Line (Direct Strike) 8/18 FPL0112 6 Classical Trigger Phase A, between pole 7 and pole 8 2001, Vertical Line (Nearby Strike) 9/5 FPL0115 Wireburn 20 m from the line north of pole 8

PAGE 161

131 Table 4-7: continued Experiment Date Event ID Number of Strokes Lightning Type Injection Point FPL0205 4 Altitude Trigger Instrument Station 1 FPL0206 Wireburn FPL0208 3 Classical Trigger 7/9 FPL0210 9 Classical Trigger Phase A, between pole 7 and pole 8 7/19 FPL0213 3 Classical Trigger 20 m north of pole 8 FPL0218 1 Classical Trigger FPL0219 3 Classical Trigger FPL0220 7 Classical Trigger 7/20 FPL0221 11 Classical Trigger 7/25 FPL0226 8 Classical Trigger FPL0228 6 Classical Trigger FPL0229 9 Classical Trigger 2002, Vertical Line (Direct Strike) 8/2 FPL0230 Wireburn Phase A, between pole 7 and pole 8 (except FPL0219-1: 20 m north of pole 8) 8/18 FPL0236 1 Classical Trigger 100 m north of pole 7 8/27 FPL0240 Wireburn 8/28 FPL0241 Wireburn FPL0244 Wireburn FPL0245 10 Classical Trigger 2002, Vertical Line (Nearby Strike) 9/13 FPL0246 2 Classical Trigger 30 m north of pole 7 FPL03013 Classical Trigger FPL0302 Wireburn 6/30 FPL0303 Wireburn FPL03054 Classical Trigger 7/6 FPL0306 Wireburn FPL0310 Wireburn 7/11 FPL0311 Wireburn FPL031216 Classical Trigger 7/13 FPL0314 Wireburn FPL03152 Classical Trigger 2003, Vertical Line (Direct Strike) 7/14 FPL03171 Classical Trigger Phase A, between pole 7 and pole 8 7/18 FPL03214 Classical Trigger FPL0326 Wireburn 7/21 FPL0327 Wireburn FPL03294 Classical Trigger 7/22 FPL03312 Classical Trigger 7 m south of pole 4 7/26 FPL0335 Wireburn 8/2 FPL03367 Classical Trigger 15 m south of pole 4 8/7 FPL03411 Classical Trigger FPL0342 Wireburn 8/11 FPL0345 Wireburn FPL03472 Classical Trigger FPL0348 Wireburn 2003, Vertical Line (Nearby Strike) 8/15 FPL03501 Classical Trigger 11 m south-east of pole 15 FPL0402 Wireburn 2004, Vertical Line with OHGW 7/24 FPL0403 2 Classical Trigger OHGW between pole 7 and pole 8

PAGE 162

132 Table 4-8: Successful triggering events during the 2005 experiment. Launcher Date Event ID Number of Strokes Lightning Type 0501 1 Classical Trigger 7/2/05 0503 4 Classical Trigger Mobile 7/3/05 0504 Wireburn 7/15/050508 Wireburn 0510 1 Classical Trigger 7/31/05 0512 2 Classical Trigger 0514 1 Classical Trigger 0517 2 Classical Trigger 8/4/05 0518 Wireburn 0520 1 Classical Trigger Stationary 8/5/05 0521 1 Classical Trigger 8/28/05 MSE0504 2 Natural

PAGE 163

133 CHAPTER 5 CHAPTER 5: DATA PRESENTATI ON The large size of the FPL data set an d the distribution of its presentation and analysis across numerous publica tions make it very difficult to see the big picture and to compare results from the different experiments. The data presented in this chapter are selected to show the essence of the FPL data that the disse rtations author was involved in collecting and to supplement the analysis of the data performed in Chapter 6 The data presented in Section 5.1 and Section 5.2 w ere selected to illust rate the light ning cu rrent path on two different line confi gurations (the vertically-config ured line with and without overhead ground wire tested in 2003 and 2004, respectively) by showing a sketch of the respective distribution line with the meas ured currents at their measurement point locations on the line. The data from the 2003 di rect strike experime nt shown in Section 5.1 are analyzed and m odeled in Section 6.4 and Section 6.5 The com plete data collec ted during the 2003 and 2004 experiments are included in Appendix B These data and all return stroke currents, phase-to-neutral currents, and ground currents measured during the 2000 and 2002 experiments, also included in Appendix B are used for the curren t consistency check and statis tical an alysis in Sec tion 6.1 and in Section 6.2, respectively. The data from the 2003 nearby strike experiment shown in Section 5.3 w ere selected to compare the fraction of the ligh t ning current that traversed the soil and entered the line through one of the groundings with the cu rrent leaving the lin e through the other groundings. Some of these data are analyzed in Section 6.8. Currents m easured on lightning protection systems (a buried counter poise and a grounded ve rtical wire) that

PAGE 164

134 were induced by nearby rocket-triggered and natural lightning are presented in this chapter. The data measured on the buried counterpoise are analyzed in Section 6.9. The large amount of data collected during the 6-years of FPL power line experim ents and the 2005 Lawrence Livermore e xperim e nt has been previously presented in two journal publications and four techni cal reports written or co-written by this dissertations author and in a number of j ournal and conference publ ications, theses, and technical reports wr itten by others (see Chapter 1 for detailed inform ation on the dissertations author origin al contributions). Data fr om the 1999/2000 direct strike experime nt are presented in C.T. Mata et al. ( 1999 2000a 2000b, 2002, 2003), and Mata ( 2000). Data from the 2001 direct strike experi ment and the 2002 direct and nearby strike experim e nts are presented in A.G. Mata et al. ( 2001, 2002) and Mata ( 2003). A review of 1993 data is found in Rakov et al. ( 2003a ). The comp lete data from the 2003 direct and nearby strike experim ents and the 2004 dir ect strike experiment are presented in Schoene et al. ( 2003a ) and Schoene et al. ( 2004b), respectively Schoene et al. ( 2006a 2006b) discuss data collected du ring the 2000, 2002, and 2003 direct strike experiments. The comp lete data from the 2005 Lawrence Livermore experiment are presented in Hanley et al. ( 2006). 5.1 2003 FPL Direct Strike Experiment : Phase-to-neutral and Ground Currents During Stroke FPL0312-5 The current division on the vertically -configured line test ed in 2003 (Section 4.6) is illustrated in Figure 5-1. The figure shows a sketch of the line with the measured currents during stroke FPL0312-5, which are analyzed and modeled in Section 6.4 and Section 6.5, and their m easurement locations. These currents were s e lected as a representative example for the lightning current divi sion between the phase-to-neutral and

PAGE 165

135 Figure 5-1: Measured currents for stroke FPL0312-5, 2 ms and 100 s time scales.

PAGE 166

136 neutral-to-ground connections during the 2003 experiment when no current bypassed the sensors due to line fl ashovers (see Section 6.3 for information on flashovers). The complete data set of the 2003 direct strike experim e nt is presented in Appendix B and in Schoene et al. ( 2003a ). Figure 5-1 shows (1) the lightning in cident curr ent inject ed in to the line (center of figure), (2) all measured phase A-arrester and termination-resistor currents (top), (3) all ground currents (bottom) and (4) the transformer current (right). Each measurement is sh own on two time scales, 100 s and 2 ms. The data shown on the 100 s time scales were sampled at 20 MHz with 8-bit LeCroy digital oscilloscopes and the data shown on the 2 ms time scales were sampled at 2 MHz with 12-bit Yokogawa digital oscilloscopes (Section 3.6.5). The current and tim e scales in Figure 5-1 were chosen for overall consistency with all line cu rrent presenta tions a nd are not optim al to present th e transformer current due to the transformer currents low magnitude and long duration. The transformer current is plotted again in Figure 5-2a on a 100 A current scale and a 5 m s time scale in or der to present it properly. Additiona lly, th e struck-phase arrester current at pole 2 and terminator current at pole 1 are shown in Figure 5-2a to demonstr ate th e current div i sion among the three types of phase-to-neutral conne ctions (arresters, terminators, and one transformer) on a 5 ms time scale. Note that in order to reduce noise all data presented in Figure 5-2 were low-pass filtered with a 5th order, 25 kHz, Butterworth digital filter. Figure 5-2a shows that the arre ster current drops ra pidly to a small positive value (it is norm ally negative ) at 0.4 ms, or so (positive polarity means positive current is flowing from the neutral into the phase A), and from there decays slowly to zero, maintaining the

PAGE 167

137 positive polarity. The terminator current exhibits a fast dip after 0.5 ms and changes to positive polarity after 2.5 ms. Figure 5-2: Stroke FPL0312-5, in jected current and phase A-to -neutral currents. a) Phase A-to-neutral currents at pol es 1 and 2 and b) the injected lightning current, the phase A current measured on the pole 1 side of the pole 6 arrester station (current flowing along the phase A c onductor from pole 6 toward pole 1), and the sum of all phase to neutral currents between the pole 6 ar rester station and pole 1. As a consistency check and to compare th e transformer current and the injected lightning current, the sum of the three currents flowing from phase A to neutral shown in Figure 5-2a along with the inje cted current and the phase A current at pole 6 (m easured after the injected current passes th e phase A arrester at pole 6) are displayed in Figure 52b. The sum of the three phase A currents is equal to the phase A current measured at pole 6 for at least 5 ms (currents larger than 100 A not displayed in Figure 5-2b, m atch

PAGE 168

138 well, too). This result is expected since phase A current measured at pole 6 flows to the neutral at the three phase A-to -neutral connections found at poles 1 and 2. After 2 ms, the sum of the three currents flowing from phase A to neutral, with the transformer current being the largest contributor, are equal to the injected lightning current. Note that after 2 ms the transformer current shown in Figure 5-2a is actually sl ightly larg er than the inject ed cu rrent shown in Figure 5-2b due to the polarity change of the arrester and terminator cu rrents. The peaks of the currents through the polemounted step-down transformer primary for the 26 strokes whose currents were injected into the vertical line in 2003 ranged from some tens of amperes at a time of a millisecond from the start of the return stroke to 200 A at a time of 4 ms. Currents larger than 200 A could not be measured accurately after a few milliseconds due to saturation of the core of the current transf ormer (CT) measuring the transformer current. The transforme r primary initially (first few hundred microseconds) does not pass appreciable cu rrent, but current builds slowly, on a millisecond time scale, in the primary of the transformer due to the lightning current still flowing into the line and the inductiveimpedance of the transformer primary. 5.2 2004 FPL Direct Strike Experiment: Complete Data Set for Stroke FPL0403-2 The current division of the vertically-configured line wi th overhead ground wire (OHGW) tested in 2004 (Section 4.7) is illustrated in Figure 5-3 by showing a sketch of the line with the measured currents du ring stroke FPL0403-1 and their measurem ent location. The currents m easured during stroke FPL0403-1 were selected as a representative example for the lightning cu rrent division on the 2004 test line. The

PAGE 169

139 complete data of the 2004 e xperiment is presented in Appendix B and in Schoene et al. ( 2004b). Figure 5-3 shows a typical example of (1) the lightning incident current in jected into the line (bottom center of figure), (2 ) OHGW currents at poles 6 and 10 (top), (3) phase A currents at pole 6 and 10, (4) phase A and phase B arrester currents at pole 6 (top), and (5) ground currents at poles 6, 7, 8, 9, and 10 (bottom). Each measurement is shown on two time scales, 100 s and 2 ms. The data shown on the 100 s time scales were sampled at 20 MHz with 8-bit LeCroy di gital oscilloscopes and the data shown on the 2 ms time scales were sampled at 20 MHz with 12-bit Yokogawa digital oscilloscopes (Section 3.6.5). The principal difference between the 2004 experim ent and the direct strike experime nts conducted from 1999 thr ough 2003 is that during the 2004 experiment the lightning current injecti on point was on the overhead gr ound wire as opposed to the lightning current injection poi nt being on the phase conduc tors during the 1999 through 2003 experiments where the lightning current fl owed to ground via the phase-to-neutral connections (arresters, transformers, or term inators) or via a flashover from the struckphase conductor to the neutra l conductor. Interestingly, ev en though the lightning current of stroke FPL0403-1 was injected into a grounded conductor the voltage between the phase conductor and the OHGW was large enough to open the arresters. Figure 5-3 shows that a current with a peak value of about 500 A flowed through the phase A and phase B arresters at po le 6. Schoene et al. ( 2004b) estimated, using the measured arres t er current and the manufacturer-provided VI-cha racteristic, a peak of about 25 kV for the voltage across the pole 6 arresters, which is well under the critical fl ashover voltage of a typical distribution line.

PAGE 170

140 Figure 5-3: Measured currents for stroke FPL0403-1, 2 ms and 100 s time scales.

PAGE 171

141 Figure 5-3 show that at the time of the in jected return stroke current peak roughly 50% of the injected currents flow through one of the two closest grounds at pole 8, 25% of the injected currents f low through the other closest ground at pole 7, and the rema ining 25% of the injected currents flow thr ough the remaining grounds. The pole 8 ground initially carries most of th e currents, although the pole 7 ground, which has a smaller low-frequency, low-current ground resistance (Section 3.3.5), is at the sam e distance from the injection point. Apparently, the transient ground resistance at pole 8 is sm aller than the one at pole 7perhaps due to the two ground rods installe d at pole 8 (only one ground rod is installed at pole 7). For later times, the magnitudes of the ground currents are determined by the value of the meas ured low-frequency, low-current ground resistances. 5.3 2003 FPL Nearby Strike Experimen t: Lightning Currents Traversing Soil And Entering Line Grounding In this section data co llected during the 2003 nearby strike experiment are presented with the focus on the lightning curren ts injected into gr ound at a distance of 11 m from pole 15 and the fractions of the lig htning currents entering the line through the pole 15 grounding and leaving the syst em at the other groundings (see Figure 4-5 in Section 4.6 f or a sketch of the test line). Some of the data shown in this section are analyzed and discussed in Section 6.8 In the f ollow i ng figures the fr action of the lightning current that traversed the soil and entered the line through one of the groundings where it was measured is compared with (1) the individual and sum of currents leaving the line through the other groundings and (2) the lightning current inject ed into ground 11 m from pole 15. Figure 5-4 through Figure 5-10 show lightning cu rrents and all ground current s during return strokes

PAGE 172

142 FPL0341-1, FPL0347-1, FPL0347-2, and FPL0350-1. Figure 5-11 through Figure 5-16 show lightning and ground currents during the initial stage of events FPL0347, FPL0348, and FPL0350. Figure 5-4, Figure 5-6, Figure 5-8, Figure 5-10, Figure 5-12, Figure 5-14, and Figure 5-16 show (1) the current injected into the pole 15 gr ounding (the polarity of this curren t was sw itched f or better comparison with the other currents), (2) the individual currents leaving the system at the pole 14, 10, 6, 2, and 1 groundings, and (3) the sum of all currents leaving the system. The currents dur ing the return stroke events are shown on 1 ms and 50 s time scales and the current during th e longer duration initial stage events are shown on 10 ms time scales. Note that th e initial offsets of the currents in these figures were removed. The good match between cu rrent injected into the line and the sum of currents leaving the line shows that the ground curre nts measured on the line are consistent. Interestingly, the risetimes of th e ground currents leaving the line are larger for ground currents that are at a larger distance from the current injection point at the pole 15 grounding. This shows that the high-frequenc y components of the current injected at pole 15 primarily flow through the groundings closer to pole 15. Figure 5-5, Figure 5-7, Figure 5-9, Figure 5-11, Figure 5-13, and Figure 5-15 com pare the norma lized lightning currents with the currents entering the line through the pole 15 grounding for all return stroke and in itial stage events lis ted above except for FPL0341-1 (a high lightning current measurem ent for this stroke was not available and the low lightning current measurem ent was saturated, see Section 4.1). The currents during the return stroke even ts are shown on 1 ms and 50 s time scales and the current during the longer duration initial stage even ts are shown on 10 ms time scales (a 10-

PAGE 173

143 millisecond window of the ICC was selected fo r illustrative purposes, the duration of the total ICC is typically hundreds of millisec onds). Additionally, the current during the initial stage of FPL0350 is show n on a 1 ms time scale to be tter resolve a relatively fast and large pulse in the ICC. The normalization factor for the lightning current given in the figure legend was obtained by dividing the peak value of the numerically integrated pole 15 ground current by the peak value of the numer ically integrated li ghtning current (in other words, the factor normalizes the maxi mum charge transfer of the lightning current with respect to the maximum charge tr ansfer of the pole 15 ground current). The integration time (or equivalently the durati on of the charge transf er) was 1 ms for the return stroke events and 10 ms for the ini tial stage events. The data shown in these figures are analyzed in Section 6.8. Figure 5-4: Individual ground currents and sum of ground currents leaving the system for stroke FPL0341-1. Displaye d on a) 1 ms and b) 50 s time scales.

PAGE 174

144 Figure 5-5: Normalized lightning current injected into ground 11 m from pole 15 (blue) and current injected into the line th rough the pole 15 ground (black) for stroke FPL0347-1. The plateau in the pole 15 ground current may be related to ground arcing at the lightning current in jection point. Displayed on a) 1 ms and b) 50 s time scales. Figure 5-6: Individual ground currents and sum of ground currents leaving the system for stroke FPL0347-1. Displaye d on a) 1 ms and b) 50 s time scales.

PAGE 175

145 Figure 5-7: Normalized lightning current injected into ground 11 m from pole 15 (blue) and current injected into the line th rough the pole 15 ground (black) for stroke FPL0347-2. Displayed on a) 1 ms and b) 50 s time scales. Figure 5-8: Individual ground currents and sum of ground currents leaving the system for stroke FPL0347-2. Displaye d on a) 1 ms and b) 50 s time scales.

PAGE 176

146 Figure 5-9: Normalized lightning current injected into ground 11 m from pole 15 (blue) and current injected into the line th rough the pole 15 ground (black) for stroke FPL0350-1. Displayed on a) 1 ms and b) 50 s time scales. Figure 5-10: Individual ground currents and sum of ground cu rrents leaving the system for stroke FPL0350-1. Displayed on a) 1 ms and b) 50 s time scales.

PAGE 177

147 Figure 5-11: Normalized lightni ng current injected into gr ound 11 m from pole 15 (blue) and current injected into the line through the pole 15 gr ound (black) for the ICC of flash FPL0347. Figure 5-12: Individual ground currents and sum of ground cu rrents leaving the system for the ICC of flash FPL0347.

PAGE 178

148 Figure 5-13: Normalized lightni ng current injected into gr ound 11 m from pole 15 (blue) and current injected into the line through the pole 15 gr ound (black) for the ICC of flash FPL0348. Figure 5-14: Individual ground currents and sum of ground cu rrents leaving the system for the ICC of flash FPL0348.

PAGE 179

149 Figure 5-15: Normalized lightni ng current injected into gr ound 11 m from pole 15 (blue) and current injected into the line through the pole 15 gr ound (black) for the ICC of flash FPL0350. Displaye d on a) 1 ms and b) 50 s time scales. Figure 5-16: Individual ground currents and sum of ground cu rrents leaving the system for the ICC of flash FPL0350.

PAGE 180

150 5.4 2005 Lawrence Livermore Experiment: Induced Currents on a Buried Counterpoise and Vertical Wire Representative data from the 2005 experiment (Section 1.9) conducted to investigate the currents induced on lightning protection systems are prese nted in this section. Induced currents on lightning protec tion systems can be potentially large enough to harm people or cause damage to electri cal devices in contact with the protection system, or to ignite explosives in explos ive storage bunkers with a lightning protection system. Currents induced on (1) a buried c ounterpoise and (2) a vertical wire are presented here. The experimental conf igurations are described in Section 4.8 and an overview of the rocket-triggered and natura l lightning events that induced currents m easured on the test stru ctures is g iven in Section 4.9. The currents measured on the counterpoise are analyzed in Section 6.9. The complete data can be found in Appendix B The experiment is also disc ussed in H anley et al. ( 2006 ). Figure 5-17 and Figure 5-18 show representative examples of currents on the counterpoise a nd vertical wire that were induced by currents of lightning triggered from the mobile launcher (Section 3.1.2 ) an d tow er launc h er (Section 3.1.1). The lightning channel base currents (top), the induced runway counterpoise currents (m iddle), and the induced verti cal wire currents (bottom) measured during the m o bile launcher (FPL0503, stroke 2) and tower launcher (FPL0517, stroke 2) experiments, respectively, are de picted. The induced count erpoise currents are characterized by a few microseconds wide V-shaped initia l pulse followed by a polarity change and a tens of microseconds wide hump of smaller magnitude th an the initial pulse (only the beginning of the hump is shown in Figure 5-17 and Figure 5-18). The induced currents on the vertical wire resem ble the first m i crosecond, or so, of dE/dt waveforms measured at distances rangi ng from 170 m to 340 m from the stationary launcher (the

PAGE 181

151 comparison is not shown in this dissertation) which indicates that the vertical wire functions as a dE/dt antenna. Figure 5-17: Measured cha nnel base current and induced currents on the runway counterpoise (290 m from th e lightning) and the verti cal wire (300 m from the lightning) for stroke 0503-2, ini tiated from the mobile launcher. Figure 5-18: Measured cha nnel base current and induced currents on the runway counterpoise (50 m from th e lightning) and the vertic al wire (100 m from the lightning) for stroke 0517-2, ini tiated from the tower launcher.

PAGE 182

152 Figure 5-19: Counterpoise and vertical wire currents during na tural lightning flash MSE0504 striking ground approximately 300 m from the vertical wire. The spikes in the vertical wire current occu rring before the return stroke initiation at t = 0 are labeled. Figure 5-19 shows the counter poise and vertical wire low currents (Section 4.1) m easured during natural lightning strike MSE050413. The spikes in the vertical wire current labeled 1 through 5 occur before the in itiation of the first re turn stroke. The first return stroke starts at zero with the beginni ng of the V-shaped pulse in the counterpoise current record. Note that spike 4 (the spike with the largest magnitude) is saturated in both the low and high current records (onl y the low record is displayed here). The magnitude of spike 4 in the slightly saturated high current record is 140 A. The 13 The exact strike location of flash MSE0504 is not known but was probably close to MSE Station 1, approximately 300 m west of the vertical wire (see Figure 4-7).

PAGE 183

153 negative polarity of the spikes indicates an upward-directed transfer of positive charges. The pre-return stroke current spikes were presumably currents of unconnected upwarddirected stepped leader emerging from the tip of the vertical wire caused by the electric field of a downward-propagati ng stepped leader. The current spikes are now compared to measured pre-return stroke dE/dt waveforms associated with downward-propagating leader steps in order to test this hypothesis. Figure 5-20 shows each of the 5 pre-return stroke cu rren t spikes ide ntified in Figure 5-19 overlaid with unfiltered and filtered dE/dt waveforms measured at stations 1, 4, 8, and 9 (see Figure 4-7 in Chapter 4 ) on 6 s time scales. Details on the dE/dt measurem ents are found in DeCarlo et al. ( 2006a ). For be tter comp arison the cur rent waveforms and dE/dt waveforms are normalized with respect to the largest magnitude current spike (spike 4) and the largest unfiltered dE/dt spike measured at each dE/dt station, respectively (the normalization factors are displayed in the figure legends). The filtered dE/dt record were obtained by low-pass filtering the unfiltered dE/dt records with a 4th order, 5 MHz Butterworth di gital filter. Note that the dE/dt waveforms and the current measured on the vertical wire ar e not properly aligned in time, since the exact locations of the steps that produced the dE/dt spikes and consequently the traveling times of the electric field waves that are needed for the proper alignment are not known. This problem make s it difficult to positively associate the current spikes of the upward-connected leaders with dE/dt sp ikes of the downward leader steps. However, the spikes measured in th e vertical wire presumably caused by step leader steps correlate reasonably well with signals induced in the counterpoise ( Figure 5-19).

PAGE 184

154 Figure 5-20: Spikes 1 through 5 in the vertical wire current overlaid with unfiltered and filtered dE/dt records measured at Stations 1, 4, 8, and 9. All data are normalized and the normalization factor s are given in the figure legends.

PAGE 185

155 Figure 5-20: continued

PAGE 186

156 Figure 5-20: continued

PAGE 187

157 CHAPTER 6 CHAPTER 6: DATA ANALYSIS, MODELING, AND DISCUSSI ON In this chapter various aspects of the experim ental data collected during the 1999 through 2004 FPL power line experiments and the 2005 Lawrence Livermore induced current experiment are examined. An outline of this chapter is provided in the following paragraphs, including some information on what motivated the investigation of the subjects presented here. A summ ary of the findings presented in this chapter can be found in Chapter 7 D ue to the co m plexity of the FPL powe r line experiments it is important to thoroughly check the consistency of the measur ed currents to ensure the quality of the data and avoid false conclusions based on invalid data. Consequently, this chapter starts with a consistency check of the pr imary data analyzed here (Section 6.1). The rela tively large num ber of rocket-tri ggered lightning return stroke currents, w hich are similar to currents from subsequent return strokes in natural lightning, recorded during the 6-years of distri bution line experiments allows a highly representative statistical characterization of rocket-triggered lightning pa rameters, which is given in Section 6.2. The large scale of the da ta set (200 return stroke currents were recorded) and the opportun ity to ensure the accuracy of mo st of the data by checking th eir consistency with currents measured on the line (Section 6.1) m akes the statistical analysis presented in this dissertation one of the most signi ficant con t ributions towards characterizing rocket-triggered lightning return stroke currents available.

PAGE 188

158 A primary objective of the FPL experiment was to evaluate the performance of the different distribution line conf igurations regarding arrester failures and line flashovers. This is important for the design of lightning protection for distribution lines, where the initial costs of arresters a nd the replacement costs for damaged arresters are weighed against the benefits of pr otection for equipment and an expected reduction in line flashovers and outage rates. The disconnector operations and flashovers observed on the test power distribution lines are ev aluated and discussed in Section 6.3. For design of lightning protection for di str ibution lines it is imperative to understand the response of the distribution lines to direct lightning strikes. The next tw o sections are concerned w ith (1) analyzi ng the experimentally-determined lightning current division between phase-t o-neutral current paths, su ch as through arresters and transformers, and between line groundings on the test distribution lines (Section 6.4), and (2) com paring the experime ntal data w ith model-predicted results (Section 6.5). Model representations of the various components of distribution lines su ch as the phase and neutral conductors, groundings, a nd arresters are readily availabl e in the literature. It is important to test the validity of these models under conditions that are as close to real world conditions as possible. The experimental data obtaine d from measurements on test distribution lines that resemble real worl d distribution lines subjected to rockettriggered lightning return stroke currents that are identical to subs equent return stroke currents in natural lightning are highly suited for this task. Additionally, the model is employed in Section 6.5 to tes t speculation raised in Section 6.4 regarding the different current division on the horizontallyand vertically-configured lines.

PAGE 189

159 Further, understanding the response of dist ribution lines to n earby lightning strikes is important as the power quality is reduced by line flashovers caused by lightninginduced overvoltages. The LIOV-EMTP96 code, one of the most advanced electromagnetic coupling/circuit-theory tools to model cu rrents and voltages induced by nearby lightning, is tested with selected data obtained during the 2003 nearby lightning strike experiment (Section 6.7). The distinction between (1) direct lightning strike inte raction with power lines, where ligh tni ng current is direc tly injected into th e line a nd the lightning can be modeled as a current source connected to the injection point (Section 2.5), and (2) nearby lightning strike interaction with power lines, where th e response of distribu tion lines to nearby lightning strikes is m odeled by calculating the currents and voltages induced by the lightnings electromagne tic f i elds (Section 2.6), is not always so clear-cut as the treatment of these tw o effects in two separa te sections of this dissertation (Section 6.5 and Section 6.7) m akes it appear to be. Specifically, if lightning strikes the ground close to a line grounding, both near by and direct strike effects contribute significantly to the currents and voltages on the line-t he coupling of the lightnings electromagnetic fields to the line induces currents and voltages on th e line (nearby strik e effects) and lightning current injected into ground at the strike point traverses the soil and enters the line grounding (direct strike effects) The latter effect, although si gnificant if lightning strikes within a few meters or possibly tens of meters of a line grounding, is typically neglected in models that calculate lightning-induced overvoltages on power lines such as the LIOVEMTP96 code tested in Section 6.7. The subject of lightning curren ts traversing soil and entering line groundings is investigated in Section 6.8 This issue has also im plications in

PAGE 190

160 many areas other than the dist ribution line design, such as lightning currents traversing soil and causing coal mine e xplosions (Novak and Fisher, 2001) or damaging electrical systems. Buried counterpoises are em ployed for th e grounding of power lines, residential houses, explosive storage bunkers, etc.. Th e interaction of a counterpoise grounding system with nearby lightning stri kes, investigated in Section 6.9, is an im portant issue as the induced current in a counter poise can potentially be harmful to people in contact with objects connected to the counter poise, dam a ge electrical devices in contact with the counterpoise, or produce sparks that can ignite explosives in explosive storage bunkers. 6.1 Current Consistency Check In this section the cons istency of the data from the 2000, 2002, 2003, and 2004 experiments14 is discussed in order to ensure the validity of the data analyzed and modeled in the following sections of this chapter. The injected high lightning current/charge (Section 4.1) is com pared with (1) the cu rrent/charg e leav ing the system (Section 6.1.1), (2) the current/charg e transfe r be tween the struck-phase and neutral conductors (Section 6.1.2), and (3) the low li ghtning current (Section 6.1.3). The charge was obtained by numerical ly integrating the m easured curren ts. Note that based on the principle of charge conservation a ny missing charge (injected charge exceeding charge leaving the system) or excessive charge (charge leaving the system exceeding injected charge) can only be attributed (a) to a current path that bypasses the current sensors or (b) to a measurement error. 14 A complete ground current or struck-phasetoneutral current data set was not obtained during the 1999 and 2001 experiments and consequently the consistency of these data is not tested.

PAGE 191

161 6.1.1 Consistency of Injected Cu rrent and Total Ground Current Generally, the measured injected curren ts and charges and the total measured currents and charges flowing to ground are in very good agreement for the 200015, 2002, and 2003 experiments16. Representative examples that show the good agreement are given in Figure 6-1a for currents on a 100 s time scale, in Figure 6-1b for charges on a 100 s time scale, and in Figure 6-2 for currents on a 4.5 ms time scale. A complete comparison of the currents on a 100 s time scale for the 2000, 2002, and 2003 experiments can be found in Appendix B Note that even though the injected charge and the total charge transfer to ground are very sim ilar, the peak value of the total ground current is typically muc h larger. The overs hoot of the sum of gr ound currents appears to be larger for injected currents with smalle r risetimes. For instance, for stroke FPL0229-7 the 10-90% risetime of the injected current is 1.5 s and the peak value of the sum of ground currents is approximately twice the pe ak value of the inje cted current (see Figure B-68 in Appendix B ), while for stroke FPL0229-6 the 10-90% risetim e of the injected current is m uch larger (2.7 s) and the peak value of the sum of ground currents is very similar to the peak value of the injected current (see Figure B-67 in Appendix B ). The overshoot of the sum of ground currents is pr obably due to the norma l ringing response of the syste m to current injection (reflections from impedance discontinuities at arresters, terminators, and line groundings) but also might be due to induced currents on the line from the electric and magnetic fields of the lightning return stroke. 15 An adjustment factor of 0.75 applied to all return stroke current data from the 2000 experiment was necessary to achieve a good agreement for this data set as discussed in Section 6.1.4. 16 Any discrepancies were usually due to a hump in the injected current that was the result of a poorly understood, non-linear measurement error as discussed in Section 6.1.3.

PAGE 192

162 Figure 6-1: Stroke FP L0315-1, ground current consistenc y check. a) Lightning current injected into phase A and the sum of ground currents and b) charge injected and the sum of charges tr ansferred to ground on a 100 s time scale. The charge displayed in b) was obtained by numerically integrating the current waveforms shown in a). Figure 6-2: Stroke FP L0315-1, ground current consistenc y check. a) Lightning current injected into phase A and the sum of ground currents and b) charge injected and the sum of charges transferre d to ground on a 4.5 ms time scale.

PAGE 193

163 6.1.2 Consistency of Injected Current and Total Struck-phasetoneutral Current The measured injected currents and charges and the total m easured currents and charges flowing from the struck-phase conduc tor to the neutral conductor are compared for the 200017, 2002, and 2003 experiments. Generally, in the absence of flashover the injected currents and the tota l struck-phasetoneutral curren t agree well for at least 2 ms. An example for the good agreement is given in Figure 6-3a for currents and in Figure 6-3b for charges, both displayed on a 100 s time scale. All currents measured during the 2000, 2002, and 2003 experiments are compared on a 100 s time scale in Appendix B The injected current often exceeded the curr ent flowing from the struck-p hase conductor to the neutral conductor for one or more of the following reasons: A measurement error caused by saturation of the magnetic core of the current transformers used to measure the phase-to-neutral currents (Section 3.6.1). An indication of a saturated core is a sudden dip in the current record. An example for this e rror is show n in Figure 6-4 which compares the injected current and the total current flowing from the struck-phase conductor to the neut ral conductor on a 4.5 ms ti m e scale. The dips at about 3 ms a nd 4 ms indicated in the figure are due to core saturations in the termination current measurement and the transformer current measurement, respectively. A result of th is measurement limitation is that we can only believe the total phase-to-neutral currents for the first 2 ms, or so. A flashover that causes current to bypa ss the current sensors installed between struck phase and neutral (e.g., Figure B-66). Flashovers ar e discussed in Section 6.3. Some struck-phasetoneutral currents were not monitored or their measurement instrumentation failed. For these current s the corresponding current at the opposite pole were used to calculate the total struck-phase toneutral current assuming symmetry (e.g., the termination resistor current at pole 15 on the verticallyconfigured line was not measured and the termination resistor current at pole 1 was used instead). A significant error is introdu ced in the calculation of the total struckphasetoneutral current if the symmetry as sumption is not accurate. For instance, the phase A arrester currents at poles 6 and 10 which are very similar most of the time are quite different for stroke FPL0220-5 as shown in Figure B-44. The 17 An adjustment factor of 0.75 applied to all return stroke current data from the 2000 experiment was necessary to achieve a good agreement for this data set as discussed in Section 6.1.4.

PAGE 194

164 calculated total struck-phasetoneutral current for this stro ke if either the pole 6 or the pole 10 arrester current would not have been measured would have resulted in a large error. Figure 6-3: Stroke FP L0315-1, phase A-to-neutral current consistency check. a) Lightning current injected into phase A and the sum of phase A-to-neutral currents and b) charge injected and the sum of charges transferred from phase A to neutral on a 100 s time scale. The charge displayed in b) was obtained by numerically integrating the cu rrent waveforms shown in a).

PAGE 195

165 Figure 6-4: Stroke FP L0315-1, phase A-to-neutral current consistency check. a) Lightning current injected into phase A and the sum of phase A-to-neutral currents and b) charge injected and the sum of charges transferred from phase A to neutral on a 4.5 ms time scale. 6.1.3 Consistency of High and Low Lightning Currents In this section the high and low lightning curren t records (Section 4.1) are compared. Both records measure the output fr om the same current sensor, therefore any discrepancies in the non-saturated records are due to errors during th e transmission of the sensor output signal to th e digital oscilloscope. These errors can be attributed to (1) errors introduced by the fiber optic links (two different receiver/transmitter pairs were used for each measurement, see Section 3.6.3), (2) coupling effects of the lightning electrom agnetic field on th e coaxial cables (Section 3.6), or (3) fai l ure or incorrect connection of the resistors conne cted to the current sensor. All high and low light ning currents measured in the 2000 experiment do not match. The reason for this inconsistency is a pparently a calibration er ror present in both current measurements, as discussed in Section 6.1.4. An adjustm ent factor of 0.75 is applied to all high lightni ng currents from 2000 analyzed in this dissertation.

PAGE 196

166 Some of the high and low lightning currents measured in the 2002, 2003, and 2004 experiments do not match. The mismatch is due to a non-linear error, that is, an error that cannot be corrected with a constant multiplicative factor (a s opposed to a linear error in the 2000 lightning currents, which could be corrected w ith an adjustment factor). The error was present in either the high current record, low current record, or both records. Figure 6-5a and Figure 6-6a compare the Yokogawa records of the high and low lightning currents and total ground current for strokes F PL0305-2 and FPL0312-8, respective l y, both triggered from the tower launcher and injected into phase A of the line. Figure 6-7a compares the high and low cu rrents for stroke FPL0336-8 triggered f rom the m obile launcher (Section 3.1 ). The high and low currents for all three strokes are different for parts of the record. The to tal ground currents shown for strokes FPL0305-2 and FPL0312-8 match the unsaturated part of the low lightning current, which gives strong evidence that the m ismatch of the hig h and low current for these two strokes is due to an error in the high current m easurement. Note that the high current of stroke FPL0312-8 exhibits an er roneous shoulder. A shoulder is also present in the high current of stroke FPL0336-7, which suggests th at the mismatch of the high and low currents for this stroke is also due to an error in the high current measurement. Subtraction of the high and low current waveforms of stroke FPL0305-2, the result being shown in Figure 6-5b, reveals a presumably erroneous positive hump in the high current record. Subtracting the high a nd low currents for strokes FPL 0312-8 and FPL0336-8 reveals a presumably erroneous ne gative hump in the high current records ( Figure 6-6b and Figure 6-7b).

PAGE 197

167 Figure 6-5: Stroke FPL03052, high/low lightning current comparison. a) High/low lightning currents injected into the line and total ground current and, b) difference of high and low lightning currents. Figure 6-6: Stroke FPL03128, high/low lightning current comparison. a) High/low lightning currents injected into the line and total ground current, and b) difference of high and low lightning currents.

PAGE 198

168 Figure 6-7: Stroke FP L0336-7, high/low lightning current comparison. a) High and low lightning currents inje cted into ground, and b) difference of high and low lightning currents. All humps present in the 2002, 2003, and 2004 lightning currents are identified in Appendix C W hat follows is a summary of so me observed h u mp characteristics: The humps have peak values ranging from a few hundreds of amperes to a few kiloamperes, risetimes ranging from a few microseconds to a few tens of microseconds, and decay to zero within a few hundreds of microseconds. The humps have negative or positive polarity. The humps are sometimes present in the low current records and sometimes present in the high current records. The humps are sometimes present in the li ghtning currents measured at the tower launcher, lightning intercepting struct ure, and mobile launcher (Section 3.1). The presence and magnitude of the hump s appear to be independent of the magnitude and the risetime of the lightning current. Different types of humps (positive and ne gative polarity) appear within different strokes of the same flash.

