Citation
Space-time reactor kinetics studies with the University of Florida, SPERT Assembly.

Material Information

Title:
Space-time reactor kinetics studies with the University of Florida, SPERT Assembly.
Series Title:
Space-time reactor kinetics studies with the University of Florida, SPERT Assembly.
Creator:
Diaz, Nils Juan,
Place of Publication:
Gainesville FL
Publisher:
University of Florida
Publication Date:
Language:
English

Subjects

Subjects / Keywords:
Analyzers ( jstor )
Decay constants ( jstor )
Instrumentation ( jstor )
Kinetics ( jstor )
Neutrons ( jstor )
Pulse duration ( jstor )
Reactivity ( jstor )
Reflectors ( jstor )
Spacetime ( jstor )
Thermal neutrons ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Nils Juan Diaz. Permission granted to the University of Florida to digitize, archive and distribute 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.
Resource Identifier:
000955726 ( alephbibnum )
16973836 ( oclc )

Downloads

This item has the following downloads:


Full Text












SPACE-TIME REACTOR KINETICS STUDIES WITH THE

UNIVERSITY OF FLORIDA SPERT ASSEMBLY














By

NILS J. DIAZ














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












UNIVERSITY OF FLORIDA 1969








































TO ZENA















ACKNOWLEDGMENTS


The author wishes to express his sincere appreciation to his graduate committee for their guidance. Special recognition is due Dr. M. J. Ohanian, whose encouragement, dedication and detailed scientific knowledge made this work possible.

The support of the Nuclear Engineering Sciences Department of the University of Florida throughout the author's graduate work is-' greatly appreciated, In particular Dr. R. E. Uhrig's support and friendship is gratefully recognized.

The author feels fortunate and proud in having studied and worked with a remarkable scientist and gentleman, Dr. R. B. Perez, now at Oak Ridge National Laboratory. The author's years of association with Dr. T. F. Parkinson, now at the University of Missouri, formed the necessary background for this work.

The continuous assistance of Mr. G. W. Fogle throughout the experimental program is sincerely appreciated. Mr. L. B. Myers designed and built the control instrumentation. Mr. R. E. Schoessow was responsible for the design and construction of the assembly. Mr. E.'Dugan. and Mr. H. Leydolt aided in the data processing. The cooperation of the staff of the Nuclear Engineering Sciences Department during the construction of the facility is acknowledged.

Most of this work was financed under subcontract No. C281 and C635 with Atomic Energy Division of the Phillips Petroleum Company,.










under a prime contract with the United States Atomic Energy Commission. The technical aid of the University of Florida Computing Center in the development of the computer programs and their financial assistance is gratefully acknowledged.

Special mention is due Messrs. S. 0. Johnson, R. W. Garner,

G. A. Mortensen and Mrs. M. E. Radd of the Nuclear Safety Research Branch, Atomic Energy Division, Phillips Petroleum Company for their' continuous assistance and invaluable suggestions throughout the research program.

To my sister, Miss Lydia Gonzalez, my sincere appreciation for. typing an elegant manuscript from my unintelligible characters.

To the fellow students, who struggled with me to reach the unreachable star, my space and time independent friendship.
















PREFACE


The text of this dissertation is divided into two related but essentially independent parts:

Part 1. The University of Florida SPERT Assembly
- Design and Calibration

Part 2. Space-Time Reactor Kinetics Studies with the
University of Florida SPERT Assembly

In Part 1, the design, operational safety and nuclear calibration of the large-in-one-space dimension, highly multiplicative University of Florida SPERT Assembly (UFSA) are described. The presentation of this material is pertinent for a complete understanding of the physical characteristics of the system to be studied in Part 2 and also because the system in itself is interesting from the nuclear engineering point of view. The data acquisition system used for the experimentation is presented in Chapter IV of this section.

In Part 2, the linear reactor kinetics studies performed with the UFSA are presented. The space-time dependent distribution of the neutrons in the assembly, following the introduction of a burst of neutrons at one end of the core, are studied in the time and in the frequency dom ain.

The study of the spatially dependent time profile of the neutron flux required a large number of figures containing the calculational and experimental results at different positions in the core. Some typical figures are included in the main text but most of the recorded (and calculated) time profiles are included as appendices to the main text.
















TABLE OF CONTENTS


Page

ACKNOWLEDGMENTS . iii PREFACE . . v LIST OF TABLES . xi LIST OF FIGURES . xiii LIST OF SYMBOLS . xviii ABSTRACT . . xix



PART I
THE UNIVERSITY OF FLORIDA SPERT ASSEMBLY
DESIGN AND CALIBRATION . . 1

CHAPTER I INTRODUCTION . 2 CHAPTER II DESCRIPTION OF THE FACILITY . 5

General Features . 5 Fuel Characteristics . 9 Mechanical Design . 10 Moderator Flow Control System . 15 Instrumentation and Interlock System . 22 Fuel Storage . 27 Neutron Sources . 30 CHAPTER III OPERATIONAL SAFETY . 32

Introduction . 32 Initial Loading . 33















TABLE OF CONTENTS (cont'd)


Page

Operating Limits . 35 Design Basis Accident Analysis . 39

CHAPTER IV THE DATA ACQUISITION SYSTEM . 45

Introduction . .45 The Neutron Detector . 47 The Electronic Instrumentation . 50 The Resolution Time Correction i . 54 The Normalization Technique . 61 Comments . 62 CHAPTER V NUCLEAR CALIBRATION OF THE UFSA SUBCRITICAL . 64

Introduction . 64 Theoretical Notes . 64 Inverse Multiplication Measurements . 68 Absolute Determination of k 73
eff
Conclusions . 75.



PART 2
SPACE-TIME REACTOR KINETICS STUDIES WITH
THE UNIVERSITY OF FLORIDA SPERT ASSEMBLY . 79 CHAPTER I INTRODUCTION . 80

Statement of the Problem . 80 Description of the Study . 81
















TABLE OF CONTENTS (cont'd)


Page

Nomenclature Used in the Description of
Pulse Propagation Phenomena. 83

CHAPTER II THEORETICAL NOTES .85

Introduction. 85 Review of the Literature .85.

The WIGLE Calculational Scheme .o.87 Neutron Wave Analysis. o.89 CHAPTER III DESCRIPTION OF THE MEASUREMENTS. o.o.91

Introduction .o.o.o.91 The Epicadmium Subtraction Method . o. 92 The Geometrical Arrangement. 94

Synopsis of the Measurements .97 CHAPTER IV EXPERIMENTAL AND THEORETICAL RESULTS IN THE
TIME DOMAIN .o.o.o.100

The Analytical Model. . o.o. 100 Flux Traverses .o.o. o.o 109 Clean Core Pulse Propagation Measurements. o. .118 Propagation of a Narrow Pulse .o.o.o.147 Propagation of a Wide Pulse .o. o. o 148 Pulse Shape vs. Input Pulse Width.o.o. o 150 Effect of Room Return at Peripheral Detector Positions .o 152


viii















TABLE OF CONTENTS (cont'd)


Page

CHAPTER V EXPERIMENTAL AND THEORETICAL RESULTS IN
THE FREQUENCY DOMAIN .156

Introduction .156 Method of Analysis. . 157

Comparison of the Theoretical and the Measured
Results of the Neutron Wave Analysis. 159 CHAPTER VI SPATIAL DEPENDENCE OF PULSED-NEUTRON
REACTIVITY MEASUREMENTS . 174

Introduction. 174 The Decay Constant . 175 The Ratio Wt/i.4. 176 The Measured Reactivity and k f Values . 179 CHAPTER VII CONCLUSIONS. 185 APPENDICES

A CALCULATIONAL PROCEDURES USED IN THE DETERMINATION
OF THE NUCLEAR PARAMETERS AND THE k f VALUES . 187

B DESCRIPTION OF THE COMPUTER PROGRAMS.191

C UFSA Ri CLEAN CORE
TIME PROFILES OF THERMAL NEUTRON FLUX AT NINETEEN CORE
POSITIONS FOR INPUT PULSES OF 0.5 AND 1.0 MSEC .196

D UFSA Ri CLEAN CORE
TIME PROFILES OF FAST NEUTRON FLUX AT SEVERAL CORE
POSITIONS FOR INPUT PULSES OF 0.5 AND 1.0 MSEC .235

E UFSA Ri CLEAN CORE
TIME PROFILES OF THERMAL NEUTRON FLUX AT FOUR POSITIONS
IN THE CORE FOR A 0.1 MSEC INPUT PULSE.248

















TABLE OF CONTENTS (cont'd)


Page

F UFSA Rl CLEAN CORE
TIME PROFILES OF THERMAL NEUTRON FLUX AT THREE
POSITIONS IN THE CORE FOR A WIDE (10 MSEC) INPUT PULSE . 253

G UFSA Ri CLEAN CORE
SHAPE OF THE PROPAGATING PULSE AS A FUNCTION OF INPUT
PULSE WIDTH. 255 LIST OF REFERENCES. 258 BIOGRAPHICAL SKETCH. 262
















LIST OF TABLES


TABLE Page

I k efvs. MODERATOR LEVEL OF UPSA REFLECTED

CORES. . 7

II UFSA INSTRUMENTATION AND CONTROL . 25

III SUMMARY OF 1/M AND PULSED MEASUREMENTS. 74

IV INPUT PARAMETERS FOR THE WIGLE CALCULATIONAL
SCHEME. 104

V TIME STEPS USED FOR THE WIGLE CALCULATIONS. 106 VI DELAY TIMES MEASURED ACROSS THE WIDTH OF THE CORE
- 0.5 MSEC INPUT PULSE .-. 119

VII CLEAN CORE PULSE PROPAGATION STUDIES EXPERIMENTAL AND THEORETICAL RESULTS
-0.5 MSEC INPUT PULSE WIDTH . . 127 VIII CLEAN CORE PULSE PROPAGATION STUDIES EXPERIMENTAL
AND THEORETICAL RESULTS
-1.0 MSEC INPUT PULSE WIDTH . . . 128

IX ASYMPTOTIC PROPAGATION VELOCITY v. .o 133
p
X DYNAMIC INVERSE RELAXATION LENGTH K d. 134

XI CHANGES IN THE NUCLEAR PARAMETERS DUE TO
CHANGES IN THE TRANSVERSE BUCKLING. .o. o. 138 XII THE CALCULATED ASYMPTOTIC VELOCITY OF PROPAGATION AND DYNAMIC INVERSE RELAXATION
LENGTH vs. CORE HEIGHT
-O0.5 MSEC INPUT PULSE WIDTH-. 146 XIII DELAY TIMES AND FWHM FOR A NARROW INPUT PULSE. 148

XIV DELAY TIMES AND FWHM FOR A WIDE INPUT PULSE.o.o.149 XVPULSE SHAPES vs. INPUT PULSE WIDTH . . 151
















LIST OF TABLES (cont'd)


TABLE Page

XVI THE REAL AN'D THE IMAGINARY COMPONENTS
OF THE COMPLEX INVERSE RELAXATION LENGTH
- 0.5 MvSEC INPUT PULSE .-. 164

XVII THE REAL AND THE IMAGINARY COMPONENTS OF THE COMPLEX INVERSE RELAXATION LENGTH
- 1.0 MSEC INPUT PULSE .-. 165

XVIII THE DECAY CONSTANT AND ka/k VALUES MEASURED
AS A FUNCTION OF INPUT PULSE WIDTH. 180

XIX REGION-WISE DEPENDENCE OF THE REACTIVITY
MEASUREMENTS. 183














LIST OF FIGURES


FIGURE Page

1 OVERALL VIEW OF THE FACILITY . 6 2A BOTTOM FUEL ROD SPACING SYSTEM . 12 2B BOTTOM FUEL ROD SPACING SYSTEM . 13

3 TOP FUEL ROD SPACING SYSTEM . 14 4 REACTIVITY-CONTROL FLOW SYSTEM . 16

5 FLOW RATE vs. HEIGHT OF WATER LEVEL ABOVE
WEIR APEX . 18

6 WEIR "BOX . . 19 7 AIR SYSTEM SCHEMATIC . 21 8 UFSA SAFETY SYSTEM LOGIC FLOW DIAGRAM . 23 9 POWER vs. TIME FOR THE DESIGN BASIS ACCIDENT . 44 10 PHYSICAL CHARACTERISTICS OF THE He3 NEUTRON COUNTERS . 49 11 MOVABLE DETECTOR DATA ACQUISITION SYSTEM . 51 12 NORMALIZING DETECTOR DATA ACQUISITION SYSTEM . 52

13 TIME PROFILES OF NEUTRON BURST RECORDED BY
CONVENTIONAL ELECTRONIC INSTRUMENTATION AND BY THE
TIME-PICKOFF SYSTEM . ? . 58

14 THE PARALYZING, NON-PARALIZING SYSTEM RESOLUTION TIME
CORRECTION AS A FUNCTION OF COUNT RATE . 60

15 DETECTOR POSITIONING SCHEME . 69 16A INVERSE MULTIPLICATION vs. MODERATOR LEVEL . 71 16B INVERSE MULTIPLICATION vs. SQUARED INVERSE HEIGHT . 72


xiii
















LIST OF FIGURES (cont'd)


FIGURE Page

17A DECAY CONSTANT vs. MODERATOR LEVEL. 76 17B ka/k, AND k vs. MODERATOR LEVEL. 77 18 THE TOTAL, EPICADMIUM AND THERMAL FLUX 117.44 CM4
FROM THE SOURCE. 95

19 UFSA SOURCE-SUBCRITICAL ASSEMBLY GEOMETRICAL
ARRANGEMENT - PLAN VIEW .-. 96

20 PLAN AND FRONT VIEW OF THE CORE REGION ENCLOSING THE
NEUTRON SOURCE. 102

21 ONE-DIMENSIONAL ARRANGEMENT OF THE UFSA CORE USED IN THE
WIGLE CALCULATIONAL SCHEME. 103

22 SPATIAL DISTRIBUTION OF SOURCE NEUTRONS INCORPORATED
INTO THE WIGLE SCHEME. . 107 23A PULSE SHAPES PREDICTED BY WIGLE AT DIFFERENT POSITIONS IN THE UFSA RI CORE. . 110 23B PULSE SHAPES PREDICTED BY WIGLE AT DIFFERENT POSITIONS IN THE UFSA Ri CORE. 111 23C PULSE SHAPES PREDICTED BY WIGLE AT DI FFERENT POSITIONS IN THE UFSA Ri CORE. 112 24A THE CALCULATED SPATIAL DISTRIBUTION OF THE THERMAL FLUX AT DIFFERENT TIMES AFTER THE PULSE. . 113 24B THE CALCULATED SPATIAL DISTRIBUTION OF THE THERMAL FLUX AT DIFFERENT TIMES AFTER THE PULSE. 114

25 THE ASYMPTOTIC STEADY-STATE VERTICAL FLUX. 116 26 THE ASYMPTOTIC STEADY-STATE HORIZONTAL FLUX. 117 27A EXPERIMENTAL PULSE SHAPES AT DIFFERENT POSITIONS IN THE UFSA Ri CORE . . . 122


xiv
















LIST OF FIGURES (cont'd)


FIGURE Page

27B EXPERIMENTAL PULSE SHAPES AT DIFFERENT POSITIONS IN THE UFSA Rl CORE. 123

27C EXPERIMENTAL PULSE SHAPES AT DIFFERENT POSITIONS IN THE UFSA Ri CORE. 124

28A EXPERIMENTALLY DETERMINED SPATIAL DISTRIBUTION OF THE NEUTRONS AT DIFFERENT TIMES AFTER THE PUSLE. 125

28B EXPERIMENTALLY DETERMINED SPATIAL DISTRIBUTION OF THE NEUTRONS AT DIFFERENT TIMES AFTER THE PULSE. 126

29 CALCULATED AND EXPERIMENTAL DELAY TIMES
-1.5 MSEC INPUT PULSE .-. 129

30 CALCULATED AND EXPERIMENTAL DELAY TIMES
- 1.0 MSEC INPUT PULSE .-. 130

31 AMPLITUDE ATTENUATION OF THE THERMAL FLUX
- 0.5 MSEC INPUT PULSE .-. 131

32 AMPLITUDE ATTENUATION OF THE THERMAL FLUX
- 1.0 MSEC INPUT PULSE .-. 132

33 DELAY TIMES OF THE THERMAL FLUX CALCULATED BY THE
WIGLE CALCULATIONAL SCHEME FOR DIFFERENT CORE HEIGHTS . 139

34 AMPLITUDE ATTENUATION OF THE THERMAL FLUX CALCULATED
BY THE WIGLE CALCULATIONAL SCHEME FOR DIFFERENT
C ORE HEIGHTS . . 140

35A THE SENSITIVITY OF THE ONE-DIMENSIONAL, TWO GROUP, SPACE-TIME KINETICS SCHEME TO CHANGES IN THE
TRANSVERSE BUCKLING. . 141

35B THE SENSITIVITY OF THE ONE-DIMENSIONAL, TWO GRO UP, SPACE-TIME KINETICS SCHEME TO CHANGES IN THE
TRANSVERSE BUCKIiNG. . 142
















LIST OF FIGURES (cont'd)


FIGURE Page

350 THE SENSITIVITY OF THE ONE-DIMENSIONAL, TWO GROUP, SPACE-TIxM KINETICS SCHEME TO CHANGES IN THE
TRANSVERSE BUCKLING. 143

36 EXPERIMENTAL PULSE SHAPES AS A FUNCTION OF CORE
HEIGHT. 144

37A EFFECT OF ROOM RETURN AT PERIPHERAL DETECTOR POSITIONS. 153

37B EFFECT OF ROOM RETURN AT PERIPHERAL DETECTOR POSITIONS. 154

38 AMPLITUDE OF ZEROTH FOURIER MOMENT vs. DISTANCE FOR
SEVERAL FREQUENCIES
-0.5 MSEC INPUT PULSE .-. 160

39 AMPLITUDE OF ZEROTH FOURIER MOMENT vs. DISTANCE FOR
SEVERAL FREQUENCIES
-1.0 MSEC INPUT PULSE .-. 161

40 PHASE OF ZEROTH FOURIER MOMENT vs. DISTANCE FOR
SEVERAL FREQUENCIES
0O.5 MSEC INPUT PULSE-.162

41 PHASE OF ZEROTH FOURIER MOMENT vs. DISTANCE FOR
SEVERAL FREQUENCIES
-1.0 MSEC INPUT PULSE - . 163

42 COMPARISON OF THE THEORETICALLY PREDICTED AND THE
MEASURED DAMPING COEFFICIENT a. 166

43 COMPARISON OF THE THEORETICALLY PREDICTED AND THE
MEASURED PHASE SHIFT PER UNIT LENGTH . 167

44 THE UFSA Rl CORE p DISPERSION LAW. 169

45 COMPARISON.OF THE THEORETICALLY PREDICTED AND THE
MEASURED (a2 - 2). . 171


Ixvi















LIST OF FIGURES (cont*'d)


FIGURE Page

46 COMPARISON OF THE THEORETICALLY PREDICTED
AND THE MEASURED 2 a F . . 172

47 THE UFSA Rl CORE p 2 DISPERSION LAW . 173

48 DECAY CONSTANT vs. AXIAL POSITION . 177

49 Wk vs. AXIAL POSITION . 178

50 REACTIVITY (-$) k eff vs. AXIAL POSITION . 182

Cl Through C19 TIME PROFILES OF THE THERMAL
NEUTRON FLUX AT NINETEEN POSITIONS IN THE
CORE FOR A 0.5 MSEC INPUT PULSE . 195-213

C20 Through C38 TIME PROFILES OF THE THERMAL NEUTRON FLUX AT NINETEEN POSITIONS IN THE CORE
FOR A 1.0 MSEC INPUT PULSE . 214-232

D! Through D6 TIME PROFILES OF THE FAST NEUTRON
FLUX AT SIX POSITIONS IN THE CORE FOR A 0.5
MSEC INPUT PULSE . 234-239

D7 Through D12 TIME PROFILES OF THE FAST NEUTRON
FLUX AT SIX POSITIONS IN THE CORE FOR A 1.0
MSEC INPUT PULSE . 240-245

El Through E4 TIME PROFILES OF THERMAL NEUTRON
FLUX AT FOUR POSITIONS IN THE CORE FOR A 0.1
MSEC INPUT PULSE . 247-250

Fl TIME PROFILES OF THERMAL NEUTRON FLUX AT THREE
POSITIONS IN THE CORE FOR A 10.0 MSEC INPUT PULSE . 252

Gl Through G2 SHAPE OF THE PROPAGATING PULSE AS A
FUNCTION OF PULSE WIDTH . 254-255


xvii


























Di.


keff


Vji

z . . . .

a . . . .


p . . . .

P . . . .


. . .


LIST OF SYMBOLS


TRANSVERSE BUCKLING DELAYED NEUTRON PRECURSOR CONCENTRATION OF THE jth GROUP

DIFFUSION COE FFICIENT OF THE jth GROUP FREQUENCY (cps) EFFECTIVE MULTIPLICATION CONSTANT NEUTRON LIFETIME NEUTRON MULTIPLICATION VELOCITY OF THE jth GROUP AXIAL COORDINATE DECAY CONSTANT (IN THE TIME DOMAIN) DAMPING COEFFICIENT (IN THE FREQUENCY DOMAIN) EFFECTIVE DELAYED NEUTRON FRACTION PHASE SHIFT PER UNIT LENGTH REACTIVITY

COMPLEX INVERSE RELAXATION LENGTH MACROSCOPIC CROSS SECTION NEUTRON FLUX


xviii














Abstract of Dissertation Presented to the Graduate Council
in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy



SPACE-TIME REACTOR KINETICS STUDIES WITH
THE UNIVERSITY OF FLORIDA SPERT ASSEMBLY By


Nils J. Diaz


March 1969


Chairman: Dr. M. J. Ohanian
Major Department: Nuclear Engineering Sciences

A large-in-one-space dimension, side reflected, highly multiplicative (k ef-0.99) subcritical assembly was designed and calibrated. The sole purpose of the facility is the experimental investigation of the dynamic behavior of large reactor cores and to provide a test for space-time kinetics models presently in use. With this facility the linear aspects of neutron physics phenomena can be investigated in the absence of inherent feedback effects. This work was conducted under a subcontract with the Nuclear Safety Research Branch, Atomic Energy Division, Phillips Petroleum Company, under a prime contract with the United States Atomic Energy Commission.

The University of Florida SPERT Assembly (UFSA) is a light-water moderated subcritical facility fueled by 4.81% enriched U0 2 pellets encased in stainless steel tubes of 0.465 inch outside diameter (SPERT F-1 Fuel). The fuel arrays are contained in a rectangular tank, 8 feet


xix










long, 39 inches high, and of variable width. In this study, the core was 6.5 inches wide and 30 inches high. The effective multiplication. constant of the assembly was determined to be 0.990+.003. The assembly is equipped with nuclear instrumentation capable of automatic scram action.

For the kinetics studies, a fast data acquisition system was

developed to handle accurately the very high, time-changing count rate encountered in the measurements. It essentially consists of a transformer-coupled pulse amplifier to produce a fast logic signal at the input of a multichannel analyzer from the input signal originating in a long, thin He 3 counter. The instrumentation adequately handled count
6-
rates up to 3 x 10 counts/sec at the peak of the pulses. A high degree of reproducibility and fidelity in following the pulse profiles was obtained with this instrumentation.

The space-time kinetics studies were performed by analyzing the

propagation of a fast neutron burst introduced at one end of the assembly, in the absence of inherent feedback effects. The experimental .results are compared with the results obtained from the two-group, space-time dependent, one-dimensional diffusion theory scheme known as the WIGLE program. A stringent test of the model is provided by a, combined analysis in the time and the frequency domain.

The WIGLE calculational scheme accurately predicts the delay times and the attenuation of the pulses when a first-flight spatial distribution is assumed for the fast source. At large distances from the source WIGLE underpredicts (- 8% in the FWHW) the spreading of the pulse. A marked sensitivity to small changes in the transverse buckling was

-found for the model, as well as the experiment., xx










A one-to-one comparison of the predicted and measured values in the frequency domain was provided by performing identical numerical Fourier transformations of the WIGLE time profiles and the measured pulse shapes. The analysis in the frequency domain confirmed the results obtained in the time domain, although discrepancies past 100 cps are found in the ultrasensitive p 2 plane. The agreement in the p plane, the system's dispersion law, is good up to 200 cps and reasonable up to 800 cps. Both theory and experiment showed a smooth behavior
2
throughout the frequency range investigated, in both the p and the p plane.

Spatial effects in large cores are clearly demonstrated in this work. The determination of the range of applicability of the onedimensional scheme requires extending the study to cases in which twodimensional effects will be noticeable and the important feedback effects can be considered.


xxi




































PART 1

THE UNIVERSITY OF FLORIDA SPERT ASSEMBLY
- DESIGN AND CALIBRATION -















CHAPTER I


INTRODUCTION


The development and construction of large power reactors focused the attention of industry and of the United States Atomic Energy Commission on the necessity of having reliable reactor dynamics analysis methods to accurately describe the spatial and temporal behavior of the neutron flux in these systems. The point-model reactor kinetics calcu-, lations seem to have been adequate for the gross evaluation of the timedependent neutron flux during the occurrence of a transient but the model can be in large error when the physical size of the system and the magnitude of the perturbation necessitates that spatial effects in the redistribution of the neutron flux be considered. Preliminary calculations done by Johnson and Garner using a one-dimensional space-time kinetics model [1] showed that the space-time dependent scheme predicts a "destructive zone" much larger than that predicted by point-model kinetics.

The necessity of experimentally determining the validity of the

various space-time kinetic analysis methods was brought out by Johnson and Garner [1], and recognized by the USAEC in establishing the Large Core Dynamics Experimental Program. The primary responsi bility for this program has been vested in the Nuclear Safety Research Branch, Atomic Energy Division, Phillips Petroleum Company as major contractor for USAE C.









The Large Core Dynamics Experimental Program is to be performed in three phases:

Phase I. Pulsed Source Experiments'in Subcritical, Multiplying Media,
Large in'One-Dimension

The first phase of the experimental program is to be conducted in a close-to-critical subcritical assembly, 8 feet long, 3 feet deep and with widths changing from 6.5 inches to 16 inches according to the core configuration and whether bare or side reflected cores are studied.

The experimental information obtained from studying the pulse

propagation phenomena in this assembly is to be used to test the validity of current space-time kinetic models in the absence of inherent feedback effects.

Phase II. Kinetic Behavior for Control-Rod-Induced Power Excursions in
Large, One-Dimensional Cores

A reactor large in one-dimension, 16 feet long, three feet deep

and with varying widths to accommodate different metal/water ratios will be used to investigate the one-dimensional kinetic behavior of large cores subjected to a large perturbation. Both non-feedback (low power experiments) and self-shutdown measurements will be conducted. Phase III. Kinetic Behavior for Control-Rod-Induced Power Excursions
in Large, Two-Dimensional Cores

The same type of measurements performed for the one-dimensional core will be conducted in a two-dimensional core.

The measurements should provide the necessary information to

establish the validity ranges for one-dimensional models, the basis for the development of a two-dimensional scheme and as a bridge to the complex, three-dimensional problem.

Phase II and Phase III of the research program will be performed









at the SPERT IV facility at the National Reactor Testing Site, Idaho.

The appropriate existing experimental equipment, as well as the

extensive kinetics studies conducted by the Nuclear Engineering Sciences Department at the University of Florida, was conducive to the granting of a subcontract by the Phillips Petroleum Company so that the basic, linear kinetics studies of Phase I could be performed at the University of Florida.

The research to be performed as Phase I of the Large Core Dynamics Experimental Program can be succinctly defined as the experimental and analytical determination of the dynamic behavior of the neutron flux in slightly subcritical water moderated assemblies of SPERT F-1 fuel rods.

The facility in which the required measurements for Phase I, Large Core Dynamics Experimental Program, are to be conducted necessitated a thorough design and safety analysis. The assemblies are to be close to critical and the core has a large U235 inventory. The nuclear capabilities of such systems were the object of a detailed study to determine their operational characteristics under normal and accident conditions.

The flexible mechanical design, the safety instrumentation, the nuclear evaluation, as well as the experimental calibration of the first configuration under study constitutes Part 1 of this dissertation.

The reactor kinetics studies performed in the first of the configurations to be studied are dealt with in Part 2 of this manuscript.















CHAPTER II


DESCRIPTION OF THE FACILITY


General Features


The University of Florida Spert Assembly is a light watermoderated subcritical facility fueled by 4.81% enriched U0 2 pellets encased in stainless steel tubes of 0.4655" outside diameter. The fuel arrays are contained in a rectangular tank, 8 feet long, 39 inches high, and of variable widths. The system is designed so that both bare and reflected cores can be studied. Only one reflected core will be dealt with in Part 2 of this manuscript; information on three reflected cores is included in this chapter. The assembly width and fuel spacing may be varied in order to:

a) have a k eff not to exceed 0.99 in all cases to be considered.

b) accommodate non-moderator/moderator ratios of

0.5, 1.0, and 1.5, respectively.

Shown in Table I are the k ef's as a function of the moderator

height, the fuel spacings, core widths, and total number of fuel elements for the different reflected cases to be considered. The calculational procedures used in the determination of the nuclear parameters and the k efvalues for the three reflected configurations of the assembly are described in Appendix A. Only the sides of the assembly will be reflected. Fig. 1 shows an overall view of the facility.

















































DUMP VALVES


FIG. 1 OVERALL VIEW OF THE FACILITY





Effective Multiplication Constant


TABLE I


keff vs- MODERATOR LEVEL OF UFSA REFLECTED CORES


Calculated Using the AIM6 Code Core Length = 243.8 cm (96") Reflector Width = 30.48 cm (12") Active Fuel Height = 91.4 cm (36")

Metal to Water Ratio 0.5
Lattice Pitch (in) 0.7152
Core Width (cm) 16-35
No. of Fuel Rods 1206 -


1.0
0.584 19.28 2132


1.5
0.5332 25.73 3420 -


Moderator Level
(cm)

20 25

30 35

40

45 50 55

60 65

70 7 5 80 85

91.4


.7270

.7941 .8416 .8761 .9020 .9218 .9372

.9494 .9592 .9672 .9738 .9793

.9840 .9880 .9922


.7695 .8288 .8715 .9022

.9249 .9420 .9551 .9655 .9738 .9805 .9870 .9906

.9944 .9977

1.0012


.7397 .8028 .8471 .8793 .9032 .9213 .9353

.9465 .9554 .9626 .9686 .9736 .9777 .9813 .9851









The assemblies are highly multiplicative; this is important for

the extrapolation of the results of the study to critical systems. The system's subcriticality is attractive because of the inherent safety of such systems and of the absence of inherent feedback effects.

In order to provide as "clean" a core as possible, a unique control system which has been successfully used on the UFAPA [2] will be employed. In this system the reactivity is controlled by adjustment of the water height in the assembly. The water height is controlled by the position of two "V"-notched weirs located in a water "box" hydraulically coupled with the assembly through flexible lines. The quantity of water discharging through a "V"-notched weir varies as H 52(H is the distance between the apex of the weir and the water level) thus providing precise control of the moderator height. The hydraulic coupling assures that under normal operating conditions (with continuous flow) there will be the same water level in the core and the reflector tanks.

The UFSA subcritical assembly is located in an isolated and

shielded room in the Nuclear Research Field Building, approximately three miles from the University campus.

The Nuclear Research Field Building consists of four bays, two of them having shielded rooms for experiments with subcritical and moderating assemblies. The shielded walls consist of stacked concrete block eight feet high and thirty-eight inches wide covered with plywood to assure that the blocks remain in place. The ceiling of this single-floor building is approximately 15 feet above the floor and consists of excelsior-filled cement bonded board. Neutron reflection from this ceiling. over the walls does not constitute a hazard to personnel operating the accelerator-type neutron source. The access door from the control room









is interlocked with the neutron generator and the subcritical assembly scram system, as is the door on the only other entrance to the shielded room from the fuel storage area. Across the front face of the assembly, a screened wire. cage with a lockable door controls access to the core.

While not in use in the assembly the fuel is stored in a room adjacent to the facility, built entirely for this purpose.

A more complete desc ription of the facility and its characteristics can be found in the Design and Hazards of the UFSA and its addenda [3, 4, 5].

The system has been licensed under Atomic Energy Commission SPECIAL N-UCLEAR MATERIAL LICENSE SNM 1050, March 1968. The license allows for the possesion of 5400 fuel rods with a total U235 inventory of 190 kgs.


Fuel, Characteristics

The UFSA is fueled with Spert F-1 type fuel elements provided by the Phillips Petroleum Company.

The fuel characteristics are:

Fuel Composition: UO 2in pellet form

U enrichment: 4.81 + .15%

Active Fuel Length: 36" + .062"

Active Fuel Diameter: .42" + .0005"

Fuel Tube Material: stainless steel

Fuel Tube Length: 41.625"

Fuel Tube o. d.: .4655" + .0025"









Mechanical, Design

The entire assembly can be divided into three components: the supporting platform and dump tank; the basic core tank and fuel rod support structure; and the combination core side walls and side reflector tanks.

The supporting platform is composed of 5 inch steel I-beams, raised

5 feet from the floor level by six steel columns 3 1/2 inches in diameter. The column footings rest within a 6x8x2 foot steel rank which serves as a reservoir for the continuous water flow system and as a dump tank. Under normal operating conditions, this represents a minimum distance of about 4 feet between the bottom of the assembly and the water surface in the reservoir. This distance is sufficiently large so that the bottom of the assembly is considered to be unreflected under all conditions.

All the core and reflector hardware is made of type 5456-H321

aluminum. The bottom of the basic tank is made of a 24x96x3/4 inch plate bolted to the steel I-bean platform. The underside of the plate is covered with a .030 inch thick Cadmium sheet. The lower fuel rod support assembly rests on the plate. The end walls of the basic tank are made of a 24 x 39 3/4 inch plate and are supported by two 2 1/2 x 2 1/2 x 1/4 inch steel angle braces welded to the I-bea'm platform. Two fourinch aluminum channels span the eight foot dimension of the tank, confining the upper fuel rod support system and detector mounts.

The side walls of the basic tank serve also as reflector tanks when the reflected cores are under study. These tanks have dimensions of 12 x 96 x 37 inches. The arrangement allows one to vary the width of the core with a sole structu ral support.









.The end walls are permanently covered with Cadmium on the outside surface while the side walls have movable Cadmium covers to define the boundaries for the bare and reflected cases. To optimize the numberof neutrons inserted into the assembly by the neutron generator, the accelerator target penetrates about 4 inches into the core. A water-tight port is provided for this purpose. The port can be removed and a blind flange inserted in its place. Several fuel rods must be taken out, the number depending on how deep the target goes into the assembly and on the. lattice pitch.

The core section of the assembly consists of an interchangeable

fuel rod spacing system made of 3/4 x 1/2 x 1/8 inch channels, 5/8 x 1/4 inch bars and aluminum shims mounted on the base plate of the tank. The bars have milled slots to accommodate the .25 inch end tip of the fuel rod and to set the pitch along the core width. The shims are placed between the channel bar units to set the pitch along the core length (96 inches) (see Fig. 2). The top fuel rod spacing system consists of an aluminum grating. The mesh is determine d by the lattice pitch under study. The grating is made of aluminum bars and spacers, as shown in Fig. 3. Thus, fuel rod removal along the length of the core is possible to locate the detector for the experimental measurements.

The one-half inch long rod tip is fully surrounded by aluminum,

with practically no reflecting characteristic, but there is a 7/8 inch length of rod between the end of the active fuel and the tip which is surrounded by water. This bottom reflector is unavoidable and will be considered in the calculations.

























FUEL ROD





WIDTH SPACE




WIDTH SPACE
CHANNEL


LENGTH SPACER
SHIM


FIG. 2A BOTTOM FUEL ROD SPACING SYSTEM



































,,WIDTH SPACER




-WIDTH
CHANNEL
SPACERS


BASE PLATE CADMIlU.M


FIG. 2B BOTTOM FUEL ROD SPACING SYSTEM


-- ).r � � � _" . ]6 .






14


ALUMINUM SUPPORT
FRAME


FUEL ROD


CORE WI DTH FUEL ROD SPACING GRATE


COPRE LENGTH
/ FUEL ROD
SPACING
GRATE


SPACER TIE ROD


FIG. 3 TOP FUEL ROD SPACING SYSTEM









ModeratotrFlow*Control System

The moderator flow control system of the UFSA can be better described by'the water flow schematic shown in Fig. 4. Besides the normal fill and drain functions for the moderator, it serves as an accurate reactivity control using adjustable moderator height by continuous flow.

The characteristic components of this system are described below.

A. Storage Tank: A 6 x 8 x 2 foot steel tank located directly

below the assembly will serve as the reservoir for the circulating light-water moderator and as a du mp tank. Normal water heights while operating will be between 6 and 12 inches. The tank also serves as a footing for the assembly supports. This arrangement makes a very convenient and compact facility.

B. Core and Reflector Tanks: As seen from the flow diagram, water is pumped from the reservoir to the core and reflector tanks through a manifold at one end of the assembly and flows from the other end of the

core and reflectors tanks to the weir "box". From the weir "~box"t, water flows over the weirs back to the storage tank through a flexible line.

The core section is equipped with two normally open solenoid

activated dump valves, 3 inches in diameter, located at each end of the core. These valves provide the reactor with a fast shutdown safety system. The reflector tanks have their own 1 1/2 inch normally open solenoid valves actuated by the same safety system.

Since the quanti ty of water discharging through a V-notched weir varies as H51 where H is the height of the water level above the apex of the V-notch, the water level and hence, the reactivity, can be controlled in a precise manner simply by varying the height of the weirs and the rate of flow of water into the tank. This is accomplished by






,WEIR DRAIN ,- DISTRIBUTION MANIFOLD


ION EXCHANGE BED


-J


FIG. 4 REACTIVITY-CONTROL FLOW SYSTEM


CONTROL VALVE


ORIFICE









an automatically operated pneumatic control valve. A plot of the flow rate versus the height of the water level above the apex of the weirs is shown in Fig. 5.

The weir plate is rigidly mounted on a "box" or small tank (see Fig. 6). The weir "box" is connected to a driv e mechanism composed Of the following: guide post, slide block, and drive screw. The guide post is a 2 inch diameter pipe attached to the support column of the crane which Is used for removal of the reflector tanks. The "box" is mounted on the guide block which slides along the vertical post and provides a rigid support for the system and is driven up and down by means of the drive screw which is fixed at the top of the guide post support and passes through the guide block. The upper limit of the position of the weir"box"is controlled by mechanical stops whose position is determined as part of the initial start-up procedure for each configuration to be considered.

The final adjustment of the position of the weir"box"is such that when the water reaches the moderator level in the assembly corresponding to k eff 99 for a given configuration it will be flowing about 2.0 inches above the apex of the V-notch weirs. At this design level the flow rate is - 7 gallons per minute with the k eff values as given in Table I for the full fuel loading. After the operating height of the weir box has been determined for.k eff < 0.99 in the initial start-up, stops are inserted to prevent raising the weirs above this height (if the height is less than the active fuel height). It should also be pointed out that the orifice in the line limits the pump capacity to a flow rate which is just sufficient to bring the weirs to full flow. A further increase in flow rate would cause discharge over the entire

















2.8


2.4
$~4
S 2.0.

0
< ~1.6.


1.2-0





0.4/



0 2 4 6 8 10 12

Flow Rate (gpm) FIG. 5 FLOW RATE vs. HEIGHT OF WATER LEVEL ABOVE WEIR APEX













4



3.5"
�L


1.5" O.D.x 1/8 WALL(3)


2" 0.D.x 1/8 WALL


FIG. 6 WEIR "BOX"









perimeter of the weir "box" into the drain line effectively preventing any further increase of the moderator level in the assembly.

Th e measurement of the water height in the core is accomplished by fixing a reference mark on the slide block at the same level as the bottom of the weir within the "box". An accurate scale is provided to read off the distance between the bottom of the core and the apex of the weirs. Continuous indication of the moderator level in the core is provided on the console by means of a recorder calibrated between the bottom of the active fuel and the maximum moderator height and by a manometer, connected directly to the core, for precise measurement of the water level. These two measuring systems insure reproducibility of .the moderator height for the experiments.

The water is pumped out of the storage tank by a constant speed centrifugal pump which has a "no load" capacity of 20 gal/min. The control valve is designed to restrict the flow to the maximum design value of 12 gal/min. A deionization system is provided to keep the water as pure as possible at all times. The pneumatic flow control system consists of two differential pressure cells, transmitters, control valve, and recorder-controller. The strip type chart recorder-controller records both flow rate and moderator level in the core. The control valve is of the air-to-open type which will close in the case of air sup ply loss, stopping the flow into the assembly y. The flow diagram for the air system is shown in Fig. 7. A pressure differential from the pressure transmitters applied to the recorder-controller allows both manual and automatic control of the flow rate through the valve operated by the controller.








SOLENOID
TO INTERLOCK


L


DUMP VALVE


DIFFERENTIAL PRESSURE TRANSMITTER


FROM PUMP


I T
--.--ORIFICE --- .TO CORE


WATER AIR


FLOW VALVE


FIG. 7 AIR SYSTEM SCHEMATIC


CORE


I I
I I









Instrumentation and Interlock System

The instrumentation and interlock system of the UFSA has been discussed extensively in the reports submitted to the Atomic Energy Commission [3,-4, 5] in conjunction with the license application. More recently, Mr. L. B. Myers submitted a detail technical report on the subject [6]. A brief descriptive explanation is given below.

A block diagram of the safety system logic flow in use at the UFSA subcritical assembly for routine monitoring is shown in Fig. 8. There are five principal channels of instrumentation:
3
a. Start-up channel using a He proportional counter, scaler, and rate meter. The counter is located at the bottom of the core, close to the geometrical center of the assembly.

b. Log power and period instrument No. 1 channel using a compensated ion chamber (operated uncompensated) as a signal to a Log N amplifier. The chamber is located along the longitudinal axis of the assembly, on~ the bottom of the core some two feet from the neutron generator end.

c. Period instrument No. 2 channel using a compensated ion chamber (operated uncompensated) as a signal to a log N amplifier. The chamber is located along the longitudinal axis of the assembly, on the bottom of the core, some six feet from the neutron generator end.

d. Linear neutron flux No. 1 channel using an uncompensated ion chamber as a signal to a micromicroammeter. The chamber is mounted on the top core support frame, close to the geometrical center of the core.

e. Linear neutron flux No. 2 channel using a compensated ion

chamber (operated uncompensated) to feed a signal to a micromicroammeter. The chamber is mounted on the top core support frame on the opposite



























II IIl
And Manual And
CKT. SCRAM I CKT.

And I And
CKT. CKT. I

And And
CKT. CKT.


uimp I And Core Dump IAnd
SCore CKT. I Valve # 2 CKT.
Width K.
Refl. Dump Pump

ValvesFIG. 8 UFSA SAFETY SYSTE LOGIC FLOW DIAGRAM

FIG. 8 UPSA SAFETY SYSTEM LOGIC FLOW DIAGRAM









side from the linear channel No. 1.

Items a. through d. are part of the safety amplifier while item f. is used to display the neutron flux on a console front panel meter.

The safety amplifier monitors the seven continuously varying input signals and provides a trip signal if any of the input signals fall outside of acceptable limits. The safety amplifier provides means of adjusting these limits over a wide range.

The duality of the scram action (see Fig. 8) is a prominent feature of the safety system. It can be said that no single failure will invalidate both automatic scram channels. Furthermore, it has been

determined that no single failure can invalidate both the manual and automatic scrams. The method of measurement and the function of each instrumentation and safety channelare shown in Table II (parts A and B).

A series of safety interlocks prevent water from flowing into or remaining in the assembly unless a proper sequence of events are followed and certain conditions are satisfied. The conditons are: a. The moderator temperature must be > 60*F. This is established by the desire to obtain the experimental data near room temperature conditions. The insertion of water at 32*F will introduce a maximum k of .00342 (based on the calculated negative temperature coefficient of reactivity) above the design k efvalue with no hazards created.

b. The four instrumentation channels must have their high voltage

*on.

c. The core width must be smaller than 26 cm.

d. The door to the assembly room must be locked.

e. The start-up channel must count more than 2 counts/sec.

f. The neutron flux, subcri tical assembly power level and period










TABLE II


Measured Parameter a. Low level neutron flux b. Linear neutron flux


c. Lg netronfluxb c. Logea neutron flux




e. Reactor period1b f. Gamma flux g. Detector power supply
voltage


h. Reactor period 2


UFSA INSTRUMENTATION AND CONTROL
A. NUCLEAR


Method of Measuirement

He3 detector - pulse discriminator; at neutron generator end of core


CICa - ammeter; on side of core


CIC a _ log N and period amp; under core near center line UIC -- ammeter; on side of core


CIC a _ log N and period amp


Ion chamber - area monitor; on front of reactor cage


Unijunction transistor oscillator and relay. Monitor detector voltage for b, c, d and e CIC a - log N and period amp; under core near center line


App-icatitop

Insure source is present before adding reactivity. Scram on low count rate

Indicate power level scram on power

Indicate power level scram on high power; log N recorder

Indicate reactor power; linear N recorder

Indicate reactor period; scram on short period

Criticality monitor for storage room. Area monitor for reactor room. Activate evacuation alarm

Scram reactor on low detector voltage


Indicate reactor period; scram on short period


To be operated in the uncompensated mode Common detector and instrument
















Application

Scram reactor on low reactor inlet water temp. (600F)

Scram reactor if water level in core exceeds top of weir height


Scram system and shut down neutron gun if reactor doors are opened or interior switches are activated

Prohibit filling reflector tanks when distance between tanks exceeds widest reflected core
width

Stops pump when water reaches llcm below weir apex

Indicates water is filling reflector tanks

Requires closing valve before starting pump


Measured Parameter a. Reactor water temperature

b. Reactor water level



c. Reactor door and personnel



d. Reactor core width




e. Reactor water level f. Reflector tank water
level (low)

g. Flow control valve
shut


TABLE II (Continued)


B. NON-NUCLEAR


Control flow rate. record flow rate


Indicate and


h. Reactor flow


Method of Measurement

Fenwall temperature switch in inlet line

Barksdale pressure switch mounted on weir"box"with sensing line connected to coreLimit switches on doors and push buttons inside reactor room



Limit switches on reflector tanks




Barnstead pressure switch senses level in weir box Float switches in reflector tanks


Limit switch on valve


Differential pressure cell and. pneumatic control D/P cell and pneumatic system


Indicate and record water level


i. Reactor water level










must be as specified under Operating Limits in this report.

g. After a normal start-up, the water height inl the core must be within 0.5" of the level set by the position of the weirs.


Fuel Storage

The large amount of fuel needed for the experimental program required a detailed criticality analysis of the fuel storage area. Criticality considerations of the fuel storage arrangements follow the Atomic Energy Commission regulations regarding the subject. Three different criteria were used to calculate the effective multiplication of the fuel storage area to assure that the array will remain subcritical under the worst circumstances. The methods and the corresponding conditions are outlined below.

The fuel storage array consists of three slabs of air-spaced fuel pins, separated by a minimum distance of 54 inch face to center. The fuel is stored in steel baskets containing 308 pins per basket. Two sets of 1/4 inch thick plastic plates, located at the bottom and top of each basket, drilled to properly position the fuel rods. The characteristics of the fuel slab are:

Slab Width = 3.73 in = 9.46 cm (corresponds to 7 fuel pins in

transverse direction)

Slab length = 14 ft

Height =3 ft (active fuel height) a. Multiregion-multigroup calculation

Thie effective multiplication factor of the fuel storage array consisting of three slabs (3.73 inch wide, 14 ft long and 3 ft wide) which are 54 inch apart (face to face) was computed for the case of flooding









the storage area to the level of the active fuel height. A 2-foot reflector on both sides was used to represent an infinite reflector. No reflector was considered on the ends, but the contribution of this to the system would be small. The calculation was done using four groups and seven regions and followed the method outlined in Appendix A of this thesis. The following configuration, which is symmetric about the indicated center line, was assumed:


24"


3.73,


Water


Fuel Water


Under these water-moderated and reflected conditions, a k =f

0.79 was obtained

b. Solid'angle criterion for slabs

To determine the interaction between the fuel stor age slabs in the proposed 3-slab array, the solid angle criterion established in 10 CFR, Part 70, 70.52, paragraph (b) was used. This establishes the maximum









total solid angle subtended by any unit in the array to be 6 steradians if the effective multiplication factor for the individual slabs is less than 0.3, as is the case here.

From 10.CFR 70.52 (d) (2) (i) the minimum required separation distance, i.e., center of one slab to face of adjacent slab, is obtained from:

3 stera dians - cross sectional area (separation distance)2


This gives, separation distance ~14 x 3 )1/2 =.4ft.
3 ) =37
We propose to establish a minimum distance between faces of adjacent slabs of 4.5 ft. This gives a total solid angle for the center slab of 3.88 steradian, which is well below the established criteria. c. Comparison with Clark's criteria

For 5% enriched, 0.4 inch diameter uranium oxide rods with a 190.13 gm/liter U-235 concentration (compared to 4.81% enriched, 0.42 inch diameter uranium oxide rods with a 200 gm/liter U-235 concentration in our case) the following data is obtained from pp. 39 and 59, respectively, of DP-1014:

Width of Slab (cm) Buckling (cm )
Critical Safe Critical Safe
11.2 10.4 0.014945 0.015912

It should be noted that the slab widths quoted on page 39 of DP-1014 are for an infinite water-moderated and reflected slab. From the buckling values given, the critical width of the infinite water-moderated unreflected slab is 25.7 cm; the corresponding safe width is 24.9 cm. When twice the reflector savings for the latter case as given on page 59 of DP-1014 is subtracted, a safe width for the infinite, water-










moderated reflected slab of 10.44 cm is obtained consistent with the 10.4 cm. value.