PAGE 199

169 6.1.4 Consistency of Lightning Currents from the 2000 Experiment Almost all total ground currents, high lightning currents, and low lightning currents measured in the 2000 experiment ar e inconsistent with each otherall total ground currents are about 25% smaller than the high lightning curren t injected into the phase conductor and most total ground curre nts are about 50% sma ller than the low lightning current. The only exceptions are for the low lightning current measurements on 8/2/00 (FPL0032, FPL0033, FPL0034) and 8/3/00 (FPL0036) for which the low current was not measured, and for the current measurement on 8/6/00 (FPL0037, the last triggered flash of the season) for which th e low lightning current matches the total ground current (but not the high lightni ng current). The errors that cause the discrepancies between the tota l ground, high, and low cu rrents are apparently linear in nature since all three curre nts have the same waveshape. In principal, the mismatch of the tota l ground current and the high and low lightning currents can be attri buted to either (1) an alte rnative path of the lightning current other than through the monitored gr ound and phase-to-neutral connections or (2) a measurement error. The mismatch of the high and low lightning currents can only be attributed to a measurement error in eith er or both current meas urements. This error must have occurred during the transmission of the current sensor output signal to the oscilloscope since the same current sensor measures both the high and the low lightning currents (Section 4.1 and Section 6.1.3). An alternative current path that woul d account for the m ism atch of total ground current s and lightning currents had to be either from the wire that connects the tower launcher with the phase conductor or from the struck-phase (phase C) conductor to ground to bypass both, the ground current m easurements and the phase-to-neutral

PAGE 200

170 measurements on the line. An a lternative current path can be established by either (1) a physical connection or (2) a flashover. The existence of a physical connection was not reported in Mata ( 2000). Mata et al. ( 2003) were not able to identify any flashovers in their video records, a lthough they did not rule out the existence of such flashovers. It appears that the presence of flashovers is not likely since the establishm ent of the flashover path would probably result in discontinu ities in the currents measured on the line, which were not observed. Furthermore, the 25% discrepancy is also present in currents with sm all m agnitude, which are unlikely to cause flashovers. Lastly, an alternative current path woul d not explain the differences between the high and low lightning current records. The total ground current appears to have been accurately measured during the 2000 experiment since it is mostly consistent w ith the total struckphasetoneutral current ( Appendix B ). Any discrepancies during the first 100 s can be attributed to flashovers on the line that reduce the total measured stru ck-phasetoneutral curr ent as discussed in Section 6.1.2. Also, the total ground currents and low lightning currents do m atch reasonably well for the two strokes in flas h FPL0037. Consequently, a measurem ent error that would explain the discrepancy between th e total injec ted current and the lightning current (high and low) would likely be an overestimation of the lightning current. Based on the above discussion, the 25 % difference between the total ground currents and high lightning currents is likely attributable to the overestimation of the high lightning current due to an undetected measurement error, which is also present in the measurement of the low current (exc ept for FPL0037). An adjustment factor of 0.75 was applied to the high lightning curr ent data measured in 2000 to correct the

PAGE 201

171 suspected measurement error. Additional s upport for the assumption that the lightning current was overestimated is found in the statistics of the return stroke current peaks (Section 6.2.1). The arithm etic mean of the 33 p eak values m easured during the 2000 experim ent is 13.8 kA with the adjustment factor (0.75) applied and 18.4 kA for the uncorrected peaks. The former value is ve ry close to the arithmetic mean of the remaining 132 peaks, 13.9 kA (not including the 2000 data), while the latter is much larger, which suggests that the arithmetic mean of the uncorr ected peak is overestimated. The alleged measurement error is possib ly due to some kind of mix-up in the connection of the 50 resistors to the lightning current shunt. Note that, according to a personal conversation with Dr Carlos Mata, a single 50 resistor was erroneously used at the output of the current shunt instead of two (one for the high current measurement and one for the low current measurement). This oversight was neither mentioned in Mata ( 2000) nor in Mata et al. ( 2003). However, using a single 50 resistor results in a 30% underestimation of the lightning current18 and not in a 25%/50% overestimation of the high/low current. Note that according to Dr. Carlos Mata a 50 resistor was added to the lightning current measurement sometimes during the 2000 experiment (the exact date is not known) to correct the error. However, correcting the error di d not seem to have affected the 25% underestimation of th e high lightning current since the underestimation is present in all the data from the 2000 experiment although it might have corrected the 50% underestimation of the low lightning current since for the last flash of the season (FPL0037) the low lightning current and total ground current do 18 The author of this dissertation confirmed this experi mentally by using a measur ement setup similar to the erroneous setup used in 2000. The current shunt was replaced by a lowlevel test voltage generator.

PAGE 202

172 match, as noted above. The discrepancies of the high and low lightning currents suggest that the division of th e current sensor output volta ge was different for the two current measurements, which could have been caused by asymmetrically connected 50 resistors, although such an error was not repor ted. Note that the 0.75 adjustment factor applied to currents from the 2000 experiment analyzed in this dissertation was not applied to lightning currents from the same experiment analyzed previously in Mata ( 2000), A.G. Mata et al. ( 2002), and Mata et al. ( 2003) and that consequently the 2000 currents presented in these docum ents were likely overestim ated by 25%. Note also th at A.G. Mata et al. ( 2002) m ultiplied the lightning current by a so-called k factor that was calculated for each value of the channel characteris tic im pedance to minimize the difference between the measured charge injected into the distributi on line and the modelpredicted charge transferred to ground (p. 42). This k-factor ranged from 0.69 to 0.82, the smaller values being associated with la rger lightning channel im pedances, effectively reducing the measured current by the same percentage we suggest above. 6.1.5 Consistency of Current from the 2004 Experiment Figure 6-8 shows data from the 2004 expe rimentthe high and low channel base currents, the current obtained by summing measured ground currents and overhead ground wire currents19, and the pole 8 ground current. The figure shows that the channel base currents and the sum of ground and overhead ground wire currents match well except for a spike in the rising edge of the latter current. Th e spike in the sum of currents is due to a spike in the pole 8 ground current as shown in Figure 6-8. 19 This current is the current leaving the system since it comprises all currents flowing to ground at poles 6 through 10 (the measured ground currents) and the currents flowing to the other grounds via the overhead ground wire (the overhead ground wire currents measured at poles 6 and 10) and is expected to match the channel base current (the current injected into the system).

PAGE 203

173 Figure 6-8: Stroke FPL0403-2, (1 ) high and (2) low channe l base currents, (3) sum of poles 6 through 10 ground currents and pol es 6 and 10 overhead ground wire currents, and (4) pole 8 ground current. Figure 6-9 shows the spike present in the pole 8 ground current overlaid with 2 spikes present in current m easuremen ts from a different experim e nt also conducted in 2004, the so-called test-house experiment20. Note that the pole 8 ground current spike and point A current spike were isolated from the rest of the data by low-pass filtering the currents and subtracting the filt ered currents from the unfiltered ones. Note also that the polarity of the pole 8 ground current spike was flipped for better comparison with the other spikes. All spikes shown in Figure 6-9 have similar magnitudes (between 6 kA and 8 kA) and widths (about 0.5 s), which indicates that the cause for the spikes in the different current measurements is the same. None of the spikes in Figure 6-9 appear in the injected lightning current s. A consistency check for the test house experim ent 20 Rocket-triggered lightning currents we re injected into the protective syst em of a test house and measured at various points in the syst em (DeCarlo et al., 2006a).

PAGE 204

174 Figure 6-9: Spikes during the FPL and test-house experiments. The spikes in the pole 8 ground current and the point A current were isolated by subtracting the lowfrequency components from the total data. currents shows that the sum of currents measur ed in the test house system considerably exceeds the injected lightning current at th e time the spikes occur (DeCarlo et al., 2006a ). A lso, the polar ity of the pole 8 ground current spike indicates that the spike is injected into the line through th e pole 8 grounding but the spike doe s not appear in other current measurements on the line. Based on these obser vations it is believed that the spikes are not due to actual currents flowing through th e current sensor but to measurement errors possibly related to electromagne tic coupling to the instrument ation in the Hoffman boxes. Note that some of the Hoffman boxes were found to be only partially latched or not latched at all during the FPL and test-house experiments, which made the instrumentation in the boxes prone to electromagnetic coupli ng. It was not reported which measurements were potentially affected by the partially-latched or non-latched boxes.

PAGE 205

175 6.2 Characterization of the Ligh tning Return Stroke Current An accurate characterization of rocket-tri ggered lightning return stroke currents which have similar, if not id entical, properties to subsequent stroke currents in natural lightning (Section 2.2) is required for the lightning prot ection design. In this se ction statistical param eters are extracted from th e lightning return stroke currents measured during the 1999 through 2004 FPL power lin e experiments. The statistical characterization of the return stroke current is highly representative due to the large sample size of the analyzed data (200 return stroke currents were recorded). The quality of the direct strike data was ensured by a thorough consistency check in Section 6.1 of the re turn stroke currents w ith currents m easured on the phase-to-neutral connections and groundings of the distribution line. Informa tion on the number of return strokes and statistical data on return stroke current pa rameters (peak value, 10-90% risetime, halfpeak width, and charge transf er within 1 ms, see Section 2.2.3 for a definition of these param eters) during the 1999 through 2004 e xperime nts is presented in Section 6.2.1 for each year, for all years combined, for the di re ct strik e experiments, and for the nearby strike experiments. Histograms that show the statistical distributions of the return stroke current parameters are presented in Section 6.2.2. The statis tical inf o rmation given in Section 6.2.1 and Section 6.2.2 is discussed in Section 6.2.3. Note that the statistical analysis of the return stroke currents measured d u ring th e 2000 experiment previously published in Mata ( 2000) has been revised by (1) applying an adjustm ent factor of 0.75 to the overestimated return stroke cu rren t peaks and charge transfers (Section 6.1.4) and (2) the risetimes determ ined f rom currents that exhibited ringing during the rising edge have been excluded from the statistics.

PAGE 206

176 6.2.1 Statistical Data of Lightning Return Stroke Current Parameters Table 6-1 shows information on the number of flashes and return strokes triggered during the 1999 through 2004 experiments, for a ll years combined, for the direct strike experiments, and for the nearby strike experiments. Flashes are dis tinguished as either containing return strokes or being composed of the initial stage (the initial continuous current) only (Section 2.2). A dditionally, Table 6-1 provides statistical inform ation extracted from the lightning return stroke curren t s including the sa mple sizes, minimum/maximum values, arithmetic/geometric means, standard de viations, and the logarithmic standard deviations21 of the return stroke current peak values, current 10-90% risetimes, current half-peak widths, and charge transfer within 1 ms22. For an illustration of the analyzed parameters refer to Figure 2-11 in Section 2.2.3. Note that for some return strokes during the 2002 and 2003 experim ents the return s troke current sometimes split between the intercepting structure (this current was injected into the test power line) and the grounded tower launcher (Section 3.1.1). The statistics in Table 6-1 and in the histogr am s presented in Section 6.2.2 were calcu lated f rom the sum of the two individual cu rrents (and not from the currents injected into the line) since the main intent here is to characterize the li ghtning return stroke current. The return stroke current parameters for each year that were used to calculate the statistics in Table 6-1 can be found in Appendix C 21 The logarithmic standard deviation of a parameter is calculated by calculating the standard deviation of the base-10 logarithm of this parameter. 22 The charge is obtained by integrating the lightning return stroke current over a 1 ms time interval.

PAGE 207

177 Vertical Line with Overhead Ground Wire 1999, direct2000, direct2001, direct2002, direct2002, nearby2003, direct2003, nearby2004, direct Total number of flashes with return strokes4635117841 04571 Total number of flashes without return strokes2415902423661 Total number of recorded return strokes2061693726 37 14 64 16 26 21 2 Sample Size1651442126 33 13 48 3 22 18 2 Minimum [kA]2.82.84.32.8 5.3 6.0 5.6 4.3 5.8 4.7 5.9 Maximum [kA]42.342.328.723.142.328.233.7 8.9 27.928.7 15.9 Arithmetic Mean [kA]13.914.013.010.113.816.716.3 5.8 12.514.2 10.9 Standard Deviation [kA]6.96.97.04.7 8.4 6.2 7.1 2.7 4.3 6.8 7.0 Geometric Mean [kA]12.212.411.19.011.915.514.7 5.5 12.012.5 9.7 Standard Deviation, log [kA]0.220.220.260.230.230.180.210.180.140.24 0.30 Sample Size8163189 13 36 1 3 17 2 Minimum [ s]0.20.40.20.8 0.9 0.4 0.4 0.5 0.2 0.8 Maximum [ s]5.75.71.62.0 1.9 5.7 0.4 1.6 1.6 0.9 Arithmetic Mean [ s]1.21.40.51.3 1.4 1.4 0.4 1.1 0.5 0.8 Standard Deviation [ s]0.80.80.40.5 0.3 0.9 0.6 0.5 0.1 Geometric Mean [ s]0.91.20.41.2 1.4 1.2 0.4 1.0 0.4 0.8 Standard Deviation, log [ s]0.320.220.310.15 0.090.27 0.270.32 0.06 Sample Size1421222093 31 34 831 71 72 Minimum [ s]4471 571 661 9471 7 Maximum [ s]93936538 90 58 93 29 20 65 42 Arithmetic Mean [ s]23232428 24 29 25 23 10 25 30 Standard Deviation [ s]17181692 01 11 9551 71 8 Geometric Mean [ s]19182026 18 28 20 23 9 20 27 Standard Deviation, log [ s]0.300.300.260.160.270.160.300.100.220.29 0.28 Sample Size1511222993 31 34 81 11 71 82 Minimum [C]0.30.30.30.4 0.3 0.5 0.3 0.3 0.3 0.4 0.5 Maximum [C]8.38.32.72.5 8.3 7.3 4.1 1.1 6.0 2.7 2.1 Arithmetic Mean [C]1.41.51.11.3 1.6 2.3 1.5 0.5 1.0 1.5 1.3 Standard Deviation [C]1.41.40.80.8 1.9 1.8 1.0 0.2 1.3 0.8 1.1 Geometric Mean [C]1.01.10.81.1 1.0 1.8 1.2 0.5 0.7 1.2 1.0 Standard Deviation, log [C]0.350.350.330.320.410.300.320.180.290.30 0.44 Return Stroke Current Half-Peak Width Return Stroke Charge Transfer within 1 ms Horizontal Line Vertical Line Return Stroke Current Peak Return Stroke Current 10-90% RisetimeTotal Total, direct Total, nearby Table 6-1: Return stroke current statistics for the 1999 through 2004 experiments.

PAGE 208

178 6.2.2 Statistical Distributions of Lightnin g Return Stroke Current Parameters Histograms of the return stroke current peak values, current half-peak widths, charge transfer within 1 ms, and 1090% risetimes from the 1999 through 2004 experiments are shown in Figure 6-10, Figure 6-11, Figure 6-12, and Figure 6-13, respectively. The arithmetic mean m standard deviation s, and sample size n for the data from each experiment and for all data combined are also given in the figures. Note that an adjustment factor of 0.75 has been applied to the current peak values and charge transfers from the 2000 experiment to account for the ap parent overestimation of the return stroke currents measured during that experiment (Section 6.1.4). A ll data in the histograms and additional statis tic al information are listed in Table 6-1 in the prev ious section and in Appendix C A discussion of the distri butions is f ound in Section 6.2.3 Figure 6-10: Histogram of return stroke curren t peaks. An adjustm e nt factor of 0.75 has been applied to the current peaks from the 2000 experime nt.

PAGE 209

179 Figure 6-11: Histograms of return stroke cu rrent 10-90% risetimes. a) Direct and nearby strokes, b) only direct strokes, and c) only nearby strokes. The horizontal scale in a) and b) is interrupted between 2.8 and 5.6 s. The vertical and horizontal scales in c) are different fr om the scales in a) and b).

PAGE 210

180 Figure 6-12: Histogram of return stroke current half-peak widths. Figure 6-13: Histogram of return stroke char ge transfers within 1 ms. An adjustment factor of 0.75 has been applied to the charge from the 2000 experiment.

PAGE 211

181 6.2.3 Discussion of Lightning Return Stroke Current Parameters There have been a number of previous analyses of triggered li ghtning return stroke current characteristics with which our analysis can be compared. Table 6-2 compares the statistical info rm ation on return stroke curren t peaks, 10-90% risetimes, half-peak widths, and charge transfers from the FPL experiments conducted from 1999 through 2004 presented in the previous sections with stat istics from experiments discussed in Depasse ( 1994), Fish er et al. ( 1993), Crawford ( 1998 ), Rakov et al. ( 1998 ), Uman et al. ( 2000 ), and Schoene et al. ( 2003a ). Note that in Table 6-2 the FPL dir ect and nearby strike experiments are considered separately to reve al any potential influe nce of the different strike objects for these two experiments on the return s t roke current characteristics. Information on the strike object and its measured low-frequency, low-current grounding resistance is also provided in the table when available. 6.2.3.1 Return stroke current peaks Table 6-2 and Figure 6-10 show that the arithmeti c mean and standard deviation of the 144 return stroke current peaks measured during the direct strike experiment ( mean: 14 kA, standard deviation: 6.9 kA) are very similar to the ones of the 21 current peaks from the nearby strike experiments (mean: 13 kA standard deviation: 7.0 kA) and appear to follow the same log-normal distribution. Thes e results are consistent with current peak statistics from the 1985 Kennedy Space Center experiment (Depasse, 1994), the 1993 Camp Blanding experi m e nt (Rakov et al., 1998), and the 1998 Camp Blanding experiment as shown in Table 6-2. The arithm etic m eans of the direct and nearby strike experiments are larger than the arithm e tic means found in the 1986, 1990 and 1991 SaintPrivat dAllier experiment, 11 kA for 54 peaks, (Depasse, 1994) and for the 1997 Camp Blanding experim ent, 12.8 kA for 11 peaks, (Crawford, 1998 ). The latter difference

PAGE 212

182 1999-2004 Direct Strike Experiment 2002 and 2003 Nearby Strike Experiment 1985-1991 Kennedy Space Center (KSC), Florida1 1986, 1990 and 1991 Saint-Privat d'Allier, France1 1990 Kennedy Space Center (KSC), Florida 1991 Fort McClellan, Alabama2 1993 Camp Blanding, Florida3 1997 Camp Blanding, Florida4 1998 Camp Blanding, Florida5 1999 and 2000 Camp Blanding, Florida6 Strikes to a distribution line phase conductor or overhead ground wire. Strikes to an 8 m high grounded wire. R = 0.1-0.15 R = 9 KSC: R = 0.12 3 R initially tens of k R = 220 R = 58 or 350 Strikes to a 70 m x 70 m buried metal grid (R = 6 ). Return Stroke Current Peak Sample Size 144 21 305 54 45 37 11 25 64 Minimum [kA] 2.8 4.3 2.5 4.5 5.3 5.3 5.9 5.0 Maximum [kA] 42.3 28.7 60.0 49.9 44.4 22.6 33.2 36.8 Arithmetic Mean [kA] 14.0 13.0 14.3 11.0 15.1 12.8 14.8 16.2 Standard Deviation [kA] 6.9 7.0 9.0 5.6 5.6 7.0 7.6 Geometric Mean [kA] 12.4 11.1 12.0 13.3 11.7 13.5 14.5 Standard Deviation, log 0.22 0.26 0.28 0.23 0.20 0.19 0.21 Return Stroke Current 10-90% Risetime Sample Size 63 18 37 43 11 63 Minimum [ s] 0.4 0.2 0.3 0.3 2.4 Maximum [ s] 5.7 1.6 4.9 4.0 0.1 Arithmetic Mean [ s] 1.4 0.5 1.1 0.9 0.3 Standard Deviation [ s] 0.8 0.4 1.1 1.2 0.4 Geometric Mean [ s] 1.2 0.4 0.4 0.6 0.2 Standard Deviation, log 0.22 0.31 0.29 0.39 0.44 Return Stroke Current HalfPeak Width Sample Size 122 20 24 41 11 64 Minimum [ s] 4 7 15 7 2 Maximum [ s] 93 65 103 100 37 Arithmetic Mean [ s] 23 24 50 36 13 Standard Deviation [ s] 18 16 22 25 9 Geometric Mean [ s] 18 20 18 29 11 Standard Deviation, log 0.30 0.26 0.30 0.29 0.32 Return Stroke Charge Transfer within 1 ms Sample Size 122 29 Minimum [C] 0.3 0.3 Maximum [C] 8.3 2.7 Arithmetic Mean [C] 1.5 1.1 Standard Deviation [C] 1.4 0.8 Geometric Mean [C] 1.1 0.8 Standard Deviation, log 0.35 0.33 1Depasse (1994), 2Fisher et al. (1993), 3Rakov et al. (1998), 4Crawford (1998), 5Uman et al. (2000) 6Schoene et al. (2003) Strike Object Information Table 6-2: Comparison of re turn stroke statistics.

PAGE 213

183 can probably be attributed to the small sa mple size of the peaks from the 1997 Camp Blanding experiment. The arithmetic mean of the current peaks from the 1999 and 2000 Camp Blanding experiments, 16.2 kA for 64 peaks, (Schoene et al., 2003a ) is som ewhat larger than the arithm etic m eans of the dire ct and nearby strike experiments, which might be attributable to the rela tively high trigger threshold (5 kA) in the 1999 Camp Blanding experiment causing a bias towards events with larger return stroke current peaks (the triggering threshold for the 2000 Camp Blanding experiment was lowered to 3.2 kA). Depasse ( 1994) and Fish er et al. ( 1993) report th a t at least 5% of their currents were below 5 kA. Note that f or the analysis of the direct and nearby strike experiments presented in this dissertation return stroke currents from full-flash records (no triggering threshold) were analyzed when currents from the segm ented records (triggering threshold present) were not available. Dependencies of peak current on lower measurement limit were studied by Rakov ( 1985), on grounding conditions by Depasse ( 1994), on trigger thresho ld level and on g r ounding conditions by Rakov et al. ( 1998), and on triggering structure height by Rakov ( 2001). Depasse ( 1994) hypothesized a dependence between peak current and grounding resistance, base d on a com parison of the data collected between 1985 and 1991 at the KSC (grounding resistance values 0.1.15 ) with the data collected between 1986 and 1991 at Saint-Privat d'Allier, France (grounding resistance of 9 ). Rakov et al. ( 1998) could not find such a dependence using various data sets obtained at the KSC, Fort McClel lan, and Camp Blandi ng, where the grounding resistance ranged from 0.1 to more than hundreds of ohms. Schoene et al. ( 2003a ) observed a tendency for larger current peaks to be associated with better grounding from

PAGE 214

184 the comparison of Camp Blanding measurements23 in different years, although the yearto-year variation of return stroke peaks from experiments with the same grounding conditions (the 1999 and 2000 Camp Blanding experiments) appears to be more significant than the variation for curren t peaks obtained under different grounding conditions. The dependence of the return stroke curre nt peaks on the transferred charge during the first millisecond of the return stroke is discussed in Section 6.2.3.4. 6.2.3.2 Return stroke current 10-90% risetime The histogram of the 10-90% risetimes of the return stroke currents measured during the 1999 through 2004 experiments shown in Figure 6-11a in Section 6.2.2 exhibits two distinguished p eaksone in the 0.2 to 0.4 s range and another in the 1.2 to 1.6 s range. Separating the data in to two histograms, one for all risetimes from the direct strike experiments (Figure 6-11b) and another for all ri setim es from the nearby strike experim e nts ( Figure 6-11c), reveals that the statistic al distributions of the risetimes for the direct and nearby strike experiment ar e d i fferent. The arithmetic mean of 1.4 s for the direct strike experiment cu rrent risetimes is considerably larger than the arithmetic mean of 0.5 s for the nearby strike experiment curre nt risetimes. The di stribution of the current risetimes from the former experiment s resemble a normal distribution, while the distribution of the current ri setimes from the latter experiments may resemble a lognormal distribution, although the sample size of 18 for the risetimes from the nearby strike experiment might be in sufficient to confidently identify the latter distribution type. 23 For the comparison of different Camp Blanding experiments any variations due to different storm types and local topography are mi nimized relative to comparisons of results from different geographical locations.

PAGE 215

185 The principal difference between the dire ct and nearby strike experiments that possibly accounts for the different risetime distributions is the strike objecta metallic object (rocket launcher or intercepting st ructure) connected to a distribution line conductor for the former experiment and an elevated, directly grounded launcher for the latter experiment (Section 3.1). The larger ris etimes in the direct s t rike experiment can likely be attributed to either the larger inductance of the struck line conductor, to reflections of the lightning current at impedan ce discontinuities such as the line arrester, or to a combination of both effects. Note that for the 2004 experiment, we injected the lightning current into the grounded overh ead ground wire, which may electrically resemble the nearby-experiment strike objec t since both strike objects were directly connected to ground as opposed to the struck-phase conduc tors in the 1999 through 2003 direct strike experiments, which were connected to ground through the arresters and line terminators. A comparison of the arithmetic/geometric means of the 10-90% risetimes from the direct strike experiment (1.4 s/1.2 s) with the means from four other non-FPL experiments where an arithmetic or a geometric mean was reported ( Table 6-2) shows that the m eans from the other experim ent s were sm aller. On the other hand, the arithmetic/geometric mean of the 10-90% ri setimes from the nearby strike experiment (0.5 s/0.4 s) is among the lowest means in Table 6-2; only the arithmetic/geometric mean from 1999 and 2000 Camp Blanding expe riment reported in Schoene et al. ( 2003a ) is slightly lo wer (0.3 s/0.2 s).

PAGE 216

186 6.2.3.3 Return stroke current half-peak width Table 6-2 and Figure 6-12 show that the arithmeti c mean and standard deviation of the 122 half-peak widths of the return stroke currents measured duri ng the direct strike experiment (m ean: 23 s, standard deviation: 18 s) are very similar to the mean and standard deviation of the 20 half-peak widths from the nearby strike experiments (mean: 24 s, standard deviation: 16 s) and appear to follow the same log-normal distribution. This similarity suggests that the half-peak width of the channel base current does not depend on the properties of the strike object (as opposed to an apparent dependence of the 10-90% risetimes on the prope rties of the strike object, as discussed in Section 6.2.3.2). On the other hand, Table 6-2 shows that the arithme tic m eans of the direct and nearby strike experim ents are si gnificantly lower than the 50 s mean of 54 half-peak widths from the Saint-Privat dAllier experiment (Depasse, 1994) and the 36 s mean of 11 half-peak widths from the 1997 Camp Blanding experiment (Crawford, 1998) and signif icantly large r than the 13 s mean of 64 half-peak widths from the 1999 and 2000 Camp Blanding experime nt (Schoene et al., 2003a ). The fact that the m eans are di fferent for different experim ents (some of which were conducted at the same geographical location as the FPL experiment) suggests that the half-peak width is influenced by the strike object, although it is not clear which property of the stri ke objects affects the halfpeak width. Generally, stri ke objects with smaller lowcurrent, low-frequency grounding resistances seem to cause smaller half-peak widths. For instance, the lowest mean, 13 s, was found in the 1999 and 2000 Camp Blanding e xperiment where the strike object was a well grounded underground launcher (the launc her was connected to a 70 m x 70 m buried metal grid).

PAGE 217

187 6.2.3.4 Return stroke charge transfer The charge transfer during the return stroke is discussed in this section. The charge transfer was obtained by numerical ly integrating the measured return stroke current over a 1 ms time interval. Table 6-2 shows that the arithmetic m ean and standard deviation of the 122 charge transfers of the return stroke currents measur ed during the direct strike experime nt (m ean: 1.5 C, standard deviation: 1.4 C) are both larger than the ones for the 29 charge transfers from the ne arby strike experiments (mean: 1.1 C, standard deviation: 0.8 C). However, the combined statistical di stribution of the charge transfers from the two experiments shown in Figure 6-13 does resemble a si ngle log-norm al distribution (as opposed to the comb ined distribut ion of the risetim es shown in Figure 6-11, which exhibits two distinguished p eaks as discussed in Section 6.2.3.2). Note that none of the 29 charge transfers from the nearby strike experi m ent is larger than 3 C while 11 of the 122 charge trans f ers from the direct strike e xperiment are larger than 3 C and 3 charge transfers are larger than 7 C. The absence of large charge transf ers during the nearby strike experiment is reflected in the smalle r mean values and standard deviations and might be attributable to statistical variation due to the relatively small number of return strokes during the nearby strike experiment. The return stroke neutralizes the charge deposited by the leader and consequently the return stroke charge and the leader charge are expected to be similar. The charge deposited by stepped leaders th at precede first return strokes in natural lightning is an important parameter for the design of lightning protection as it can be used with a model that specifies the charge dist ribution along the leader channe l to calculate the electric

PAGE 218

188 field on objects, which determines the striking distance24 (Uman, 1987). However, very few statistical data on charge deposited by l eaders preced ing first strokes are available, which motivates the investigation of the corr elation between leader charge and return stroke current peak (the stat istical distribution of return stroke current peaks is well known), since the leader charge can be estimated if such a correlation exists. Berger ( 1972) and Berger et al. (1975) ana l yzed 89 first return stroke current peaks from natural lightning and the corresponding charge transfer red to ground during the first millisecond, or so, and found a weak correlation between the current peak and charge transfer (R2 = 0.5925, where R2 is the coefficient of determination). The current peaks and charge transferred within 1 ms of 143 rocket-triggered lightning re turn strokes during the FPL experiments shown in Figure 6-14 on a log-log scale are better correlated (R2 = 0.76). The power regression equation for the FPL data is y = 12.3x0.54. The regression equation, y = 10.6x0.7, given in Uman ( 1987) for the first return stroke data of Berger ( 1972 ), is also plotted in Figure 6-14 for com parison. Note that the current peaks of the first return strokes in Bergers data range from 8 kA to about 85 kA and the current peaks in our rocket-triggered lightn i ng data range from 4 kA to 40 kA. Interestingly, the regression equations fo r our rocket-triggered lightning data and Bergers first return stroke data are somewhat similar although a comparison is difficult due to the considerable scatter in the two data sets. A consequence of the scatter is 24 The striking distance is the distance between the desce nding leader tip and the to -be-struck object at the time it is certain that the object will be struck. This time is when the electric field enhancement on the object exceeds a critical breakdown va lue so that an upward -directed leader emerges from the object that initiates a return stroke if it connects to the downward-directed leader. The striking distance is important, for instance, for the design of power line protection. 25 Berger plots the charge as a function of peak current (as opposed to the peak current as a function of charge in Figure 6-14). Consequently, the R2 value from Berger (1972) displayed in Figure 6-14 may be slightly different than the actual R2 for Bergers regression equation given in Uman (1987).

PAGE 219

189 probably the fact that the re gression equation for the rocket -triggered lightning strokes for currents larger than 20 kA gives larger charge transfers than the charge transfer obtained from the regression equation for first return stro kes, since this is not an expected result (waveshapes of first return st roke currents are typically wider than the waveshapes of rocket-triggered lightning currents26 and consequently a first re turn stroke current would transfer more charge than a rocket-triggered lightning current with the same current peak, see Section 2.2.3). Figure 6-14: Current peaks as a function of charg e transfer within 1 ms for rockettriggered lightning return strokes (143 individual values and regression line are shown). The regression power equation and R2 value are given. The regression line for 89 negative first return strokes in natural lightning found by Berger is also displayed. The shaded area represents an envelope that encompasses all Berger data points (onl y the outside values that confine the shaded area are shown). 26 The median value of the 2 kA to half-peak width for natural lightning first return stroke currents is 75 s. The median value of the 2 kA to half-peak width for natural lightning subsequent return stroke currents, which are similar to return stroke current s from rocket-triggered lightning, is 32 s (Berger, 1972).

PAGE 220

190 6.3 Arrester Disconnector Operation and Flashovers Information is now presented on the perfor mance of the line configurations tested from 2000 through 2003 regarding (1) arresters rendered inoperative due to disconnector operation and (2) line flashovers. The arresters were not examined for internal damage to the MOV blocks although external arrester damage was often observed. Note that disconnectors are designed not to activate on lightning current, wh ether the arrester operates normally or whether the MOV blocks fail, but rather on the larger rms charge transfer associated with 60 Hz fault current in the event of MOV block failure. Typically, there was more than one lightning flash triggere d to a line per day, but the line could not be inspected for disconnector operation until th e end of the overall triggering session so that there may have been lightning flashe s that did not cause disconnector operation followed by flashes on the same day that did cause disconnector operation. Consequently, it is not possible to determine, in general, the individual flashes which caused disconnector operation. Therefore, the aver age number of disconnector operation per triggering day is given as the measure of the susceptibility of the line configuration to disconnector operation ( Table 6-3). Note that during the 2000 through 2002 experim ents all arresters rendered inoperative were on the struck phase (2000: phase C, 2001: phase A) and no disconnector operated in 2003. For the 2000 horizonta l line experime nt, the disconnector of one arrester (the arrester at pole 8one of the two arresters closest to the strike point) operated during each of the 5 triggering da ys during which ICCs, return stroke currents, and possibly continuing currents were injected into the line. Additionally, one flash without return str okes (ICC only) was positively identified to have operated the disconnector at this location (it was the only flash triggered during that day). For the 2001 vertical line experiment (ICC and return strokes possibly followed by continuing current

PAGE 221

191 Table 6-3: Disconnector ope rations and flashovers dur ing the 2000 experiments. Vertical Line Horizontal Line No OHGW OHGW 2000 2001 2002 2003 2004 Average number of disconnector operations per triggering day (# of triggering days) 1 (5) 2.5 (2) 2 (4) 0 (4) 0 (2) Percentage of return strokes causing flashovers (# of return strokes) 24% (34) 92% (13) 91% (43) 85% (26) 0% (2) Peak currents of return strokes causing flashovers, Arithmetic Mean / Standard Deviation (Sample Size) 18.5 kA / 15.1 kA (7) 17.6 kA / 5.5 kA (12) 15.8 kA / 6.7 k A (39) 13.2 kA / 4.5 kA (18) / (0) Peak currents of return strokes not causing flashovers, Arithmetic Mean / Standard Deviation (Sample Size) 13.1 kA / 5.7 k A (26) 6.0 kA / (1) 9.0 kA / 3.3 kA (4) 9.6 kA / 1.9 kA (4) 10.9 kA / 7.0 kA (2) injected into the line), an average 2.5 disconnectors of the 4 disconnectors installed on the struck-phase arresters operated during each of the 2 triggering days during which ICC, return stroke current, and possibly continui ng current were injected into the line. Arresters at poles 2, 6, and 10 had operate d disconnectors after th e first triggering day and arresters at poles 2, 10, and 14 had ope rated disconnectors after the second triggering day. Additionally, two flashes without return strokes (ICCs only) were positively identified to not have operated any disconnectors (they were the two only flashes triggered during that day). For the 2002 e xperiment (ICC diverted), an average 2 disconnectors of the 4 struckphase arresters operated duri ng each of the 4 triggering days. No disconnector operated during the firs t triggering day, all 4 disconnectors on the struck-phase operated during th e second triggering day, 3 disconnectors operated during

PAGE 222

192 the third triggering day, and 1 disconnector operated during th e fourth triggering day. For the 2003 experiment (ICC diverted and transfor mer on the line), no disconnector operated during the 4 triggering days. For the 2004 expe riment the lightning cu rrent was injected into the overhead ground wire (Section 4.7). No disconnector operation was observed during the unusually large wire burn (9 kA peak value) or during the 2 return strokes triggered in 2004. Table 6-3 also gives information about the percentage of return strokes that caused phase-to-phase flashovers, which were eviden ced by appreciable st roke current being m easured in a phase not subjected to dire ct lightning current injection. For the 2000 horizontal line experiment 24% of the 34 strokes caused phase-to-phase flashovers. For the 2001, 2002, and 2003 vertical line experi m ents flashovers occurred much more frequentlyabout 90% of the 13, 43, and 26 st rokes, respectively, caused flashovers. All 3 flashes that contained stroke s which did not cause any fl ashovers occurred during the 2000 experiment. Statistical information on the peak currents of strokes causing flashovers and of strokes not causi ng flashovers are also given in Table 6-3. On average, the peak cu rrents of the former are larger th en the peak currents of the latter, as expected. However, during all years strokes with very small peak currents (w ell below 10 kA down to 5.6 kA) caused flashover. Note that the total sample size of the return stroke peak currents does not always add up to the tota l number of strokes due to failed current measurements. Detailed information for wh ich strokes/flashes during the 2000 through 2003 experiments flashovers and operated discon nectors were identified can be found in Appendix D (the 2004 experim ent is not include d in the appendix since for this experiment neither flashover nor disconnector operation occurred).

PAGE 223

193 6.3.1 Discussion of Arrester Disconnector Operation The disconnector operations per lightni ng triggering day is discussed now. The likely reason for the more frequent disconnector operati on during the 2001/2002 experiment than during the 2000 experiment ( Table 6-3) was the different number of arrester stations on the line ( 2000: 6 arrester stations, 2001/2002: 4 arrester stations); that is, the therm al heating that caused detonati on of the disconnector was less during the 2000 experime nt since the lightning current divided am ong more arresters. Although the average number of operated disconnectors per triggering day on th e vertical line was slightly reduced from 2001 to 2002 (2001: 2.5 disconnector operations, 2002: 2 disconnector operations), the difference is not st atistically significant. Therefore, there is no experimental evidence that the two ch anges made in 2002, that is, employing two arresters in parallel (vs. single arresters on phase A in 2001) and not injecting the ICC into the line helped reduce the likelihood of disconnector operation. The reason for the insensitivity of the disconnect or operation to whether single or parallel arresters are being used is probably related to the known diffi culties of matching two MOVs connected in parallel due to the intrinsi c nonuniformity of the MOV disks microstructure (He et al. 2005). Ideally, the lightning current through two arresters connected in parallel is equally shared between them which would double th eir joint energy handling capability. This was apparently not the case for the arrest ers connected in para llel during the 2002 experiment. On the other hand, Hitoshi et al. ( 1999 ) inves tigated the effectiveness of using two arresters in parallel by injecting a curr e nt impulse from a surge generator into a 430 m long test distribution line and found that the two arresters equally shared the currents. Note that even though no ICC was injected into the line in 2002, continuing current which has properties similar to the I CC (current of a few hundreds of amperes for

PAGE 224

194 some milliseconds to hundreds of milliseconds after return stroke initiation) followed some of the strokes. Note also that the averag e return stroke charge injected into the line in the initial millisecond was significantly la rger in 2001 than in 2002 (2001: 2.3 C, 2002: 1.2 C, see Table C-3 and Table C-4) which also may acc ount for the slightly larger num ber of disconnector operations per tr iggering day in 2001. The absence of disconnector operation in 2003 vs. 2000, 2001, and 2002, during which there was common disconnector operation, can likely be a ttr ibuted to the pres ence of a transformer on the line in 2003, which shunted the lowmagnitude, low-frequency lightning current components through the transformer prim ary to earth as shown in Section 5.1 and Section 6.4. In other words, the transform er ma y ha ve served to reduce the low-frequency currents through the arresters pr eventing excessive thermal he ating of the disconnectors cartridges, whereas the arresters initially pr otected the transformer from damage due to high-voltage transients. Note th at even though the average return stroke charge injected into the line in the initial m illisecond was slightly larger in 2002 than in 2003 (2002: 1.2 C, 2003: 1.0 C, see Table C-4 and Table C-6) the maximum charge transfer was larger during the 2003 experim ent (2002: 4.1 C, 2003: 6.0 C). Also, the 2003 experime nt contain ed the flash with the largest number of return strokes of all experiments (a 16stroke flash). It is an important result of this study th at disconnectors frequently operated during the 2000 through 2002 experiments even though 60 Hz fault current was not present (disconnectors are designed to operate only on the 60 Hz power frequency fault current that follows MOV block failure). There is direct evidence that long duration currents caused disconnector operation: some disconne ctors operated during events that contained

PAGE 225

195 ICC only. There is also evid ence that disconnector operation on the test distribution lines was not exclusively caused by long-duration currents: the disconnectors of arresters closest to the strike point, which pass the bulk of the impulsive lightning currents during the initial tens of mi croseconds after the return stroke initiation (see part 2 and Mata et al., 2003) operated m ost often. It appears th at di s connector operation on the test distribution lines was typically caused by excessive heating due to the combined charge transferred during both the return stroke current transients and the long duration currents. Significantly, disconnector opera tion on the 2003 line was eliminated, apparently by the absence of the long duration currents through th e arresters due to th e alternative current path provided by the transformer. This imp lies that even though the energy absorbed during both long duration currents and return stroke transients contribute significantly to disconnector operation, as noted above, the energy absorbed during the return stroke current transients alone is t ypically not large enough to cause disconnector operation. It is important to note that this statement app lies to disconnector operation from rockettriggered lightning on the test di stribution lines only and not necessarily to real world distribution lines, where natura l lightning first return stroke currents have typically a much larger energy content than rocket-trigg ered return stroke currents, where the longduration currents divide among many more arresters, and where distribution and zone substations transformers are present. The energy absorbed in the cl osest arresters during the first return stroke transients in natural lightning is estimated in Section 6.4.2. Further, the division of the lightning current on the test distribution li ne is investigated in Section 6.4 and Section 6.5 using experim ental and modeling results and implica tio ns of this investigation to the long-

PAGE 226

196 duration current division on rea l world distribution lines ar e discussed. The absence of disconnector operation during the 2004 experime nt is apparently due to injecting the lightning current into the overhead ground wire, which provided the lightning current with a direct path to ground (during all other experiments the lightning current path to ground included the struck-phase arresters). 6.3.2 Discussion of Flashovers Phase-to-phase flashovers occurred much less frequently during the horizontal line experiment than during the vertical line experiments (Table 6-3) even though the distance of the struck phase to the next closest phase was smaller f o r the horizontal line than it was for the vertical line (horizontal line: 70 cm, vertical line: 80 cm). The reason for the fewer phase-to-phase flashovers on the horizontal line is likely related to one or more of the following differences between the horizontal and vertical line experiments: 1) The differences in arrester spacings (horizontal line: arre ster stations every 3 spans, vertical line: arrester stations every 4 sp ans). The arresters reduce the voltage on the struck phase thereby preventing flashovers. However, the voltage reduction does not occur instantaneously, but is delayed by the time it takes for the voltage signal to travel from the injection point to the arrester, be reversed in polarity, and from there travel back to any poi nt between the arrester sta tions. For the horizontal line, the distances from the lightning current in jection point to each of the two closest arresters were 1.5 spans (span length: 50 m). Due to symmetry, the weakest point on the horizontal line (that is the point on the line where the voltage relief wave from the arresters arrives la st and therefore is most lik ely to experience a flashover) is the current injection point. This mean s that, for the horizontal line, the voltage relief wave has arrived everywhere on the struck phase after 0.5 s (the roundtrip distance from the weakest point to the strike point, 3 sp ans or 150 m, divided by the speed of the traveling wave, we assume c = 3x108 m/s although the actual speed of the wave is slightly less than c). For the vertical line, the distance from the lightning current injection poi nt to the closest arrester in one direction was 1.5 spans and to the closest arrester in the other direction was 2. 5 spans (span length: 58 m). A calculation similar to the one performed for the horizontal line above shows that the voltage relief wave has ar rived everywhere on the line after 0.7 s (that is, 0.2 s later than for the horizontal line). The larger delay time for the vertical line can have a significant impact on the voltage at the weakest point. At the time the relief wave a rrives (horizontal line: 0.5 s, vertical line: 0.7 s) the

PAGE 227

197 voltage at the weakest point on the vertical line is estimat ed to be about 40% larger than the voltage at the weakest point on the horizontal line (assuming the voltage builds up linearly du ring the first 0.7 s27), which increases the probability of flashovers on the vertical line. 2) The presence of voltage measuremen t equipment on the horizontal line. The voltage dividers might have helped to pr event phase-to-phase flashovers by passing current or by facilitating phase-to-neut ral flashovers which could not easily be detected, thereby reducing the potential di fference between the struck phase and the next closest phase. Note th at according to Mata ( 2003) an instrum entation device installed on the vertical li ne in 2002 has likely facilita ted flashovers at pole 7 during flashes FPL0208 to FPL0226. Howe ver, the flashover locations and frequencies determined for the 2002 experim e nt do not confirm this possibility. In fact the flashover frequency between poles 6 and 10 increased after the instrumentation device was removed, that is, flashovers between poles 6 and 10 occurred in 18 out of 28 strokes (64% ) during flashes FPL0208 to FPL0226 when the instrumentation device was present a nd in 12 out of 15 strokes (80%) during flashes FPL0228 and FPL0229 after the in strumentation device was removed ( Appendix D ). 3) The differences in phase arrangem ents. In the p resence of both direct current injection and the electric and magnetic fields of the leader/return stroke combination, the horizontal arrangement of the phases in the horizontal line might suffer a smaller potential difference between the struck phase and the next closest phase than the potential difference of the vertically arranged phases in the vertical line. However, the induced voltages have been calculated using the LIOV-EMTP96 code presented in Section 2.6.3 (the calcu lation results are not presented h ere) and were found to be very sm all compared to the voltages caused by the direct lightning current injection. Interestingly, the percentages of stroke s causing flashovers were essentially the same for all three years during which the ver tical line was tested (2001: 92%, 2002: 91%, and 2003: 85%) even though the disconnector operation during the 2001/2002 experiments and the 2003 experiment was quite different (2.5 and 2 disconnector operations per triggering day for the 2001 and 2002 experiments, respectively, vs. no disconnector operation for the 2003 experiment). Apparently, the tenden cy of the vertical line configuration to experience flashovers is neither significantly influenced by the 27 This is a reasonable assumption since the injected current waveforms, which are expected to have the same waveshape as the voltages, have an average 10-90% riseti me of well above 1 s (see Section 6.2).

PAGE 228

198 number of disconnected arrester s on the line nor by the presen ce of the transformer on the line (a transformer was only present during the 2 003 experiment). The vertical line with overhead ground wire tested in 2004 was the only line configuration which did not experience flashovers. However, the absence of flashovers was not necessarily due to the experimental configuration but might as well be attributed to the small sample size of the return stroke currents injected into the line (the currents of one two-stroke flash were injected, the first stroke had a peak value of 16 kA and the second stroke had a peak value of 6 kA). On the other hand, during the 2001, 2002, and 2003 vertical line experiments all strokes with cu rrent peak values of 16 kA or larger (the largest peak value during the 2004 experiment) caused flashovers. 6.4 Measured Lightning Current Division on the Test Distribution Lines The measured division of the lightning current and charge among the phase-toneutral connections and among th e neutral-to-ground connections is analyzed in the next two sections. The analysis applies primarily to data from strokes that did not cause flashovers or disconnector operation on the li ne. A comparison of the measured division with the model-predicted divi sion can be found in Section 6.5. A usable set of arrester currents during the 2001 experiment was not obt ained due to instrum e ntation problems. 6.4.1 Measured Phase-to-neutral and Ground Current Division Two observed modes of operation for the arrester and ground current divisions during the horizontal and vertical line experiments are defined below: The transient mode indicated by the dark-shaded area in Figure 6-15. This mode occurs during the first tens of microseconds and is characterized by a fast, large magnitude impulsive current throug h each of the two closest struck-phase arresters/grounds and much slower, sma ller magnitude currents through the other struck-phase arresters/grounds. The transient mode ends when the rate of change of all arrester/ground currents becomes similar. The duration of the transient mode

PAGE 229

199 (the width of the dark shaded area in Figure 6-15) is defined as the equilibration time and its determination is discusse d below. The steady-state mode is identified by th e more lightly-shaded area to the right of the dark-shaded area in Figure 6-15. The steady state is characterized by a similar, approxima tely linear rate of change of all arrester/ground currents. Typically, all arrester/ground currents slowly (within hundreds of microseconds) decay to zero during the steady-state mode. Figure 6-15: Struck-phase arre ster currents with the transient mode (dark shaded area) and steady-state mode (light shaded area) indicated. a) 2000 experiment (phase C arrester curr ents during return stroke FPL0032-4) and b) 2003 experiment (phase A arrester currents during return stroke FPL0312-5).

PAGE 230

200 Table 6-4 includes statistical information (arith metic mean, standard deviation, minimum, maximum) of the arrester and gr ound currents equilibration times for the 2000, 2002, and 2003 experiments28. The equilibration time is determined by (1) subtracting the current at one of the two clos est struck-phase arresters/ground s and the current at the next closest struck-phase arrester /ground (horizontal line experiment: phase C arrester/ground current at pole 11 and pole 14, vertical line experiment: phase A arrester/ground current at pole 10 and pole 14), (2) removing the offset (current difference) in case the arrester/ground currents at the two poles do not converge during th e steady-state mode, and (3) measuring the 5%-width of the resul ting pulse (the 5% was chosen so the result would not be influenced by b it noise of the measurement). Table 6-4: Statistical information on the arrester and ground current equilibration times. Horizontal Line Vertical Line 2000 2002 2003 Arithmetic Mean / Stand. Dev. (Sample Size) 39 s / 33 s (n = 11) 43 s / 14 s (n = 23) 29 s / 10 s (n = 17) Arrester Current Equilibration Time Minimum / Maximum 10 s / 114 s 23 s / 74 s 18 s / 51 s Arithmetic Mean / Stand. Dev. (Sample Size) 13 s / 12 s (n = 11) 38 s / 13 s (n = 22) 28 s / 8 s (n = 17) Ground Current Equilibration Time Minimum / Maximum 6 s / 49 s 21 s / 67 s 15 s / 45 s 28 Note that most of the currents analyzed in Table 6-4 were measured during return strokes that caused flashovers and/or disconnector operations. However, we compared the arrester and ground current divisions during strokes that caused flashovers/disconnector operations with the division during strokes that did not cause flashovers/disconnector operations and found that neither occurrence appears to have affected the equilibration time significantly.

PAGE 231

201 Figure 6-15 shows representative examples of arrester current divisions on the struck phase for the horizontal line experiment and the vertical line experiment. Figure 6-15a shows all phase C (the struck phase) arrester currents for stroke FPL0032-4 from the 2000 horizontal line experiment and Figure 6-15b shows all phase A (the struck phase) arrester currents for stroke FPL03125 from the 2003 vertical line experim e nt. Figure 6-15 illust r ates that current variati ons du ring the transient mode in both the horizontal and vertical line expe ri m ents are similar, that is, the two arresters closest to the lightning current injection point (poles 8 and pole 11 arresters for the horizontal line experiment and pole 6 and pole 10 arresters for the vertical line experiment) initially pass the bulk of the return stroke current. However, the steady-state modes are quite different in the two experimentsfor th e horizontal line experiment ( Figure 6-15a) the arrester curren ts at the two closest arresters (pole 8 and po le 11) af ter the equilibration time are much larger than the arrester currents at the other arresters (poles 2, 5, 14, and 17), while for the vertical line experiment ( Figure 6-15b) all f our arrester currents have converged af ter the equilibration tim e. Figure 6-16 shows the individual currents flowing between phase A and neutral divided by the total phase A-to-neutral curr ent for FPL0312, stroke 5 for a period of 100 s. At the time of the peak value of the return stroke current (about 2 s after the return stroke initiation) the two closest arresters pass about 90% of the total phase-to-neutral current. By about 50 s, or so, the total a rrester current is even ly distributed among all arresters. For the first 100 s, the transformer at pole 2 ca rries essentially no current and the terminator at pole 1 carries very little current. The arrester current equilibration time

PAGE 232

202 determined for this stroke is 20 s at which time the two closes t arresters pass 60% of the total lightning current. Figure 6-17 shows the individual currents to ground divided by the total ground current. The ground currents initially behave sim ila r to the arrester curr ents, that is, at the time of peak value the two closest grounds pass about 90% of the total ground current. The percentages of the i ndividual currents at 100 s displayed in Figure 6-17 show that the total ground current at that time is more or less evenly distributed am ong all grounds. The ground current equilibration time determined for this stroke is 20 s at which time the two closest grounds pass 40% of the total lig htning current. Figure 6-18 shows the individual charges transferred from phase A to neutral divided by the total phase A-to-neutral charge f o r the initial 2 ms of the lightning current. At 2 ms, the phase A arrester at pole 6, which is the arrester closest to the lightning current injection point, had carried the most charge (approximatel y 30% of the total charge transferred from phase A to neutral). The transformer and terminator, which transfer essentially no or very little charge during the first 100 s, passed considerable charge by 2 ms (the transformer and the te rminator transferred each 10% of the total charge). Additional information on these currents is found in Section 5.1. Figure 6-19 shows the individual charges tran sferred to ground di vided by the total ground charge for the initial 2 m s of the lightning curre nt. The distribution of the individual charges transferred to ground at 2 ms appears to be inversely proportional to the low-frequency, low-current grounding resi stance of the individua l grounds (e.g., the leas t charg e, 10% of the total charge, is tran sferred at pole 14, the pole with the largest ground resistance, 28 ). The features of the first 100 s of ground current and charge

PAGE 233

203 distributions for the 2003 stroke to the vertical line discussed above are consistent with similar features for other 2003 strokes a nd for strokes in th e 2002 vertical line experiment. The ground current division on the vert ical line is also similar to that on the horizontal line. Alt hough there was no transformer on the line in 2002, the relative distribution of the lightning cu rrents and charges among the phase-to-neutral paths are similar to the 2003 experiment, that is, the individual phase-to-neutral paths in 2003 take the same percentages of the injected li ghtning current and charge as in the 2002 experiment. Figure 6-16: Stroke FPL0312-5, the individual current s flowing from phase A to neutral divided by the sum of all phase A-to-neutral currents on a 100 s time scale. The pole 15 terminator current was not measured and was assumed to be equal to the pole 1 terminator current.

PAGE 234

204 Figure 6-17: Stroke FPL0312-5, the individual curre nts flowing to ground divided by the sum of all ground currents on a 100 s time scale. The low-frequency, lowcurrent ground resistance of each of the pole grounds is given in the parentheses. The percentages of the individual currents at 100 s are displayed on the right side. Figure 6-18: Stroke FPL0312-5, the individual charges flowin g from phase A to neutral divided by the sum of all phase A-to-neu tral charges, on a 2 ms time scale. The pole 15 terminator charge was not m easured and was assumed to be equal to the pole 1 terminator charge.

PAGE 235

205 Figure 6-19: Stroke FPL0312-5, the individual charges flow ing to ground divided by the sum of all ground charges, on a 2 ms time scale. The dc ground resistance of each of the pole grounds is given in the parentheses. The percentages of the individual charges at 2 ms are displayed on the right side. 6.4.2 Discussion of the Measured Lightning Current Division During the horizontal line experiment the two closest struck-phase arresters passed the bulk of the lightning current during both the transient mode and the steady-state mode, these modes being defined in Section 6.4.1. During the vertical line experim ent the transient mode behavior of the arresters wa s sim ilar to tha t during the horizontal line experiment, but the steady-state mode behavi or was quite different that is, during the steady-state mode the lightning current in the vert ical line was uniformly divided among all struck-phase arresters ( Figure 6-15a) while it was not for the horizontal line where the closes t arresters carried the m o st current ( Figure 6-15b). The reas on for the different steady-state mode behavior of th e horizontal and vertical lin es is not known but is likely rela ted to one of the three following diffe rences between the two experiments:

PAGE 236

206 A change of the VI-characteristics of the two closest arresters (model 2 in Section 6.5). The VI-characteristics of the two arre st ers installed on the struck ph ase of the 2000 horizontal line that were cl osest to the strike point might have changed due to large energy absorption. The arresters on the 2000 horizontal line experiment were exposed to initial continuous currents (ICCs), to the follo wing fast return stroke currents, and sometimes to return stroke continuing currents, while the arresters on the 2002/2003 vertical line were exposed to the latter two currents only and not to ICCs. Thus the two closest arresters on the 2000 horizontal line absorbed more energy than the other arresters, the other arresters being the remote arresters on the 2000 horizontal line (these arresters do not pass much current during the transient mode as shown in Section 6.4.1) and the arresters on the 2002/2003 vertical line (these arresters were not exposed to the ICC). Modeling results presented in Section 6.5 show that (a) a good overall m atch between the modeled and m easured arres ter currents on the horizontal line can be achieved by reducing the residual voltage of the two closest arresters by 20% (model 2) and (b) the arrester voltage measured at pole 11 and at pole 8 is better reproduced by model 2 than by model 1. Both observations support the view that th e two closest arrester s changed their VIcharacteristics. The use of different types of arresters (horizontal line experiment: the arresters at the two closest arrester stations were Ohio -Brass arresters, the arresters at all other arrester stations were Cooper arrester s; 2002/2003 vertical line experiment: only Cooper or only Ohio-Brass arresters were us ed at all stations). Note that the two types of arresters have the same rate d voltage and very similar published VIcharacteristics for currents above 1.5 kA (based on an 8/20 s waveform, see Table 3-9). However, it is known that if two similar arresters that are not carefully matched are connected in parallel can be have such that one arrester passes considerably more current than the other arrester (Section 6.3.1). For the c ase of the horizontal line during the st eady-state mode the impeda nce of the line conductors separating the 6 struck-phase arres t ers is very small and the arresters can be viewed as connected in parallel. It is possible that the two clos est arresters (Ohio-Brass) on the horizontal line pass more current duri ng the steady-state mode since they are not matched with the other arresters (Coope r). Note that during the transient mode the impedance of the phase conductors separa ting the arrester st ations is large and consequently the arresters ca nnot be viewed as connected in parallel (the primary current paths during the transient mode ar e the two closest arresters, matched or unmatched, due to the large impedance separating them from the other arresters on the line, as discussed below). The presence of voltage measurement equi pment at the two closest arresters on the struck phase (phase C) for the horizonta l line experiment that was not present on the vertical line. The voltage dividers possibly have fac ilitated flashovers resulting in an additional phase-to-neut ral current path at the two closest arrester stations. This would have resulted in an overestim ation of the arrester current since the arrester current sensor measures the sum of the current through the arrester and the current through the voltage divider. Howeve r, flashovers were not evidenced in the current, voltage, or optical records.

PAGE 237

207 We now discuss the implications of the m easured and modeled current divisions for real world distribution lines. On both the horizontal and vert ical test distribution line the arresters closest to the strike point pass the bulk of the curr ent during the transi ent mode (Section 6.4.1). The reason f or this is the indu ctance of the line segment that separates the closest arrester station from the next-closest station (that is, the impedance of the segment is large during the transient mode, which delays current flow to the next-closest arrester). The current division during the transient mode determin ed on the test distribution lines can be expected to be similar on real world dist ribution lines. Note that the equilibration time depends strongly on the length of the line se gments between arrester stations, for instance, the smaller inductance on shorter se gments causes the current to equilibrate faster. This has been previously shown by McDermott (2006) with EMTP m odeling. It is presently unclear if, on a real worl d distribution line, (a) the lower-f requency current components would be evenly divided among all arresters as the experimentallydetermined current division in the 2002/ 2003 vertical line experiment reproduced by model 1 in Section 6.5 or (b) the lower-frequency cu rrent components would flow prim arily through the two closest arresters as the experime ntally-determ ined arrester current division in the horizontal line experiment reproduced by model 2 in Section 6.5. It is possible that, sim ilar to the effect hypothesized as the reason for the model 2 current division on the horizontal line, the energy absorbed in the arresters clos es t to the strike point of a natural lightning st rike to a real world dist ribution line is large enough to change their VI-characteristics. Support for this view is found in Nakada et al al. ( 2003) where the reduction of th e measured arrest er voltage by 50% during a natural strike is

PAGE 238

208 attributed to a change in the arresters V I-characteristic due to energy absorption and a simple arrester model in which the manufact urer-provided residual voltage is reduced by 50% is introduced. Also, an argument for model 2 is made in Fernandez et al. ( 1999) from the fact that Barker et al. ( 1993) measured a significan t low-frequency arrester curren t (app roxim ately 2 kA after 1 ms) in a 10 kV MOV arrester installed on an actual power distribution line. Note that during a natural lightning strike to a real world distribution line the arresters closest to the strike point are expect ed to absorb the largest amount of energy of all arresters on the line (the closest arresters pass the bulk of the current during the transient mode) and that the en ergy content of a natural lightning return stroke is typically significantly larger than the energy conten t of return strokes from rocket-triggered lightning. It was estimat ed by Mata et al. ( 2003), based on the m odel 2 lightning return stroke current division determined in the hor izontal line experiment and the available statistics on the current amplit udes and waveshapes of first strokes in natural lightning, that, within about 450 s of the initiation of the first retu rn stroke current flow, the energy input from about half of all natural lightni ng first strokes delivered to each of the two closest arresters exceeds 70 kJ and thus would likely damage them (in the absence of flashovers or other alternative paths for the re turn stroke current to bypass the arresters). If model 1 applies to the current division on real world distribution lines, then the arrester damage rate on real world lines would be smaller than the arrester damage rate on the horizontal test line es timated by Mata et al. ( 2003) sin ce the low-frequency current components are divi ded evenly am ong multiple a rresters on the line (that is, the arresters further away from the lightni ng strike point absorb a significant portion of the lightning

PAGE 239

209 energy and thus help protect the arresters closer to the lightning strike point from damage and degradation). For instance, McDermott ( 2006) using the EMTP found a considerably lower arrester-absorbed energy (30 kJ) than Mata et al. (70 kJ) for a typical natu ral lightning first return stroke cu rrent injected into the phase conductor of a distribution line, which is m ostly attributable to the fact that McDermott adopted the model 1 current division and Mata et al. adopted the model 2 current division. It is important to note that for both model 1 and model 2 the MOV block of arresters on real world lines closest to the lightning strike poi nt may be damaged since they pass most of the impulsive lightning current and therefore absorb most of the ener gy during the transient mode. This energy is defined here as the minimum arrester-absorbe d energy and is invest igated further in the next section. It was found for the vertical line that in itially the bulk of th e total ground current goes through the two closest grounds and at 100 s the total ground current is more or less evenly distributed among all grounds (Section 6.4.1). Interestingly, Mata et al. ( 2003) found ground currents on the horizontal line after 25 s (and the charge transfer within 100 s, 500 s, and 1 ms) to be roughly inversely proportional to grounding resistance. The reason why the inverse proportionality for ground currents at 100 s on the vertical line was not obs erved is likely the relative small variance of the lowfrequency grounding resistances for the vertical line (the measured low-frequency, lowcurrent grounding resistances ranged from 22 to 55 for the horizontal line and from 18 to 28 for the vertical line, see Section 3.3). It was found that fo r the vertical line the individual charges transf erred to ground (the integrated ground currents) at 2 ms are roughly inversely proportional to the measured low-fre quency, low-current grounding

PAGE 240

210 resistance of the individual ground (Section 6.4). Apparently, the tre nd of the individual ground currents to be distributed inversely to the individual gr ounding resistances, not observed for the first 100 s, is revealed by integrati ng the ground currents over 2 ms. 6.5 Modeled Lightning Current Divisions on the Test Distribution Lines In this section the divisions of the lightning current on th e horizontallyand vertically-configured test distribution lin es are modeled using the Electromagnetic Transient Program 1996 (EMTP96), version 3.2d. The model-predicted results are compared to the experimentally-determined data to (1) test the validity of the model representations of distribution lines, arresters, and line groundings that are commonly employed for the distribution li ne design and (2) test specu lations raised in Section 6.4 regarding the different current divisions on the horizontallyand vertically-configured lines. The lightning was represented in the mode l as a parallel com bination of Type-13 current sources (ram p functions with linear decay) to genera te a piecewise linear approximation of the measured chan nel base lightning current (Section 2.5.2). Idea l current sources were used, which is equivalent to the lightning cha nnels characteristic impedance being infinity. The lightning current was either inje cted in to the phase C conductor midway between poles 9 and 10 of the model of the horizontally-configured distribution line system or into the phase A conductor midway between poles 7 and 8 of the model of the vertically-configured line sy stem. The modeled distribution line systems consisted of distribution line sections represented by the frequency-dependent (JMARTI) transmission line model (Section 2.5.4), line groundings m odele d using a distributed ground mo del (Section 2.5.6 ) and the measured low-fre quency, low-current grounding

PAGE 241

211 resistances (Section 3.3), and a gapless arrester model with the manuf acturer-provided V I-characteristic (Section 3.4) or with a modified VI-cha racteristic. Note that the frequency-dependent transmi ssion line m odel and the arrester m odel are build-in components of the EMTP. The model-predicted results are compared to currents measured on the distribution line. The division of rocket-triggered lightning curre nts injected into the phase conductors during the 2000 and 2003 experiments was prev iously modeled by Dr. Carlos Mata (2000 experiment: Mata, 2000 and A.G. Mata et al., 2002; 2003 experim ent: Schoene et al., 2003a ) using the Advanced Transient Program (ATP). 6.5.1 Modeled Lightning Current Division on the Horizontal Line Currents successfully measured at six of the eight phase C-to-neutral connections (five arresters and one line terminato r) and at the eight line groundings (see Figure 4-2 in Section 4.3) on the horizontally-conf igured line test ed in 2000 for stroke FPL0036-1 are compared to m odeled results. The curren t division during stroke FPL0036-1 was previously modeled in Mata ( 2000) and A.G. Mata et al. ( 2002). The sa me return stroke was chosen f o r modeling in this dissertation to better compare the modeling results presented here with the mode ling results published previous ly. The lightning current was measured at the channel base and then in jected into the phase C conductor midway between poles 9 and 10. An adjustment factor of 0.75 was applied to the lightning current due to the apparent over estimation of the measured lightning current (Section 6.1.4). The m easured lightning current w as low -pass filtered with a digital 3rd order, 3 MHz Butterworth filter in order to facilitate the reconstruction of the lightning current in the model using a piecewise-linear approximation. Figure 6-20 shows the measured lightning

PAGE 242

212 Figure 6-20: Stroke FPL0036-1, Unfiltered a nd filtered channel base currents and its piecewise-linear approximation. The pi ecewise linear approximation of the filtered current was used as input to the model. Displayed on a) 5 s and b) 700 s time scales.

PAGE 243

213 current of stroke FPL0036-1 (unfiltered and filtered) and the light ning current used as model input. Two models are considered here: (1) model 1 us es the published VIcharacteristics for all modeled arresters and (2 ) model 2 uses a modifi ed VI-characteristic for the two arresters closest to the lightn ing current injection points (poles 8 and 11 arresters installed at phase C) and the published VI-characteris tic for all other arresters. The modified VI-characteristic is the manufacturer-provided VI-characteristic of the Ohio-Brass arresters ( Table 3-9) with the voltage values reduced by 20%. The 20% voltage reduction was selected by trial and erro r so that an optim al match between modeled and measured arrest er currents at pole 11 of th e horizontal line was achieved (the current through the other closest phase C arrester at pole 8 was not successfully measured). For the successful calculation of the observed current division on the vertical line, the published VI-characteristic was used for all arresters, that is, model 1; whereas for the horizontal line, model 2 provided a successful match between experimental and modeled results. Figure 6-21 compares the overall division of the m easured struck-phasetoneutral currents and ground currents with the results from model 1 and model 2. Contrarily to model 1, which poorly reproduces the overall di vision of arrester and terminator currents (left side of Figure 6-21), m odel 2 reproduces the ove rall division very well. The m ost significant difference between model 1 and mode l 2 results is that the current through one of the two closest arres ters e qualizes during the steady-state mode in model 1 (similar to the closest arrester currents during the 2002 and 2003 experiments) while in model 2 the closest arrester current is considerably la rger during the steady-state mode than the currents through the other arrester (similar to the measured clos est arrester current). Both,

PAGE 244

214 model 1 and model 2, reproduce well the overall measured di vision of the ground currents (right side of Figure 6-21). Note that both m odels reproduce the equilibration time, defined in Section 6.4, very well for the arrester currents (about 50 s) and for the ground currents (about 30 s). For a more accurate comparison of the measured and modeled results the individual measured and modeled phase-to-neutral cu rrents were overlaid and displayed on 50 s and 500 s time scales in Figure 6-22 and Figure 6-23, respectively. The comparison shows that both models reproduce th e measured currents very well on a 50 s time scale with the model 2 results matching the measured currents slightly better. Figure 6-23 shows that model 1 does not re p roduce the m easured curren t well on a 500 s time scale while model 2 reproduces the m easured arrester currents at poles 11, 14, and 17 very well during the total 500 s duration displayed in the figure. The decay after the peak of the model 2 arrester currents at poles 5 and 2 is cons iderably faster than the decay of the measured currents. After the steep decay, during 250 s and 500 s, the modeled and measured pole 5 and pole 2 arrester currents match very well. A similar comparison of the measured and modeled ground currents, displayed in Figure 6-24 and Figure 6-25, shows that the model 1 resu lts and the model 2 resu lts ar e essentially identical and that both models reproduce the measured ground currents well on 50 s and 500 s time scales. Note that the m odeled ground currents shown in Figure 6-25 on a 500 s time scale are generally slightly larger than the measured ground currents.

PAGE 245

215 Figure 6-21: Stroke FPL0036-1, comparison of the overall phase Cto-neutral current divisions (left) and ground current divisions (right) of measured currents (top), model 1 currents (center), and model 2 currents (bottom) on a 500 s time scale.

PAGE 246

216 Figure 6-22: Stroke FPL0036-1, individual co mparison of all successfully measured phase C-to-neutral currents with model 1and model 2-predicted results on a 50 s time scale.

PAGE 247

217 Figure 6-23: Stroke FPL0036-1, individual co mparison of all successfully measured phase C-to-neutral currents with model 1and model 2-predicted results on a 500 s time scale.

PAGE 248

218 Figure 6-24: Stroke FPL0036-1, individual co mparison of all measured ground currents with model 1 and model 2 predicted results on a 50 s time scale.

PAGE 249

219 Figure 6-25: Stroke FPL0036-1, individual co mparison of all measured ground currents with model 1and model 2-predicted results on a 500 s time scale.

PAGE 250

220 6.5.2 Modeled Lightning Current Division on the Vertical Line Currents measured at six of the seven phase A-to-neutral connections (four arresters, one of the two line terminators, and one transformer) and at the six line groundings (see Figure 4-5 in Section 4.6) on the vertically-configur ed line tested in 2003 for stroke F PL0312-5 and stroke FPL0315-1 are com p ared to modeled results. The lightning currents of both strokes were measured at the channel base and then injected into the phase A conductor midway between pol es 7 and 8. To keep the model simple only strokes that did not cause flashovers on the line were chosen for modeling. Furthermore, a small peak current, short dur ation stroke (FPL0312-5) and a large peak current, long duration stroke (FPL0315-1) were chosen to test the model for strokes with different characteristics. The peak current of stroke FPL0312-5 was 9 kA and current ceased to flow after about 350 s. Stroke FPL0315-1 was the st roke with the largest peak current (12 kA) of all 2002 and 2003 strokes that neither caused flashovers on the line nor flashed over from the intercepting structur e on the tower to the tower launcher ( Appendix D ). The return stroke current of stro ke FPL0315-1 had continuing current of about 100 A, which ceased to flow after abou t 125 ms The m easured lightning current of stroke FPL0312-5 was low-pass filtered with a digital 4th order, 3 MHz Butterworth filter in orde r to reduce the noise thereby facilitating the reconstruction of the lightning curre nt in the model using piecewise-linear approximation. Figure 6-26 shows the measured light ning current of stroke FPL0312-5 (unf iltered and f ilter ed) sampled at 20 MHz with the 8-bit LeCroy oscilloscopes (Section 3.6.5 and Section 4.6) and the lightning current used as m odel input. Figure 6-26 shows the measured lightning current of st roke FPL0315-1 sampled at 2 MHz with the

PAGE 251

221 12-bit Yokogawa oscilloscopes (Section 3.6.5 and Section 4.6) and the piecewise-linear approxima tion of the measured lightning curren t, which was used as model input. Note that reconstructing the lightning current from the filtered LeCroy data, as it was done for stroke FPL0312-5, takes advantage of the higher bandwidth of the LeCroy data while reconstructing the lightning current from the Yo kogawa data, the approach chosen for the reconstruction of the return stroke current of stroke FPL0315-1, utilizes the higher vertical resolution and longer dur ation of the Yokogawa records. Figure 6-26: Stroke FPL0312-5, unfiltered and filtered channel base currents and its piecewise-linear approximation. The pi ecewise linear approximation of the filtered current was used as input to the model. Displayed on a) 5 s and b) 350 s time scales.

PAGE 252

222 Figure 6-27: Stroke FPL0315-1, measured chan nel base current and its piecewise-linear approximation. Displayed on a) 10 s and b) 800 s time scales. Figure 6-28 and Figure 6-29 compare the overall divi sion of the measured struckphasetoneutral currents during stroke s FPL0312-5 and FPL0315-1 with the modelpredicted results. The current divisions of the phase-to-neutral connections between the current injection point and pole 1 (poles 6, 2, and 1), and betw een the current injection point and pole 15 (poles 10 and 14) are displayed on the left and right side, respectively. Note that the pole 15 terminator currents ar e not included since they were not measured. The overall measured current divisions in Figure 6-28 resemble the model-predicted curren t divi s ions very well. Note that the equilibration times (Section 6.4) for the

PAGE 253

223 measured and modeled arrester current divisions are simila r (FPL0312-5: about 40 s for both measured divisi ons and about 35 s for both modeled divisi ons, FPL0315-1: about 50 s for both measured divisions and about 40 s for both modeled divisions). Figure 6-28: Stroke FPL0312-5, comparison of the overall measured and modeled phase A-to-neutral current divisions at (left) poles 1, 2, and 6 and (right) poles 10 and 14 on a 100 s time scale.

PAGE 254

224 Figure 6-29: Stroke FPL0315-1, comparison of the overall measured and modeled phase A-to-neutral current divisions at (left) poles 1, 2, and 6 and (right) poles 10 and 14 on a 100 s time scale. Figure 6-30 and Figure 6-31 compare the overall di vision of the measured ground currents during strokes FPL0312-5 and FPL0315-1 with the model-pred icted results. The ground current divisions between the current injection point and pole 1, and between the current injection point and pole 15 are displaye d on the left and right side, respectively. The overall m easured ground current division s resemble the model-predicted divisions very well. Note that the equilibration times (Section 6.4) for the m easured and modeled

PAGE 255

225 ground current divisions are very similar (FPL0312-5: about 20 s for the modeled and measured divisions between the lightning curr ent injection point a nd pole 1 and about 35 s for the modeled and measured divisions be tween the lightning current injection point and pole 15, FPL0315-1: about 30 s for the modeled and measured divisions between the lightning current injecti on point and pole 1 and about 40 s for the modeled and measured divisions between the lightni ng current injection point and pole 15). Figure 6-30: Stroke FPL0312-5, comparison of the overall measured and modeled ground current divisions at (left) poles 1, 2, and 6 and (right) poles 10, 14, and 15 on a 100 s time scale.

PAGE 256

226 Figure 6-31: Stroke FPL0315-1, comparison of the overall measured and modeled ground current divisions at (left) poles 1, 2, and 6 and (right) poles 10, 14, and 15 on a 100 s time scale. For a more accurate comparison of the measured and modeled results the individual measured and modeled arrester currents during strokes FPL0312-5 and FPL0315-1 were overlaid and displayed on a short time scale (50 s) in Figure 6-32 and Figure 6-33, respectively, and on larger time scales (500 s and 800 s) in Figure 6-34 and Figure 6-35, respectively. The comparison show s that generally the model reproduces the measured arrester currents very well on both tim e scales. Th e only exception being some oscillations after the initial peak present in the modele d currents of stroke FPL0312-5

PAGE 257

227 through the two closest arresters at pole 6 and pole 10, which are not well resolved in the measured closest arrester current waveforms ( Figure 6-32). Figure 6-32: Stroke FPL0312-5, individual co mparison of all measured phase A-toneutra l cur rents with mode l-predicted results on a 50 s time scale.

PAGE 258

228 Figure 6-33: Stroke FPL0315-1, individual co mparison of all measured phase A-toneutral currents with mode l-predicted results on a 50 s time scale.

PAGE 259

229 Figure 6-34: Stroke FPL0312-5, individual co mparison of all measured phase A-toneutral currents with mode l-predicted results on a 500 s time scale. Note that the pole 15 terminator current was not measured.

PAGE 260

230 Figure 6-35: Stroke FPL0315-1, individual co mparison of all measured phase A-toneutral currents with mode l-predicted results on an 800 s time scale. Note that the pole 15 terminator current was not measured.

PAGE 261

231 A similar comparison of the measured and modeled ground currents during stroke FPL0312-5 and FPL0315-1 displayed in Figure 6-36, Figure 6-37, Figure 6-38, and Figure 6-39 shows that the model reproduces the m easured ground currents well on the short and the larger tim e scales. Oscillations after the peak similar to the ones present in the modeled arrester currents are present in both the modeled and measured ground currents. Note that the frequency of the osci llations in the measured ground currents is slightly larger than the frequency of the oscillations in the m odeled ground currents. Figure 6-36: Stroke FPL0312-5, individual co mparison of all measured ground currents with model-predicted results on a 50 s time scale. Note that the pole 15 terminator current was not measured.

PAGE 262

232 Figure 6-37: Stroke FPL0315-1, individual co mparison of all measured ground currents with model-predicted results on a 50 s time scale. Note that the pole 15 terminator current was not measured.

PAGE 263

233 Figure 6-38: Stroke FPL0312-5, individual co mparison of all measured ground currents with model-predicted results on a 500 s time scale.

PAGE 264

234 Figure 6-39: Stroke FPL0315-1, individual co mparison of all measured ground currents with model-predicted results on an 800 s time scale.

PAGE 265

235 6.5.3 Simple Model of the Lightning Current Division It has been shown in Section 6.5.1 and Section 6.5.2 that the EMTP model employed reproduces the m easured currents on the horizontal and vertical line very well for the case that no flashover or disconnect or operation occurs and the manufacturerprovided VI-characteristic is assumed. However, the question why the lightning current initially flows through the closest arrester has not been answered yet in this dissertation. The purpose of this section is to gain some insight and intuitiv e understanding of the arrester current division. This is accomplis hed by reducing the comp lex test distribution line to a distribution line with two arrester stations without a neutral conductor, with the lightning current being injected into one end of the line. This idealized line is modeled both with the sophisticated EMTP technique employed in the previ ous two sections and with a simple RL-circuit that produc es essentially the same result. Figure 6-40a shows the idealized d ist ribu tion line. The idealized line is the vertical test line with the following simplifications: The idealized line does not have a neutral conductor (the vertical test line has a neutral conductor present). The idealized line does not have line term inators (the vertical test line has line terminators present). The idealized line consists of only two arrester stations that are separated by n spans (n being 3, 4, or 5) each span having a length of 50 m (the vertical test line has 4 arrester stations with any two adjacent stations separated by 4 spans). The lightning current injection point is at the closest arrester station (the vertical test line has the lightning current injection point at midspan). Figure 6-40b shows the simplest possible circuit of the idealized lin e. In this sim ple circuit the complexity of the EMTP model has been reduced as follows: The distributed parameters of the FD transmission line model (Section 2.5.4) that represent the distribution line condu ctors be tween arrester stations are replaced by a

PAGE 266

236 single inductance (inductance L in Figure 6-40b). This simplification assumes that the struck distribution line conductor behaves primarily as an inductor for the duration of the simulation (100 s), that is, the resistan ce and capacitance of the line conductor are negligible. The non-linear resistance of the arrester s and groundings are represented as a single, constant resistor at each arrester station (resistance R in Figure 6-40b). It is assumed that the value of this res istor is 20 Figure 6-40: Idealized distribution line and its circuit repres entation. a) Sketch of the idealized three-phase distri bution line consisting of a close and distant arrester station separated by n spans. b) Ci rcuit of the model representing the distribution line shown in a). The cu rrents in each branch of the line representations are indicated. The lightning current ii in the simple model is repres ented as an ideal current source (as it is in the EMTP model). The current iR is the current flowing through the close arrester and then to ground (in the simple circuit the sum of arrester resistance and grounding resistance is represented as resistance R). The current iLR is the current flowing

PAGE 267

237 (1) through the n spans of the line conductor (represented in the simple circuit as inductance L), (2) thro ugh the far arrester, and (3) to ground through resistance R (two resistances, arrester and ground, are again represented in th e simple model as resistance R). The differential equation th at describes the circuit in Figure 6-40b can be obtained as f ollows: Applying Kirchoffs current law yields LRRiiii 6-23 Applying Kirchoffs voltage law yields dt di LRiRiLR LR R. 6-24 Eliminating iR by combining Equations 6-23 and 6-24 yields the differential equation dt di iiLR LR i21 6-25 where = L/R. The solution of Equation 6-25 is given by t i t LRe dtie i 2 21 6-26 Combining Equations 6-23 and 6-26 yields t i t iRe dtie ii 2 21 6-27 If is is a step function u(t) with amplitude I0, then Equations 6-26 and 6-27 become )( 2 1 2 12 0tueIit LR 6-28 and )( 2 1 2 12 0tueIit R 6-29

PAGE 268

238 As discussed in Section 6.4, during a lightning strike to a real world distribution line the minimum energy absorbed by the arrest er closest to the strike point is the energy absorbed during the impulsiv e current flow through the cl osest arrester (the lowfrequency lightning current may or may not contribute significantly to the absorbed energy). The impulsive current can be defined as the difference of the current iR (this current consists of highand low-fr equency components) and the current iLR (this current consists of only low-frequency components) Also, the charge transferred through the closest arrester can be used as a measure for the arresters energy absorption, since the transferred charge and the absorbed energy are strongly linear corre lated, as shown by Mata ( 2000). The transferred ch arge Q due to the impulsive current is defined as dtiiQLRR. 6-30 For the case of a step function with amplitude I0, Equation 6-30 integrated from t = 0 to t = becomes 20 IQ 6-31 The current through th e close arrester iR and the current thro ugh the far arrester iRL calculated with the simple model (Equations 6-28 and 6-29) for I0 = 1 kA and = 10.5 s are illustrated in Figure 6-41. The yellow area enclosed by the two currents represen ts the defined charge Q (Equation 6-31). It is apparent from Figure 6-41 that initially the cl ose arrester passes the total lightning current a nd no current flows through the far arrester due to the inductance L in the circ uit. The current through the close arrester decreases at the sam e rate as the current through the far ar rester increases until both currents are essentially the same after a few tens of microseconds. The tim e constant defined above

PAGE 269

239 as the ratio of the inductance L and the resi stance R governs the time it takes to equalize both currents. Note that in the simple model is proporti onal to the length of the line since L can be calculated as the product of the line length and the inductance per unit length. Figure 6-41: Currents through the close arrester (iR) and the far arrester (iRL) calculated using the simple model. The yellow-sh aded area enclosed by the two currents represents the charge transfer Q through the close arrester due to the impulsive current. Figure 6-42 compares the current through the close arrester and far arrester calcu lated for the idealized dis t ribution line w ith the simple circuit with the currents calculated using the EMTP model for 3, 4, a nd 5 span lengths between the two arrester stations. The time constant per span was used as a free parameter and was chosen to be 3.5 s per span, that is, for 3, 4, and 5 spans the time constant displa yed in the figure is 10.5 s, 14 s, and 17.5 s, respectively. Additionally, Figure 6-42 displays the values for t he defined Q obtained from the sim p le circuit and the EMTP model by numerically integrating the difference of the clos e and far arrester currents over 50 s.