Thus the slab width of 9.46 cm proposed by us compares favorably from the safety viewpoint with the safe width for an infinite, watermoderated and reflected slab and is considerably narrower than the safe width for an infinite, water-moderated and unreflected slab. Within the present context it should also be pointed out that as indicated on page 54 of the Design and Hazards Report [3], no flooding of the storage area seems possible from natural causes.


Neutron Sources

Two types of neutron sources were used throughout this work.

1) Two Pu-Be sources mounted in an aluminum cylinder which can be driven remotely through a plastic pipe from a shielded box located in one corner of the facility room to underneath the center of the core. Neon lights provide indication at the console of the position of the sources. These sources, which have a combined yield of - 3.2 x 10 6 n/sec are used for start-ups and for the inverse multiplication m easurements.

2) A Texas Nuclear Neutron Generator which is used in continuous mode for static measurements and in the pulsing mode for the pulse propagation measurements. The generator is of the Cockcroft-Walton type, TNC Model 150-1H with continuously variable high voltage from 0-150 kv and has been modified to obtain larger currents by removing the einzel lenses and installing a new 22 electrode accelerator tube and gap lense. Pre- and post-acceleration beam deflection produces sharp, low-residual pulses.






31


The accelerator was used with a 4-5 curie tritium target.

The position of the target can be changed to keep the source centered on the target-end of the assembly for any given moderator level.















CHAPTER III


OPERATIONAL SAFETY


Introduction


The University of Florida SPERT Assembly, due to its large size, enriched uranium-oxide fuel and nuclear potentialities required a thorough study of its capabilities, operational characteristics, initial loading procedures and of the behavior of the assembly under accident conditions. The study was part of the requirements established by the Division of Material Licensing of the USAEC prior to the granting of an operating license.

Legally, a subcritical assembly has to comply with regulations

under 10 CFR Part 70 "Licensing of Special Nuclear Materials" since no self-sustaining nuclear reaction is envisioned. In the case of the UFSA, however, the Commission felt-that certain technical sections of 10 CFR Part 50,, which deals with nuclear reactor licensing, should apply and serve as a guide for the design and the safety analysis.

The basic philosophies employed in the design of the system were:

a) The UFSA facility has been designed to remain subcritical under normal operating conditions.

b) The safety instrumentation (see Part 1, Chapter I) has been

designed such that a single failure will not invalidate both the manual and automatic scram and will not cause subsequent failures.

c) The design basis accidents were postulated on a single failure 32









criterion.

d) Operating limits have been set to delineate the normal operating ranges of the assembly.

e) Initial loading procedures have been established to determine the safe operating multiplication factor of each configuration.

A series of administrative controls are necessarily applied to all segments of the experimental program and strictly enforced.


Initial Loading

A series of calculations were done to determine which of the two following schemes should be employed for the initial loading of UFSA:

I) Step loading of the fuel from the center out, accompanied by step increases in water level with the usual inverse multiplication determination.

II) Loading all the fuel into the dry tank and proceeding with a careful evaluation of the multiplication as a function of water level.

Since the.UFSA core is very loosely coupled as far as the lumped

reactivity parameter is concerned, the second method was selected due to the fact that a better determination of the multiplication was possible from a basic moderator height-zero loading inverse counts determination. The slope of the k eff vs- water height curve has a slope substantially smoother than the k eff vs. per cent fuel loading (full water height) curve.

The moderator level control system in operation at the facility

provides a extremely reliable and safe mode of adjusting the water level without safety compromises.

The following regulations were followed for the initial fuel









loading, and will be followed for subsequent cores:

1) After the fuel has been loaded, prior to each new incremental change in the water level, the water is drained completely and the weir (and water level scram) adjusted to prevent a level increase beyond the desired value.

2) The first three measurements of the inverse multiplication are obtained at water heights of approximately 20 cm, 25 cm, and 30 cm above the bottom of the active fuel. As shown in Table I, the maximum k eff calculated for a water height of 30 cm is 0.87. Subsequent filling increments are not to exceed the least of the following:

a. An increase in water height of 10 cm.

b. An increase in water height which, by extrapolation of the

inverse multiplication curve, would increase the k by one-half of the eff
amount required to make the assembly delayed critical.

c. An increase in water height which would, by extrapolation of

the inverse multiplication curve, result in a k - of 0.990. For values eff
of k eff above 0.95, the k eff 0 f the assembly arealso determined by pulsed source techniques.

3) At each filling step, the measured k eff.of the assembly is

compared with the calculated value. If significant deviations of the experimental values from the calcula ted k eff vs. water height curves occur, the experiments are to be discontinued, and a detailed analysis of the results obtained performed. If it is determined that the dimensions of the assembly should be changed in order to achieve the desired k at maximum water height, the University of Florida will apply for eff
and obtain written approval from the AtomicEnergy Division of the Phillips Petroleum Company before such changes are made.










Operating Limits

A description of the operating limits of the subcritical assembly, including the basis for such limits ate listed below. Effective Multiplication Factor

Specification: the maximum allowable k efwill be 0.985+.005.

The absolute value of k efas well as the slope of the k vs. " water

height will be carefully measured so as not to exceed the limiting value.

Basis: 'the upper limit of k =f 0.985+.005 is established by:

the accuracy with which k efcan be measured, the reported [7] differences between calculations and experiments in similar cores and the value of the multiplication factor required to make a meaningful study of the dynamic properties of large cores. Comparison [7] between 29 cal~ulations and the corresponding critical experiments (on cores similar to the UPSA) established that an overestimate of k efis generally made; the standard deviation for these cases was + 0.00175 and the maximum understimate of the multiplication was for a case yielding kef 0.9966, a

0.34% deviation.

The absolute value of k efwill be determined for water levels

yielinga keff :*95by pulsed techniques independently of the inverse multiplication measurements. The method to be used is the GarelisRussell[8J method which, when appropriate corrections are made for the reflector, has been shown to give good results. In this method both
a kf~0 and are determined; then keff may be obtained by

independently obtaining $ or k. Since of the two parameters changes the most slowly with k eff' it is valid to use a value based on theoretical calculations.









Reactivity Addition Rate

Specification: the reactivity addition rate is controlled by the

water flow rate into the subcritical assembly. The maximum flow is fixed to be 12 gpm. At this maximum flow rate, the rates of addition of water, and consequently or reactivity, computed between water heights of 30 and 45 cm. are:


0. 5332"


Lattice Pitch

0.584"


0. 7152"


Rate of increase of
water level 0.0435 0.0439 0.0431

Ak/cm 0.0093 0.00865 0.008

Ak/sec 0.0004 0.0038 0.000344

$/sec. 0.057 0.054 0.049

It should be noted that these reactivity addition rates are a large overstimate compared to the calculated rates at k eff - . 98.

Basis: the maximum rate of addition of reactivity was established by the calculated values of k efvs. water height and the maximum flow rate. The values specified above constitute an upper limit in the region of interest and are considered to be safe under circumstances.

The flow rate is a function of the capacity of the pump, the orifice and the pneumatic control valve and cannot exceed 12 gpm. Reactivity Removal Rate

Specification: a conservative value for the reactivity removal rate is taken from the slope of the k efvs. water height curves near the maximum designed k -- values. Since no difference is detected for the scram times of the three configurations, only one rate of removal will be specified,, corresponding to the smallest slope.










Drain rate, including the system

reaction time . 2.65 cm./sec,

Rate of removal of reactivity . 0.00037 Ak/cm.

. 0.00098 Ak/see

. 0.140 $/sec

it should be noted that the value of Ak/cm used to specify the reactivity removal rate is - 20 times smaller than the corresponding value specified for the reactivity insertion rate.

Basis: the rate of drainage from the assembly established the reactivity removal rate. Consistent with the approach taken when specifying the reactivity, the time to drain 45 cm of water from the assembly was measured to establish a lower limit on the rate of decrease of water height. Even under these extreme assumptions, i.e., using the maximum slope of-the k eff vs- water height curve for the insertion rate, the small calculated slope around k efC* 99 for the ,removal rate and the reduced pressure head, the reactivity removal rate is three times larger than the insertion rate. Subcritical Assembly Power Level and Power Level Scram

Specification:

Power Level . 0.5 watt Power Level Scram . 1.0 watt

Basis: with the maximum source strength available and with k eff

0.985 the maximum average power was calculated to be .130 watts by taking into account the spatial dependence of the flux and a steady source at one end.

Assuming that a reactivity accident occurs, based on the maximum reactivity addition rate specified above, the design basis accident









predicts-that from a k eff of 0.993 it would take ap proximately 15 seconds to double the power level. Even if the power level indicator were not to scram the system until the power level reached 10 kw, assuming a one second del ay time to actuate the scram, the power would increase to only 69 kw by the time the scram actually begins. Based on the reactivity removal rate described above, the power level would then decrease rapidly to a very low value. Period Scram

Specification:

Period scram . 15 sec

Basis: A positive period will be obtained in the subcritical assembly for any addition of reactivity beyond a given steady state condition. if the maximum reactivity insertion rate of 0.05 $/sec is considered, the initial positive period is about 50 sec and decreases monotonically with time. The period channel has been determined to respond reliably to periods = 50 sec. Originally, the period scram was set at 50 sec but repeated scrams caused by the start-up of the neutron generator forced the reduction of the scram period to 15 sec for operational reasons. Average Neutron Flux and Neutron Flux Scram

Specification:

Ave neutron flux for

most reactive core . 1.5 x 105 n/cm 2 _ sec Neutron flux scram . 3.0 x 10 5 n/cm 2 _ sec

Basis: the above are based on the average power calculated for the assembly. Doubling of the flux will occur when 0.75 $ worth of reactivity is added to the system from any design subcriticality level. Under these conditions, even at the maximum k eff status, the facility










will still be subcritical and the instrumentation will have sufficient time to scram the system before accidental criticality occurs.


Design Basis Accident Analysis

Two types of accidents have been postulated to occur in conjunction with the UFSA experimental program. The design basis accident analysis will deal with the following two cases a) a dropped fuel rod accident in which a fuel rod is broken releasing the fission products accumulated during previous operation of the assembly; b) an accidental criticality resulting in a power excursion. It will be shown below that the potential consequences of either accident are well within the radiation dose limits specified by Title 10 CFR Part 20. Dropped Fuel Rod Accident

For the purposes of this analysis, it'is assumed that a dropped

fuel rod results in a broken pin releasing the accumulated fission products. The following bases have been established to calculate the radiation dose from such an accident:

a. It is stipulated that the assembly has been operated continuously for 200 hr at an average power level of 0.065 watt. It should be noted that this is a conservative value since the assembly will not be operated on a continuous basis; the 200 hours of operating time for each configuration is essentially spread over a period of several weeks.

b. The assembly contains 1200 rods. Since the actual loadings for the three reflected cases are 1206, 2132 and 3430 rods respectively, this will result in conservative results.

c. The peak to space-averaged flux ratio during operation is a maximum of 120, with the rod at peak flux contributing 1/10 of the









total power of the assembly.

d. The fuel rod subjected to the specified highest specific power is broken open and the fuel is completely fragmented, releasing 100% of the gaseous fission products.

e. The iodine (in elementary form) diffuses uniformly throughout the 18' x 32' x 15' room to calculate the on-site internal dose. The off-site dose was calculated using the very conservative Pasqual's principle of atmospheric diffusion.

The fission products inventory was calculated by means of the RSAC code [9). The results indicated that the iodine isotopes were the only significant contributor to the inhalation dose. Using the active worker breathing rate of 3.47 X 10 m /sec specified in 10 CFR 20, a person remaining in the room would accumulate a thyroid dose of 0.023 mr for each minute he remains in the room, after the hottest rod breaks open. The total dose to a person that inhales all of the iodine contained in the rod would be 0.2 r. The established radiation safeguards at the University of Florida require the personnel to abandon the area immediately and notify the Radiation Safety Officer. The maximum time required to evacuate the assembly and thefuel storage area is 10 sec with 30 sec needed to evacua te the entire building; these times have been measured during practice evacuation of the building.

The off-site total inhalation dose, calculated assuming the iodine is released in one puff with zero wind velocity, cloud inverted condition, were typically:









DistAnd&'fr6-, ite (m) Total dose (mrem)

200 4.5 x 10-4

500 2 x 10-4
1000 7 x 10-6

The direct radiation dose from all the fission products of the irradiated fuel elements is calculated to be 1.3 mr/hr. Accidental Criticality and Subsequent Power Excursion

Accidental criticality and a subsequent power excursion could occur only by the uncontrolled addition of water to the assembly and gross error in the calculations and/or procedures. The occurrence of such an accident is highly improbably and would require:

a. Setting a core width corresponding to the 1.5 metal/water ratio and proceed to install the fuel spacer system for the 0.5 metal/water ratio, load the fuel under these conditions and disregard small water height increments and l/M measurements. To do this, several administrative rules would have to be wilfully ignored. and/or

b. Improper setting of the following: the weir height, the water height scram system for any configuration and a gross error in the calculations causing criticality at about half the design core height (91.44 cm). It should be noted that the calculational method used to. predict that k eff values has been tested successfully against the results of 29 different critical assemblies [7]. and/or

c. An obstructed 2 inch line from the core to the weir drain system and a simultaneous failure of the water height scram together with the referred to error in the calculations.









d. Violation by the facility operator of the administrative

procedures requiring a visual check of the assembly before start-up and continuous attention to the control console instrumentation to determine the status of the assembly at all times.

The consequences of such an accident were determined by assuming the following:

1. The assembly is initially at a steady state power level of 10 watts (normal average power is .065 watt).

2. Water flows into the system at the maximum rate of 12 gpm.

3. The reactivity addition rate is 0.05 $/sec. This rate is larger than the calculated rate (.04 $/sec) for the case discussed above and more than that calculated to occur near critical for properly loaded assembly.

4. The power scram is set, through calibration or other error., at 10 kw. Corresponding error settings occur for the neutron flux and period scram.

5. A scram occurs 1 sec after a power level of 10 kw is reached. This time has been determined as the elapsed time from the initiation of a scram signal to the opening of the dump valves.

6. No feedback effects are considered. This is a good approximation to our case due to the low power levels involved and again is a conservative assumption.

The calculations were made using the IREKIN code, described in reference [10]. IREKIN numerically solves the point model kinetics equations.

Starting with the assembly one dollar subcritical (k 993),
eff
and proceeding with the described excursion, the accident would yield










a peak power of 69 kw, the total energy release is 0.5 Mwsec. The power vs. time behavior of the assembly for the postulated power excursion is shown in Fig. 9.

The combined neutron and gamma dose to the operator is 1.5 rem

assuming that: the energy release is instantaneous, all neutrons have an energy of 1 MeV and there is no attenuation in the assembly. The proper RBE factors were taken into consideration.

It is concluded that, even if such an improbable accident would occur, the hazards to personnel and the general population are not significant.







7x104


3. - At t=0: k eff=0.993, Power=O

Ramp Insertion of .05$/sec
2 - Scram 1 sec after 10 kw
Scram Rate = .14$/sec
Total Energy Release=0.5Mwse


Time (secs)


FIG. 9 POWER vs. TIME FOR THE DESIGN BASIS ACCIDENT















CHAPTER IV


THE DATA ACQUISITION SYSTE24


Introduction


In the delicate and laborious task of performing nuclear experiments the most common source of difficulties and errors lies in the acquisition of the data. Modern nuclear instrumentation with its excellent time-energy resolution has enhanced detection sensitivity to the extent that deviations previously masked by the poorer resolution of the equipment are easily distinguishable. The major problem now lies in the degree of reproducibility of the results. Although in general the instrumentation is very reliable the enhanced sensitivity demands continual standard ization for the sake of reproducibility.

The "brain" of the data acquisition system for pulsed neutron

measurements is the multichannel analyzer (MCA) which is now available with a large number of data channels and narrower channel widths for increased time resolution. In general, the mode of data acquisition for a pulsing experiment differs from that of many other types of nuclear experiments. In particular, neutron interactions with a suitable detector are fed into the analyzer while it is time-sweeping. Ideally every neutron interaction should be counted regardless of the an-plitude of the collected pulse. This implies that the linear signal originated at the neutron detector should be converted to a logic.signal so that its probability of being recorded is independent of its amplitude.

45









Logic signals have fixed amplitude-shape characteristic and convey information by their presence, absence of time relationship with respect to a reference signal. Only gamma and noise discrimination are necessary; this process will inherently eliminate.the counting of weak neutron interactions.

The process of registering events whose density per unit time

varies considerably requires an electronic system with extremely high t ime resolution and a broad frequency response. Typically the counting rates at the peak of the neutron burst in pulsed neutron measurements vary from 10 5 to > 10 6 event/sec. This high, time-varying count rate caused drastic losses in conventional instrumentation due to pulse pile-up and circuit induced distortion when the dynamic range of the circuit is exceeded causing a net amplitude shift in excess of that tolerable by the system. The alternative modus operandi is to use low count rates so that resolution losses are minimized; this, however, increases the probability of systematic errors due to much longer running times.

The count rates encountered during the initial experiments with a
3
He counter close to the neutron generator target in the SPERT assembly exceeded 10 6 event/sec at the peak. S saturation effects were observed in modular instrumentation incorporating the latest F.E.T. preamplifier and double-delay line pulse shaping amplifier at about 10 5 event/sec. The fast pulse train caused a drop in the base voltage of the Field Effect Transistor which prevented further counting until a significant reduction in the count rate allowed the voltage to recover. Changing to a fast scintillation prototype preamplifier improved the situation somewhat. However, if the potential count rate in the UFSA assembly









were efficiently handled with few losses the actual running time of a particular experiment could be reduced to a few minutes. This possibility forced the development of the ultra-fast instrumentation channels used in this work.

The electronic counting system used throughout the experimentation used a transformer-coupled pulse amplifier to produce a fast logic signal to the input of the MCA from a slower input signal originating in a ' 3
He counter. The system has superior counting and stability characteristics. The resolution time of the system is practically nil compared to that of the MCA. In general, data acquisition times were reduced to a few minutes; in that time it was possible to obtain, at most of the detector locations, the following counting statistics:

Peak Channel Counts: 2 17

Counting Time: 2 - 10 minutes

No. of data channels: 1024

Channel Width:, 20 micro sec

Pulse repetition rate: 30 hz

Noise Level: 40 event/min

Background Level: 140 event/min


The Neutron Detector

The neutron detectors used throughout the experimental program were
3
proportional counters filled with He . The counters are of the Texlium variety made by Texas Nuclear Corporation and were specially built to conform to the UFSA core grid. Detectors of this type have been



1. The assistance of Mr. Joel B. Ayers of ORTEC, Inc.,who suggested
this electronic instrumentation is gratefully acknowledged.









preferred over the BF 3 variety at the University of Florida because of the larger neutron absorption cross section and operational reliability.
3
He undergoes the following reaction

He 3 + n H 3 + p + 0.764 Mev.

The cross section for this reaction is 5327 + 10 barn at v 2200 10 3
m/sec compared to 3840 +.11 barn for B . He follows a 1/v law in the energy range from 0 to 200 kev. The pulse heights yielded by a He 3 filled counter are proportional to the energy of the neutrons plus 764 kev. The reaction has been used for neutron spectroscopy from the 100 kev to 2 Mev energy range. Gamma discrimination can be easily accomplished by the use of a biased integral discriminator.

The detectors were long and thin and each took the place of a fuel element in the core. The active length of the counter is slightly less than the active length of the fuel. 'Two one atmosphere (predominantly thermal detection) and one ten atmosphere (more responsive to higher energy neutrons) detectors were used in the experimental program. A sketch of the physical characteristics of the counters is shown in Fig. 10. The thin, long cylindrical shape enhances the tire characteristics'of the counter. Although normally a 1 atmosphere He 3 detector operates with a bias voltage of 1000 volt, the minimum input pulse voltage requirement of the pulse transformer was such that the operating voltage of the counters had to be raised to 1200 volt. At this voltage the slope of the counts vs. high voltage curve was about 8% per

volt; therefore very stable, low ripple high voltage supplies were used to insure reproducibility of the detector response.













HV BNC Connector


.- 5LI.25" I


ive Volume Stainless Casing









.~ .247"



33" 2.375"


40.5" 0. 5" -


FIG. 10 PHYSICAL CHARACTERISTICS OF THE He3 NEUTRON COUNTERS


dw










The'Electronic Instrumentation

A block diagram of the instrumentation used in the pulse propagation experiments is shown in Figs. 11 and 12. Two independent data acquisition systems with a common start-stop clock are necessary to carry out the measurements: (1) a system connected to a movable detector that obtains the time profile of the propagating pulse at a given position and (2) the all important normalizing detector, fixed at one position. Two 1 atmosphere He detectors with the characteristics described above were used for these purposes. The system is essentially composed of signal transmitting devices and data registering and handling units.

The movable detector data acquisition system (MDDAS) consists of:

1. Detector - High Voltage Power Supply

The high voltage power supply was an ultrastable FLUKE 405 B with superior stability and negligible ripple. Manufacturer's specifications state the stability at .005% per hour and the ripp le at less than 1 my RMS.

2. ORTEC Model 260 Time-Pickoff unit, with 3000 volt isolation

The time-pickoff units are normally used to detect the time of

arrival of a detected particle, usually with subnanosecond precision. The use of this characteristic and the electronic arrangement shown in Figs. 11 and 12 permitted the counting of neutron events with excellent time resolution. To our knowledge this is the first time that the timepickoff units have been used for this application.

Briefly, the system operates as follows: the primary of a toroidal transformer, having a bandpass for very high frequencies only, is inserted between the detector and the bias power supply. Fast















RTC 260 TIME-PICKO
LORTEC 260 TIME-PI CKOFj


FIG. 11 MOVABLE DETECTOR DATA ACQUISITION SYSTEM













I 2 M_ I
ORTEC 260 TIME-PICKOFFJ


FIG. 12 NORMALIZING DETECTOR DATA ACQUISITION SYSTEM









components of the detector signal will actuate a wide band transistor amplifier and tunnel diode discriminator from the secondary of the transformer. A line driver buffer is also provided. The power and control bias are provided by the ORTEC 403 Time Pickoff Control.

3. Modified ORTEC 403 Time Pickoff-Control

The 403 Time-Pickoff Control provides control and fan-out buffering for time derivation units such as the 260 Time Pickoff Unit. The fan-out buffer accepts the fast negative logic signal from the 260 unit and provides either fast negative or slower positive logic output signals.

The 403 Time-Pickoff Control had to be modified to make its output signal compatible with the input requirements of the 212 Pulsed Neutron Logic Unit used with the MCA. The 212 plug-in unit requires pulses with rise times longer than 50 nsec while the positive output from the 403 control unit has a rise time 1 10 nsec. Furthermore, the positive logic signal from the 403 control has a 0.5 microsec pulse width which is too wide for the time resolution required. Therefore two modifications were made: a capacitor2 (C12) was changed from 270 pf to 68 pf reducing the pulse width to 0.12 microsec and an inductor coil (Ll) was removed from the system changing the rise time to 50 nsec.

The modifications made the system compatible with the input requirements of the MCA and avoided overdriving the analyzer.





2. Refer to ORTEC, Inc. "Instruction Manual for 403 Time-Pickoff
Control."









4. Technical Measurement Corp. 1024 Multichannel Analyzer

(MCA) and Data Handling UnitsThe time analysis of the pulse and the corresponding data output are obtained from the following coupled instrumentation:

Pulsed Neutron Logic Unit, TIMC Model 212

Digital Computer Unit, TMC Model CN1024

Data Output Unit, TMC Model 220C

Digital Recorder, Hewlett-Packard Model 561B

Binary Tape Perforator, Tally MAodel 420

A trigger from an external pulse generator starts the sweep of the analyzer as controlled by the pulsed neutron logic unit and initiates, the burst at the accelerator. The 212 has variable channel widths from 10 to 2560 microsec. The full 1024 channels were used in all the measurements. The storage capacity of the CN1024 is 2 17counts. Both printed paper tape and perforated binary tape were obtained as output.

5. ORTEC 430 Scaler

For some of the flux traverses and the 1/M measurements, the integral counts were recorded in a 10 Mhz scaler.

The normalizing detector data acquisition system (NDDAS) consists of the same components as the movable detection system except that a 256 multichannel analyzer, TMC Model CN110 was used for data handling and the integral counts were accumulated in an ORTEC 429 Scaler, modified for 7 Mhz counting rate.


The Resolution Time'Correction

The resolution time of the pulsed neutron data acquisition system was the parameter under consideration while searching for a well-matched









and fast electronic instrumentation set-up. Some of the findings were surprising and may explain some of the discrepancies found in previous pulse propagation and neutron wave experiments which used an accelerator type neutron source. In these experiments it has been customary to increase the neutron yield of the genetator as the movable detector is positioned farther away from the source to keep running times within tolerable limits. The count rate at the movable detector is then maintained at a level which has been determined to be acceptable; however the normalizing detector - usually fixed in a position close to the source (normally at the "thermalizing tank" if one is used) - may then be exposed to count rates b eyond the capabilities of the detector and the associated instrumentation. Since the movable detector signal was always fed into an analyzer, saturation effects could be easily detected; the signal from the normalizing counter, however, was always being fed into a scal er where saturation effects may go unnoticed since a time profile is not available.

During most of the present work, two MCA's were available and the phenomenon described above was observed and count rates throughout the experimental program limited to those permissible by the time characteristics of the counting system.

The experimental program being-carried out with the UFSA subcritical calls for the one to one comparison between theoretically calculated and exp erimentally determined time profiles of the neutron flux as a function of space. The shape of the pulse, rather than conventional integral parameters extracted from it, is therefore the significant and required information. For this reason the influence of the resolution time correction on the shape of the pulse came under close









scrutiny.

The resolution time of several, submicrosecond preamplifier-linear amplifier combinations was shown to be a function of the mode of operation and of the count rate. For this type of instrumentation, there is a significant difference between the resolution time as determined by a steady-state technique (two source method) and a dynamic method (maximum count rate method [11]), the latter method yielding a much larger resolution time.

For the instrumentation finally chosen to carry out the pulsed experiments in this work no significant difference was found between the results of the two methods. Furthermore the pulse transformeramplifier combination behaves like an ideal non-paralyzing system. The resolution time of the overall data acquisition system was found to be primarily determined by the multichannel analyzer.

As a matter of illustration, when the system depicted below was used the resolution time changed from -0.24 microsec as determined under low count rate steady-state conditions to -8 microsec as determined under high count rate, pulsing mode conditions.









Shown in Fig. 13 are two time profiles obtained at the same position in the UFSA core for the same count rate. The conventional preamplifier-linear amLplifier system shown above was used to obtain one of the profiles and the instrumentation selected for this work was used to obtain the other. Both were corrected for resolution time losses with the best available value for the resolution time. The distortion observed in the pulse shape obtained with the conventional preamplifierlinear amplifier system is non-linear due to the count-rate dependent resolution time. It should be noted that the count rate near the peak of the pulse was less than 10 count sec. The conventional system response "flattens out" when it reaches complete saturation - in this case above 10 count sec. Saturation effects are not observed in the
6
pulse transformer system until the count rate exceeds 3.3 x 10 count sec. An erroneous resolution time correction, or one which used a resolution time that does not characterize the system throughout the range of count rates will change the shape of the neutron pulse and consequently affect any analysis done on the pulse shape obtained.

As pointed out by Bierman, Garlid and Clark [11], in a pulsing experiment it is necessary to determine whether the counting system has the characteristics of a purely non-paralyzing system or those of a mixture of paralyzing and non-paral yzing systems. Using the method described by Bierman and co-workers, it was established that the data acquisition system being used in the present work is, as close as can be determined, non-paralyzing. The resolving time of the system is essentially determined by the width of the input pulse to the analyzer and the MCA characteristics.

It should be noted that for a wide range of count rates, depending





58



40












30


Conventional








Time-Pickoff r 204





10




10







30 microsec between channels




I I I I I I I I I
0 4 12 16 24 32 40 48 56

Channel Number FIG. 13 TIME PROFILES OF NEUTRON BURST RECORDED BY CONVENTIONAL
ELECTRONIC INSTRUMENTATION AND BY THE TIME-PICKOFF SYSTEM










on the magnitude of the resolving time, it is not necessary to properly identify the counting system since no significant difference in the corrected counts are observed when using the paralyzing or nonparalyzing correction. This is shown in Fig. 14.

The resolution time correction for the UFSA data acquisition system is then given by: No
T 1 Noxtr
1
CWxTrs

where

NT = true number of events in a given channel

N = observed number of events in a given channel

tr = resolution time of the system

CW = channel width

Trs = number of sweeps of the analyzer and, for a non-paralyzing system

tr = reciprocal of the maximum observable count rate

Using this method the following resolution or resolving time factors were determined:

A) MDDAS including 1024 Multichannel Analyzer tr = 0.31 + .015 microsec B) MDDAS excluding Multichannel Analyzer tr 0.20 + .01 microsec C) NDDAS including 256 Multichannel Analyzer tr = 0.56 + .02 microsec D) NDDAS excluding Multichannel Analyzer tr 0.20 + .01 microsec




I-w


700










0
600




S500



0
z 400

S- Non-paralizable

o 300- D Paralizable

S. O 80% non-paralizable
20% paralizable

0
g 200




100





0 100 200 300 400 500 600 700 800 900 1000

True Count Rate NT (x 103) FIG. 14 THE PARALIZING, NON-PARALIZING SYSTEM RESOLUTION TIME CORRECTION AS A FUNCTION OF COUNT RATE









The Notmalization'Technique

The analysis of a pulse propagation experiment will yield information on the velocity of propagation, the attenuation and pulse shape as a function of position of a propagating disturbance. To determine the attenuation or relative amplitude of the pulse at different positions in an assembly, the data must be normalized to a fixed reference so that variations in the source strength, data acquisition time, etc., can be properly accounted for. The usual technique requires accumulating integral counts in a scaler for each measurement and reducing all measurements to the fixed reference afterwards. As was mentioned previously, the resolution time correction can have a significant effect on the normalizing factor since widely differing count rates are employed. It is extremely hard, if not impossible, to obtain a proper resolution time correction based on integral counts determined from a time-varying count rate.

Two methods of normalization and their relative merits are discussed below.

The Integral Count Method is the normally employed method of normalization. The total counts accumulated with the NDDAS for all runs is referred to a predetermined one, with the normalizing detector in a fixed position while the movable detector changes position.

The Analyzer Method, which is essentially the same as the above

except that the counts arising from the normalizing detector are stored as a function of time in a multichannel analyzer. The average of a series of ratios obtained by dividing the corrected channel counts by the corresponding channel counts of a predetermined measurement gives the normalizing factor.









The Analyzer Method is intrinsically more accurate than the Integral Count Method because it permits an "exact" correction for resolution time losses by the use of the expression previously given applied to the recorded time profile. The correction for the integral counts is, on the other band, inaccurate since the count rate is continuously changing and no base exists for a resolution time correction.

It was found, however, that as long as the count rate near the

peak of the pulse is kept well within the capabilities of the NDDAS no significant difference is observed in the results of the two methods. This is due to the fact that ratios are being taken'in both cases; this tends to minimize whatever differences there might be. Thus, it is concluded that with proper care the Integral Method is adequate whenever an analyzer is not available for normalization purposes.

It should be noted that an "effective" resolution time can be used to improve the results of the Integral Method. This "effective" resolution time can be found by forcing the normalizing ratio obtained from two runs by the Integral Method to match the normalizing ratio obtained from the Analyzer Method for the same two runs by adjusting the resolution time correction applied to the integral counts.


Comments

Certain inconsistencies in the results of the first few experi-m ments prompted a careful inspection of the multichannel analyzer modus operandi. For the sake of completeness the significant findings are listed below.

A) Operation with the 10 microsec channel width (selected by the settings of the 212 plug-in-unit) proves to be unreliable due to instabilities, in the clock and gating circuits. Channel widths of 20









microsec or longer are stable.

B) The address current setting of the CN1024 is extremely critical, markedly so for high count rates.

C) An optimum pulse into the analyzer should have a rise time of 50 nanosec, a total width of .1 microsec and an amplitude of 3-5 volt.

D) Above noise level, the discriminator setting of the 212 logic unit becomes irrelevant when a constant pulse height is used as input.

E) Reproducibility tests were performed on the analyzer with the neutron generator in continuous mode.

The statistical analysis of the channel counts gave:

72% were less than 1 from the mean

25% were between 1 and 2 from the mean 3% were between.2 and 3 from the mean Th e system is statistically well-behaved.
















CHAPTER V


NUCLEAR CALIBRATION OF THE UFSA SUBCRITICAL


Introduction


The nuclear calibration of the University of Florida SPERT Assembly was performed prior to the space-time kinetics studies. The calibration involved conventional inverse multiplication measurements, absolute k determination by pulsing techniques and comparison with multigroup-multiregion diffusion theory calculations. The GarelisRussell tec hnique was employed to determine kaIZ and this result used to calculate k efby coupling-it with the experimentally determined decay constant and the theoretically calculated effective delayed neutron fraction. During this phase of the experimentation, "spatial effects" were noted in both a and k~t These effects and other interesting kinetic phenomena involving these basic reactor parameters were considered worthy of further study and were investigated during the main part of the research. They are discussed in Part 2, Chapter V. In this section the results pertinent to the necessary calibration of the system are given.


Theoretical Notes

The Inverse Multiplication Method

Under ideal conditions, usually met only in small-fast assemblies, the reactivity can be represented by 64











1k I lorlk l/M

where M is the net neutron multiplication in the assembly with a centrally located source. In practice, the multiplication is obtained from the ratio of multiplied to unmultiplied counts with a centrallyr located source. The unmultiplied counts are obtained with the fissile material removed and all other conditions undisturbed. In watermoderated cores it is difficult to match neutron spectra for multiplied and unmultiplied counts and deviations from the ideal M are to be expected. If possible, a search for detector locations should be conducted so as to obtain curves that follow the expected behavior of l/M. Even if k can not be directly inferred from the lIM determination, the curve of reciprocal count rate vs.'th e parameter that controls reactivity (fuel loading or moderator height or % control rod withdrawal) is a useful guide for safely approaching criticality if a well-behaved curve can be obtained.

The inverse multiplication curve can be obtained as a function of moderator height by first obtaining a series of unmultiplied counts at

various water levels and the multiplied counts as the water level is raised with the assembly originally air-spaced. Sensitivity to geometrical configuration (source-detector-water level) requires an empirical determination of "well-behaved" detector positions. Reactivity Measurements by the Pulsing Technique

The pulsed-neutron technique has been used successfully for several years to measure reactivity. The transient neutron density following a burst of neutrons is used to determine the reactivity of the system by

either the Simmons and King method [12], Sjostrand's area ratio method









[13], Gozani' s extrapolated area-ratio analysis [14] or the GarelisRussel technique [8]. In all these techniques it is essential that a fundamental spatial distribution of the neutrons be established for a correct determination of the decay constant and, therefore, the reactivity of the system.

The Siramons and King method established that a value for the

reactivity can be obtained directly if a prompt fundamental decay constant can be measured at delayed critical. The value of a at delayed critical determines /2. and if these parameters are assumed constant over the reactivity range of interest a value of a can be obtained. The technique has given good results up to - $20 subcritical in small multiplicative systems. The method strongly depends on being able to establish the prompt fundamental decay mode; it suffers from the inconvenient necessity of a delayed critical measurement and the assumed constancy of a/ throughout the ranges of reactivity.

The Sjostrand method improves the Simmons and King method in that the delayed critical measurement is no longer necessary but the resul ts

are shadowed by the strong influence of higher spatial harmonics. The method is based on the premise that the impulse response curve of the system is dominated by the prompt fundamental mode.

Gozani's treatment is a significant improvement over Sjostrand's method. Gozani proposed the extraction of the fundamental mode of prompt neutron decay from the impulse response curve and the extrapolation of this curve to zero time. The reactivity in dollars can be found by integrating under this curve; the method is independent of the presence of higher prompt spatial modes.

The Garelis-Russeiltechnique, similar to Gozani's extrapolated










area-ratio method, is of practical value because of its intrinsic elimination of the effect of prompt higher harmonics. This method, which was used in the present work, was postulated originally for a repetitively pulsed (with a delta function source in time), bare, monoenergetic reactor but has proven to be of broader application. Garelis and Russellpostulate, that for the conditions specified above: 1/R 1/R
fNp exp((k /Z)t)dt = fNpdt + Nd/R where

Np = prompt contribution to the neutron density Nd = delayed contribution to the neutron density

R = pulsing rate

The following conditions should be satisfied for the correct application of the method.

a) R >>X, where X is the decay constant of the shortest lived precursor group.

b) R >> a, where a is the prompt fundamental decay constant.

c) The system must be pulsed a sufficient number of times so that exp (ma sn/R) � exp (-a sn/R), where m + 1 is the total number of pulses and the a are the roots of the inhour equation.
sn

d) The prompt root dominates the decay.

The Garelis-Russeltreatment permits the determination of p ($) when all the above conditions are satisfied, by the relation: p($) = k 1
ka /P,

An absolute value of p is obtained by the use of a calculated effective delayed neutron fraction. Garelis has discussed the use of the method in reflected systems; the technique seems to be of practical value in









these systems [15].

Becker and Quisenberry were able to compute a correction [16] for the observed spatial dependence of the reactivity in two-region systems by recognizing the differences in the spatial distributions of prompt and'~delayed neutrons. Their excellent comparative study of the above techniques emphasized the need for their recommended spatial correction unless the neutron detector is properly positioned to minimize this correction.

The study of Garlid and Bierman [17] correctly points out that in very large systems "an asymptotic spatial distribution cannot be established before the pulse has decayed away, since the asymptotic mode is one that is uniform everywhere in space." They proceed to apply a combination of first flight, age, and time-dependent diffusion theory to the study of pulsed measurements in large aqueous media; their conclusion is that their measured apparent decay constant is a good approximation to the asymptotic value and that pulsed measurements in very large multiplying systems may also give good results.


Inverse Multiplication Measurements

The safe approach to the design value of k <.f:i99 was undertaken

with the conventional l/M measurements until k -f 0.95 and then by both

the l/M and the pulsing technique.

To establish detector positions free from geom~etrical effects

(source-detect or-water level), six different locations were used until a water level of 60 cm (k - .98) was reached and four locations afterwxards. Two of the detector locations, the closest to the neutron source, failed to describe the multiplication of the system. Shown in Fig. 15 are the




- I ',

















GRID POSITION


34 101 84 59 42 1
A PI

DB0- gIc TNC
WEIRS F 0- GENERATOR
I II g
HO- *I &
G .- 0
H 0-


241.61

DISTANCE IN CM.


FIG. 15 DETECTOR POSITIONING SCHEME










locations employed for these measurements and for the absolute k ff determination by pulsing. Two 1 cu, centrally located Pu-Be sources were used for these measurements.

The measurements were performed in the following sequence:

a) Uninultiplied counts vs. water level were obtained from a water level of 20 cm. to 91.4 cm above the bottom of the active fuel in steps of 5 cm.

b) The fuel was loaded in the assembly in the presence of two centrally located Pu-Be sources, with proper monitoring.

c) Multiplied counts vs. water level were obtained from a water level of 20 cm in 5 cm steps until an effective multiplication factor <.99 was reached. This procedure follows the criteria established for initial Loading of the assembly, Part 2,-Chapter III of this work; the

5 cm steps were more conservative increments than those specified for the Initial Loading of the assembly.

The results of these measurements are shown in Figs. 16 (A, B). All the curves were well behaved in the sense that none "nose-dived." Position 2 and 4 failed to properly describe the multiplication of the assemblies because of their nearness to the source. Position 1 seems to overestimate the multiplication, mainly because the unmultiplied count rate was extremely low at this position thus an apparently high ratio of multiplied/unmultiplied counts was obtained. A detector more sensitive to high energy neutrons was used in locations 1 and 6.

It should be noted that when the inverse multiplication is plotted vs. 1/H 2the predictions become qu ite linear much earlier than when plotted versus HI. Better predictions are therefore made with the 1/H 2 curves but the approach can become less conservative by underestimating













.6



.5 Rl Core
0 Position No 1 No 2
ANo 3
4 No 4
0 No 5
A No 6













0)





020 30 40 50 60 70 80


Moderator Level (cm)


FIG. 16A INVERSE MULTIPLICATION vs. MODERATOR LEVEL




w


V I


w


Moderator Level (cm)


30.5


4 8 12 16

(1/Moderator Level)2 x 10-4 (cm-2) FIG. 16B INVERSE MULTIPLICATION vs. SQUARED INVERSE HEIGHT









the multiplication.

Shown in Table III are the values-estimated for k effor the four locations that seemed to represent the system best. Position 3 gives a lower limit and Position 1 an upper bound. No attempt was made to establish the error associated with measured k ff' but it is believed that the 0.99 value obtained at the last water level is within +.005 of the true value.


Absolute Determination of keff

After an estimated value of k ef>.95 was obtained from the l/M measurements, an independent determination of k was required by the operating license at every new increment in the moderator level (as

determined by the criteria established in Part 1, Chapter 3). The technique chosen for this determination was the Garelis-Russeilmethod of measuring Wt/ and a simultaneous determination of the prompt fundamental decay constant.

Different detector locations were used to determine the influence of the source and of higher order harmonics contamination. Strong spatiall effects" were observed in both a and k /P. This phenomena will be discussed in detail in Part 2, Chapter V because of its importance. A seemingly true fundamental decay constant and "spaced-converged"

k$%were obtained at large distances from the source and were used to determine the reactivity of the system.

A Fortran IV, IBM 360 computer program named UNIPUL was coded to perform a unified analysis of the pulsed neutron data (see Appendix B). The program calculates the decay constant using Peierl's statistical analysis [18] and ko/2. using the Garelis-Russellapproach after the data













TABLE III


SUMMARY OF l/M AN'D PULSED MEASUREMENTS


UFSA R1 Core 0.5 MIW Ratio
16.35 cm wide reflected core

INVERSE MULTIPLICATION Moderatora k f


Height (cm) Pos. 1 Pos. 3


PULSED EXP
k ffb


Pos. 5 Pos. 6


20 .415 .481 .425

25 .744 .629 .705

30 .8671 .737 .824

35 .914 .794 .881

40 .944 .84 .908

45 .9612 .873 .934

50 .973 .895 .950

55 .98 .98 .962

60 .985 .936 .973

65 .9891 .95 .980

70 .9925 .96 .9853

75 .9949 .971 .989

80

85

91.4

Above bottom of active fuel Averaged from 3 detector positions


.581 .771 .865 .900 .927

.946 .961 .969 .971 .979 .986 .990


.948+. 01 .965+. 007 .972+. 006 .9 796+. 005 .9855+. 004 .990+.003


PREDICTED
k eff


.7675 .8287 .871 .9022 .925

.9419 .955 .9655

.974' .980 .986 .9906

.9944 .9976 1. 00117










has been resolution time corrected and background subtracted. A "pure" delayed neutron background is statistically calculated and used to determine Wt The data can be normalized to a reference detector position for the analysis of the pulse propagation measurements in the time and in the frequency domain. A Fourier analysis of the pulse can also be performed if required.

Shown in Fig. 17 (A, B) are the experimentally determined a, kO/t and k as a function of moderator height obtained by averaging results from three chosen detector location's in the "asymptotic" region. The results are summarized, together with the 1/14 measurements and theoreticlycalculated values in Table III. The excellent agreement between the experimental and theoretical results should be considered somewhat fortuitous. The calculations were done following the method outlined in Appendix A. Some later calculations [19] done by the Phillips Petroleum Co.,showed more disagreement, especially at low water levels.

The last calculations tried to account for the fact that there is a fissionable reflector above each experimental moderator height. This fact was disregarded in the calculational results shown in this work. The agreement at the 75 cm water level is good for all calculational. methods.


Conclusions

The University ofE Florida SPERT Assembly has been operated for several months with very few operational problems. The system has proven to be extremely reliable and the instrumentation has performed adequately. The calibration of the system has established that meaningful values for the reactivity can be obtained when applying the




w


1200


N


RI Core


N.


800


N
N


-S


400 -


200 -


I S I I


i i i i I


0 60 70 R1


Moderator Level (cm)


FIG. 17A DECAY CONSTANT vs. MODERATOR LEVEL


w


1000 -


0
5


* g t I




w


210 -


RI Core


/
1/


170 -


w


1.00




0.99 0.98




0.97




0.96 0.95


I I I I I


150


I I I S I


I I I


Moderator Level (cm)


FIG. 17B k8/ AND k vs. MODERATOR LEVEL


200 -


190 -


0


4=


J3


160 -


~1 /1


110


I I II !









Garelis-Russell technique to a reflected slab assembly when proper care is taken. Agreement between calculated and experimentally determined values of the reactivity is termed excellent but since only one case has been studied judgement on the overall applicability of the technique

to this type* of reactor configurations should be reserved until the other assemblies to be studied in the UFSA facility are duly analyzed. It should be pointed out that multiple detector positions are necessary to establish when an asymptotic decay constant is obtained, and that the value of k /k is affected by the input pulse width. As mentioned previously, a more detailed analysis of the pulsed neutron reactivity measurements is conducted later in this thesis.





































PART 2

SPACE-TIME REACTOR KINETICS STUDIES WITH THE UNIVERSITY OF FLORIDA SPERT ASSENE3LY




Full Text
rLUX (RELATIVE UNITS)
FIG. 23C PULSE SHAPES PREDICTED BY WIGLE AT DIFFERENT POSITIONS IN THE UFSA R1 CORE
112


FLUX (RELATIVE UNITS)
FIG. G1 SHAPE OF THE PROPAGATING PULSE AS A FUNCTION OF PULSE WIDTH
- EXPERIMENTAL -
256


FLUX (RELATIVE UNITS
TIME (MSEC)
FIG- E4 TIME PROFILE' OF THERMAL NEUTRON FLUX 142-8E3 CM FROM THE SOURCE
252


41
Distncfrom sit (m)
200
500
1000
The direct radiation dose from all the fission products of the
irradiated fuel elements is calculated to be 1.3 mr/hr.
Accidental Criticality and Subsequent Power Excursion
Accidental criticality and a subsequent power excursion could occur
only by the uncontrolled addition of water to the assembly and gross
error in the calculations and/or procedures. The occurrence of such an
accident is highly improbably and would require:
a. Setting a core width corresponding to the 1.5 metal/water ratio
and proceed to install the fuel spacer system for the 0.5 metal/water
ratio, load the fuel under these conditions and disregard small water
height increments and 1/M measurements. To do this, several administra
tive rules would have to be wilfully ignored.
and/or
b. Improper setting of the following: the weir height, the water
height scram system for any configuration and a gross error in the
calculations causing criticality at about half the design core height
(91.44 cm). It should be noted that the calculational method used to
predict that values has been tested successfully against the
results of 29 different critical assemblies [7].
and/or
c. An obstructed 2 inch line from the core to the weir drain
system and a simultaneous failure of the water height scram together
with the referred to error in the calculations.
Total dose (mrem)
4.5 x lO-4
2 x 10"4
7 x 10-6


36
Reactivity Addition Rate
Specification: the reactivity addition rate is controlled by the
water flow rate into the subcritical assembly. The maximum flow is fixed
to be 12 gpm. At this maximum flow rate, the rates of addition of water,
and consequently or reactivity, computed between water heights of 30 and
45 cm. are:
Lattice Pitch
0.5332"
0.584"
0.7152"
Rate of increase of
water level
0.0435
0.0439
0.0431
Ak/cm
0.0093
0.00865
0.008
Ak/sec
0.0004
0.0038
0.000344
$/sec
0.057
0.054
0.049
It should be noted that these reactivity addition rates are a large
overstimate compared to the calculated rates at k^^ ~ .98.
Basis: the maximum rate of addition of reactivity was established
by the calculated values of k^^ vs. water height and the maximum flow
rate. The values specified above constitute an upper limit in the
region of interest and are considered to be safe under circumstances.
The flow rate is a function of the capacity of the pump, the ori
fice and the pneumatic control valve and cannot exceed 12 gpm.
Reactivity Removal Rate
Specification: a conservative value for the reactivity removal
rate is taken from the slope of the k ^ vs. water height curves near
the maximum desiened k values. Since no difference is detected for
w err
the scram times of the three configurations, only one rate of removal
will be specified, corresponding to the smallest slope.