PAGE 270

240 Figure 6-42: Currents through the close and far arresters and Q calculated with the simple model and the EMTP model for 3, 4, and 5 span lengths between the two arrester stations. The time constants used in the simple model are given.

PAGE 271

241 The following observations can be made from Figure 6-42: The values for Q obtained from the simple circu it and the EMTP model are very similar for 3, 4, and 5 span lengths, show ing that a single time constant per span length can be found for the simple circui t to achieve a very good agreement of simple-circuit and EMTP-model results. The good agreement of Q values obtained from the simple circuit and the EMTP circuit implies that the simple circuit is sufficient to estimate the absorbed energy of the arresters if the time constant per span length is known. The Q values obtained by numerically integrating the simple circuit and EMTP model results over 50 s (that is, 3 to 5 time constants) agree with the Q values calculated with Equation 6-31. 6.5.4 Discussion of the Modeled Lightning Current Division Mata ( 2000) modeled the current division during the first stroke in flash FPL0036. The measured and mode led ground currents in Mata ( 2000) match reasonably well if the measured low-frequency, low-current resist ances of the line groundings are used in the model and match better if the measured channel base cu rren t that is used as an input to the model is scaled by a factor of 0.68 a nd a set of adjusted grounding resistances are used in the model. An optim ization algorithm that varies the grounding resistances in the model to find the best match between modeled and measured ground currents was used by Mata ( 2000) to calculate the adjusted grounding resistances. The scaling factor was obtained by adjusting the char ge inje cted into the line to match the to tal c harge transferred through all line groundings during the first 100 s. Note that the scaling factor for this stroke varies considerably with th e duration of the charge transfer (a scaling factor of 0.68, 0.57, and 0.56 is necessary to match the injected charge and charge transferred to ground during 100 s, 500 s, and 1 ms, respectively). An adjustment factor for the lightning current of 0.75 was c hosen in this dissertation (as opposed to 0.68 in Mata, 2000). A factor of 0.75 generally pr ovides a good match between injected

PAGE 272

242 charge, total charge to ground, and total phase-to-neutral charge (in the absence of flashovers) for most strokes during the 2000 experiment (Section 6.1.4 and Appendix B ). It was not deemed necessary to adjust the measured grounding resistances in the models considered here, since using the m easured grounding resistances in m odel 1 and model 2 already provides a good match between measured and modeled results. The measured struck-phase arrester curre nts on the horizontally-configured line for stroke FPL0036-1 were not successfully modeled in Mata ( 2000). The modeled arrester current at pole 11 (one of the two arresters closest to the lightni ng current injection point between poles 9 and 10) exhibited damped ringing (period about 1 s; la rgest magnitude about 7 kA) during the first tens of microseconds that were not present in the measured arrester current. The other modeled arrester cu rrents at poles 2, 5, 14, and 17 presented in Mata ( 2000) did not exhibit ringing but generally m atched the measured currents poorly if the measured grounding resistances were us ed in the m odel and better, but still not satisfactory, if the measured channel base curr ent that is used as an input to the model was scaled by a factor of 0.68 and the adjusted grounding resistances were used in the model. According to Mata ( 2000) the ringing is no t present when (1) a sl o wer current source is used, (2) a crude lumped corona mode l is implemented (a resistor in parallel with a capacitor to ground at the strike poi nt), and/or the characteristic impedance (between 500 and 2000 ) of the lightning channel is taken into account. Mata ( 2003) modeled the current division on the horizontal line for various values of the lightning channels ch aracteristic impedance. The va lue of the characteristic impedance does no t appear to have a sign if icant effect on the modeled curre nt division except for the ringing in the modeled closest arrester currentthe ringing was damped considerably for low

PAGE 273

243 values (below 500 or so) of the characteristic impedance. The reason for the ringing in the modeled arrester currents in Mata (2000) and A.G. Mata ( 2002) is unknown. None of the ar rester curren ts modeled in this dissertation exhibited similar ringing. For the horizontally-configured line the phase-to-neutral cu rrents calculated using model 1 poorly match the measured currents during the steady-state mode (the steadystate mode is defined in Section 6.4.1). The overall phase-to-neutral current division in model 1 does not resem ble the measured cu rrent division since a ll m odeled arrester currents equalize during the steady-state mode, which is similar to the arrester current division observed during the 2002 and 2003 experiments on the vertically-configured line, while the measured currents on the horiz ontally-configured line do not equalize (the closest arresters pass the bulk of the current during the steady-state mode). On the other hand, in model 2 a generally good match be tween all measured and modeled phase-toneutral currents is achieved by modifying th e VI-characteristics of both closest arresters (that is, reducing the residual vol tage by 20%) so that the current measured in one of the two closest arresters (the cu rrent through the other closest arrester was not measured) matches the modeled results. The fact that m odifying the VI-characteristic for one of the two closest arresters not only improves the match of model-predicted and measured currents through that arrester but also improves consider ably the match of modelpredicted and measured current s through the other four arre ster provides support for the hypothesis that the VI-characteristic of the two closest arresters did actually change. Significantly, it was not necessary to modify th e VI-characteristic in the model to achieve good agreement between modeled and measured results for the current division during the 2003 experiment on the vertically-configured lin e. This indicates that if it is true that

PAGE 274

244 the VI-characteristic of the arresters actua lly changed then this change was probably caused by the initial continuous current (ICC) which was flowing through the arresters during the 2000 experiment but not during th e 2003 experiment (the ICC was diverted during the 2002 and 2003 experiment, see Section 3.1.1). N ote that a reduction of the residual voltage for the VI-cha racter istics of the two closest arresters by 20% could be explained if there are 4 or 5 MOV disks in the arresters and one of the disks failed in short circuit mode. Additional discussion regarding this issue can be found in Section 6.4.2. Note that other explanations for the poor model 1 results are possible. For instance, the m easured current division is reasonably well modeled if the m a nufacturer-provided VI-characteristics are used for all arresters in the model and a flas hover across the highresistance section to the low-resistance section of the voltage dividers installed at the two closest arrester stations is simulated by connecting a 50 resistor and 1 F capacitor (the low-voltage section of the divide rs) in parallel to the closest arrester. Interestingly, the adjustment of the VI-characteristic in mode l 2 has essentially no effect on the modeled currents to ground (that is, both model 1 and model 2 ground currents are in good agreement with the measured ground currents). This is probably due to the fact that the modified VI-characteristic ma inly affects the arrester cu rrents during the steady-state mode, and that during this time the ground cu rrent division is determined by the DC grounding resistance of the individual pole groundings, as found for the horizontallyconfigured line by Mata et al. ( 2003) and confirmed with data from the verticallyconfigured line experime nt by Schoene et al. ( 2006b). For the vertically-configured line tested in 2003 the model-predicted currents match the m easured arrester and ground currents very well. The good agreement achieved

PAGE 275

245 without tweaking the mode l (that is, optimizing the match between measured and modeled results by making assumptions regarding corona effect s, lightning channel impedance, etc.) is perhaps surprising cons idering the simplicity of the employed model vs. the complexity of the experiment (c orona effects and the lightning channels characteristic impedance are not taken into account in the model) and the limited/inaccurate information about mode l parameters (the line groundings are represented by a simple model, inaccuracies in the dete rmination of the DC grounding resistance). It has been shown in Section 6.5.3 that the inductance of the distribution line conductors between arrester stations determ ines the duration of the im pulsive current flow through the arrester closest to th e lightning strike poi nt significantly and consequently also the energy absorb ed by this ar rester during the impulsive current. This means that the absorbed energy due to the impul sive current in the closest arrester (and thereby the likelihood of damage to that arrester) can be sign ificantly reduced by minimizing the distance between a rrester stations on the line. 6.6 Estimation of the Arrester-absorbed Energy It was estimated by Mata et al. ( 2003), based on the model 2 lightning return stroke current division determined in the horizontal line experiment and the available s tatistics on the current amplitudes and waveshapes of firs t strokes in natural lightning, that, within about 450 s of the initiation of the first return st roke current flow, the energy input from about half of all natural li ghtning first strokes delivered to each of the two closest arresters exceeds 70 kJ and thus would likely damage them (in the absence of flashovers or other alternative pa ths for the return stroke current to bypass the arrest ers). If model 1 applies to the current division on real world distribution lines, then the arrester damage

PAGE 276

246 rate on real world lines would be smalle r than the arrester damage rate on the horizontal test line estimated by Mata et al. ( 2003) since the low-frequency current components are divi ded evenly am ong multiple a rresters on the line (that is, the arresters further away from the lightni ng strike point absorb a significant portion of the lightning energy and thus help protect the arresters closer to the lightni ng strike point from damage and degradation). For instance, McDermott ( 2006) using the EMTP found a considerably lower arrester-absorbed energy (30 kJ) than Mata et al. (70 kJ) f or a typical natu ral lightning first return stroke cu rrent injected into the phase conductor of a distribution line, which is m ostly attributable to the fact th at McDermott adopted the model 1 and Mata et al. adopted the model 2 current division. It is important to note that for both model 1 and model 2 the MOV block of arre sters on real world lines closest to the lightning strike point may be damaged since they pass most of the impulsive lightning current and therefore absorb most of the energy during the transient mode. We now calculate the energy absorbed in each of the two arresters closest to the strike point during a typical natural lightni ng first return stroke29 using model 1 and lines of different lengths and different number of arrester stations. As noted in Section 6.4.2 the arrester-absorbed energy will be considerably larger if the model 2 arrester current division is adopted. The energy is calculated using the EMTP30 for the following line configurations: (1) the vertical distribution line tested from 2001 through 2003 (4 arrester stations), (2) the vertical line extended to about 1.5 km (8 a rrester stations), and (3) the 29 The current waveshape found in Berger et al. (1975) with a peak valu e of 30 kA, which is the median value. The same current waveshape was used in Mata et al. (2003) to estimate the arrester-absorbed energy based on the horizontal line experiment results. 30 The EMTP model has been verified to some extend in Section 6.5 by successfully modeling the arrester currents on the vertical line.

PAGE 277

247 vertical line extended to about 3 km (16 arrester stations). The distance between arrester stations is the same for all line config urations (that is, 4 spans or 200 m). The manufacturer-provided VI-characteristic of the Ohio-Brass arre ster was used in the model (Section 3.4). The energ y capability of this arr e ster is rated at 40 kJ. The calcu lated energies are displayed in Figure 6-43 for 1 ms. The equilibration time in the figure (the duration of the transient m ode as defined in Section Figure 6-4) was determined for case (a) to b e 130 s. Figure 6-43 : EMTP-calculated absorbed energy in one of the two closest arresters during a typic al natural lightning first return stroke current injected into the phase conductor at midspan. The vertical lin e contained (a) 4, (b) 8, or (c) 16 arrester stations. The transient mode (d ark shaded area) and steady-state mode (light shaded area) determined from the arrester currents for case (a) are indicated. The following information can be gleaned from Figure 6-43: The arrester-absorbed energies for cases (b ) and case (c) are very similar (50 kJ and 45 kJ, respectively). This demonstrates th at the arrester-absorbed energy becomes insensitive to the increase of the number of stations and that c onsequently cases (b) and (c) are good representations of long real world dist ribution lines with a large number of stations. The mode ling results of McDermott ( 2006) conf irm that the arrester-absorbed energy converges with the increasing number of arresters. For cases (b) and (c) almost all of th e total arrester-abs orbed energy during a natural lightning first return stroke with the median peak value found by Berger et al. ( 1975), 30 kA, is abs orbed during the transient m ode. The absorbed energy

PAGE 278

248 during the transient mode, 40 kJ, can be viewed as the minimum arrester-absorbed energy, since this energy will not be reduced by adding additional arrester stations to the line (the energy becomes insensitive to the increase of the number of stations, as noted in the previous item) or by the presence of transforme rs on the line (the transformer current during the transient mode is negligible, as shown in part 1). Therefore, it can be concluded that for a real world distribution line of any length with 4 spans between arrester stations a bout 50% of natural lightning first strokes dissipate at least 40 kJ into the closest ar rester, which is a value identical to the energy capability of the Ohio-Brass arrester used in the model. More energy will be absorbed in the closest arrest ers due to (1) larger first return stroke currents, (2) subsequent stroke currents, (3) continuing cu rrents if model 2 app lies, and (4) strike locations not equidistant between the two arre ster stations (for this case the energy would not divide equally and the closer ar rester would absorb more energy). Less energy will be absorbed for lines with shorter equilibration times, which can be achieved by reducing the line length between arrester stations, as noted above and shown in McDermott (2006). The arrester-absorbed energy during the tran sient mode is similar or the same for all three cases (case (a): 45 kJ, case (b) and case (c): 40 kJ). This indicates that the data obtained from our test distribution line are suitable to estimate the minimum arrester-absorbed energy for the clos est arresters on real world lines. 6.7 Modeling of Induced Currents on the Test Line due to Nearby Lightning The LIOV-EMTP96 code (Section 2.6.3) is one of the most a dvanced tools used for the power distribution line design to predict lightni ng-induced overvoltages on distribution lines that potent ially can cause line flashovers thereby reducing the power quality. In this section the ability of the c ode to a ccurately model induced currents on distribution lines is tested by comparing model-predicted cu rrents to experimentallydetermined currents for one selected str oke, that is, FPL0336-6. Additional comparison and discussion of modeling results and expe rimental data from the 2003 nearby strike experiment can be found in Schoene et al. ( 2003a ) and in P aolone et al. ( 2004b). The measurements on the line during str oke FPL 0336-6, a seven-stroke flash 15 m from the line at pole 4, include ground current s, currents in the termination resistances, line currents in the neutral and the phase conductors, and arrester currents. The LIOVEMTP96 code with the code input being the retu rn stroke current measured at the channel

PAGE 279

249 base approximated by two Heidler functions is used to calculate the model results. The measured grounding resistances used in the model were 23 20 19 18 29 and 24 at poles 1, 2, 6, 10, 14, and 15, respectively (Section 3.3). Ohio-Brass a rresters with the m a nufacturer-provided VI-chara cteristic in response to an 8/20 s current pulse were used in the model (Section 3.4). Discrepancies between m easurement and m odel could be due to inaccuracies in the VI-c haracteris tics of the arresters. The 8 s risetime of the 8/20 s test pulse is much slow er than typical risetimes of lightning return stroke currents and the resultant induced currents. Therefore, the re sponse of the arresters to a fast current pulse may not be adequately m odeled. Additionally, the VI-characteristic of low currents is not well approximated. The modified transmission line model with exponential decay (MTLE, see Section 2.3) with an assum ed return stroke speed of v = 1.3 x 108 m/s and an assumed current decay constant of = 2000 m was employed for the LIOV-EMTP96 calculation here. The ground conductivity in the model was used as an adjustable parameter and assumed to be 1.7 x 10-3 S/m. The conductivity is used in the calculation of the horizontal electric field from the model-derived vert ical electric field and in accounting for propagation effects. Figure 6-44 shows a comparison of the curren ts measured on th e distribu tion line with the model results. The measured and modeled ground and neutral curr ents generally match reasona bly well. The measured and modeled arrester and phase cu rrents generally match reasona bly well for the duration of the first peak. The modeled arrester and pha se currents are typical ly larger than the measured currents for times after the first pe ak. In general, the model works reasonably well for the first microsecond after the beginning of the return stroke and not so well for later times.

PAGE 280

250 Figure 6-44: Comparison of m easured data with model-pr edicted results for stroke FPL0336-6.

PAGE 281

251 Figure 6-44: continued

PAGE 282

252 6.8 Lightning Currents Traversing Soil and Entering Line Grounding The direct lightning current injection through distribution line groundings during a lightning strike in the vicinity of a power line, which contributes significantly to the lightning caused currents and voltages on th e line if the lightning strikes ground some meters from the line grounding, is inves tigated in this section. During the 2003 experiment lightning currents were injected into ground and some fraction of the lightning current entered the line through the pole 15 grounding located 11 m from the lightning current injection poi nt and superimposed on light ning-induced currents on the line. Lightning currents, curre nts injected into the line through the pole 15 grounding, and currents leaving the line through the other groundings are shown and compared in Section 5.3. These data are discu ssed a nd modeled in this section. 6.8.1 Analysis of Lightning Currents Traversing Soil The data presented in Section 5.3 are analyzed and discussed in this section. Table 6-5 shows the lightning current peak values and lightning charges and compares them with the peak values of the currents entering the line through the pole 15 grounding located 11 m from the lightning current injec tion point and the char ges entering the line through the pole 15 grounding. The charge transfer was obtained by numerically integrating the measured return stroke curr ents and the initial c ontinuous currents (ICCs) over a 1 ms time interval and a 10 ms time inte rval, respectively (a longer time interval was chosen to determine the charge transfer during the ICC due to the longer duration of the ICC compared to the return stroke current). The percentage of the return stroke curre nt peaks entering the line through the pole 15 groundings are about 50% smaller than the percentage of the ICC peaks entering the line as seen in Table 6-5 (return stroke events: 7%, initial stage events: between 12% and

PAGE 283

253 Table 6-5: Comparison of peak values and charge transfers of lightning currents and currents entering the line through the pole 15 grounding. The charge transfers were obtained by numerically integra ting the measured currents over 1 ms (return stroke current) and 10 ms (initi al continuous current) time intervals. Peak Current Value [kA] Char g e transfer within 1ms / 10ms [C] Lightning Pole 15 Ground Percenta g e Of Li g htnin g Current Peak Entering Line [%] Lightning Pole 15 Ground Percenta g e Of Li g htnin g Char g e Entering Line [%] FPL0347-1 20.1 1.4 7 1.79 0.30 17 FPL0347-2 6.1 0.5 7 0.41 0.08 19 FPL0350-1 8.4 0.6 7 0.58 0.11 19 FPL0347-ICC 5.2 0.7 14 16.82 2.73 16 FPL0348-ICC 2.9 0.5 17 10.84 2.02 19 FPL0350-ICC 11.1 1.4 12 18.25 3.18 17 17%). This trend is illustrated in Figure 6-45 where the lightning current peaks (x valu es) are plo tted against the peaks of the pole 15 ground currents (y values). The return stroke current peaks and ICC peaks both are very well linearly correlated with the corresponding pole 15 ground current peaks (return stroke events: R2 = 1, initial stage events: R2 = 1, where R2 is the coefficient of determination), but the linear regression equation is quite different for the return stroke and initial stage ev ents (return stroke events: y = 0.07x + 0.02, initia l stage events: y = 0.11x + 0.17). A similar comparison of the charge transfers within 1 ms (return stroke events) and 10 ms (initial stage events) gives a different resultthe fractions of the lightning charge entering the line through the pole 15 grounding are very similar for the return stroke and initial stage events (between 16% and 19%, see Table 6-5). Consequently, only one regression equation is needed to obtain good correlation between the lightning ch arge transfer to ground and the charge transferred into the line through the pole 15 grounding (y = 0.17x + 0.02, R2 = 0.99, see Figure 6-46). Note that the integration time us ed to obtain the charge transfers from the return strok e currents and ICCs was 1 ms and 10 ms, respectively.

PAGE 284

254 Figure 6-45: Peak value of cu rrent injected into the line through the pole 15 grounding as a function of peak value of lightning cu rrent injected into ground 11 m from pole 15. The linear regression equations and R2 values are given. Figure 6-46: Charge injected into the line through the pole 15 grounding as a function of lightning charge injected into ground 11 m from pole 15. The linear regression equation and R2 value are given. The integration time used to obtain the charge transfers from the re turn stroke currents and initial continuous currents was 1 ms and 10 ms, respectively.

PAGE 285

255 Table 6-6 shows the 10-90% risetimes and ha lf-peak widths of the lightning return stroke currents injected into ground and co mpares them with the 10-90% risetimes and half-peak widths of the currents entering the line through the pole 15 grounding. The table includ es the ratio between 10-90% risetimes/half-peak wi dths of the lightning return stroke currents and the currents entering th e line through the pole 15 grounding. The 10-90% risetimes of the pole 15 ground current s are 2.1 to 3.0 times larger than the risetimes of the lightning re turn stroke currents. The ha lf-peak widths of the pole 15 ground currents are 2.5 to 4 times larger than the half-peak widths of the lightning return stroke currents. Table 6-6: Comparison of 10-90% risetimes and half-peak widths of lightning currents and currents entering the lin e through the pole 15 grounding. 10-90% Risetime Of Current [ s] Half-Peak Width Of Current [ s] Lightning Pole 15 Ground Ratio Of Pole 15 Ground Current And Lightning Current Risetimes Lightning Pole 15 Ground Ratio Of Pole 15 Ground Current And Lightning Current HalfPeak Widths FPL0347-1 0.9 2.7 3.0 54 135 2.5 FPL0347-2 1.0 2.6 2.5 17 85 5.0 FPL0350-1 1.4 2.9 2.1 24 98 4.0 6.8.2 Discussion of Lightnin g Currents Traversing Soil It was shown in the previous section that for a lightning strike 11 m from pole 15 a significant fraction of the lightning return str oke current is injected into the neutral conductor of the vertically-configured line th rough the pole 15 grounding. Schoene et al. ( 2003a ) shows that during stroke FP L0347-1 the pha se A arrester clos est to the strike point (the arrester at pole 14) does not conduc t appreciable current while the other phase A arresters (the arresters at pol es 10, 6, and 2) conduct curren ts with peak values between 150 A and 50 A. This result is apparently due to the lightning current directly injected

PAGE 286

256 into the neutral conductor at pole 15, which prevents the po le 14 arrester from opening by reducing the lightning-induced voltage be tween the phase conductors and neutral conductor at pole 14. The inject ed current into the neutra l also reduces the lightninginduced phase-to-neutral voltage at the other arrester poles, but to a lesser degree since a large fraction of the injected current leaves the line th rough the pole 14 grounding. This interpretation of the experimental result imp lies that the fraction of the lightning current injected into the neutra l conductor through the line gr ounding closest to the nearby lightning can reduce the lightni ng-induced voltage between pha se and neutral conductors thereby reducing flashovers due to induced voltages. Realistic models that estimate flashovers on distribution lines due to nearby lightning need to take this effect into account. The following paragraphs discuss th e dependencies of the fraction of the lightning current entering the pole 15 grounding in order to properly characterize this current to implement this effect in to a nearby strike flashover model. The peak values of the fractions of th e lightning currents entering the line through the pole 15 grounding are 7% of the lightning re turn stroke current peak values and, dissimilarly, between 12% and 17% of the ICC peak values ( Table 6-5). This difference is also reflected in the dif ferent regr es sion e quations for the return stroke events and the initial stage events that relate the light ning current peaks to the pole 15 ground current peaks ( Figure 6-45). The percentage value is apparently invers ely dependent to the highfrequency content of the current, that is, the larger the high-frequency content of the curren t the sm aller the p ercentage of the curr ent peak entering the line. For instance, the peak values of the fraction of the lightning cu rrent entering the line ar e 7% of the return stroke current peak values (the fastest curr ents), 12% of the ICC peak value of flash

PAGE 287

257 FPL0350 (the fastest ICC pulse containing the maximum ICC), and 17% of the ICC peak value of flash FPL0348 (the slowest ICC pul se containing the maximum ICC). Also, a comparison of 10-90% risetimes in the previous section showed that th e fronts of the pole 15 ground currents are considerably slower than the fronts of the return stroke currents ( Table 6-6). The frequency dependence of th e pole 15 ground current p eaks for the return stroke and initial stage events and the sl ower fronts of the pole 15 ground currents com pared to the fronts of the retu rn stroke currents indicates that the lightning return stroke current injected into ground has lost high-frequency content when it enters the line at pole 15. Apparently, the ground the lightning current traverses and/or the line system the lightning current enters (the line system being mainly the neutral conductor grounded at the current injection point at pole 15 and at poles 14, 10, 6, 2, and 1) act as low-path filters. The low-path filter behavior affects the faster return stroke currents more than the slower ICCs, which is reflected in the lower fraction of the return stroke current peak entering the line compared to the fracti on of the ICC peak entering the line. On the other hand, the fractions of th e lightning charges entering the line through the pole 15 grounding are between 17% and 19% of the lightning return stroke charges and, similarly, between 16% and 19% of the lightning charges for the initial stage events (note that the different current integration times, 1 ms for the return stroke currents and 10 ms for the ICCs, does not affect the percen tage values). This similarity is also reflected in the fact that for the return stroke events and the initial stage events the same regression equation correlates well the lightning charge and th e charge injected into the line through the pole 15 grounding ( Figure 6-46). The observed similarity of the percen tages of the charge tran sferred during th e fast return stroke events and the slower

PAGE 288

258 initial stage events shows that, in contrast to the lightning currents discussed above, the percentage of the lightning charge entering th e line is not significantly influenced by their frequency content and that therefore the ground and/or the line conductors show purely resistive behavior for the charge transfer Note that the charges were obtained by integrating the lightning currents over a 1 ms (return stroke currents) and 10 ms (ICCs) time intervals and that the integration pro cess accentuates the low-frequency components in the currents. Interestingly, the fraction of the lightning current injected into the line through the pole 15 grounding for stroke FPL0347-1 exhibits a plateau that is not present in the lightning return stroke current for this event and in the ground currents during other return stroke or initial stage events (see Figure 5-5 in Section 5.3). N ote that the return stroke current peak of this stroke (20 kA) was considerably larger than the currents during the other return stroke and initial stage events. The pole 15 ground current peak during this stroke (1.4 kA) was considerably la rger than the pole 15 ground current peaks during the other two retu rn stroke events and similar to the largest pole 15 ground current peaks during the initial stage events. The r eason for the presence of the plateau in the pole 15 ground current of stroke FPL0347-1 is no t known but it can be speculated that the plateau is related to ground ar cing at the lightning current injection point enhancing the non-uniform current distribution in the soil and diverting the pa th of the lightning current away from the pole 15 grounding. A similar analysis was conducted at Camp Blanding by Fernandez et al. ( 1998b) on a 730 m long test distribution line with tw o vertically stacked conductors supported by 15 poles. The bottom conductor simu lated the neut ral and was grounded at poles 1, 9, and

PAGE 289

259 15. The line was terminated at both line e nds in its characteri stic impedance of 500 No arresters or other equipment was installed on the line. It was found that for a 17 kA rocket-triggered lightning st roke at 20 m from the line about 890 A (5% of the lightning return stroke current peak) entered the neutral conductor thro ugh the pole 9 grounding, which was located 40 m from the strike point. The fraction of the total lightning return stroke current peak entering the line through the pole 15 grounding in our experiment was 7% ( Table 6-5), which is similar to the 5% found in the experim ent of Fernandez et al.. This similarity is som ewhat surp rising cons idering that the distan ce between the lightning current injection point and the closest line grounding in the experiment of Fernandez et al. was almost four times larger than the distance between the injection point and the closest grounding in our experiment (40 m in th e experiment of Fernandez et al. vs. 11 m in our experiment). Note that both experiments were conducte d at the same experimental site so variations of the experimental result s due to different soil pr operties are minimal, although some soil variation is expected since the experiment locations were separated by hundreds of meters. The current injected into the pole 9 ground in the experiment of Fernandez et al. exhibited damped oscillation during the first 10 s after the initial crest, which had a 3.2 s period and a maximum peak-to-peak value of 400 A (the peak value of the pole 9 ground current was 890A, as noted above). Fernandez et al. attributed the oscillation to wave reflections at the poles 1, 9, and 15 gr ounds. Interestingly, oscillations were not seen in any of the pole 15 gr ound currents in our experiment (Section 5.3). The low-pass filtering of the lightning return stroke current entering the line grounding observed in our experiment does not appear to be present in the retu rn stroke current entering the line in the experim ent of Fern andez et al.. On the contrary, the current

PAGE 290

260 entering the line grounding of Fernandez et al appears to have a faster front and consequently a higher frequency content than the lightning return stroke current, as if this line grounding shows capacitive behavior while the low-pass filtering of the current entering the line grounding in our experi ment indicates inductive behavior. In a different experiment also conducted at Camp Blanding, 18% of the current from a 20 kA nearby rocket-triggered lightning stroke entered the ground of a simulated house 19 m from the strike point and 10% of the currents from a nearby rocket-triggered lightning two-stroke flash entered the ground of a transformer 60 m fr om the strike point (Fernandez, 1997). N ote that for both events larger percentages of the lightning currents (18% and 10%) entered the line grounding comp ared to the 7% in our ex perim ent even though the distances of the lightning current injection points to the groundings were larger (19 m and 60 m vs. 11 m in our experiment). 6.9 Induced Currents on a Buried Counterpoise The interaction of nearby lightning with a buried counterpoise is analyzed in this section to investigate if lightning can induce currents large enough to ignite explosives stored in explosive storage bunkers that use counterpoises as their grounding systems or to cause damage to electric devices connected to counterpoises. Th e correlation of peak values of induced currents measured on the counterpoise and the peak values of the causative rocket-triggered lightning return stroke currents is investigated. This experiment is introduced in Section 1.9 and the experim ental c onfiguration is described in Section 4.8. An overview of the rocket-triggered and natural lightning events that induced cu rrents m easured on the counterpoise is given in Section 4.9 The experim ental data is presented in Section 5.4.

PAGE 291

261 Figure 6-47: Magnitudes of counterpoise curren t peaks vs. channel base current peaks for the mobile launcher experiment. The linear regression equation and R2 value are given. Distance from the lightning to the north-west corner of the counterpoise is 290 m. Figure 6-48: Magnitudes of counterpoise curren t peaks vs. channel base current peaks for the tower launcher experiment. Th e linear regression equation and R2 value are given. Distance from the lightning to the north-west corner of the counterpoise is 50 m. The induced counterpoise cu rrents are characterized by a few microseconds wide V-shaped initial pulse followe d by a polarity change and a tens of microseconds wide hump of smaller magnitude than the initial pulse. Figure 6-47 and Figure 6-48 show that

PAGE 292

262 the peaks of the initial pulse (y-axis) and the channel base curren t peaks (x-axis) are strongly linearly correlated for both, the mobile launcher experiment and the stationary launcher experiment (R2=0.99, where R2 is the coefficient of determination). Interestingly, the polarity of the counter poise currents induced by rocket-triggered lightning from the mobile launcher and stat ionary launcher are different (only the magnitudes of the counterpoise current peak values are shown in Figure 6-47 and Figure 6-48). The polarities of all co unterpoise current peak values during the m obile launcher experim ent are negative and the polar ities of all counterpoise current peak values during the stationary launcher experiment are positive. The different polarities are apparently related to the different polarities of the lightnings horizontal electric field components at the counterpoise due to the different locations of the mobile and stationary launchers.

PAGE 293

263 CHAPTER 7 CHAPTER 7: SUMMARY OF ORIGINAL RESULTS The origin al, most important results presen ted in this dissertation are summarized here. 7.1 Data Consistency The consistency and hence th e validity of the data coll ected during the 2000, 2002, 2003, and 2004 FPL experiments has been thoroughly tested in this dissertation. The data used for the consistency check are the high and low lightning current injected into one of the line conductors, the sum of curre nts flowing from the struck conductor to neutral, and the sum of currents flowing from the neutral to ground. A match between the lightning current (the current injected into the system) and the sum of ground currents (the current leaving the sy stem) gives high confidence in the validity of the data. Discrepancies between the lightning current and the sum of struck-phasetoneutral currents can be due to flashovers that cau sed currents to bypass the sensors or to unmeasured currents through failed arresters. Typically, the sum of the ground currents exhibited an overshoot that appears to be due to reflections from impedance discontinuities on the line ( 6.1.1). Mata ( 2003) previously observed this overshoot in a com parison of the lightning currents of 4 st rokes measured during the 2002 experime nt with the total ground currents a nd concluded that there is no balance between the current entering the system and the sum of current s leaving the system (ground currents), the peak value of the sum of ground currents being greater than th e peak value of the incident current (p. 83).

PAGE 294

264 7.1.1 Data Consistency during 2000 Ho rizontal Line Experiment The lightning currents and sum of ground currents measured during the 2000 experiment do not always match. Reasons for the discrepancies are discussed in Section 6.1.2. Evide nce is presented in Section 6.1.4 that during the 2000 FPL experiment a measurement erro r cau sed a 25% overestima tion of all measured high lightning currents and a 50% overestimation of most low lightning currents. The only low lightning currents that were measured correctly were the currents during the two strokes of flash FPL0037. An adjustment factor of 0.75 has been applied to all high lightning currents measured during the 2000 experiment presented and analyzed in this dissertation. The high lightni ng current, with the adjustment factor applied, matches the sum of ground currents reasonably well on a 100 s time scale (Section 6.1.1 and Appendix B ). N ote that Mata ( 2000 ) and Mata et al. ( 2003) previously observed a 20% to 30% discrepancy between the to tal charge injected into the system (the integrated high lightning current) and the total charge transferred to ground (the integrated sum of ground currents). Mata ( 2000) stated th at this diffe rence m i ght be the result of limited resolution of the measuring systems, calibration errors ( overestimation of the incident current and/or underestimation of currents to ground), and appreciable capacitive and resistive leakage currents (p. 137). Mata et al. ( 2003) did not rule out undetected flashovers as the reason for the differences. As stated above, it is now believed based, on the discussion in Section 6.1.4, that th e differences are due to a meas urement error. No adjustm ent factor correcting the alleged measurement error was applied to the lightning currents from the 2000 experiment presented and analyzed in Mata ( 2000), Mata et al. ( 2003), and A.G. Mata ( 2002).

PAGE 295

265 7.1.2 Data Consistency during 2002, 2003, and 2004 Vertical Line Experiments Generally, the lightning currents and the sum of ground currents measured during the 2002, 2003, and 2004 experiment are in good agreement (Section 6.1.1). The exceptions are: Inconsistencies of the lightning curre nts and sum of ground currents measured during the first triggering day of the 2002 season (the currents of flash FPL0208 and FPL0210 on 7/9/02). An unusual waveshap e of the lightning current indicates an erroneous lightning current measurement during that day ( Appendix B and Appendix C ). All return stroke currents measured during the first triggering day of the 2003 season (flash FPL0301 on 6/30/06) exhibit ringing of a few tens of kiloamperes magnitude that starts before the return stroke initiation and lasts for a few microseconds. The ringing also appear s with much lower magnitude on the lightning currents measured during str okes 1, 3, 6, and 10 of flash FPL0312 ( Appendix B and Appendix C ). The ground current measured at pole 8 during stroke FPL0403-2 of the 2004 season exhibits a spike that has a peak va lue of roughly 6 kA and a width of 0.5 s. Apparently, this spike was caused by electromagnetic coupling to the instrumentation in the unlatched Hoffma n box. Spikes with similar waveshape and magnitude appear in the Point A and Point B current measurements of the 2004 test house experiment (Section 6.1.5). The lightning currents and the sum of ground currents measured during the 2002 and 2003 experime nts do not always m atch. R easons for the discrepancies are discussed in Section 6.1.2. A com parison of the high and l ow lightning currents in Section 6.1.3 shows that erroneous hum ps frequently appear in the high and /or low lig h tning currents measured at the tower, lightning intercepti ng structure, and mob ile launcher during the 2002, 2003, and 2004 experiment. These humps can have either positive or negative polarity, their peak values range from a fe w hundreds of amperes to a few kiloamperes, and they have widths of a few hundreds of microseconds.

PAGE 296

266 7.2 Characterization of the Ligh tning Return Stroke Current A statistical analysis of the lightning re turn stroke currents measured during the 1999 through 2004 FPL experiment has been conducted in Section 6.2. The accuracy of the analyzed data from the 2000, 2002, 2003, and 2004 experiments has been tested in Section 6.1 by checking the consistency of the re turn stroke cu rrents w ith the total phaseto-neutral currents and the total currents flowing to ground and/or by comparing the high return stroke currents with the low currents. Inconsistent data have been excluded from the statistical analysis. The return stroke currents measured during the 2000 experiment have been analyzed previously in Mata ( 2000). This an alysis has been revised in this dissertation by excluding return stroke current risetime s determ ined from currents that exhibited ringing during the current front and by introducing an adjustment factor of 0.75 for all return stroke currents measured in 2000. Statistical distributions of the return stroke current peak, 10-90% risetimes, halfpeak widths, and charge transfer during th e first millisecond after the return stroke initiation have been presented in Section 6.2.2. The statistical data have been com pared to previous analys is o f rocket-triggered lightning return stroke characteristics. This comparison can be summarized as follows: The statistical characterization of the return stroke current peaks determined in this dissertation is in general consistent with statistical characterizations from other studies. In particular, the arithmetic mean of the return stroke current peaks found here (14.0 kA for 144 samples) is very si milar to the arithmetic mean found in the study with the largest sample size (that is, the arithmetic mean of the peak values of 305 currents measured at the Kennedy Space Center from 1985 to 1991 was 14.3 kA; Depasse, 1994). D iscrepancies between the return stroke peak current statistics presented in this dissertation and in othe r publication m a y be attributable to (1) small sample size in the other studies (2) differences in the triggering thresholds between the experiments discussed here and experiments discussed in other publications, and/or (3) erroneous data an alyzed in the other studies. A dependency of the return stroke curren t peak and the electrical prope rties of the strike object was not observed.

PAGE 297

267 The statistical distribution of the 10-90% rise times of return stroke currents directly injected into the line is different than the distribution of the 10-90% risetimes of return stroke currents during the nearby st rike experiment. This discrepancy is likely due to the different strike object and grounding system, that is, the generally larger risetimes observed duri ng the direct strike experime nt are apparently due to the inductance of the struck line conductor and/or re flections of the lightning current at impedance discon tinuities on the line such as the line arresters. A comparison of the arithmetic/geometric means of the 10-90% risetimes from the direct strike experiment with the mean s from four other non-FPL experiments shows that the means from the other experiments were smaller. The arithmetic/geometric mean of the 10-90% risetimes from the FPL nearby strike experiment is among the lowest means. A dependency of the return stroke current half-peak width on the electrical properties of the strike object was not observed for the direct and nearby strike experiments. However, the difference betw een mean values determined during the FPL experiments and in other studies s uggests that such a dependency exists. It was determined that the charge transferred during the first millisecond after the return stroke initiation and the return str oke current peaks are correlated by a power regression equation (y = 12.3x0.54, where y is the peak current and x is the charge transfer, R2 = 0.76, where R2 is the coefficient of determination). 7.3 Arrester Disconnector Operation and Arrester Energy Absorption A performance assessment regarding the frequency of disconnector operation on the distribution lines tested from 200 0 through 2004 is conducted in Section 6.3. Also, the minim u m arrester-absorbed energy for natu ral lightning first strokes to real world distribution lines was estimated in Section 6.4.2 based on the experim entally-determined arrester current division on th e test distribution lines dur ing the trans i ent mode. The results can be summarized as follows: The disconnector operation common during the 2001/2002 ve rtical line experiment was absent during the 2003 ve rtical line experiment, pr obably due to a transformer on the line which protected the arrester s by shunting the low-frequency current components to ground. Disconnector operation during the 2000 horizontal line experiment was considerably less frequent than during the 2001/2002 vertical line experiment, which was possibly due to the larger nu mber of arrester stations on the 2000 horizontal line reducing the long-duration current through each individu al arrester.

PAGE 298

268 Typically, the disconnectors of arresters closest to the lightning current injection point, which conduct the bulk of the lightni ng return stroke tr ansients, operated. The results summarized in the previous three items indicate that the combined energy input of the return stroke transients and long-dur ation currents in triggered lightning is sufficient to activate disc onnectors. However, the absence of disconnector operation in 2003 where the l ong-duration current was considerably reduced by a transformer indicates that th e return stroke current transients in rocket-triggered lightning alone do not commonly activate disconnectors. No statistically significan t experimental evidence was found that employing two arresters in parallel instead of si ngle arresters reduces the likelihood of disconnector operation. No statistically significant experimental evidence was found that not injecting the initial continuous curre nt of triggered lightning into th e distribution line reduces the likelihood of disconnector operation. The absence of disconnector operation dur ing the 2004 experiment is apparently due to injecting the light ning current into the overhead ground wire, which provided the lightning current wi th a direct path to ground. The minimum arrester-absorbed energy dur ing the transient mode for natural lightning first strokes to r eal world distribution lines with a large number of arrester stations; each separated by 4 spans, was estimated. Based on the assumptions given in the previous section, at least 40 kJ of energy is absorbed in each of the two arresters clos est to the strike point for 50% of all natural lightning first strokes to the line. This estimate doe s not take energy absorbed in the closest arrester during the steady-state mode and during subsequent stroke currents into account. Also, this energy is larger for a lightning strike point not equidistant to two arrester stations and smaller if th e inductance of the line segment separating arrester stations is reduced by, for inst ance, reducing the number of spans between stations. 7.4 Flashover Occurrence A performance assessment regarding the frequency of flashovers on the distribution lines tested from 2000 through 2004 is conducted in Section 6.3. The results can be summarized as follows: Theory shows that the voltage between the struck phase and the next closest phase at the weakest point on the vertical lin e tested in 2001, 2002, and 2003 (arresters every 4 spans) is about 40% smaller than the voltage at the weakest point on the horizontal line (arresters ev ery 3 spans) due to the diffe rent arrester spacing. The smaller number of flashovers on the horizon tal line compared to the number of

PAGE 299

269 flashovers on the vertical line appears to reflect this difference in voltage although other explanations for the different flas hover behavior of the lines are possible. The tendency of the 2001, 2002, and 2003 vertical line configuration to experience frequent flashovers was appa rently neither influenced by the number of operated disconnectors nor by the presence of a transformer on the line. The vertical line with overhead ground wire tested in 2004 was the only line configuration which did not experience flashovers. 7.5 Phase-to-neutral Current Division The phase-to-neutral current division during the 2000 horizontal line experiment previously studied in Mata (2000), A.G. Mata et al. (2002), and Mata et al. ( 2003) is reexamined in Section 6.4 and Section 6.5 and compared to the phase-to -neutral current division on the 2002/2003 verti cal line experiment. The resu lts can be summarized as f o llows: During the 2002/2003 vertical lin e experiment the primary pa th of the return stroke current for the first tens of microseconds was through the two arresters closest to the lightning current injection point. This finding is consistent with the arrester current division for the first tens of microseconds during th e horizontal line experiment discussed in Mata et al. ( 2003) and with the m odeled results presented here using model 1 (that is the arres t ers are modeled with the manufacturerprovided VI-characteristic). During the 2002/2003 vertical line experiment (no initia l continuous current) the lightning current was evenly distributed among all arrester s after an equilibration time that ranged from 18 s to 74 s. This finding is in contrast with the arrester current division during the horizontal line experiment (ICC was injected into the line) where after a few hundreds of micros econds the two closest arresters still passed the bulk of the lightning current (Mata et al., 2003). The m odel 1 arrester currents match well the currents m easur ed during the 2002/2003 experim ent. The published arresters VI-characteristic wa s used in model 1 for all arresters. The model-predicted arrester currents do not match all arrester currents measured during the 2000 horizontal line experiment if the published VI-characteristic is used for all arresters in the model, that is, mode l 1. The model-predicted arrester currents match well all arrester currents measured on the horizontal line if a modified VIcharacteristic is used for the two arresters closest to the lightning current injection point (that is, model 2). The findings summarized in the previous two items of this summary suggest that the current divisi on on the horizontal li ne during the steadystate mode is caused by the large energies absorbed in the two arresters closest to

PAGE 300

270 the lightning current injection point that resulted in a reduction of their residual voltages. Other explanations for the current division on the horizontal line are that the division is caused by unmatched arresters installed on the horizontal line or is essentially an artifact caused by the presen ce of voltage dividers in the horizontal line experiment. The ringing present in the st ruck-phase arrester curre nts on the 2000 horizontal line modeled in Mata ( 2000) and A.G. Mata et al. ( 2002) when the lightning was represented in the model as an ide al cu rrent source was not present in m odeling results with ideal current sources for th e struck-phasetoneutral currents on the horizontal and vertical li nes presented in Section 6.5. It has been shown in Section 6.5.4 that the inductance of the distribution line conductors between arrester stations de termines the duration of the impulsive current flow through the arrester closest to the lightning strike point significantly and consequently also the energy ab sorbed by this arrester during the impulsive current. This means that the absorbed en ergy due to the impulsive current in the closest arrester (and thereby the likelihood of damage to that arrester) can be significantly reduced by minimizing the dist ance between arrester stations on the line. 7.6 Neutral-to-ground Current Division The neutral-to-ground current division dur ing the 2000 horizontal line experiment previously studied in Mata (2000) and Mata et al. ( 2003) is re-exami ned in Section 6.4 and Section 6.5 and compared to the neutral-to -ground current division on the 2002/2003 vertical line experiment. The result s can be summarized as follows: During the 2002/2003 vertical lin e experiment, the primary pa th of the return stroke current to ground for the first tens of microseconds was through the two grounds closest to the lightning current injection point. This findi ng is consistent with the ground current division for the first tens of microseconds duri ng the horizontal line experiment discussed in Mata et al. ( 2003). The equilib ration times for the ground currents for both the horizontal and vert ical line experiment ranged from 15 s to 67 s. It appears that the ch arge transfer to ground within 2 ms is inversely proportional to the grounding resistance in the vertical line experiments. This trend was previously found for ground currents after 25 s, or so, and for the charge transfer within 100 s, 500 s, and 1 ms for the horizontal line e xperiment, as reported in Mata et al. ( 2003). Generally, model-predicte d ground currents for the horiz ontallyand verticallyconfigured lines presented in Section 6.5 match the measured ground currents very

PAGE 301

271 well. An adjustment of the measured lo w-frequency, low-curr ent ground resistance, which was done in Mata ( 2000) was not necessary to achieve good modeling results. The modeled ground currents are in sensitive to whether model 1 or model 2 is used. 7.7 Lightning Currents Traversing Soil and Entering Vertical Line Grounding Currents from lightning stri kes in the vicinity of the vertically-configured distribution line tested in 2003 that are traversing the soil and are injected into the distribution line groundings are investigated in Section 5.3 and Section 6.8. The results can be summ arized as follows: The fraction of the lightning current direc tly injected into the neutral conductor through the line grounding cl osest to a nearby lightni ng strike can reduce the lightning-induced voltage between phase conductors and neutral conductor and thereby help to prevent flashovers due to induced voltages. Real istic models that estimate flashovers on distribu tion lines due to nearby lightning need to take this effect into account. The peak values of the lightning curren t/charge injected into ground 11 m from pole 15 and the peak values of the current/charge injected into the li ne through the pole 15 grounding are strongly linear correlated. The peak values of the return stroke currents entering the li ne through the pole 15 grounding are 7% of the peak values of the return stroke curren ts injected into ground 11 m from pole 15. The peak values of the ICCs entering the line through the pole 15 grounding are between 12% and 17% of the peak values of the ICCs injected into ground 11 m from pole 15. The percentage values are apparently inversely dependent on the high-frequency content of the current, that is, the fraction of the fast return stroke currents (large high-frequenc y content) injected into the line is smaller than the fraction of the slow ICCs (small high-frequency content). The charge entering the line through the pole 15 grounding during the return strokes and the initial stages are between 16% and 19% of the peak values of the lightning return stroke and initial stage charge injected into ground 11 m from pole 15. The similarity of the pe rcentages shows that the frac tion of the lightning charge entering the line is not influenced signif icantly by their fre quency content (the return stroke charge has a larger high -frequency content than the initial stage charge) and that therefore the ground and/or the line system show purely resistive behavior for the charge transfer. The 10-90% risetimes of the return stroke currents entering the line through the pole 15 grounding are 2 to 3 times larger th an the risetimes of the return stroke

PAGE 302

272 currents injected into ground 11 m from pole 15. Apparently, the ground the lightning current traverses a nd/or the line grounding the li ghtning current enters act as low-pass filters. The half-peak widths of the return stroke curr ents entering the line through the pole 15 grounding are 2.5 to 5 times larger than th e half-peak widths of the return stroke currents injected into ground 11 m from pole 15. 7.8 Induced Currents on Distribution Lines The LIOV-EMTP96 code (Section 2.6.3) has been employed in Section 6.7 to predict lightning-induced current s on the ver tica lly-configured test distribution line. The model-predicted currents are compared to experimentally-determined currents. The results can be summarized as follows: The measured and modeled ground and neutral currents generally match reasonably well. The measured and modeled arrester and phase currents generally match reasonably well for the duration of the first peak. The modeled arrester and phase currents are typically larger than the measured cu rrents for times after the first peak. In general, the model works reasonably well for the first microsecond after the beginning of the return stroke and not so well for later times. 7.9 Induced Currents on a Buried Counterpoise and Vertical Wire The nearby rocket-triggered and natural light ning strike interaction with a test runway counterpoise and a grounded vertical wire of 7 m height is investigated in Section 5.4 and Section 6.9. The results can b e summarized as follows: The peak values of lightning return stroke currents and the peak values of currents in the buried counterpoise are strongly linear correla ted at distances of both 50 m and 290 m from the lightning strike point. The largest current induced in the counterpoise from rocket-triggered lightni ng at a distance of 290 m was 70 A. The peak current associated with this event was 26 kA. The largest current induced in the buried counterpoise from rocket-trigg ered lightning at a distance of 50 m was 160 A. The peak current associated with this event was 15 kA. The induced counterpoise cu rrents are characterized by a few microseconds wide V-shaped initial pulse followed by a polar ity change and a tens of microseconds wide hump of smaller magnit ude than the initial pulse.

PAGE 303

273 The vertical wire functions as a dE/dt an tenna during the rocket -triggered lightning strikes. The largest induced current in the grounded vertical wire associated with natural lightning striking ground a bout 300 m away was 140 A. This current was likely associated with an upward-directed unconnect ed leader generated in response to an overhead downward-propagating stepped leader step. In fact, a number of current pulses in the wire occurred, likely associat ed with leader steps occurring during the 100 s before the return stroke connected to ground. It is difficult to positively associate the current spikes of the upwardconnected leaders with dE/dt spikes of the downward-directed leader steps due to difficulties in properly aligning the measured current spikes and dE/dt spikes in time although the spikes measured in the vertical wire presumably caused by step leader steps correlate reasonably with signals induced in the counterpoise.

PAGE 304

274 CHAPTER 8 CHAPTER 8: RECOMMENDATIONS FO R FUTURE RESEARCH The previous chapter gave a summary of the numerous original and important results presented in this dissertation and illust rated the large scope of the subjects treated here. One consequence of the large scope is that not all subjects could be treated in sufficient depths. What follows is an overview of the problems that were not fully solved in this dissertation and the consequent quest ions that were raised. Hopefully, future research will shed additional light on these subjects. Arrester disconnector operation during the 2000, 2001, and 2002 FPL experiments: Frequent arrester disconnector operation was observed during the 2000 through 2002 experiments. Unfortunate ly, it was not determined if the arresters MOV discs were damaged or if the disconnectors on healthy arresters operated (a point against the use of disconne ctors can be made if the latter was the case). It is important to clarify this issue with additional rocket-triggered lightning experiments or high-voltage laboratory experiments. Effect of the transformer during the 2003 FPL experiments: More work has to be done on how transformers on distribut ion lines affect the lightning current division. Apparently, the transformer inst alled on the vertical line tested in 2003 prevented disconnector operation by conducting the low-frequency lightning currents. However, resistive loads of 4 and 6 Ohms were connected to the transformers secondary, which should have caused a relatively large resistance on the transformers primary and should have pr evented any significant current flow in the primary. The currents measured in the primary show that this was not the case and the transformer was conducting current s of hundreds of amperes after a few hundreds of microseconds. EMTP modeling re sults not presented here show that the measured results are reproduced if the transformers secondary is shortcircuited. It is possible th at a protective spark gap on the secondary indeed cause a short-circuit connection of the secondar y. However, additional experiments are needed to clarify this issue. Arrester current division dur ing the 2000 FPL experiment: It is not clear why during the 2000 horizontal line experiment the arresters clos est to the lightning current injection point co nducted the bulk of the li ghtning current during the steady-state mode. The likely explanations are that the closest arresters passed more

PAGE 305

275 current during the steady-state mode becau se of (1) a reduction of the residual voltage of the closest arresters due to excessive energy absorption or (2) the mismatch of the closest arresters and the remote arresters, which were from different manufacturers. Both possibilities ha ve, if true, important implications for the design of the lightning protection of distribution lines and need to be investigated further. Arrester energy absorption: Modeling results with distribution lines of different lengths, different number of arrester st ations, and the same number of spans between arrester stations are presented in Section 6.6. The modeling results show that the energy absorbed during th e impulsi ve lightning currents in the two closes t arresters on the line that re sem ble the vertically-configured line tested from 2001 through 2002 (812 m length and 4 arrester st ations) is similar to the arresterabsorbed energy in arresters installed on longer, more real istic lines with a larger number of arrester stations. However, the arrester-absorbed energy during both the impulsive and long-duration lightning currents is considerably larger on the line with 4 arrester stations than on the lines with 8 and 16 arrester stations. Consequently, the vertical test line was too short to be an accura te representation of a long real world distributi on line with many arrester st ations for the purpose of collecting experimental data on arrester damage and disconnector operation due to excessive energy absorption. Modeli ng results presented in Section 6.6 sh ow that in order to conduct a mo re realistic exp eriment to investigate effects related to arrester energy absorption, the test di stribution line has to be extended to a length of about 1.5 km so it can accommodated 8 arrest er stations separated by 4 spans. Flashover occurrence during the FPL experiments: We found some indication that distribution lines with arrester stations every 3 spans exhi bit considerably less phase-to-phase flashovers than arrester stations every 4 sp ans. This possibility has, if true, important implications for th e design of the lightning protection of distribution lines and needs to be confirmed with additional experiments. Currents from nearby lightning strikes entering line groundings: It was determined experimentally that a signif icant fraction of currents from lightning strikes close to a distribu tion line grounding enters the grounding. The current entering the grounding can reduce the light ning-induced voltage between phase conductors and neutral conduc tor and thereby help to prevent flashovers due to induced voltages. Realistic models that ta ke this effect into account should be developed. Direct strikes to distribution lines: The distribution line tested in 2004 (the vertically-configured line w ith overhead ground wire) was the line that showed the best lightning performan ce (neither disconnector ope ration nor flashovers were detected). However, it was also the line that was, compared to the lines tested from 1999 through 2003, the most expensive to build, since arrester stations were installed and the overhead ground wire wa s grounded at every pole. For the 2004 studies, we had planned thr ee experiments of which only the first was completed due to the several hurrican es and the time needed to reconfigure the line. The

PAGE 306

276 additional two experiments on the distribution line i nvolved two modified less expensive configurations: (1) The same configuration as the original tested configuration with the arresters removed (t he original configur ation had 4 arrester stations) and (2) the same configuration as the original tested configuration with the overhead ground wire grounded only at the ar rester stations (for the original configuration the ground wire was gr ounded at each of the 15 poles). The additional experiments need to be performed to determine an optimum costeffective distribution line with sufficient li ghtning protection. Additionally, further testing of the 2004 line configur ation is needed to expose this line configuration to several currents larger than the maximum 16 kA observed to date. Induced currents on a buried counterpoise and vertical wire: The experimentally-determined induced current s on a buried counter poise and vertical wired need to be modeled. Dr. Carlo Al berto Nucci and Dr. Mario Paolone of the University of Bologna, Italy, attempted to model the induced currents in the buried counterpoise with the LIOV-EMTP code. Compared to the measured currents presented here, the model results have similar waveshapes but are of considerably larger magnitude. The likely reason for the mismatch is that the bare buried counterpoise was represented in the model as an insulated wi re. According to a personal communicatio n with Dr. Mario Paolone modeling the case of a buried bare wire is considerably more difficult. However, modeling the case of a bare buried wire is needed to potentially achieve accurate modeling results.

PAGE 307

277 APPENDIX A APPENDIX A: MEASUREMENT SETTINGS The measurement settings for the 20 02 nearby strike experiment, the 2003 direct and nearby strike experiments, and the 2004 dire ct strike experiment are listed in this appendix. The measurement settings for the 1999 and 2000 direct strike experiments are found in Mata ( 2000) an d the measurement settings for the 2001 and 2002 direct strike experiments are found in Mata ( 2003 ). The reference to th e measurement settings for each flash in the 1999, 2000, 2001, 2002, 2003, and 2004 experim e nts are listed in Table 4-1, Table 4-2, Table 4-3, Table 4-4, Table 4-5, and Table 4-6, respectively. The tables in this appendix contain infor mation about the measurement settings for each param e ter including the oscilloscopes ID, channel, and sampling rate, the fiber optic cables ID and length, the colors of the two fibers used for each measurement, the fiber optic link delay, the current sensors ID, model, and V/A rating, the PICs ID and attenuation setting, the RF PICs ID, the fiber optic transmitters/receivers ID and setting, and the measured peak-to-peak va lue of the one-volt square wave used to calibrate the fiber optic link. A calibration si gnal was typically recorded before the first triggering attempt of the day and after the last triggering attempt of the day. The peak-topeak value of the calibration signal was obtained from the latter cal ibration record and from the former if a calibration record after the last triggering attempt was not available. The range for each parameter was calculated from Equation A-1: 2010 2AS CF R A-1

PAGE 308

278 where F is the Isobe setting (0 .1 V, 1 V, or 10 V), C is the measured peak-to-peak value of the calibration square wave, S is the current sensors V/A rating, and A is the PICs attenuation setting. Each measurement was terminated in 50 which reduced the signal by a factor of two (hence the factor two in Equation A-1). A.1 2002 Measurement Settings, Nearby Strikes The measurement settings for the strikes to ground 100 m north of pole 7 are found in Table A-1 (FPL0236 on 8/18/2002) and the se ttings for the strikes to ground 30 m north of pole 7 are found in Table A-2 (FPL0240 on 8/27/2002) and Table A-3 (FPL0244FPL0246 on 9/13/2002). Note that flash FPL0241 (a triggered lightning without return strokes) did not trigger the os cilloscopes and consequently data for this flash are no t available. A lso, note that the pe ak-topeak value of th e m easured calibration signals for parameters IG2, IG15, IAN2, and IAN6 in Table A-1, for parameters IG15, IN3, and IC6 in Table A-2, and for parameters IG15 in Table A-3 are off. The calibration signals for parameters IB6 in Table A-2 and IN3 in Table A-3 were saturate d and consequently the range for these parameters could not be accura tely calculated and is not included in the table s.

PAGE 309

279 Measured Parameter LeCroy Oscilloscope Yokogawa Oscilloscope Fiber Optic Cable Sensor PIC RF PIC ID Isobe Cal. Range [A/V] ID Ch. Sampling Rate [MHz] ID Ch. Sampling Rate [MHz] ID Color low / high Length [m] Delay [ s] ID Model V/A ID Att. [dB] Trans. ID Rec. ID Setting [V] IAN14 6 1 20 18 13 1 p14 y/r 567 2.90 110A#5 110A 0.1 39 26 80 R6B R6B 1 0.99 394 Ibucket, low 6 4 20 18 4 1 bt2 o/b ? ? MS#10 R-2800-4 0.0025 35 14 91 10 10 1 0.96 3861 Ibucket, high 11 1 20 18 3 1 bt1 o/b ? ? MS#10 R-2800-4 0.0025 3F 34 91 25 25 1 0.96 38611 IG1 11 2 20 18 5 1 p1b o/b 565 2.89 MS#3 R-7000-10 0.001 4 13 87 11 11 0.1 0.99 883 IG2 11 3 20 18 6 1 p2 o/b 435 2.24 MS#6 R-5600-8 0.00125 24 13 86 R2A R2A 0.1 0.42 302 IG6 11 4 20 18 7 1 p6c o/b 245 1.29 MS#7 R-5600-8 0.00125 16 27 94 12 12 0.1 0.97 3471 IG10 12 1 20 18 8 1 p10 o/b 375 1.94 MS#2 R-7000-10 0.001 3A 9 82 18 18 0.1 0.96 543 IG14 12 2 20 18 9 1 p14 o/b 567 2.90 MS#5 R-5600-8 0.00125 1 13 80 14 14 0.1 0.96 688 IG15 12 3 20 18 10 1 p15 o/b 595 3.04 MS#4 R-7000-10 0.001 15 13 80 13 13 0.1 0.84 754 IBN1 12 4 20 18 11 1 p1a o/b 542.5 2.78 110A#1 110A 0.1 2F 13 87 R8B R8B 1 0.96 85 IAN1 13 1 20 18 12 1 p1a y/r 542.5 2.78 110A#6 110A 0.1 2B 13 87 R1C R1C 1 0.96 86 ICN1 13 2 20 18 13 1 p1a g/br 542.5 2.78 110A#7 110A 0.1 6 9 87 3 3 1 0.97 55 IAN2 13 3 20 18 14 1 p2 g/br 435 2.24 110A#4 110A 0.1 38 23 86 R1B R1B 1 0.65 184 IN3 13 4 20 18 15 1 p3a o/b 558 2.86 6801#1 3025C 0.025 10 14 88 R7A R7A 1 0.96 383 IAN6 14 1 20 7 15 1 p6b y/r 397 2.05 6801#3 3025C 0.025 8 20 94 R2B R2B 1 0.87 695 IBN6 14 2 20 7 1 1 p6b g/br 397 2.05 110A#3 110A 0.1 3C 33 94 R3B R3B 1 0.96 854 ICN6 14 3 20 7 2 1 p6b o/b 397 2.05 110A#2 110A 0.1 41 30 94 R4B R4B 1 0.98 620 IA6 14 4 20 7 3 1 p6a y/r 267 1.40 6801#2 3025C 0.025 19 20 94 R3C R3C 1 0.96 770 IB6 15 1 20 7 4 1 p6a g/br 267 1.40 5179#3 3525 0.1 3 26 94 R4C R4C 1 0.95 379 IC6 15 2 20 7 5 1 p6a o/b 267 1.40 5179#2 3525 0.1 42 26 94 R5C R5C 1 0.96 381 IN6 15 3 20 7 6 1 p6c y/r 245 1.29 5179#6 3525 0.1 2D 27 94 R8A R8A 1 0.96 428 IA7 15 4 20 7 7 1 p7 o/b 233 1.23 6801#4 3025C 0.025 3E 20 81 R6C R6C 1 0.99 795 IB7 16 1 20 7 9 1 p7 y/r 233 1.23 5179#5 3525 0.1 12 26 81 R7C R7C 1 0.99 394 IC7 16 2 20 7 10 1 p7 g/br 233 1.23 5179#4 3525 0.1 40 26 81 R8C R8C 1 0.94 377 IAN10 16 3 20 7 11 1 p10 y/r 375 1.94 6801#5 3025C 0.025 2C 6 82 R5B R5B 1 0.96 154 IN10 16 4 20 7 12 1 p10 g/br 375 1.94 6801#6 3025C 0.025 2E 17 82 R7B R7B 1 0.96 541 Ibucket, ac ,low 18 16 1 bt2 o/b ? ? MS#10 R-2800-4 0.0025 35 14 91 10 10 1 0.96 3861 Table A-1: 2002 measurement settings for flash FPL0236 (8/18/2002).

PAGE 310

280 Measured Parameter LeCroy Oscilloscope Yokogawa Oscilloscope Fiber Optic Cable Sensor PIC RF PIC ID Isobe Cal. Range [A/V] ID Ch. Sampling Rate [MHz] ID Ch. Sampling Rate [MHz] ID Color low / high Length [m] Delay [ s] ID Model V/A ID Att. [dB] Trans. ID Rec. ID Setting [V] IAN14 6 1 20 7 13 1 p14 y/r 567 2.90 110A#5 110A 0.1 39 19 80 R6B R6B 1 0.99 177 Ibucket, low 6 4 20 18 4 1 bt2 o/b ? ? MS#10 R-2800-4 0.0025 35 14 91 2C 10 1 0.97 3889 Ibucket, high 11 1 20 18 3 1 bt1 o/b ? ? MS#10 R-2800-4 0.0025 3F 34 91 25 25 1 0.96 38491 IG1 11 2 20 18 5 1 p1b o/b 565 2.89 MS#3 R-7000-10 0.001 4 6 87 11 11 0.1 0.98 389 IG2 11 3 20 18 6 1 p2 o/b 435 2.24 MS#6 R-5600-8 0.00125 24 14 86 R2A R2A 0.1 0.99 794 IG6 11 4 20 18 7 1 p6c o/b 245 1.29 MS#7 R-5600-8 0.00125 16 23 94 12 12 0.1 0.99 2226 IG10 12 1 20 18 8 1 p10 o/b 375 1.94 MS#2 R-7000-10 0.001 3A 16 82 18 18 0.1 0.96 1211 IG14 12 2 20 18 9 1 p14 o/b 567 2.90 MS#5 R-5600-8 0.00125 1 9 80 14 14 0.1 0.97 435 IG15 12 3 20 18 10 1 p15 o/b 595 3.04 MS#4 R-7000-10 0.001 15 9 80 13 13 0.1 0.70 392 IBN1 12 4 20 18 11 1 p1a o/b 542.5 2.78 110A#1 110A 0.1 2F 10 87 R8B R8B 1 0.96 61 IAN1 13 1 20 18 12 1 p1a y/r 542.5 2.78 110A#6 110A 0.1 2B 10 87 R1C R1C 1 0.96 61 ICN1 13 2 20 18 13 1 p1a g/br 542.5 2.78 110A#7 110A 0.1 6 10 87 3 3 1 0.97 61 IAN2 13 3 20 18 14 1 p2 g/br 435 2.24 110A#4 110A 0.1 38 20 86 R1B R1B 1 1.00 200 IN3 13 4 20 18 15 1 p3a o/b 558 2.86 6801#1 3025C 0.025 10 10 88 R7A R7A 1 0.77 195 IAN6 14 1 20 7 15 1 p6b y/r 397 2.05 6801#3 3025C 0.025 8 16 94 R2B R2B 1 0.97 490 IBN6 14 2 20 7 1 1 p6b g/br 397 2.05 110A#3 110A 0.1 3C 29 94 R3B R3B 1 0.97 544 ICN6 14 3 20 7 2 1 p6b o/b 397 2.05 110A#2 110A 0.1 41 29 94 R4B R4B 1 0.96 543 IA6 14 4 20 7 3 1 p6a y/r 267 1.40 6801#2 3025C 0.025 19 17 94 R3C R3C 1 0.97 549 IB6 15 1 20 7 4 1 p6a g/br 267 1.40 5179#3 3525 0.1 3 24 94 R4C R4C 1 ? ? IC6 15 2 20 7 5 1 p6a o/b 267 1.40 5179#2 3525 0.1 42 24 94 R5C R5C 1 0.51 162 IN6 15 3 20 7 6 1 p6c y/r 245 1.29 5179#6 3525 0.1 2D 26 94 R8A R8A 1 0.96 383 IA7 15 4 20 7 7 1 p7 o/b 233 1.23 6801#4 3025C 0.025 3E 19 81 R6C R6C 1 1.00 710 IB7 16 1 20 7 9 1 p7 y/r 233 1.23 5179#5 3525 0.1 12 24 81 R7C R7C 1 1.00 316 IC7 16 2 20 7 10 1 p7 g/br 233 1.23 5179#4 3525 0.1 40 24 81 R8C R8C 1 0.95 302 IAN10 16 3 20 7 11 1 p10 y/r 375 1.94 6801#5 3025C 0.025 2C 9 82 R5B R5B 1 0.97 218 IN10 16 4 20 7 12 1 p10 g/br 375 1.94 6801#6 3025C 0.025 2E 10 82 R7B R7B 1 0.96 242 Ibucket, ac ,low 18 16 1 bt2 o/b ? ? MS#10 R-2800-4 0.0025 35 14 91 2C 10 1 0.97 3889 Table A-2: 2002 measurement settings for flash FPL0240 (8/27/2002).

PAGE 311

281 Measured Parameter LeCroy Oscilloscope Yokogawa Oscilloscope Fiber Optic Cable Sensor PIC RF PIC ID Isobe Cal. Range [A/V] ID Ch. Sampling Rate [MHz] ID Ch. Sampling Rate [MHz] ID Color low / high Length [m] Delay [ s] ID Model V/A ID Att. [dB] Trans. ID Rec. ID Setting [V] IAN14 6 1 20 7 13 1 p14 y/r 567 2.90 110A#5 110A 0.1 39 19 80 R6B R6B 1 0.99 177 Ibucket, low 6 4 20 18 4 1 bt2 o/b ? ? MS#10 R-2800-4 0.0025 35 14 91 2C 10 1 0.97 3897 Ibucket, high 11 1 20 18 3 1 bt1 o/b ? ? MS#10 R-2800-4 0.0025 3F 34 91 25 25 1 0.96 38571 IG1 11 2 20 18 5 1 p1b o/b 565 2.89 MS#3 R-7000-10 0.001 4 6 87 11 11 0.1 0.92 367 IG2 11 3 20 18 6 1 p2 o/b 435 2.24 MS#6 R-5600-8 0.00125 24 14 86 R2A R2A 0.1 0.99 794 IG6 11 4 20 18 7 1 p6c o/b 245 1.29 MS#7 R-5600-8 0.00125 16 23 94 12 12 0.1 0.99 2228 IG10 12 1 20 18 8 1 p10 o/b 375 1.94 MS#2 R-7000-10 0.001 3A 16 82 18 18 0.1 0.96 1213 IG14 12 2 20 18 9 1 p14 o/b 567 2.90 MS#5 R-5600-8 0.00125 1 9 80 14 14 0.1 0.97 435 IG15 12 3 20 18 10 1 p15 o/b 595 3.04 MS#4 R-7000-10 0.001 15 9 80 13 13 0.1 0.67 379 IBN1 12 4 20 18 11 1 p1a o/b 542.5 2.78 110A#1 110A 0.1 2F 10 87 R8B R8B 1 0.96 61 IAN1 13 1 20 18 12 1 p1a y/r 542.5 2.78 110A#6 110A 0.1 2B 10 87 R1C R1C 1 0.96 61 ICN1 13 2 20 18 13 1 p1a g/br 542.5 2.78 110A#7 110A 0.1 6 10 87 3 3 1 0.97 61 IAN2 13 3 20 18 14 1 p2 g/br 435 2.24 110A#4 110A 0.1 38 20 86 R1B R1B 1 1.00 200 IN3 13 4 20 18 15 1 p3a o/b 558 2.86 6801#1 3025C 0.025 10 10 88 R7A R7A 1 ? ? IAN6 14 1 20 7 15 1 p6b y/r 397 2.05 6801#3 3025C 0.025 8 16 94 R2B R2B 1 0.97 490 IBN6 14 2 20 7 1 1 p6b g/br 397 2.05 110A#3 110A 0.1 3C 29 94 R3B R3B 1 0.97 544 ICN6 14 3 20 7 2 1 p6b o/b 397 2.05 110A#2 110A 0.1 41 29 94 R4B R4B 1 0.96 543 IA6 14 4 20 7 3 1 p6a y/r 267 1.40 6801#2 3025C 0.025 19 17 94 R3C R3C 1 0.97 550 IB6 15 1 20 7 4 1 p6a g/br 267 1.40 5179#3 3525 0.1 3 24 94 R4C R4C 1 1.00 317 IC6 15 2 20 7 5 1 p6a o/b 267 1.40 5179#2 3525 0.1 42 24 94 R5C R5C 1 0.96 305 IN6 15 3 20 7 6 1 p6c y/r 245 1.29 5179#6 3525 0.1 2D 26 94 R8A R8A 1 0.96 384 IA7 15 4 20 7 7 1 p7 o/b 233 1.23 6801#4 3025C 0.025 3E 19 81 R6C R6C 1 1.00 710 IB7 16 1 20 7 9 1 p7 y/r 233 1.23 5179#5 3525 0.1 12 24 81 R7C R7C 1 1.00 316 IC7 16 2 20 7 10 1 p7 g/br 233 1.23 5179#4 3525 0.1 40 24 81 R8C R8C 1 0.96 304 IAN10 16 3 20 7 11 1 p10 y/r 375 1.94 6801#5 3025C 0.025 2C 9 82 R5B R5B 1 0.97 218 IN10 16 4 20 7 12 1 p10 g/br 375 1.94 6801#6 3025C 0.025 2E 10 82 R7B R7B 1 0.96 242 Ibucket, ac ,low 18 16 1 bt2 o/b ? ? MS#10 R-2800-4 0.0025 35 14 91 2C 10 1 0.97 3897 Table A-3: 2002 measurement settings fo r flashes FPL0244, FPL0245, and FPL0246 (9/13/2002).

PAGE 312

282 A.2 2003 Measurement Settings, Direct Strikes The measurement settings for the strikes to phase A between poles 7 and 8 at midspan are found in Table A-4 (FPL0301, FPL0302, FPL0303 on 6/30/2003), in Table A-5 (FPL0305 and FPL0306 on 7/6/2003), in Table A-6 (FPL0312 and FPL0314 on 7/13/2003), and in Table A-7 (FPL0315 and FPL0317 on 7/ 14/2003). Note that flashes FPL0310 and FPL0311 on 7/11/2003 (both triggere d lightning strikes without return strokes) did not trigger the os cilloscopes and consequently data for these flashes were not record ed. Also, note that the peak-to-peak va lue s of the measured calibration signals for parameters IG1 and IB7 in Table A-4 and for parameter IB7 in Table A-6 and Table A-7 are off.

PAGE 313

283 Measured Parameter LeCroy Oscilloscope Yokogawa Oscilloscope Fiber Optic Cable Sensor PIC RF PIC ID Isobe Cal. Range [A/V] ID Ch. Sampling Rate [MHz] ID Ch. Sampling Rate [MHz] ID Color low / high Length [m] Delay [ s] ID Model V/A ID Att. [dB] Trans. ID Rec. ID Setting [V] Isr, high 11 1 20 7 1 1 Tower B y/r 92 0.53 MS#8 R-5600-8 0.00125 1 29 84 23 R1A 1 1.01 45455 Itower, high 11 2 20 7 2 1 Tower A o/b 103 0.58 MS#9 R-5600-8 0.00125 3F 29 84 R2A R2A 1 1.01 45635 IG2 11 3 20 7 3 1 p2 o/b 536 2.75 MS#6 R-5600-8 0.00125 3 9 90 24 R3A 1 1.03 4627 IG14 11 4 20 7 4 1 p14 gb 541 2.77 MS #5 R-5600-8 0.00125 4 6 80 R4A R4A 1 1.01 3221 IG1 12 1 20 7 5 1 p1a g/b 569 2.91 MS#3 R-7000-10 0.001 5 9 9A R5A R5A 1 0.85 4786 IG6 12 2 20 7 6 1 p6c y/r 397 2.05 MS#7 R-5600-8 0.00125 2F 19 8D 21 R6A 1 1.01 14374 IG10 12 3 20 7 7 1 p10 g/b 349 1.81 MS#2 R-7000-10 0.001 17 17 97 15 R5C 1 1.00 14088 IG15 12 4 20 7 8 1 p15 g/b 559 2.86 MS#4 R-7000-10 0.001 6 6 87 R8A R8A 1 0.97 3883 IAN2 13 1 20 7 9 1 p2 y/r 536 2.75 110A#4 110A 0.1 0C 40 90 R1B R1B 1 0.99 1982 IAN6 13 2 20 7 10 1 p6b o/b 400 2.07 6801#3 3025C 0.025 10 43 8D R2B R2B 1 1.00 11278 Itower, ct 13 3 20 18 12 1 Tower A g/b 103 0.58 3025 0.0125 12 29 97 R3B R3B 1 1.01 4545 IAN10, ct 13 4 20 7 12 1 p10 o/b 349 1.81 6801#5 3025C 0.025 16 43 97 12 10 1 1.01 11391 IAN1 14 1 20 7 13 1 p1b o/b 560 2.87 110A#6 110A 0.1 19 26 9A R5B R5B 1 1.02 405 IAN14 14 2 20 7 14 1 p14 o/b 541 2.77 110A#5 110A 0.1 22 39 80 R6B R6B 1 1.00 1781 IA6 14 3 20 7 15 1 p6a o/b 378 1.96 6801#2 3025C 0.025 24 37 8D R7B R7B 1 1.00 5686 IA7 14 4 20 7 16 1 p7 o/b 306 1.60 6801#4 3025C 0.025 08 47 94 1 R8B 1 1.01 18071 IC6 15 1 20 18 1 1 p6a y/r 378 1.96 5179#2 3525 0.1 2A 43 8D R1C R1C 1 0.99 2791 IB6 15 2 20 18 2 1 p6a g/b 378 1.96 5179#3 3525 0.1 2C 43 8D R2C R2C 1 1.01 2839 IB7 15 3 20 18 3 1 p7 r/y 306 1.60 5179#5 3525 0.1 2E 53 94 R3C R3C 1 0.93 8344 IC7 15 4 20 18 4 1 p7 g/b 306 1.60 5179#4 3525 0.1 1A 47 94 R4C R4C 1 0.99 4415 IBN1 16 1 20 18 5 1 p1b y/r 560 2.87 110A#1 110A 0.1 33 23 9A R7A R7A 1 1.00 283 ICN1 16 2 20 18 6 1 p1b g/b 560 2.87 110A#7 110A 0.1 35 23 9A R6C R6C 1 1.01 285 IBN6 16 3 20 18 7 1 p6b y/r 400 2.07 110A#3 110A 0.1 36 43 8D R7C R7C 1 1.01 2853 ICN6 16 4 20 18 8 1 p6b g/b 400 2.07 110A#2 110A 0.1 38 50 8D R8C R8C 1 1.02 6457 IN3 17 1 20 18 9 1 p3 o/b 475 2.44 6801#1 3025C 0.025 39 34 92 2 2 1 0.99 3969 IN6 17 2 20 18 10 1 p6c o/b 397 2.05 5179#6 3525 0.1 3A 53 8D 20 20 1 0.99 8800 IN10 17 3 20 18 11 1 p10 y/r 349 1.81 6801#6 3025C 0.025 3C 43 97 10 4 1 0.99 11165 Itrans + IAN2 17 4 20 7 11 1 p2 g/b 536 2.75 SAT 110A 0.1 3E 40 90 5 5 1 1.00 1996 Isr, ac, low 18 13 1 Tower B g/b 92 0.53 MS#8 R-5600-8 0.00125 29 9 84 7 7 1 0.98 4437 Isr, dc, low 18 14 1 Tower B g/b 92 0.53 MS#8 R-5600-8 0.00125 29 9 84 7 7 1 0.99 4451 Itower, ac, low 18 15 1 Tower A y/r 103 0.58 MS#9 R-5600-8 0.00125 40 9 84 8 8 1 1.01 4545 Itower, dc, low 18 16 1 Tower A y/r 103 0.58 MS#9 R-5600-8 0.00125 40 9 84 8 8 1 1.01 4573 Table A-4: 2003 measurement settings for flashes FPL0301, FPL0302, and FPL0303 (6/30/2003).