165
TABLE XVII
THE REAL AND THE IMAGINARY COMPONENTS OF
THE COMPLEX INVERSE RELAXATION LENGTH
- 1.0 MSEC INPUT PULSE -
FREQUENCY a (cm"1) ? (rad/cm)
(cps)
Theory
Exp
Theory
. Exp
0
.02444
.02346
,
10
.02330
.02324
.001500
.001499
20
.02343
.02344
.002987
.002980
30
.02365
.02361
.004438
.004427
40
.02394
.02393
.005963
.005951
50
.02429
.02437
.007338
.007319
60
.02469
.02484
.008655
.008623
70
.02512
.02535
.009920
.009864
80
.02558
.02597
.01115
.01104
100
.02665
.02724
.01343
.01320
120
.02775
.02861
.01559
.01516
140
.02883
.03004
.01739
.01688
160
.02995
.03148
.01918
.01846
180
.03111
.03292
.02081
.01995
200
.03220
.03438
.02230
.02129
200
.03330
.03577
.02380
.02257
240
.03444
.03714
.02515
.02377
260
.03547
.03834
.02640
.02493
280
.03650
.03959
.02770
.02605
300
.03764
.04068
.02883
.02712
350
.04013
.04288
.03160
.02925
400
.04246
.04465
.03409
.03138
450
.04496
.04896
.03634
.03362
500
.04708
.05121
.03840
.03568
550
.04920
.05420
.04031
.03771
600
.05122
.05625
.04200
.03814
650
.05329
.05813
.04363
.03881
700
.05528
.06113
.04501
.04008
750
.05746
.06276
.04638
.04389
800
.06000
.06396
.04777
.04848
900
.06530
.06692
.05230
.05410


148
TABLE XIII
DELAY TIMES AND FWHM FOR A NARROW INPUT PULSE
DISTANCE TO
SOURCE (cm)
PEAK AT
Theory
(msec)
Exp
FWHM
Theory
(msec)
ExP
53.9
0.60
0.580
1.35
1.62
79.3
1.08
1.06
2.06
2.28
130.2
2.27
2.26
3.20
3.51
142.9
2.50
2.53
3.47
3.75
Significant discrepancies between theoretical and experimental
pulse shapes were noted in the region where source effects are still
noticeable; in contrast, excellent agreement was found in this region
for the 0.5 and 1.0 msec input pulse cases. Deep in the asymptotic
region the measured pulse shapes are identical to those determined for
the 0.5 and 1.0 msec input pulses.
The results are somewhat inconclusive, however, since near the
source the measured pulse shapes appeared wider than the corresponding
shapes for the 0.5 and 1.0 msec input pulsed, although they occur
earlier in time. Unfortunately, the assembly was dismantled before
these comparisons were completed. The observed discrepancies deserve
further experimental study.
It should also be noted that noticeable oscillations were observed
in the WIGLE results in the spatial region close to the source.
Propagation of Wide Pulse
A series of measurements were also made in the clean core with a
10 msec wide input pulse. This pulse is much wider than the


55
and fast electronic instrumentation set-up. Some of the findings were
surprising and may explain some of the discrepancies found in previous
pulse propagation and neutron wave experiments which used an acceler
ator type neutron source. In these experiments it has been customary
to increase the neutron yield of the genetator as the movable detector
is positioned farther away from the source to keep running times within
tolerable limits. The count rate at the movable detector is then
maintained at a level which has been determined to be acceptable; how
ever the normalizing detector usually fixed in a position close to
the source (normally at the "thermalizing tank" if one is used) may
then be exposed to count rates beyond the capabilities of the detector
and the associated instrumentation. Since the movable detector signal
was always fed into an analyzer, saturation effects could be easily
detected; the signal from the normalizing counter, however, was always
being fed into a scaler where saturation effects may go unnoticed since
a time profile is not available.
During most of the present work, two MCA's were available and the
phenomenon described above was observed and count rates throughout the
experimental program limited to those permissible by the time charac
teristics of the counting system.
The experimental program being carried out with the UFSA subcrit-
ical calls for the one to one comparison between theoretically calcu
lated and experimentally determined time profiles of the neutron flux
as a function of space. The shape of the pulse, rather than conven
tional integral parameters extracted from it, is therefore the signif
icant and required information. For this reason the influence of the
resolution time correction on the shape of the pulse came under close


FLUX (RELATIVE UNITS)
FIG C32 TIME PROFILE OF THERMAL NEUTRON FLUX 155-53 CM FROM THE SOURCE
228


158
time domain were transformed to the frequency domain, with identical
numerical and fitting procedures, setting a firm basis for a one-to-
one comparison.
The following interesting numerical problems are worthy of note:
a) To determine to what extent the numerical transformation is
affected by the accuracy of the operations of the IBM 360 computer,
sample problems were run in single precision (4 byte words) and double
precision (8 byte words) using the MORWIG program. No differences were
found in the results up to the sixth significant figure, but the double
precision computations consumed almost twice the amount of time as
the single precision ones.
b) Initially, the results of every third WIGLE time calculation
was punched on cards for the comparison in the time domain and for the
input to the Fourier transformation code. Less than 250 time points,
with time increments of up to 150 microsec, were available for each space
point. The numerical transformation performed on these results was
somewhat unsatisfactory since not all the amplitude and phase curves
2
were smooth and in the p plane the scattering of points was significant.
The WIGLE calculations were then performed again, with smaller time in
crements, carried farther in time and every time step punched for input
to the MORWIG program. A total of 997 and 998 points respectively were
now available for the Fourier transformation for the 0.5 and 1.0 msec
input pulse cases. The largest time increment was 32 microsec. The
calculation was carried to about 19.2 msec after the initiation of the
pulse. A significant improvement was noted in the results of the numer
ical transformation. The results were smooth throughout up to ~ 800
cps, where a certain amount of scattering was observed.


Observed Count Rate
FIG. 14 THE PARALIZING, NON-PARALIZING SYSTEM RESOLUTION TIME CORRECTION AS A FUNCTION
OF COUNT RATE


101
"effective" core height is very close to 76 cm. The transverse leakage
was taken into account in the WIGLE scheme by changing the absorption
2
cross section of the fast and the thermal group with the proper D^B^
obtained from the above calculations.
The aluminum port in which the target section of the accelerator
is located created a large void in the source section of the core.
Shown in Fig. 20 is a plan and front view of this arrangement. To mock-
up the physical situation as closely as possible the input to the WIGLE
code consisted of parameters for two different types of regions: one
for the normal core and one for the region depicted in Fig. 20. The
parameters calculated for the UFSA core were then volume-weighted
together with two group parameters for aluminum and light water as a
first approximation to the parameters in the source region (Region 1 in
the calculations).
In Fig. 21 the type of material, number of mesh points and the
mesh spacing used in the calculations is shown. Table IV shows the
value of the parameters for Region 1 and for Regions 2, 3, 4, 5 respec
tively. Before the final runs were made the importance of the mesh
spacing was investigated by running two identical cases with 166 and 99
mesh points respectively. No difference was found between results of
these two cases.
The choice of time increments was more critical. When time steps
of 5 ysec were used in the vicinity of the time where the value of the
4
fast source was changed from 10 to 0 severe oscillations developed at
the first space point. To avoid this problem, time steps of 0.2 to 0.5
ysec were employed in this time region and the magnitude of the source
was reduced in two steps; first to one-half the original value and then-


CHAPTER II
DESCRIPTION OF THE FACILITY
General Features
The University of Florida Spert Assembly is a light water
moderated subcritical facility fueled by 4.81% enriched U0£ pellets
encased in stainless steel tubes of 0.4655" outside diameter. The fuel
arrays are contained in a rectangular tank, 8 feet long, 39 inches high,
and of variable widths. The system is designed so that both bare and
reflected cores can be studied. Only one reflected core will be dealt
with in Part 2 of this manuscript; information on three reflected cores
is included in this chapter. The .assembly width and fuel spacing may be
varied in order to:
a) have a k not to exceed 0.99 in all cases to be
eff
considered.
b) accommodate non-moderator/moderator ratios of
0.5, 1.0, and 1.5, respectively.
Shown in Table I are the k _'s as a function of the moderator
eff
height, the fuel spacings, core widths, and total number of fuel ele
ments for the different reflected cases to be considered. The calcu-
lational procedures used in the determination of the nuclear parameters
and the k^^ values for the three reflected configurations of the
assembly are described in Appendix A. Only the sides of the assembly
will be reflected. Fig. 1 shows an overall view of the facility.
5


LIST OF FIGURES (cont'd)
FIGURE Page
35C THE SENSITIVITY OF THE ONE-DIMENSIONAL, TOO GROUP,
SPACE-TIME KINETICS SCHEME TO CHANGES IN THE
TRANSVERSE BUCKLING 143
36 EXPERIMENTAL PULSE SHAPES AS A FUNCTION OF CORE
HEIGHT 144
37A EFFECT OF ROOM RETURN AT PERIPHERAL DETECTOR
POSITIONS 153
37B EFFECT OF ROOM RETURN AT PERIPHERAL DETECTOR
POSITIONS 154
38 AMPLITUDE OF ZEROTH FOURIER MOMENT vs. DISTANCE FOR
SEVERAL FREQUENCIES
- 0.5 MSEC INPUT PULSE 160
39 AMPLITUDE OF ZEROTH FOURIER MOMENT vs. DISTANCE FOR
SEVERAL FREQUENCIES
- 1.0 MSEC INPUT PULSE 161
40 PHASE OF ZEROTH FOURIER MOMENT vs. DISTANCE FOR
SEVERAL FREQUENCIES
- 0.5 MSEC INPUT PULSE 162
41 PHASE OF ZEROTH FOURIER MOMENT vs. DISTANCE FOR
SEVERAL FREQUENCIES
- 1.0 MSEC INPUT PULSE 163
42 COMPARISON OF THE THEORETICALLY PREDICTED AND THE
MEASURED DAMPING COEFFICIENT a 166
43 COMPARISON OF THE THEORETICALLY PREDICTED AND THE
MEASURED PHASE SHIFT PER UNIT LENGTH 5 167
44 THE UFSA R1 CORE p DISPERSION LAW 169
45 COMPARISON OF THE THEORETICALLY PREDICTED AND THE
MEASURED (a2 £2) 171
xvi


CHAPTER V
NUCLEAR CALIBRATION OF THE UFSA SUBCRITICAL
Introduction
The nuclear calibration of the University of Florida SPERT Assem
bly was performed prior to the space-time kinetics studies. The cali
bration involved conventional inverse multiplication measurements,
absolute k^.,. determination by pulsing techniques and comparison with
multigroup-multiregion diffusion theory calculations. The Garelis-
Russell technique was employed to determine kg/l and this result used to
calculate ke^ by coupling it with the experimentally determined decay
constant and the theoretically calculated effective delayed neutron
fraction. During this phase of the experimentation, "spatial effects"
were noted in both a and kg/£. These effects and other interesting
kinetic phenomena involving these basic reactor parameters were con
sidered worthy of further study and were investigated during the main
part of the research. They are discussed in Part 2, Chapter V. In
this section the results pertinent to the necessary calibration of the
system are given.
Theoretical Notes
The Inverse Multiplication Method
Under ideal conditions, usually met only in small-fast assemblies,
the reactivity can be represented by
64


TABLE V
TIME STEPS USED FOR THE WIGLE CALCULATIONS
UFSA R1 Core
0.5 M/W Ratio
16.35 cm wide reflected core
Input Pulse Width = 0.5 msec
Time Step
(Microsec)
2
4
1
5
1
2
5
10
25
50
To Time
Step No.
40
140
160
170
205
235
315
415
655
795
Time Step
(Microsec)
To Time
Step No.
Input Pulse Width = 1.0 msec
3
5
1
.5
1
2
5
10
25
50
30
210
220
230
265
295
375
475
695
834
106


59
on the magnitude of the resolving time, it is not necessary to prop
erly identify the counting system since no significant difference in
the corrected counts are observed when using the paralyzing or non
paralyzing correction. This is shown in Fig. 14.
The resolution time correction for the UFSA data acquisition
system is then given by:
N.
N =
T N0xtr
1
CWxTrs
where
N^, = true number of events in a given channel
N = observed number of events in a given channel
tr = resolution time of the system
CW = channel width
Trs = number of sweeps of the analyzer
and, for a non-paralyzing system
tr = reciprocal of the maximum observable count rate
Using this method the following resolution or resolving time fan
tors were determined:
A) MDDAS including 1024 Multichannel Analyzer
tr = 0.31 + .015 microsec
B) MDDAS excluding Multichannel Analyzer
tr 0.20 + .01 microsec
C) NDDAS including 256 Multichannel Analyzer
tr = 0.56 + .02 microsec
D) NDDAS excluding Multichannel Analyzer
tr 0.20 + .01 microsec


FLUX (RELATIVE UNITS)
FIG- C24 TIME PROFILE OF THERMAL NEUTRON FLUX 53-8G CM FROM THE SOURCE
220


260
30. M. A. Perks, "FREAK A Fast Reactor Multigroup Kinetics Program,"
ANL-7050 (1965).
31. R. W. Garner, private communication (1968).
32. S. Kaplan, 0. J. Marlower and J. Bewick, "Application of
Synthesis Techniques to Problems Involving Time Dependance,"
Nuc. Sci. Eng., 18, 163 (1964).
33. J. B. Yasinsky and A. F. Henry, "Some Numerical Experiments
Concerning Space-Time Reactor Kinetics Behavior, Nuc. Sci. Eng.,
22, 171 (1965).
34. W. M. Stacey, Jr., "A Variational Multichannel Space-Time
Synthesis Method for Nonseparable Reactor Transients," Nuc. Sci.
Eng., 34, 45 (1968).
35. A.Foderaro and H. L. Garabedian, "Two Group Reactor Kinetics,"
Nuc. Sci. Eng., 14, 22 (1962).
36. K. 0. Ott, "Quasistatic Treatment of Spatial Phenomena in Reactor
Dynamics," Nuc. Sci. Eng., 26, 559 (1966).
37. K. 0. Ott and D. A. Meneley, "Accuracy of the Quasistatic Treat
ment of Spatial Reactor Kinetics," Proceedings of the Brookhaven
Conference on Industrial Needs and Academic Research in Reactor
Kinetics, BNL 50117 (T-497) (1968).
38. A. F. Henry, "The Application of Reactor Kinetics to the Analysis
of Experiments," Nuc. Sci. Eng., _3, 52 (1958).
39. C. D. Kylstra and R. E. Uhrig, "Spatially Dependent Transfer
Function of Nuclear Systems," Nuc. Sci. Eng., 22, 191 (1965).
40. S. R. Kavipurapu, "Energy and Spatially Dependent Impulse Response
and Transfer Function of Nuclear Systems, Unpublished Ph.D.
Dissertation, University of Florida (1964).
41. A. M. Weinberg and H.. C. Schweinler, "Theory of Oscillating
Absorber in a Chain Reactor," Physical Review, 74, 851 (1948).
42. R. B. Perez and R. E. Uhrig, "Propagation of Neutron Waves in
Moderating Media," NuC. Sci. Eng., 17, 90 (1963).
43. M. N. Moore, "The Determination of Reactor Dispersion Laws from
Modulated Neutron Experiments," Nuc. Sci. Eng., 21, 565 (1965).
44. R. L. Brehm, "Analysis of Neutron Wave Experiments," Proceedings
of the Symposium on Neutron Noise, Waves, and Pulse Propagation,
AEC Symposium Series No. 9 (1967).


31
The accelerator was used with a 4-5 curie tritium target.
The position of the target can be changed to keep the source
centered on the target-end of the assembly for any given moderator
level


FLUX (RELATIVE UNITS)
FIG- C17 TIME PROFILE OF THERMAL NEUTRON FLUX PCS-46 CM FROM THE SOURCE


82
of this work. The subcritical assembly consists of a square array of
4.8% enriched UC^ SPERT F-l fuel elements moderated by light water and
with essentially an infinite light water reflector on the sides. The
core is 243 cm long, 16.35 cm wide, has an effective height of 76 cm
with 30.5 cm wide side reflectors. The effective multiplication con
stant of the assembly has been determined to be 0.990+.003.
A fast, reliable set of electronic instrumentation was developed
and used to record the very high, time-varying count rate as a function
of position in the core of the assembly. A technique was developed to
subtract the contribution of epicadmium neutrons from the recorded
time profile of the neutron flux.
The study of the pulse propagation phenomena was conducted in a
two-fold manner: in the time and in the frequency domain. The study
in the time domain consisted of several sections, focusing the attention
on basic aspects of the space and time dependence of propagating dis
turbances. The following aspects were investigated:
a) The propagation of 0.5 and 1.0 msec wide input pulse intro
duced at one end of the assembly. This investigation constituted the
main part of the research; it involved a detailed comparison of the
experimental results with the predictions of the WIGLE calculational
scheme. The sensitivity of the one-dimensional model to small varia
tions in the transverse leakage was also investigated.
b) Static flux traverses were conducted to determine the steady-
state neutron distribution in the asymptotic region. Dynamic flux
traverses were performed to determine whether any propagation occurs in
the transverse direction.
c) The effect that the input pulse width has on the propagation


APPENDIX E
UFSA R1 CLEAN CORE
TIME PROFILES OF THERMAL NEUTRON FLUX
AT FOUR POSITIONS IN THE CORE
FOR A 0.1 MSEC INPUT PULSE


47
were efficiently handled with few losses the actual running time of a
particular experiment could be reduced to a few minutes. This possi
bility forced the development of the ultra-fast instrumentation chan
nels used in this work.
The electronic counting system^" used throughout the experimentation
used a transformer-coupled pulse amplifier to produce a fast logic sig
nal to the input of the MCA from a slower input signal originating in a
3
He counter. The system has superior counting and stability character
istics. The resolution time of the system is practically nil compared
to that of the MCA. In general, data acquisition times were reduced
to a few minutes; in that time it was possible to obtain, at most of
the detector locations, the following counting statistics:
ol7
Peak Channel Counts:
Counting Time:
No. of data channels:
Channel Width:
Pulse repetition rate:
Noise Level:
Background Level:
2-10 minutes
1024
20 micro sec
30 hz
40 event/min
140 event/min
The Neutron Detector
The neutron detectors used throughout the experimental program were
3
proportional counters filled with He The counters are of the Texlium
variety made by Texas Nuclear Corporation and were specially built to
conform to the UFSA core grid. Detectors of this type have been
1. The assistance of Mr. Joel B. Ayers of ORTEC, Inc.,who suggested
this electronic instrumentation is gratefully acknowledged.


86
methods presently in use are:
a) Direct solution of the multigroup, space-time diffusion
equations, referred to as the "exact" method. This method constitutes
a benchmark for the other techniques. The most prominent of these
calculational schemes are:
1) WIGLE, a two-group, one-dimensional space-time diffusion
theory computer program [24, 27, 28].
2) TWIGLE, a two-dimensional version of WIGLE [29].
3) FREAK, a fast reactor multigroup kinetics code [30].
4) A four-group, one-dimensional space-time diffusion theory
computer program [31] developed by the Phillips Petroleum Co.
b) Flux Synthesis methods, which used the idea that the flux
shapes which occur during a transient can be bracketed by a set of
shape functions [32, 33, 34].
c) Conventional Modal methods, exemplified by the Foderaro-
Garabedian technique [35] in which the flux is expanded in terms of
the eigenfunctions of the wave equation (Helmholtz modes).
d) The Quasistatic approach of Ott [36, 37], a more sophisticated
"factorizing" technique (the flux is factorized into an amplitude and
a shape function). Use is made of the fact that the time dependence
of the shape function is less important than the time dependence of
the amplitude function.
e) The Adiabatic approximation [38], a "factorizing" technique
in which, besides ignoring the time dependence of the shape function,
no distinction is made between the prompt and delayed neutron sources.
Other related studies, on a somewhat different context,were performed
by Kystra and Uhrig [39] and by Kavipurapu [40]. They studied the


SPACE-TIME REACTOR KINETICS STUDIES WITH THE
UNIVERSITY OF FLORIDA SPERT ASSEMBLY
By
NILS J. DIAZ
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1969


179
pulse.
Tabulated in Table XVIII are values of kg/5 determined using the
Garelis-Russell technique on pulses arising from widely differing input
pulses; the measurements were made at one position far from the neutron
source. The effect of the finite pulse width on kg/5, is significant.
An analysis of the results shown in Fig. 49 points to the fact that at
very large distances from the source a value of kg/5- close to the one
that would be obtained from a delta function input pulse is obtained.
Some comments on observed facts about the Garelis-RusseU tech
nique are pertinent. A dominant factor in the determination of kg/5, is
the delayed neutron background, which is strongly dependent on the
repetition rate of the neutron generator (which should be determined
very accurately). Meaningful values of kg/5, are obtained far from the
source for not too wide input pulses. A correction on the determined
kg/5, values to account for the finite width of the input pulses was
developed by G. A. Mortensen, Nuclear Safety Division, Phillips
Petroleum Co. [48] after the experimental observation of the phenomena.
The suggested method was applied to several of the cases under consid
eration but an overcorrection seems to result; the "corrected" values
keep changing drastically at long distances from the source where the
influence of the input pulse is negligible. Further work in this cor
rection should improve the results.
The Measured Reactivity and kef£ Values
The reactivity in dollars is obtained, following the Garelis-
Russell method, by the expression
P($) -
a
kg/5,
- 1


PART 1
THE UNIVERSITY OF FLORIDA SPERT ASSEMBLY
- DESIGN AND CALIBRATION -


90
^0(r, f) = A1(J (f) Q23 (r2, r3, B^)
exp {[-o^Cf, B) + i (f, b£)]z}
where:
A1q specifies the source condition.
2
Q23 (r2> r3> B ) specifies the transverse space dependence.
and are respectively the real and imaginary components of
the complex inverse relaxation length.
The Fourier transform (r, f) of the unit impulse response
(r, t) is the transfer function of the system. Therefore the ampli
tude and phase curves obtained from a numerical Fourier transformation
of the pulse propagation data corresponds to those measured directly
by the wave experiment.
It should be mentioned that, although most common nomenclature
refers to the pulse propagation phenomena and/or the neutron wave
propagation as "thermal" neutron propagation because of the "close-to-
thermal" source employed and the characteristics of the media studied,
there is no such "pure" phenomena in a multiplying medium. The dis
turbance propagates whether it started as a "fast" or "thermal" pulse
and quickly losses memory of its origin (specially in water moderated
media). The only difference between inserting a thermal or a fast
burst of neutrons lies in the spatial distribution.of source neutrons.
The thermal source will appear essentially localized at the point of
insertion while the fast source neutrons will penetrate deep into the
assembly.


FLUX (RELATIVE UNITS)
FIG- C20 TIME PROFILE OF THERMAL NEUTRON FLUX
3-0 CM FROM THE SOURCE


180
TABLE XVIII
THE DECAY CONSTANT AND kg/A VALUES
MEASURED AS A FUNCTION OF INPUT PULSE WIDTH
- 0.5 MSEC INPUT PULSE WIDTH -
UFSA R1 Clean Core
0.5 M/W Ration
16.35 cm wide reflected core
PULSE
WIDTH
(msec)
a
(sec
k8/£
(sec
0.1
511+10
213+4
0.5
488+11
202+5
1.0
501+6
193+5
2.0
503+8
183+7
3.0
51.0+4
172+9
4.0
523+16
169+9
5.0
494+6
144+10
6.0
504+5
136+10
7.0
505+5
126+9
8.0
522+17
118+11
9.0
524+13
111+8
10.0
496+16
108+9


This dissertation was prepared under the direction of the chairman
of the candidate's supervisory committee and has been approved by all
members of that committee. It was submitted to the Dean of the College
of Engineering and to the Graduate Council, and was approved as partial
fulfillment of the requirements for the degree of Doctor of Philosophy.
March 1969
Dean, Graduate School
Supervisory Committee:
Chairman


188
Determination of the Four Group Parameters
A four-group diffusion approach was used to calculate the effective
multiplication factor of the six UFSA configurations to be studied as
Phase I of the large core dynamics experimental program. The calcula
tion of the parameters was performed by the Nuclear Safety Research
Branch of the Phillips Petroleum Company [31].
The four energy groups were divided as follows:
Group 1 8.21 X 105 107 EV
Group 2 5.53 X 103 8.21 X 105 EV
Group 3 0.532 5.53 X 103 EV
Group 4 0 0.532 EV.
To obtain the four-group constants, the following sequence of cal
culations was performed for each of the six cases mentioned.
1. Fast group (groups 1-3) constants were calculated by the PHROG
[50] computer program. These calculations include resonance integral
and Dancoff correction calculations as provided by the RAVEN [51] cal-
culational scheme.
2. Thermal constants were calculated by the TOTEM [52] computer
program.
The constants used for the core object of this work are given
below. It should be mentioned that the procedure described was also
employed to determine the two-group parameters used for the WIGLE
calculations.


VALVE
FIG. 7 AIR SYSTEM SCHEMATIC


P'LX (RELATIVE UNITS)
FIG*. C12 TIME PROFILE OF THERMAL NEUTRON FLUX 145-BB CM FROM THE SOURCE


192
General
A set of computer programs was coded for the data processing and
the analysis of the results of the pulsed measurements.^ The programs
were written for the IBM 360-50 computer, with the exception of
MULTIPLOT which was coded for the IBM 1800 computer.
UNIPUL
The FORTRAN IV, IBM 360 computer program UNIPUL is used for a
unified processing of the data obtained from conventional pulsing or
pulse propagation measurements. A complete description of the program
and the input requirements are available in Ref. 54.
The main operations of UNIPUL are:
1. Resolution time correction of the data, using a non-paralyzing
counting correction.
2. Statistical determination of the neutron background. If the
system under study is multiplicative, the delayed neutron background
is calculated and stored in COMMON for the calculation of k8/JL.
3. Determination of the fundamental decay constant using Peierl's
statistical method [18]. The decay constants can be calculated at dif
ferent waiting times from the peak of the pulse to determine the time
convergence. The a's are stored in COMMON for the reactivity determina
tion.
4. The background is subtracted from the data. A test is made to
find if any channel has a negative number of counts. If a few channels
have negative counts, their value is set to zero. The number of
1. The cooperation of Dr. M. J. Ohanian in the coding of these
programs is gratefully acknowledged.


120
were conducted to investigate the propagation, throughout the core, of
0.5 and 1.0 msec input pulses. As mentioned in the previous section,
particular aspects of the propagation phenomena investigated are dealt
with separately.
The experimental technique used has already been described.
Briefly, the time profile of the normalized thermal neutron flux is
obtained at selected positions by the integral count method of normal
ization and the epicadmium subtraction technique. Nineteen detector
positions were used for these measurements. For each pulse width two
runs were made at each position, except at the far end of the core where
data acquisition times became significantly longer. Most of the ex
perimental errors assigned to the data were obtained by analyzing the
results of these two measurements.
Following the nomenclature developed by Doshi and Miley [20] at
the University of Illinois, the delay times t^, the dynamic inverse
relaxation length k^, the asymptotic velocity of propagation v the
full-width at half-maximum (FWHM) of the propagating pulse and the
pulse shapes will be analyzed to describe the propagation character
istics of a neutron burst in the very close to critical assembly. The
experimental results are presented together with the calculated values.
It should be noted that practically all the figures showing the
pulse shapes were plotted using a Calcomp plotter which is part of the
IBM 1800 Computer facility in the Nuclear Engineering Sciences Depart
ment at the University of Florida. The data is plotted in a continuous
manner and the points are superimposed later for convenience.' Typi
cally the experimental pulse shapes were plotted from 333 time steps;
the corresponding calculational results have 231-245 time steps.


FLUX (RELATIVE UNITS)
FIG C31 TIME PROFILE OF THERMAL NEUTRON FLUX 142-09 CM FROM THE SOURCE
227


138
TABLE XI
CHANGES IN THE NUCLEAR PARAMETERS
DUE TO CHANGES IN THE TRANSVERSE BUCKLING
UFSA R1 Core
0.5 M/W Ratio
16.35 cm wide reflected core
WIGLE X-Section
(cm )
Reactor
Height (cm)
Estimated
k ,,
eff
Vertical
Buckling
(cm-2)
S
70.0
.986
20.14
E-04
.054671
.12223
72.5
.988
18.78
E-04
.054526
.122207
76
.990
17.08
E-04
.054345
.12217
77.5
.992
16.43
E-04
.054277
.122158
80.0
.995
15.42
E-04
.054169
.122137
85.0
.998
13.66
E-04
.053981
.122096


40
total power of the assembly.
d. The fuel rod subjected to the specified highest specific power
is broken open and the fuel is completely fragmented, releasing 100% of
the gaseous fission products.
e. The iodine (in elementary form) diffuses uniformly throughout
the 18 x 32 x 15 room to calculate the on-site internal dose. The
off-site dose was calculated using the very conservative Pasqual's
principle of atmospheric diffusion.
The fission products inventory was calculated by means of the RSAC
code [9]. The results indicated that the iodine isotopes were the only
significant contributor to the inhalation dose. Using the active worker
breathing rate of 3.47 x 10 in /sec specified in 10 CFR 20, a person
remaining in the room would accumulate a thyroid dose of 0.023 mr for
each minute he remains in the room, after the hottest rod breaks open.
The total dose to a person that inhales all of the iodine contained in
the rod would be 0.2 r. The established radiation safeguards at the
University of Florida require the personnel to abandon the area imme
diately and notify the Radiation Safety Officer. The maximum time
required to evacuate the assembly and the fuel storage area is 10 sec
with 30 sec needed to evacuate the entire building; these times have
been measured during practice evacuation of the building.
The off-site total inhalation dose, calculated assuming the iodine
is released in one puff with zero wind velocity, cloud inverted con
dition, were typically:


210
200
190
180
170
160
150
O 0.5 msec input pulse width
O 1.0 msec input pulse width x)
'O
c/
/
/
/
X '
p
p"
if
/
/
9'
M M O* C?.. ' 1
__ .Q o
o o
-Q-
UFSA Ri Clean Core
12.72 cm between data points
? i ill I l
6 20 34 48
I I i 1 t ! I I I
62 76 90 104 118 132
Position Number
49 vs. AXIAL POSITION
178


53
components of the detector signal will actuate a wide band transistor
amplifier and tunnel diode discriminator from the secondary of the
transformer. A line driver buffer is also provided. The power and
control bias are provided by the ORTEC 403 Time Pickoff Control.
3. Modified ORTEC 403 Time Pickoff-Control
The 403 Time-Pickoff Control provides control and fan-out buf
fering for time derivation units such as the 260 Time Pickoff Unit.
The fan-out buffer Accepts the fast negative logic signal from the 260
unit and provides either fast negative or slower positive logic output
signals.
The 403 Time-Pickoff Control had to be modified to make its out
put signal compatible with the input requirements of the 212 Pulsed
Neutron Logic Unit used with the MCA. The 212 plug-in unit requires
pulses with rise times longer than 50 nsec while the positive output
from the 403 control unit has a rise time ~ 10 nsec. Furthermore, the
positive logic signal from the 403 control has a 0.5 microsec pulse
width which is too wide for the time resolution required. Therefore
2
two modifications were made: a capacitor (C12) was changed from 270
pf to 68 pf reducing the pulse width to 0.12 microsec and an inductor
coil (LI) was removed from the system changing the rise time to 50
nsec.
The modifications made the system compatible with the input
requirements of the MCA and avoided overdriving the analyzer.
2. Refer to ORTEC, Inc. "Instruction Manual for 403 Time-Pickoff
Control."


FLUX (RELATIVE UNITS)
FIG. 23A PULSE SHAPES PREDICTED BY WIGLE AT DIFFERENT POSITIONS IN THE UFSA R1 CORE
110


Decay Constant (sec
FIG. 48 DECAY CONSTANT VS. AXIAL POSITION
177


84
Asymptotic Velocity of Propagation
The asymptotic velocity of propagation, v is defined as the
inverse of the slope of the delay time vs. distance curve in the
asymptotic region.
The Asymptotic Dynamic Inverse Relaxation Length
The asymptotic dynamic inverse relaxation length, k^, is defined
as the inverse of the distance required for the amplitude of the pulses
to attenuate by a factor e.
The Damping Coefficient of the Neutron Wave
The damping coefficient of the neutron wave, a, is a measure of
the exponential attenuation of the amplitude of the wave for a given
frequency. The damping coefficient is determined from the amplitudes
of the zeroth Fourier moment by numerical transformation for each
frequency of interest.
The Phase-Shift per Unit Length
The phase-shift per unit length of path, £, is determined from
the phase angles of the zeroth Fourier moment by numerical transforma
tion for each frequency of interest.
The damping coefficient and the phase shift per unit length of
path are the real and the imaginary components of the complex inverse
relaxation length, p, respectively.
It should be noted that in this work pulse and wave propagation
from the "conventional" viewpoint, i.e., propagation of a residual
disturbance rather than the propagation of a pulse or wave front, is
being studied.


FLUX (RELATIVE UNITE)
TIME (MSEC)
TIME FRCFILE CF THERMAL NEUTRON FLUX
FI CL- C6
B6-5S CM FROM THE SDJRCE
mu)


REGION NO.
1
2
3
A
5
6
MATERIAL NO.
2
1
1
1
1
1
NO. OF MESH POINTS
12
9
A
7
6A
1
MESH SIZE (cm)
.9916
1.0
1.179
1.8166
3.17905
2.777
MESH POINT NO.
DETECTOR POSITION
13
25
32
36
AO
AA
A8
52
56
60
6A
68
72
76
80
8A
88
92
96
NO.
6
13
20
27
3A
A1
A8
55
62
69
76
83
90
97
10 A
111
118
125
132
FIG. 21 ONE-DIMENSIONAL ARRANGEMENT OF THE UFSA CORE USED IN THE WIGLE CALCULATIONAL
SCHEME


FIG. 9 POWER vs. TIME FOR THE DESIGN BASIS ACCIDENT


Amplitude (Relative Units)
FIG. 39 AMPLITUDE OF ZEROTH FOURIER MOMENT vs. DISTANCE FOR
SEVERAL FREQUENCIES
- 1.0 MSEC INPUT PULSE -


FLUX (RELATIVE UNITS)
TIME (MSEC)
FIG* El
TIME PROFILE OF THERMAL NEUTRON FLUX
53*BS CM FROM THE SOURCE


APPENDIX B
DESCRIPTION OF THE COMPUTER PROGRAMS
UNIPUL
MORE
MORWIG
ALXILS
MULTIPLOT


Flux (Relative Units)
131
FIG. 31 AMPLITUDE ATTENUATION OF THE THERMAL NEUTRON FLUX
- 0.5 MSEC INPUT PULSE -


42
d. Violation by the facility operator of the administrative
procedures requiring a visual check of the assembly before start-up and
continuous attention to the control console instrumentation to determine
the status of the assembly at all times.
The consequences of such an accident were determined by assuming
the following:
1. The assembly is initially at a steady state power level of 10
watts (normal average power is .065 watt).
2. Water flows into the system at the maximum rate of 12 gpm.
3. The reactivity addition rate is 0.05 $/sec. This rate is
larger than the calculated rate (.04 $/sec) for the case discussed
above and more than that calculated to occur near critical for properly
loaded assembly.
4. The power scram is set, through calibration or other error, at
10 kw. Corresponding error settings occur for the neutron flux and
period scram.
5. A scram occurs 1 sec after a power level of 10 kw is reached.
This time has been determined as the elapsed time from the initiation
of a scram signal to the opening of the dump valves.
6. No feedback effects are considered. This is a good approxima
tion to our case due to the low power levels involved and again is a
conservative assumption.
The calculations were made using the IREKIN code, described in
reference [10]. IREKIN numerically solves the point model kinetics
equations.
Starting with the assembly one dollar subcritical (ke^ = .993),
and proceeding with the described excursion, the accident would yield


kg/£ (sec
FIG. 17B kg/i, AND k vs. MODERATOR LEVEL


173
looo
>
FIG. 47- THE UFSA R1 CORE p2 DISPERSION LAW


87
space dependent reactor transfer function, within the framework of a
combination of time-dependent Fermi-Age and diffusion thories.
A parallel field of study has been the neutron wave technique
[41, 42, 43]. A major improvement in the neutron wave field was
achieved when Moore proved [25] that the pulsed technique is equivalent
to the neutron wave method when the pulsed data is analyzed in the
Fourier transform plane. The work by Booth [26] confirmed this simpli
fying and time-saving approach to the powerful but time-consuming neu
tron wave measurements.
The study of multiplying media by neutron wave (or pulse) propaga
tion has not been developed as extensively as in non-multiplying media.
Brehm made a very elegant analysis of the problem, showing the excita
tion of slowing down modes whose relaxation lengths he was able to
compute [44]. Dunlap and Perez studied the dispersion law of a heavy
water-moderated, natural uranium assembly [45].
To date, no complete analysis of a pulse propagation experiment,
i.e., in the time and in the frequency domain is available in the
literature.
The WIGLE Calculational Scheme
The version of the WIGLE code used for the calculations performed
in this work is an IBM 360 version of the WIGLE-40 program described
by Radd [27]. The following description is extracted from Ref. 27.
WIGLE is a one-dimensional, two-energy-group, time-dependent dif
fusion program. It calculates the space-time behavior of the neutron
flux in a reactor during a transient. The calculation is restricted
to one-dimensional slab geometry with zero gradient or zero flux


FLUX (RELATIVE UNITS)
1-0
FIG* FI PROPAGATION OF A PULSE WIDER THAN THE SYSTEM PROPAGATION TIME
- EXPERIMENTAL -
254


FLUX (RELATIVE UNITS)
FIG' E2
TIME PRDFILE OF THERMAL NEUTRON FLUX
73-30 CM FROM THE SOURCE


LIST OF TABLES
TABLE Page
Ik vs. MODERATOR LEVEL OF UFSA REFLECTED
eff
CORES 7
II UFSA INSTRUMENTATION AND CONTROL 25
III SUMMARY OF 1/M AND PULSED MEASUREMENTS 74
IVINPUT PARAMETERS FOR THE WIGLE CALCULATIONAL
SCHEME 104
V TIME STEPS USED FOR THE WIGLE CALCULATIONS 106
VIDELAY TIMES MEASURED ACROSS THE WIDTH OF
THE CORE
- 0.5 MSEC INPUT PULSE 119
VIICLEAN CORE PULSE PROPAGATION STUDIES EXPERIMENTAL
AND THEORETICAL RESULTS
- 0.5 MSEC INPUT PULSE WIDTH 127
VIIICLEAN CORE PULSE PROPAGATION STUDIES EXPERIMENTAL
AND THEORETICAL RESULTS
- 1.0 MSEC INPUT PULSE WIDTH 128
IX ASYMPTOTIC PROPAGATION VELOCITY v 133
P
X DYNAMIC INVERSE RELAXATION LENGTH k, 134
a
XICHANGES IN THE NUCLEAR PARAMETERS DUE TO
CHANGES IN THE TRANSVERSE BUCKLING 138
XIITHE CALCULATED ASYMPTOTIC VELOCITY OF
PROPAGATION AND DYNAMIC INVERSE RELAXATION
LENGTH vs. CORE HEIGHT
- 0.5 MSEC INPUT PULSE WIDTH 146
XIII DELAY TIMES AND FWHM FOR A NARROW INPUT PULSE 148
XIV DELAY TIMES AND FWHM FOR A WIDE INPUT PULSE 149
XV PULSE SHAPES vs. INPUT PULSE WIDTH 151
xi


136
in which the largest number of counts was recorded and in many in
stances is not the best value for the location of the peak. The values
quoted in Tables VII and VIII represent refined values obtained by
inspection of the experimental data.
c) The Input Pulse Width
d) The Experimental and Theoretical FWHM
e) The Location with Respect to the Source is given in the
figure title.
A comparison between the theoretical and experimental pulse shapes
reveals the following:
a) Shapes agree very closely in the spatial region dominated by
the neutron source.. Small discrepancies start to show up near P41.
b) Past P41 the theoretical pulses become consistently narrower
than the experimental one. As discussed above, the other main charac
teristics describing the propagation of the pulses are well matched.
Discussion of the Clean Core Results
The fact that good agreement was found between the theoretical and
experimental results deserves some elabotation since the agreement may
seem fortuitous.
The calculational scheme used to obtain the WIGLE input parameters
is well established. The parameters are therefore considered to be as
reliable as the present state of the art permits. The main uncertainty
as far as the calculational model is concerned is in the allocation of
a transverse buckling.
To investigate the sensitivity of the two-group, one-dimensional
model to changes in the transverse buckling (incorporated into the
WIGLE scheme as a change in the absorption cross-section) a series of


FLUX (RELATIVE UNITS)
FIG- D12 TIME PRDFILE OF FAST NEUTRON FLUX 101-03 CM FROM THE SOURCE


FLUX (RELATIVE UNITS)
FIG* C21 TIME PROFILE GF THERMAL NEUTRON FLUX
15-71 CM FROM THE SCURCE


PART 2
SPACE-TIME REACTOR KINETICS STUDIES WITH
THE UNIVERSITY OF FLORIDA SPERT ASSEMBLY


19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
Exp
O

s=
20
FIG. 45
Frequency
COMPARISON OF THE THEORETICALLY PREDICTED AND THE MEASURED (cx2-£2)
171


10
Mechanical Design
The entire assembly can be divided into three components: the
supporting platform and dump tank; the basic core tank and fuel rod
support structure; and the combination core side walls and side
reflector tanks.
The supporting platform is composed of 5 inch steel I-beams, raised
5 feet from the floor level by six steel columns 31/2 inches in diam
eter. The column footings rest within a 6x8x2 foot steel rank which
serves as a reservoir for the continuous water flow system and as a dump
tank. Under normal operating conditions, this represents a minimum
distance of about 4 feet between the bottom of the assembly and the
water surface in the reservoir. This distance is sufficiently large so
that the bottom of the assembly is considered to be unreflected under
all conditions.
All the core and reflector hardware is made of type 5456-H321
aluminum. The bottom of the basic tank is made of a 24x96x3/4 inch plate
bolted to the steel I-bean platform. The underside of the plate is
covered with a .030 inch thick Cadmium sheet. The lower fuel rod sup
port assembly rests on the plate. The end walls of the basic tank are
made of a 24 x 39 3/4 inch plate and are supported by two 2 1/2 x 2 1/2
x 1/4 inch steel angle braces welded to the I-beam platform. Two four-
inch aluminum channels span the eight foot dimension of the tank, con
fining the upper fuel rod support system and detector mounts.
The side walls of the basic tank serve also as reflector tanks when
the reflected cores are under study. These tanks have dimensions of
12 x 96 x 37 inches. The arrangement allows one to vary the width of the
core with a sole structural support.


FLUX (RELATIVE UNITS)
FIG* C13 TIME PROFILE DF THERMAL NEUTRDN FLUX 155-53 CM FROM THE SOURCE


92
by the pump or the low operating average power level of the system
( 0.13 watt). The large volume of circulating water aided in main
taining the temperature constant. In one 14-hour period of operation
the temperature change was about 1C.
The Epicadmium Subtraction Method
The use of a thermalizing tank and the cadmium difference method
is widespread for pulse propagation measurements. In the particular
case of the UFSA assembly this technique is not pertinent, since the
primary reason for the measurement was to test the analytical model
under stringent, viz. non-asymptotic, conditions; furthermore, with
the WIGLE model it is possible to account for the spatial and energy
dependence of the fast source rather well within a two-energy group,
slab geometry scheme. Therefore, no thermalizing tank was used and no
cadmium plate was inserted between the source and the assembly.
3
The He detectors used for the measurements respond to epicadmium
neutrons, although with decreasing efficiency of detection as the
neutron energy increases. Therefore, to obtain a true basis of compar
ison between the model and experiment, the contribution of the
epicadmium neutrons was subtracted from the time profile recorded by
the detector.
The thermal and epicadmium neutron time profiles measured in the
assembly do not differ significantly in shape in the asymptotic region
but show significant differences in the time scale, the relative atten
uation and the shapes near the source. Although the detectors used
are about ten times more efficient for detecting thermal neutrons, the
epicadmium flux is much larger than the thermal flux throughout the
system and, in particular, close to the source. Therefore, the


GRID POSITION
WEIRS
FIG. 15 DETECTOR POSITIONING SCHEME


70
locations employed for these measurements and for the absolute
determination by pulsing. Two 1 cu, centrally located Pu-Be sources
were used for these measurements.
The measurements were performed in the following sequence:
a) Unmultiplied counts vs. water level were obtained from a water
level of 20 cm to 91.4 cm above the bottom of the active fuel in steps
of 5 cm.
b) The fuel was loaded in the assembly in the presence of two
centrally located Pu-Be sources, with proper monitoring.
c) Multiplied counts vs. water level were obtained from a water
level of 20 cm in 5 cm steps until an effective multiplication factor
<_.99 was reached. This procedure follows the criteria established for
Initial Loading of the assembly, Part 2, Chapter III of this work; the
5 cm steps were more conservative increments than those specified for
the Initial Loading of the assembly.
The results of these measurements are shown in Figs. 16 (A, B).
All the curves were well behaved in the sense that none "nose-dived."
Position 2 and 4 failed to properly describe the multiplication of the
assemblies because of their nearness to the source. Position 1 seems
to overestimate the multiplication, mainly because the unmultiplied
count rate was extremely low at this position thus an apparently high
ratio of multiplied/unmultiplied counts was obtained. A detector more
sensitive to high energy neutrons was used in locations 1 and 6.
It should be noted that when the inverse multiplication is plotted
2
vs. 1/H the predictions become quite linear much earlier than when
2
plotted versus H. Better predictions are therefore made with the 1/H
curves but the approach can become less conservative by underestimating


151
TABLE XV
PULSE SHAPES VS. INPUT PULSE WIDTH
UFSA Rl Core
0.5 H/W Ratio
16.35 cm wide reflected
core
INPUT PULSE
PEAK AT
FWHM
WIDTH (msec)
(msec)
(msec!
.1
2.53
3.75
.5
2.65
3.69
1.0
3.04
3.69
2.0
3.79
4.20
3.0
4.39
4.68
4.0
4.72
5.28
5.0
6.02
5.95
6.0
6.97
6.70
7.0
7.67
7.40
8.0
8.62
8.30
9.0
9.67
9.20
10.0
10.22
10.5


Amplitude (Relative Units)
160
FIG. 38 AMPLITUDE OF ZEROTH FOURIER MOMENT vs. DISTANCE FOR
SEVERAL FREQUENCIES
- 0.5 MSEC INPUT PULSE -


30
moderated reflected slab of 10.44 cm is obtained consistent with the
10.4 cm value.
Thus the slab width of 9.46 cm proposed by us compares favorably
from the safety viewpoint with the safe width for an infinite, water
moderated and reflected slab and is considerably narrower than the safe
width for an infinite, water-moderated and unreflected slab. Within the
present context it should also be pointed out that as indicated on page
54 of the Design and Hazards Report [3], no flooding of the storage area
seems possible from natural causes.
Neutron Sources
Two types of neutron sources were used throughout this work.
1) Two Pu-Be sources mounted in an aluminum cylinder which can be
driven remotely through a plastic pipe from a shielded box located in
one corner of the facility room to underneath the center of the core.
Neon lights provide indication at the console of the position of the
sources. These sources, which have a combined yield of 3.2 x 10^
n/sec are used for start-ups and for the inverse multiplication meas
urements .
2) A Texas Nuclear Neutron Generator which is used in continuous
mode for static measurements and in the pulsing mode for the pulse
propagation measurements. The generator is of the Cockcroft-Walton
type, TNC Model 150-1H with continuously variable high voltage from
0-150 kv and has been modified to obtain larger currents by removing
the einzel lenses and installing a new 22 electrode accelerator tube and
gap lense. Pre- and post-acceleration beam deflection produces sharp,
low-residual pulses.


43
a peak power of 69 kw, the total energy release is 0.5 Mwsec. The
power vs. time behavior of the assembly for the postulated power excur
sion is shown in Fig. 9.
The combined neutron and gamma dose to the operator is 1.5 rem
*
assuming that: the energy release is instantaneous, all neutrons have
an energy of 1 MeV and there is no attenuation in the assembly. The
proper RBE factors were taken into consideration.
It is concluded that, even if such an improbable accident would
occur, the hazards to personnel and the general population are not
significant.


88
boundary conditions. Up to six delayed neutron groups can be con
sidered in the calculations. The feedback subroutine may be used to
introduce arbitrary changes in the reactor parameters, either for
inherent feedback or for other time-dependent variations.
The basic equations solved by the WIGLE program are as follows:
D1 v*l Z1 h + X1 (vi:f *1 + vZf2 4>2) (1 Y1B)
1 1 3*1
+ E i + si = ^-ir
1=1 1
(1)
VD2 V2 E2<*>2 + x2 ^vEf1 ^1 + vZf2 ^ y2B^
1 i 92
+ S + Cl 6) C2iXi + S, = (2)
3C
Xi Cli + Y1 Bi vEfl *1
li
3t
3C
Xi C21 + Y2 Bi vEf2 ^2 = "3t
2i
(3a)
(i = 1, 2, ..., I)
(3b)
The number (I) of delayed neutron groups may be 0, 1, or 6. 6 can
be 1 or 0. When 5=1 and = y2> these equations are the conventional,
two-group, time-dependent diffusion equations. .
The time-dependent equations, and those equations necessary to
represent the inherent feedback effects if they are considered, are
reduced to finite (time) difference equations and numerically solved.
WIGLE can handle up to 60 regions, 251 mesh, points and 999 time steps.


145
evaluation of the nuclear parameters.
The asymptotic propagation velocity and dynamic inverse relaxation
length as a.function of core height ar shown in Table XII. As ex
pected, the velocity of propagation and the attenuation decrease as
criticality is approached.
The above analysis puts the good agreement between experiment and
theory on a firmer footing. When the observed phenomenon is very sen
sitive to small variations in system characteristics the probability
that fortuitous agreement may be obtained between theory and experi
ment diminishes with the degree of sensitivity.
Comments on the Fast Group Results
The following comments are pertinent regarding the fast group
results as calculated by the two-group WIGLE model:
1. The curve of delay times vs. distance remains flatter than
the corresponding curve for the thermal group near the source region;
it then bends rapidly to yield the same asymptotic velocity of propaga
tion as that of the thermal group.
2. The fast group flux peaks between positions P6 and P13; this
has been confirmed experimentally. In contrast the thermal group flux
peaks at P6. The asymptotic inverse relaxation length of the fast
group is, as expected, identical to the one determined for the thermal
group.
3. The fast group pulse profiles are slightly wider than those
for the thermal group.
The time profiles from the cadmium-covered detector measurements,
which were used to determine the thermal group pulse profiles by the
epicadmium subtraction method were compared with the calculated fast


155
distance of 5 cm from the outer wall. At the far end of the assembly
the reflected pulse "propagated" into the assembly, attenuating and
dispersing with distance to the end wall. This pulse quickly dissi
pated into the.assembly with no significant effect on the measurements


APPENDIX D
UFSA R1 CLEAN CORE
TIME PROFILES OF FAST NEUTRON FLUX
AT SEVERAL CORE POSITIONS
FOR INPUT PULSES OF 0.5 AND 1.0 MSEC


190
N0G ( Ef)i<¡>i(r)
s(r) = z L-
i=l X
where X is the eigenvalue. The code then sets and solves a set of
finite difference equations.


Water Level Above Weir Apex (cm)
Flow Rate (gpm)
FIG. 5 FLOW RATE vs. HEIGHT OF WATER LEVEL ABOVE WEIR APEX


APPENDIX A
CALCULATIONAL PROCEDURES USED IN THE
DETERMINATION OF THE NUCLEAR PARAMETERS
AND THE k VALUES
eff


FLUX (RELATIVE UNITS)
FIG* C36 TIME PROFILE OF THERMAL NEUTRON FLUX E0G-4S CM FROM THE SOURCE
232


FLUX (RELATIVE UNITS)
1-0
0*9
0-E3
0-7
0-G
0 *5
0-4
0-3
ft
I
0-
0-1
T
0
3
3 4
=F
5
RIFS 30-1
PEAK AT 0-430 MSECS
pulse: width = o-s msecs
THEORY FWHM = 0-747 MSEC
EXP- FWHM = 0-750 MSEC
to
OJ
0
9
f
10
TIME (MSEIO
FIG- D2
TIME PROFILE OF FAST NEUTRON FLUX 38-43 CM FROM THE SOURCE
UVJ1


FLUX (RELATIVE UNITS)
FIG* C37 TIME PROFILE OF THERMAL NEUTRON FLUX 219*17 CM FROM THE SOURCE
233


54
4. Technical Measurement Corp. 1024 Multichannel Analyzer
(MCA) and Data Handling Units
The time analysis of the pulse and the corresponding data output
are obtained from the following coupled instrumentation:
Pulsed Neutron Logic Unit, TMC Model 212
Digital Computer Unit, TMC Model CN1024
Data Output Unit, TMC Model 220C
Digital Recorder, Hewlett-Packard Model 561B
Binary Tape Perforator, Tally Model 420
A trigger from an external pulse generator starts the sweep of the
analyzer as controlled by the pulsed neutron logic unit and initiates
the burst at the accelerator. The 212 has variable channel widths from
10 to 2560 microsec. The full 1024 channels were used in all the meas
urements. The storage capacity of the CN1024 is 2^ counts. Both
printed paper tape and perforated binary tape were obtained as output.
5. ORTEC 430 Scaler
For some of the flux traverses and the 1/M measurements, the
integral counts were recorded in a 10 Mhz scaler.
The normalizing detector data acquisition system (NDDAS) consists
of the same components as the movable detection system except that a
256 multichannel analyzer, TMC Model CN110 was used for data handling
and the integral counts were accumulated in an ORTEC 429 Scaler, modi
fied for 7 Mhz counting rate.
The Resolution Time Correction
The resolution time of the pulsed neutron data acquisition system
was the parameter under consideration while searching for a well-matched


FLUX (RELATIVE UNITS)
FIG* C19 TIME PROFILE OF THERMAL NEUTRON FLUX 031-QQ CM FROM THE SOURCE


104
TABLE IV
INPUT PARAMETERS FOR THE WIGLE
CALCULATIONAL SCHEME
UFSA R1 Core
0.5 M/W Ratio
16.35 cm wide reflected core
CORE
Region Regions
1
2,3,4,5,6
ALUMINUM
WATER
D^(cm) 1
.5211
1.06379
1.9338
1.1370
EaiCcnf1)
.055588
.054350
.000228
.000589
Ir^(cm
.02116
.025224
.000159
.0483
vEfl(cm-1)
.003247
.007337
.00
.00
D2(cm-^)
.2895
.20809
3.6078
.14938
Ia2(cm-1)
.4234
.12217
.01048
.019242
vlfzicm-^)
.10027
.22662
.00
.00
[l/v^3(cm ^sec)
5647E-7
3324E-7
[l/v2](cm "'sec)
.3324E-7
.3324E-5
Note: Thermal group
: 0-0.53
ev

Fast group:
0.53 ev
- 10.0 MeV
V


78
Garelis-RusseHtechnique to a reflected slab assembly when proper care
is taken. Agreement between calculated and experimentally determined
values of the reactivity is termed excellent but since only one case
has been studied judgement on the overall applicability of the technique
to this type of reactor configurations should be reserved until the
other assemblies to be studied in the UFSA facility are duty analyzed.
It should be pointed out that multiple detector positions are necessary
to establish when an asymptotic decay constant is obtained, and that the
value of k£/Jl is affected by the input pulse width. As mentioned previ
ously, a more detailed analysis of the pulsed neutron reactivity meas
urements is conducted later in this thesis.


5 (rad/cm)
169
FIG. 44 THE UFSA R1 CORE p DISPERSION LAW


CHAPTER VII
CONCLUSIONS
The space-time kinetics behavior of a large-in-one-space dimension,
side-reflected, highly multiplicative (k^^ ~ 0.99) subcritical assembly
has been studied by the use of the pulse propagation technique. The
experimental results were used to test the predictions of the one
dimensional two-group, space-time diffusion theory calculational scheme
(WIGLE). The analysis of both the theoretically predicted and experi
mentally measured results was performed in the time as well as the fre
quency domain.
The comparison of theory and experiment in the time domain shows
good agreement. WIGLE accurately predicted the delay times and the
attenuation of the peak of the pulse when a first flight kernel was used
to describe the spatial distribution of the fast source. At large dis
tances from the source the time profiles predicted by WIGLE are con
sistently narrower than the measured pulse shapes. The sensitivity of
the one-dimensional model to small changes in the transverse buckling
was studied; WIGLE (and experiment) showed a large sensitivity to these
changes. These results point to the necessity of obtaining the best pos
sible estimate of the nuclear parameters to be used in space-time ki
netics calculations.
The comparison of theory and experiment in the sensitive frequency
domain confirms the good agreement found in the time domain. The
185


CHAPTER III
OPERATIONAL SAFETY
Introduction
The University of Florida SPERT Assembly, due to its large size,
enriched uranium-oxide fuel and nuclear potentialities required a
thorough study of its capabilities, operational characteristics, initial
loading procedures and of the behavior of the assembly under accident
conditions. The study was part of the requirements established by the
Division of Material Licensing of the USAEC prior to the granting of an
operating license.
Legally, a subcritical assembly has to comply with regulations
under 10 CFR Part 70 "Licensing of Special Nuclear Materials" since no
self-sustaining nuclear reaction is envisioned. In the case of the
UFSA, however, the Commission felt that certain technical sections of
10 CFR Part 50, which deals with nuclear reactor licensing, should apply
and serve as a guide for the design and the safety analysis.
The basic philosophies employed in the design of the system were:
a) The UFSA facility has been designed to remain subcritical under
normal operating conditions.
b) The safety instrumentation (see Part 1, Chapter I) has been
designed such that a single failure will not invalidate both the manual
and automatic scram and will not cause subsequent failures.
c) The design basis accidents were postulated on a single failure
32


FLUX (RELATIVE UNITS)
TIME (MSEC)
TIME PROFILE OF FAST NEUTRON FLUX 53-SB CM FROM THE SOURCE
FIG* D9
(JILO


40
30
20
10
58
Conventional
24 32
Channel Number
IG. 13 TIME PROFILES OF NEUTRON BURST RECORDED BY CONVENTIONAL
ELECTRONIC INSTRUMENTATION AND BY THE TIME-PICKOFF SYSTEM


FLUX (RELATIVE UNITE)
FIG. 273 EXPERIMENTAL PULSE SHAPES AT DIFFERENT POSITIONS IN THE TJFSA R1 CORE
123


FLUX (RELATIVE UNITS)
FIG* D8
TIME PROFILE OF FAST NEUTRON FLUX 41*15 EM FROM THE SOURCE
243


22
Instrumentation and Interlock System
The instrumentation and interlock system of the UFSA has been
discussed extensively in the reports submitted to the Atomic Energy
Commission [3, 4, 5] in conjunction with the license application. More
recently, Mr. L. B. Myers submitted a detail technical report on the
subject [6]. A brief descriptive explanation is given below.
A block diagram of the safety system logic flow in use at the UFSA
subcritical assembly for routine monitoring is shown in Fig. 8. There
are five principal channels of instrumentation:
3
a. Start-up channel using a He proportional counter, scaler, and
rate meter. The counter is located at the bottom of the core, close to
the geometrical center of the assembly.
b. Log power and period instrument No. 1 channel using a compen
sated ion chamber (operated uncompensated) as a signal to a Log N
amplifier. The chamber is located along the longitudinal axis of the
assembly, on the bottom of the core some two feet from the neutron
generator end.
c. Period instrument No. 2 channel using a compensated ion chamber
(operated uncompensated) as a signal to a log N amplifier. The chamber
is located along the longitudinal axis of the assembly, on the bottom
of the core, some six feet from the neutron generator end.
d. Linear neutron flux No. 1 channel using an uncompensated ion
chamber as a signal to a micromicroammeter. The chamber is mounted on
the top core support frame, close to the geometrical center of the core.
e. Linear neutron flux No. 2 channel using a compensated ion
chamber (operated uncompensated) to feed a signal to a micromicroammeter.
The chamber is mounted on the top core support frame on the opposite


CHAPTER II
THEORETICAL NOTES
Introduction
The main concern of this work has been the experimental investiga
tion of space-time kinetics effects on a large core subjected to a
perturbation; the basic aspects of the neutron physics phenomena were
studied. The analysis of the experiment has been conducted in both
the time and the frequency domain. The theoretical model used for
comparison with experiment, in the time and the frequency domain, was
the two-group, space-time dependent diffusion equations numerically
solved by the WIGLE code [24]. A one-to-one comparison of theory and
experiment was performed in the time domain. For the corresponding
study in the frequency domain both the WIGLE and experimental time
profiles as a function of position were analyzed into its wave compo
nents by a numerical Fourier transformation; this method was suggested
by Moore [25] and confirmed by Booth [26]. In this work only the
zeroth Fourier moment is analyzed; the amplitude and phase angle of
the zeroth Fourier moment correspond to those measured by the conven
tional neutron wave experiment.
Review of the Literature
A considerable amount of theoretical effort has been devoted to
the space-time kinetics problem in nuclear reactors. The more important
85


FLUX (RELATIVE UNITS)
FIG- C18 TIME PROFILE OF THERMAL NEUTRON FLUX 219-17 CM FROM THE SOURCE
214


FLUX (RELATIVE UNITS)
FIG- C38 TIME PROFILE OF THERMAL NEUTRON FLUX E31-8B CM FROM THE SOURCE
234


61
The Normalization Technique
The analysis of a pulse propagation experiment will yield informa
tion on the velocity of propagation, the attenuation and pulse shape as
a function of position of a propagating disturbance. To determine the
attenuation or relative amplitude of the pulse at different positions
in an assembly, the data must be normalized to a fixed reference so
that variations in the source strength, data acquisition time, etc.,
can be properly accounted for. The usual technique requires accumulating
integral counts in a scaler for each measurement and reducing all meas
urements to the fixed reference afterwards. As was mentioned previously,
the resolution time correction can have a significant effect on the nor
malizing factor since widely differing count rates are employed. It is
extremely hard, if not impossible, to obtain a proper resolution time
correction based on integral counts determined from a time-varying count
rate.
Two methods of normalization and their relative merits are discussed
below.
The Integral Count Method is the normally employed method of nor
malization. The total counts accumulated with the NDDAS for all runs
is referred to a predetermined one, with the normalizing detector in a
fixed position while the movable detector changes position.
The Analyzer Method, which is essentially the same as the above
except that the counts arising from the normalizing detector are stored
as a function of time in a multichannel analyzer. The average of a
series of ratios obtained by dividing the corrected channel counts by
the corresponding channel counts of a predetermined measurement gives
the normalizing factor.


175
yielded the decay constants at different times after the pulse and
computed a value of p ($), as well as absolute p and k^^ using a
calculated fc>ef£ for each decay constant.
The Decay Constant
A series of decay constants were obtained by analyzing the tail
of the pulse at increasing times from the peak. The convergence of the
decay constant in time was sought, before background is reached, and
the value that gave the least deviation according to Peierls statis
tical analysis was considered to give the decay constant at that posi
tion.
The decay constants determined close to the source decreased
monotonically as the fitting was performed farther away in time from
the peak. Once the "asymptotic" source region, as defined in the pulse
propagation measurements, was reached, an apparent decay constant was
found at each position. This decay constant is observed to converge
or to vary very slowly with respect to time and decreased in magnitude
as the distance to the source increased. At large distances from the
source a fundamental decay constant is observed, seemingly when an
asymptotic distribution of the neutrons is established in the assembly,
more than 9 msec after the burst.
An analysis of the time-space behavior of the neutron flux (ref. .
Fig. 28) reveals that, indeed, more than 9 msec are required for the
assembly to achieve an asymptotic neutron distribution. Therefore,
only at large distances from the source can a true fundamental decay
constant be measured; prior to P83 (142.9 cm) background is reached
before a uniform spatial distribution is established.


FLUX (RELATIVE UNITS)
FIG- C22 TIME PROFILE OF THERMAL NEUTRON FLUX
2B-43 CM FROM THE SOURCE
218


UX (RELATIVE: UNITS)
FIG C4 TIME PROFILE OF THERMAL NEUTRON FLUX 4115 EM FROM THE SOURCE


24
side from the linear channel No. 1.
Items a. through d. are part of the safety amplifier while item f.
is used to display the neutron flux on a console front panel meter.
The safety amplifier monitors the seven continuously varying input
signals and provides a trip signal if any of the input signals fall
outside of acceptable limits. The safety amplifier provides means of
adjusting these limits over a wide range.
The duality of the scram action (see Fig. 8) is a prominent feature
of the safety system. It can be said that no single failure will
invalidate both automatic scram channels. Furthermore, it has been
determined that no single failure can invalidate both the manual and
automatic scrams. The method of measurement and the function of each
instrumentation and safety channel are shown in Table II (parts A and B).
A series of safety interlocks prevent water from flowing into or
remaining in the assembly unless a proper sequence of events are fol
lowed and certain conditions are satisfied. The conditons are:
a. The moderator temperature must be >_ 60F. This is established
by the desire to obtain the experimental data near room temperature con
ditions. The insertion of water at 32F will introduce a maximum k of
.00342 (based on the calculated negative temperature coefficient of
reactivity) above the design k^^ value with no hazards created.
b. The four instrumentation channels must have their high voltage
on.
c. The core width must be smaller than 26 cm.
d. The door to the assembly room must be locked.
e. The start-up channel must count more than 2 counts/sec.
f. The neutron flux, subcritical assembly power level and period


CHAPTER I
INTRODUCTION
The development and construction of large power reactors focused
the attention of industry and of the United States Atomic Energy Com
mission on the necessity of having reliable reactor dynamics analysis
methods to accurately describe the spatial and temporal behavior of the
neutron flux in these systems. The point-model reactor kinetics calcu-
lations seem to have been adequate for the gross evaluation of the time-
dependent neutron flux during the occurrence of a transient but the
model can be in large error when the physical size of the system and the
magnitude of the perturbation necessitates that spatial effects in the
redistribution of the neutron flux be considered. Preliminary calcula
tions done by Johnson and Garner using a one-dimensional space-time
kinetics model [1] showed that the space-time dependent scheme predicts
a "destructive zone" much larger than that predicted by point-model
kinetics.
The necessity of experimentally determining the validity of the
various space-time kinetic analysis methods was brought out by Johnson
and Garner [1], and recognized by the USAEC in establishing the Large
Core Dynamics Experimental Program. The primary responsibility for this
program has been vested in the Nuclear Safety Research Branch, Atomic
Energy Division, Phillips Petroleum Company as major contractor for
USAEC.
2


FLUX (RELATIVE UNITS)
FIG- C25 TIME PROFILE OF THERMAL NEUTRON FLUX
EG-58 CM FROM THE SOURCE
221


- 14
ALUMINUM SUPPORT
FIG. 3 TOP FUEL ROD SPACING SYSTEM


35
Operating Limits
A description of the operating limits of the subcritical assembly,
including the basis for such limits ate listed below.
Effective Multiplication Factor
Specification: the maximum allowable will be 0.985+.005.
The absolute value of k as well as the slope of the k vs. water
eff r eff
height will be carefully measured so as not to exceed the limiting value.
Basis: the upper limit of kg^^ = 0.985+.005 is established by:
the accuracy with which k^^ can be measured, the reported [7] differ
ences between calculations and experiments in similar cores and the value
of the multiplication factor required to make a meaningful study of the
dynamic properties of large cores. Comparison [7] between 29 calcula
tions and the corresponding critical experiments (on cores similar to
the UFSA) established that an overestimate of kg^^ is generally made; the
standard deviation for these cases was + 0.00175 and the maximum under-
stimate of the multiplication was for a case yielding kg^^ = 0.9966, a
0.34% deviation.
The absolute value of k will be determined for water levels
eff
yielding a k^^ >.95 by pulsed techniques independently of the inverse
multiplication measurements. The method to be used is the Garelis-
Russell[8] method which, when appropriate corrections are made for the
reflector, has been shown to give good results. In this method both
l-keff(l-$) keff8
a = and are determined; then ke££ may be. obtained by
independently obtaining B or L Since of the two parameters 8 changes
the most slowly with k^^, it is valid to use a value based on theoret
ical calculations.


FLUX (RELATIVE UNITS)
FIG C14 TIME PROFILE OF THERMAL NEUTRON FLUX IBS-31 CM FROM THE SOURCE


133
"asymptotic region of the assembly. In this region a linear rela
tionship between delay times and distance from the source exists.
The asymptotic velocity of propagation for the clean UFSA core is
given in Table IX.
TABLE IX
ASYMPTOTIC PROPAGATION VELOCITY v
P
UFSA R1 Clean Core
0.5 M/W Ratio
16.35 cm reflected core
Input Pulse
Width (msec)
Theory
vp (m/sec)
Experiment
0.5
456
448+9
1.0
456
452+7
It should be mentioned that the asymptotic velocity of propaga
tion is independent of the assumed distribution of the source. When
a localized source in space was introduced into the WIGLE scheme the
results yielded the same asymptotic velocity as the calculations with
a spatially distributed source. This is to be expected since v^ is a
characteristic of the system and has been defined in a region removed
from source effects.
The Dynamic Inverse Relaxation Length
. The dynamic inverse relaxation length, k^, is defined as the
inverse of the distance required for the amplitude of the pulses to
attenuate by a factor e. is determined by the relationship between
the logarithm of the magnitude of the peak of the propagating pulses
and distance in the "asymptotic" region where this relationship is
linear.


Flux (Relative Units)
140
FIG. 34 AMPLITUDE ATTENUATION OF THE THERMAL FLUX CALCULATED BY
THE WICLE CALCULATIONAL SCHEME FOR DIFFERENT CORE HEIGHTS


TABLE OF CONTENTS (cont'd)
Operating Limits
Design Basis Accident Analysis
CHAPTER IV THE DATA ACQUISITION SYSTEM
Introduction .
The Neutron Detector
The Electronic Instrumentation
The Resolution Time Correction ..
The Normalization Technique
Comments
CHAPTER V NUCLEAR CALIBRATION OF THE UFSA SUBCRITICAL
Introduction
Theoretical Notes
Inverse Multiplication Measurements
Absolute Determination of k
eff
Conclusions
Page
35
39
45
45
47
50
54
61
62
64
64
64
68
73
75,
PART 2
SPACE-TIME REACTOR KINETICS STUDIES WITH
THE UNIVERSITY OF FLORIDA SPERT ASSEMBLY 79
CHAPTER I INTRODUCTION 80
Statement of the Problem 80
Description of the Study 81
vii


FLUX (RELATIVE UNITS)
FIG- C34 TIME PROFILE OF THERMAL NEUTRON FLUX 181-03 CM FROM THE SOURCE
230


125
e
!=>
0)
>
te
t-H
Pi
X
d
iH
tn
Position Number
FIG. 28A EXPERIMENTALLY DETERMINED SPATIAL DISTRIBUTION OF
THE NEUTRONS AT DIFFERENT TIMES AFTER THE PULSE


UX (RELATIVE UNITS)
FIG- C8
TIME PROFILE GF THERMAL NEUTRGN FLUX
35"01 CM FROM THE SOURCE


17
an automatically operated pneumatic control valve. A plot of the flow
rate versus the height of the water level above the apex of the weirs
is shown in Fig. 5.
The weir plate is rigidly mounted on a "box" or small tank (see
Fig. 6). The weir "box" is connected to a drive mechanism composed
of the following: guide post, slide block, and drive screw. The guide
post is a 2 inch diameter pipe attached to the support column of the
crane which is used for removal of the reflector tanks. The "box" is
mounted on the guide block which slides along the vertical post and
provides a rigid support for the system and is driven up and down by
means of the drive screw which is fixed at the top of the guide post
support and passes through the guide block. The upper limit of the
position of the weir"box"is controlled by mechanical stops whose posi
tion is determined as part of the initial start-up procedure for each
configuration to be considered.
The final adjustment of the position of the weir"box"is such that
when the water reaches the moderator level in the assembly corresponding
to *99 for a given configuration it will be flowing about 2.0
inches above the apex of the V-notch weirs. At this design level the
flow rate is ~ 7 gallons per minute with the k^^ values as given in
Table I for the full fuel loading. After the operating height of the
weir box has been determined for kg^^ <^0.99 in the initial start-up,
stops are inserted to prevent raising the weirs above this height (if
the height is less than the active fuel height). It should also be
pointed out that the orifice in the line limits the pump capacity to a
flow rate which is just sufficient to bring the weirs to full flow. A
further increase in flow rate would cause discharge over the entire


9
8
7
6
5
4
3
2
1
O
163
Position Number
. 41 PHASE OF ZEROTH FOURIER MOMENT vs. DISTANCE FOR
SEVERAL FREQUENCIES
- 1.0 MSEC INPUT PULSE -


TABLE OF CONTENTS (cont'd)
Page
Nomenclature Used in the Description of
Pulse Propagation Phenomena 83
CHAPTER II THEORETICAL NOTES 85
Introduction 85
Review of the Literature 85
The WIGLE Calculational Scheme 87
Neutron Wave Analysis 89
CHAPTER III DESCRIPTION OF THE MEASUREMENTS 91
Introduction 91
The Epicadmium Subtraction Method 92
The Geometrical Arrangement 94
Synopsis of the Measurements 97
CHAPTER IV EXPERIMENTAL AND THEORETICAL RESULTS IN THE
TIME DOMAIN 100
The Analytical Model 100
Flux Traverses 109
Clean Core Pulse Propagation Measurements 118
Propagation of a Narrow Pulse 147
Propagation of a Wide Pulse 148
Pulse Shape vs. Input Pulse Width 150
Effect of Room Return at Peripheral Detector Positions 152
viii


FLUX (RELATIVE UNITS)
1-0
0*3
0-E3
0*7
0-S
0*5
0-4
0*3
0-3
0-1
RIFS 41-1
PEAK AT 0-340 MSECS
PULSE WIDTH =0-5 MSECS
THEGRY FWHM = 1-7BS MSECS
EXP- FWHM = 1*710 MSECS
to
u>
MO
10
TIME (MSE:C)
TIME PROFILE OF FAST NEUTRDN FLUX 66-5B CM FRDM THE SOURCE
FIG- D4


TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS iii
PREFACE V
LIST OF TABLES xi
LIST OF FIGURES xiii
LIST OF SYMBOLS xviii
ABSTRACT xix
i '
PART 1
THE UNIVERSITY OF FLORIDA SPERT ASSEMBLY
- DESIGN AND CALIBRATION 1
CHAPTER I INTRODUCTION 2
CHAPTER II DESCRIPTION OF THE FACILITY 5
General Features .. 5
Fuel Characteristics 9
Mechanical Design 10
Moderator Flow Control System 15
Instrumentation and Interlock System 22
Fuel Storage 27
Neutron Sources 30
CHAPTER III OPERATIONAL SAFETY 32
Introduction 32
Initial Loading 33
vi


FLUX (RELATIVE UNITS)
FIG- D5
TIME PROFILE OF FAST NEUTRON FLUX 79-30 CM FROM THE SOURCE
240


99
the Johnson criterion [25] a wide, 10 msec square pulse was introduced
into the assembly. Three selected positions were used for these meas
urements.
Pulse Propagation vs. Pulse Width
During the experimentation described above, the observation was
made that narrow input pulses will yield the same pulse shape in the
asymptotic region, with the peaks displaced in time. To gain an
insight into this predicted occurrence, measurements were taken at one
detector location (P83) with input pulse widths of 0.1, 0.5, 1.0, 2.0,
3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0 and 10.0 msec.
Effect of Room Return at Peripheral Detector Positions
The effect of neutrons reflected from the concrete walls on the
measurements was investigated by recording pulse shapes at peripheral
detector positions. The measurements were done with and without a
shield of borated paraffin in the region in which the detector was
located.


FLUX (RELATIVE UNITS)
FIG- C28 TIME PROFILE OF THERMAL NEUTRON FLUX 104-73 CM FROM THE SOURCE
224


:~LUX (RELATIVE UNITS)
R1REFREF7G
PEAK AT 0-4S0 MSECS
PULSE WIDTH = 05 MSECS
EXP" FWHV1 = 0-570 MSECS
UNSHIELDED SHIELDED
M
Ui
-C~
4

3 4 5 S 7
a
8
CJ
10
TIME (MSEC)
FIG- 37B EFFECT DF ROOM RETURN AT PERIPHERAL DETECTOR POSITIONS


63
microsec or longer are stable.
B) The address current setting of the CN1024 is extremely criti
cal, markedly so for high count rates.
C) An optimum pulse into the analyzer should have a rise time of
50 nanosec, a total width of .1 microsec and an amplitude of 3-5
volt.
D) Above noise level, the discriminator setting of the 212 logic
unit becomes irrelevant when a constant pulse height is used as input.
E) Reproducibility tests were performed on the analyzer with the
neutron generator in continuous mode.
The statistical analysis of the channel counts gave:
72% were less than 1 from the mean
25% were between 1 and 2 from the mean
3% were between 2 and 3 from the mean
The system is statistically well-behaved.


189
UFSA R1 CORE PARAMETERS
0.5 M/W ratio
16.35 cm wide reflected core
Group
1
2
3
4
D
1.6079
0.94239
0.61401
0.20809
^a
0.0035411
0.0023701
0.23109
0.20809
Zr
0.090927
0.096727
0.084015
_
vlf
0.0068929
0.0012609
0.015301
0.22662
1/v
5.25 Ox 10-10
2.637 x 10"9
1.843 x 10~7
3.324 x 10"6
_3
Void Coefficient of Reactivity: 3.14 x 10 Ak/% void
Temperature Coefficient of Reactivity: 1.03 x 10 ^ Ak/C
Determination of k
eff
Using the constants tabulated above, eigenvalue calculations were
performed with the AIM-6 computer code [53].
' AIM-6 is a multigroup, multiregion, one-dimensional diffusion
theory code that has been widely used for criticality calculations.
The code will handle as many as 18 energy groups, 101 space points
and 20 regions. Basically, the code solves the equation
- D^Vir) + 4 4(r) = x^ir) + ^ E j+i j(r)
i J-q s
1 < i < NOG < 18. NOG = number of groups
The symbols follow conventional notation and
X* = the integral of the fission spectrum over the lethargy range
represented by group i.


97
penetrated 10 cm into the core. The purpose of this arrangement was
to increase the solid angle subtended by the target in the assembly to
augment the number of neutrons going into the system and to reduce the
number of unscattered fast neutrons dispersing into the room and
creating a shielding problem.
Synopsis of the Measurements
Flux .Traverses
A series of flux traverses were done on the assembly to determine
the flux shapes across the width and the height of the assembly. The
traverses were done dynamically and statically.
Static traverses were performed with Indium foils at a distance of
66.6 cm from the neutron source. Measurements were done on the width
and height of the assembly with the neutron generator in the continuous
mode for a period of 6 hours to achieve the saturation activity of the
foils. The foils were 5 mil thick, 7/16" in diameter, and 99.97+%
Indium.
Dynamic traverses were performed across the width of the core at 41,
54, 130 cm from the neutron source with the generator in the pulsing
mode to determine whether any propagation occurs in the tranverse
3
direction. The long He detectors were used for the measurements,
which employed pulse widths of 0.5 and 1.0 msec as input.
Clean Core Pulse Propagation Measurements
A whole series of measurements using narrow, square, fast bursts
of neutrons as a disturbance were carried out to determine the most
important characteristics of the pulse propagation phenomena. Most
of the measurements utilized the entire length of the assembly. Certain


LIST OF FIGURES (cont'd)
FIGURE Page
46 COMPARISON OF THE THEORETICALLY PREDICTED
AND THE MEASURED 2 a F, ... 172
47 THE UFSA R1 CORE p2 DISPERSION LAW 173
48 DECAY CONSTANT vs. AXIAL POSITION 177
49 kB/A vs. AXIAL POSITION 178
50 REACTIVITY (-$) k f vs. AXIAL POSITION 182
Cl Through C19 TIME PROFILES OF THE THERMAL
NEUTRON FLUX AT NINETEEN POSITIONS IN THE
CORE FOR A 0.5 MSEC INPUT PULSE 195-213
C20 Through C38 TIME PROFILES OF THE THERMAL
NEUTRON FLUX AT NINETEEN POSITIONS IN THE CORE
FOR A 1.0 MSEC INPUT PULSE 214-232
Dl Through D6 TIME PROFILES OF THE FAST NEUTRON
FLUX AT SIX POSITIONS IN THE CORE FOR A 0.5
MSEC INPUT PULSE 234-239
D7 Through D12 TIME PROFILES OF THE FAST NEUTRON
FLUX AT SIX POSITIONS IN THE CORE FOR A 1.0
MSEC INPUT PULSE 240-245
El Through E4 TIME PROFILES OF THERMAL NEUTRON
FLUX AT FOUR POSITIONS IN THE CORE FOR A 0.1
MSEC INPUT PULSE 247-250
FI TIME PROFILES OF THERMAL NEUTRON FLUX AT THREE
POSITIONS IN THE CORE FOR A 10.0 MSEC INPUT PULSE 252
G1 Through G2 SHAPE OF THE PROPAGATING PULSE AS A
FUNCTION OF PULSE WIDTH 254-255
xvii


66
[13], Gozani's extrapolated area-ratio analysis [14] or the Garelis-
Russel technique [8]. In all these techniques it is essential that a
fundamental spatial distribution of the neutrons be established for a
correct determination of the decay constant and, therefore, the reac
tivity of the system.
The Simmons and King method established that a value for the
reactivity can be obtained directly if a prompt fundamental decay con
stant can be measured at delayed critical. The value of a at delayed
critical determines B/£ and if these parameters are assumed constant
over the reactivity range of interest a value of a can be obtained. The
technique has given good results up to $20 subcritical in small mul
tiplicative systems. The method strongly depends on being able to
establish the prompt fundamental decay mode; it suffers from the incon
venient necessity of a delayed critical measurement and the assumed
constancy of 8/£ throughout the ranges of reactivity.
The Sjostrand method improves the Simmons and King method in that
the delayed critical measurement is no longer necessary but the results
are shadowed by the strong influence of higher spatial harmonics. The
method is based on the premise that the impulse response curve of the
system is dominated by the prompt fundamental mode.
Gozani's treatment is a significant improvement over Sjostrand's
method. Gozani proposed the extraction of the fundamental mode of
prompt neutron decay from the impulse response curve and the extrapola
tion of this curve to zero time. The reactivity in dollars can be found
by integrating under this curve; the method is independent of the pres
ence of higher prompt spatial modes.
The Garelis-Russell technique, similar to Gozani's extrapolated


152
"characteristic asymptotic" pulse shape already exists. Characteristic
"dumbbell" shapes are observed however until the HWHM ^ propagation time.
For wider input pulses, the "propagating" pulses are unable to achieve
the smooth shape observed for narrower pulses. Still, a certain
spatial "re-arrangement" of the shape occurs, as was discussed above
for the 10 msec input pulse.

Effect of Room Return at Peripheral Detector Positions
At positions farthest removed from the source, it was observed
that a secondary peak was occurring at times corresponding to the input
pulse width. This was also observed at peripheral positions in the
reflector tanks. To prove that neutrons reflected from the concrete
walls were the cause of this secondary peak, measurements were made at
the outer edge of the reflector tanks with the neutron detector
shielded by borated paraffin placed outside the reflector tank and with
the detector unshielded.
The results of these measurements are shown in Fig. 37A. A
pronounced peak at M).5 msec was found when the detector was not
shielded from the walls and practically disappeared when about 6" of
borated paraffin was placed between the reflector tank and the concrete
wall. The input pulse width was 0.5 msec. Shown in Fig. 37B is the
result of subtracting the "shielded" pulse from the "unshielded" one.
A sharp 0.5 msec pulse is obtained. The "early" peaks are therefore
blamed on fast neutrons reflected from the walls.
The penetration of these neutrons into the reflectors was analyzed
by positioning a detector at increasing distances from the outer reflec
tor tank wall. It was found that the effect was negligible at a


FLUX (RELATIVE UNITS)
FIG- C15 TIME PROFILE OF THERMAL NEUTRON FLUX 181-03 CM FROM THE SOURCE


TABLE OF CONTENTS (cont'd)
Page
F UFSA R1 CLEAN CORE
TIME PROFILES OF THEKMAL NEUTRON FLUX AT THREE
POSITIONS IN THE CORE FOR A WIDE (10 MSEC) INPUT PULSE ... 253
G UFSA R1 CLEAN CORE
SHAPE OF THE PROPAGATING PULSE AS A FUNCTION OF INPUT
PULSE WIDTH 255
LIST OF REFERENCES 258
BIOGRAPHICAL SKETCH 262
x


Flux (Relative Units)
R1 Clean Core
66.6 cm from the source
In Foil Activation
FIG. 26 THE ASYMPTOTIC STEADY-STATE HORIZONTAL FLUX
117


Reactivity (-$)
FIG. 50 REACTIVITY (-$) AND k .. VS. AXIAL POSITION
eff
Effective Multiplication Factor


108
configuration of the source and the assembly. The planar nature of the
source was confirmed experimentally by obtaining pulse propagation data
across the width of the core close to and far from the source.
The removal cross-section for the uncollided source neutrons in
region 1 was chosen to be the removal cross-section of neutrons from
the fast group in that region as obtained from the two-group scheme
(ref. Table IV). For regions 2,3,4 and 5 the corresponding removal
cross-section was taken to be the removal cross-section of neutrons
from the fast group in the four-group scheme.
It should be noted that the results are not sensitive to the value
of the removal cross-section in region 1. The above choice of the
\
removal cross-section for the remainder of the assembly is consistent
with the overall calculational scheme used and with the value deter
mined empirically for 14 Mev neutrons in light water.
This choice of a first-flight source gave calculational results
in surprinsingly good agreement with experiment, particularly for the
delay times and the spatial attenuation. The predicted pulse shapes
are somewhat narrower than the pulse shapes obtained from experiment.
No significant differences between the results of the localized sources
case mentioned above and a point kernel source distribution were ob
served. The point kernel source distribution underpredicts the pene
tration of the first collision fast neutrons into the assembly.
At this point a few comments on the energy distribution of the
source neutrons are pertinent. A pure tritium target bombarded with a
beam of monoenergetic deuteron ions will produce essentially mono-
energetic neutrons, since the relationship between the energy of the
ions, the neutron energy and the angle of emission is unique. However,


FIG. 10 PHYSICAL CHARACTERISTICS OF THE He3 NEUTRON COUNTERS


146
TABLE XII
THE CALCULATED ASYMPTOTIC VELOCITY OF PROPAGATION AND
DYNAMIC INVERSE RELAXATION LENGTH VS. CORE HEIGHT
- 0.5 MSEC INPUT PULSE WIDTH -
UFSA R1 Core
0.5 M/W Ratio
16.35 cm wide reflected core
Core
Kd
V
Height
(cm)
Q
(cm X)
P
(msec )
70
0.03565
571
72.5
0.03233
506
76
0.03027
456
77.5
0.02772
435
80
0.02515
397
85
0.02047
321


WEIR SYSTEM
6
CORE TANK
REFLECTOR
TANK
RESERVOIR
DUMP
VALVES
FIG. 1 OVERALL VIEW OF THE FACILITY


FLUX (RELATIVE UNITS)
FIG* D7
TIME PROFILE OF FAST NEUTRON FLUX 15-71 CM FROM THE SOURCE
in in


75
has been resolution time corrected and background subtracted. A "pure"
delayed neutron background is statistically calculated and used to
determine k$/£. The data can be normalized to a reference detector
position for the analysis of the pulse propagation measurements in the
time and in the frequency domain. A Fourier analysis of the pulse can
also be performed if required.
Shown in Fig. 17 (A, B) are the experimentally determined a, kg/i,
and k as a function of moderator height obtained by averaging results
from three chosen detector locations in the "asymptotic" region. The
results are summarized, together with the 1/M measurements and theoreti
cally calculated values in Table III. The excellent agreement between
the experimental and theoretical results should be considered somewhat
fortuitous. The calculations were done following the method outlined
in Appendix A. Some later calculations [19] done by the Phillips
Petroleum Co.,showed more disagreement, especially at low water levels.
The last calculations tried to account for the fact that there is a
fissionable reflector above each experimental moderator height. This
fact was disregarded in the calculational results shown in this work.
The agreement at the 75 cm water level is good for all calculational
methods.
Conclusions
The University of Florida SPERT Assembly has been operated for
several months with very few operational problems. The system has
proven to be extremely reliable and the instrumentation has performed
adequately. The calibration of the system has established that mean
ingful values for the reactivity can be obtained when applying the


Delay Time (msec)
130
ft
ft
FIG. 30 CALCULATED AND EXPERIMENTAL DELAY TIMES
- 1.0 MSEC INPUT PULSE -


Flux (Relative Units)
FIG. 25 THE ASYMPTOTIC STEADY-STATE VERTICAL FLUX
116


261
45. J. H. Dunlap and R. B. Perez, "Dispersion Law for a Subcritical
Assembly," Proceedings of the Symposium on Neutron Noise, Waves,
and Pulse Propagation, AEC Symposium Series No. 9 (1967).
46. CORA, an improved version of the IMP computer program. IMP is
described in IDO-17199.
47. S. 0. Johnson, SPERT PROJECT Quarterly Progress Report, Philips
Petroleum Co., IDO-17123, 43 (1966).
48. G. Mortensen, private communication (1968).
49. A. E. Waltar and L. Ruby, "Interpretation of Pulsed-Source Experi
ments in a Reflected Reactor," Internal Report, Department of
Nuclear Engineering, University of California (1964).
50. PHROG is a Phillips-Hanford revision of GAM-1 (G. D. Joanou and
J. S. Dudek, GAM-1, GA-1850 (1961).
51. F. J. Wheeler, "RAVEN, A Computer Package for Evaluating Resolved
and Unresolved Resonance Absorption Including Pin Shadowing,"
IDO-17212 (1964).
52. TOTEM links the TOPIC and TEMPEST codes for calculating thermal
constants (G. E. Putnam, TOPIC, IDO-16968 (1964), and R. H. Shudde
and J. Dyer, TEMPEST II, TID-18284 (1961)).
53. H. P. Flatt and D. C. Bailer, "AIM-6, A Multigroup, One-Dimensional
Diffusion Equation Code (1961).
54. N. J. Diaz and M. J. Ohanian, "UNIPUL, A Unified Data Processing
Program for Pulsed Measurements," Internal Report, University of
Florida (1969).
55. M. J. Ohanian, "MORE, A Conventional Fourier Transform Program for
Pulsed Measurements," Internal Report, University of Florida (1969).
56. M. J. Ohanian, "MORWIG, A Conventional Fourier Transform Program
with Variable Time Increments," Internal Report, University of
Florida (1969).
57. N. J. Diaz, "ALXILS, A Linear Least-Squares Program for the Fitting
of Alpha and Xi," Internal Report, University of Florida (1969).


119
TABLE VI
DELAY TIMES MEASURED ACROSS THE WIDTH OF THE CORE
- 0.5 MSEC INPUT PULSE -
UFSA R1 core
0.5 M/W ratio
16.35 cm wide reflected core
Delay Times (msec)
Position Distance to Distance to the source
No Core Center (cm) 53.9 cm 130.2 cm
A
7.27
.860
2.42
B
5.45
.830
2.51
C
3.63
.830
2.48
D
1.82
.830
2.45
E.
0.0
.830
2.42
F-
3.63
.830
2.54
G
5.45
.830
2.42
H
7.27
.860
2.51
I
10.9
.860
2.42
RE1
12.5
.890
2.45
RE2
14.3
.920
2.48
RE3
16.1
.950

RE4
17.9
.920
2.57
RE5
28.9
.920

RE6
31.8

2.45
RE7
34.3
.890



105
to zero with 2-5 ysec (10 time steps) between the two steps. It was
also observed that a long time after the input pulse was cut-off oscil
lations developed at spatial points close to the source when time
imcrements larger than 50 ysec were used.
The time increments used for the 0.5 and 1.0 msec input pulse
problems are shown in Table V. Stable operation of the code was ob
tained with these time steps. Running times were of the order of 20
minutes in the UF IBM 360-50.
The main problem in establishing a fair comparison between the
calculational and experimental results lies in an adequate physical
description of the spatial distribution of the external fast source.
For example, when calculations were done with a spatially localized,
1 cm wide source placed at the physical location of the neutron target
(Fig. 20) the calculated pulse shapes were found to agree reasonably
well with experiment but the peaks of the pulses were delayed by 50 to
500 microsec with respect to the experimental values. Furthermore the
calculations predicted a much sharper attenuation of the peak of the
pulse near the source, although predicting an asymptotic relaxation
length in agreement with experimental results. Therefore, a careful
study was undertaken to obtain the "best" description of the spatial
distribution of the source.
The best approximation to the source that could logically be
assumed based on physical grounds was that of a spatial distribution
I Z
given by a first-flight kernel of the form e r where £r is the removal
cross-section for source neutrons. The distribution used is shown in
Fig. 22. The choice of a plane source kernel is reasonable especially
at points a few cm removed from the source due to the physical


FLUX (RELATIVE UNITS)
X
FIG* C9 time PROFILE CF THERMAL NEUTRON FLUX 104-73 CM FROM THE SOURCE


LIST OF SYMBOLS
TRANSVERSE BUCKLING
DELAYED NEUTRON PRECURSOR CONCENTRATION OF THE
ith GROUP
DIFFUSION COEFFICIENT OF THE ith GROUP
FREQUENCY (cps)
EFFECTIVE MULTIPLICATION CONSTANT
NEUTRON LIFETIME
NEUTRON MULTIPLICATION
VELOCITY OF THE ith GROUP
AXIAL COORDINATE
DECAY CONSTANT (IN THE TIME DOMAIN)
DAMPING COEFFICIENT (IN THE FREQUENCY DOMAIN)
EFFECTIVE DELAYED NEUTRON FRACTION
PHASE SHIFT PER UNIT LENGTH
REACTIVITY .
COMPLEX INVERSE RELAXATION LENGTH
MACROSCOPIC CROSS SECTION
NEUTRON FLUX
xviii


TABLE VIII
CLEAN CORE PULSE PROPAGATION STUDIES
EXPERIMENTAL AND THEORETICAL RESULTS
- 1.0 MSEC INPUT PULSE WIDTH -
UFSA R1 Core
0.5 M/W Ration
16.35 cm wide reflected core
POSITION
DISTANCE TO
PEAK
AT (msec)
FWHM
(msec)
PEAK
COUNTS3
NO
SOURCE (cm)
Theory
Exp
Theory
Exp
Theory
6
3
1.000
0.97+.03
1.020
0.930+.06
48.25
56.5+1.5
13
15.71
1.000
1.00+.03
1.065
0.960+.06
47.62
56.7+1.7
20
28.43
1.00
1.03+.03
1.140
1.080+.06
36.52
37.9+.9
27
41.15
1.046
1.04+.03
1.365
1.290+.06 .
23.60
29.6+1.6
34
53.86
1.215
1.19+.03
1.620
1.62+.06
14.99
14.9+.2
41
66.58
1.410
1.33+.03
1.875
1.92+.06
9.720
10.1+.3
48
79.30
1.630
1.65+.06
2.175
2.19+.06
6.431
6.43+.05
55
92.01
1.880
1.81+.06
2.468
2.49+.09
4.332
4.32+.04
62
104.73
2.150
2.10+.06
2.760
2.82+.09
2.955
3.11+.04
69
117.44
2.420
2.41+.06
3.030
3.12+.09
2.041
2.15+.05
76
130.16
2.725
2.71+.06
3.315
3.45+.09
1.423
1.41+.02
83
142.88
3.025
3.01+.06
3.585
3.69+.09
1.000
1.00+.03
90
155.60
3.275
3.30+.06
3.750
3.96+.09
0.7070
0.663+.01
97
168.31
3.575
3.58+.06
3.945
4.29+.09
0.5019
0.489+.009
104
181.03
3.850
3.85+.06
4.110
4.44+.12
0.3553
0.356+.007
111
193.74
4.075
4.10+.08
4.185
4.59+.12
0.2471
0.254+.003
118
206.46
4.275
4.30+.08
4.200
4.56+.12
0.1640
0.184+.003
125
219.17
4.425
4.475+.09
4.275
4.53+.12
0.0962
0.121+.0025
132
231.88
4.500
4.600+.12
4.275
4.56+.15
0.0369
0.0585+.00012
a Normalized to Position 83
k Errors assigned from the deviation from the mean of two measurements except for the last 5 space
points. This error is usually larger than the counting statistics error. The error of the last 5
points was calculated from counting statistics.
128


164
TABLE XVI
THE REAL AND THE
IMAGINARY COMPONENTS OF
THE COMPLEX INVERSE RELAXATION
LENGTH
- 0.5 MSEC INPUT PULSE -
FREQUENCY
a (cm
'1)
5
(rad/cm)
(cps)
Theory
Exp
Theory
Exp
0
.02286
.02248
10
.02290
.02254
.001557
.001558
20
.02305
.02270
.003095
.003098
30
.02328
.02296
.004595
.004600
40
.02359
.02331
.006046
.006054
50
.02396
.02373
.007435
.007451
60
.02438
.02423
.008765
.008780
70
.02483
.02476
.01004
.01006
80
.02532
.02536
.01128
.01124
100
.02640
.02663
.01358
.01349
120
.02755
.02797
.01563
.01541
140
.02866
.02937
.01753
.01714
160
.02986
.03081
.01932
.01868
180
.03101
.03223
.02092
.02026
200
.03215
.03359
.02244
.02148
220
.03330
.03495
.02391
.02278
240
.03442
.03627
.02521
.02396
260
.03545
.03757
.02650
.02508
280
.03658
.03878
.02775
.02619
300
.03766
.04002
.02883
.02726
350
.04028
.04285
.03153
.02979
400
.04272
.04576
.03401
.03213
450
.04505
.04813
.03627
.03434
500
.04724
.05018
.03841
.03661
550
.04941
.05272
.04037
.03803
600
.05146
.05475
.04224
.04060
650
.05344
.05754
.04391
.04138
700
.05534
.05934
.04550
.04234
750
.05727
.06095
.04695
.04118
800
.05910
.06237
.04829
.04425
900
.06276
.06852
.05091
.05150
1000
.06636
.07404
.0530
.05650


APPENDIX F
UFSA R1 CLEAN CORE
TIME PROFILES OF THERMAL NEUTRON FLUX
AT THREE POSITIONS IN THE CORE
FOR A WIDE (10 MSEC) INPUT PULSE


CHAPTER IV
THE DATA ACQUISITION SYSTEM
Introduction
In the delicate and laborious task of performing nuclear experi
ments the most common source of difficulties and errors lies in the
acquisition of the data. Modern nuclear instrumentation with its excel
lent time-energy resolution has enhanced detection sensitivity to the
extent that deviations previously masked by the poorer resolution of the
equipment are easily distinguishable. The major problem now lies in the
degree of reproducibility of the results. Although in general the in
strumentation is very reliable the enhanced sensitivity demands contin
ual standardization for the sake of reproducibility.
The "brain" of the data acquisition system for pulsed neutron
measurements is the multichannel analyzer (MCA) which is now available
with a large number of data channels and narrower channel widths for
increased time resolution. In general, the mode of data acquisition
for a pulsing experiment differs from that of many other types of
nuclear experiments. In particular, neutron interactions with a suitable
detector are fed into the analyzer while it is time-sweeping. Ideally
every neutron interaction should be counted regardless of the ampli
tude of the collected pulse. This implies that the linear signal
originated at the neutron detector should be converted to a logic signal
so that its probability of being recorded is independent of its amplitude
45


FLUX (RELATIVE UNITS)
FIG. 35B THE SENSITIVITY OF THE ONE-DIMENSIONAL, TWO GROUP, SPACE-TIME KINETICS SCHEME
TO CHANGES IN THE TRANSVERSE BUCKLING
142


FLUX (RELATIVE UNITS)
FIG- Cl
TIME PROFILE OF THERMAL NEUTRON FLUX 3-0 CM FROM THE SOURCE


55
Sr1' >rhf r 1 -k'1/M
where M is the net neutron multiplication in the assembly with a
centrally located source. In practice, the multiplication is obtained
from the ratio of multiplied to unmultiplied counts with a centrally
located source. The unmultiplied counts are obtained with the fissile
material removed and all other conditions undisturbed. In water
moderated cores it is difficult to match neutron spectra for multiplied
and unmultiplied counts and deviations from the ideal M are to be ex
pected. If possible, a search for detector locations should be con
ducted so as to obtain curves that follow the expected behavior of 1/M.
Even if k can not be directly inferred from the 1/M determination, the
curve of reciprocal count rate vs. the parameter that controls reac
tivity (fuel loading or moderator height or % control rod withdrawal)
is a useful guide for safely approaching criticality if a well-behaved
curve can be obtained.
The inverse multiplication curve can be obtained as a function of
moderator height by first obtaining a series of unmultiplied counts at
various water levels and the multiplied counts as the water level is
raised with the assembly originally air-spaced. Sensitivity to geo
metrical configuration (source-detector-water level) requires an empir
ical determination of "well-behaved" detector positions.
Reactivity Measurements by the Pulsing Technique
The pulsed-neutron technique has been used successfully for several
years to measure reactivity. The transient neutron density following a
burst of neutrons is used to determine the reactivity of the system by
either the Simmons and King method [12], Sjostrand's area ratio method


28
the storage area to the level of the active fuel height. A 2-foot
reflector on both sides was used to represent an infinite reflector.
No reflector was considered on the ends, but the contribution of this
to the system would be small. The calculation was done using four
groups and seven regions and followed the method outlined in Appendix
A of this thesis. The following configuration, which is symmetric about
the indicated center line, was assumed:
54"
Water
24" -
Water
Under these water-moderated and reflected conditions, a k =
eff
0.79 was obtained
b. Solid angle criterion for slabs
To determine the interaction between the fuel storage slabs in the
proposed 3-slab array, the solid angle criterion established in 10 CFR,
Part 70, 70.52, paragraph (b) was used. This establishes the maximum


ACKNOWLEDGMENTS
The author wishes to express his sincere appreciation to his
graduate committee for their guidance. Special recognition is due
Dr. M. J. Ohanian, whose encouragement, dedication and detailed
scientific knowledge made this work possible.
The support of the Nuclear Engineering Sciences Department of
the University of Florida throughout the authors graduate work is
greatly appreciated. In particular Dr. R. E. Uhrigs support and
friendship is gratefully recognized.
The author feels fortunate and proud in having studied and worked
with a remarkable scientist and gentleman, Dr. R. B. Perez, now at
Oak Ridge National Laboratory. The author's years of association with
Dr. T. F. Parkinson, now at the University of Missouri, formed the
necessary background for this work.
The continuous assistance of Mr. G. W. Fogle throughout the ex
perimental program is sincerely appreciated. Mr. L. B. Myers designed
and built the control instrumentation. Mr. R. E. Schoessow was respon
sible for the design and construction of the assembly. Mr. E. Dugan
and Mr. H. Leydolt aided in the data processing. The cooperation of
the staff of the Nuclear Engineering Sciences Department during the
construction of the facility is acknowledged.
Most of this work was financed under subcontract No. C281 and
C635 with Atomic Energy Division of the Phillips Petroleum Company,
iii


118
long detector in one position and rotating the fuel elements that
surround it. The narrow core (6.5 in wide) is obviously very suscep
tible to this effect and to the exact centering of the neutron source
whose location could not be determined to better than 0.5 cm accuracy.
Dynamic Flux'Traverses
Dynamic flux traverses across the width of the core and reflector
were carried out at distances of 41, 54, and 130 cm from the source.
As mentioned above, these measurements are important since the theoret
ical model is one-dimensional and planar propagation is conceptually
desirable.
Shown in Table VI are the "peaking times" obtained for several
positions across the width of the core and side reflectors. The exper
iment was performed with both 0.5 and 1.0 msec input pulses. Only the
results for the 0.5 msec input pulse are shown. The results for the
1 msec case display the same behavior. All the peaks occurred well
within the experimental accuracy. Comparison of pulse shapes at dif
ferent positions across the core and reflector revealed no differences
in the basic shapes. Near the outer wall of the reflector a secondary
pulse, which peaked at a time corresponding to the input pulse width,
was observed. It was confirmed experimentally that this was due to
neutron reflections from the walls of the facility room.
From the pulse propagation measurements, reactivity data was also
obtained. These results are analyzed in Part 2, Chapter VI.
Clean Cor Pulse Propagation Measurements
The main portion of the research propagation measurements in the
clean, cold, side-reflected assembly. These most detailed experiments


193
channels is now reduced to an odd number of points terminating where
the background has been determined to start.
5. Determination of the ratio kf3/£ following the Garelis-Russel
method. Simpson's integration is used for the pulse and the Regula-
Falsi iterative method is used to find the root.
6. The data is normalized according to a selected normalization
scheme (see Part I, Chapter IV). The normalized data can be punched
in cards for further analysis.
7. If a neutron wave type analysis is desired, a numerical
Fourier transformation is performed and the amplitude and phase of the
pulse for each frequency selected is printed and punched for further
processing.
UNIPUL was used to process the pulse data punched in cards from
original perforated paper tape (at the IBM 1800 computer of the
Department of Nuclear Engineering Sciences, University of Florida);
the paper tape is the output of the multichannel analyzer used in the
experimentation.
MORE
The FORTRAN IV, IBM 360 computer program MORE performs a conven
tional numerical Fourier transformation of the pulse neutron data
(zeroth moment only) for the analysis of the experiment in the fre
quency domain. The transformation used equal time steps, a maximum
of 1023 time points and Simpson's integration scheme. An odd number
of points are required by the program. Thermal neutron data can be
entered directly as input or the total and epicadmium neutron flux
are required to obtain the thermal flux by subtraction.