PAGE 314

284 Measured Parameter LeCroy Oscilloscope Yokogawa Oscilloscope Fiber Optic Cable Sensor PIC RF PIC ID Isobe Cal. Range [A/V] ID Ch. Sampling Rate [MHz] ID Ch. Sampling Rate [MHz] ID Color low / high Length [m] Delay [ s] ID Model V/A ID Att. [dB] Trans. ID Rec. ID Setting [V] IG1 11 1 20 7 5 1 p1a g/b 569 2.91 MS#3 R-7000-10 0.001 5 9 9A R5A R5A 1 0.97 5445 IG6 11 2 20 7 6 1 p6c y/r 397 2.05 MS#7 R-5600-8 0.00125 2F 19 8D 21 R6A 1 1.01 14331 IG10 11 3 20 7 7 1 p10 g/b 349 1.81 MS#2 R-7000-10 0.001 17 17 97 15 R5C 1 0.99 14060 IG15 11 4 20 7 8 1 p15 g/b 559 2.86 MS#4 R-7000-10 0.001 6 6 87 R8A R8A 1 0.97 3879 Isr, high 12 1 20 7 1 1 Tower B y/r 92 0.53 MS#8 R-5600-8 0.00125 1 29 84 23 R1A 1 1.00 45229 Itower, high 12 2 20 7 2 1 Tower A o/b 103 0.58 MS#9 R-5600-8 0.00125 3F 29 84 R2A R2A 1 1.01 45455 IG2 12 3 20 7 3 1 p2 o/b 536 2.75 MS#6 R-5600-8 0.00125 3 9 90 24 R3A 1 1.03 4622 IG14 12 4 20 7 4 1 p14 gb 541 2.77 MS#5 R-5600-8 0.00125 4 6 80 R4A R4A 1 1.01 3218 IAN2 13 1 20 7 9 1 p2 y/r 536 2.75 110A#4 110A 0.1 0C 40 90 R1B R1B 1 0.99 1982 IAN6 13 2 20 7 10 1 p6b o/b 400 2.07 6801#3 3025C 0.025 10 43 8D R2B R2B 1 1.00 11255 Itower, ct 13 3 20 18 12 1 Tower A g/b 103 0.58 3025 0.0125 12 29 97 R3B R3B 1 1.01 4541 IAN10, ct 13 4 20 7 12 1 p10 o/b 349 1.81 6801#5 3025C 0.025 16 43 97 12 10 1 1.01 11379 IAN1 14 1 20 7 13 1 p1b o/b 560 2.87 110A#6 110A 0.1 19 26 9A R5B R5B 1 1.01 405 IAN14 14 2 20 7 14 1 p14 o/b 541 2.77 110A#5 110A 0.1 22 39 80 R6B R6B 1 1.00 1779 IA6 14 3 20 7 15 1 p6a o/b 378 1.96 6801#2 3025C 0.025 24 37 8D R7B R7B 1 1.00 5675 IA7 14 4 20 7 16 1 p7 o/b 306 1.60 6801#4 3025C 0.025 08 47 94 1 R8B 1 1.01 18035 IC6 15 1 20 18 1 1 p6a y/r 378 1.96 5179#2 3525 0.1 2A 43 8D R1C R1C 1 0.98 2780 IB6 15 2 20 18 2 1 p6a g/b 378 1.96 5179#3 3525 0.1 2C 43 8D R2C R2C 1 1.00 2831 IB7 15 3 20 18 3 1 p7 r/y 306 1.60 5179#5 3525 0.1 2E 53 94 R3C R3C 1 0.98 8755 IC7 15 4 20 18 4 1 p7 g/b 306 1.60 5179#4 3525 0.1 1A 47 94 R4C R4C 1 0.99 4419 IBN1 16 1 20 18 5 1 p1b y/r 560 2.87 110A#1 110A 0.1 33 23 9A R7A R7A 1 1.00 282 ICN1 16 2 20 18 6 1 p1b g/b 560 2.87 110A#7 110A 0.1 35 23 9A R6C R6C 1 1.01 284 IBN6 16 3 20 18 7 1 p6b y/r 400 2.07 110A#3 110A 0.1 36 43 8D R7C R7C 1 1.01 2842 ICN6 16 4 20 18 8 1 p6b g/b 400 2.07 110A#2 110A 0.1 38 50 8D R8C R8C 1 1.02 6445 IN3 17 1 20 18 9 1 p3 o/b 475 2.44 6801#1 3025C 0.025 39 34 92 2 2 1 0.99 3953 IN6 17 2 20 18 10 1 p6c o/b 397 2.05 5179#6 3525 0.1 3A 53 8D 20 20 1 0.98 8782 IN10 17 3 20 18 11 1 p10 y/r 349 1.81 6801#6 3025C 0.025 3C 43 97 10 4 1 1.00 11323 Itrans + IAN2 17 4 20 7 11 1 p2 g/b 536 2.75 SAT 110A 0.1 3E 40 90 5 5 1 1.00 1994 Isr, ac, low 18 13 1 Tower B g/b 92 0.53 MS#8 R-5600-8 0.00125 29 9 84 7 7 1 0.98 4415 Isr, dc, low 18 14 1 Tower B g/b 92 0.53 MS#8 R-5600-8 0.00125 29 9 84 7 7 1 0.98 4433 Itower, ac, low 18 15 1 Tower A y/r 103 0.58 MS#9 R-5600-8 0.00125 40 9 84 8 8 1 1.01 4550 Itower, dc, low 18 16 1 Tower A y/r 103 0.58 MS#9 R-5600-8 0.00125 40 9 84 8 8 1 1.01 4550 Table A-5: 2003 measurement settings for flashes FPL0305 and FPL0306 (7/6/2003).

PAGE 315

285 Measured Parameter LeCroy Oscilloscope Yokogawa Oscilloscope Fiber Optic Cable Sensor PIC RF PIC ID Isobe Cal. Range [A/V] ID Ch. Sampling Rate [MHz] ID Ch. Sampling Rate [MHz] ID Color low / high Length [m] Delay [ s] ID Model V/A ID Att. [dB] Trans. ID Rec. ID Setting [V] IG1 11 1 20 7 5 1 p1a g/b 569 2.91 MS#3 R-7000-10 0.001 5 9 9A R5A R5A 1 0.99 5569 IG6 11 2 20 7 6 1 p6c y/r 397 2.05 MS#7 R-5600-8 0.00125 2F 19 8D 21 R6A 1 1.01 14346 IG10 11 3 20 7 7 1 p10 g/b 349 1.81 MS#2 R-7000-10 0.001 17 17 97 15 R5C 1 0.99 14003 IG15 11 4 20 7 8 1 p15 g/b 559 2.86 MS#4 R-7000-10 0.001 6 6 87 R8A R8A 1 0.97 3887 Isr, high 12 1 20 7 1 1 Tower B y/r 92 0.53 MS#8 R-5600-8 0.00125 1 29 84 23 R1A 1 1.00 45275 Itower, high 12 2 20 7 2 1 Tower A o/b 103 0.58 MS#9 R-5600-8 0.00125 3F 29 84 R2A R2A 1 1.01 45365 IG2 12 3 20 7 3 1 p2 o/b 536 2.75 MS#6 R-5600-8 0.00125 3 9 90 24 R3A 1 1.02 4582 IG14 12 4 20 7 4 1 p14 gb 541 2.77 MS#5 R-5600-8 0.00125 4 6 80 R4A R4A 1 1.01 3212 IAN2 13 1 20 7 9 1 p2 y/r 536 2.75 110A#4 110A 0.1 0C 40 90 R1B R1B 1 0.99 1980 IAN6 13 2 20 7 10 1 p6b o/b 400 2.07 6801#3 3025C 0.025 10 43 8D R2B R2B 1 1.00 11244 Itower, ct 13 3 20 18 12 1 Tower A g/b 103 0.58 3025 0.0125 12 29 97 R3B R3B 1 1.004 4527 IAN10, ct 13 4 20 7 12 1 p10 o/b 349 1.81 6801#5 3025C 0.025 16 43 97 12 10 1 1.01 11368 IAN1 14 1 20 7 13 1 p1b o/b 560 2.87 110A#6 110A 0.1 19 26 9A R5B R5B 1 1.02 406 IAN14 14 2 20 7 14 1 p14 o/b 541 2.77 110A#5 110A 0.1 22 39 80 R6B R6B 1 1.00 1775 IA6 14 3 20 7 15 1 p6a o/b 378 1.96 6801#2 3025C 0.025 24 37 8D R7B R7B 1 1.00 5675 IA7 14 4 20 7 16 1 p7 o/b 306 1.60 6801#4 3025C 0.025 08 47 94 1 R8B 1 1.01 18017 IC6 15 1 20 18 1 1 p6a y/r 378 1.96 5179#2 3525 0.1 2A 43 8D R1C R1C 1 0.99 2786 IB6 15 2 20 18 2 1 p6a g/b 378 1.96 5179#3 3525 0.1 2C 43 8D R2C R2C 1 1.00 2836 IB7 15 3 20 18 3 1 p7 r/y 306 1.60 5179#5 3525 0.1 2E 53 94 R3C R3C 1 0.91 8148 IC7 15 4 20 18 4 1 p7 g/b 306 1.60 5179#4 3525 0.1 1A 47 94 R4C R4C 1 0.98 4406 IBN1 16 1 20 18 5 1 p1b y/r 560 2.87 110A#1 110A 0.1 33 23 9A R7A R7A 1 1.00 282 ICN1 16 2 20 18 6 1 p1b g/b 560 2.87 110A#7 110A 0.1 35 23 9A R6C R6C 1 1.01 284 IBN6 16 3 20 18 7 1 p6b y/r 400 2.07 110A#3 110A 0.1 36 43 8D R7C R7C 1 1.01 2839 ICN6 16 4 20 18 8 1 p6b g/b 400 2.07 110A#2 110A 0.1 38 50 8D R8C R8C 1 1.02 6432 IN3 17 1 20 18 9 1 p3 o/b 475 2.44 6801#1 3025C 0.025 39 34 92 2 2 1 0.98 3945 IN6 17 2 20 18 10 1 p6c o/b 397 2.05 5179#6 3525 0.1 3A 53 8D 20 20 1 0.98 8782 IN10 17 3 20 18 11 1 p10 y/r 349 1.81 6801#6 3025C 0.025 3C 43 97 10 4 1 1.00 11289 Itrans 17 4 20 7 11 1 p2 g/b 536 2.75 SAT 110A 0.1 3E 40 90 5 5 1 1.00 1994 Isr, ac, low 18 13 1 Tower B g/b 92 0.53 MS#8 R-5600-8 0.00125 29 9 84 7 7 1 0.98 4406 Isr, dc, low 18 14 1 Tower B g/b 92 0.53 MS#8 R-5600-8 0.00125 29 9 84 7 7 1 0.98 4406 Itower, ac, low 18 15 1 Tower A y/r 103 0.58 MS#9 R-5600-8 0.00125 40 9 84 8 8 1 0.98 4397 Itower, dc, low 18 16 1 Tower A y/r 103 0.58 MS#9 R-5600-8 0.00125 40 9 84 8 8 1 0.97 4388 Table A-6: 2003 measurement settings for flashes FPL0312 and FPL0314 (7/13/2003).

PAGE 316

286 Measured Parameter LeCroy Oscilloscope Yokogawa Oscilloscope Fiber Optic Cable Sensor PIC RF PIC ID Isobe Cal. Range [A/V] ID Ch. Sampling Rate [MHz] ID Ch. Sampling Rate [MHz] ID Color low / high Length [m] Delay [ s] ID Model V/A ID Att. [dB] Trans. ID Rec. ID Setting [V] IG1 11 1 20 7 5 2 p1a g/b 569 2.91 MS#3 R-7000-10 0.001 5 9 9A R5A R5A 1 0.99 5575 IG6 11 2 20 7 6 2 p6c y/r 397 2.05 MS#7 R-5600-8 0.00125 2F 19 8D 21 R6A 1 1.01 14346 IG10 11 3 20 7 7 2 p10 g/b 349 1.81 MS#2 R-7000-10 0.001 17 17 97 15 R5C 1 0.99 14074 IG15 11 4 20 7 8 2 p15 g/b 559 2.86 MS#4 R-7000-10 0.001 6 6 87 R8A R8A 1 0.97 3883 Isr, high 12 1 20 7 1 2 Tower B y/r 92 0.53 MS#8 R-5600-8 0.00125 1 29 84 23 R1A 1 1.01 45365 Itower, high 12 2 20 7 2 2 Tower A o/b 103 0.58 MS#9 R-5600-8 0.00125 3F 29 84 R2A R2A 1 1.01 45545 IG2 12 3 20 7 3 2 p2 o/b 536 2.75 MS#6 R-5600-8 0.00125 3 9 90 24 R3A 1 1.02 4609 IG14 12 4 20 7 4 2 p14 gb 541 2.77 MS#5 R-5600-8 0.00125 4 6 80 R4A R4A 1 1.01 3218 IAN2 13 1 20 7 9 2 p2 y/r 536 2.75 110A#4 110A 0.1 0C 40 90 R1B R1B 1 0.99 1984 IAN6 13 2 20 7 10 2 p6b o/b 400 2.07 6801#3 3025C 0.025 10 43 8D R2B R2B 1 1.00 11278 Itower, ct 13 3 20 18 12 2 Tower A g/b 103 0.58 3025 0.0125 12 29 97 R3B R3B 1 1.01 4541 IAN10, ct 13 4 20 7 12 2 p10 o/b 349 1.81 6801#5 3025C 0.025 16 43 97 16 16 1 1.01 11391 IAN1 14 1 20 7 13 2 p1b o/b 560 2.87 110A#6 110A 0.1 19 26 9A R5B R5B 1 1.02 406 IAN14 14 2 20 7 14 2 p14 o/b 541 2.77 110A#5 110A 0.1 22 39 80 R6B R6B 1 1.00 1781 IA6 14 3 20 7 15 2 p6a o/b 378 1.96 6801#2 3025C 0.025 24 37 8D R7B R7B 1 1.00 5686 IA7 14 4 20 7 16 2 p7 o/b 306 1.60 6801#4 3025C 0.025 08 47 94 1 R8B 1 1.01 18053 IC6 15 1 20 18 1 2 p6a y/r 378 1.96 5179#2 3525 0.1 2A 43 8D R1C R1C 1 0.99 2788 IB6 15 2 20 18 2 2 p6a g/b 378 1.96 5179#3 3525 0.1 2C 43 8D R2C R2C 1 1.01 2839 IB7 15 3 20 18 3 2 p7 r/y 306 1.60 5179#5 3525 0.1 2E 53 94 R3C R3C 1 0.90 8058 IC7 15 4 20 18 4 2 p7 g/b 306 1.60 5179#4 3525 0.1 1A 47 94 R4C R4C 1 0.99 4415 IBN1 16 1 20 18 5 2 p1b y/r 560 2.87 110A#1 110A 0.1 33 23 9A R7A R7A 1 1.00 283 ICN1 16 2 20 18 6 2 p1b g/b 560 2.87 110A#7 110A 0.1 35 23 9A R6C R6C 1 1.01 285 IBN6 16 3 20 18 7 2 p6b y/r 400 2.07 110A#3 110A 0.1 36 43 8D R7C R7C 1 1.01 2848 ICN6 16 4 20 18 8 2 p6b g/b 400 2.07 110A#2 110A 0.1 38 50 8D R8C R8C 1 1.01 6375 IN3 17 1 20 18 9 2 p3 o/b 475 2.44 6801#1 3025C 0.025 39 34 92 2 2 1 0.99 3961 IN6 17 2 20 18 10 2 p6c o/b 397 2.05 5179#6 3525 0.1 3A 53 8D 20 20 1 0.99 8800 IN10 17 3 20 18 11 2 p10 y/r 349 1.81 6801#6 3025C 0.025 3C 43 97 10 4 1 1.00 11323 Itrans 17 4 20 7 11 2 p2 g/b 536 2.75 SAT 110A 0.1 3E 40 90 5 5 1 1.00 1996 Isr, ac, low 18 13 2 Tower B g/b 92 0.53 MS#8 R-5600-8 0.00125 29 9 84 7 7 1 0.98 4428 Isr, dc, low 18 14 2 Tower B g/b 92 0.53 MS#8 R-5600-8 0.00125 29 9 84 7 7 1 0.99 4442 Itower, ac, low 18 15 2 Tower A y/r 103 0.58 MS#9 R-5600-8 0.00125 40 9 84 8 8 1 0.98 4401 Itower, dc, low 18 16 2 Tower A y/r 103 0.58 MS#9 R-5600-8 0.00125 40 9 84 8 8 1 0.98 4401 Table A-7: 2003 measurement settings for flashes FPL0315 and FPL0317 (7/14/2003).

PAGE 317

287 A.3 2003 Measurement Settings, Nearby Strikes The measurement settings for the strikes to ground at 7 m south of pole 4 are found in Table A-8 (FPL0321 on 7/18/2003), Table A-9 (FPL0329 and FPL0331 on 7/22/2003), to ground at 15 m south of pole 4 are found in Table A-10 (FPL0336 on 8/2/2003), and to ground 11 m south-east of pole 15 are found in Table A-11 (FPL0341 on 8/7/2003), Table A-12 (FPL0342 and FPL 0345 on 8/11/2003), and Table A-13 (FPL0347, FPL0348, and FPL0350 on 8/15/2003). Note that fl ashes FP L0326, FPL0327, and FPL0335 (all triggered lightning strikes wit hout return strokes) did not trigger the oscilloscopes and consequently data for these flashes are not av ailable. Also, no te that the peak-to peak values of the measured calibration signals for parameter IB7 in all tables are approximately 10% too low.

PAGE 318

288 Measured Parameter LeCroy Oscilloscope Yokogawa Oscilloscope Fiber Optic Cable Sensor PIC RF PIC ID Isobe Cal. Range [A/V] ID Ch. Sampling Rate [MHz] ID Ch. Sampling Rate [MHz] ID Color low / high Length [m] Delay [ s] ID Model V/A ID Att. [dB] Trans. ID Rec. ID Setting [V] IG1 11 1 20 7 5 2 p1a g/b 569 2.91 MS#3 R-7000-10 0.001 5 9 9A R5A R5A 0.1 0.98 554 IG6 11 2 20 7 6 2 p6c y/r 397 2.05 MS#7 R-5600-8 0.00125 2F 3 8D 21 R6A 1 1.01 2276 IG10 11 3 20 7 7 2 p10 g/b 349 1.81 MS#2 R-7000-10 0.001 17 16 97 15 R5C 0.1 0.98 1239 IG15 11 4 20 7 8 2 p15 g/b 559 2.86 MS#4 R-7000-10 0.001 6 9 87 R8A R8A 0.1 0.97 546 Ibucket, high 12 1 20 7 1 2 bkt w/g 588 3.01 MS#10 R-2800-4 0.0025 1 34 89 23 R1A 1 1.01 40376 IG2 12 3 20 7 3 2 p2 o/b 536 2.75 MS#6 R-5600-8 0.00125 3 6 90 24 R3A 1 1.03 3272 IG14 12 4 20 7 4 2 p14 gb 541 2.77 MS#5 R-5600-8 0.00125 4 9 80 R4A R4A 0.1 1.00 452 IAN2 13 1 20 7 9 2 p2 y/r 536 2.75 110A#4 110A 0.1 0C 36 90 R1B R1B 1 0.99 1252 IAN6 13 2 20 7 10 2 p6b o/b 400 2.07 6801#3 3025C 0.025 10 17 8D R2B R2B 1 1.00 566 IAN10, ct 13 4 20 7 12 2 p10 o/b 349 1.81 6801#5 3025C 0.025 16 14 97 12 10 1 1.01 404 IAN1 14 1 20 7 13 2 p1b o/b 560 2.87 110A#6 110A 0.1 19 10 9A R5B R5B 1 1.02 64 IAN14 14 2 20 7 14 2 p14 o/b 541 2.77 110A#5 110A 0.1 22 26 80 R6B R6B 1 1.00 399 IA6 14 3 20 7 15 2 p6a o/b 378 1.96 6801#2 3025C 0.025 24 23 8D R7B R7B 1 1.00 1135 IA7 14 4 20 7 16 2 p7 o/b 306 1.60 6801#4 3025C 0.025 08 19 94 1 R8B 1 1.01 719 IC6 15 1 20 18 1 2 p6a y/r 378 1.96 5179#2 3525 0.1 2A 29 8D R1C R1C 1 0.99 558 IB6 15 2 20 18 2 2 p6a g/b 378 1.96 5179#3 3525 0.1 2C 29 8D R2C R2C 1 1.01 566 IB7 15 3 20 18 3 2 p7 r/y 306 1.60 5179#5 3525 0.1 2E 27 94 R3C R3C 1 0.90 402 IC7 15 4 20 18 4 2 p7 g/b 306 1.60 5179#4 3525 0.1 1A 27 94 R4C R4C 1 0.99 441 IBN1 16 1 20 18 5 2 p1b y/r 560 2.87 110A#1 110A 0.1 33 10 9A R7A R7A 1 1.00 63 ICN1 16 2 20 18 6 2 p1b g/b 560 2.87 110A#7 110A 0.1 35 10 9A R6C R6C 1 1.01 64 IBN6 16 3 20 18 7 2 p6b y/r 400 2.07 110A#3 110A 0.1 36 27 8D R7C R7C 1 1.01 451 ICN6 16 4 20 18 8 2 p6b g/b 400 2.07 110A#2 110A 0.1 38 24 8D R8C R8C 1 1.02 324 IN3 17 1 20 18 9 2 p3 o/b 475 2.44 6801#1 3025C 0.025 39 19 92 2 2 1 0.99 704 IN6 17 2 20 18 10 2 p6c o/b 397 2.05 5179#6 3525 0.1 3A 29 8D 20 20 1 0.99 555 IN10 17 3 20 18 11 2 p10 y/r 349 1.81 6801#6 3025C 0.025 3C 16 97 10 4 1 1.00 506 Ibucket, ac, low 18 13 2 bkt o/b 588 3.01 MS#10 R-2800-4 0.0025 29 14 89 7 7 1 0.99 3949 Ibucket, dc, low 18 14 2 bkt o/b 588 3.01 MS#10 R-2800-4 0.0025 29 14 89 7 7 1 0.99 3953 Table A-8: 2003 measurement settings for flash FPL0321 (7/18/2003).

PAGE 319

289 Measured Parameter LeCroy Oscilloscope Yokogawa Oscilloscope Fiber Optic Cable Sensor PIC RF PIC ID Isobe Cal. Range [A/V] ID Ch. Sampling Rate [MHz] ID Ch. Sampling Rate [MHz] ID Color low / high Length [m] Delay [ s] ID Model V/A ID Att. [dB] Trans. ID Rec. ID Setting [V] IG1 11 1 20 7 5 2 p1a g/b 569 2.91 MS#3 R-7000-10 0.001 5 9 9A R5A R5A 0.1 0.98 554 IG6 11 2 20 7 6 2 p6c y/r 397 2.05 MS#7 R-5600-8 0.00125 2F 3 8D 21 R6A 1 1.01 2276 IG10 11 3 20 7 7 2 p10 g/b 349 1.81 MS#2 R-7000-10 0.001 17 16 97 15 R5C 0.1 0.98 1239 IG15 11 4 20 7 8 2 p15 g/b 559 2.86 MS#4 R-7000-10 0.001 6 9 87 R8A R8A 0.1 0.97 546 Ibucket, high 12 1 20 7 1 2 bkt w/g 588 3.01 MS#10 R-2800-4 0.0025 1 34 89 23 R1A 1 1.00 40215 IG2 12 3 20 7 3 2 p2 o/b 536 2.75 MS#6 R-5600-8 0.00125 3 6 90 24 R3A 1 1.02 3269 IG14 12 4 20 7 4 2 p14 gb 541 2.77 MS#5 R-5600-8 0.00125 4 9 80 R4A R4A 0.1 1.00 452 IAN2 13 1 20 7 9 2 p2 y/r 536 2.75 110A#4 110A 0.1 0C 36 90 R1B R1B 1 0.99 1252 IAN6 13 2 20 7 10 2 p6b o/b 400 2.07 6801#3 3025C 0.025 10 17 8D R2B R2B 1 1.00 566 IAN10, ct 13 4 20 7 12 2 p10 o/b 349 1.81 6801#5 3025C 0.025 16 14 97 12 10 1 1.01 404 IAN1 14 1 20 7 13 2 p1b o/b 560 2.87 110A#6 110A 0.1 19 10 9A R5B R5B 1 1.02 64 IAN14 14 2 20 7 14 2 p14 o/b 541 2.77 110A#5 110A 0.1 22 26 80 R6B R6B 1 1.00 399 IA6 14 3 20 7 15 2 p6a o/b 378 1.96 6801#2 3025C 0.025 24 23 8D R7B R7B 1 1.00 1135 IA7 14 4 20 7 16 2 p7 o/b 306 1.60 6801#4 3025C 0.025 08 19 94 1 R8B 1 1.01 719 IC6 15 1 20 18 1 2 p6a y/r 378 1.96 5179#2 3525 0.1 2A 29 8D R1C R1C 1 0.99 558 IB6 15 2 20 18 2 2 p6a g/b 378 1.96 5179#3 3525 0.1 2C 29 8D R2C R2C 1 1.01 566 IB7 15 3 20 18 3 2 p7 r/y 306 1.60 5179#5 3525 0.1 2E 27 94 R3C R3C 1 0.90 403 IC7 15 4 20 18 4 2 p7 g/b 306 1.60 5179#4 3525 0.1 1A 27 94 R4C R4C 1 0.99 442 IBN1 16 1 20 18 5 2 p1b y/r 560 2.87 110A#1 110A 0.1 33 10 9A R7A R7A 1 1.00 63 ICN1 16 2 20 18 6 2 p1b g/b 560 2.87 110A#7 110A 0.1 35 10 9A R6C R6C 1 1.01 64 IBN6 16 3 20 18 7 2 p6b y/r 400 2.07 110A#3 110A 0.1 36 27 8D R7C R7C 1 1.01 451 ICN6 16 4 20 18 8 2 p6b g/b 400 2.07 110A#2 110A 0.1 38 24 8D R8C R8C 1 1.02 324 IN3 17 1 20 18 9 2 p3 o/b 475 2.44 6801#1 3025C 0.025 39 19 92 2 2 1 0.99 704 IN6 17 2 20 18 10 2 p6c o/b 397 2.05 5179#6 3525 0.1 3A 29 8D 20 20 1 0.99 555 IN10 17 3 20 18 11 2 p10 y/r 349 1.81 6801#6 3025C 0.025 3C 16 97 10 4 1 1.00 506 Ibucket, ac, low 18 13 2 bkt o/b 588 3.01 MS#10 R-2800-4 0.0025 29 14 89 7 7 1 0.99 3949 Ibucket, dc, low 18 14 2 bkt o/b 588 3.01 MS#10 R-2800-4 0.0025 29 14 89 7 7 1 0.99 3961 Table A-9: 2003 measurement settings for flashes FPL0329 and FPL0331 (7/22/2003).

PAGE 320

290 Measured Parameter LeCroy Oscilloscope Yokogawa Oscilloscope Fiber Optic Cable Sensor PIC RF PIC ID Isobe Cal. Range [A/V] ID Ch. Sampling Rate [MHz] ID Ch. Sampling Rate [MHz] ID Color low / high Length [m] Delay [ s] ID Model V/A ID Att. [dB] Trans. ID Rec. ID Setting [V] IG1 11 1 20 7 5 2 p1a g/b 569 2.91 MS#3 R-7000-10 0.001 5 9 9A R5A R5A 0.1 0.98 555 IG6 11 2 20 7 6 2 p6c y/r 397 2.05 MS#7 R-5600-8 0.00125 2F 3 8D 21 R6A 1 1.01 2276 IG10 11 3 20 7 7 2 p10 g/b 349 1.81 MS#2 R-7000-10 0.001 17 16 97 15 R5C 0.1 0.98 1238 IG15 11 4 20 7 8 2 p15 g/b 559 2.86 MS#4 R-7000-10 0.001 6 9 87 R8A R8A 0.1 0.97 546 Ibucket, high 12 1 20 7 1 2 bkt w/g 588 3.01 MS#10 R-2800-4 0.0025 1 34 89 23 R1A 1 1.00 40135 IG2 12 3 20 7 3 2 p2 o/b 536 2.75 MS#6 R-5600-8 0.00125 3 6 90 24 R3A 1 1.02 3266 IG14 12 4 20 7 4 2 p14 gb 541 2.77 MS#5 R-5600-8 0.00125 4 9 80 R4A R4A 0.1 1.00 451 IAN2 13 1 20 7 9 2 p2 y/r 536 2.75 110A#4 110A 0.1 0C 36 90 R1B R1B 1 0.99 1251 IAN6 13 2 20 7 10 2 p6b o/b 400 2.07 6801#3 3025C 0.025 10 17 8D R2B R2B 1 1.00 565 IAN10, ct 13 4 20 7 12 2 p10 o/b 349 1.81 6801#5 3025C 0.025 16 14 97 12 10 1 1.01 404 IAN1 14 1 20 7 13 2 p1b o/b 560 2.87 110A#6 110A 0.1 19 16 9A R5B R5B 1 1.02 128 IAN14 14 2 20 7 14 2 p14 o/b 541 2.77 110A#5 110A 0.1 22 26 80 R6B R6B 1 1.00 398 IA6 14 3 20 7 15 2 p6a o/b 378 1.96 6801#2 3025C 0.025 24 23 8D R7B R7B 1 1.00 1133 IA7 14 4 20 7 16 2 p7 o/b 306 1.60 6801#4 3025C 0.025 08 19 94 1 R8B 1 1.01 718 IC6 15 1 20 18 1 2 p6a y/r 378 1.96 5179#2 3525 0.1 2A 29 8D R1C R1C 1 0.99 558 IB6 15 2 20 18 2 2 p6a g/b 378 1.96 5179#3 3525 0.1 2C 29 8D R2C R2C 1 1.01 566 IB7 15 3 20 18 3 2 p7 r/y 306 1.60 5179#5 3525 0.1 2E 27 94 R3C R3C 1 0.90 402 IC7 15 4 20 18 4 2 p7 g/b 306 1.60 5179#4 3525 0.1 1A 27 94 R4C R4C 1 0.99 441 IBN1 16 1 20 18 5 2 p1b y/r 560 2.87 110A#1 110A 0.1 33 16 9A R7A R7A 1 1.00 126 ICN1 16 2 20 18 6 2 p1b g/b 560 2.87 110A#7 110A 0.1 35 16 9A R6C R6C 1 1.01 127 IBN6 16 3 20 18 7 2 p6b y/r 400 2.07 110A#3 110A 0.1 36 27 8D R7C R7C 1 1.01 451 ICN6 16 4 20 18 8 2 p6b g/b 400 2.07 110A#2 110A 0.1 38 24 8D R8C R8C 1 1.02 323 IN3 17 1 20 18 9 2 p3 o/b 475 2.44 6801#1 3025C 0.025 39 19 92 2 2 1 0.99 703 IN6 17 2 20 18 10 2 p6c o/b 397 2.05 5179#6 3525 0.1 3A 29 8D 20 20 1 0.99 555 IN10 17 3 20 18 11 2 p10 y/r 349 1.81 6801#6 3025C 0.025 3C 16 97 10 4 1 1.00 506 Ibucket, ac, low 18 13 2 bkt o/b 588 3.01 MS#10 R-2800-4 0.0025 29 14 89 7 7 1 0.99 3949 Ibucket, dc, low 18 14 2 bkt o/b 588 3.01 MS#10 R-2800-4 0.0025 29 14 89 7 7 1 0.99 3953 Table A-10: 2003 measurement setti ngs for flash FPL0336 (8/2/2003).

PAGE 321

291 Measured Parameter LeCroy Oscilloscope Yokogawa Oscilloscope Fiber Optic Cable Sensor PIC RF PIC ID Isobe Cal. Range [A/V] ID Ch. Sampling Rate [MHz] ID Ch. Sampling Rate [MHz] ID Color low / high Length [m] Delay [ s] ID Model V/A ID Att. [dB] Trans. ID Rec. ID Setting [V] IG1 11 1 20 7 5 2 p1a g/b 569 2.91 MS#3 R-7000-10 0.001 5 9 9A R5A R5A 0.1 0.98 554 IG6 11 2 20 7 6 2 p6c y/r 397 2.05 MS#7 R-5600-8 0.00125 2F 19 8D 21 R6A 0.1 0.99 1415 IG10 11 3 20 7 7 2 p10 g/b 349 1.81 MS#2 R-7000-10 0.001 17 27 97 15 R5C 0.1 0.98 4383 IG15 11 4 20 7 8 2 p15 g/b 559 2.86 MS#4 R-7000-10 0.001 6 19 87 R8A R8A 1 0.98 17379 Ibucket, high 12 1 20 7 1 2 bkt w/g 588 3.01 MS#10 R-2800-4 0.0025 1 34 89 23 R1A 1 1.00 39975 IG2 12 3 20 7 3 2 p2 o/b 536 2.75 MS#6 R-5600-8 0.00125 3 14 90 24 R3A 0.1 0.99 797 IG14 12 4 20 7 4 2 p14 gb 541 2.77 MS#5 R-5600-8 0.00125 4 13 80 R4A R4A 1 1.00 7176 IAN2 13 1 20 7 9 2 p2 y/r 536 2.75 110A#4 110A 0.1 0C 29 90 R1B R1B 1 0.99 559 IAN6 13 2 20 7 10 2 p6b o/b 400 2.07 6801#3 3025C 0.025 10 17 8D R2B R2B 1 1.00 564 IAN10, ct 13 4 20 7 12 2 p10 o/b 349 1.81 6801#5 3025C 0.025 16 23 97 12 10 1 1.01 1137 IAN1 14 1 20 7 13 2 p1b o/b 560 2.87 110A#6 110A 0.1 19 16 9A R5B R5B 1 1.02 128 IAN14 14 2 20 7 14 2 p14 o/b 541 2.77 110A#5 110A 0.1 22 37 80 R6B R6B 1 1.00 1410 IA6 14 3 20 7 15 2 p6a o/b 378 1.96 6801#2 3025C 0.025 24 13 8D R7B R7B 1 1.00 358 IA7 14 4 20 7 16 2 p7 o/b 306 1.60 6801#4 3025C 0.025 08 24 94 1 R8B 1 1.01 1274 IC6 15 1 20 18 1 2 p6a y/r 378 1.96 5179#2 3525 0.1 2A 17 8D R1C R1C 1 0.99 140 IB6 15 2 20 18 2 2 p6a g/b 378 1.96 5179#3 3525 0.1 2C 27 8D R2C R2C 1 1.00 449 IB7 15 3 20 18 3 2 p7 r/y 306 1.60 5179#5 3525 0.1 2E 27 94 R3C R3C 1 0.93 415 IC7 15 4 20 18 4 2 p7 g/b 306 1.60 5179#4 3525 0.1 1A 30 94 R4C R4C 1 0.99 623 IBN1 16 1 20 18 5 2 p1b y/r 560 2.87 110A#1 110A 0.1 33 16 9A R7A R7A 1 1.00 126 ICN1 16 2 20 18 6 2 p1b g/b 560 2.87 110A#7 110A 0.1 35 16 9A R6C R6C 1 1.00 127 IBN6 16 3 20 18 7 2 p6b y/r 400 2.07 110A#3 110A 0.1 36 29 8D R7C R7C 1 1.01 566 ICN6 16 4 20 18 8 2 p6b g/b 400 2.07 110A#2 110A 0.1 38 29 8D R8C R8C 1 1.02 573 IN3 17 1 20 18 9 2 p3 o/b 475 2.44 6801#1 3025C 0.025 39 9 92 2 2 1 0.99 222 IN6 17 2 20 18 10 2 p6c o/b 397 2.05 5179#6 3525 0.1 3A 19 8D 20 20 1 0.98 175 IN10 17 3 20 18 11 2 p10 y/r 349 1.81 6801#6 3025C 0.025 3C 19 97 10 4 1 0.99 706 Ibucket, ac, low 18 13 2 bkt o/b 588 3.01 MS#10 R-2800-4 0.0025 29 14 89 7 7 1 0.99 3949 Ibucket, dc, low 18 14 2 bkt o/b 588 3.01 MS#10 R-2800-4 0.0025 29 14 89 7 7 1 0.99 3949 Table A-11: 2003 measurement setti ngs for flash FPL0341 (8/7/2003).

PAGE 322

292 Measured Parameter LeCroy Oscilloscope Yokogawa Oscilloscope Fiber Optic Cable Sensor PIC RF PIC ID Isobe Cal. Range [A/V] ID Ch. Sampling Rate [MHz] ID Ch. Sampling Rate [MHz] ID Color low / high Length [m] Delay [ s] ID Model V/A ID Att. [dB] Trans. ID Rec. ID Setting [V] IG1 11 1 20 7 5 2 p1a g/b 569 2.91 MS#3 R-7000-10 0.001 5 9 9A R5A R5A 0.1 0.98 554 IG6 11 2 20 7 6 2 p6c y/r 397 2.05 MS#7 R-5600-8 0.00125 2F 19 8D 21 R6A 0.1 0.99 1416 IG10 11 3 20 7 7 2 p10 g/b 349 1.81 MS#2 R-7000-10 0.001 17 27 97 15 R5C 0.1 0.98 4388 IG15 11 4 20 7 8 2 p15 g/b 559 2.86 MS#4 R-7000-10 0.001 6 19 87 R8A R8A 1 0.97 17362 Ibucket, high 12 1 20 7 1 2 bkt w/g 588 3.01 MS#10 R-2800-4 0.0025 1 34 89 23 R1A 1 1.00 39975 IG2 12 3 20 7 3 2 p2 o/b 536 2.75 MS#6 R-5600-8 0.00125 3 14 90 24 R3A 0.1 1.00 804 IG14 12 4 20 7 4 2 p14 gb 541 2.77 MS#5 R-5600-8 0.00125 4 13 80 R4A R4A 1 1.01 7183 IAN2 13 1 20 7 9 2 p2 y/r 536 2.75 110A#4 110A 0.1 0C 29 90 R1B R1B 1 0.99 558 IAN6 13 2 20 7 10 2 p6b o/b 400 2.07 6801#3 3025C 0.025 10 17 8D R2B R2B 1 1.00 564 IAN10, ct 13 4 20 7 12 2 p10 o/b 349 1.81 6801#5 3025C 0.025 16 23 97 12 10 1 1.01 1137 IAN1 14 1 20 7 13 2 p1b o/b 560 2.87 110A#6 110A 0.1 19 16 9A R5B R5B 1 1.00 126 IAN14 14 2 20 7 14 2 p14 o/b 541 2.77 110A#5 110A 0.1 22 37 80 R6B R6B 1 1.00 1409 IA6 14 3 20 7 15 2 p6a o/b 378 1.96 6801#2 3025C 0.025 24 13 8D R7B R7B 1 1.00 358 IA7 14 4 20 7 16 2 p7 o/b 306 1.60 6801#4 3025C 0.025 08 24 94 1 R8B 1 1.01 1274 IC6 15 1 20 18 1 2 p6a y/r 378 1.96 5179#2 3525 0.1 2A 17 8D R1C R1C 1 0.99 140 IB6 15 2 20 18 2 2 p6a g/b 378 1.96 5179#3 3525 0.1 2C 27 8D R2C R2C 1 1.00 450 IB7 15 3 20 18 3 2 p7 r/y 306 1.60 5179#5 3525 0.1 2E 27 94 R3C R3C 1 0.92 413 IC7 15 4 20 18 4 2 p7 g/b 306 1.60 5179#4 3525 0.1 1A 30 94 R4C R4C 1 0.99 623 IBN1 16 1 20 18 5 2 p1b y/r 560 2.87 110A#1 110A 0.1 33 16 9A R7A R7A 1 1.00 126 ICN1 16 2 20 18 6 2 p1b g/b 560 2.87 110A#7 110A 0.1 35 16 9A R6C R6C 1 1.01 127 IBN6 16 3 20 18 7 2 p6b y/r 400 2.07 110A#3 110A 0.1 36 29 8D R7C R7C 1 1.01 566 ICN6 16 4 20 18 8 2 p6b g/b 400 2.07 110A#2 110A 0.1 38 29 8D R8C R8C 1 1.02 574 IN3 17 1 20 18 9 2 p3 o/b 475 2.44 6801#1 3025C 0.025 39 9 92 2 2 1 0.99 222 IN6 17 2 20 18 10 2 p6c o/b 397 2.05 5179#6 3525 0.1 3A 19 8D 20 20 1 0.98 175 IN10 17 3 20 18 11 2 p10 y/r 349 1.81 6801#6 3025C 0.025 3C 19 97 10 4 1 0.98 699 Ibucket, ac, low 18 13 2 bkt o/b 588 3.01 MS#10 R-2800-4 0.0025 29 14 89 7 7 1 0.98 3945 Ibucket, dc, low 18 14 2 bkt o/b 588 3.01 MS#10 R-2800-4 0.0025 29 14 89 7 7 1 0.98 3945 Table A-12: 2003 measurement settings fo r flashes FPL0342 and FPL0345 (8/11/2003).