FLUX (RELATIVE UNITS)
FIG* C23 TIME PROFILE OF THERMAL NEUTRON FLUX 41*15 CM FROM THE SOURCE
219


FLUX (RELATIVE UNITE)
FIG. 36 EXPERIMENTAL PULSE SHAPES AS A FUNCTION OF CORE HEIGHT


27
must be as specified under Operating Limits in this report.
g. After a normal start-up, the water height in the core must be
within 0.5" of the level set by the position of the weirs.
Fuel Storage
The large amount of fuel needed for the experimental program
required a detailed criticality analysis of the fuel storage area.
Criticality considerations of the fuel storage arrangements follow the
Atomic Energy Commission regulations regarding the subject. Three dif
ferent criteria were used to calculate the effective multiplication of
the fuel storage area to assure that the array will remain subcritical
under the worst circumstances. The methods and the corresponding con
ditions are outlined below.
The fuel storage array consists of three slabs of air-spaced fuel
pins, separated by a minimum distance of 54 inch face to center. The
fuel is stored in steel baskets containing 308 pins per basket. Two
sets of 1/4 inch thick plastic plates, located at the bottom and top of
each basket, drilled to properly position the fuel rods. The charac
teristics of the fuel slab are:
Slab Width = 3.73 in = 9.46 cm (corresponds to 7 fuel pins in
transverse direction)
Slab length = 14 ft
Height = 3 ft (active fuel height)
a. Multiregion-multigroup calculation
The effective multiplication factor of the fuel storage array con
sisting of three slabs (3.73 inch wide, 14 ft long and 3 ft wide) which
are 54 inch apart (face to face) was computed for the case of flooding


APPENDIX C
UFSA Rl-CLEAN CORE
TIME PROFILES OF THERMAL NEUTRON FLUX
AT NINETEEN CORE POSITIONS
FOR INPUT PULSES OF 0.5 AND 1.0 MSEC


195
the pulses. A CALCOMP plotter is used in connection with the 1800
computer for the actual plotting of the pulse. The program is capable
of normalizing to the peak counts of a reference pulse for the super
position of related time profiles.


98
aspects of pulse propagation phenomena were studied using a few or a
single detector location in the clean core; these measurements will be
discussed separately from the main part of the research as complementary
experiments. .
Pulse propagation measurements across the entire length of the
core were conducted with input pulse widths of 0.5 and 1.0 msec. The
time profile of the thermal neutron flux was obtained by the cadmium
subtraction method at nineteen locations in the core. The first
detector position was 3 cm from the neutron target and measurements
were taken every 12.72 cm thereafter. The data thus obtained was then
analyzed in the time domain.
With the exception of the measurements at the last four detector
locations, 2^ or 2^ counts were accumulated in the peak channel of
the analyzer and two measurements conducted at each position. The full
1024 available channels of the analyzer were used. The channel width
was 20 microsec with an intrinsic 10 microsec storage dead time giving
an effective time between channels of 30 microsec. Running times
varied from 3 minutes close to the source to two hours at the position
farthest from the source.
Propagation of a Narrow Pulse
To test the analytical model under more severe conditions, a
narrow, 100 microsec wide input pulse was introduced into the assembly
and the pulse profile recorded at a few detector locations. The exper
iment is identical to the one described above except that fewer detec
tor locations were employed.
Propagation of a Wide Pulse
To illustrate the propagation characteristics of a pulse violating


Flux (Relative Units)
132
FIG. 32 AMPLITUDE ATTENUATION OF THE THERMAL NEUTRON FLUX
- 1.0 MSEC INPUT PULSE -


FLUX (RELATIVE UNITS)
FID- C26 TIME PROFILE OF THERMAL NEUTRON FLUX 73-30 CM FROM THE SOURCE


FLUX (RELATIVE UNITS)
1-0
0*9
0-9
0-7
I
,*2 .
* '
\
THERM--5- 34 '
PEAK AT 0-950 MSE.'CS
PULSE WIDTH = 0-55 MSECS
THECRY FWHM = 1-3559 MSECS
....... EXP- FWHM = 1-470 MSECS
0-G
I
S
s
0-55
0-4
0*3
03
0-1
0 1
*
%
9
%
hO
O
{1
V
******
***.
-i-
-L.
*"****..*
....[
10
3
4 5
TIME (MSECD
9
FIG- C5
TIME PRDFILE GF THERMAL NEUTRGN FLUX
53 BS GM FRDM THE SOURCE


3
The Large Core Dynamics Experimental Program is to be performed
in three phases:
Phase I. Pulsed Source Experiments in Subcritical, Multiplying Media,
Large in One-Dimension
The first phase of the experimental program is to be conducted in
a close-to-critical subcritical assembly, 8 feet long, 3 feet deep and
with widths changing from 6.5 inches to 16 inches according to the core
configuration and whether bare or side reflected cores are studied.
The experimental information obtained from studying the pulse
propagation phenomena in this assembly is to be used to test the valid
ity of current space-time kinetic models in the absence of inherent
feedback effects.
Phase II. Kinetic Behavior for Control-Rod-Induced Power Excursions in
Large, One-Dimensional Cores
A reactor large in one-dimension, 16 feet long, three feet deep
and with varying widths to accommodate different metal/water ratios will
be used to investigate the one-dimensional kinetic behavior of large cores
subjected to a large perturbation. Both non-feedback (low power exper
iments) and self-shutdown measurements will be conducted.
Phase III. Kinetic Behavior for ContrOl-Rod-Induced Power Excursions
in Large, Two-Dimensional Cores
The same type of measurements performed for the one-dimensional
core will be conducted in a two-dimensional core.
The measurements should provide the necessary information to
establish the validity ranges for one-dimensional models, the basis for
the development of a two-dimensional scheme and as a bridge to the
complex, three-dimensional problem.
Phase II and Phase III of the research program will be performed


FLUX (RELATIVE UNITS)
FID* C30 time PROFILE OF THERMAL NEUTRON FLUX ISO*IB CM FROM THE SOURCE
226


TABLE II
UFSA INSTRUMENTATION AND CONTROL
A. NUCLEAR
Measured Parameter
Method of Measurement
Application
a. Low level neutron flux
He detector pulse discriminator;
at neutron generator end of core
Insure source is present before
adding reactivity. Scram on low
count rate
b. Linear neutron flux
c. Log neutron flux*5
CICa ammeter; on side of core
Indicate power level scram on
power
3.
CIC log N and period amp; under
core near center line
Indicate power level scram on
high power; log N recorder
d. Linear neutron flux
UIC ammeter; on side of core
Indicate reactor power; linear N
recorder
e. Reactor period lc
CIC log N and period amp
Indicate reactor period; scram on
short period
f. Gamma flux
g. Detector power supply
voltage
h. Reactor period 2
Ion chamber area monitor; on front
of reactor cage
Unijunction transistor oscillator
and relay. Monitor detector
voltage for b, c, d and e
CICa log N and period amp; under
core near center line
Criticality monitor for storage
room. Area monitor for reactor
room. Activate evacuation alarm
Scram reactor on low detector
voltage
Indicate reactor period; scram on
short period
N3
cn
*
a
b
To be operated in the uncompensated mode
Common detector and instrument


CHAPTER VI
SPATIAL DEPENDENCE OF PULSED-NEUTRON
REACTIVITY MEASUREMENTS
Introduction
A resume of pulsed-neutron reactivity measurements was presented
in Part 2, Chapter V of this work. During the nuclear calibration of
the UFSA assembly certain "spatial" effects were observed in the deter
mination of the reactivity of the long assembly. In this chapter these
effects are investigated.
The pulse propagation measurements performed in the UFSA R1 clean
core were analyzed using the Garelis-RusseUtechnique [8] to determine
the ratio kB/,; the "fundamental" decay constant was obtained by ap
plying Peierls statistical analysis to the tail of the pulse and the
effective delayed neutron fraction was calculated and assumed constant
with respect to reactivity. A strong "spatial" dependence of the
measurements conducted along the center line of the core was observed.
An asymptotic spatial distribution was obtained long times after the
pulse; at these times only the positions farthest from the neutron
source had not reached background level. The Garelis-Russell technique
was found to be very sensitive to the finite width of the input pulse.
A computer program, UNIPUL, was coded and used to perform the
Garelis-RusseU type analysis on the impulse response curves determined
as a function of position in the UFSA R1 core. The same program
174


12
FIG. 2A BOTTOM FUEL ROD SPACING SYSTEM


8
The assemblies are highly multiplicative; this is important for
the extrapolation of the results of the study to critical systems. The
system's subcriticality is attractive because of the inherent safety
of such systems and of the absence of inherent feedback effects.
In order to provide as "clean" a core as possible, a unique control
system which has been successfully used on the UFAPA [2] will be em
ployed. In this system the reactivity is controlled by adjustment of
the water height in the assembly. The water height is controlled by
the position of two "V"-notched weirs located in a water "box" hydrau
lically coupled with the assembly through flexible lines. The quantity
5/2
of water discharging through a "V"-notched weir varies as H (H is the
distance between the apex of the weir and the water level) thus pro
viding precise control of the moderator height. The hydraulic coupling
assures that under normal operating conditions (with continuous flow)
there will be the same water level in the core and the reflector tanks.
The UFSA subcritical assembly is located in an isolated and
shielded room in the Nuclear Research Field Building, approximately
three miles from the University campus.
The Nuclear Research Field Building consists of four bays, two of
them having shielded rooms for experiments with subcritical and moder
ating assemblies. The shielded walls consist of stacked concrete block
eight feet high and thirty-eight inches wide covered with plywood to
assure that the blocks remain in place. The ceiling of this single-floor
building is approximately 15 feet above the floor and consists of excel
sior-filled cement bonded board. Neutron reflection from this ceiling .
over the walls does not constitute a hazard to personnel operating the
accelerator-type neutron source. The access door from the control room


89
Neutron Wave Analysis
The neutron wave technique has proven to be a very powerful method
of determining the deficiences a model has in predicting propagation
phenomena or in establishing the "quality" of the results from a corre
sponding experiment.
The study of the space-time data in the frequency domain "looks"
at the overall behavior of the asymptotic spatial region of the assem
bly with parameters that include information on all spatial points for
each frequency of interest. The comparison of the predicted and meas
ured real and imaginary components of the complex inverse relaxation
length, 5, is therefore a more comprehensive evaluation of the discrep
ancies than the point-wise comparison in the time domain. The analysis
in the 6^ plane is an especially stringent test of the model and the
experimental data.
In this work, the comparisons in the frequency domain were made
by numerically determining the zeroth Fourier moment of the data meas
ured in the time domain. This method was suggested by Moore.^
Moore expressed Fourier moments of the space-time data as
. n -2-rrift .
i|>n(r, f) = / dt t e ij>(r, t)
(A)
where iKr, t) is the neutron pulse at space point r as a function of
time, t, and f is the frequency in cycles per second.
The fundamental space and energy mode propagating through the
medium is given by:
1. The author acknowledges Dr. M. N. Moore for several enlightening
discussions on the subject of neutron waves propagation.


A one-to-one comparison of the predicted and measured values in
the frequency domain was provided by performing identical numerical
Fourier transformations of the WIGLE time profiles and the measured
pulse shapes. The analysis in the frequency domain confirmed the
results obtained in the time domain, although discrepancies past 100
2
cps are found in the ultrasensitive p plane. The agreement in the p
plane, the system's dispersion law, is good up to 200 cps and reasonable
up to 800 cps. Both theory and experiment showed a smooth behavior
2
throughout the frequency range investigated, in both the p and the p
plane.
Spatial effects in large cores are clearly demonstrated in this
work. The determination of the range of applicability of the one
dimensional scheme requires extending the study to cases in which two-
dimensional effects will be noticeable and the important feedback
effects can be considered.
xxi


xml record header identifier oai:www.uflib.ufl.edu.ufdc:UF0009022600001datestamp 2009-02-23setSpec [UFDC_OAI_SET]metadata oai_dc:dc xmlns:oai_dc http:www.openarchives.orgOAI2.0oai_dc xmlns:dc http:purl.orgdcelements1.1 xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.openarchives.orgOAI2.0oai_dc.xsd dc:title Space-time reactor kinetics studies with the University of Florida, SPERT Assembly.dc:creator Diaz, Nils Juan,dc:type Bookdc:identifier http://www.uflib.ufl.edu/ufdc/?b=UF00090226&v=00001000955726 (alephbibnum)16973836 (oclc)dc:source University of Florida


LIST OF FIGURES (contd)
FIGURE Page
27B EXPERIMENTAL PULSE SHAPES AT DIFFERENT POSITIONS
IN THE UFSA R1 CORE 123
27C EXPERIMENTAL PULSE SHAPES AT DIFFERENT POSITIONS
IN THE UFSA R1 CORE 124
28A EXPERIMENTALLY DETERMINED SPATIAL DISTRIBUTION OF
THE NEUTRONS AT DIFFERENT TIMES AFTER THE PUSLE 125
28B EXPERIMENTALLY DETERMINED SPATIAL DISTRIBUTION OF
THE NEUTRONS AT DIFFERENT TIMES AFTER THE PULSE 126
29 CALCULATED AND EXPERIMENTAL DELAY TIMES
- 1.5 MSEC INPUT PULSE 129
30 CALCULATED AND EXPERIMENTAL DELAY TIMES
- 1.0 MSEC INPUT PULSE 130
31 AMPLITUDE ATTENUATION OF THE THERMAL FLUX
- 0.5 MSEC INPUT PULSE 131
32 AMPLITUDE ATTENUATION OF THE THERMAL FLUX
- 1.0 MSEC INPUT PULSE 132
33 DELAY TIMES OF THE THERMAL FLUX CALCULATED BY THE
WIGLE CALCULATIONAL SCHEME FOR DIFFERENT CORE HEIGHTS 139
34 AMPLITUDE ATTENUATION OF THE THERMAL FLUX CALCULATED
BY THE WIGLE CALCULATIONAL SCHEME FOR DIFFERENT
CORE HEIGHTS 140
35A THE SENSITIVITY OF THE ONE-DIMENSIONAL, TWO GROUP,
SPACE-TIME KINETICS SCHEME TO CHANGES IN THE
TRANSVERSE BUCKLING 141
35B THE SENSITIVITY OF THE ONE-DIMENSIONAL, TWO GROUP,
SPACE-TIME KINETICS SCHEME TO CHANGES IN THE
TRANSVERSE BUCKLING 142
xv


FLUX (RELATIVE UNITS)
TIME PROFILE OF THERMAL NEUTRON FLUX 130IS CM FROM THE SOURCE
FID- E3


flux (relative: units)
1-0 +
0-0
0-0
0-7-
0-G
0-5
0-4
0-3
V

UNSHIELDED
ft
*
ft
*
ft
ft
f
ft

ft

I \
/****'*?>> A A**
fit'* y?
\a^ f
* y
&
4
a-
§r
SHIELDED
r
%
*h
? rw
i n:
r
&
5.
2f.\
0-1
i
i
a
if /
f/
-UJ
*H\ft
/
VA
Vni1
n\
NORMALIZED TO
R1REFREF7G
pulse: width = 0-5 msecs
EXP- FWHM = 3-570 MSECS
V**4<

10
0
J.
4
G
7
B
3
TIME (MSEC)
FIG- 37A EFFECT OF ROOM RETURN AT PERIPHERAL DETECTOR POSITIONS
153


67
area-ratio method, is of practical value because of its intrinsic
elimination of the effect of prompt higher harmonics. This method,
which was used in the present work, was postulated originally for a
repetitively pulsed (with a delta function source in time), bare,
monoenergetic reactor but has proven to be of broader application.
Garelis and Russellpostulate, that for the conditions specified above:
1/R 1/R
.fNp exp( (k|3/£,) t)dt = £Npdt + Nj/R
where
Np = prompt contribution to the neutron density'
= delayed contribution to the neutron density
R = pulsing rate
The following conditions should be satisfied for the correct ap
plication of the method.
a) R>>X, where X is the decay constant of the shortest lived
precursor group.
b) R a, where a is the prompt fundamental decay constant.
c) The system must be pulsed a sufficient number of times so that
exp (magn/R) << exp (-agn/R) where m + 1 is the total number of pulses
and the agn are the roots of the inhour equation.
d) The prompt root dominates the decay.
The Garelis-RusseUtreatment permits the determination of p ($)
when all the above conditions are satisfied, by the relation:
($) =
k e/£
- 1
An absolute value of p is obtained by the use of a calculated effective
delayed neutron fraction. Garelis has discussed the use of the method
in reflected systems; the technique seems to be of practical value in


FLUX IRELATIVE UNITS)
FIG* 18 THE TOTALtEPICADMIUM AND THERMAL FLUX 117-44 CM FROM THE SOURCE


13
I
FIG. 2B BOTTOM FUEL ROD SPACING SYSTEM


7
TABLE I
keff vs. MODERATOR LEVEL OF UFSA REFLECTED CORES
Calculated Using the AIM6 Code
Core Length = 243.8 cm (96")
Reflector Width = 30.48 cm (12")
Active Fuel Height = 91.4 cm (36")
Metal to Water Ratio = 0.5
Lattice Pitch (in) = 0.7152
Core Width (cm) = 16.35
No. of Fuel Rods = 1206
1.0
0.584
19.28
2132
1.5
0.5332
25.73
3420
Moderator Level
(cm)
20
25
30
35
40
45
50
55
60
65
70
75
80
85
Effective Multiplication Constant
.7695
.7397
. 7276
.8288
.8028
.7941
.8715
.8471
.8416
.9022
.8793
.8761
.9249
.9032
.9020
.9420
.9213
.9218
.9551
.9353
.9372
.9655
.9465
.9494
.9738
.9554
.9592
.9805
.9626
.9672
.9870
.9686
.9738
.9906
.9736
.9793
.9944
.9777
.9840
.9977
.9813
.9880
1.0012
.9851
.9922
91.4


109
the (d,t) reaction rapidly becomes contaminated with the (d,d) reaction
due to the accumulation of the deuteron beam on target. After ~ 600
microamp-hr per unit area of accumulated beam on target the original
neutron yield from the (d,t) reaction has dropped to half of its
original value while (d,d) neutrons account for 1% of the neutron
beam. After ~ 2000 microamp-hr per unit area of accumulated bean on
target, the number of (d,d) neutrons has increased to 2%; this .repre
sents a detectable contamination, and may have some effect on the
removal cross-section of source neutrons. In the present case, how
ever, due to the coarse energy mesh used in the calculation scheme the
effect is not considered to be important.
Typical results obtained with the WIGLE code are shown in Fig. 23
(A, B, C) for a 0.5 msec input pulse. The general characteristics of
the pulse propagation phenomena are clearly displayed. The calculated
spatial distribution of the neutrons as a function of time is shown for
times of 1,2,3,5,7 and 9 msec in Fig. 24 (A, B). The spatial dispersion
increases with time while the pulse attenuates severely in time and
space.
The WIGLE results are discussed simultaneously with the results of
the clean core measurements.
Flux Traverses
Static and dynamic flux traverses were obtained in the clean core
to establish the following:
a) The asymptotic steady-state flux shapes in the transverse
direction.
b) The planar propagation of the neutron pulse. This is extremely


Flux (Relative Units)
126
UFSA R1 Clean Core
12.716 cm between points
Position Number
FIG. 28B EXPERIMENTALLY DETERMINED SPATIAL DISTRIBUTION OF
THE NEUTRONS AT DIFFERENT TIMES AFTER THE PULSE


CHAPTER IV
EXPERIMENTAL AND THEORETICAL RESULTS
IN THE TIME DOMAIN
The Analytical Model
The analytical results were obtained using an IBM 360 version of
the WIGLE code^ supplied by the Phillips Petroleum Company, and made
operative on the IBM 360-50 at the University of Florida. Certain modi
fications were made to WIGLE to obtain punched card output of the time
profile at selected spatial points. In this manner, the pulse shapes
at the detector positions corresponding to the experimental measurements
were readily available for direct comparison with experiment.
The WIGLE scheme, being a one-dimensional model, requires that the
transverse leakage be taken into account by the proper adjustment of
the parameters. The two-group nuclear parameters for WIGLE were obtained
by the procedure outlined in Appendix A. The eigenvalue, three-di
mensional flux shapes and the bucklings were obtained from the CORA [46]
computer program. The core width was set at 16.35 cm, the reflector
width at 30.5 cm and the reactor height at 76 cm. The experimental
measurements were made with 75 cm of water above the bottom of the ac
tive fuel but there was a small water reflector at the bottom; thus the
1. The author is indebted to Mrs. M. E. Radd, Nuclear Safety Research
Branch, Phillips Petroleum Company, for her valuable cooperation in
familiarizing the author in the use of the WIGLE code.
100


50
The Electronic Instrumentation
A block diagram of the instrumentation used in the pulse propaga
tion experiments is shown in Figs. 11 and 12. Two independent data
acquisition systems with a common start-stop clock are necessary to
carry out the measurements: (1) a system connected to a movable detec
tor that obtains the time profile of the propagating pulse at a given
position and (2) the all important normalizing detector, fixed at one
3
position. Two 1 atmosphere He detectors with the characteristics de
scribed above were used for these purposes. The system is essentially
composed of signal transmitting devices and data registering and
handling units.
The movable detector data acquisition system (MDDAS) consists of:
1. Detector High Voltage Power Supply
The high voltage power supply was an ultrastable FLUKE 405 B with
superior stability and negligible ripple. Manufacturer's specifications
state the stability at .005% per hour and the ripple at less than 1 mv
RMS.
2. ORTEC Model 260 Time-Pickoff unit, with 3000 volt isolation
The time-pickoff units are normally used to detect the time of
arrival of a detected particle, usually with subnanosecond precision..
The use of this characteristic and the electronic arrangement shown in
Figs. 11 and 12 permitted the counting of neutron events with excellent
time resolution. To our knowledge this is the first time that the time-
pickoff units have been used for this application.
Briefly, the system operates as follows: the primary of a toroidal
transformer, having a bandpass for very high frequencies only, is
inserted between the detector and the bias power supply. Fast


Delay Times (msec)
139
FIG. 33 DELAY TIMES OF THE THERMAL FLUX CALCULATED BY THE
WIGLE CALCULA!IONAL SCHEME FOR DIFFERENT CORE HEIGHTS


38
predicts that from a k^^ of 0.993 it would take approximately 15
seconds to double the power level. Even if the power level indicator
were not to scram the system until the power level reached 10 kw,
assuming a one second delay time to actuate the scram, the power would
increase to only 69 kw by the time the scram actually begins. Based on
the reactivity removal rate described above, the power level would then
decrease rapidly to a very low value.
Period Scram
Specification;
Period scram 15 sec
Basis; A positive period will be obtained in the subcritical assem
bly for any addition of reactivity beyond a given steady state condition.
If the maximum reactivity insertion rate of 0.05 $/sec is considered,
the initial positive period is about 50 sec and decreases monotonically
with time. The period channel has been determined to respond reliably
to periods = 50 sec. Originally, the period scram was set at 50 sec but
repeated scrams caused by the start-up of the neutron generator forced
the reduction of the scram period to 15 sec for operational reasons.
Average Neutron Flux and Neutron Flux Scram
Specification:
Ave neutron flux for
5 2
most reactive core 1.5 x 10 n/cm sec
5 2
Neutron flux scram 3.0 x 10 n/cm sec
Basis: the above are based on the average power calculated for the
assembly. Doubling of the flux will occur when 0.75 $ worth of reac
tivity is added to the system from any design subcriticality level.
Under these conditions, even at the maximum k^^ status, the facility


LIST OF FIGURES
FIGURE Page
1 OVERALL VIEW OF THE FACILITY .. 6
2A BOTTOM FUEL ROD SPACING SYSTEM 12
2B BOTTOM FUEL ROD SPACING SYSTEM 13
3 TOP FUEL ROD SPACING SYSTEM 14
4 REACTIVITY-CONTROL FLOW SYSTEM 16
5 FLOW RATE vs. HEIGHT OF WATER LEVEL ABOVE
WEIR APEX 18
6 WEIR "BOX 19
7 AIR SYSTEM SCHEMATIC 21
8 UFSA SAFETY SYSTEM LOGIC FLOW DIAGRAM 23
9 POWER vs. TIME FOR THE DESIGN BASIS ACCIDENT 44
10 PHYSICAL CHARACTERISTICS OF THE He3 NEUTRON COUNTERS 49
11 MOVABLE DETECTOR DATA ACQUISITION SYSTEM 51
12 NORMALIZING DETECTOR DATA ACQUISITION SYSTEM 52
13 TIME PROFILES OF NEUTRON BURST RECORDED BY
CONVENTIONAL ELECTRONIC INSTRUMENTATION AND BY THE
TIME-PICKOFF SYSTEM 58
14 THE PARALIZING, NON-PARALIZING SYSTEM RESOLUTION TIME
CORRECTION AS A FUNCTION OF COUNT RATE 60
15 DETECTOR POSITIONING SCHEME 69
16A INVERSE MULTIPLICATION vs. MODERATOR LEVEL 71
16B INVERSE MULTIPLICATION vs. SQUARED INVERSE HEIGHT 72
xiii


.WEIR DRAIN
DISTRIBUTION MANIFOLD
PUMP
FIG. 4 REACTIVITY-CONTROL FLOW SYSTEM
CONTROL
VALVE
-0
ORIFICE


62
The Analyzer Method is intrinsically more accurate than the
Integral Count Method because it permits an "exact" correction for
resolution time, losses by the use of the expression previously given
applied to the recorded time profile. The correction for the integral
counts is, on the other hand, inaccurate since the count rate is con
tinuously changing and no base exists for a resolution time correction.
It was found, however, that as long as the count rate near the
peak of the pulse is kept well within the capabilities of the NDDAS no
significant difference is observed in the results of the two methods.
This is due to the fact that ratios are being taken in both cases; this
tends to minimize whatever differences there might be. Thus, it is
concluded that with proper care the Integral Method is adequate whenever
an analyzer is not available for normalization purposes.
It should be noted that an "effective" resolution time can be used
to improve the results of the Integral Method. This "effective" reso
lution time can be found by forcing the normalizing ratio obtained from
two runs by the Integral Method to match the normalizing ratio obtained
from the Analyzer Method for the same two runs by adjusting the reso
lution time correction applied to the integral counts.
Comments
Certain inconsistencies in the results of the first few experi
ments prompted a careful inspection of the multichannel analyzer modus
operandi. For the sake of completeness the significant findings are
listed below.
A) Operation with the 10 microsec channel width (selected by the
settings of the 212 plug-in-unit) proves to be unreliable due to insta
bilities in the clock and gating circuits. Channel widths of 20


157
Method of Analysis
The following method of analysis was applied to the calculated
and measured results of the UFSA R1 clean core space-time studies:
1. A conventional zeroth moment Fourier transformation was
performed on the experimental data. The numerical integration employed
Simpson's rule of integration. Equal time steps (30 microsec) were
used throughout this analysis. The number of data points varied from
500 to 800, depending on where a stable neutron background was reached,
as determined by an statistical analysis performed in the UNIPUL pro
gram. A computer code, MORE, was coded for this purpose.^"
2. A conventional zeroth moment Fourier transformation was
performed on the space-time results obtained from the WIGLE calcula
tions. It should be recalled that the WIGLE calculation employs a
series of time increments in different time intervals. To conform to
this scheme Simpson's rule of integration with different time incre
ments was used for each time interval. Trapezoidal integration was
used to bridge the gap between the different time intervals. A computer
program, MORWIG, was coded for this purpose.
3. The amplitudes and phases of the zeroth Fourier moment as a
function of frequency, obtained from the MORE and MORWIG numerical
transformation, were least-squares fitted to obtain the damping coef
ficient and the phase shift per unit length for the frequencies of
interest.
In this manner, the theoretical and experimental results in the
1. The MORE and MORWIG programs were coded by Dr. M. J. Ohanian,
University of Florida.


83
phenomena was studied.
The space-time data was also analyzed in the frequency domain by
numerically obtaining the amplitude and phase of the zeroth Fourier
moment of the pulse. The experimentally determined dispersion law of
the system is compared to that predicted by the WIGLE scheme. The
results of the WIGLE calculations performed for the analysis in the
time domain were Fourier analyzed so that a one-to-one comparison could
be performed.
Space dependent effects on reactivity measurements were studied
by analyzing the pulsed data using the Garelis-Russellmodel.
Nomenclature- Used in the Description of Pulse
Propagation Phenomena
A series of directly measurable or inferred parameters are used
to describe the pulse propagation phenomena, in the time and in the
frequency domain. The most important of these parameters are defined
below. The pulse shapes and the conventional full-width-at-half-
maximum also form part of the analysis.
Delay Time
The delay time, t^, is defined as the time displacement of the
peak of the pulse from a reference position to the position under con
sideration. In this work all delay times are referred to zero time;
this corresponds physically to the initiation of the pulse at the
neutron generator.
Propagation Time
The propagation time of a narrow pulse in the assembly is defined
as the delay time between two extreme positions in the assembly.


68
these systems [15].
Becker and Quisenberry were able to compute a correction [16] for
the observed spatial dependence of the reactivity in two-region systems
by recognizing the differences in the spatial distributions of prompt
andjdelayed neutrons. Their excellent comparative study of the above
techniques emphasized the need for their recommended spatial correction
unless the neutron detector is properly positioned to minimize this
correction.
The study of Garlid and Bierman [17] correctly points out that in
very large systems "an asymptotic spatial distribution cannot be estab
lished before the pulse has decayed away, since the asymptotic mode is
one that is uniform everywhere in space." They proceed to apply a
combination of first flight, age, and time-dependent diffusion theory
to the study of pulsed measurements in large aqueous media; their con
clusion is that their measured apparent decay constant is a good ap
proximation to the asymptotic value and that pulsed measurements in very
large multiplying systems may also give good results.
Inverse Multiplication Measurements
The safe approach to the design value of k^^ <^99 was undertaken
with the conventional 1/M measurements until k ~0.95 and then by both
eff
the 1/M and the pulsing technique.
To establish detector positions free from geometrical effects
(source-detector-water level), six different locations were used until
a water level of 60 cm (k ~ .98) was reached and four locations afterwards
Two of the detector locations, the closest to the neutron source, failed
to describe the multiplication of the system. Shown in Fig. 15 are the


FLUX (RELATIVE UNITS)
FIG. 23B PULSE SHAPES PREDICTED BY WIGLE AT DIFFERENT POSITIONS IN THE UFSA R1 CORE
111


181
Using a calculated value of = 0.0079 for this assembly configura
tion, the absolute value of the reactivity p, is obtained. The cor
responding value of k^, is then found by the relation
k =
eff 1 + p
Shown in Fig. 50 are the measured absolute reactivity and ke££ as a
function of position in the lattice. It is obvious from what has been
said above regarding the behavior of a and kg/£, that a meaningful
value for either k ,, or p cannot be obtained in this case until 143
cm separate the source and detector. The assigned value of k^^ was
0.990+.003. This value compares very favorably with a value of
0.990+.0025 found by inverse multiplication measurements and a calcu
lated value of 0.9906. Although the agreement obtained could be some
what fortuitous, it can be said that a representative value of the
reactivity was found. It seems that, if proper care is taken, the
pulsed reactivity measurements will yield good results in large multi
plying media.
Region-wise Dependence of the Reactivity Measurements
Becker and Quisenberry, as well as Waltar and Ruby [49] have
pointed out the spatial dependence of pulsed source reactivity measure
ments in two-media systems. Differences in the measured reactivity can
be present whether the detector is located in the core or in the reflec
tor.
A series of pulsed source traverses were conducted across the UFSA
R1 clean core and reflectors, at three different distances from the
source. No differences were found for this core in the reactivity
measured in the core or in the reflector. Shown in Table XIX is a


102
16.35
i r
i i
i i
K?.5-h
i i
-^pss
in
co
r\
-1
vJ
m
m
co
i
\o
FIG. 20 PLAN AND FRONT VIEW OF THE CORE REGION ENCLOSING THE
NEUTRON SOURCE


194
MORE punches in cards the amplitude and phase of the zeroth
Fourier moment of the pulse for input to the ALXILS program. A com
plete description of the program and input requirements can be found
in Ref. 55.
MORWIG
The MORWIG program is a version of the MORE code written to
perform a numerical Fourier transformation on a pulse with unequal
time steps. It is intended specifically for the neutron wave type
analysis of the space-time data calculated by the WIGLE scheme. See
Ref. 56 for a complete description.
ALXILS
The FORTRAN IV, IBM 360 program ALXILS performs a least-squares
fit of the amplitudes and phases obtained from the Fourier transforma
tion performed by either the MORE of the MORWIG program. A linear
least-squares scheme is- used. The values of ALPHA are obtained by fit
ting the log amplitude vs. distance for each frequency. The values of
XI are obtained by a linear fit of the phase angle vs. distance. The
channel chopping technique is used to determine the convergence of a
and £ in space. The weighting of the points can be set to (1/observed
value) or to equal weights. Statistical quantities determining the
"goodness of the fit" are calculated for each fit. A complete descrip
tion can be found in Reg. 57.
MULTIPLOT
The FORTRAN IV, IBM 1800 computer program MULTIPLOT used the out
put of the UNIPUL and/or the WIGLE codes to plot the time profiles of


FLUX (RELATIVE UNITS)
FIG- DIO TIME PROFILE OF FAST NEUTRON FLUX EG-58 CM FROM THE SOURCE


57
Shown in Fig. 13 are two time pi-ofiles obtained at the same posi
tion in the UFSA core for the same count rate. The conventional
preamplifier-linear amplifier system shown above was used to obtain one
of the profiles and the instrumentation selected for this work was used
to obtain the other. Both were corrected for resolution time losses
with the best available value for the resolution time. The distortion
observed in the pulse shape obtained with the conventional preamplifier-
linear amplifier system is non-linear due to the count-rate dependent
resolution time. It should be noted that the count rate near the peak
of the pulse was less than 10^ count sec. The conventional system
response "flattens out" when it reaches complete saturation in this
case above 10^ count sec. Saturation effects are not observed in the
pulse transformer system until the count rate exceeds 3.3 x 10^ count
sec. An erroneous resolution time correction, or one which used a
resolution time that does not characterize the system throughout the
range of count rates will change the shape of the neutron pulse and
consequently affect any analysis done on the pulse shape obtained.
As pointed out by Bierman, Garlid and Clark [11], in a pulsing
experiment it is necessary to determine whether the counting system
has the characteristics of a purely non-paralyzing system or those
of a mixture of paralyzing and non-paralyzing systems. Using the
method described by Bierman and co-workers, it was established that
the data acquisition system being used in the present work is, as
close as can be determined, non-paralyzing. The resolving time of the
system is essentially determined by the width of the input pulse to the
analyzer and the MCA characteristics.
It should be noted that for a wide range of count rates, depending


176
Shown in. Fig. 48 are the experimentally determined decay constants
as a function of position in the assembly, for a 0.5 and 1.0 msec input
pulse. As distance to the source increased, more "waiting" time was
available to extract the decay constant. The values found past P83 are
believed to represent the fundamental decay constant of the assembly.
It should be noted that the instantaneous decay constant at dif
ferent times after the pulse was calculated for all the detector posi
tions in the assembly, using the WIGLE code. The results agree closely
with those measured experimentally at different times after the peak.
The two-group, space-time dependent model describes well the observed
phenomena.
The Ratio k8/5,
The Garelis-RusseUtechnique was applied to the two-medium system
referred to as the UFSA R1 core. The measured kg/5, values, determined
as a function of distance from the source, showed a "spatial" depend
ence. This was somewhat surprising since the detector was supposedly
located to minimize the Becker-Quisenberry correction [16] and a more
uniform value of kg/5, was expected. Although the model lacks a good
energy representation of the system being studied such a strong spatial
variation in kg/5, was not expected.
Shown in Fig. 49 are the measured kg/5, values for input pulses of
0.5 and 1.0 msec. A prominent feature is the noticeable difference
between the results obtained with the two different, finite width
input pulses. Attention was focused on the sensitivity of the Garelis-
RusselLmodel to the finite input pulse widths use in the experiments.
As mentioned previously, the model utilizes a delta function input


74
TABLE III
SUMMARY OF 1/M AND PULSED MEASUREMENTS
UFSA R1 Core
0.5 M/W Ratio
16.35 cm wide reflected core
INVERSE MULTIPLICATION PULSED EXP PREDICTED
Moderator3 keff k0ffb keff
Height (cm) Pos. 1 Pos. 3 Pos. 5 Pos. 6
20
.415
.481
.425
.581
.7675
25
.744
.629
.705
.771
.8287
30
.867 .
.737
.824
.865
.871
35
.914
.794
.881
.900
.9022
40
.944
.84
.908
.927
.925
45
.9612
.873
.934
.946
.9419
50
.973
.895
.950
.961
.948+.01
.955
55
.98
.918
.962
.969
.965+.007
.9655
60
.985
.936
.973
.971
.972+.006
.974
65
.9891
.95
.980
.979
.9796+.005
.980
70
.9925
.96
.9853
.986
.9855+.004
.986
75
.9949
.971
.989
.990
.990+.003
.9906
80
.9944
85
.9976
91.4
1.00117
Above
bottom of ,
active fuel
Averaged from 3 <
detector positions


135
The Full Width at Half-Maximm (FWHM)
The FWHM is defined as the time-width of the pulse at half the
peak amplitude. The FWHM is one of the parameters conventionally used
to compare pulse shapes and is therefore included in this study.
An analysis of the experimental and calculated values of the FWHM
tabulated in Tables VII and VIII showed the following.
a) Close to the source, theory predicts slightly larger FWHM
values than experiment, the difference decreasing with distance from
the source. As discussed above the discrepancy near the source is due
to difficulties in adequately describing the void region behind the
source.
b) Between P27 and P34 the FWHM are practically identical.
c) Past P34 the experimentally determined FWHM becomes increasing
ly larger than the WIGLE prediction. The maximum difference is about
350 microsec which corresponds to ~ 8% of the experimental value.
The agreement is still considered to be very good.
The Flux Shapes
A detailed comparison of the theoretical and experimental flux
shapes obtained for the UFSA assembly was also made.
Plots of the time profiles of the thermal flux at each detector
position are included in Appendix C for the 0.5 and the 1.0 msec wide
source pulses. The peaks are normalized to unity.
The following information is supplied in each figure:
a) Identifying Run Number
b) Location of the Experimental Peak in Time
This parameter is determined directly by a computer routine which is
part of the plotter program. The routine simply finds the time channel


15
Moderator Flow Control System
The moderator flow control system of the UFSA can be better des
cribed by'the water flow schematic shown in Fig. 4. Besides the normal
fill and drain functions for the moderator, it serves as an accurate
reactivity control using adjustable moderator height by continuous flow.
The characteristic components of this system are described below,
A. Storage Tank: A 6 x 8 x 2 foot steel tank located directly
below the assembly will serve as the reservoir for the circulating
light-water moderator and as a dump tank. Normal water heights while
operating will be between 6 and 12 inches. The tank also serves as a
footing for the assembly supports. This arrangement makes a very con
venient and compact facility.
B. Core and Reflector Tanks: As seen from the flow diagram, water
is pumped from the reservoir to the core and reflector tanks through a
manifold at one end of the assembly and flows from the other end of the
core and reflectors tanks to the weir "box". From the weir "box", water
flows over the weirs back to the storage tank through a flexible line.
The core section is equipped with two normally open solenoid
activated dump valves, 3 inches in diameter, located at each end of the
core. These valves provide the reactor with a fast shutdown safety
system. The reflector tanks have their own 1 1/2 inch normally open
solenoid valves actuated by the same safety system.
Since the quantity of water discharging through a V-notched weir
5/2
varies as H where H is the height of the water level above the apex
of the V-notch, the water level and hence, the reactivity, can be con
trolled in a precise manner simply by varying the height of the weirs
and the rate of flow of water into the tank. This is accomplished by


TABLE VII
CLEAN CORE PULSE PROPAGATION STUDIES
EXPERIMENTAL AND THEORETICAL RESULTS
- 0.5 MSEC INPUT PULSE WIDTH -
UFSA R1 Core
0.5 M/W Ration
16.35 cm wide reflected core
POSITION
DISTANCE TO
PEAK
AT (msec)
FWHM
(msec)
PEAK
COUNTS3
NO
SOURCE (cm)
Theory
Exp
Theory
Exp
Theory
Expb
6
3
0.500
0.50+.01
0.542
0.540+.03
81.22
94.8+6
13
15.71
0.501
050+.01
0.615
0.540+.03
72.68
81.9+5
20
28.43
0.504
0.52+.01
0.779
0.779+.03
48.13
49.9+.8
27
41.15
0.635
0.63+.01
1.098
1.080+.03
27.35
32.7+1.6
34
53.86
0.840
0.82+.03
1.398
1.470+.06
16.37
16.5+.4
41
66.58
1.060
1.04+.03
1.766
1.800+.06
10.27
10.6+.5
48
79.30
1.300
1.28+.03
2.055
2.13+.06
6.65
6.4+.3
55
92.01
1.560
1.50+.04
2.355
2.52+.09
4.42
4.39+.11
62
104.73
1.840
1.82+.04
2.655
2.82+.09
3.00
2.97+.10
69
117.44
2.125
2.11+.04
2.925
3.11+.09
2.055
2.03+.06
76
130.16
2.400
2.40+.05
3.195
3.42+.09
1.427
1.53+.07
83
142.88
2.700
2.65+.06
3.465
3.72+.09
1.000
1.00+.005
90
155.60
3.000
2.96+.06
3.735
3.90+.12
0.7056
0.7062+.03
97
168.31
3.300
3.29+.06
3.900
4.17+.12
0.5000
0.5065+.02
104
181.03
3.570
3.50+.06
4.050
4.38+.12
0.3534
0.362+.007
111
193.74
3.810
3.85+.08
4.125
4.44+.12
0.2460
0.2512+.005
118
206.46
4.000
4.00+.08
4.200
4.47+.12
0.1630
0.1764+.004
125
219.17
4.150
4.18+.08
4.200
4.47+.15
0.0957
0.1161+.0035
132
231.88
4.210
4.21+.09
4.200
4.53+.15
0.0367
0.0437+.003
a Normalized to Position 83
b Errors assigned from the deviation from the mean of two measurements except for the last 5 space
points. This error is usually larger than the counting statistics error. The error of the last 5
points was calculated from counting statistics.
127


149
characteristic propagation time (^3.7 msec) across the assembly of a
narrow (0.5 msec) pulse. It should be recalled that the Johnson
criterion [47]postulates that to be able to observe spatially dependent
effects, the HWHM of the input pulse should be smaller than the charac
teristic propagation time across the assembly. For example, the experi
ments of Miley and co-workers at the University of Illinois [20, 21, 22]
showed that for an initially asymptotic very wide input pulse no shape
changes were observed as the pulse "propagated" across the assembly.
The pulse shapes resulting from these measurements at three posi
tions in the assembly are shown in Appendix F. The delay times, the
FWHM and area under the pulses, normalized to unity at the first posi
tion, are given below in Table XIV.
TABLE XIV
DELAY TIMES AND FWHM FOR A WIDE INPUT PULSE
DISTANCE TO
SOURCE (cm)
PEAK AT
(msec)
FWHM
(msec)
AREA UNDER
THE CURVE
41.2
10.07
10.1
1.0
79.3
10.17
10.7
1.15
142.9
10.22
10.5
0.92
From these results it seems that, although the delay times between
the recorded peaks are quite small a certain "rearrangement" of the
initially non-asymptotic pulse takes place. Close to the source the
pulse is practically square, at a distance of 79 cm it has widened some
what and at large distances from the source the pulse seems to be trying
to achieve the normal "dumbbell" shape. Johnson's criterion appears to


73
the multiplication.
Shown in Table III are the values estimated for k for the four
eff
locations that seemed to represent the system best. Position 3 gives
a lower limit and Position 1 an upper bound. No attempt was made to
establish the error associated with measured k but it is believed
eff
that the 0.99 value obtained at the last water level is within +.005
of the true value.
Absolute Determination of kef£
After an estimated value of k >.95 was obtained from the 1/M
eft
measurements, an independent determination of k was required by the
operating license at every new increment in the moderator level (as
determined by the criteria established in Part 1, Chapter 3). The
technique chosen for this determination was the Garelis-Russell method
of measuring k3/& and a simultaneous determination of the prompt funda
mental decay constant.
Different detector locations were used to determine the influence
of the source and of higher order harmonics contamination. Strong
"spatial effects" were observed in both a and k3/£. This phenomena will
be discussed in detail in Part 2, Chapter V because of its importance.
A seemingly true fundamental decay constant and "spaced-converged"
kS/i were obtained at large distances from the source and were used to
determine the reactivity of the system.
A Fortran IV, IBM 360 computer program named UNIPUL was coded to
perform a unified analysis of the pulsed neutron data (see Appendix B).
The program calculates the decay constant using Peierl's statistical
analysis [18] and k8/£ using the Garelis-Russellapproach after the data


56
scrutiny.
The resolution time of several, submicrosecond preamplifier-linear
amplifier combinations was shown to be a function of the mode of opera
tion and of the count rate. For this type of instrumentation, there is
a significant difference between the resolution time as determined by a
steady-state technique (two source method) and a dynamic method (maximum
count rate method [11]), the latter method yielding a much larger reso
lution time.
For the instrumentation finally chosen to carry out the pulsed
experiments in this work no significant difference was found between
the results of the two methods. Furthermore the pulse transformer-
amplifier combination behaves like an ideal non-paralyzing system. The
resolution time of the overall data acquisition system was found to be
primarily determined by the multichannel analyzer.
As a matter of illustration, when the system depicted below was
used the resolution time changed from -0.24 microsec as determined
under low count rate steady-state conditions to 8 microsec as deter
mined under high count rate, pulsing mode conditions.
ORTEC 410
Linear Amo
ORTEC 429
Scaler


4
at the SPERT IV facility at the National Reactor Testing Site, Idaho.
The appropriate existing experimental equipment, as well as the
extensive kinetics studies conducted by the Nuclear Engineering Sciences
Department at the University of Florida, was conducive to the granting
of a subcontract by the Phillips Petroleum Company so that the basic,
linear kinetics studies of Phase I could be performed at the University
of Florida.
The research to be performed as Phase I of the Large Core Dynamics
Experimental Program can be succinctly defined as the experimental and
analytical determination of the dynamic behavior of the neutron flux in
slightly subcritical water moderated assemblies of SPERT F-l fuel rods.
The facility in which the required measurements for Phase I, Large
Core Dynamics Experimental Program, are to be conducted necessitated a
thorough design and safety analysis. The assemblies are to be close to
critical and the core has a large U235 inventory. The nuclear capabil
ities of such systems were the object of a detailed study to determine
their operational characteristics under normal and accident conditions.
The flexible mechanical design, the safety instrumentation, the
nuclear evaluation, as well as the experimental calibration of the
first configuration under study constitutes Part 1 of this dissertation.
The reactor kinetics studies performed in the first of the con
figurations to be studied are dealt with in Part 2 of this manuscript.


Page 2 of 2
Internet Distribution Consent Agreement
In reference to the following dissertation:
AUTHOR: Diaz,Nils
TITLE: Space-time reactor kinetics studies with the University of Florida SPERT
Assembly, (record number: 955726)
PUBLICATION DATE: 1969
i. K\vW A. as copyright holder for the aforementioned
dissertation, hereby grant specific and limited archive and distribution rights to the Board of Trustees of
the University of Florida and its agents. I authorize the University of Florida to digitize and distribute
the dissertation described above for nonprofit, educational purposes via the Internet or successive
technologies.
This is a non-exclusive grant of permissions for specific off-line and on-line uses for an indefinite term.
Off-line uses shall be limited to those specifically allowed by "Fair Use" as prescribed by the terms of
United States copyright legislation (cf, Title 17, U.S. Code) as well as to the maintenance and
preservation of a digital archive copy. Digitization allows the University of Florida to generate image-
and text-based versions as appropriate and to provide and enhance access using search software.
This grant of permissions prohibits use of the digitized versions for commercial use or profit.
personal information blurred
Date of Signature
Please print, sign and return to:
Cathleen Martyniak
UF Dissertation Project
Preservation Department
University of Florida Libraries
P.O. Box 117007
Gainesville, FL 32611-7007
5/28/2008


Frequency
FIG. 46 COMPARISON OF THE THEORETICALLY PREDICTED AND THE MEASURED 2 a K
172


FLUX (RELATIVE UNITS)
FIG- D3
TIME PROFILE OF FAST NEUTRON FLUX 53-BG EM FROM THE SOURCE
238


37
Drain rate, including the system
reaction time 2.65 cm/sec
Rate of removal of reactivity 0.00037 Ak/cm
0.00098 Ak/sec
0.140 $/sec
it should be noted that the value of Ak/cm used to specify the reac
tivity removal rate is ~ 20 times smaller than the corresponding value
specified for the reactivity insertion rate.
Basis: the rate of drainage from the assembly established the
reactivity removal rate. Consistent with the approach taken when
specifying the reactivity, the time to drain 45 cm of water from the
assembly was measured to establish a lower limit on the rate of de
crease of water height. Even under these extreme assumptions, i.e.,
using the maximum slope of the k^^ vs. water height curve for the
insertion rate, the small calculated slope around k^^.99 for the
removal rate and the reduced pressure head, the reactivity removal rate
is three times larger than the insertion rate.
Subcritical Assembly Power Level and Power Level Scram
Specification:
Power Level 0.5 watt
Power Level Scram 1.0 watt
Basis: with the maximum source strength available and with k^^ =
0.985 the maximum average power was calculated to be .130 watts by
taking into account the spatial dependence of the flux and a steady
source at one end.
Assuming that a reactivity accident occurs, based on the maximum
reactivity addition rate specified above, the design basis accident


34
loading, and will be followed for subsequent cores:
1) After the fuel has been loaded, prior to each new incremental
change in the water level, the water is drained completely and the weir
(and water level scram) adjusted to prevent a level increase beyond the
desired value.
2) The first three measurements of the inverse multiplication are
obtained at water heights of approximately 20 cm, 25 cm, and 30 cm above
the bottom of the active fuel. As shown in Table I, the maximum k
eff
calculated for a water height of 30 cm is 0.87. Subsequent filling
increments are not to exceed the least of the following:
a. An increase in water height of 10 cm.
b. An increase in water height which, by extrapolation of the
inverse multiplication curve, would increase the kg^^ by one-half of the
amount required to make the assembly delayed critical.
c. An increase in water height which would, by extrapolation of
the inverse multiplication curve, result in a kg_^ of 0.990. For values
of k^^ above 0.95, the k^^ of the assembly are also determined by
pulsed source techniques.
3) At each filling step, the measured kej£ of the assembly is
compared with the calculated value. If significant deviations of the
experimental values from the calculated kg^^ vs. water height curves
occur, the experiments are to be discontinued, and a detailed analysis
of the results obtained performed. If it is determined that the dimen
sions of the assembly should be changed in order to achieve the desired
ke^ at maximum water height, the University of Florida will apply for
and obtain written approval from the Atomic.Energy Division of the
Phillips Petroleum Company before such changes are made.


94
however for the fast group. Although the results for the epicadmium
flux obtained in these measurements will be presented, they should be
looked on with the reservation that the energy-dependent neutron
counting statistics invalidate the one-to-one comparison with calcu-
lational results. The epicadmium flux was obtained primarily for the
necessary correction of the time profiles. The analysis of the thermal
neutron group is duly emphasized.
Shown in Fig. 18 is a typical plot of the total and epicadmium
time profiles recorded in the experiment and the thermal pulse shape
obtained by subtraction. The magnitude of the correction is space-
dependent, since no constant thermal/fast ratio (as detected by the
counter) was observed. The correction varied from 3-10% of the total
flux values.
The Geometrical Arrangement
Shown in Fig. 19 is a plan view of the source-subcritical assembly
arrangement used throughout this work. The detector positions used for
the measurements are clearly indicated. The numbers shown correspond
to the positions of fuel pins with respect to the end of the assembly
facing the neutron generator. There were 7 grid-holes between detector
positions, giving a distance of 12.72 cm between data points. Through
out the rest of this report the positions of the data points will be
referred to as PXX, where XX is the grid-hole number at which the data
was recorded, i.e., P41 refers to position number 41 starting from the
front face of the assembly. A total of 134 rows of fuel pins were in
the assembly. Data was taken at nineteen spatial points for the main
part of the research. The target assembly of the neutron generator


FLUX (RELATIVE UNITS)
FIG. 35C THE SENSITIVITY OF THE ONE-DIMENSICNAL, TWO GROUP, SPACE-TIME KINETICS SCHEME
TO CHANGES IN THE TRANSVERSE BUCKLING
143


FLUX (RELATIVE UNITS)
FIG- D6
TIME PROFILE OF FAST NEUTRON FLUX 133-7-4 CM FROM THE SOURCE


FLUX (RELATIVE UNITS)
FIG* G2 SHAPE DF THE PROPAGATING PULSE AS A FUNCTION OF PULSE WIDTH
- EXPERIMENTAL -
257


FLUX (RELATIVE UNITS)
FIG C33 TIME PROFILE OF THERMAL NEUTRON FLUX IBB*31 CM FROM THE SOURCE


Chapter v
EXPERIMENTAL AND THEORETICAL
RESULTS IN THE FREQUENCY DOMAIN
Introduction
The results of the pulse propagation measurements and of the space
time-dependent WIGLE calculations were analyzed in the sensitive Fourier
transform plane. This neutron wave type analysis probes into the defi-
ciences of the model and/or the experimental data that the analysis in
the time domain might fail to reveal. In this study the analysis is
performed with the sole intention of testing the model and not to ex
tract parameters.
The WIGLE predicted time profiles of the propagating disturbance as
a function of space and the pulse propagation data were Fourier analyzed
in an identical manner; the amplitude and phase angle of the zeroth
Fourier moment thus obtained were least-squares fitted (log amplitude vs.
z and phase angle vs. z for each frequency of interest) yielding the
damping coefficient, a, and the phase shift per unit length, £. The
predicted and measured dispersive characteristics of the system are then
2
compared in the p and p plane.
The neutron wave analysis was applied to both the 0.5 and 1.0 msec
input pulse space-time measurements and calculations.
156


TO
ZENA


FLUX (RELATIVE UNITS)
NORMALIZED TD
THERM-5- 20
PULSE WIDTH = 0-5 MSECS
TIME CM;
E)
FIG. 27A EXPERIMENTAL PULSE SHAPES AT DIFFERENT POSITIONS IN THE UFSA R1 CORE
122


FLUX (RELATIVE UNITS)
FIG- C35 TIME PROFILE OF THERMAL NEUTRON FLUX 193-74 CM FROM THE SOURCE
231


FLUX (RELATIVE UNITS)
FIG* C27 TIME PROFILE OF THERMAL NEUTRON FLUX 32-01 CM FROM THE SOURCE
223


183
TABLE XIX
REGION-WISE DEPENDENCE OF THE REACTIVITY MEASUREMENTS
Measurement across the core width
UFSA R1 Clean Core
0.5 M/W Ratio
16.35 cm wide reflected core
POSITION
NO
DISTANCE TO
CORE CENTER
(cm)
a
, -Is
vsec )
k B/i
(sec
P
k
eff
A
7.27
508
205
.01169
.9884
B
5.45
504
2C7
.01132
.9888
C
3.63
497
208
.01102
.9891
D
1.82
505
207
.01137
.9887
E
0.0
517
203
.01227
.9878
F
3.63
504
209
.01119
.9889
G
5.45
510
206
.01168
.9884
H
7.27
503
207
.01124
.9889
I
10.9
512
203
.01205
.9881
RE2
14.3
501
205
.01147
.9887
RE4
17.9
490
199
.01154
.9886
RE6
31.8
474
187
.01214
.9886


29
total solid angle subtended by any unit in the array to be 6 steradians
if the effective multiplication factor for the individual slabs is less
than 0.3, as is the case here.
From 10.CFR 70.52 (d) (2) (i) the minimum required separation dis
tance, i.e., center of one slab to face of adjacent slab, is obtained
from:
3 steradians =
cross sectional area
2
(separation distance)
14 x 3
This gives, separation distance = ( ^) = 3.74 ft.
We propose to establish a minimum distance between faces of adja
cent slabs of 4.5 ft. This gives a total solid angle for the center
slab of 3.88 steradian, which is well below the established criteria,
c. Comparison with Clarks criteria
For 5% enriched, 0.4 inch diameter uranium oxide rods with a 190.13
gm/liter U-235 concentration (compared to 4.81% enriched, 0.42 inch
diameter uranium oxide rods with a 200 gm/liter U-235 concentration in
our case) the following data is obtained from pp. 39 and 59, respectively,
of DP-1014:
Width of Slab (cm)
Critical Safe
11.2 10.4
_2
Buckling (cm )
Critical Safe
0.014945 0.015912
It should be noted that the slab widths quoted on page 39 of DP-1014
are for an infinite water-moderated and reflected slab. From the buck
ling values given, the critical width of the infinite water-moderated
unreflected slab is 25.7 cm; the corresponding safe width is 24.9 cm.
When twice the reflector savings for the latter case as given on page
59 of DP-1014 is subtracted, a safe width for the infinite, water-


52
FIG. 12 NORMALIZING DETECTOR DATA ACQUISITION SYSTEM


115
important since the one-dimensional WIGLE calculations cannot account
for transverse propagation of a disturbance.
Static Flux Traverses
Shown in Figs.25 and 26 are the mapped steady-state fluxes across
the height and the width of the core respectively, at a distance of
66.6 cm from the neutron source. This location was chosen because'the
delay time and attenuation of the pulses showed it to be at the begin
ning of the "asymptotic region", viz. the region in which both the
velocity of propagation and the relaxation length have reached their
asymptotic values. Both flux maps were done with Indium foils and
appropriate corrections were applied to account for self-shielding and
interference (shadowing) effects between foils.
The fluxes calculated from four-group diffusion theory using the
AIM-6 code are displayed together with the experimental data. A cosine
fit was made on the vertical flux and is shown as a solid line in
Fig. 25; the fit obtained was excellent. The analysis of the results
in this direction gives an extrapolation distance between 5 and 7 cm.
The results were not as consistent across the width of the core.
A sophisticated aluminum foil holder was built to permit the location
of the foils between two rows of fuel; the foils were inserted in slots
parallel to the axis of the assembly. Due to the mechanical difficulty
of positioning the foils in the assembly, the physical impossibility of
having them perfectly horizontal and the arduous, significant correction
due to the short distance between foils, only one irradiation was per
formed. When all this is taken into account the results look reason
able. It was found that the bowing of the fuel rods affected these
measurements significantly. This was determined by positioning the


114
FIG. 24B THE CALCULATED SPATIAL DISTRIBUTION OF THE THERMAL
FLUX AT DIFFERENT TIMES AFTER THE PULSE
13
12
11
10
9
8
7
6
5
4
3
2
1
0


147
group pulse shapes. Several typical examples are shown in Appendix
D. The following features were observed:
1. The calculated pulse shapes are delayed in time in comparison
to the measured shapes throughout most of the length of the assembly.
2. The measured and calculated attenuation of the peaks is in
poor agreement.
3. The measured and calculated pulse shapes are practically
identical.
The validity of the comparison between theory and experiment in
this instance is not clear, however. The main question is whether the
experimental measurement corresponds to the fast group used in the
calculations, since the sensitivity of the detector rapidly diminishes
with increasing neutron energies. Therefore the conclusions based on
the above comparison must be taken with some reservation, until a more
detailed study can be conducted.
Propagation of a Narrow Pulse
A series of measurements were made in the clean core utilizing a
100 microsec wide input pulse. Only 4 positions along the longitudinal
axis were studied. The purpose of these measurements was to obtain a
qualitative idea of how well the WIGLE scheme can predict the space-
time kinetic behavior of a system perturbed by a very narrow pulse.
From the experimental viewpoint the narrow input pulse resulted in
much poorer count rates, thus requiring longer data acquisition times.
Listed below are the calculated and measured FWHM and delay times
for these measurements; the pulse shapes at the different positions
are included in Appendix E.


flux (relative units)
TIME PROFILE OF THERMAL NEUTRON FLUX 15*71 EM FROM THE SOURE'E
FIG- C2


PREFACE
The text of this dissertation is divided into two related but
essentially independent parts:
Part 1. The University of Florida SPERT Assembly
- Design and Calibration -
Part 2. Space-Time Reactor Kinetics Studies with the
University of Florida SPERT Assembly
In Part 1, the design, operational safety and nuclear calibration
of the large-in-one-space dimension, highly multiplicative University
of Florida SPERT Assembly (UFSA) are described. The presentation of
this material is pertinent for a complete understanding of the physical
characteristics of the system to be studied in Part 2 and also because
the system in itself is interesting from the nuclear engineering point
of view. The data acquisition system used for the experimentation is
presented in Chapter IV of this section.
In Part 2, the linear reactor kinetics studies performed with the
UFSA are presented. The space-time dependent distribution of the
neutrons in the assembly, following the introduction of a burst of
neutrons at one end of the core, are studied in the time and in the
frequency domain.
The study of the spatially dependent time profile of the neutron
flux required a large number of figures containing the calculational
and experimental results at different positions in the core. Some
typical figures are included in the main text but most of the recorded
(and calculated) time profiles are included as appendices to the main text
v


APPENDIX G
UFSA R1 CLEAN CORE
SHAPE OF THE PROPAGATING PULSE
AS A FUNCTION OF INPUT PULSE WIDTH
Pulse Width (msec) = 0.1,0.5,1.0,2.0,3.0,
4.0,5.0,6.0,7.0,8.0,
9.0,10.0


LIST OF TABLES (cont'd)
TABLE Page
XVITHE REAL AND THE IMAGINARY COMPONENTS
OF THE COMPLEX INVERSE RELAXATION LENGTH
- 0.5 MSEC INPUT PULSE 164
XVIITHE REAL AND THE IMAGINARY COMPONENTS OF
THE COMPLEX INVERSE RELAXATION LENGTH
- 1.0 MSEC INPUT PULSE 165
XVIIITHE DECAY CONSTANT AND kB/i VALUES MEASURED
AS A FUNCTION OF INPUT PULSE WIDTH 180
XIXREGION-WISE DEPENDENCE OF THE REACTIVITY
MEASUREMENTS 183
xii


FLUX (RELATIVE. UNITS)
FIG D1
TIME PROFILE OF FAST NEUTRON FLUX 1571 CM FROM THE SOURCE


TABLE OF CONTENTS (contd)
Page
CHAPTER V EXPERIMENTAL AND THEORETICAL RESULTS IN
THE FREQUENCY DOMAIN 156
Introduction 156
Method of Analysis ... 157
Comparison of the Theoretical and the Measured
Results of the Neutron Wave Analysis 159
CHAPTER VI SPATIAL DEPENDENCE OF PULSED-NEUTRON
REACTIVITY MEASUREMENTS 174
Introduction ... 174
The Decay Constant 175
The Ratio kS/& 176
The Measured Reactivity and ke^ Values 179
CHAPTER VII CONCLUSIONS 185
APPENDICES
A CALCULATIONAL PROCEDURES USED IN THE DETERMINATION
OF THE NUCLEAR PARAMETERS AND THE k _ VALUES 187
eff
B DESCRIPTION OF THE COMPUTER PROGRAMS 191
C UFSA R1 CLEAN CORE
TIME PROFILES OF THERMAL NEUTRON FLUX AT NINETEEN CORE
POSITIONS FOR INPUT PULSES OF 0.5 AND 1.0 MSEC 196
D UFSA R1 CLEAN CORE
TIME PROFILES OF FAST NEUTRON FLUX AT SEVERAL CORE
POSITIONS FOR INPUT PULSES OF 0.5 AND 1.0 MSEC 235
E UFSA R1 CLEAN CORE
TIME PROFILES OF THERMAL NEUTRON FLUX AT FOUR POSITIONS
IN THE CORE FOR A 0.1 MSEC INPUT PULSE 248
ix


81
is the experimental verification of present space-time kinetics
models and to guide the development of new calculational techniques.
Phase I of the Large Core Dynamics Experimental Program is being
conducted at the University of Florida SPERT Assembly, under Subcon
tracts C281 and C635 with the Phillips Petroleum Co. This part of the
program will provide the most fundamental neutron physics type of in
formation as the redistribution characteristics of the neutron flux in
a close-to-critical assembly are studied in the absence of inherent
feedback effects. The adequacy of basic parts of the analytical models
can be tested accurately by concentrating on the neutron physics
problem.
In this work, the propagation of a narrow, non-asymptotic neutron
burst is studied in the time and in the frequency domain. The experi
mental results are compared with results obtained from the two-group,
space-time dependent, one-dimensional diffusion theory calculational
scheme known as the WIGLE program [24]. A stringent test of the model
is provided by a combined analysis in the time and frequency domain.
Clean, unambiguous experimental information will be shown demonstrating
the presence of spatial effects in large cores.
Description of the Study
Pulse propagation phenomena in a large-in-one-space dimension side
reflected core were studied by introducing a fast neutron burst through
one face of the assembly. A Cockcroft-Walton (TNC) type generator was
employed to produce square neutron pulses of selected widths using the
(d,t) neutron reaction.
The assembly which was studied is described in detail in Part 1


113
FIG. 24A THE CALCULATED SPATIAL DISTRIBUTION OF THE THERMAL
FLUX AT DIFFERENT TIMES AFTER THE PULSE


46
Logic signals have fixed amplitude-shape characteristic and convey
information by their presence, absence of time relationship with
respect to a reference signal. Only gamma and noise discrimination
are necessary; this process will inherently eliminate the counting of
weak neutron interactions.
The process of registering events whose density per unit time
varies considerably requires an electronic system with extremely high
time resolution and a broad frequency response. Typically the counting
rates at the peak of the neutron burst in pulsed neutron measurements
vary from 10^ to > 10^ event/sec. This high, time-varying count rate
caused drastic losses in conventional instrumentation due to pulse
pile-up and circuit induced distortion when the dynamic range of the
circuit is exceeded causing a net amplitude shift in excess of that
tolerable by the system. The alternative modus operandi is to use low
count rates so that resolution losses are minimized; this, however,
increases the probability of systematic errors due to much longer run
ning times.
The count rates encountered during the initial experiments with a
3
He counter close to the neutron generator target in the SPERT assembly
exceeded 10^ event/sec at the peak. Saturation effects were observed
in modular instrumentation incorporating the latest F.E.T. preamplifier
and double-delay line pulse shaping amplifier at about 10"* event/sec.
The fast pulse train caused a drop in the base voltage of the Field
Effect Transistor which prevented further counting until a significant
reduction in the count rate allowed the voltage to recover. Changing
to a fast scintillation prototype preamplifier improved the situation
somewhat. However, if the potential count rate in the UFSA assembly


FLUX (RELATIVE UNITS)
FIG- C29 TIME PROFILE OF THERMAL NEUTRON FLUX 117-44 CM FROM THE SOURCE


FLUX (RELATIVE UNITS)
FIG- C16 TIME PROFILE OF THERMAL NEUTRON FLUX 133-74 CM FROM THE SOURCE
212


BIOGRAPHICAL SKETCH
Nils J. Diaz was born in Moron, Cuba on April 7, 1938. He was
graduated in June, 1955 from LaSalle High School in Havana, Cuba. In
July, I960, he obtained the degree of Professional Mechanical Engine
ering from the University of Villanova, Havana, Cuba, with honors* He
worked as a plant design engineer from January, 1960 to April, 1961
and also as instructor of Machine and Plant Design at the University of
Villanova for two semesters after his graduation. He arrived in the
United States in October, 1961 and worked as a machine designer until
September, 1962. He then entered the Graduate School of the University
of Florida and received a Master of Science in Engineering in June, 1964.
He held a graduate assistantship and a Fellowship from the Organization
of American States until December, 1965. From January, 1966 to date he
held a Junior Faculty appointment, as an Engineering Assistant, in the
Nuclear Engineering Sciences Department of the University of Florida,
while working toward the degree of Doctor of Philosophy.
Nils J. Diaz is married to the former Zena G. Gonzalez and is the
father of three children, Nils, Ariadne, and Aliene. He is a member of
the American Nuclear Society and the Society of the SIGMA XI.
262


159
Comparison of the Theoretical and the Measured
Results of the Neutron Wave Analysis
To determine the effect that the input pulse width could have on
the results of the wave analysis, the study was performed for both the
0.5 and 1.0 msec input pulse cases. These experimental and theoretical
results in the time domain were presented in Part 2, Chapter IV of this
work.
Shown in Figs. 38 through 41 are the amplitude and phase of the ex
perimental data obtained by numerical Fourier transformation, for both
the 0.5 and the 1.0 msec input pulses. The results for the 0.5 msec
case are consistently "smoother" than those for the 1.0 msec case. In
the 0.5 msec case the results "blow-up" past 1000 cps while the 1 msec
case seems to be good only to 800 cps. The behavior of the amplitude and
phase is as expected. The data was still good in a frequency region were
most wave type measurements in multiplying media reported to date had
already broken down. For example the experiments of Dunlap were only
good to ~ 250 cps [23]. Undoubtedly, the frequency content of the nar
row, fast burst is superior to the smeared pulse that is normally ob
tained if a thermalizing tank is used. The wave analysis on the WIGLE
space-time results showed an identical behavior throughout the fre
quency range.
The real and the imaginary components of the complex inverse
relaxation length are shown for both theory and experiment in Tables
XVI and XVII, for the 0.5 and the 1.0 msec input pulses respectively.
A graphical comparison of these results is shown in Figs. 42 and 43.
The following comments are pertinent:
a) A distinct but not significant difference is found between the
results of the 0.5 and the 1.0 msec cases in both theory and experiment.


134
The attenuation of the amplitude of the peaks for the UFSA assem
bly were shown in Figs. 31 and 32; the normalized peak values were
tabulated in Tables VII and VIII. Again excellent agreement was ob
tained between theory and experiment, not only in the asymptotic region
but throughout most of the assembly. The first two data points (P6 and
P13) are considerably larger in magnitude than the values predicted by
WIGLE. This is probably caused by an undercorrection for the epicadmium
flux detected at these positions and by not accounting for the signifi
cant neutron streaming in the large void created by the aluminum port
in which the target is located. Position 27 is probably the only "bad"
data point and no explanation has been found for its behavior since
four consecutive runs were made at that position and all four showed
identical behavior. The WIGLE results reflect the influence of end
effects more markedly than experiment. It should be noted that reflec
tions from the concrete walls in the assembly room are significant at
peripheral detector positions.
TABLE X
DYNAMIC INVERSE RELAXATION LENGTH <.
d
UFSA R1 Clean Core
0.5 M/W Ratio
16.35 cm wide reflected core
Input Pulse
Width (msec)
Theory
Experiment
0.5
.03025
.0297+.0012
1.0
.03025
.0301+.0010


Inverse Multiplication
Moderator Level (cm)
FIG. 16A INVERSE MULTIPLICATION vs. MODERATOR LEVEL


Delay Time (msec)
129

FIG. 29 CALCULATED AND EXPERIMENTAL DELAY TIMES
- 0.5 MSEC INPUT PULSE -


184
list of the decay constants, kg/£, p and ke^ measured across the core
and side reflector at a distance of 130.2 cm from the source.
The core-reflector configuration investigated is too narrow to
observe any transverse propagation of the pulse; consequently no dif
ference in the value of the reactivity was measured across the system.
The value of a and of kg/£ obtained near the outer edge of the reflector
was low compared to the other values but somehow compensated each other
to maintain p approximately constant.


39
will still be subcritical and the instrumentation will have sufficient
time to scram the system before accidental criticality occurs.
Design Basis Accident Analysis
Two types of accidents have been postulated to occur in conjunction
with the UFSA experimental program. The design basis accident analysis
will deal with the following two cases: a) a dropped fuel rod accident
in which a fuel rod is broken releasing the fission products accumulated
during previous operation of the assembly; b) an accidental criticality
resulting in a power excursion. It will be shown below that the poten
tial consequences of either accident are well within the radiation dose
limits specified by Title 10 CFR Part 20.
Dropped Fuel Rod Accident
For the purposes of this analysis, it is assumed that a dropped
fuel rod results in a broken pin releasing the accumulated fission pro
ducts. The following bases have been established to calculate the
radiation dose from such an accident:
a. It is stipulated that the assembly has been operated continu
ously for 200 hr at an average power level of 0.065 watt. It should be
noted that this is a conservative value since the assembly will not be
operated on a continuous basis; the 200 hours of operating time for each
configuration is essentially spread over a period of several weeks.
b. The assembly contains 1200 rods. Since the actual loadings for
the three reflected cases are 1206, 2132 and 3430 rods respectively, this
will result in conservative results.
c. The peak to space-averaged flux ratio during operation is a
maximum of 120, with the rod at peak flux contributing 1/10 of the



PAGE 1

63$&(7,0( 5($&725 .,1(7,&6 678',(6 :,7+ 7+( 81,9(56,7< 2) )/25,'$ 63(57 $66(0%/< %\ 1,/6 ',$= $ ',66(57$7,21 35(6(17(' 72 7+( *5$'8$7( &281&,/ 2) 7+( 81,9(56,7< 2) )/25,'$ ,1 3$57,$/ )8/),//0(17 2) 7+( 5(48,5(0(176 )25 7+( '(*5(( 2) '2&725 2) 3+,/2623+< 81,9(56,7< 2) )/25,'$

PAGE 2

72 =(1$

PAGE 3

$&.12:/('*0(176 7KH DXWKRU ZLVKHV WR H[SUHVV KLV VLQFHUH DSSUHFLDWLRQ WR KLV JUDGXDWH FRPPLWWHH IRU WKHLU JXLGDQFH 6SHFLDO UHFRJQLWLRQ LV GXH 'U 0 2KDQLDQ ZKRVH HQFRXUDJHPHQW GHGLFDWLRQ DQG GHWDLOHG VFLHQWLILF NQRZOHGJH PDGH WKLV ZRUN SRVVLEOH 7KH VXSSRUW RI WKH 1XFOHDU (QJLQHHULQJ 6FLHQFHV 'HSDUWPHQW RI WKH 8QLYHUVLW\ RI )ORULGD WKURXJKRXW WKH DXWKRUf¬V JUDGXDWH ZRUN LV JUHDWO\ DSSUHFLDWHG ,Q SDUWLFXODU 'U 5 ( 8KULJf¬V VXSSRUW DQG IULHQGVKLS LV JUDWHIXOO\ UHFRJQL]HG 7KH DXWKRU IHHOV IRUWXQDWH DQG SURXG LQ KDYLQJ VWXGLHG DQG ZRUNHG ZLWK D UHPDUNDEOH VFLHQWLVW DQG JHQWOHPDQ 'U 5 % 3HUH] QRZ DW 2DN 5LGJH 1DWLRQDO /DERUDWRU\ 7KH DXWKRUnV \HDUV RI DVVRFLDWLRQ ZLWK 'U 7 ) 3DUNLQVRQ QRZ DW WKH 8QLYHUVLW\ RI 0LVVRXUL IRUPHG WKH QHFHVVDU\ EDFNJURXQG IRU WKLV ZRUN 7KH FRQWLQXRXV DVVLVWDQFH RI 0U * : )RJOH WKURXJKRXW WKH H[n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

PAGE 4

XQGHU D SULPH FRQWUDFW ZLWK WKH 8QLWHG 6WDWHV $WRPLF (QHUJ\ &RPPLVVLRQ 7KH WHFKQLFDO DLG RI WKH 8QLYHUVLW\ RI )ORULGD &RPSXWLQJ &HQWHU LQ WKH GHYHORSPHQW RI WKH FRPSXWHU SURJUDPV DQG WKHLU ILQDQFLDO DVVLVWDQFH LV JUDWHIXOO\ DFNQRZOHGJHG 6SHFLDO PHQWLRQ LV GXH 0HVVUV 6 -RKQVRQ 5 : *DUQHU * $ 0RUWHQVHQ DQG 0UV 0 ( 5DGG RI WKH 1XFOHDU 6DIHW\ 5HVHDUFK %UDQFK $WRPLF (QHUJ\ 'LYLVLRQ 3KLOOLSV 3HWUROHXP &RPSDQ\ IRU WKHLU FRQWLQXRXV DVVLVWDQFH DQG LQYDOXDEOH VXJJHVWLRQV WKURXJKRXW WKH UHVHDUFK SURJUDP 7R P\ VLVWHU 0LVV /\GLD *RQ]DOH] P\ VLQFHUH DSSUHFLDWLRQ IRU W\SLQJ DQ HOHJDQW PDQXVFULSW IURP P\ XQLQWHOOLJLEOH FKDUDFWHUV 7R WKH IHOORZ VWXGHQWV ZKR VWUXJJOHG ZLWK PH WR UHDFK WKH XQn UHDFKDEOH VWDU P\ VSDFH DQG WLPH LQGHSHQGHQW IULHQGVKLS LY

PAGE 5

35()$&( 7KH WH[W RI WKLV GLVVHUWDWLRQ LV GLYLGHG LQWR WZR UHODWHG EXW HVVHQWLDOO\ LQGHSHQGHQW SDUWV 3DUW 7KH 8QLYHUVLW\ RI )ORULGD 63(57 $VVHPEO\ 'HVLJQ DQG &DOLEUDWLRQ 3DUW 6SDFH7LPH 5HDFWRU .LQHWLFV 6WXGLHV ZLWK WKH 8QLYHUVLW\ RI )ORULGD 63(57 $VVHPEO\ ,Q 3DUW WKH GHVLJQ RSHUDWLRQDO VDIHW\ DQG QXFOHDU FDOLEUDWLRQ RI WKH ODUJHLQRQHVSDFH GLPHQVLRQ KLJKO\ PXOWLSOLFDWLYH 8QLYHUVLW\ RI )ORULGD 63(57 $VVHPEO\ 8)6$f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f WLPH SURILOHV DUH LQFOXGHG DV DSSHQGLFHV WR WKH PDLQ WH[W Y

PAGE 6

7$%/( 2) &217(176 3DJH $&.12:/('*0(176 LLL 35()$&( 9 /,67 2) 7$%/(6 [L /,67 2) ),*85(6 [LLL /,67 2) 6<0%2/6 [YLLL $%675$&7 [L[ L fµ n 3$57 7+( 81,9(56,7< 2) )/25,'$ 63(57 $66(0%/< '(6,*1 $1' &$/,%5$7,21 &+$37(5 , ,1752'8&7,21 &+$37(5 ,, '(6&5,37,21 2) 7+( )$&,/,7< *HQHUDO )HDWXUHV )XHO &KDUDFWHULVWLFV 0HFKDQLFDO 'HVLJQ 0RGHUDWRU )ORZ &RQWURO 6\VWHP ,QVWUXPHQWDWLRQ DQG ,QWHUORFN 6\VWHP )XHO 6WRUDJH 1HXWURQ 6RXUFHV &+$37(5 ,,, 23(5$7,21$/ 6$)(7< ,QWURGXFWLRQ ,QLWLDO /RDGLQJ YL

PAGE 7

7$%/( 2) &217(176 FRQWnGf 2SHUDWLQJ /LPLWV 'HVLJQ %DVLV $FFLGHQW $QDO\VLV &+$37(5 ,9 7+( '$7$ $&48,6,7,21 6<67(0 ,QWURGXFWLRQ 7KH 1HXWURQ 'HWHFWRU 7KH (OHFWURQLF ,QVWUXPHQWDWLRQ 7KH 5HVROXWLRQ 7LPH &RUUHFWLRQ 7KH 1RUPDOL]DWLRQ 7HFKQLTXH &RPPHQWV &+$37(5 9 18&/($5 &$/,%5$7,21 2) 7+( 8)6$ 68%&5,7,&$/ ,QWURGXFWLRQ 7KHRUHWLFDO 1RWHV ,QYHUVH 0XOWLSOLFDWLRQ 0HDVXUHPHQWV $EVROXWH 'HWHUPLQDWLRQ RI N HII &RQFOXVLRQV 3DJH fµ 3$57 63$&(7,0( 5($&725 .,1(7,&6 678',(6 :,7+ 7+( 81,9(56,7< 2) )/25,'$ 63(57 $66(0%/< &+$37(5 , ,1752'8&7,21 6WDWHPHQW RI WKH 3UREOHP 'HVFULSWLRQ RI WKH 6WXG\ YLL

PAGE 8

7$%/( 2) &217(176 FRQWnGf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

PAGE 9

7$%/( 2) &217(176 FRQWf¬Gf 3DJH &+$37(5 9 (;3(5,0(17$/ $1' 7+(25(7,&$/ 5(68/76 ,1 7+( )5(48(1&< '20$,1 ,QWURGXFWLRQ 0HWKRG RI $QDO\VLV m &RPSDULVRQ RI WKH 7KHRUHWLFDO DQG WKH 0HDVXUHG 5HVXOWV RI WKH 1HXWURQ :DYH $QDO\VLV &+$37(5 9, 63$7,$/ '(3(1'(1&( 2) 38/6('1(87521 5($&7,9,7< 0($685(0(176 ,QWURGXFWLRQ 7KH 'HFD\ &RQVWDQW 7KH 5DWLR N6t r 7KH 0HDVXUHG 5HDFWLYLW\ DQG NHA 9DOXHV &+$37(5 9,, &21&/86,216 $33(1',&(6 $ &$/&8/$7,21$/ 352&('85(6 86(' ,1 7+( '(7(50,1$7,21 2) 7+( 18&/($5 3$5$0(7(56 $1' 7+( N B B 9$/8(6 HII % '(6&5,37,21 2) 7+( &20387(5 352*5$06 & 8)6$ 5 &/($1 &25( 7,0( 352),/(6 2) 7+(50$/ 1(87521 )/8; $7 1,1(7((1 &25( 326,7,216 )25 ,1387 38/6(6 2) $1' 06(& ' 8)6$ 5 &/($1 &25( 7,0( 352),/(6 2) )$67 1(87521 )/8; $7 6(9(5$/ &25( 326,7,216 )25 ,1387 38/6(6 2) $1' 06(& ( 8)6$ 5 &/($1 &25( 7,0( 352),/(6 2) 7+(50$/ 1(87521 )/8; $7 )285 326,7,216 ,1 7+( &25( )25 $ 06(& ,1387 38/6( L[

PAGE 10

7$%/( 2) &217(176 FRQWnGf 3DJH ) 8)6$ 5 &/($1 &25( 7,0( 352),/(6 2) 7+(.0$/ 1(87521 )/8; $7 7+5(( 326,7,216 ,1 7+( &25( )25 $ :,'( 06(&f ,1387 38/6( * 8)6$ 5 &/($1 &25( 6+$3( 2) 7+( 3523$*$7,1* 38/6( $6 $ )81&7,21 2) ,1387 38/6( :,'7+ /,67 2) 5()(5(1&(6 %,2*5$3+,&$/ 6.(7&+ [

PAGE 11

/,67 2) 7$%/(6 7$%/( 3DJH ,N f± YV 02'(5$725 /(9(/ 2) 8)6$ 5()/(&7(' HII &25(6 ,, 8)6$ ,167580(17$7,21 $1' &21752/ ,,, 6800$5< 2) 0 $1' 38/6(' 0($685(0(176 ,9,1387 3$5$0(7(56 )25 7+( :,*/( &$/&8/$7,21$/ 6&+(0( 9 7,0( 67(36 86(' )25 7+( :,*/( &$/&8/$7,216 9,'(/$< 7,0(6 0($685(' $&5266 7+( :,'7+ 2) 7+( &25( 06(& ,1387 38/6( 9,,&/($1 &25( 38/6( 3523$*$7,21 678',(6 (;3(5,0(17$/ $1' 7+(25(7,&$/ 5(68/76 06(& ,1387 38/6( :,'7+ 9,,,&/($1 &25( 38/6( 3523$*$7,21 678',(6 (;3(5,0(17$/ $1' 7+(25(7,&$/ 5(68/76 06(& ,1387 38/6( :,'7+ ,; $6<03727,& 3523$*$7,21 9(/2&,7< Y 3 ; '<1$0,& ,19(56( 5(/$;$7,21 /(1*7+ N D ;,&+$1*(6 ,1 7+( 18&/($5 3$5$0(7(56 '8( 72 &+$1*(6 ,1 7+( 75$169(56( %8&./,1* ;,,7+( &$/&8/$7(' $6<03727,& 9(/2&,7< 2) 3523$*$7,21 $1' '<1$0,& ,19(56( 5(/$;$7,21 /(1*7+ YV &25( +(,*+7 06(& ,1387 38/6( :,'7+ ;,,, '(/$< 7,0(6 $1' ):+0 )25 $ 1$552: ,1387 38/6( ;,9 '(/$< 7,0(6 $1' ):+0 )25 $ :,'( ,1387 38/6( ;9 38/6( 6+$3(6 YV ,1387 38/6( :,'7+ [L

PAGE 12

/,67 2) 7$%/(6 FRQWnGf 7$%/( 3DJH ;9,7+( 5($/ $1' 7+( ,0$*,1$5< &20321(176 2) 7+( &203/(; ,19(56( 5(/$;$7,21 /(1*7+ 06(& ,1387 38/6( ;9,,7+( 5($/ $1' 7+( ,0$*,1$5< &20321(176 2) 7+( &203/(; ,19(56( 5(/$;$7,21 /(1*7+ 06(& ,1387 38/6( ;9,,,7+( '(&$< &2167$17 $1' N%L 9$/8(6 0($685(' $6 $ )81&7,21 2) ,1387 38/6( :,'7+ ;,;5(*,21:,6( '(3(1'(1&( 2) 7+( 5($&7,9,7< 0($685(0(176 [LL

PAGE 13

/,67 2) ),*85(6 ),*85( 3DJH 29(5$// 9,(: 2) 7+( )$&,/,7< $ %27720 )8(/ 52' 63$&,1* 6<67(0 % %27720 )8(/ 52' 63$&,1* 6<67(0 723 )8(/ 52' 63$&,1* 6<67(0 5($&7,9,7<&21752/ )/2: 6<67(0 )/2: 5$7( YV +(,*+7 2) :$7(5 /(9(/ $%29( :(,5 $3(; :(,5 %2;f° $,5 6<67(0 6&+(0$7,& 8)6$ 6$)(7< 6<67(0 /2*,& )/2: ',$*5$0 32:(5 YV 7,0( )25 7+( '(6,*1 %$6,6 $&&,'(17 3+<6,&$/ &+$5$&7(5,67,&6 2) 7+( +H 1(87521 &2817(56 029$%/( '(7(&725 '$7$ $&48,6,7,21 6<67(0 1250$/,=,1* '(7(&725 '$7$ $&48,6,7,21 6<67(0 7,0( 352),/(6 2) 1(87521 %8567 5(&25'(' %< &219(17,21$/ (/(&7521,& ,167580(17$7,21 $1' %< 7+( 7,0(3,&.2)) 6<67(0 7+( 3$5$/,=,1* 1213$5$/,=,1* 6<67(0 5(62/87,21 7,0( &255(&7,21 $6 $ )81&7,21 2) &2817 5$7( '(7(&725 326,7,21,1* 6&+(0( $ ,19(56( 08/7,3/,&$7,21 YV 02'(5$725 /(9(/ % ,19(56( 08/7,3/,&$7,21 YV 648$5(' ,19(56( +(,*+7 [LLL

PAGE 14

/,67 2) ),*85(6 FRQWf¬Gf ),*85( 3DJH $ '(&$< &2167$17 YV 02'(5$725 /(9(/ % N6e $1' N YV 02'(5$725 /(9(/ 7+( 727$/ (3,&$'0,80 $1' 7+(50$/ )/8; &0 )520 7+( 6285&( 8)6$ 6285&(68%&5,7,&$/ $66(0%/< *(20(75,&$/ $55$1*(0(17 3/$1 9,(: 3/$1 $1' )5217 9,(: 2) 7+( &25( 5(*,21 (1&/26,1* 7+( 1(87521 6285&( 21(',0(16,21$/ $55$1*(0(17 2) 7+( 8)6$ &25( 86(' ,1 7+( :,*/( &$/&8/$7,21$/ 6&+(0( 63$7,$/ ',675,%87,21 2) 6285&( 1(875216 ,1&25325$7(' ,172 7+( :,*/( 6&+(0( $ 38/6( 6+$3(6 35(',&7(' %< :,*/( $7 ',))(5(17 326,7,216 ,1 7+( 8)6$ 5 &25( % 38/6( 6+$3(6 35(',&7(' %< :,*/( $7 ',))(5(17 326,7,216 ,1 7+( 8)6$ 5 &25( ,OO & 38/6( 6+$3(6 35(',&7(' %< :,*/( $7 ',))(5(17 326,7,216 ,1 7+( 8)6$ 5 &25( $ 7+( &$/&8/$7(' 63$7,$/ ',675,%87,21 2) 7+( 7+(50$/ )/8; $7 ',))(5(17 7,0(6 $)7(5 7+( 38/6( % 7+( &$/&8/$7(' 63$7,$/ ',675,%87,21 2) 7+( 7+(50$/ )/8; $7 ',))(5(17 7,0(6 $)7(5 7+( 38/6( 7+( $6<03727,& 67($'<67$7( 9(57,&$/ )/8; 7+( $6<03727,& 67($'<67$7( +25,=217$/ )/8; N (;3(5,0(17$/ 38/6( 6+$3(6 $7 ',))(5(17 326,7,216 ,1 7+( 8)6$ 5 &25( [LY

PAGE 15

/,67 2) ),*85(6 FRQWf¬Gf ),*85( 3DJH % (;3(5,0(17$/ 38/6( 6+$3(6 $7 ',))(5(17 326,7,216 ,1 7+( 8)6$ 5 &25( & (;3(5,0(17$/ 38/6( 6+$3(6 $7 ',))(5(17 326,7,216 ,1 7+( 8)6$ 5 &25( $ (;3(5,0(17$//< '(7(50,1(' 63$7,$/ ',675,%87,21 2) 7+( 1(875216 $7 ',))(5(17 7,0(6 $)7(5 7+( 386/( % (;3(5,0(17$//< '(7(50,1(' 63$7,$/ ',675,%87,21 2) 7+( 1(875216 $7 ',))(5(17 7,0(6 $)7(5 7+( 38/6( &$/&8/$7(' $1' (;3(5,0(17$/ '(/$< 7,0(6 06(& ,1387 38/6( &$/&8/$7(' $1' (;3(5,0(17$/ '(/$< 7,0(6 06(& ,1387 38/6( $03/,78'( $77(18$7,21 2) 7+( 7+(50$/ )/8; 06(& ,1387 38/6( $03/,78'( $77(18$7,21 2) 7+( 7+(50$/ )/8; 06(& ,1387 38/6( '(/$< 7,0(6 2) 7+( 7+(50$/ )/8; &$/&8/$7(' %< 7+( :,*/( &$/&8/$7,21$/ 6&+(0( )25 ',))(5(17 &25( +(,*+76 $03/,78'( $77(18$7,21 2) 7+( 7+(50$/ )/8; &$/&8/$7(' %< 7+( :,*/( &$/&8/$7,21$/ 6&+(0( )25 ',))(5(17 &25( +(,*+76 $ 7+( 6(16,7,9,7< 2) 7+( 21(',0(16,21$/ 7:2 *5283 63$&(7,0( .,1(7,&6 6&+(0( 72 &+$1*(6 ,1 7+( 75$169(56( %8&./,1* % 7+( 6(16,7,9,7< 2) 7+( 21(',0(16,21$/ 7:2 *5283 63$&(7,0( .,1(7,&6 6&+(0( 72 &+$1*(6 ,1 7+( 75$169(56( %8&./,1* [Y

PAGE 16

/,67 2) ),*85(6 FRQWnGf ),*85( 3DJH & 7+( 6(16,7,9,7< 2) 7+( 21(',0(16,21$/ 722 *5283 63$&(7,0( .,1(7,&6 6&+(0( 72 &+$1*(6 ,1 7+( 75$169(56( %8&./,1* (;3(5,0(17$/ 38/6( 6+$3(6 $6 $ )81&7,21 2) &25( +(,*+7 $ ())(&7 2) 5220 5(7851 $7 3(5,3+(5$/ '(7(&725 326,7,216 % ())(&7 2) 5220 5(7851 $7 3(5,3+(5$/ '(7(&725 326,7,216 $03/,78'( 2) =(527+ )285,(5 020(17 YV ',67$1&( )25 6(9(5$/ )5(48(1&,(6 06(& ,1387 38/6( $03/,78'( 2) =(527+ )285,(5 020(17 YV ',67$1&( )25 6(9(5$/ )5(48(1&,(6 06(& ,1387 38/6( 3+$6( 2) =(527+ )285,(5 020(17 YV ',67$1&( )25 6(9(5$/ )5(48(1&,(6 06(& ,1387 38/6( 3+$6( 2) =(527+ )285,(5 020(17 YV ',67$1&( )25 6(9(5$/ )5(48(1&,(6 06(& ,1387 38/6( &203$5,621 2) 7+( 7+(25(7,&$//< 35(',&7(' $1' 7+( 0($685(' '$03,1* &2()),&,(17 D &203$5,621 2) 7+( 7+(25(7,&$//< 35(',&7(' $1' 7+( 0($685(' 3+$6( 6+,)7 3(5 81,7 /(1*7+ 7+( 8)6$ 5 &25( S ',63(56,21 /$: &203$5,621 2) 7+( 7+(25(7,&$//< 35(',&7(' $1' 7+( 0($685(' D ef [YL

PAGE 17

/,67 2) ),*85(6 FRQWnGf ),*85( 3DJH &203$5,621 2) 7+( 7+(25(7,&$//< 35(',&7(' $1' 7+( 0($685(' D ) 7+( 8)6$ 5 &25( S ',63(56,21 /$: '(&$< &2167$17 YV $;,$/ 326,7,21 N%$ YV $;,$/ 326,7,21 5($&7,9,7< f N I YV $;,$/ 326,7,21 &O 7KURXJK & 7,0( 352),/(6 2) 7+( 7+(50$/ 1(87521 )/8; $7 1,1(7((1 326,7,216 ,1 7+( &25( )25 $ 06(& ,1387 38/6( & 7KURXJK & 7,0( 352),/(6 2) 7+( 7+(50$/ 1(87521 )/8; $7 1,1(7((1 326,7,216 ,1 7+( &25( )25 $ 06(& ,1387 38/6( 'O 7KURXJK ' 7,0( 352),/(6 2) 7+( )$67 1(87521 )/8; $7 6,; 326,7,216 ,1 7+( &25( )25 $ 06(& ,1387 38/6( ' 7KURXJK ' 7,0( 352),/(6 2) 7+( )$67 1(87521 )/8; $7 6,; 326,7,216 ,1 7+( &25( )25 $ 06(& ,1387 38/6( (O 7KURXJK ( 7,0( 352),/(6 2) 7+(50$/ 1(87521 )/8; $7 )285 326,7,216 ,1 7+( &25( )25 $ 06(& ,1387 38/6( ), 7,0( 352),/(6 2) 7+(50$/ 1(87521 )/8; $7 7+5(( 326,7,216 ,1 7+( &25( )25 $ 06(& ,1387 38/6( * 7KURXJK * 6+$3( 2) 7+( 3523$*$7,1* 38/6( $6 $ )81&7,21 2) 38/6( :,'7+ [YLL

PAGE 18

/,67 2) 6<0%2/6 75$169(56( %8&./,1* '(/$<(' 1(87521 35(&85625 &21&(175$7,21 2) 7+( LWK *5283 ',))86,21 &2()),&,(17 2) 7+( LWK *5283 )5(48(1&< FSVf ())(&7,9( 08/7,3/,&$7,21 &2167$17 1(87521 /,)(7,0( 1(87521 08/7,3/,&$7,21 9(/2&,7< 2) 7+( LWK *5283 $;,$/ &225',1$7( '(&$< &2167$17 ,1 7+( 7,0( '20$,1f '$03,1* &2()),&,(17 ,1 7+( )5(48(1&< '20$,1f ())(&7,9( '(/$<(' 1(87521 )5$&7,21 3+$6( 6+,)7 3(5 81,7 /(1*7+ 5($&7,9,7< &203/(; ,19(56( 5(/$;$7,21 /(1*7+ 0$&526&23,& &5266 6(&7,21 1(87521 )/8; [YLLL