PAGE 323

293 Measured Parameter LeCroy Oscilloscope Yokogawa Oscilloscope Fiber Optic Cable Sensor PIC RF PIC ID Isobe Cal. Range [A/V] ID Ch. Sampling Rate [MHz] ID Ch. Sampling Rate [MHz] ID Color low / high Length [m] Delay [ s] ID Model V/A ID Att. [dB] Trans. ID Rec. ID Setting [V] IG1 11 1 20 7 5 2 p1a g/b 569 2.91 MS#3 R-7000-10 0.001 5 9 9A R5A R5A 0.1 0.98 554 IG6 11 2 20 7 6 2 p6c y/r 397 2.05 MS#7 R-5600-8 0.00125 2F 19 8D 21 R6A 0.1 0.99 1416 IG10 11 3 20 7 7 2 p10 g/b 349 1.81 MS#2 R-7000-10 0.001 17 27 97 15 R5C 0.1 0.98 4388 IG15 11 4 20 7 8 2 p15 g/b 559 2.86 MS#4 R-7000-10 0.001 6 19 87 R8A R8A 1 0.97 17362 Ibucket, high 12 1 20 7 1 2 bkt w/g 588 3.01 MS#10 R-2800-4 0.0025 1 34 89 23 R1A 1 1.00 40095 IG2 12 3 20 7 3 2 p2 o/b 536 2.75 MS#6 R-5600-8 0.00125 3 14 90 24 R3A 0.1 1.01 809 IG14 12 4 20 7 4 2 p14 gb 541 2.77 MS#5 R-5600-8 0.00125 4 13 80 R4A R4A 1 1.01 7197 IAN2 13 1 20 7 9 2 p2 y/r 536 2.75 110A#4 110A 0.1 0C 29 90 R1B R1B 1 0.99 559 IAN6 13 2 20 7 10 2 p6b o/b 400 2.07 6801#3 3025C 0.025 10 17 8D R2B R2B 1 1.00 564 IAN10, ct 13 4 20 7 12 2 p10 o/b 349 1.81 6801#5 3025C 0.025 16 23 97 12 10 1 1.01 1137 IAN1 14 1 20 7 13 2 p1b o/b 560 2.87 110A#6 110A 0.1 19 16 9A R5B R5B 1 0.98 123 IAN14 14 2 20 7 14 2 p14 o/b 541 2.77 110A#5 110A 0.1 22 37 80 R6B R6B 1 1.00 1410 IA6 14 3 20 7 15 2 p6a o/b 378 1.96 6801#2 3025C 0.025 24 13 8D R7B R7B 1 1.00 358 IA7 14 4 20 7 16 2 p7 o/b 306 1.60 6801#4 3025C 0.025 08 24 94 1 R8B 1 1.01 1276 IC6 15 1 20 18 1 2 p6a y/r 378 1.96 5179#2 3525 0.1 2A 17 8D R1C R1C 1 0.99 140 IB6 15 2 20 18 2 2 p6a g/b 378 1.96 5179#3 3525 0.1 2C 27 8D R2C R2C 1 1.00 449 IB7 15 3 20 18 3 2 p7 r/y 306 1.60 5179#5 3525 0.1 2E 27 94 R3C R3C 1 0.91 407 IC7 15 4 20 18 4 2 p7 g/b 306 1.60 5179#4 3525 0.1 1A 30 94 R4C R4C 1 0.99 623 IBN1 16 1 20 18 5 2 p1b y/r 560 2.87 110A#1 110A 0.1 33 16 9A R7A R7A 1 1.00 126 ICN1 16 2 20 18 6 2 p1b g/b 560 2.87 110A#7 110A 0.1 35 16 9A R6C R6C 1 1.01 127 IBN6 16 3 20 18 7 2 p6b y/r 400 2.07 110A#3 110A 0.1 36 29 8D R7C R7C 1 1.01 567 ICN6 16 4 20 18 8 2 p6b g/b 400 2.07 110A#2 110A 0.1 38 29 8D R8C R8C 1 1.02 574 IN3 17 1 20 18 9 2 p3 o/b 475 2.44 6801#1 3025C 0.025 39 9 92 2 2 1 0.99 222 IN6 17 2 20 18 10 2 p6c o/b 397 2.05 5179#6 3525 0.1 3A 19 8D 20 20 1 0.98 175 IN10 17 3 20 18 11 2 p10 y/r 349 1.81 6801#6 3025C 0.025 3C 19 97 10 4 1 1.00 712 Ibucket, ac, low 18 13 2 bkt o/b 588 3.01 MS#10 R-2800-4 0.0025 29 14 89 7 7 1 0.98 3909 Ibucket, dc, low 18 14 2 bkt o/b 588 3.01 MS#10 R-2800-4 0.0025 29 14 89 7 7 1 0.98 3913 Table A-13: 2003 measurement settings for flashes FPL0347, FPL0348, and FPL0350 (8/15/2003).

PAGE 324

294 A.4 2004 Measurement Settings, Direct Strikes The measurement settings for the strikes to the overhead ground wire between poles 7 and 8 at midspan are found in Table A-14 (FPL0402 and FPL0403 on 7/24/2004). Note that the calibra tion signal for parameter IO10 shows considerable distortion (the rising edge of the square wa ve overshoots), which indicates that the measured parameter IO10 was also distorted by the fiber optic link.

PAGE 325

295 Measured Parameter LeCroy Oscilloscope Yokogawa Oscilloscope Fiber Optic Cable Filter Sensor PIC Isobe Cal. Range [A/V] ID Ch. Sampling Rate [MHz] ID Ch. Sampling Rate [MHz] ID Color Length [m] Delay [ s] ID -3 dB Point [MHz] ID Model V/A ID Att. [dB] Trans. ID Rec. ID Setting [V] IG9 12 2 20 7 1 2 10B bl/gr 225.5 1.19 10 5 LS006 R-7000-10 0.00103 22 13 2 2 1 1.01 8726 IG7 12 3 20 7 2 2 6B br/gr 400.9 2.07 19 5 LS007 R5600-8 0.00127 2C 23 2A 2A 1 1.00 22245 IG8 12 4 20 7 3 2 Tower B bl/o 128.5 0.71 29 5 LS003 R5600-8 0.00126 2F 23 8C 8C 1 1.01 22556 IA6 13 1 20 7 4 2 7 y/r 307 1.60 33 5 80080 3525 0.1004 01 47 8A 8A 1 0.99 4406 IAN6 13 2 20 7 5 2 6C br/gr 397 2.05 23 5 90573 110A 0.101 29 47 7C 7C 1 1.00 4433 IBN6 13 3 20 7 6 2 7 bl/o 303.2 1.58 15 5 90845 110A 0.1016 38 47 4C 4C 1 0.99 4380 IO10 13 4 20 7 7 2 10C bl/o 204 1.09 32 5 80187 3025-C 0.025 39 40 5 5 1 1.07a 8528 Isr, high 14 1 20 18 1 2 Tower B y/r 96.7 0.55 12 5 #8 R5600-8 0.00124 33 26 18 18 1 0.95 30573 Itower, high 14 2 20 18 2 2 Tower A y/r 102.6 0.58 6 5 #9 R5600-8 0.00124 40 26 21 21 1 0.95 30573 IA10 17 1 20 7 8 2 10C br/gr 206.2 1.10 3 5 80081 3525 0.1002 3C 47 11 11 1 1.00 4469 IG6 17 2 20 7 9 2 6C y/r 398 2.06 25 5 LS009 R-2800-4 0.00246 36 20 4A 4A 1 1.00 8130 IG10 17 3 20 7 10 2 10C y/r 202.3 1.08 8 5 LS0010 R-2800-4 0.00254 2A 20 5A 5A 1 1.00 7874 IO6 17 4 20 7 11 2 7 br/gr 308.1 1.61 30 5 80185 3025-C 0.0244 16 40 6C 6C 1 1.00 8197 Isr, low 18 9 2 Tower B br/gr 96.7 0.55 24 5 #8 R5600-8 0.00124 5 9 16 16 1 0.98 4464 Itower, low 18 10 2 Tower A bl/o 107.5 0.60 11 5 #9 R5600-8 0.00124 10 9 6 6 1 0.97 4409 Table A-14: 2004 Measurement settings fo r flashes FPL0402 and FPL0403 (7/24/2004).

PAGE 326

296 APPENDIX B APPENDIX B: DATA PRESENTATION The com plete data sets of the 2003 and 2004 direct strike ex periments (lightning current injected into the phase A of the vertically-configur ed test line) and the 2002 and 2003 nearby strike experiments (lightning curren t injected into ground near the verticallyconfigured test line) are pres ented in this section. Additionally, channel base currents, individual and total ground cu rrents, and individual and total struck-phasetoneutral currents are shown for the 2000, 2002, and 2003 direct strike experiments. The total ground currents and struck-phasetoneutral currents were obtained by summing all individual ground and struck-pha setoneutral currents, respec tively. If an individual current was not available (due to a missi ng or failed measurement) the corresponding individual current due to the symmetry of the test system with respect to the current injection point was used to obtain the total current (e.g., during the 2002 and 2003 experiments the pole 15 phase A-to-neutr al current was not measured and the corresponding current, the pole 1 phase A-to-n eutral current, was used instead to calculate the total phase A-to-neutral current). Note that the indivi dual currents measured during the 2000 and 2002 direct strike e xperiments are also shown in Mata ( 2000) and Mata ( 2003) respectively.

PAGE 327

297 B.1 2000 Data, Direct Strikes In 2000 lightning currents were injected into the phase C of the horizontallyconfigured test distribution line. The 2000 e xperiment is described in detail in Mata ( 2000) and s ummarized in Section 4.3 of this diss erta tion. LeCr oy return stroke data of all channel base currents, individual and to tal ground currents, and individual and total struck-phasetoneutral currents re corded in 2000 are shown on a 100 s time scale. Figure B-1: Stroke FPL0011-1, pha se C-to-neutral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 328

298 Figure B-2: Stroke FPL0011-2, pha se C-to-neutral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-3: Stroke FPL0011-3, phase C to-neutral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 329

299 Figure B-4: Stroke FPL0011-4, pha se C-to-neutral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-5: Stroke FPL0011-5, pha se C-to-neutral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 330

300 Figure B-6: Stroke FPL0014-1, pha se C-to-neutral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-7: Stroke FPL0014-2, pha se C-to-neutral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 331

301 Figure B-8: Stroke FPL0014-3, pha se C-to-neutral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-9: Stroke FPL0018-1, pha se C-to-neutral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 332

302 Figure B-10: Stroke FPL0018-2, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-11: Stroke FPL0018-3, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 333

303 Figure B-12: Stroke FPL0018-4, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-13: Stroke FPL0018-5, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 334

304 Figure B-14: Stroke FPL0018-6, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-15: Stroke FPL0032-1, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 335

305 Figure B-16: Stroke FPL0032-2, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-17: Stroke FPL0032-3, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 336

306 Figure B-18: Stroke FPL0032-4, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-19: Stroke FPL0032-5, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 337

307 Figure B-20: Stroke FPL0032-6, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-21: Stroke FPL0032-7, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 338

308 Figure B-22: Stroke FPL0033-1, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-23: Stroke FPL0034-1, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 339

309 Figure B-24: Stroke FPL0034-2, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-25: Stroke FPL0034-3, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 340

310 Figure B-26: Stroke FPL0034-4, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-27: Stroke FPL0034-5, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 341

311 Figure B-28: Stroke FPL0036-1, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-29: Stroke FPL0036-2, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 342

312 Figure B-30: Stroke FPL0036-3, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-31: Stroke FPL0036-4, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 343

313 Figure B-32: Stroke FPL0036-5, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-33: Stroke FPL0037-1, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 344

314 Figure B-34: Stroke FPL0037-2, phase C-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). B.2 2002 Data, Direct Strikes In 2002 lightning currents were injected into the phase A of the verticallyconfigured test distribution line. The 2002 e xperiment is described in detail in Mata ( 2003) and s ummarized in Section 4.5 of this diss erta tion. LeCr oy return stroke data of all channel base currents, individual and to tal ground currents, and individual and total struck-phasetoneutral currents re corded in 2002 are shown on a 100 s time scale with the exception of data from stroke FPL0226-6 that includes an unusual slow return stroke current Data from this return stroke are shown on a 1 ms time scale.

PAGE 345

315 Figure B-35: Stroke FPL0208-1, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-36: Stroke FPL0210-1, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 346

316 Figure B-37: Stroke FPL0218-1, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-38: Stroke FPL0219-2, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 347

317 Figure B-39: Stroke FPL0219-3, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-40: Stroke FPL0220-1, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 348

318 Figure B-41: Stroke FPL0220-2, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-42: Stroke FPL0220-3, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 349

319 Figure B-43: Stroke FPL0220-4, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-44: Stroke FPL0220-5, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 350

320 Figure B-45: Stroke FPL0221-2, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-46: Stroke FPL0221-3, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 351

321 Figure B-47: Stroke FPL0221-4, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-48: Stroke FPL0221-5, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 352

322 Figure B-49: Stroke FPL0221-6, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-50: Stroke FPL0226-1, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 353

323 Figure B-51: Stroke FPL0226-2, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-52: Stroke FPL0226-3, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 354

324 Figure B-53: Stroke FPL0226-4, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-54: Stroke FPL0226-5, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 355

325 Figure B-55: Stroke FPL0226-6, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right), 1 ms time scale. Figure B-56: Stroke FPL0228-1, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 356

326 Figure B-57: Stroke FPL0228-2, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-58: Stroke FPL0228-3, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 357

327 Figure B-59: Stroke FPL0228-4, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-60: Stroke FPL0228-5, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 358

328 Figure B-61: Stroke FPL0228-6, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-62: Stroke FPL0229-1, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 359

329 Figure B-63: Stroke FPL0229-2, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-64: Stroke FPL0229-3, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 360

330 Figure B-65: Stroke FPL0229-4, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-66: Stroke FPL0229-5, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 361

331 Figure B-67: Stroke FPL0229-6, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-68: Stroke FPL0229-7, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 362

332 Figure B-69: Stroke FPL0229-8, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right). Figure B-70: Stroke FPL0229-9, phase A-to-neu tral currents (top, left), ground currents (top, right), and the respec tive sum of the individual currents overlaid with the channel base current (bottom, left and right).

PAGE 363

333 B.3 2003 Data, Direct Strikes In 2003 lightning currents were injected into the phase A of the verticallyconfigured test distribution line. The 2003 experiment is de scribed in detail in Section 4.6. All LeCroy return stroke data recorded in 2003 are show n on a 100 s time scale. Additionally, the total ground and phase-to-neutral currents (the sum s of the individual ground and phase-to-neutral currents, respectiv ely) are shown and compared with the lightning currents.

PAGE 364

334 Figure B-71: Data matrix for st roke FPL0301-1 (direct strike), 100 s time windows.

PAGE 365

335 Figure B-72: Data matrix for st roke FPL0301-2 (direct strike), 100 s time windows.

PAGE 366

336 Figure B-73: Data matrix for st roke FPL0301-3 (direct strike), 100 s time windows.

PAGE 367

337 Figure B-74: Data matrix for st roke FPL0305-1 (direct strike), 100 s time windows.

PAGE 368

338 Figure B-75: Data matrix for st roke FPL0305-2 (direct strike), 100 s time windows.

PAGE 369

339 Figure B-76: Data matrix for st roke FPL0305-3 (direct strike), 100 s time windows.

PAGE 370

340 Figure B-77: Data matrix for st roke FPL0305-4 (direct strike), 100 s time windows.

PAGE 371

341 Figure B-78: Data matrix for st roke FPL0312-1 (direct strike), 100 s time windows.

PAGE 372

342 Figure B-79: Data matrix for st roke FPL0312-2 (direct strike), 100 s time windows.

PAGE 373

343 Figure B-80: Data matrix for st roke FPL0312-3 (direct strike), 100 s time windows.

PAGE 374

344 Figure B-81: Data matrix for st roke FPL0312-4 (direct strike), 100 s time windows.

PAGE 375

345 Figure B-82: Data matrix for st roke FPL0312-5 (direct strike), 100 s time windows.

PAGE 376

346 Figure B-83: Data matrix for st roke FPL0312-6 (direct strike), 100 s time windows.

PAGE 377

347 Figure B-84: Data matrix for st roke FPL0312-7 (direct strike), 100 s time windows.

PAGE 378

348 Figure B-85: Data matrix for st roke FPL0312-8 (direct strike), 100 s time windows.

PAGE 379

349 Figure B-86: Data matrix for st roke FPL0312-9 (direct strike), 100 s time windows.

PAGE 380

350 Figure B-87: Data matrix for st roke FPL0312-10 (direct strike), 100 s time windows.

PAGE 381

351 Figure B-88: Data matrix for st roke FPL0315-1 (direct strike), 100 s time windows.

PAGE 382

352 Figure B-89: Data matrix for st roke FPL0315-2 (direct strike), 100 s time windows.

PAGE 383

353 Figure B-90: Data matrix for st roke FPL0317-1 (direct strike), 100 s time windows.

PAGE 384

354 B.4 2004 Data, Direct Strikes In 2004 lightning currents were injected into the overhead ground wire of the vertically-configured test distribution line. The complete data set for all measured currents is shown in this section. Figure B-91 shows the complete data for the wireburn FPL0402 on a 600 s tim e scale. and Figure B-92 and Figure B-93 show the complete data for the first and second return strokes, respectively, of the tw o stroke flash FPL 0403 on 2 m s (blue waveforms) and 100 s (pink waveform s) time scales. All data shown on 600 s and 2 ms time scales are 12-bit Yokogawa data sampled at 2 MHz. All data shown on 100 s time scales are 8bit LeCroy data sampled at 20 MHz, except the phase A and ground currents at pole 10 and the overhead ground wire and ground currents at pole 6, which were recorded with the Yokogawa oscilloscopes. LeCroy data are not available for these measurements due to an instrumentation failure.

PAGE 385

355 Figure B-91: Data for wireburn FPL0402, 600 s time scale.

PAGE 386

356 Figure B-92: Data for st roke FPL0403-1, 2 ms and 100 s time scales.

PAGE 387

357 Figure B-93: Data for st roke FPL0403-2, 2 ms and 100 s time scales.

PAGE 388

358 B.5 2005 Data, Induced Currents Figure B-94: Measured channel base curr ent and induced currents on the runway counterpoise (290 m from th e lightning) and the verti cal wire (300 m from the lightning) for flash 0501, stroke 1 duri ng the mobile launcher experiment. Figure B-95: Measured channel base curr ent and induced currents on the runway counterpoise (290 m from th e lightning) and the verti cal wire (300 m from the lightning) for flash 0503, stroke 1 duri ng the mobile launcher experiment.

PAGE 389

359 Figure B-96: Measured channel base curr ent and induced currents on the runway counterpoise (290 m from th e lightning) and the verti cal wire (300 m from the lightning) for flash 0503, stroke 2 duri ng the mobile launcher experiment. Figure B-97: Measured channel base curr ent and induced currents on the runway counterpoise (290 m from th e lightning) and the verti cal wire (300 m from the lightning) for flash 0503, stroke 3 duri ng the mobile launcher experiment.

PAGE 390

360 Figure B-98: Measured channel base curr ent and induced currents on the runway counterpoise (290 m from th e lightning) and the verti cal wire (300 m from the lightning) for flash 0503, stroke 4 duri ng the mobile launcher experiment. Figure B-99: Measured channel base curr ent and induced curre nt on the runway counterpoise (50 m from the lightning) for flash 0510, stroke 1 during the tower launcher experiment.

PAGE 391

361 Figure B-100: Measured channel base cu rrent and induced cu rrent on the runway counterpoise (50 m from the lightning) for flash 0512, stroke 1 during the tower launcher experiment. Figure B-101: Measured channel base cu rrent and induced cu rrent on the runway counterpoise (50 m from the lightning) for flash 0512, stroke 2 during the tower launcher experiment.

PAGE 392

362 Figure B-102: Measured channel base current and induced currents on the runway counterpoise (50 m from th e lightning) and the vertic al wire (100 m from the lightning) for flash 0517, stroke 1 du ring the tower launcher experiment. Figure B-103: Measured channel base current and induced currents on the runway counterpoise (50 m from th e lightning) and the vertic al wire (100 m from the lightning) for flash 0517, stroke 2 du ring the tower launcher experiment.

PAGE 393

363 Figure B-104: Measured channel base current and induced currents on the runway counterpoise (50 m from th e lightning) and the vertic al wire (100 m from the lightning) for flash 0520, stroke 1 du ring the tower launcher experiment. Figure B-105: Measured channel base current and induced currents on the runway counterpoise (50 m from th e lightning) and the vertic al wire (100 m from the lightning) for flash 0521, stroke 1 du ring the tower launcher experiment.

PAGE 394

364 Figure B-106: Currents during natural lightning flash MSE0504, stroke 1. Figure B-107: Currents during natural lightning flash MSE0504, stroke 2.

PAGE 395

365 APPENDIX C APPENDIX C: RETURN STROKE CURRENT PARAMETERS C.1 1999 Experiment Table C-1: Return stroke st atistics for the 1999 experiment. Date Flash ID Return Stroke Current Comments ID Peak [kA] 10-90% Risetime [ s] Half-Peak Width [ s] Charge, 1 ms [C] 8/16/99 UF9904 1 15.1 1.6 36 2.5 Stroke 4 :Peak current from hard copy of tape data. 2 8.4 0.8 18 0.5 3 23.1 1.3 37 2.5 4 2.8 5 16.7 0.9 27 1.6 6 7.9 0.8 15 0.4 8/24/99 UF9911 1 11.0 1.9 38 1.5 No Nicolet data. Positve ICC. 2 11.3 2.0 36 1.8 UF9912 1 7.1 1.4 21 0.7 2 7.5 1.1 20 0.5 9/6/99 UF9914 1 3.1 2 7.6 No LeCroy and Nicolet data. UF9915 1 6.7 Peak currents from hard copy of tape data. 2 2.9 3 10.1 9/10/99 UF9916 1 9.4 2 14.7 3 9.3 4 6.2 LeCroy data files not found 5 13.2 6 10.3 Peak currents from hard copy of tape data 7 16.6 UF9917 1 13.6 2 8.6 3 11.9 4 7.8 Sample Size 26 9 9 9 Minimum 2.8 0.8 15 0.4 Maximum 23.1 2.0 38 2.5 Arithmetic Mean 10.1 1.3 28 1.3 Standard Deviation 4.7 0.5 9 0.8 Geometric Mean 9.0 1.2 26 1.1 Standard Deviation, log 0.23 0.15 0.16 0.32

PAGE 396

366 C.2 2000 Experiment Table C-2: Return stroke st atistics for the 2000 experiment. Date Flash ID Return Stroke Current Comments ID Peak [kA] 10-90% Risetime [ s] Half-Peak Width [ s] Charge, 1 ms [C] 7/16/00 FPL0011a 1 12.9 85 2.2 Stroke 4 : Saturated at 24 kA. 2 20.8 90 4.7 3 14.6 13 2.3 4 5 9.5 8 0.7 FPL0014b 1 8.3 17 0.6 2 18.9 26 1.6 3 10.0 8 0.5 7/20/00 FPL0018b 1 10.1 8 0.7 2 6.6 15 0.3 3 5.6 7 0.4 4 7.4 7 0.5 5 5.3 7 0.5 6 9.3 7 0.7 8/2/00 FPL0032a 1 8.3 22 0.3 2 8.7 20 0.8 3 6.3 13 0.3 4 14.1 18 1.4 5 12.9 14 1.1 6 7.3 17 0.3 7 6.9 19 0.4 FPL0033a 1 42.3 50 7.6 FPL0034b 1 21.8 38 2.6 2 14.8 25 1.0 3 14.5 25 1.0 4 14.5 13 1.1 5 20.7 13 2.3 8/3/00 FPL0036b 1 20.4 57 2.8 2 20.7 36 3.0 3 18.7 41 2.0 4 5.8 18 0.3 5 6.8 18 0.4 8/6/00 FPL0037b 1 36.1 27 8.3 2 15.8 12 1.3 Sample Size 33 33 33 Minimum 5.3 7 0.3 Maximum 42.3 90 8.3 Arithmetic Mean 13.8 24 1.6 Standard Deviation 8.4 20 1.9 Geometric Mean 11.9 18 1.0 Standard Deviation, log 0.23 0.27 0.41 a Return stroke currents exhibit ringing. The ringing has a magnitude of a few kA to a few 10s of kA a frequency of a few 10s to a few 100s of nanoseconds, starts a few microseconds before the return stroke, and ends a few microseconds after the return stroke initiation. b Return stroke currents exhibit a double peak or step on the rising edge. Slight ringing is also p resent on the risin g ed g e.

PAGE 397

367 C.3 2001 Experiment Table C-3: Return stroke st atistics for the 2001 experiment. Date Flash ID Return Stroke Current Comments ID Peak [kA] 10-90% Risetime [ s] Half-Peak Width [ s] Charge, 1 ms [C] 7/27/01 FPL0107 1 11.7 0.9 29 1.4 2 10.7 1.3 24 1.3 FPL0108 1 21.1 1.3 42 1.9 2 21.7 1.4 32 2.3 3 23.7 1.5 18 7.3 4 15.9 1.0 28 2.2 5 16.6 1.3 26 1.4 8/18/01 FPL0110 1 9.7 1.6 24 0.7 FPL0112 1 28.2 1.5 58 4.3 One stroke recorded by Yokogawa and not by LeCroy. It was assumed that this stroke was stroke 6. 2 19.9 1.4 33 2.4 3 16.9 1.5 35 1.7 4 15.7 1.7 16 2.7 5 6.0 1.9 17 0.5 6 Sample Size 13 13 13 13 Minimum 6.0 0.9 16 0.5 Maximum 28.2 1.9 58 7.3 Arithmetic Mean 16.7 1.4 29 2.3 Standard Deviation 6.2 0.3 11 1.8 Geometric Mean 15.5 1.4 28 1.8 Standard Deviation, log 0.18 0.09 0.16 0.30

PAGE 398

368 C.4 2002 Experiment Table C-4: Return stroke statistics for the 2002 direct strike experiment. Date Flash ID Return Stroke Current Comments ID Injection Point Peak [kA] 10-90% Risetime [ s] Half-Peak Width [ s] Charge, 1 ms [C] Line Total Line Total Line Total Line Total 7/9/02 FPL0205 1 IS1 2 IS1 Altitude Trigger terminated on IS1.3 IS1 No LeCroy and Yokogawa data. 4 IS1 FPL0208 1 line Return Stroke currents have an initial spike followed by a hump. Waveshape does not match sum of ground currents, which suggest an error in the return stroke current data. No Yokogawa Scope 7 data. LeCroy Data only for first strokes. 2 tower/line 3 tower/line FPL0210 1 line 2 line 3 line 4 line 5 tower/line 6 line 7 line 8 tower/line 9 tower/line 7/19/02 FPL0213 1d tower 20.0 0.7 57 3.7 All strokes : No current into the line. Stroke 2 : No LeCroy data. 2d tower 14.8 40 2.1 3a,d tower 22.4 0.4 93 4.1 7/20/02 FPL0218 1 tower/line 13.2 19.5 2.0 2.0 41 32 1.2 1.8 FPL0219 1d tower 8.4 91 1.3 Stroke 1 : No Current into the line and no LeCroy data. 2 tower/line 14.6 21.2 2.6 0.5 51 21 2.3 3.4 3a line 22.4 22.4 1.6 1.6 9 9 2.2 2.2 FPL0220 1 tower/line 11.0 13.8 1.9 0.4 23 23 0.8 1.0 Stroke 7 : No LeCroy data. 2b line 18.7 18.7 1.5 1.5 17 17 1.1 1.1 3d line 17.5 17.5 1.3 1.3 8 8 1.3 1.3 4a line 14.1 14.1 1.5 1.5 6 6 0.8 0.8 5 line 7.0 7.0 1.9 1.9 14 14 0.4 0.4 6b line 18.4 18.4 1.6 1.6 14 14 1.0 1.0 7b line 16.1 16.1 14 14 0.9 0.9 FPL0221 1d tower 5.6 45 0.9 Strokes 1,7,8,9,10,11: No LeCroy data. 2 tower/line 10.4 17.3 1.9 0.4 42 26 1.3 2.0 3b tower/line 11.7 13.8 0.5 0.5 20 14 0.7 1.0 4a line 12.3 12.3 0.4 0.4 33 33 0.8 0.8 5 tower/line 8.5 10.0 1.7 0.9 9 8 0.3 0.4 6a tower/line 23.2 26.5 1.8 0.9 45 41 2.2 2.7 7a tower/line 13.2 19.7 22 20 0.9 1.4 8 tower/line 19.2 23.7 41 31 0.2 2.8 9b tower/line 13.3 18.6 47 31 1.4 1.8 10a line 19.3 19.3 13 13 1.8 1.8 11 line 7.4 7.4 6 6 0.3 0.3 7/25/02 FPL0226 1a line 28.0 28.0 2.5 2.5 33 33 3.8 3.8 Strokes 6,7,8: No LeCroy data. 2 line 10.4 10.4 1.5 1.5 11 11 0.6 0.6 3 line 6.4 6.4 1.6 1.6 6 6 0.5 0.5 4 line 9.4 9.4 1.4 1.4 13 13 0.4 0.4 5b line 27.3 27.3 1.2 1.2 14 14 1.9 1.9 6 line 8.2 8.2 9 9 0.5 0.5 7 line 15.3 15.3 21 21 0.8 0.8 8 line 15.9 15.9 8 8 1.1 1.1 8/2/02 FPL0228 1 tower/line 21.2 28.7 0.7 55 42 2.3 2.7 Stroke 6 : No LeCroy data for Isr,low, Itower,low, Itower,high, and IAN14. 2 line 13.9 13.9 1.3 1.3 18 18 0.9 0.9 3 line 9.1 9.1 1.4 1.4 7 7 0.4 0.4 4 line 25.1 25.1 1.5 1.5 19 19 2.4 2.4 5 line 33.7 33.7 1.3 1.3 31 31 3.1 3.1 6 line 21.4 21.4 1.2 1.2 10 10 2.2 2.2 FPL0229 1 tower/line 13.5 20.3 0.5 74 49 1.8 1.9 Strokes 8,9: No LeCroy data for Isr,low, Itower,low, Itower,high, and IAN14. 2 line 9.4 9.4 2.0 2.0 28 28 0.6 0.6 3c line 20.6 20.6 1.5 1.5 25 25 1.3 1.3 4 line 12.9 12.9 1.5 1.5 18 18 0.7 0.7 5 line 6.8 6.8 6.8 6.8 39 39 1.0 1.0 6 line 6.7 6.7 2.7 2.7 18 18 0.3 0.3 7 line 9.7 9.7 1.5 1.5 12 12 1.2 1.2 8 line 7.9 7.9 2.1 2.1 21 21 0.5 0.5 9c line 27.1 27.1 1.5 1.5 32 32 1.9 1.9 Sample Size 43 48 32 36 43 48 43 48 Minimum 6.4 5.6 0.4 0.4 6 6 0.2 0.3 Maximum 33.7 33.7 6.8 6.8 74 93 3.8 4.1 Arithmetic Mean 15.1 16.3 1.8 1.4 23 25 1.2 1.5 Standard Deviation 6.7 7.1 1.0 1.1 16 19 0.8 1.0 Geometric Mean 13.8 14.7 1.6 1.2 19 20 1.0 1.2 Standard Deviation, log 0.19 0.21 0.21 0.27 0.29 0.30 0.31 0.32 a Large hump (over 1 kA) in Itower,high that does not appear in Itower,low. b Small hump (under 1 kA) in Itower,high that does not appear in Itower,low. c Small hump (under 1 kA) in Itower,low that does not appear in Itower,high. d Small hump (under 1 kA) in Isr,high that does not appear in Isr,low.

PAGE 399

369 Table C-5: Return stroke statistics for the 2002 nearby strike experiment. Date Flash ID Return Stroke Current Comment ID Peak [kA] 10-90% Risetime [ s] Half-Peak Width [ s] Charge, 1 ms [C] 8/18/02 FPL0236 1 8.9 0.4 29 0.6 No Yokogawa data. Part of Isr,high larger than Isr,low. 9/13/02 FPL0245 1 0.7 2 0.5 3 4.3 21 0.3 4 0.5 All strokes : No Ihigh data and no LeCroy data. 5 1.1 6 4.4 19 0.3 7 0.4 8 0.5 9 0.4 10 0.4 FPL0246 1 No lightning current data. Only Yokogawa scope 7 data available. 2 Sample Size 3 1 3 11 Minimum 4.3 0.4 19 0.3 Maximum 8.9 0.4 29 1.1 Arithmetic Mean 5.8 0.4 23 0.5 Standard Deviation 2.7 5 0.2 Geometric Mean 5.5 0.4 23 0.5 Standard Deviation, log 0.18 0.10 0.18 All strokes except strokes 3,6: Charge calculated from saturated Ibucket,low.

PAGE 400

370 C.5 2003 Experiment Table C-6: Return stroke statistics for the 2003 direct strike experiment. Date Flash ID Return Stroke Current Comments ID Peak [kA] 10-90% Risetime [ s] Half-Peak Width [ s] Charge, 1 ms [C] 6/30/03 FPL0301 1 All return stroke currents exhibit ringing. The ringing has a magnitude of a few 10s of kA, a frequency of a few 100s of nanoseconds, starts a few microseconds before the return stroke, and ends a few microseconds after the return stroke initiation. 2 3 7/6/03 FPL0305 1a,c Stroke 1 : Split between tower and line. 2a,c 11.3 3c 10.6 6 0.5 4b,c 12.3 7/13/03 FPL0312 1c 16.0 11 0.9 2c 5.8 4 0.3 3c 12.5 10 0.7 4c 7.3 4 0.4 5c 8.9 5 0.4 6c 10.4 10 0.5 7c 10.4 6 0.5 Strokes 11 16: No LeCroy data. 8b,c 15.3 Strokes 1, 3, 6, and 10 : Ringing similar to the one observed for strokes of FPL0301 except much lower magnitude (a few kA). 9c 10.9 11 0.6 10c 12.2 5 1.0 11c 14.8 16 0.8 12c 9.9 10 0.5 13b,c 13.8 14c 9.8 8 0.5 15b,c 15.0 16c 14.8 8 0.9 7/14/03 FPL0315 1 11.7 1.6 20 1.1 2 14.6 1.2 16 0.8 FPL0317 1 27.9 0.5 17 6.0 Sample Size 22 3 17 17 Minimum 5.8 0.5 4 0.3 Maximum 27.9 1.6 20 6.0 Arithmetic Mean 12.5 1.1 10 1.0 Standard Deviation 4.3 0.6 5 1.3 Geometric Mean 12.0 1.0 9 0.7 Standard Deviation, log 0.14 0.27 0.22 0.29 a Large negative hump (under -1 kA) in Isr,high that does not appear in Isr,low. b Large positive hump (over 1 kA) in Isr,high that does not appear in Isr,low. c Double peak or step during the rising edge.

PAGE 401

371 Table C-7: Return stroke statistics for the 2003 nearby strike experiment. Date Flash ID Return Stroke Current Comment ID Peak [kA] 10-90% Risetime [ s] Half-Peak Width [ s] Charge, 1 ms [C] 7/18/03 FPL0321 1 8.3 1.0 16 0.6 Stroke 2 : Unusual current waveshape possibly due to arcing between launcher and line. Stroke 3 : Ilow between 15% and 30% smaller than Ihigh. Stroke 4 : Ilow between 20% and 55% smaller than Ihigh. 2 3 18.9 0.3 21 2.7 4 14.4 0.2 15 1.8 7/22/03 FPL0329 1 12.7 1.1 33 1.0 2 4.7 1.6 25 0.5 3 22.7 0.3 18 2.3 4 15.4 0.2 65 1.8 FPL0331 1 Stroke 1 : Unusual current waveshape possibly due to arcing between launcher and line. Stroke 2 : Positive polarity (bipolar flash). 2 5.0 1.1 8/2/03 FPL0336 1 28.7 0.3 56 2.6 Stroke 2 : Ilow between 20% and 35% smaller than Ihigh. Stroke 5 : Ilow between 15% and 40% smaller than Ihigh. Stroke 7 : Ilow between 30% and 65% smaller than Ihigh. 2 16.2 0.2 17 2.5 3 5.8 0.4 7 0.4 4 13.5 0.2 19 0.8 5 19.4 0.2 23 1.9 6 19.7 0.2 13 1.5 7 13.8 0.2 7 2.0 8/7/03 FPL0341 1 No LeCroy data. No Ihigh. 8/15/03 FPL0347 1 20.5 0.3 54 1.8 2 6.6 0.6 10 0.4 FPL0350 1 8.9 1.4 19 0.6 Sample Size 18 17 17 18 Minimum 4.7 0.2 7 0.4 Maximum 28.7 1.6 65 2.7 Arithmetic Mean 14.2 0.5 25 1.5 Standard Deviation 6.8 0.5 17 0.8 Geometric Mean 12.5 0.4 20 1.2 Standard Deviation, log 0.24 0.32 0.29 0.30

PAGE 402

372 C.6 2004 Experiment Table C-8: Return stroke st atistics for the 2004 experiment. Date Flash ID Return Stroke Current ID Peak [kA] 10-90% Risetime [ s] Half-Peak Width [ s] Charge, 1 ms [C] 7/24/04 FPL0403 1 15.9 0.8 42 2.1 2 5.9 0.9 17 0.5 Sample Size 2 2 2 2 Minimum 5.9 0.8 17 0.5 Maximum 15.9 0.9 42 2.1 Arithmetic Mean 10.9 0.8 30 1.3 Standard Deviation 7.0 0.1 18 1.1 Geometric Mean 9.7 0.8 27 1.0 Standard Deviation, lo g 0.300.060.280.44

PAGE 403

373 APPENDIX D APPENDIX D: FLASHOVERS AND DISCONNECTOR OPERATION Flashovers are iden tified for each stroke a nd disconnector operations are identified for each flash during the 2000 through 2003 expe riments. A discussion of the flashovers and disconnector operations during e ach year can be found in Section 6.3. Table D-1, Table D-2, Table D-3, and Table D-4 include information about the triggering date, the stroke ID (com posed of the fl ash ID and the stroke number), and the return stroke current peak value. Note that WB stands for W ir e Burn and indicates flashes without return strokes (composed of th e initial stage only). Table D-1, Table D-2, and Table D-3 contain inform ation about which arresters had activ ated disconnectors. Disconnector operation was identified after each trigge ring d ay by a blown safety device at the bottom of the arrester which disconnected the lead betw een the arrester and neutral connection. Additionally, Table D-1, Table D-3, and Table D-4 contain information on the percentage of the charg e deficit at 90 s for the struck-phasetone utral and neutral-to-ground connections and the percen tage of charge at 90 s in the non-struck-phase conductor and the non-struck-phasetoneutral conn ection. The charge deficit at 90 s for the s truckphasetoneutral connections is the difference of charge injected in to the struck phase (the return stroke current nu merically integrated over a 90 s time interval) and the total charge transferred from the struck-phase c onductor to the neutral (t he integrated total struck-phasetoneutral current) relativ e to the injected charge at 90 s for the 2000, 2002, and 2003 experiments, respecti vely. The charge deficit at 90 s for the neutral-to-

PAGE 404

374 ground connections was calculate d in a similar fashion. Th e currents in phase B were measured at pole 9 (column B9 in the table) for the horizontal line experiment and at poles 6 and 7 (B6 and B7) for the vertical line experiments. The phase B arrester currents were measured at pole 8 (B8 to N) for the hor izontal line experiment and at pole 6 (B6 to N) for the vertical line experiments. Table D-1, Table D-3, and Table D-4 show the charge trans fer within 90 s at thes e measurement points (the phase B currents and phase B arrester currents integrated over 90 s) relative to the tota l injected charge at 90 s. The data presented in Table D-1, Table D-3, and Table D-4 are used to identify flashovers and their locations as follow: A large charge deficit for the struckphasetoneutral current and not for the neutral-to-ground current is indicative for current bypassing the struck-phaseto neutral measurement points via a flashover. Note that a missing struck-phaseto neutral current was substituted by the corre sponding current due to the symmetry of the system to calculate the total struck -phasetoneutral current. This symmetry assumption does not apply if the disconnector of a struck-phase was activated and the disconnector of the corresponding arrest er was not activated, as it was often the case during the 2000 experiment (the disconnector of the st ruck-phase arrester at pole 8 was commonly activated and the disc onnector of the corresponding arrester at pole 11 never operated). The large char ge deficit during the 2000 experiment for strokes with a failed pole 8 arrester is likel y not due to flashovers but to errors in the calculation of the total struck-phase toneutral current caused by not meeting the symmetry assumption. A large fraction of the total charge present in a non-struck phase is indicative for a flashover. A large fraction of the total charge pr esent in a non-stru ck-phasetoneutral connection is indicative for a flashover. The polarity of the charge percentages was used to identify flashover locations. The cells in Table D-1 through Table D-4 highlighted in yellow indicate flashovers and the inf orm ation used to iden tify them. Note that the currents of the 2001 experiment ( Table D-2) were examined for flashovers by analyzing the hard-copies of the nonstruck -phase current s.

PAGE 405

375 D.1 2000 Experiment Table D-1: Flashovers and disconnector operation dur ing the 2000 horizontal line experiment. Date Stroke ID Peak Current [kA] Charge Deficit, 90 s [%] Charge, 90 s [%] Phase C to B Flashover Damaged Arrester Injected C to N N to Ground B9 B8 to N Between p8 & p11 After p8 7/16/00 FPL0011 1 12.9 -4.6 -11.8 0.0 0.0 NO NO none 2 20.8 -22.7 2.4 >0.9 0.1 YES NO C8 3 14.6 -53.9 -12.2 -0.1 -0.2 NO NO 4 >24.1 >-49.0 >10.4 <0.1 <0.1 NO NO 5 9.5 -59.3 -11.8 0.0 -0.1 NO NO FPL0014 1 8.3 -13.0 0.8 0.2 YES NO 2 18.9 -7.9 0.0 -0.1 NO NO 3 10.0 -1.4 -0.1 47.9 NO YES 7/20/00 FPL0018 1 10.1 13.4 -1.8 -13.5 18.0 NO YES none 2 6.6 13.6 1.8 -13.7 17.5 NO YES 3 5.6 82.4 -1.3 -0.9 0.0 NO YES C8 4 7.4 79.3 1.6 -0.1 -0.1 NO NO 5 5.3 85.9 4.2 -0.1 -0.2 NO NO 6 9.3 78.9 2.4 0.0 -0.1 NO NO 7/28/00 FPL0025 WB 8/2/00 FPL0032 1 8.3 8.5 34.2 -0.1 -0.1 NO NO none 2 8.7 10.0 5.9 -0.1 -0.2 NO NO 3 6.3 10.7 3.0 0.0 0.1 NO NO 4 14.1 6.2 5.0 0.0 -0.2 NO NO 5 12.9 3.0 1.9 0.0 -0.2 NO NO 6 7.3 10.6 11.9 -0.1 -0.2 NO NO 7 6.9 15.6 11.8 -0.1 -0.1 NO NO FPL0033 1 42.3 21.0 17.8 >6.5 >2.5 YES ? FPL0034 1 21.8 63.3 2.5 0.0 -0.4 NO NO C8 2 14.8 72.5 4.1 -0.1 -0.3 NO NO 3 14.5 74.6 5.5 -0.1 -0.3 NO NO 4 14.5 76.0 3.0 -0.2 -0.3 NO NO 5 20.7 70.1 5.8 0.1 0.1 NO NO 8/3/00 FPL0035 WB none FPL0036 1 20.4 -3.7 8.2 0.1 -0.2 NO NO 2 20.7 -1.8 7.1 0.2 0.1 NO NO 3 18.7 4.1 8.9 0.2 0.1 NO NO 4 5.8 7.7 7.8 -0.1 -0.2 NO NO 5 6.8 0.1 1.5 -0.1 -0.2 NO NO 6 C8 7 8 8/6/00 FPL0037 1 36.1 70.9 14.1 >3.0 0.4 YES NO C8 2 15.8 80.813.50.10.0 NO NO

PAGE 406

376 D.2 2001 Experiment Table D-2: Flashovers and disconnector operation dur ing the 2001 vertical line experiment. Date Stroke ID Peak Current [kA] Phase A to B Flashover Damaged Arrester Injected Between p6 & p10 After p10 7/26/01 FPL0101 WB none FPL0102 WB 7/27/01 FPL0105 WB 2A, 6A FPL0107 1 11.7 YES 2 10.7 YES FPL0108 1 21.1 YES 2 21.7 YES 3 23.7 YES 4 15.9 YES 5 16.6 YES 8/18/01 FPL0110 1 9.7 NO YES 2A, 10A, 14A FPL0111 WB FPL0112 1 28.2 YES 2 19.9 YES 3 16.9 YES 4 15.7 YES 5 6.0 NO NO 6

PAGE 407

377 D.3 2002 Experiment Table D-3: Flashovers and disconnector operation dur ing the 2002 vertical line experiment. Date Stroke ID Peak Current [kA] Charge Deficit, 90 s [%] Charge, 90 s [%] Phase A to B Flashover Damaged Arrester Injected A to N N to Ground B7 B6 B6 to N Between p6 & p10 After p6 7/9/02 FPL0208 (3 strokes) none FPL0210 (9 strokes) 7/20/02 FPL0218 1 13.2 43.0 13.3 -7.7 <<-2.5 8.7 NO YES A2, A6, A10, A14 FPL0219 2 14.6 40.9 16.0 -5.1 <<-2.2 6.1 NO YES 3 22.4 40.0 13.3 <<-2.5 >6.1 YES YES FPL0220 1 11.0 51.7 18.4 -5.5 <<-3.7 4.4 NO YES 2 18.7 32.6 10.2 <<-2.8 >11.7 YES YES 3 17.5 29.7 11.1 <<-3.4 >7.6 YES YES 4 14.1 37.5 15.7 <<-15.1 10.1 YES YES 5 7.0 31.5 11.5 -7.5 <<-7.9 0.1 NO YES 6 18.4 32.0 8.7 <<-2.9 >8.3 YES YES 7 16.1 <<-3.9 12.3 YES YES FPL0221 2 10.4 34.8 20.1 -30.3 <<-3.1 9.9 NO YES 3 11.7 33.8 15.0 -12.3 <<-4.4 10.3 NO YES 4 12.3 29.4 14.6 -6.4 <<-3.4 7.2 NO YES 5 8.5 42.7 20.0 -13.9 <<-8.2 8.5 NO YES 6 23.2 41.9 12.5 <<-1.5 11.5 YES YES 7 13.2 <<-4.2 11.5 YES YES 8 19.2 <<-2.1 7.7 YES YES 9 13.3 <<-2.9 12.5 YES YES 10 19.3 <<-3.0 8.1 YES YES 11 7.4 -0.8 <<-11.4 9.4 NO YES 7/25/02 FPL0226 1 28.0 16.3 4.8 10.1 7.0 2.0 YES NO A2, A6, A10 2 10.4 17.7 6.9 6.0 6.0 1.8 YES NO 3 6.4 1.8 2.0 0.4 0.0 0.2 NO NO 4 9.4 10.6 7.0 4.3 3.3 0.5 YES NO 5 27.3 22.6 4.6 12.9 8.0 3.2 YES NO 6 8.2 11.0 6.4 3.4 YES NO 7 15.3 20.0 10.2 7.6 YES NO 8 15.9 17.7 9.0 6.6 YES NO 8/2/02 FPL0228 1 21.2 21.0 2.5 11.1 6.8 2.9 YES NO A2 2 13.9 22.9 -0.7 10.2 5.5 3.6 YES NO 3 9.1 16.9 1.1 5.7 3.7 1.3 YES NO 4 25.1 26.7 -3.3 11.8 6.5 3.7 YES NO 5 33.7 42.2 6.4 12.1 5.3 5.2 YES NO 6 21.4 <27.3 -1.4 3.3 1.2 1.8 YES NO FPL0229 1 13.5 5.0 1.6 0.3 0.1 0.3 NO NO 2 9.4 20.2 1.9 0.3 0.0 0.0 NO NO 3 20.6 25.6 -0.3 12.7 7.2 3.5 YES NO 4 12.9 22.9 2.2 10.4 5.6 2.9 YES NO 5 6.8 46.9 8.1 16.4 7.7 5.9 YES NO 6 6.7 -3.0 0.6 0.0 0.1 0.1 NO NO 7 9.7 14.6 -2.0 7.9 4.0 1.8 YES NO 8 7.9 17.3 1.5 6.2 3.8 1.5 YES NO 9 27.1 <31.9-0.9 17.2 8.6 >6.0 YESNO

PAGE 408

378 D.4 2003 Experiment Table D-4: Flashovers during the 2003 vertical line experiment No disconnector operated during this experiment. Date Stroke ID Peak Current [kA] Charge Deficit, 90 s [%] Charge, 90 s [%] Phase A to B Flashover Injected A to N N to Ground B7 B6 B6 to N Between p6 & p10 After p6 6/30/03 FPL0301 1 11.1 -13.8 5.1 3.4 2.2 YES NO 2 <47.6 <36.1 >2.3 >1.5 >1.0 YES NO 3 7.6 -7.9 3.7 2.9 1.7 YES NO 7/6/03 FPL0305 1 >-40.1 >-63.1 -4.9 5.5 YES YES 2 11.3 -14.9 -15.5 2.4 2.1 0.1 YES NO 3 10.6 1.7 -2.1 3.7 3.2 1.3 YES NO 4 12.3 <18.1 <25.8 >2.2 >1.4 >0.2 YES NO 7/13/03 FPL0312 1 16.0 7.9 2.9 4.0 2.5 2.0 YES NO 2 5.8 -1.6 -11.1 5.5 5.1 1.2 YES NO 3 12.5 16.6 3.1 8.8 5.2 4.5 YES NO 4 7.3 -13.1 -8.9 -0.8 -0.2 -0.2 NO NO 5 8.9 -11.5 -10.0 0.1 0.0 0.1 NO NO 6 10.4 -2.8 -1.0 0.2 -0.1 0.0 NO NO 7 10.4 3.1 -0.9 1.7 1.8 0.0 YES NO 8 15.3 <40.9 <36.5 >2.8 >1.8 >1.2 YES NO 9 10.9 2.8 -2.5 2.7 2.0 1.0 YES NO 10 12.2 7.9 3.1 3.6 2.4 1.7 YES NO 11 14.8 6.7 4.4 3.4 YES NO 12 9.9 8.6 5.2 4.8 YES NO 13 13.8 >4.7 >2.9 >2.5 YES NO 14 9.8 3.7 2.7 1.5 YES NO 15 15.0 >3.4 >2.0 >2.0 YES NO 16 14.8 6.8 3.8 4.0 YES NO 7/14/03 FPL0315 1 11.7 -5.0 -1.0 0.0 -0.1 0.0 NO NO 2 14.6 1.9 1.1 1.5 1.4 0.5 YES NO FPL0317 1 27.9 14.5 7.8 3.5 2.7 1.3 YES NO

PAGE 409

379 LIST OF REFERENCES Agrawal, A.K., H.J. Price, and S.J. Gurbax ani, Transient resp onse of a multiconductor transm ission line excited by a nonunifo rm electromagnetic field, IEEE Transactions on Electrom agnetic Co mpatibility, Vol. 22, 119, 1980. Barker, P.P., R.T. Mancao, D.J. Kvaltine, and D.E. Parrish, Characteristics of lightning surges measured at metal oxide distri bution arresters, IE EE Transactions on Power Delivery, Vol. 8, No. 1, 301, January 1993. Barker, P.P., T.A. Short, A.R. Eybert-Berard, and J.P. Berlan dis, Induced voltage measurements on an experimental distribu tion line during near by rocket-triggered lightning flashes, IEEE Tran sactions on Power Delivery, Vol. 11, No. 2, April 1996. Bazelyan, E.M., B.N. Gorin, and V.I. Levit ov, Physical and engi neering foundations of lightning protection, Gidrometeoizdat, St. Petersburg, Russia, 1978. Bejleri, M., V.A. Rakov, M.A. Uman, K.J. Rambo, C.T. Mata, and M.I. Fernandez, Triggered lightning testing of an air port runway lighting system, IEEE Transactions on Electromagnetic Comp atibility, Vol. 46, No. 1, 96, February, 2004. Berger, K., Methoden und Resultate der Bl itzforschung auf dem Monte San Salvatore bei Lugano in den Jahren 1963, Bulletin des Schweizerischen Elektrotechnischen Vereins, Vol. 63, No. 24, 1403, 1972. Berger, K., R.B. Anderson, and H. Kroninger, Parameters of lightni ng flashes Electra, No. 41, 23, July 1975. Berger, K., Blitzstrom-Parameter von Aufwaertsblitzen, Bulletin des Schweizerischen Elektrotechnischen Vereins, 353, 1978. Brown, K., Metal oxi de varistor degradation, IAEI Magazine, http://www.iaei.org/subscrib er/m agazine/04_b/04_b_brown.htm last accessed October 2006, March/April 2004. Bruce, C.E.R. and R.H. Golde, The lightni ng discharge, Journal of the Institution of Electrical Engineers, Pt. 2, Vol. 88, 487, 1941. Chowdhuri, P. and E.T.B. Gross, Voltage surges induced on overhead lines by lightning strokes, Proceedings IEE, Vol. 114, 1899, 1967.

PAGE 410

380 Chrzan, K.L. and V.A. Rakov, "Lightning protection of distribution lines using metal oxide surge arresters" (in Polish), Pr oceedings of Conference on Outdoor HighVoltage Insulation (NIWE'2000), Bi elsko-Biala, Poland, 2000. Cooray, V., Horizontal fields generated by return stroke st rokes, Radio Science, Vol. 27, 529, 1992. Cooray, V., Calculating lightning-induced ove rvoltages in power lines: a comparison of two coupling models, IEEE Transactions on Electromagnetic Compatibility, Vol. 36, 179, 1994. Cooray, V. and V. Scuka, Lightning-induced overvoltages in power lines: validity of various approximations made in overvoltage calculations, IEEE Transactions on Electromagnetic Compatibility, Vol. 40, 355, 1998. Crawford, D.E., Multiple-station measurem ents of triggered lightning electric and magnetic fields, Masters thesis, University of Florida, Gainesville, USA, 1998. DeCarlo, B.A., J. Howard, J.E. Jerauld, G.H. Schnetzer, J. Schoene, V.A. Rakov, M.A. Uman, K.J. Rambo, D.M. Jordan, Rocke t-triggered lightning experiment final report for EDOT in summer of 2005, Los Alamos National Laboratory Statement of Work #98497, May 2006a. DeCarlo, B.A., V.A. Rakov, J. E. Jerauld, G.H. Schnetzer, J. Schoene, M.A. Uman, K.J. Rambo, V. Kodali, D.M. Jordan, G. Maxwell, S. Humeniuk, and M. Morgan, Triggered-lightning testing of the protective system of a residential building: 2004 and 2005 results, Proceedings of In ternational Conference on Lightning Protection, Kanazawa, Japan, 2006b. Depasse, P., Statistics on artificially triggered lightning, Journal of Geophysical Research, Vol. 99, No. D9, 18515, 1994. Diendorfer, G. and M.A. Uman, An improve d return stroke model with specified channel-base current, Journal of Ge ophysical Research, Vol. 95, No. D9, 13621 13644, 1990. Dommel, H.W., EMTP Theory Book B.P.A., Portland, Oregon, 1986. Dommel, H.W. and W.S. Meyer, Computation of electromagnetic transients, IEEE Proceedings, Vol 62, No 7, July 1974. EMTP Rule Book, EPRI/DCG Version 3, December 2001. Eriksson, A.J., M.F. Stringfellow, and D.V. Meal, Lightning-induced overvoltages on overhead distribution lines, IEEE Tr ansactions on Power Apparatus and Systems, Vol. 101, No.4, 960, April 1982.

PAGE 411

381 Fernandez, M.I., Responses of an unenergized test power distributi on system to direct and nearby lightning strikes, Masters thesis, University of Florida, Gainesville, Florida, 1997. Fernandez, M.I., V.A. Rakov, and M.A. Um an, Overvoltages in underground systems, phase I results, Technical Report TR-109669, Electric Power Research Institute (EPRI), Palo Alto, CA, December 1997a. Fernandez, M.I., V.A. Rakov, and M.A. Uman, Testing of lightning arresters and improved lightning protection, preliminary results, Technical Report TR-109670, Electric Power Research Institute (EPRI), Palo Alto, CA, December 1997b. Fernandez, M.I., C.T. Mata, V.A. Rakov, M.A. Uman, K.J. Rambo, M.V. Stapleton, and M. Bejleri, Improved lightning arrester protection results, final results, Technical Report TR-109670-R1, Electric Po wer Research Institute (EPRI), Palo Alto, CA, December 1998a. Fernandez, M.I., K.J. Rambo, M.V. Staple ton, V.A. Rakov, M.A. Uman, Review of triggered lightning experiments performe d on a power distribution system at Camp Blanding, Florida, during 1996 and 1997, Proceedings 24th International Conference on Lightning Protecti on, Birmingham, UK, 29, 1998b. Fernandez, M.I., V.A. Rakov, M.A. Uman, T ransient currents and voltages in a power distribution system due to natu ral lightning, Proceedings 24th International Conference on Lightning Protecti on, Birmingham, UK, 622, 1998c. Fernandez, M.I., K.J Rambo, V.A. Rakov, M.A. Uman, Performance of MOV arresters during very close, direct lightning strikes to a power distribution system, IEEE Transactions on Power Delivery, Vol. 14, No. 2, 411, April 1999. Fisher, R.J., G.H. Schnetzer, R. Thottappi llil, V.A. Rakov, M.A. Uman, and J.D. Goldberg, Parameters of triggered-light ning flashes in Florida and Alabama, Journal of Geophysical Research, Vol. 98, No. D12, 22887, 1993. Georgiadis, N., M. Rubinstein, M.A. Um an, P.J. Medullas, and E.M. Thomson, Lightning induced voltages at both e nds of a 448-m power-distribution line, IEEE Transactions on Electromagnetic Compatibility, Vol. 34, No. 4, 452, November 1992. Hanley, B., J. Schoene, M.A. Uman, J.E. Jerauld, K.J. Rambo, J. Howard, and B. DeCarlo, Testing of the OBO Betterma nn peak current system for Lawrence Livermore national laboratory-phase II, Technical Report, University of Florida, June 2006. Harnden, J.D. Jr., F.D. Martzloff, W.G. Mo rris, F.B. Golden, Metal-oxide varistor: A new way to suppress transients, Electronics, October 9, 1972.

PAGE 412

382 He, J., S. Chen, R. Zeng, X. Liang, Z. Gu an, S. Han, H. Cho, Electrical parameter statistic analysis and parallel coordination of ZnO varistors in low-voltage protection devices, IEEE Transactions on Power Delivery, Vol. 20, No.1, January 2005. Henry, J. A., K.M. Portier, and J. Coyne, The climate and weather of Florida Pineapple Press, Sarasota, Florida, 1994. Hitoshi, S., A. Asakawa, S. Yokoyama, and K. Nakada, Effectiven ess of installing two pairs of distribution surge arresters in parallel, High Vo ltage Engineering Symposium, 22, August 1999. Horii, K., Experiment of artificial lightni ng triggered with rocket, Memoirs of the Faculty of Engineering, Nagoya Un iversity Japan Vol. 34, 77, 1982. Horii, K. and M. Nakano, Artificially triggered lightning Handbook of Atmospheric Electrodynamics, Vol. 1, 151, ed. H. Volland, CRC Press, Boca Raton, Florida, 1995. Hillman, C., Metal oxide varistor (MOV) failures, DfR Solutions, www.dfrsolutions.com/Information/MOV%20Failures.pdf last accessed March 2005. Im ece, A.F., D.W. Durbak, H. Elahi, S. Kolluri A. Lux, D. Mader, T.E. McDerm ott, A. Morched, A.M. Mousa, R. Natarajan, L. Rugeles, and E. Ta rasiewicz, Modeling guidelines for fast front tr ansients, IEEE Transactions on Power Delivery, Vol. 11, No. 1, 493, January 1996. Jacob, P.B., S. Grzybowski, and E.R. Ross, An estimation of lightning insulation level of overhead distribution lines , IEEE Transactions on Power Delivery, Vol.6, No. 1, January 1991. Jerauld, J.E., A multiple-station experime nt to examine the close electromagnetic environment of natural and triggered light ning, Masters Thesis, University of Florida, 2003. Jerauld, J.E., Camp Blandi ng experiments, Internal Report, University of Florida, October 2005. Kobayashi, M., H. Sasaki, and K. Nakamura Rocket-triggered lightning experiments on zinc oxide arresters and thei r applications to power transmission lines, Electrical Engineering in Japan, Vol. 122, No. 4, 25, March 1998. Kuehr, W., Der Privatflugzeugfuehrer Schiffmann Verlag, Bergisch-Gladbach 1996. Lenk, D.W., New distribution class arrester ground lead disconnector design enhances detonation reliability and improves a rrester performance, Power System Conference and Exposition, October 2004.

PAGE 413

383 MacGorman, D.R. and W.D. Rust, The electrical nature of thunderstorms Oxford University Press, New York, January 1998. Marti, J.R., L. Marti, and H.W. Dommel, Transmission line models for steady-state and transients analysis, IEEE/NTUA Power Tech Conference, Athens, Greece, Paper APT IS-333, September 1993. Master, M.J., M.A. Uman, W.H. Beasley, and M. Darveniza, Lightning induced voltages in power lines: Experiment, IEEE Transactions on Power Apparatus and Systems, Vol. 103, No. 9, 2519, September 1984. Master, M.J., M.A. Uman, W.H. Beasley, a nd M. Darveniza, Voltages induced in an overhead line by the lightning steppe d leader, IEEE Transactions on Electromagnetic Compatibility, Vol. 28, No. 3, 158, August 1986. Mata, A.G., Interaction of lightning with power di stribution lines: 2001 and 2002 experiments at the International Cent er for Lightning Research and Testing (ICLRT), Masters Thesis, Un iversity of Florida, 2003. Mata, A.G., V.A. Rakov, K.J. Rambo, M.V. Stapleton, and M.A. Uman, UF/FPL study of triggered lightning stri kes to FPL distribution lin es: 2001 experiments, phase III report, University of Florida, December 2001. Mata, A.G., C.T. Mata, V.A. Rakov, M.A. Um an, J. Schoene, K.J. Rambo, D.M. Jordan, and J.E. Jerauld, UF/FPL study of trigge red lightning strikes to FPL distribution lines, phase IV report, Final Report, Un iversity of Florida, December 2002. Mata, C.T., Interaction of li ghtning with power distribution lines, Ph.D. Dissertation, University of Florida, 2000. Mata, C.T., M.I. Fernandez, V.A. Rakov, M.A. Uman, M. Bejleri, K.J. Rambo, and M.V. Stapleton, Overvoltages in underground systems, phase 2 results, Technical Report TR-109669-R1, Electric Power Research Institute (EPRI), Palo Alto, CA, December 1998. Mata, C.T., V.A. Rakov, K.J. Rambo, M.V. Stapleton, and M.A. Uman, UF/FPL study of triggered lightning strike s to FPL distribution lines, Technical Report, Florida Power and Light, Miami, Florida, December 1999. Mata, C.T., M.I. Fernandez, V.A. Rakov, and M.A. Uman, EMTP modeling of a triggered lightning strike to the phase conductor of an overhead distribution line, IEEE Transactions on Power Delivery, Vol. 15, No. 4, 1175, 2000a. Mata, C.T., V.A. Rakov, K.J. Rambo, and M.A. Uman, Final Report, UF/FPL study of triggered lightning strikes to FPL di stribution lines: 2000 experiments, University of Florida., December 2000b.

PAGE 414

384 Mata, C.T., V.A. Rakov, and M.A. Uman, "D ivision of lightning current and charge among multiple arresters and grounds of a power distribution line, Proceedings of 26th International Conference on Li ghtning Protection, Krakow, Poland, 585 590, September 2, 2002. Mata, C.T., V.A. Rakov, K.J. Rambo, P. Diaz, R. Rey, and M.A. Uman., "Measurement of the division of lightning return stroke current among the multiple arresters and grounds of a power distribution line, I EEE Transactions on Power Delivery, Vol. 18, No. 4, 1203, 2003. Matsumoto, Y., O. Sakuma, K. Shinjo, M. Saik i, T. Wakai, T. Sakai, H. Nagasaka, H. Motoyama, and M. Ishii, Measurement of lightning surges on test transmission line equipped with arresters struck by natural and triggered lightning, IEEE Transactions on Power Delivery, Vol. 11, No. 2, 996, April 1996. McDermott, T.E., Line Arrester Energy Di scharge Duties, Proceedings of IEEE PES T&D Conference 2005/2006, Dallas, 450, May 2006. Meliopoulos, P.A. and M.G. Moharam, Trans ient analysis of grounding systems, IEEE Transactions on Power Apparatus and Systems, Vol. 102, No.2, 389, 1983. Motoyama, H., Y. Matsumoto, and N. Itamoto, Observation and analysis of multiphase back flashover on the okushishiku test transmission line caused by winter lightning, IEEE Transactions on Power Delivery, Vol. 13, No. 4, 1391, October 1998. Nakada, K., H. Sugimoto, S. Yokoyama, Expe rimental facility for investigation of lightning performance of distribution lines, IEEE Transactions on Power Delivery, Vol. 18, No. 1, January 2003. Newton, D. J., Peak Current Sensor Card M easurement Trials using Lightning Simulator at AEA Culham, Technical Report, AWE/DSE01/B/RP/123, October 2004. Novak, T., T.J. Fisher, Lightning propagation through the earth a nd its potential for methane ignitions in abandoned area s of underground coal mines, IEEE Transactions on Industry Applications, Vol. 37, No. 6, November/December, 2001. Nucci, C.A., G. Diendorfer, M.A. Uman, F. Rachidi, M. Ianoz, and C. Mazzetti, "Lightning return stroke current models with specified channel-base current: A review and comparison, J ournal of Geophysical Research, Vol. 95, No. D12, 20395, 1990. Nucci, C.A., F. Rachidi, M. Ianoz, and C. Mazzetti, Lightning-induced voltages on overhead lines, IEEE Transactions on El ectromagnetic Compatibility, Vol. 35, No. 1, 75, February 1993.

PAGE 415

385 Nucci, C.A., M. Ianoz, F. Rachidi, M. Rubinstein, F.M. Tesche, M.A. Uman, and C. Mazzetti, Modeling of lightning-induced voltages on overhead lines: Recent developments, Elektrotechnik und Informationstechnik, Vol. 112, No. 6, 290 296, 1995. Paolone, M., C.A. Nucci, and F. Rachidi, M itigation of lightninginduced overvoltages in medium voltage distribution line s by means of periodical grounding of shielding wires and of surge arresters, Proceedings of 4th European Symposium on Electromagnetic Compatibility, Brugge, Belgium, 11 Sept. 2000. Paolone, M., C. A. Nucci, A. Borghetti, E. Petrache, F. Rachidi, A new interface for lightning induced overvoltages calcula tion between EMTP and LIOV code, EMTP user group meetingIEEE Summer M eeting, Vancouver, Canada, July 15 19, 2001. Paolone, M., C.A. Nucci, E. Petrache, and F. Rachidi, Mitigati on of lightning-induced overvoltages in medium voltage dist ribution lines by means of periodical grounding of shielding wires and of surge arresters: Modeling and experimental validation, IEEE Transactions on Power Delivery, Vol. 19, No. 1, January 2004a. Paolone, M., J. Schoene, M.A. Uman, V.A. Rakov D. Jordan, K.J. Rambo, J.E. Jerauld, C.A. Nucci, A. Borghetti, F. Rachidi, and E. Petrache, Testing of the LIOVEMTP96 code for computing lightning-indu ced currents on real distribution lines: Triggered-lightning experiments, Proceedings of International Conference on Lightning Protection, Avignon, France, 2004b. Paul, C.R., Analysis of Multiconductor Transmission Lines Wiley series in microwave and optical engineering, ISBN 0-471-02080-X, 1994. Rachidi, F., C.A. Nucci, M. Ianoz, C. M azzetti, Response of multic onductor power lines to nearby lightning return stroke electrom agnetic fields, IEEE Transactions on Power Delivery, Vol.12, 1404, July 1997a. Rachidi, F., M. Rubinstein, S. Guerrieri, and C.A. Nucci, Voltages induced on overhead lines by dart leaders and subsequent return strokes in natural and triggered lightning, IEEE Transactions on Electromagnetic Compatibility, Vol. 39, No. 2, 160, May 1997b. Rakov, V.A., On estimating the lightning peak current distribution parameters taking into account the lower measurement limit, Elektrichestvo, No. 2, 57, 1985. Rakov, V.A., Lightning electromagnetic fields: Modeling and measurements, Proceedings of 12th International Zurich Symposium on Electromagnetic Compatibility, Zurich, Switzerland, 59, February 1997.

PAGE 416

386 Rakov, V.A., "Rocket-triggered lightning experiments at Camp Blanding, Florida, Proceedings of 1999 International Conferen ce on Lightning and Static Electricity, Toulouse, France, 469, June 22, 1999a. Rakov, V.A., "Lightning discharg es triggered using rocket-and-wire techniques, Recent Research Development in Geophysics, 141, 1999b. Rakov, V.A., "Positive and bipolar lightning discharges: A review , Proceedings of 25th International Conference on Lightning Protection Rhodes, Greece, 10308, September 18, 2000. Rakov, V.A., Lightning, Classnotes of course EEL 5490, Spring Term, University of Florida, 2001. Rakov, V.A., "Review of triggered-lightni ng experiments at the ICLRT at Camp Blanding, Florida, Proceedings of 2nd International Seminar on Lightning Physics and Protection in the South of Brazil, Porto Alegre, Brazil, May 9, 2003. Rakov, V.A. and A.A. Dulzon, Calculated electromagnetic fields of lightning return stroke, Tekh. Elektrodinam. No. 1, 87, 1987. Rakov, V.A. and A.A. Dulzon, A modified transmission line model for lightning return stroke field calculations, Proceedings of 9th International Zurich Symposium Electromagnetic Compatibility, Zuri ch, Switzerland, 229, March 1991. Rakov, V.A. and M.A. Uman, Review and eval uation of lightning return stroke models including some aspects of their ap plication, IEEE Transactions on Electromagnetic Compatibility Vol. 40, 403, 1998. Rakov, V.A., M.A. Uman, K.J. Rambo, M.I. Fe rnandez, R.J. Fisher, G.H. Schnetzer, R. Thottapillil, A. Eybert-Berard, J.P. Berlan dis, P. Lalande, A. Bonamy, P. Laroche, and A. Bondiou-Clergerie, New insights into lightning processes gained from triggered-lightning experiments in Florid a and Alabama, Journal of Geophysical Research, Vol. 103, No. D12, 14117, 1998. Rakov, V.A., D.E. Crawford, K.J. Rambo, G.H. Schnetzer, M.A. Uman, and R. Thottappillil, M-component mode of ch arge transfer to ground in lightning discharges, Journal of Geophysical Research, Vol. 106, 22817, 2001. Rakov, V.A. and M.A. Uman, Lightning Physics and Effects ISBN 0-521-58327-6, Cambridge University Press, 2003. Rakov, V.A., C.T. Mata, M.A. Uman, K.J. Ra mbo, and A.G. Mata, "Review of triggeredlightning experiments at the ICLRT at Camp Blanding, Florida, Proceedings of 5th IEEE Power Tech Conference, Paper 381, Bologna, Italy, 2003a.

PAGE 417

387 Rakov, V.A., C.T. Mata, A.G. Mata, M.A. Uman, K.J. Rambo, "Recent triggeredlightning experiments at the ICLRT at Camp Blanding, Florida, Proceedings of International Symposium on Lightning Protection (VII SI PDA), Curitiba, Brazil, 144, November 17, 2003b. Rubinstein, M., An approximate formula for the calculation of the horizontal electric field from lightning at close, intermedia te, and long range, IEEE Transactions. on Electromagnetic Compatibility, Vol. 38, 531, 1996. Rubinstein, M., M. A. Uman, P.J. Medelius, and E.M. Thomson, Measurements of the voltage induced on an overhead power line 20 m from triggered lightning, IEEE Transactions on Electromagnetic Compa tibility, Vol. 36, No. 2, 134, May 1994. Rusck, S., Induced lightning overvoltages on power transmission lines with special reference to the overvoltage protection of low voltage networks, Transactions of the Royal Institute of Technology, No. 120, Stockholm, 1958. Rusck, S., Protection of distribution systems, Lightning Vol. 2, Lightning Protection ed. R.H. Golde, New York: Academic, 747, 1977. Schneider, H. M. and H. R. Stillwell, Measurement of lightning current waveshapes on distribution systems, IEEE PES Summer Meeting, Paper A 79, 526, July 15, 1979. Schoene, J., M.A. Uman, K.J. Rambo, D.M. Jordan, V.A. Rakov, G.H. Schnetzer, J.E. Jerauld, M.V. Stapleton, A.G. Mata, C. T. Mata, UF/FPL study of the interaction of triggered lightning with FPL distribution lines-pha se V report, Technical Report, University of Florida, December 2003a. Schoene, J., M.A. Uman, V.A. Rakov, V. Kodali, K.J. Rambo, G.H. Schnetzer, Statistical characteristic s of the electric and magnetic fields and their time derivatives 15 m and 30 m from trigge red lightning, Journal of Geophysical Research, Vol. 108, No. D6, 4192, 2003b. Schoene, J., M.A. Uman, M. Aurele, K.J. Ra mbo, J.E. Jerauld, and G.H. Schnetzer, Testing of the OBO Bettermann peak current sensor system for Lawrence Livermore national laboratory, Techni cal Report, University of Florida, November 2004a. Schoene, J., M.A. Uman, J.E. Jerauld, G.H. Schnetzer, K.J. Rambo, D.M. Jordan, V.A. Rakov, UF/FPL study of the interacti on of triggered lightning with FPL distribution lines-phase VI report, Tec hnical Report, University of Florida, December 2004b.

PAGE 418

388 Schoene, J., M.A. Uman, V.A. Rakov, A.G. Mata, C.T. Mata, K.J. Rambo, J.E. Jerauld, D.M. Jordan, and G.H. Schnetzer, Direct lightning strikes to test power distribution linespart 1: Experiment a nd overall Results, IEEE Transactions on Power Delivery (accepted) Manuscript No. TPWRD-00032-2006, 2006a. Schoene, J., M.A. Uman, V.A. Rakov, A.G. Mata, C.T. Mata, K.J. Rambo, J.E. Jerauld, D.M. Jordan, and G.H. Schnetzer, Direct lightning strikes to test power distribution linespart 2: Measured a nd modeled current division among multiple arresters and grounds, IEEE Transactions on Power Delivery (accepted) Manuscript No. TPWRD-00207-2006b, 2006b. Simpson, G. and F.J. Scrase, The distri bution of electricity in the thunderclouds, Proceedings of the Royal Society London, Ser. A 161, 309, 1937. Thottappillil, R., D.K. McLain, M.A. Uman and G. Diendorfer, Extension of the Diendorfer-Uman lightning return stroke m odel to the case of a variable upward return stroke speed and a variable downwa rd discharge current speed, Journal of Geophysical Research, Vol. 96, 17143, 1991. Thottappillil, R. and M.A. Uman, "Compari son of lightning return-stroke models, Journal of Geophysical Research, Vol. 98, No. D12, 22903, 1993. Thottappillil, R. and M.A. Uman, "Lightning return stroke model with height-variable discharge time constant, Journal of Geophysical Research, Vol. 99, No. D11, 22773, 1994. Thottappillil, R., V.A. Rakov, and M.A. Uman, "Distribution of charge along the lightning channel: Relation to remote electr ic and magnetic fields and to returnstroke models, Journal of Geophysical Research, Vol. 102, No. D6, 6987-7006, 1997. Uman, M.A., The lightning discharge Dover Publications, Mineola, New York, 1987. Uman, M.A. and D.K. McLain, Magnetic field of the lightning return stroke, Journal of Geophysical Research, Vol. 74, 6899, 1969. Uman, M.A., D.K. McLain, and E.P. Krider, T he electromagnetic radiation from a finite antenna, American Journal of Physics, No. 43, 33, 1975. Uman, M.A. and V.A. Rakov, "Triggered light ning effects on a test power distribution system consisting of an overhead line a nd an underground cable, Electric Power Research Institute (EPRI), Palo Alto, Ca lifornia, Final Report on Performance of Storm Test, 1995. Uman, M.A., V.A. Rakov, K.J. Rambo, T.W. Vaught, M.I. Fernandez, D.J. Cordier, R.M. Chandler, R. Bernstein, and C. Gold en, "Triggered-lightning experiments at Camp Blanding, Florida (1993)," Transactions IEE Japan, Vol. 117-B, 446 452, 1997.

PAGE 419

389 Uman, M.A., V.A. Rakov, G.H. Schnetzer, K. J. Rambo, D.E. Crawford, and R.J. Fisher, Time derivative of the electric field 10, 14, and 30 m from triggered lightning strokes, Journal of Geophysical Re search, Vol. 105, No. D12, 15577, 2000. Uman, M.A., J. Schoene, V.A. Rakov, K.J. Ra mbo, and G.H. Schnetzer, Correlated time derivatives of current, el ectric field intensity, and magnetic flux density for triggered lightning at 15 m, Journal of Geophysical Research, Vol. 107, No. D13, 2002. Verma, R. and D. Mukhedkar, Impulse impedance of buried ground wire, IEEE Transactions on Power Apparatus and Systems, Vol. 99, No. 5, 2003, 1980. Willett, J.C., V.P. Idone, R.E. Orville, C. Lete inturier, A. Eybert-Berard, L. Barret, and E.P. Krider, "An experimental test of the transmission-line model of electromagnetic radiation from triggered lightning return strokes, Journal of Geophysical Research, Vol. 93, No. D4, 3867, 1988. Willett, J.C., J.C. Bailey, V.P. Idone, A. Eybert-Berard, and L. Barret, "Submicrosecond intercomparison of radiation fields an d currents in triggered lightning return strokes based on the transmission-line model, Journal of Geophysical Research, Vol. 94, No. D11, 13275, 1989. Williams, E.R., Das Gewitter als elektrischer Generator, Spektrum der Wissenschaft, January 1989. Wilson, C.T.R., On some determination of th e sign and magnitude of electric discharges in lightning flashes, Proceedings of the Royal Society, Vol. A 92, 555, 1916. Wilson, C.T.R., Investigation on lightning di scharges and on the electric field of thunderstorms, Philosophical Transactions of the Royal Society, Vol. A 221, 73 115, 1920. Yokoyama, S., Calculation of lightning-i nduced voltages on ove rhead multiconductor systems, IEEE Transactions on Power Appa ratus and Systems, Vol. 103, No. 1, 2420, January 1984. Yokoyama, S., Lightning prot ection of MV overhead distribution lines, 7th International Symposium on Lightning Pr otection, Curitiba, Brazil, November 17, 2003. Yokoyama, S., K. Miyake, H. Mitani, and A. Takanishi, Simultaneous measurement of lightning induced voltages with associated stroke currents, IEEE Transactions on Power Apparatus and Systems, Vol. 102, 2420, 1983.

PAGE 420

390 Yokoyama, S., K. Miyake, H. Mitani, and N. Yamazaki, Advanced observations of lightning induced voltages on power distribution lines, IEEE Transactions on Power Delivery, Vol. 1, 129, 1986. Yokoyama, S., K. Miyake, and S. Fukui, Advanced observations of lightning induced voltage on power distribution lines (II),, IEEE Transactions on Power Delivery, Vol. 4, No. 4, 2196, October 1989.

PAGE 421

391 BIOGRAPHICAL SKETCH Jens D. Schoene received the Diplom-Ingenieur degree from the University of Paderborn, Department Soest, Germany in 1999 and the M.S. degree from the University of Florida in 2002. In 1998, he became involve d in lightning research during a five month practical training at the International Center for Lightning Research and Testing (ICLRT) at Camp Blanding, Florida, USA. His re search areas are the responses of power distribution systems to direct and nearby lightning strikes, the modeling of the lightning return stroke process, and the character ization of the lightning electromagnetic environment. He is the author or coauthor of over 30 papers or t echnical reports, with 11 papers being published in reviewed journals.