PAGE 19

$EVWUDFW RI 'LVVHUWDWLRQ 3UHVHQWHG WR WKH *UDGXDWH &RXQFLO LQ 3DUWLDO )XOILOOPHQW RI WKH 5HTXLUHPHQWV IRU WKH 'HJUHH RI 'RFWRU RI 3KLORVRSK\ 63$&(7,0( 5($&725 .,1(7,&6 678',(6 :,7+ 7+( 81,9(56,7< 2) )/25,'$ 63(57 $66(0%/< %\ 1LOV 'LD] 0DUFK &KDLUPDQ 'U 0 2KDQLDQ 0DMRU 'HSDUWPHQW 1XFOHDU (QJLQHHULQJ 6FLHQFHV $ ODUJHLQRQHVSDFH GLPHQVLRQ VLGH UHIOHFWHG KLJKO\ PXOWLSOLFDn WLYH NJII a f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f LV D OLJKWZDWHU PRGHUDWHG VXEFULWLFDO IDFLOLW\ IXHOHG E\ b HQULFKHG 8&A SHOOHWV HQFDVHG LQ VWDLQOHVV VWHHO WXEHV RI LQFK RXWVLGH GLDPHWHU 63(57 )O )XHOf 7KH IXHO DUUD\V DUH FRQWDLQHG LQ D UHFWDQJXODU WDQN IHHW [L[

PAGE 20

ORQJ LQFKHV KLJK DQG RI YDULDEOH ZLGWK ,Q WKLV VWXG\ WKH FRUH ZDV LQFKHV ZLGH DQG LQFKHV KLJK 7KH HIIHFWLYH PXOWLSOLFDWLRQ FRQVWDQW RI WKH DVVHPEO\ ZDV GHWHUPLQHG WR EH M. 7KH DVVHPEO\ LV HTXLSSHG ZLWK QXFOHDU LQVWUXPHQWDWLRQ FDSDEOH RI DXWRPDWLF VFUDP DFWLRQ )RU WKH NLQHWLFV VWXGLHV D IDVW GDWD DFTXLVLWLRQ V\VWHP ZDV GHYHORSHG WR KDQGOH DFFXUDWHO\ WKH YHU\ KLJK WLPHFKDQJLQJ FRXQW UDWH HQFRXQWHUHG LQ WKH PHDVXUHPHQWV ,W HVVHQWLDOO\ FRQVLVWV RI D WUDQVn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n EO\ LQ WKH DEVHQFH RI LQKHUHQW IHHGEDFN HIIHFWV 7KH H[SHULPHQWDO UHVXOWV DUH FRPSDUHG ZLWK WKH UHVXOWV REWDLQHG IURP WKH WZRJURXS VSDFHWLPH GHSHQGHQW RQHGLPHQVLRQDO GLIIXVLRQ WKHRU\ VFKHPH NQRZQ DV WKH :,*/( SURJUDP $ VWULQJHQW WHVW RI WKH PRGHO LV SURYLGHG E\ D FRPELQHG DQDO\VLV LQ WKH WLPH DQG WKH IUHTXHQF\ GRPDLQ 7LLH :,*/( FDOFXODWLRQDO VFKHPH DFFXUDWHO\ SUHGLFWV WKH GHOD\ WLPHV DQG WKH DWWHQXDWLRQ RI WKH SXOVHV ZKHQ D ILUVWIOLJKW VSDWLDO GLVWULEXn WLRQ LV DVVXPHG IRU WKH IDVW VRXUFH $W ODUJH GLVWDQFHV IURP WKH VRXUFH :,*/( XQGHUSUHGLFWV a b LQ WKH ):+:f WKH VSUHDGLQJ RI WKH SXOVH $ PDUNHG VHQVLWLYLW\ WR VPDOO FKDQJHV LQ WKH WUDQVYHUVH EXFNOLQJ ZDV nIRXQG IRU WKH PRGHO DV ZHOO DV WKH H[SHULPHQW [[

PAGE 21

$ RQHWRRQH FRPSDULVRQ RI WKH SUHGLFWHG DQG PHDVXUHG YDOXHV LQ WKH IUHTXHQF\ GRPDLQ ZDV SURYLGHG E\ SHUIRUPLQJ LGHQWLFDO QXPHULFDO )RXULHU WUDQVIRUPDWLRQV RI WKH :,*/( WLPH SURILOHV DQG WKH PHDVXUHG SXOVH VKDSHV 7KH DQDO\VLV LQ WKH IUHTXHQF\ GRPDLQ FRQILUPHG WKH UHVXOWV REWDLQHG LQ WKH WLPH GRPDLQ DOWKRXJK GLVFUHSDQFLHV SDVW FSV DUH IRXQG LQ WKH XOWUDVHQVLWLYH S SODQH 7KH DJUHHPHQW LQ WKH S SODQH WKH V\VWHPnV GLVSHUVLRQ ODZ LV JRRG XS WR FSV DQG UHDVRQDEOH XS WR FSV %RWK WKHRU\ DQG H[SHULPHQW VKRZHG D VPRRWK EHKDYLRU WKURXJKRXW WKH IUHTXHQF\ UDQJH LQYHVWLJDWHG LQ ERWK WKH S DQG WKH S SODQH 6SDWLDO HIIHFWV LQ ODUJH FRUHV DUH FOHDUO\ GHPRQVWUDWHG LQ WKLV ZRUN 7KH GHWHUPLQDWLRQ RI WKH UDQJH RI DSSOLFDELOLW\ RI WKH RQHn GLPHQVLRQDO VFKHPH UHTXLUHV H[WHQGLQJ WKH VWXG\ WR FDVHV LQ ZKLFK WZR GLPHQVLRQDO HIIHFWV ZLOO EH QRWLFHDEOH DQG WKH LPSRUWDQW IHHGEDFN HIIHFWV FDQ EH FRQVLGHUHG [[L

PAGE 22

3$57 7+( 81,9(56,7< 2) )/25,'$ 63(57 $66(0%/< '(6,*1 $1' &$/,%5$7,21

PAGE 23

&+$37(5 , ,1752'8&7,21 7KH GHYHORSPHQW DQG FRQVWUXFWLRQ RI ODUJH SRZHU UHDFWRUV IRFXVHG WKH DWWHQWLRQ RI LQGXVWU\ DQG RI WKH 8QLWHG 6WDWHV $WRPLF (QHUJ\ &RPn PLVVLRQ RQ WKH QHFHVVLW\ RI KDYLQJ UHOLDEOH UHDFWRU G\QDPLFV DQDO\VLV PHWKRGV WR DFFXUDWHO\ GHVFULEH WKH VSDWLDO DQG WHPSRUDO EHKDYLRU RI WKH QHXWURQ IOX[ LQ WKHVH V\VWHPV 7KH SRLQWPRGHO UHDFWRU NLQHWLFV FDOFX ODWLRQV VHHP WR KDYH EHHQ DGHTXDWH IRU WKH JURVV HYDOXDWLRQ RI WKH WLPH GHSHQGHQW QHXWURQ IOX[ GXULQJ WKH RFFXUUHQFH RI D WUDQVLHQW EXW WKH PRGHO FDQ EH LQ ODUJH HUURU ZKHQ WKH SK\VLFDO VL]H RI WKH V\VWHP DQG WKH PDJQLWXGH RI WKH SHUWXUEDWLRQ QHFHVVLWDWHV WKDW VSDWLDO HIIHFWV LQ WKH UHGLVWULEXWLRQ RI WKH QHXWURQ IOX[ EH FRQVLGHUHG 3UHOLPLQDU\ FDOFXODn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

PAGE 24

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n LW\ RI FXUUHQW VSDFHWLPH NLQHWLF PRGHOV LQ WKH DEVHQFH RI LQKHUHQW IHHGEDFN HIIHFWV 3KDVH ,, .LQHWLF %HKDYLRU IRU &RQWURO5RG,QGXFHG 3RZHU ([FXUVLRQV LQ /DUJH 2QH'LPHQVLRQDO &RUHV $ UHDFWRU ODUJH LQ RQHGLPHQVLRQ IHHW ORQJ WKUHH IHHW GHHS DQG ZLWK YDU\LQJ ZLGWKV WR DFFRPPRGDWH GLIIHUHQW PHWDOZDWHU UDWLRV ZLOO EH XVHG WR LQYHVWLJDWH WKH RQHGLPHQVLRQDO NLQHWLF EHKDYLRU RI ODUJH FRUHV VXEMHFWHG WR D ODUJH SHUWXUEDWLRQ %RWK QRQIHHGEDFN ORZ SRZHU H[SHUn LPHQWVf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

PAGE 25

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n LWLHV RI VXFK V\VWHPV ZHUH WKH REMHFW RI D GHWDLOHG VWXG\ WR GHWHUPLQH WKHLU RSHUDWLRQDO FKDUDFWHULVWLFV XQGHU QRUPDO DQG DFFLGHQW FRQGLWLRQV 7KH IOH[LEOH PHFKDQLFDO GHVLJQ WKH VDIHW\ LQVWUXPHQWDWLRQ WKH QXFOHDU HYDOXDWLRQ DV ZHOO DV WKH H[SHULPHQWDO FDOLEUDWLRQ RI WKH ILUVW FRQILJXUDWLRQ XQGHU VWXG\ FRQVWLWXWHV 3DUW RI WKLV GLVVHUWDWLRQ 7KH UHDFWRU NLQHWLFV VWXGLHV SHUIRUPHG LQ WKH ILUVW RI WKH FRQn ILJXUDWLRQV WR EH VWXGLHG DUH GHDOW ZLWK LQ 3DUW RI WKLV PDQXVFULSW

PAGE 26

&+$37(5 ,, '(6&5,37,21 2) 7+( )$&,/,7< *HQHUDO )HDWXUHV 7KH 8QLYHUVLW\ RI )ORULGD 6SHUW $VVHPEO\ LV D OLJKW ZDWHUn PRGHUDWHG VXEFULWLFDO IDFLOLW\ IXHOHG E\ b HQULFKHG 8e SHOOHWV HQFDVHG LQ VWDLQOHVV VWHHO WXEHV RI RXWVLGH GLDPHWHU 7KH IXHO DUUD\V DUH FRQWDLQHG LQ D UHFWDQJXODU WDQN IHHW ORQJ LQFKHV KLJK DQG RI YDULDEOH ZLGWKV 7KH V\VWHP LV GHVLJQHG VR WKDW ERWK EDUH DQG UHIOHFWHG FRUHV FDQ EH VWXGLHG 2QO\ RQH UHIOHFWHG FRUH ZLOO EH GHDOW ZLWK LQ 3DUW RI WKLV PDQXVFULSW LQIRUPDWLRQ RQ WKUHH UHIOHFWHG FRUHV LV LQFOXGHG LQ WKLV FKDSWHU 7KH DVVHPEO\ ZLGWK DQG IXHO VSDFLQJ PD\ EH YDULHG LQ RUGHU WR Df KDYH D N QRW WR H[FHHG LQ DOO FDVHV WR EH HII FRQVLGHUHG Ef DFFRPPRGDWH QRQPRGHUDWRUPRGHUDWRU UDWLRV RI DQG UHVSHFWLYHO\ 6KRZQ LQ 7DEOH , DUH WKH N BnV DV D IXQFWLRQ RI WKH PRGHUDWRU HII KHLJKW WKH IXHO VSDFLQJV FRUH ZLGWKV DQG WRWDO QXPEHU RI IXHO HOHn PHQWV IRU WKH GLIIHUHQW UHIOHFWHG FDVHV WR EH FRQVLGHUHG 7KH FDOFX ODWLRQDO SURFHGXUHV XVHG LQ WKH GHWHUPLQDWLRQ RI WKH QXFOHDU SDUDPHWHUV DQG WKH NAA YDOXHV IRU WKH WKUHH UHIOHFWHG FRQILJXUDWLRQV RI WKH DVVHPEO\ DUH GHVFULEHG LQ $SSHQGL[ $ 2QO\ WKH VLGHV RI WKH DVVHPEO\ ZLOO EH UHIOHFWHG )LJ VKRZV DQ RYHUDOO YLHZ RI WKH IDFLOLW\

PAGE 27

:(,5 6<67(0 &25( 7$1. 5()/(&725 7$1. 5(6(592,5 '803 9$/9(6 ),* 29(5$// 9,(: 2) 7+( )$&,/,7<

PAGE 28

7$%/( , NHII YV 02'(5$725 /(9(/ 2) 8)6$ 5()/(&7(' &25(6 &DOFXODWHG 8VLQJ WKH $,0 &RGH &RUH /HQJWK FP f 5HIOHFWRU :LGWK FP f $FWLYH )XHO +HLJKW FP f 0HWDO WR :DWHU 5DWLR /DWWLFH 3LWFK LQf &RUH :LGWK FPf 1R RI )XHO 5RGV 0RGHUDWRU /HYHO FPf (IIHFWLYH 0XOWLSOLFDWLRQ &RQVWDQW

PAGE 29

7KH DVVHPEOLHV DUH KLJKO\ PXOWLSOLFDWLYH WKLV LV LPSRUWDQW IRU WKH H[WUDSRODWLRQ RI WKH UHVXOWV RI WKH VWXG\ WR FULWLFDO V\VWHPV 7KH V\VWHPnV VXEFULWLFDOLW\ LV DWWUDFWLYH EHFDXVH RI WKH LQKHUHQW VDIHW\ RI VXFK V\VWHPV DQG RI WKH DEVHQFH RI LQKHUHQW IHHGEDFN HIIHFWV ,Q RUGHU WR SURYLGH DV FOHDQ D FRUH DV SRVVLEOH D XQLTXH FRQWURO V\VWHP ZKLFK KDV EHHQ VXFFHVVIXOO\ XVHG RQ WKH 8)$3$ >@ ZLOO EH HPn SOR\HG ,Q WKLV V\VWHP WKH UHDFWLYLW\ LV FRQWUROOHG E\ DGMXVWPHQW RI WKH ZDWHU KHLJKW LQ WKH DVVHPEO\ 7KH ZDWHU KHLJKW LV FRQWUROOHG E\ WKH SRVLWLRQ RI WZR 9QRWFKHG ZHLUV ORFDWHG LQ D ZDWHU ER[ K\GUDXn OLFDOO\ FRXSOHG ZLWK WKH DVVHPEO\ WKURXJK IOH[LEOH OLQHV 7KH TXDQWLW\ RI ZDWHU GLVFKDUJLQJ WKURXJK D 9QRWFKHG ZHLU YDULHV DV + + LV WKH GLVWDQFH EHWZHHQ WKH DSH[ RI WKH ZHLU DQG WKH ZDWHU OHYHOf WKXV SURn YLGLQJ SUHFLVH FRQWURO RI WKH PRGHUDWRU KHLJKW 7KH K\GUDXOLF FRXSOLQJ DVVXUHV WKDW XQGHU QRUPDO RSHUDWLQJ FRQGLWLRQV ZLWK FRQWLQXRXV IORZf WKHUH ZLOO EH WKH VDPH ZDWHU OHYHO LQ WKH FRUH DQG WKH UHIOHFWRU WDQNV 7KH 8)6$ VXEFULWLFDO DVVHPEO\ LV ORFDWHG LQ DQ LVRODWHG DQG VKLHOGHG URRP LQ WKH 1XFOHDU 5HVHDUFK )LHOG %XLOGLQJ DSSUR[LPDWHO\ WKUHH PLOHV IURP WKH 8QLYHUVLW\ FDPSXV 7KH 1XFOHDU 5HVHDUFK )LHOG %XLOGLQJ FRQVLVWV RI IRXU ED\V WZR RI WKHP KDYLQJ VKLHOGHG URRPV IRU H[SHULPHQWV ZLWK VXEFULWLFDO DQG PRGHUn DWLQJ DVVHPEOLHV 7KH VKLHOGHG ZDOOV FRQVLVW RI VWDFNHG FRQFUHWH EORFN HLJKW IHHW KLJK DQG WKLUW\HLJKW LQFKHV ZLGH FRYHUHG ZLWK SO\ZRRG WR DVVXUH WKDW WKH EORFNV UHPDLQ LQ SODFH 7KH FHLOLQJ RI WKLV VLQJOHIORRU EXLOGLQJ LV DSSUR[LPDWHO\ IHHW DERYH WKH IORRU DQG FRQVLVWV RI H[FHOn VLRUILOOHG FHPHQW ERQGHG ERDUG 1HXWURQ UHIOHFWLRQ IURP WKLV FHLOLQJ RYHU WKH ZDOOV GRHV QRW FRQVWLWXWH D KD]DUG WR SHUVRQQHO RSHUDWLQJ WKH DFFHOHUDWRUW\SH QHXWURQ VRXUFH 7KH DFFHVV GRRU IURP WKH FRQWURO URRP

PAGE 30

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rFAOLQHQW b $FWLYH )XHO /HQJWK $FWLYH )XHO 'LDPHWHU )XHO 7XEH 0DWHULDO VWDLQOHVV VWHHO )XHO 7XEH /HQJWK )XHO 7XEH R G

PAGE 31

0HFKDQLFDO 'HVLJQ 7KH HQWLUH DVVHPEO\ FDQ EH GLYLGHG LQWR WKUHH FRPSRQHQWV WKH VXSSRUWLQJ SODWIRUP DQG GXPS WDQN WKH EDVLF FRUH WDQN DQG IXHO URG VXSSRUW VWUXFWXUH DQG WKH FRPELQDWLRQ FRUH VLGH ZDOOV DQG VLGH UHIOHFWRU WDQNV 7KH VXSSRUWLQJ SODWIRUP LV FRPSRVHG RI LQFK VWHHO ,EHDPV UDLVHG IHHW IURP WKH IORRU OHYHO E\ VL[ VWHHO FROXPQV LQFKHV LQ GLDPn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n SRUW DVVHPEO\ UHVWV RQ WKH SODWH 7KH HQG ZDOOV RI WKH EDVLF WDQN DUH PDGH RI D [ LQFK SODWH DQG DUH VXSSRUWHG E\ WZR [ [ LQFK VWHHO DQJOH EUDFHV ZHOGHG WR WKH ,EHDP SODWIRUP 7ZR IRXU LQFK DOXPLQXP FKDQQHOV VSDQ WKH HLJKW IRRW GLPHQVLRQ RI WKH WDQN FRQn ILQLQJ WKH XSSHU IXHO URG VXSSRUW V\VWHP DQG GHWHFWRU PRXQWV 7KH VLGH ZDOOV RI WKH EDVLF WDQN VHUYH DOVR DV UHIOHFWRU WDQNV ZKHQ WKH UHIOHFWHG FRUHV DUH XQGHU VWXG\ 7KHVH WDQNV KDYH GLPHQVLRQV RI [ [ LQFKHV 7KH DUUDQJHPHQW DOORZV RQH WR YDU\ WKH ZLGWK RI WKH FRUH ZLWK D VROH VWUXFWXUDO VXSSRUW

PAGE 32

7KH HQG ZDOOV DUH SHUPDQHQWO\ FRYHUHG ZLWK &DGPLXP RQ WKH RXWVLGH VXUIDFH ZKLOH WKH VLGH ZDOOV KDYH PRYDEOH &DGPLXP FRYHUV WR GHILQH WKH ERXQGDULHV IRU WKH EDUH DQG UHIOHFWHG FDVHV 7R RSWLPL]H WKH QXPEHU RI n QHXWURQV LQVHUWHG LQWR WKH DVVHPEO\ E\ WKH QHXWURQ JHQHUDWRU WKH DFFHOn HUDWRU WDUJHW SHQHWUDWHV DERXW LQFKHV LQWR WKH FRUH $ ZDWHUWLJKW SRUW LV SURYLGHG IRU WKLV SXUSRVH 7KH SRUW FDQ EH UHPRYHG DQG D EOLQG IODQJH LQVHUWHG LQ LWV SODFH 6HYHUDO IXHO URGV PXVW EH WDNHQ RXW WKH QXPEHU GHSHQGLQJ RQ KRZ GHHS WKH WDUJHW JRHV LQWR WKH DVVHPEO\ DQG RQ WKH ODWWLFH SLWFK 7KH FRUH VHFWLRQ RI WKH DVVHPEO\ FRQVLVWV RI DQ LQWHUFKDQJHDEOH IXHO URG VSDFLQJ V\VWHP PDGH RI [ [ LQFK FKDQQHOV [ LQFK EDUV DQG DOXPLQXP VKLPV PRXQWHG RQ WKH EDVH SODWH RI WKH WDQN 7KH EDUV KDYH PLOOHG VORWV WR DFFRPPRGDWH WKH LQFK HQG WLS RI WKH IXHO URG DQG WR VHW WKH SLWFK DORQJ WKH FRUH ZLGWK 7KH VKLPV DUH SODFHG EHWZHHQ WKH FKDQQHO EDU XQLWV WR VHW WKH SLWFK DORQJ WKH FRUH OHQJWK LQFKHVf VHH )LJ f 7KH WRS IXHO URG VSDFLQJ V\VWHP FRQVLVWV RI DQ DOXPLQXP JUDWLQJ 7KH PHVK LV GHWHUPLQHG E\ WKH ODWWLFH SLWFK XQGHU VWXG\ 7KH JUDWLQJ LV PDGH RI DOXPLQXP EDUV DQG VSDFHUV DV VKRZQ LQ )LJ 7KXV IXHO URG UHPRYDO DORQJ WKH OHQJWK RI WKH FRUH LV SRVn VLEOH WR ORFDWH WKH GHWHFWRU IRU WKH H[SHULPHQWDO PHDVXUHPHQWV 7KH RQHKDOI LQFK ORQJ URG WLS LV IXOO\ VXUURXQGHG E\ DOXPLQXP ZLWK SUDFWLFDOO\ QR UHIOHFWLQJ FKDUDFWHULVWLF EXW WKHUH LV D LQFK OHQJWK RI URG EHWZHHQ WKH HQG RI WKH DFWLYH IXHO DQG WKH WLS ZKLFK LV W VXUURXQGHG E\ ZDWHU 7KLV ERWWRP UHIOHFWRU LV XQDYRLGDEOH DQG ZLOO EH FRQVLGHUHG LQ WKH FDOFXODWLRQV

PAGE 33

),* $ %27720 )8(/ 52' 63$&,1* 6<67(0

PAGE 34

, ),* % %27720 )8(/ 52' 63$&,1* 6<67(0

PAGE 35

$/80,180 6833257 ),* 723 )8(/ 52' 63$&,1* 6<67(0

PAGE 36

0RGHUDWRU )ORZ &RQWURO 6\VWHP 7KH PRGHUDWRU IORZ FRQWURO V\VWHP RI WKH 8)6$ FDQ EH EHWWHU GHVn FULEHG E\n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n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n WUROOHG LQ D SUHFLVH PDQQHU VLPSO\ E\ YDU\LQJ WKH KHLJKW RI WKH ZHLUV DQG WKH UDWH RI IORZ RI ZDWHU LQWR WKH WDQN 7KLV LV DFFRPSOLVKHG E\

PAGE 37

:(,5 '5$,1 ',675,%87,21 0$1,)2/' 3803 ),* 5($&7,9,7<&21752/ )/2: 6<67(0 &21752/ 9$/9( f§ 25,),&(

PAGE 38

DQ DXWRPDWLFDOO\ RSHUDWHG SQHXPDWLF FRQWURO YDOYH $ SORW RI WKH IORZ UDWH YHUVXV WKH KHLJKW RI WKH ZDWHU OHYHO DERYH WKH DSH[ RI WKH ZHLUV LV VKRZQ LQ )LJ 7KH ZHLU SODWH LV ULJLGO\ PRXQWHG RQ D ER[ RU VPDOO WDQN VHH )LJ f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n WLRQ LV GHWHUPLQHG DV SDUW RI WKH LQLWLDO VWDUWXS SURFHGXUH IRU HDFK FRQILJXUDWLRQ WR EH FRQVLGHUHG 7KH ILQDO DGMXVWPHQW RI WKH SRVLWLRQ RI WKH ZHLUER[LV VXFK WKDW ZKHQ WKH ZDWHU UHDFKHV WKH PRGHUDWRU OHYHO LQ WKH DVVHPEO\ FRUUHVSRQGLQJ WR r IRU D JLYHQ FRQILJXUDWLRQ LW ZLOO EH IORZLQJ DERXW LQFKHV DERYH WKH DSH[ RI WKH 9QRWFK ZHLUV $W WKLV GHVLJQ OHYHO WKH IORZ UDWH LV a JDOORQV SHU PLQXWH ZLWK WKH NAA YDOXHV DV JLYHQ LQ 7DEOH , IRU WKH IXOO IXHO ORDGLQJ $IWHU WKH RSHUDWLQJ KHLJKW RI WKH ZHLU ER[ KDV EHHQ GHWHUPLQHG IRU NJAA A LQ WKH LQLWLDO VWDUWXS VWRSV DUH LQVHUWHG WR SUHYHQW UDLVLQJ WKH ZHLUV DERYH WKLV KHLJKW LI WKH KHLJKW LV OHVV WKDQ WKH DFWLYH IXHO KHLJKWf ,W VKRXOG DOVR EH SRLQWHG RXW WKDW WKH RULILFH LQ WKH OLQH OLPLWV WKH SXPS FDSDFLW\ WR D IORZ UDWH ZKLFK LV MXVW VXIILFLHQW WR EULQJ WKH ZHLUV WR IXOO IORZ $ IXUWKHU LQFUHDVH LQ IORZ UDWH ZRXOG FDXVH GLVFKDUJH RYHU WKH HQWLUH

PAGE 39

:DWHU /HYHO $ERYH :HLU $SH[ FPf )ORZ 5DWH JSPf ),* )/2: 5$7( YV +(,*+7 2) :$7(5 /(9(/ $%29( :(,5 $3(;

PAGE 40

),* :(,5 %2;

PAGE 41

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n WUROOHU UHFRUGV ERWK IORZ UDWH DQG PRGHUDWRU OHYHO LQ WKH FRUH 7KH FRQWURO YDOYH LV RI WKH DLUWRRSHQ W\SH ZKLFK ZLOO FORVH LQ WKH FDVH RI DLU VXSSO\ ORVV VWRSSLQJ WKH IORZ LQWR WKH DVVHPEO\ 7KH IORZ GLDJUDP IRU WKH DLU V\VWHP LV VKRZQ LQ )LJ $ SUHVVXUH GLIIHUHQWLDO IURP WKH SUHVVXUH WUDQVPLWWHUV DSSOLHG WR WKH UHFRUGHUFRQWUROOHU DOORZV ERWK PDQXDO DQG DXWRPDWLF FRQWUROn RI WKH IORZ UDWH WKURXJK WKH YDOYH RSHUDWHG E\ WKH FRQWUROOHU

PAGE 42

9$/9( ),* $,5 6<67(0 6&+(0$7,&

PAGE 43

,QVWUXPHQWDWLRQ DQG ,QWHUORFN 6\VWHP 7KH LQVWUXPHQWDWLRQ DQG LQWHUORFN V\VWHP RI WKH 8)6$ KDV EHHQ GLVFXVVHG H[WHQVLYHO\ LQ WKH UHSRUWV VXEPLWWHG WR WKH $WRPLF (QHUJ\ &RPPLVVLRQ > @ LQ FRQMXQFWLRQ ZLWK WKH OLFHQVH DSSOLFDWLRQ 0RUH UHFHQWO\ 0U / % 0\HUV VXEPLWWHG D GHWDLO WHFKQLFDO UHSRUW RQ WKH VXEMHFW >@ $ EULHI GHVFULSWLYH H[SODQDWLRQ LV JLYHQ EHORZ $ EORFN GLDJUDP RI WKH VDIHW\ V\VWHP ORJLF IORZ LQ XVH DW WKH 8)6$ VXEFULWLFDO DVVHPEO\ IRU URXWLQH PRQLWRULQJ LV VKRZQ LQ )LJ 7KHUH DUH ILYH SULQFLSDO FKDQQHOV RI LQVWUXPHQWDWLRQ D 6WDUWXS FKDQQHO XVLQJ D +H SURSRUWLRQDO FRXQWHU VFDOHU DQG UDWH PHWHU 7KH FRXQWHU LV ORFDWHG DW WKH ERWWRP RI WKH FRUH FORVH WR WKH JHRPHWULFDO FHQWHU RI WKH DVVHPEO\ E /RJ SRZHU DQG SHULRG LQVWUXPHQW 1R FKDQQHO XVLQJ D FRPSHQn VDWHG LRQ FKDPEHU RSHUDWHG XQFRPSHQVDWHGf DV D VLJQDO WR D /RJ 1 DPSOLILHU 7KH FKDPEHU LV ORFDWHG DORQJ WKH ORQJLWXGLQDO D[LV RI WKH DVVHPEO\ RQ WKH ERWWRP RI WKH FRUH VRPH WZR IHHW IURP WKH QHXWURQ JHQHUDWRU HQG F 3HULRG LQVWUXPHQW 1R FKDQQHO XVLQJ D FRPSHQVDWHG LRQ FKDPEHU RSHUDWHG XQFRPSHQVDWHGf DV D VLJQDO WR D ORJ 1 DPSOLILHU 7KH FKDPEHU LV ORFDWHG DORQJ WKH ORQJLWXGLQDO D[LV RI WKH DVVHPEO\ RQ WKH ERWWRP RI WKH FRUH VRPH VL[ IHHW IURP WKH QHXWURQ JHQHUDWRU HQG G /LQHDU QHXWURQ IOX[ 1R FKDQQHO XVLQJ DQ XQFRPSHQVDWHG LRQ FKDPEHU DV D VLJQDO WR D PLFURPLFURDPPHWHU 7KH FKDPEHU LV PRXQWHG RQ WKH WRS FRUH VXSSRUW IUDPH FORVH WR WKH JHRPHWULFDO FHQWHU RI WKH FRUH H /LQHDU QHXWURQ IOX[ 1R FKDQQHO XVLQJ D FRPSHQVDWHG LRQ FKDPEHU RSHUDWHG XQFRPSHQVDWHGf WR IHHG D VLJQDO WR D PLFURPLFURDPPHWHU 7KH FKDPEHU LV PRXQWHG RQ WKH WRS FRUH VXSSRUW IUDPH RQ WKH RSSRVLWH

PAGE 44

12 ),* 8)6$ 6$)(7< 6<67(0 /2*,& )/2: ',$*5$0

PAGE 45

VLGH IURP WKH OLQHDU FKDQQHO 1R ,WHPV D WKURXJK G DUH SDUW RI WKH VDIHW\ DPSOLILHU ZKLOH LWHP I LV XVHG WR GLVSOD\ WKH QHXWURQ IOX[ RQ D FRQVROH IURQW SDQHO PHWHU 7KH VDIHW\ DPSOLILHU PRQLWRUV WKH VHYHQ FRQWLQXRXVO\ YDU\LQJ LQSXW VLJQDOV DQG SURYLGHV D WULS VLJQDO LI DQ\ RI WKH LQSXW VLJQDOV IDOO RXWVLGH RI DFFHSWDEOH OLPLWV 7KH VDIHW\ DPSOLILHU SURYLGHV PHDQV RI DGMXVWLQJ WKHVH OLPLWV RYHU D ZLGH UDQJH 7KH GXDOLW\ RI WKH VFUDP DFWLRQ VHH )LJ f LV D SURPLQHQW IHDWXUH RI WKH VDIHW\ V\VWHP ,W FDQ EH VDLG WKDW QR VLQJOH IDLOXUH ZLOO LQYDOLGDWH ERWK DXWRPDWLF VFUDP FKDQQHOV )XUWKHUPRUH LW KDV EHHQ GHWHUPLQHG WKDW QR VLQJOH IDLOXUH FDQ LQYDOLGDWH ERWK WKH PDQXDO DQG DXWRPDWLF VFUDPV 7KH PHWKRG RI PHDVXUHPHQW DQG WKH IXQFWLRQ RI HDFK LQVWUXPHQWDWLRQ DQG VDIHW\ FKDQQHO DUH VKRZQ LQ 7DEOH ,, SDUWV $ DQG %f $ VHULHV RI VDIHW\ LQWHUORFNV SUHYHQW ZDWHU IURP IORZLQJ LQWR RU UHPDLQLQJ LQ WKH DVVHPEO\ XQOHVV D SURSHU VHTXHQFH RI HYHQWV DUH IROn ORZHG DQG FHUWDLQ FRQGLWLRQV DUH VDWLVILHG 7KH FRQGLWRQV DUH D 7KH PRGHUDWRU WHPSHUDWXUH PXVW EH !B r) 7KLV LV HVWDEOLVKHG E\ WKH GHVLUH WR REWDLQ WKH H[SHULPHQWDO GDWD QHDU URRP WHPSHUDWXUH FRQn GLWLRQV 7KH LQVHUWLRQ RI ZDWHU DW r) ZLOO LQWURGXFH D PD[LPXP N RI EDVHG RQ WKH FDOFXODWHG QHJDWLYH WHPSHUDWXUH FRHIILFLHQW RI UHDFWLYLW\f DERYH WKH GHVLJQ NAA YDOXH ZLWK QR KD]DUGV FUHDWHG E 7KH IRXU LQVWUXPHQWDWLRQ FKDQQHOV PXVW KDYH WKHLU KLJK YROWDJH RQ F 7KH FRUH ZLGWK PXVW EH VPDOOHU WKDQ FP G 7KH GRRU WR WKH DVVHPEO\ URRP PXVW EH ORFNHG H 7KH VWDUWXS FKDQQHO PXVW FRXQW PRUH WKDQ FRXQWVVHF I 7KH QHXWURQ IOX[ VXEFULWLFDO DVVHPEO\ SRZHU OHYHO DQG SHULRG

PAGE 46

7$%/( ,, 8)6$ ,167580(17$7,21 $1' &21752/ $ 18&/($5 0HDVXUHG 3DUDPHWHU 0HWKRG RI 0HDVXUHPHQW $SSOLFDWLRQ D /RZ OHYHO QHXWURQ IOX[ +H GHWHFWRU SXOVH GLVFULPLQDWRU DW QHXWURQ JHQHUDWRU HQG RI FRUH ,QVXUH VRXUFH LV SUHVHQW EHIRUH DGGLQJ UHDFWLYLW\ 6FUDP RQ ORZ FRXQW UDWH E /LQHDU QHXWURQ IOX[ F /RJ QHXWURQ IOX[r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r D E 7R EH RSHUDWHG LQ WKH XQFRPSHQVDWHG PRGH &RPPRQ GHWHFWRU DQG LQVWUXPHQW

PAGE 47

7$%/( ,, &RQWLQXHGf % 12118&/($5 0HDVXUHG 3DUDPHWHU D 5HDFWRU ZDWHU WHPSHUn DWXUH 0HWKRG RI 0HDVXUHPHQW )HQZDOO WHPSHUDWXUH VZLWFK LQ LQOHW OLQH E 5HDFWRU ZDWHU OHYHO F 5HDFWRU GRRU DQG SHUn VRQQHO %DUNVGDOH SUHVVXUH VZLWFK PRXQWHG RQ ZHLUER[ZLWK VHQVLQJ OLQH FRQQHFWHG WR FRUH /LPLW VZLWFKHV RQ GRRUV DQG SXVK EXWWRQV LQVLGH UHDFWRU URRP G 5HDFWRU FRUH ZLGWK /LPLW VZLWFKHV RQ UHIOHFWRU WDQNV H 5HDFWRU ZDWHU OHYHO %DUQVWHDG SUHVVXUH VZLWFK VHQVHV OHYHO LQ ZHLU ER[ I 5HIOHFWRU WDQN ZDWHU )ORDW VZLWFKHV LQ UHIOHFWRU WDQNV OHYHO ORZf J )ORZ FRQWURO YDOYH /LPLW VZLWFK RQ YDOYH VKXW K 5HDFWRU IORZ 'LIIHUHQWLDO SUHVVXUH FHOO DQG SQHXPDWLF FRQWURO L 5HDFWRU ZDWHU OHYHO '3 FHOO DQG SQHXPDWLF V\VWHP $SSOLFDWLRQ 6FUDP UHDFWRU RQ ORZ UHDFWRU LQOHW ZDWHU WHPS r)f 6FUDP UHDFWRU LI ZDWHU OHYHO LQ FRUH H[FHHGV WRS RI ZHLU KHLJKW 6FUDP V\VWHP DQG VKXW GRZQ QHXWURQ JXQ LI UHDFWRU GRRUV DUH RSHQHG RU LQWHULRU VZLWFKHV DUH DFWLn YDWHG 3URKLELW ILOOLQJ UHIOHFWRU WDQNV ZKHQ GLVWDQFH EHWZHHQ WDQNV H[FHHGV ZLGHVW UHIOHFWHG FRUH ZLGWK 6WRSV SXPS ZKHQ ZDWHU UHDFKHV FP EHORZ ZHLU DSH[ ,QGLFDWHV ZDWHU LV ILOOLQJ UHIOHFn WRU WDQNV 5HTXLUHV FORVLQJ YDOYH EHIRUH VWDUWLQJ SXPS &RQWURO IORZ UDWH ,QGLFDWH DQG UHFRUG IORZ UDWH ,QGLFDWH DQG UHFRUG ZDWHU OHYHO

PAGE 48

PXVW EH DV VSHFLILHG XQGHU 2SHUDWLQJ /LPLWV LQ WKLV UHSRUW J $IWHU D QRUPDO VWDUWXS WKH ZDWHU KHLJKW LQ WKH FRUH PXVW EH ZLWKLQ RI WKH OHYHO VHW E\ WKH SRVLWLRQ RI WKH ZHLUV )XHO 6WRUDJH 7KH ODUJH DPRXQW RI IXHO QHHGHG IRU WKH H[SHULPHQWDO SURJUDP UHTXLUHG D GHWDLOHG FULWLFDOLW\ DQDO\VLV RI WKH IXHO VWRUDJH DUHD &ULWLFDOLW\ FRQVLGHUDWLRQV RI WKH IXHO VWRUDJH DUUDQJHPHQWV IROORZ WKH $WRPLF (QHUJ\ &RPPLVVLRQ UHJXODWLRQV UHJDUGLQJ WKH VXEMHFW 7KUHH GLIn IHUHQW FULWHULD ZHUH XVHG WR FDOFXODWH WKH HIIHFWLYH PXOWLSOLFDWLRQ RI WKH IXHO VWRUDJH DUHD WR DVVXUH WKDW WKH DUUD\ ZLOO UHPDLQ VXEFULWLFDO XQGHU WKH ZRUVW FLUFXPVWDQFHV 7KH PHWKRGV DQG WKH FRUUHVSRQGLQJ FRQn GLWLRQV DUH RXWOLQHG EHORZ 7KH IXHO VWRUDJH DUUD\ FRQVLVWV RI WKUHH VODEV RI DLUVSDFHG IXHO SLQV VHSDUDWHG E\ D PLQLPXP GLVWDQFH RI LQFK IDFH WR FHQWHU 7KH IXHO LV VWRUHG LQ VWHHO EDVNHWV FRQWDLQLQJ SLQV SHU EDVNHW 7ZR VHWV RI LQFK WKLFN SODVWLF SODWHV ORFDWHG DW WKH ERWWRP DQG WRS RI HDFK EDVNHW GULOOHG WR SURSHUO\ SRVLWLRQ WKH IXHO URGV 7KH FKDUDFn WHULVWLFV RI WKH IXHO VODE DUH 6ODE :LGWK LQ FP FRUUHVSRQGV WR IXHO SLQV LQ WUDQVYHUVH GLUHFWLRQf 6ODE OHQJWK IW +HLJKW IW DFWLYH IXHO KHLJKWf D 0XOWLUHJLRQPXOWLJURXS FDOFXODWLRQ 7KH HIIHFWLYH PXOWLSOLFDWLRQ IDFWRU RI WKH IXHO VWRUDJH DUUD\ FRQn VLVWLQJ RI WKUHH VODEV LQFK ZLGH IW ORQJ DQG IW ZLGHf ZKLFK DUH LQFK DSDUW IDFH WR IDFHf ZDV FRPSXWHG IRU WKH FDVH RI IORRGLQJ

PAGE 49

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f ZDV XVHG 7KLV HVWDEOLVKHV WKH PD[LPXP

PAGE 50

WRWDO VROLG DQJOH VXEWHQGHG E\ DQ\ XQLW LQ WKH DUUD\ WR EH VWHUDGLDQV LI WKH HIIHFWLYH PXOWLSOLFDWLRQ IDFWRU IRU WKH LQGLYLGXDO VODEV LV OHVV WKDQ DV LV WKH FDVH KHUH )URP &)5 Gf f Lf WKH PLQLPXP UHTXLUHG VHSDUDWLRQ GLVn WDQFH LH FHQWHU RI RQH VODE WR IDFH RI DGMDFHQW VODE LV REWDLQHG IURP VWHUDGLDQV FURVV VHFWLRQDO DUHD VHSDUDWLRQ GLVWDQFHf [ 7KLV JLYHV VHSDUDWLRQ GLVWDQFH Af§f IW :H SURSRVH WR HVWDEOLVK D PLQLPXP GLVWDQFH EHWZHHQ IDFHV RI DGMDn FHQW VODEV RI IW 7KLV JLYHV D WRWDO VROLG DQJOH IRU WKH FHQWHU VODE RI VWHUDGLDQ ZKLFK LV ZHOO EHORZ WKH HVWDEOLVKHG FULWHULD F &RPSDULVRQ ZLWK &ODUNf¬V FULWHULD )RU b HQULFKHG LQFK GLDPHWHU XUDQLXP R[LGH URGV ZLWK D JPOLWHU 8 FRQFHQWUDWLRQ FRPSDUHG WR b HQULFKHG LQFK GLDPHWHU XUDQLXP R[LGH URGV ZLWK D JPOLWHU 8 FRQFHQWUDWLRQ LQ RXU FDVHf WKH IROORZLQJ GDWD LV REWDLQHG IURP SS DQG UHVSHFWLYHO\ RI '3 :LGWK RI 6ODE FPf &ULWLFDO 6DIH B %XFNOLQJ FP f &ULWLFDO 6DIH ,W VKRXOG EH QRWHG WKDW WKH VODE ZLGWKV TXRWHG RQ SDJH RI '3 DUH IRU DQ LQILQLWH ZDWHUPRGHUDWHG DQG UHIOHFWHG VODE )URP WKH EXFNn OLQJ YDOXHV JLYHQ WKH FULWLFDO ZLGWK RI WKH LQILQLWH ZDWHUPRGHUDWHG XQUHIOHFWHG VODE LV FP WKH FRUUHVSRQGLQJ VDIH ZLGWK LV FP :KHQ WZLFH WKH UHIOHFWRU VDYLQJV IRU WKH ODWWHU FDVH DV JLYHQ RQ SDJH RI '3 LV VXEWUDFWHG D VDIH ZLGWK IRU WKH LQILQLWH ZDWHU

PAGE 51

PRGHUDWHG UHIOHFWHG VODE RI FP LV REWDLQHG FRQVLVWHQW ZLWK WKH FP YDOXH 7KXV WKH VODE ZLGWK RI FP SURSRVHG E\ XV FRPSDUHV IDYRUDEO\ IURP WKH VDIHW\ YLHZSRLQW ZLWK WKH VDIH ZLGWK IRU DQ LQILQLWH ZDWHUn PRGHUDWHG DQG UHIOHFWHG VODE DQG LV FRQVLGHUDEO\ QDUURZHU WKDQ WKH VDIH ZLGWK IRU DQ LQILQLWH ZDWHUPRGHUDWHG DQG XQUHIOHFWHG VODE :LWKLQ WKH SUHVHQW FRQWH[W LW VKRXOG DOVR EH SRLQWHG RXW WKDW DV LQGLFDWHG RQ SDJH RI WKH 'HVLJQ DQG +D]DUGV 5HSRUW >@ QR IORRGLQJ RI WKH VWRUDJH DUHD VHHPV SRVVLEOH IURP QDWXUDO FDXVHV 1HXWURQ 6RXUFHV 7ZR W\SHV RI QHXWURQ VRXUFHV ZHUH XVHG WKURXJKRXW WKLV ZRUN f 7ZR 3X%H VRXUFHV PRXQWHG LQ DQ DOXPLQXP F\OLQGHU ZKLFK FDQ EH GULYHQ UHPRWHO\ WKURXJK D SODVWLF SLSH IURP D VKLHOGHG ER[ ORFDWHG LQ RQH FRUQHU RI WKH IDFLOLW\ URRP WR XQGHUQHDWK WKH FHQWHU RI WKH FRUH 1HRQ OLJKWV SURYLGH LQGLFDWLRQ DW WKH FRQVROH RI WKH SRVLWLRQ RI WKH VRXUFHV 7KHVH VRXUFHV ZKLFK KDYH D FRPELQHG \LHOG RI [ A QVHF DUH XVHG IRU VWDUWXSV DQG IRU WKH LQYHUVH PXOWLSOLFDWLRQ PHDVn XUHPHQWV f $ 7H[DV 1XFOHDU 1HXWURQ *HQHUDWRU ZKLFK LV XVHG LQ FRQWLQXRXV PRGH IRU VWDWLF PHDVXUHPHQWV DQG LQ WKH SXOVLQJ PRGH IRU WKH SXOVH SURSDJDWLRQ PHDVXUHPHQWV 7KH JHQHUDWRU LV RI WKH &RFNFURIW:DOWRQ W\SH 71& 0RGHO + ZLWK FRQWLQXRXVO\ YDULDEOH KLJK YROWDJH IURP NY DQG KDV EHHQ PRGLILHG WR REWDLQ ODUJHU FXUUHQWV E\ UHPRYLQJ WKH HLQ]HO OHQVHV DQG LQVWDOOLQJ D QHZ HOHFWURGH DFFHOHUDWRU WXEH DQG JDS OHQVH 3UH DQG SRVWDFFHOHUDWLRQ EHDP GHIOHFWLRQ SURGXFHV VKDUS ORZUHVLGXDO SXOVHV

PAGE 52

7KH DFFHOHUDWRU ZDV XVHG ZLWK D FXULH WULWLXP WDUJHW 7KH SRVLWLRQ RI WKH WDUJHW FDQ EH FKDQJHG WR NHHS WKH VRXUFH FHQWHUHG RQ WKH WDUJHWHQG RI WKH DVVHPEO\ IRU DQ\ JLYHQ PRGHUDWRU OHYHO

PAGE 53

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f 7KH 8)6$ IDFLOLW\ KDV EHHQ GHVLJQHG WR UHPDLQ VXEFULWLFDO XQGHU QRUPDO RSHUDWLQJ FRQGLWLRQV Ef 7KH VDIHW\ LQVWUXPHQWDWLRQ VHH 3DUW &KDSWHU ,f KDV EHHQ GHVLJQHG VXFK WKDW D VLQJOH IDLOXUH ZLOO QRW LQYDOLGDWH ERWK WKH PDQXDO DQG DXWRPDWLF VFUDP DQG ZLOO QRW FDXVH VXEVHTXHQW IDLOXUHV Ff 7KH GHVLJQ EDVLV DFFLGHQWV ZHUH SRVWXODWHG RQ D VLQJOH IDLOXUH

PAGE 54

FULWHULRQ Gf 2SHUDWLQJ OLPLWV KDYH EHHQ VHW WR GHOLQHDWH WKH QRUPDO RSHUn DWLQJ UDQJHV RI WKH DVVHPEO\ Hf ,QLWLDO ORDGLQJ SURFHGXUHV KDYH EHHQ HVWDEOLVKHG WR GHWHUPLQH WKH VDIH RSHUDWLQJ PXOWLSOLFDWLRQ IDFWRU RI HDFK FRQILJXUDWLRQ $ VHULHV RI DGPLQLVWUDWLYH FRQWUROV DUH QHFHVVDULO\ DSSOLHG WR DOO VHJPHQWV RI WKH H[SHULPHQWDO SURJUDP DQG VWULFWO\ HQIRUFHG ,QLWLDO /RDGLQJ $ VHULHV RI FDOFXODWLRQV ZHUH GRQH WR GHWHUPLQH ZKLFK RI WKH WZR IROORZLQJ VFKHPHV VKRXOG EH HPSOR\HG IRU WKH LQLWLDO ORDGLQJ RI 8)6$ ,f 6WHS ORDGLQJ RI WKH IXHO IURP WKH FHQWHU RXW DFFRPSDQLHG E\ VWHS LQFUHDVHV LQ ZDWHU OHYHO ZLWK WKH XVXDO LQYHUVH PXOWLSOLFDWLRQ GHWHUPLQDWLRQ ,,f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f FXUYH 7KH PRGHUDWRU OHYHO FRQWURO V\VWHP LQ RSHUDWLRQ DW WKH IDFLOLW\ SURYLGHV D H[WUHPHO\ UHOLDEOH DQG VDIH PRGH RI DGMXVWLQJ WKH ZDWHU OHYHO ZLWKRXW VDIHW\ FRPSURPLVHV 7KH IROORZLQJ UHJXODWLRQV ZHUH IROORZHG IRU WKH LQLWLDO IXHO

PAGE 55

ORDGLQJ DQG ZLOO EH IROORZHG IRU VXEVHTXHQW FRUHV f $IWHU WKH IXHO KDV EHHQ ORDGHG SULRU WR HDFK QHZ LQFUHPHQWDO FKDQJH LQ WKH ZDWHU OHYHO WKH ZDWHU LV GUDLQHG FRPSOHWHO\ DQG WKH ZHLU DQG ZDWHU OHYHO VFUDPf DGMXVWHG WR SUHYHQW D OHYHO LQFUHDVH EH\RQG WKH GHVLUHG YDOXH f 7KH ILUVW WKUHH PHDVXUHPHQWV RI WKH LQYHUVH PXOWLSOLFDWLRQ DUH REWDLQHG DW ZDWHU KHLJKWV RI DSSUR[LPDWHO\ FP FP DQG FP DERYH WKH ERWWRP RI WKH DFWLYH IXHO $V VKRZQ LQ 7DEOH , WKH PD[LPXP N f¬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f $W HDFK ILOOLQJ VWHS WKH PHDVXUHG NHMe RI WKH DVVHPEO\ LV FRPSDUHG ZLWK WKH FDOFXODWHG YDOXH ,I VLJQLILFDQW GHYLDWLRQV RI WKH H[SHULPHQWDO YDOXHV IURP WKH FDOFXODWHG NJAA YV ZDWHU KHLJKW FXUYHV RFFXU WKH H[SHULPHQWV DUH WR EH GLVFRQWLQXHG DQG D GHWDLOHG DQDO\VLV RI WKH UHVXOWV REWDLQHG SHUIRUPHG ,I LW LV GHWHUPLQHG WKDW WKH GLPHQn VLRQV RI WKH DVVHPEO\ VKRXOG EH FKDQJHG LQ RUGHU WR DFKLHYH WKH GHVLUHG NHA DW PD[LPXP ZDWHU KHLJKW WKH 8QLYHUVLW\ RI )ORULGD ZLOO DSSO\ IRU DQG REWDLQ ZULWWHQ DSSURYDO IURP WKH $WRPLF(QHUJ\ 'LYLVLRQ RI WKH 3KLOOLSV 3HWUROHXP &RPSDQ\ EHIRUH VXFK FKDQJHV DUH PDGH

PAGE 56

2SHUDWLQJ /LPLWV $ GHVFULSWLRQ RI WKH RSHUDWLQJ OLPLWV RI WKH VXEFULWLFDO DVVHPEO\ LQFOXGLQJ WKH EDVLV IRU VXFK OLPLWV DWH OLVWHG EHORZ (IIHFWLYH 0XOWLSOLFDWLRQ )DFWRU 6SHFLILFDWLRQ WKH PD[LPXP DOORZDEOH ZLOO EH 7KH DEVROXWH YDOXH RI N DV ZHOO DV WKH VORSH RI WKH N YV ZDWHU HII U HII KHLJKW ZLOO EH FDUHIXOO\ PHDVXUHG VR DV QRW WR H[FHHG WKH OLPLWLQJ YDOXH %DVLV WKH XSSHU OLPLW RI NJAA LV HVWDEOLVKHG E\ WKH DFFXUDF\ ZLWK ZKLFK NAA FDQ EH PHDVXUHG WKH UHSRUWHG >@ GLIIHUn HQFHV EHWZHHQ FDOFXODWLRQV DQG H[SHULPHQWV LQ VLPLODU FRUHV DQG WKH YDOXH RI WKH PXOWLSOLFDWLRQ IDFWRU UHTXLUHG WR PDNH D PHDQLQJIXO VWXG\ RI WKH G\QDPLF SURSHUWLHV RI ODUJH FRUHV &RPSDULVRQ >@ EHWZHHQ FDOFXODn WLRQV DQG WKH FRUUHVSRQGLQJ FULWLFDO H[SHULPHQWV RQ FRUHV VLPLODU WR WKH 8)6$f HVWDEOLVKHG WKDW DQ RYHUHVWLPDWH RI NJAA LV JHQHUDOO\ PDGH WKH VWDQGDUG GHYLDWLRQ IRU WKHVH FDVHV ZDV DQG WKH PD[LPXP XQGHU VWLPDWH RI WKH PXOWLSOLFDWLRQ ZDV IRU D FDVH \LHOGLQJ NJAA D b GHYLDWLRQ 7KH DEVROXWH YDOXH RI N ZLOO EH GHWHUPLQHG IRU ZDWHU OHYHOV HII \LHOGLQJ D NAA ! E\ SXOVHG WHFKQLTXHV LQGHSHQGHQWO\ RI WKH LQYHUVH PXOWLSOLFDWLRQ PHDVXUHPHQWV 7KH PHWKRG WR EH XVHG LV WKH *DUHOLV 5XVVHOO>@ PHWKRG ZKLFK ZKHQ DSSURSULDWH FRUUHFWLRQV DUH PDGH IRU WKH UHIOHFWRU KDV EHHQ VKRZQ WR JLYH JRRG UHVXOWV ,Q WKLV PHWKRG ERWK ONHIIOf NHII D DQG DUH GHWHUPLQHG WKHQ NHee PD\ EH REWDLQHG E\ LQGHSHQGHQWO\ REWDLQLQJ % RU / 6LQFH RI WKH WZR SDUDPHWHUV FKDQJHV WKH PRVW VORZO\ ZLWK NAA LW LV YDOLG WR XVH D YDOXH EDVHG RQ WKHRUHWn LFDO FDOFXODWLRQV

PAGE 57

5HDFWLYLW\ $GGLWLRQ 5DWH 6SHFLILFDWLRQ WKH UHDFWLYLW\ DGGLWLRQ UDWH LV FRQWUROOHG E\ WKH ZDWHU IORZ UDWH LQWR WKH VXEFULWLFDO DVVHPEO\ 7KH PD[LPXP IORZ LV IL[HG WR EH JSP $W WKLV PD[LPXP IORZ UDWH WKH UDWHV RI DGGLWLRQ RI ZDWHU DQG FRQVHTXHQWO\ RU UHDFWLYLW\ FRPSXWHG EHWZHHQ ZDWHU KHLJKWV RI DQG FP DUH /DWWLFH 3LWFK 5DWH RI LQFUHDVH RI ZDWHU OHYHO $NFP $NVHF VHF ,W VKRXOG EH QRWHG WKDW WKHVH UHDFWLYLW\ DGGLWLRQ UDWHV DUH D ODUJH RYHUVWLPDWH FRPSDUHG WR WKH FDOFXODWHG UDWHV DW NAA a %DVLV WKH PD[LPXP UDWH RI DGGLWLRQ RI UHDFWLYLW\ ZDV HVWDEOLVKHG E\ WKH FDOFXODWHG YDOXHV RI NAA YV ZDWHU KHLJKW DQG WKH PD[LPXP IORZ UDWH 7KH YDOXHV VSHFLILHG DERYH FRQVWLWXWH DQ XSSHU OLPLW LQ WKH UHJLRQ RI LQWHUHVW DQG DUH FRQVLGHUHG WR EH VDIH XQGHU FLUFXPVWDQFHV 7KH IORZ UDWH LV D IXQFWLRQ RI WKH FDSDFLW\ RI WKH SXPS WKH RULn ILFH DQG WKH SQHXPDWLF FRQWURO YDOYH DQG FDQQRW H[FHHG JSP 5HDFWLYLW\ 5HPRYDO 5DWH 6SHFLILFDWLRQ D FRQVHUYDWLYH YDOXH IRU WKH UHDFWLYLW\ UHPRYDO UDWH LV WDNHQ IURP WKH VORSH RI WKH N A YV ZDWHU KHLJKW FXUYHV QHDU WKH PD[LPXP GHVLHQHG N YDOXHV 6LQFH QR GLIIHUHQFH LV GHWHFWHG IRU Z HUU WKH VFUDP WLPHV RI WKH WKUHH FRQILJXUDWLRQV RQO\ RQH UDWH RI UHPRYDO ZLOO EH VSHFLILHG FRUUHVSRQGLQJ WR WKH VPDOOHVW VORSH

PAGE 58

'UDLQ UDWH LQFOXGLQJ WKH V\VWHP UHDFWLRQ WLPH FPVHF 5DWH RI UHPRYDO RI UHDFWLYLW\ $NFP $NVHF VHF LW VKRXOG EH QRWHG WKDW WKH YDOXH RI $NFP XVHG WR VSHFLI\ WKH UHDFn WLYLW\ UHPRYDO UDWH LV a WLPHV VPDOOHU WKDQ WKH FRUUHVSRQGLQJ YDOXH VSHFLILHG IRU WKH UHDFWLYLW\ LQVHUWLRQ UDWH %DVLV WKH UDWH RI GUDLQDJH IURP WKH DVVHPEO\ HVWDEOLVKHG WKH UHDFWLYLW\ UHPRYDO UDWH &RQVLVWHQW ZLWK WKH DSSURDFK WDNHQ ZKHQ VSHFLI\LQJ WKH UHDFWLYLW\ WKH WLPH WR GUDLQ FP RI ZDWHU IURP WKH DVVHPEO\ ZDV PHDVXUHG WR HVWDEOLVK D ORZHU OLPLW RQ WKH UDWH RI GHn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

PAGE 59

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n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n WLYLW\ LV DGGHG WR WKH V\VWHP IURP DQ\ GHVLJQ VXEFULWLFDOLW\ OHYHO 8QGHU WKHVH FRQGLWLRQV HYHQ DW WKH PD[LPXP NAA VWDWXV WKH IDFLOLW\

PAGE 60

ZLOO VWLOO EH VXEFULWLFDO DQG WKH LQVWUXPHQWDWLRQ ZLOO KDYH VXIILFLHQW WLPH WR VFUDP WKH V\VWHP EHIRUH DFFLGHQWDO FULWLFDOLW\ RFFXUV 'HVLJQ %DVLV $FFLGHQW $QDO\VLV 7ZR W\SHV RI DFFLGHQWV KDYH EHHQ SRVWXODWHG WR RFFXU LQ FRQMXQFWLRQ ZLWK WKH 8)6$ H[SHULPHQWDO SURJUDP 7KH GHVLJQ EDVLV DFFLGHQW DQDO\VLV ZLOO GHDO ZLWK WKH IROORZLQJ WZR FDVHV Df D GURSSHG IXHO URG DFFLGHQW LQ ZKLFK D IXHO URG LV EURNHQ UHOHDVLQJ WKH ILVVLRQ SURGXFWV DFFXPXODWHG GXULQJ SUHYLRXV RSHUDWLRQ RI WKH DVVHPEO\ Ef DQ DFFLGHQWDO FULWLFDOLW\ UHVXOWLQJ LQ D SRZHU H[FXUVLRQ ,W ZLOO EH VKRZQ EHORZ WKDW WKH SRWHQn WLDO FRQVHTXHQFHV RI HLWKHU DFFLGHQW DUH ZHOO ZLWKLQ WKH UDGLDWLRQ GRVH OLPLWV VSHFLILHG E\ 7LWOH &)5 3DUW 'URSSHG )XHO 5RG $FFLGHQW )RU WKH SXUSRVHV RI WKLV DQDO\VLV LW LV DVVXPHG WKDW D GURSSHG IXHO URG UHVXOWV LQ D EURNHQ SLQ UHOHDVLQJ WKH DFFXPXODWHG ILVVLRQ SURn GXFWV 7KH IROORZLQJ EDVHV KDYH EHHQ HVWDEOLVKHG WR FDOFXODWH WKH UDGLDWLRQ GRVH IURP VXFK DQ DFFLGHQW D ,W LV VWLSXODWHG WKDW WKH DVVHPEO\ KDV EHHQ RSHUDWHG FRQWLQXn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

PAGE 61

WRWDO SRZHU RI WKH DVVHPEO\ G 7KH IXHO URG VXEMHFWHG WR WKH VSHFLILHG KLJKHVW VSHFLILF SRZHU LV EURNHQ RSHQ DQG WKH IXHO LV FRPSOHWHO\ IUDJPHQWHG UHOHDVLQJ b RI WKH JDVHRXV ILVVLRQ SURGXFWV H 7KH LRGLQH LQ HOHPHQWDU\ IRUPf GLIIXVHV XQLIRUPO\ WKURXJKRXW WKH f¬ [ f¬ [ f¬ URRP WR FDOFXODWH WKH RQVLWH LQWHUQDO GRVH 7KH RIIVLWH GRVH ZDV FDOFXODWHG XVLQJ WKH YHU\ FRQVHUYDWLYH 3DVTXDOn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n GLDWHO\ DQG QRWLI\ WKH 5DGLDWLRQ 6DIHW\ 2IILFHU 7KH PD[LPXP WLPH UHTXLUHG WR HYDFXDWH WKH DVVHPEO\ DQG WKH IXHO VWRUDJH DUHD LV VHF ZLWK VHF QHHGHG WR HYDFXDWH WKH HQWLUH EXLOGLQJ WKHVH WLPHV KDYH EHHQ PHDVXUHG GXULQJ SUDFWLFH HYDFXDWLRQ RI WKH EXLOGLQJ 7KH RIIVLWH WRWDO LQKDODWLRQ GRVH FDOFXODWHG DVVXPLQJ WKH LRGLQH LV UHOHDVHG LQ RQH SXII ZLWK ]HUR ZLQG YHORFLW\ FORXG LQYHUWHG FRQn GLWLRQ ZHUH W\SLFDOO\

PAGE 62

'LVW£QF«f¬IURP VLW« Pf 7KH GLUHFW UDGLDWLRQ GRVH IURP DOO WKH ILVVLRQ SURGXFWV RI WKH LUUDGLDWHG IXHO HOHPHQWV LV FDOFXODWHG WR EH PUKU $FFLGHQWDO &ULWLFDOLW\ DQG 6XEVHTXHQW 3RZHU ([FXUVLRQ $FFLGHQWDO FULWLFDOLW\ DQG D VXEVHTXHQW SRZHU H[FXUVLRQ FRXOG RFFXU RQO\ E\ WKH XQFRQWUROOHG DGGLWLRQ RI ZDWHU WR WKH DVVHPEO\ DQG JURVV HUURU LQ WKH FDOFXODWLRQV DQGRU SURFHGXUHV 7KH RFFXUUHQFH RI VXFK DQ DFFLGHQW LV KLJKO\ LPSUREDEO\ DQG ZRXOG UHTXLUH D 6HWWLQJ D FRUH ZLGWK FRUUHVSRQGLQJ WR WKH PHWDOZDWHU UDWLR DQG SURFHHG WR LQVWDOO WKH IXHO VSDFHU V\VWHP IRU WKH PHWDOZDWHU UDWLR ORDG WKH IXHO XQGHU WKHVH FRQGLWLRQV DQG GLVUHJDUG VPDOO ZDWHU KHLJKW LQFUHPHQWV DQG 0 PHDVXUHPHQWV 7R GR WKLV VHYHUDO DGPLQLVWUDn WLYH UXOHV ZRXOG KDYH WR EH ZLOIXOO\ LJQRUHG DQGRU E ,PSURSHU VHWWLQJ RI WKH IROORZLQJ WKH ZHLU KHLJKW WKH ZDWHU KHLJKW VFUDP V\VWHP IRU DQ\ FRQILJXUDWLRQ DQG D JURVV HUURU LQ WKH FDOFXODWLRQV FDXVLQJ FULWLFDOLW\ DW DERXW KDOI WKH GHVLJQ FRUH KHLJKW FPf ,W VKRXOG EH QRWHG WKDW WKH FDOFXODWLRQDO PHWKRG XVHG WR SUHGLFW WKDW YDOXHV KDV EHHQ WHVWHG VXFFHVVIXOO\ DJDLQVW WKH UHVXOWV RI GLIIHUHQW FULWLFDO DVVHPEOLHV >@ DQGRU F $Q REVWUXFWHG LQFK OLQH IURP WKH FRUH WR WKH ZHLU GUDLQ V\VWHP DQG D VLPXOWDQHRXV IDLOXUH RI WKH ZDWHU KHLJKW VFUDP WRJHWKHU ZLWK WKH UHIHUUHG WR HUURU LQ WKH FDOFXODWLRQV 7RWDO GRVH PUHPf [ O2 [ [

PAGE 63

G 9LRODWLRQ E\ WKH IDFLOLW\ RSHUDWRU RI WKH DGPLQLVWUDWLYH SURFHGXUHV UHTXLULQJ D YLVXDO FKHFN RI WKH DVVHPEO\ EHIRUH VWDUWXS DQG FRQWLQXRXV DWWHQWLRQ WR WKH FRQWURO FRQVROH LQVWUXPHQWDWLRQ WR GHWHUPLQH WKH VWDWXV RI WKH DVVHPEO\ DW DOO WLPHV 7KH FRQVHTXHQFHV RI VXFK DQ DFFLGHQW ZHUH GHWHUPLQHG E\ DVVXPLQJ WKH IROORZLQJ 7KH DVVHPEO\ LV LQLWLDOO\ DW D VWHDG\ VWDWH SRZHU OHYHO RI ZDWWV QRUPDO DYHUDJH SRZHU LV ZDWWf :DWHU IORZV LQWR WKH V\VWHP DW WKH PD[LPXP UDWH RI JSP 7KH UHDFWLYLW\ DGGLWLRQ UDWH LV VHF 7KLV UDWH LV ODUJHU WKDQ WKH FDOFXODWHG UDWH VHFf IRU WKH FDVH GLVFXVVHG DERYH DQG PRUH WKDQ WKDW FDOFXODWHG WR RFFXU QHDU FULWLFDO IRU SURSHUO\ ORDGHG DVVHPEO\ 7KH SRZHU VFUDP LV VHW WKURXJK FDOLEUDWLRQ RU RWKHU HUURU DW NZ &RUUHVSRQGLQJ HUURU VHWWLQJV RFFXU IRU WKH QHXWURQ IOX[ DQG SHULRG VFUDP $ VFUDP RFFXUV VHF DIWHU D SRZHU OHYHO RI NZ LV UHDFKHG 7KLV WLPH KDV EHHQ GHWHUPLQHG DV WKH HODSVHG WLPH IURP WKH LQLWLDWLRQ RI D VFUDP VLJQDO WR WKH RSHQLQJ RI WKH GXPS YDOYHV 1R IHHGEDFN HIIHFWV DUH FRQVLGHUHG 7KLV LV D JRRG DSSUR[LPDn WLRQ WR RXU FDVH GXH WR WKH ORZ SRZHU OHYHOV LQYROYHG DQG DJDLQ LV D FRQVHUYDWLYH DVVXPSWLRQ 7KH FDOFXODWLRQV ZHUH PDGH XVLQJ WKH ,5(.,1 FRGH GHVFULEHG LQ UHIHUHQFH >@ ,5(.,1 QXPHULFDOO\ VROYHV WKH SRLQW PRGHO NLQHWLFV HTXDWLRQV 6WDUWLQJ ZLWK WKH DVVHPEO\ RQH GROODU VXEFULWLFDO NHA f DQG SURFHHGLQJ ZLWK WKH GHVFULEHG H[FXUVLRQ WKH DFFLGHQW ZRXOG \LHOG

PAGE 64

D SHDN SRZHU RI NZ WKH WRWDO HQHUJ\ UHOHDVH LV 0ZVHF 7KH SRZHU YV WLPH EHKDYLRU RI WKH DVVHPEO\ IRU WKH SRVWXODWHG SRZHU H[FXUn VLRQ LV VKRZQ LQ )LJ 7KH FRPELQHG QHXWURQ DQG JDPPD GRVH WR WKH RSHUDWRU LV UHP r DVVXPLQJ WKDW WKH HQHUJ\ UHOHDVH LV LQVWDQWDQHRXV DOO QHXWURQV KDYH DQ HQHUJ\ RI 0H9 DQG WKHUH LV QR DWWHQXDWLRQ LQ WKH DVVHPEO\ 7KH SURSHU 5%( IDFWRUV ZHUH WDNHQ LQWR FRQVLGHUDWLRQ ,W LV FRQFOXGHG WKDW HYHQ LI VXFK DQ LPSUREDEOH DFFLGHQW ZRXOG RFFXU WKH KD]DUGV WR SHUVRQQHO DQG WKH JHQHUDO SRSXODWLRQ DUH QRW VLJQLILFDQW

PAGE 65

),* 32:(5 YV 7,0( )25 7+( '(6,*1 %$6,6 $&&,'(17

PAGE 66

&+$37(5 ,9 7+( '$7$ $&48,6,7,21 6<67(0 ,QWURGXFWLRQ ,Q WKH GHOLFDWH DQG ODERULRXV WDVN RI SHUIRUPLQJ QXFOHDU H[SHULn PHQWV WKH PRVW FRPPRQ VRXUFH RI GLIILFXOWLHV DQG HUURUV OLHV LQ WKH DFTXLVLWLRQ RI WKH GDWD 0RGHUQ QXFOHDU LQVWUXPHQWDWLRQ ZLWK LWV H[FHOn OHQW WLPHHQHUJ\ UHVROXWLRQ KDV HQKDQFHG GHWHFWLRQ VHQVLWLYLW\ WR WKH H[WHQW WKDW GHYLDWLRQV SUHYLRXVO\ PDVNHG E\ WKH SRRUHU UHVROXWLRQ RI WKH HTXLSPHQW DUH HDVLO\ GLVWLQJXLVKDEOH 7KH PDMRU SUREOHP QRZ OLHV LQ WKH GHJUHH RI UHSURGXFLELOLW\ RI WKH UHVXOWV $OWKRXJK LQ JHQHUDO WKH LQn VWUXPHQWDWLRQ LV YHU\ UHOLDEOH WKH HQKDQFHG VHQVLWLYLW\ GHPDQGV FRQWLQn XDO VWDQGDUGL]DWLRQ IRU WKH VDNH RI UHSURGXFLELOLW\ 7KH EUDLQ RI WKH GDWD DFTXLVLWLRQ V\VWHP IRU SXOVHG QHXWURQ PHDVXUHPHQWV LV WKH PXOWLFKDQQHO DQDO\]HU 0&$f ZKLFK LV QRZ DYDLODEOH ZLWK D ODUJH QXPEHU RI GDWD FKDQQHOV DQG QDUURZHU FKDQQHO ZLGWKV IRU LQFUHDVHG WLPH UHVROXWLRQ ,Q JHQHUDO WKH PRGH RI GDWD DFTXLVLWLRQ IRU D SXOVLQJ H[SHULPHQW GLIIHUV IURP WKDW RI PDQ\ RWKHU W\SHV RI QXFOHDU H[SHULPHQWV ,Q SDUWLFXODU QHXWURQ LQWHUDFWLRQV ZLWK D VXLWDEOH GHWHFWRU DUH IHG LQWR WKH DQDO\]HU ZKLOH LW LV WLPHVZHHSLQJ ,GHDOO\ HYHU\ QHXWURQ LQWHUDFWLRQ VKRXOG EH FRXQWHG UHJDUGOHVV RI WKH DPSOLn WXGH RI WKH FROOHFWHG SXOVH 7KLV LPSOLHV WKDW WKH OLQHDU VLJQDO RULJLQDWHG DW WKH QHXWURQ GHWHFWRU VKRXOG EH FRQYHUWHG WR D ORJLF VLJQDO VR WKDW LWV SUREDELOLW\ RI EHLQJ UHFRUGHG LV LQGHSHQGHQW RI LWV DPSOLWXGH

PAGE 67

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n QLQJ WLPHV 7KH FRXQW UDWHV HQFRXQWHUHG GXULQJ WKH LQLWLDO H[SHULPHQWV ZLWK D +H FRXQWHU FORVH WR WKH QHXWURQ JHQHUDWRU WDUJHW LQ WKH 63(57 DVVHPEO\ H[FHHGHG A HYHQWVHF DW WKH SHDN 6DWXUDWLRQ HIIHFWV ZHUH REVHUYHG LQ PRGXODU LQVWUXPHQWDWLRQ LQFRUSRUDWLQJ WKH ODWHVW )(7 SUHDPSOLILHU DQG GRXEOHGHOD\ OLQH SXOVH VKDSLQJ DPSOLILHU DW DERXW r HYHQWVHF 7KH IDVW SXOVH WUDLQ FDXVHG D GURS LQ WKH EDVH YROWDJH RI WKH )LHOG (IIHFW 7UDQVLVWRU ZKLFK SUHYHQWHG IXUWKHU FRXQWLQJ XQWLO D VLJQLILFDQW UHGXFWLRQ LQ WKH FRXQW UDWH DOORZHG WKH YROWDJH WR UHFRYHU &KDQJLQJ WR D IDVW VFLQWLOODWLRQ SURWRW\SH SUHDPSOLILHU LPSURYHG WKH VLWXDWLRQ VRPHZKDW +RZHYHU LI WKH SRWHQWLDO FRXQW UDWH LQ WKH 8)6$ DVVHPEO\

PAGE 68

ZHUH HIILFLHQWO\ KDQGOHG ZLWK IHZ ORVVHV WKH DFWXDO UXQQLQJ WLPH RI D SDUWLFXODU H[SHULPHQW FRXOG EH UHGXFHG WR D IHZ PLQXWHV 7KLV SRVVLn ELOLW\ IRUFHG WKH GHYHORSPHQW RI WKH XOWUDIDVW LQVWUXPHQWDWLRQ FKDQn QHOV XVHG LQ WKLV ZRUN 7KH HOHFWURQLF FRXQWLQJ V\VWHPA XVHG WKURXJKRXW WKH H[SHULPHQWDWLRQ XVHG D WUDQVIRUPHUFRXSOHG SXOVH DPSOLILHU WR SURGXFH D IDVW ORJLF VLJn QDO WR WKH LQSXW RI WKH 0&$ IURP D VORZHU LQSXW VLJQDO RULJLQDWLQJ LQ D +H FRXQWHU 7KH V\VWHP KDV VXSHULRU FRXQWLQJ DQG VWDELOLW\ FKDUDFWHUn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

PAGE 69

SUHIHUUHG RYHU WKH %)A YDULHW\ DW WKH 8QLYHUVLW\ RI )ORULGD EHFDXVH RI WKH ODUJHU QHXWURQ DEVRUSWLRQ FURVV VHFWLRQ DQG RSHUDWLRQDO UHOLDELOLW\ +H XQGHUJRHV WKH IROORZLQJ UHDFWLRQ +HA Q +A S 0HY 7KH FURVV VHFWLRQ IRU WKLV UHDFWLRQ LV EDUQ DW Y f§ R PVHF FRPSDUHG WR EDUQ IRU % +H IROORZV D Y ODZ LQ WKH HQHUJ\ UDQJH IURP WR NHY 7KH SXOVH KHLJKWV \LHOGHG E\ D +H ILOOHG FRXQWHU DUH SURSRUWLRQDO WR WKH HQHUJ\ RI WKH QHXWURQV SOXV NHY 7KH UHDFWLRQ KDV EHHQ XVHG IRU QHXWURQ VSHFWURVFRS\ IURP WKH NHY WR 0HY HQHUJ\ UDQJH *DPPD GLVFULPLQDWLRQ FDQ EH HDVLO\ DFFRPn SOLVKHG E\ WKH XVH RI D ELDVHG LQWHJUDO GLVFULPLQDWRU 7KH GHWHFWRUV ZHUH ORQJ DQG WKLQ DQG HDFK WRRN WKH SODFH RI D IXHO HOHPHQW LQ WKH FRUH 7KH DFWLYH OHQJWK RI WKH FRXQWHU LV VOLJKWO\ OHVV WKDQ WKH DFWLYH OHQJWK RI WKH IXHO 7ZR RQH DWPRVSKHUH SUHGRPLQDQWO\ WKHUPDO GHWHFWLRQf DQG RQH WHQ DWPRVSKHUH PRUH UHVSRQVLYH WR KLJKHU HQHUJ\ QHXWURQVf GHWHFWRUV ZHUH XVHG LQ WKH H[SHULPHQWDO SURJUDP $ VNHWFK RI WKH SK\VLFDO FKDUDFWHULVWLFV RI WKH FRXQWHUV LV VKRZQ LQ )LJ 7KH WKLQ ORQJ F\OLQGULFDO VKDSH HQKDQFHV WKH WLPH FKDUDFWHU LVWLFV RI WKH FRXQWHU $OWKRXJK QRUPDOO\ D DWPRVSKHUH +H GHWHFWRU RSHUDWHV ZLWK D ELDV YROWDJH RI YROW WKH PLQLPXP LQSXW SXOVH YROWDJH UHTXLUHPHQW RI WKH SXOVH WUDQVIRUPHU ZDV VXFK WKDW WKH RSHUDWLQJ YROWDJH RI WKH FRXQWHUV KDG WR EH UDLVHG WR YROW $W WKLV YROWDJH WKH VORSH RI WKH FRXQWV YV KLJK YROWDJH FXUYH ZDV DERXW b SHU YROW WKHUHIRUH YHU\ VWDEOH ORZ ULSSOH KLJK YROWDJH VXSSOLHV ZHUH XVHG WR LQVXUH UHSURGXFLELOLW\ RI WKH GHWHFWRU UHVSRQVH

PAGE 70

),* 3+<6,&$/ &+$5$&7(5,67,&6 2) 7+( +H 1(87521 &2817(56

PAGE 71

7KH (OHFWURQLF ,QVWUXPHQWDWLRQ $ EORFN GLDJUDP RI WKH LQVWUXPHQWDWLRQ XVHG LQ WKH SXOVH SURSDJDn WLRQ H[SHULPHQWV LV VKRZQ LQ )LJV DQG 7ZR LQGHSHQGHQW GDWD DFTXLVLWLRQ V\VWHPV ZLWK D FRPPRQ VWDUWVWRS FORFN DUH QHFHVVDU\ WR FDUU\ RXW WKH PHDVXUHPHQWV f D V\VWHP FRQQHFWHG WR D PRYDEOH GHWHFn WRU WKDW REWDLQV WKH WLPH SURILOH RI WKH SURSDJDWLQJ SXOVH DW D JLYHQ SRVLWLRQ DQG f WKH DOO LPSRUWDQW QRUPDOL]LQJ GHWHFWRU IL[HG DW RQH SRVLWLRQ 7ZR DWPRVSKHUH +H GHWHFWRUV ZLWK WKH FKDUDFWHULVWLFV GHn VFULEHG DERYH ZHUH XVHG IRU WKHVH SXUSRVHV 7KH V\VWHP LV HVVHQWLDOO\ FRPSRVHG RI VLJQDO WUDQVPLWWLQJ GHYLFHV DQG GDWD UHJLVWHULQJ DQG KDQGOLQJ XQLWV 7KH PRYDEOH GHWHFWRU GDWD DFTXLVLWLRQ V\VWHP 0''$6f FRQVLVWV RI 'HWHFWRU +LJK 9ROWDJH 3RZHU 6XSSO\ 7KH KLJK YROWDJH SRZHU VXSSO\ ZDV DQ XOWUDVWDEOH )/8.( % ZLWK VXSHULRU VWDELOLW\ DQG QHJOLJLEOH ULSSOH 0DQXIDFWXUHUnV VSHFLILFDWLRQV VWDWH WKH VWDELOLW\ DW b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

PAGE 72

),* 029$%/( '(7(&725 '$7$ $&48,6,7,21 6<67(0

PAGE 73

),* 1250$/,=,1* '(7(&725 '$7$ $&48,6,7,21 6<67(0

PAGE 74

FRPSRQHQWV RI WKH GHWHFWRU VLJQDO ZLOO DFWXDWH D ZLGH EDQG WUDQVLVWRU DPSOLILHU DQG WXQQHO GLRGH GLVFULPLQDWRU IURP WKH VHFRQGDU\ RI WKH WUDQVIRUPHU $ OLQH GULYHU EXIIHU LV DOVR SURYLGHG 7KH SRZHU DQG FRQWURO ELDV DUH SURYLGHG E\ WKH 257(& 7LPH 3LFNRII &RQWURO 0RGLILHG 257(& 7LPH 3LFNRII&RQWURO 7KH 7LPH3LFNRII &RQWURO SURYLGHV FRQWURO DQG IDQRXW EXIn IHULQJ IRU WLPH GHULYDWLRQ XQLWV VXFK DV WKH 7LPH 3LFNRII 8QLW 7KH IDQRXW EXIIHU $FFHSWV WKH IDVW QHJDWLYH ORJLF VLJQDO IURP WKH XQLW DQG SURYLGHV HLWKHU IDVW QHJDWLYH RU VORZHU SRVLWLYH ORJLF RXWSXW VLJQDOV 7KH 7LPH3LFNRII &RQWURO KDG WR EH PRGLILHG WR PDNH LWV RXWn SXW VLJQDO FRPSDWLEOH ZLWK WKH LQSXW UHTXLUHPHQWV RI WKH 3XOVHG 1HXWURQ /RJLF 8QLW XVHG ZLWK WKH 0&$ 7KH SOXJLQ XQLW UHTXLUHV SXOVHV ZLWK ULVH WLPHV ORQJHU WKDQ QVHF ZKLOH WKH SRVLWLYH RXWSXW IURP WKH FRQWURO XQLW KDV D ULVH WLPH a QVHF )XUWKHUPRUH WKH SRVLWLYH ORJLF VLJQDO IURP WKH FRQWURO KDV D PLFURVHF SXOVH ZLGWK ZKLFK LV WRR ZLGH IRU WKH WLPH UHVROXWLRQ UHTXLUHG 7KHUHIRUH WZR PRGLILFDWLRQV ZHUH PDGH D FDSDFLWRU &f ZDV FKDQJHG IURP SI WR SI UHGXFLQJ WKH SXOVH ZLGWK WR PLFURVHF DQG DQ LQGXFWRU FRLO /,f ZDV UHPRYHG IURP WKH V\VWHP FKDQJLQJ WKH ULVH WLPH WR QVHF 7KH PRGLILFDWLRQV PDGH WKH V\VWHP FRPSDWLEOH ZLWK WKH LQSXW UHTXLUHPHQWV RI WKH 0&$ DQG DYRLGHG RYHUGULYLQJ WKH DQDO\]HU 5HIHU WR 257(& ,QF ,QVWUXFWLRQ 0DQXDO IRU 7LPH3LFNRII &RQWURO

PAGE 75

7HFKQLFDO 0HDVXUHPHQW &RUS 0XOWLFKDQQHO $QDO\]HU 0&$f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n XUHPHQWV 7KH VWRUDJH FDSDFLW\ RI WKH &1 LV A FRXQWV %RWK SULQWHG SDSHU WDSH DQG SHUIRUDWHG ELQDU\ WDSH ZHUH REWDLQHG DV RXWSXW 257(& 6FDOHU )RU VRPH RI WKH IOX[ WUDYHUVHV DQG WKH 0 PHDVXUHPHQWV WKH LQWHJUDO FRXQWV ZHUH UHFRUGHG LQ D 0K] VFDOHU 7KH QRUPDOL]LQJ GHWHFWRU GDWD DFTXLVLWLRQ V\VWHP 1''$6f FRQVLVWV RI WKH VDPH FRPSRQHQWV DV WKH PRYDEOH GHWHFWLRQ V\VWHP H[FHSW WKDW D PXOWLFKDQQHO DQDO\]HU 70& 0RGHO &1 ZDV XVHG IRU GDWD KDQGOLQJ DQG WKH LQWHJUDO FRXQWV ZHUH DFFXPXODWHG LQ DQ 257(& 6FDOHU PRGLn ILHG IRU 0K] FRXQWLQJ UDWH 7KH 5HVROXWLRQ 7LPH &RUUHFWLRQ 7KH UHVROXWLRQ WLPH RI WKH SXOVHG QHXWURQ GDWD DFTXLVLWLRQ V\VWHP ZDV WKH SDUDPHWHU XQGHU FRQVLGHUDWLRQ ZKLOH VHDUFKLQJ IRU D ZHOOPDWFKHG

PAGE 76

DQG IDVW HOHFWURQLF LQVWUXPHQWDWLRQ VHWXS 6RPH RI WKH ILQGLQJV ZHUH VXUSULVLQJ DQG PD\ H[SODLQ VRPH RI WKH GLVFUHSDQFLHV IRXQG LQ SUHYLRXV SXOVH SURSDJDWLRQ DQG QHXWURQ ZDYH H[SHULPHQWV ZKLFK XVHG DQ DFFHOHUn DWRU W\SH QHXWURQ VRXUFH ,Q WKHVH H[SHULPHQWV LW KDV EHHQ FXVWRPDU\ WR LQFUHDVH WKH QHXWURQ \LHOG RI WKH JHQHWDWRU DV WKH PRYDEOH GHWHFWRU LV SRVLWLRQHG IDUWKHU DZD\ IURP WKH VRXUFH WR NHHS UXQQLQJ WLPHV ZLWKLQ WROHUDEOH OLPLWV 7KH FRXQW UDWH DW WKH PRYDEOH GHWHFWRU LV WKHQ PDLQWDLQHG DW D OHYHO ZKLFK KDV EHHQ GHWHUPLQHG WR EH DFFHSWDEOH KRZn HYHU WKH QRUPDOL]LQJ GHWHFWRU XVXDOO\ IL[HG LQ D SRVLWLRQ FORVH WR WKH VRXUFH QRUPDOO\ DW WKH WKHUPDOL]LQJ WDQN LI RQH LV XVHGf PD\ WKHQ EH H[SRVHG WR FRXQW UDWHV EH\RQG WKH FDSDELOLWLHV RI WKH GHWHFWRU DQG WKH DVVRFLDWHG LQVWUXPHQWDWLRQ 6LQFH WKH PRYDEOH GHWHFWRU VLJQDO ZDV DOZD\V IHG LQWR DQ DQDO\]HU VDWXUDWLRQ HIIHFWV FRXOG EH HDVLO\ GHWHFWHG WKH VLJQDO IURP WKH QRUPDOL]LQJ FRXQWHU KRZHYHU ZDV DOZD\V EHLQJ IHG LQWR D VFDOHU ZKHUH VDWXUDWLRQ HIIHFWV PD\ JR XQQRWLFHG VLQFH D WLPH SURILOH LV QRW DYDLODEOH 'XULQJ PRVW RI WKH SUHVHQW ZRUN WZR 0&$nV ZHUH DYDLODEOH DQG WKH SKHQRPHQRQ GHVFULEHG DERYH ZDV REVHUYHG DQG FRXQW UDWHV WKURXJKRXW WKH H[SHULPHQWDO SURJUDP OLPLWHG WR WKRVH SHUPLVVLEOH E\ WKH WLPH FKDUDFn WHULVWLFV RI WKH FRXQWLQJ V\VWHP 7KH H[SHULPHQWDO SURJUDP EHLQJ FDUULHG RXW ZLWK WKH 8)6$ VXEFULW LFDO FDOOV IRU WKH RQH WR RQH FRPSDULVRQ EHWZHHQ WKHRUHWLFDOO\ FDOFXn ODWHG DQG H[SHULPHQWDOO\ GHWHUPLQHG WLPH SURILOHV RI WKH QHXWURQ IOX[ DV D IXQFWLRQ RI VSDFH 7KH VKDSH RI WKH SXOVH UDWKHU WKDQ FRQYHQn WLRQDO LQWHJUDO SDUDPHWHUV H[WUDFWHG IURP LW LV WKHUHIRUH WKH VLJQLIn LFDQW DQG UHTXLUHG LQIRUPDWLRQ )RU WKLV UHDVRQ WKH LQIOXHQFH RI WKH UHVROXWLRQ WLPH FRUUHFWLRQ RQ WKH VKDSH RI WKH SXOVH FDPH XQGHU FORVH

PAGE 77

VFUXWLQ\ 7KH UHVROXWLRQ WLPH RI VHYHUDO VXEPLFURVHFRQG SUHDPSOLILHUOLQHDU DPSOLILHU FRPELQDWLRQV ZDV VKRZQ WR EH D IXQFWLRQ RI WKH PRGH RI RSHUDn WLRQ DQG RI WKH FRXQW UDWH )RU WKLV W\SH RI LQVWUXPHQWDWLRQ WKHUH LV D VLJQLILFDQW GLIIHUHQFH EHWZHHQ WKH UHVROXWLRQ WLPH DV GHWHUPLQHG E\ D VWHDG\VWDWH WHFKQLTXH WZR VRXUFH PHWKRGf DQG D G\QDPLF PHWKRG PD[LPXP FRXQW UDWH PHWKRG >@f WKH ODWWHU PHWKRG \LHOGLQJ D PXFK ODUJHU UHVRn OXWLRQ WLPH )RU WKH LQVWUXPHQWDWLRQ ILQDOO\ FKRVHQ WR FDUU\ RXW WKH SXOVHG H[SHULPHQWV LQ WKLV ZRUN QR VLJQLILFDQW GLIIHUHQFH ZDV IRXQG EHWZHHQ WKH UHVXOWV RI WKH WZR PHWKRGV )XUWKHUPRUH WKH SXOVH WUDQVIRUPHU DPSOLILHU FRPELQDWLRQ EHKDYHV OLNH DQ LGHDO QRQSDUDO\]LQJ V\VWHP 7KH UHVROXWLRQ WLPH RI WKH RYHUDOO GDWD DFTXLVLWLRQ V\VWHP ZDV IRXQG WR EH SULPDULO\ GHWHUPLQHG E\ WKH PXOWLFKDQQHO DQDO\]HU $V D PDWWHU RI LOOXVWUDWLRQ ZKHQ WKH V\VWHP GHSLFWHG EHORZ ZDV XVHG WKH UHVROXWLRQ WLPH FKDQJHG IURP PLFURVHF DV GHWHUPLQHG XQGHU ORZ FRXQW UDWH VWHDG\VWDWH FRQGLWLRQV WR PLFURVHF DV GHWHUn PLQHG XQGHU KLJK FRXQW UDWH SXOVLQJ PRGH FRQGLWLRQV 257(& /LQHDU $PR 257(& 6FDOHU

PAGE 78

6KRZQ LQ )LJ DUH WZR WLPH SLRILOHV REWDLQHG DW WKH VDPH SRVLn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

PAGE 79

&RQYHQWLRQDO &KDQQHO 1XPEHU ,* 7,0( 352),/(6 2) 1(87521 %8567 5(&25'(' %< &219(17,21$/ (/(&7521,& ,167580(17$7,21 $1' %< 7+( 7,0(3,&.2)) 6<67(0

PAGE 80

RQ WKH PDJQLWXGH RI WKH UHVROYLQJ WLPH LW LV QRW QHFHVVDU\ WR SURSn HUO\ LGHQWLI\ WKH FRXQWLQJ V\VWHP VLQFH QR VLJQLILFDQW GLIIHUHQFH LQ WKH FRUUHFWHG FRXQWV DUH REVHUYHG ZKHQ XVLQJ WKH SDUDO\]LQJ RU QRQn SDUDO\]LQJ FRUUHFWLRQ 7KLV LV VKRZQ LQ )LJ 7KH UHVROXWLRQ WLPH FRUUHFWLRQ IRU WKH 8)6$ GDWD DFTXLVLWLRQ V\VWHP LV WKHQ JLYHQ E\ 1 1 7 1[WU f§ &:[7UV ZKHUH 1A WUXH QXPEHU RI HYHQWV LQ D JLYHQ FKDQQHO 1 REVHUYHG QXPEHU RI HYHQWV LQ D JLYHQ FKDQQHO WU UHVROXWLRQ WLPH RI WKH V\VWHP &: FKDQQHO ZLGWK 7UV QXPEHU RI VZHHSV RI WKH DQDO\]HU DQG IRU D QRQSDUDO\]LQJ V\VWHP WU UHFLSURFDO RI WKH PD[LPXP REVHUYDEOH FRXQW UDWH 8VLQJ WKLV PHWKRG WKH IROORZLQJ UHVROXWLRQ RU UHVROYLQJ WLPH IDQ WRUV ZHUH GHWHUPLQHG $f 0''$6 LQFOXGLQJ 0XOWLFKDQQHO $QDO\]HU WU PLFURVHF %f 0''$6 H[FOXGLQJ 0XOWLFKDQQHO $QDO\]HU WU PLFURVHF &f 1''$6 LQFOXGLQJ 0XOWLFKDQQHO $QDO\]HU WU PLFURVHF 'f 1''$6 H[FOXGLQJ 0XOWLFKDQQHO $QDO\]HU WU PLFURVHF

PAGE 81

2EVHUYHG &RXQW 5DWH ),* 7+( 3$5$/,=,1* 1213$5$/,=,1* 6<67(0 5(62/87,21 7,0( &255(&7,21 $6 $ )81&7,21 2) &2817 5$7(

PAGE 82

7KH 1RUPDOL]DWLRQ 7HFKQLTXH 7KH DQDO\VLV RI D SXOVH SURSDJDWLRQ H[SHULPHQW ZLOO \LHOG LQIRUPDn WLRQ RQ WKH YHORFLW\ RI SURSDJDWLRQ WKH DWWHQXDWLRQ DQG SXOVH VKDSH DV D IXQFWLRQ RI SRVLWLRQ RI D SURSDJDWLQJ GLVWXUEDQFH 7R GHWHUPLQH WKH DWWHQXDWLRQ RU UHODWLYH DPSOLWXGH RI WKH SXOVH DW GLIIHUHQW SRVLWLRQV LQ DQ DVVHPEO\ WKH GDWD PXVW EH QRUPDOL]HG WR D IL[HG UHIHUHQFH VR WKDW YDULDWLRQV LQ WKH VRXUFH VWUHQJWK GDWD DFTXLVLWLRQ WLPH HWF FDQ EH SURSHUO\ DFFRXQWHG IRU 7KH XVXDO WHFKQLTXH UHTXLUHV DFFXPXODWLQJ LQWHJUDO FRXQWV LQ D VFDOHU IRU HDFK PHDVXUHPHQW DQG UHGXFLQJ DOO PHDVn XUHPHQWV WR WKH IL[HG UHIHUHQFH DIWHUZDUGV $V ZDV PHQWLRQHG SUHYLRXVO\ WKH UHVROXWLRQ WLPH FRUUHFWLRQ FDQ KDYH D VLJQLILFDQW HIIHFW RQ WKH QRUn PDOL]LQJ IDFWRU VLQFH ZLGHO\ GLIIHULQJ FRXQW UDWHV DUH HPSOR\HG ,W LV H[WUHPHO\ KDUG LI QRW LPSRVVLEOH WR REWDLQ D SURSHU UHVROXWLRQ WLPH FRUUHFWLRQ EDVHG RQ LQWHJUDO FRXQWV GHWHUPLQHG IURP D WLPHYDU\LQJ FRXQW UDWH 7ZR PHWKRGV RI QRUPDOL]DWLRQ DQG WKHLU UHODWLYH PHULWV DUH GLVFXVVHG EHORZ 7KH ,QWHJUDO &RXQW 0HWKRG LV WKH QRUPDOO\ HPSOR\HG PHWKRG RI QRUn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

PAGE 83

7KH $QDO\]HU 0HWKRG LV LQWULQVLFDOO\ PRUH DFFXUDWH WKDQ WKH ,QWHJUDO &RXQW 0HWKRG EHFDXVH LW SHUPLWV DQ H[DFW FRUUHFWLRQ IRU UHVROXWLRQ WLPH ORVVHV E\ WKH XVH RI WKH H[SUHVVLRQ SUHYLRXVO\ JLYHQ DSSOLHG WR WKH UHFRUGHG WLPH SURILOH 7KH FRUUHFWLRQ IRU WKH LQWHJUDO FRXQWV LV RQ WKH RWKHU KDQG LQDFFXUDWH VLQFH WKH FRXQW UDWH LV FRQn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n OXWLRQ WLPH FDQ EH IRXQG E\ IRUFLQJ WKH QRUPDOL]LQJ UDWLR REWDLQHG IURP WZR UXQV E\ WKH ,QWHJUDO 0HWKRG WR PDWFK WKH QRUPDOL]LQJ UDWLR REWDLQHG IURP WKH $QDO\]HU 0HWKRG IRU WKH VDPH WZR UXQV E\ DGMXVWLQJ WKH UHVRn OXWLRQ WLPH FRUUHFWLRQ DSSOLHG WR WKH LQWHJUDO FRXQWV &RPPHQWV &HUWDLQ LQFRQVLVWHQFLHV LQ WKH UHVXOWV RI WKH ILUVW IHZ H[SHULn PHQWV SURPSWHG D FDUHIXO LQVSHFWLRQ RI WKH PXOWLFKDQQHO DQDO\]HU PRGXV RSHUDQGL )RU WKH VDNH RI FRPSOHWHQHVV WKH VLJQLILFDQW ILQGLQJV DUH OLVWHG EHORZ $f 2SHUDWLRQ ZLWK WKH PLFURVHF FKDQQHO ZLGWK VHOHFWHG E\ WKH VHWWLQJV RI WKH SOXJLQXQLWf SURYHV WR EH XQUHOLDEOH GXH WR LQVWDn ELOLWLHV LQ WKH FORFN DQG JDWLQJ FLUFXLWV &KDQQHO ZLGWKV RI

PAGE 84

PLFURVHF RU ORQJHU DUH VWDEOH %f 7KH DGGUHVV FXUUHQW VHWWLQJ RI WKH &1 LV H[WUHPHO\ FULWLn FDO PDUNHGO\ VR IRU KLJK FRXQW UDWHV &f $Q RSWLPXP SXOVH LQWR WKH DQDO\]HU VKRXOG KDYH D ULVH WLPH RI QDQRVHF D WRWDO ZLGWK RI PLFURVHF DQG DQ DPSOLWXGH RI YROW 'f $ERYH QRLVH OHYHO WKH GLVFULPLQDWRU VHWWLQJ RI WKH ORJLF XQLW EHFRPHV LUUHOHYDQW ZKHQ D FRQVWDQW SXOVH KHLJKW LV XVHG DV LQSXW (f 5HSURGXFLELOLW\ WHVWV ZHUH SHUIRUPHG RQ WKH DQDO\]HU ZLWK WKH QHXWURQ JHQHUDWRU LQ FRQWLQXRXV PRGH 7KH VWDWLVWLFDO DQDO\VLV RI WKH FKDQQHO FRXQWV JDYH b ZHUH OHVV WKDQ IURP WKH PHDQ b ZHUH EHWZHHQ DQG IURP WKH PHDQ b ZHUH EHWZHHQ DQG IURP WKH PHDQ 7KH V\VWHP LV VWDWLVWLFDOO\ ZHOOEHKDYHG

PAGE 85

&+$37(5 9 18&/($5 &$/,%5$7,21 2) 7+( 8)6$ 68%&5,7,&$/ ,QWURGXFWLRQ 7KH QXFOHDU FDOLEUDWLRQ RI WKH 8QLYHUVLW\ RI )ORULGD 63(57 $VVHPn EO\ ZDV SHUIRUPHG SULRU WR WKH VSDFHWLPH NLQHWLFV VWXGLHV 7KH FDOLn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e 7KHVH HIIHFWV DQG RWKHU LQWHUHVWLQJ NLQHWLF SKHQRPHQD LQYROYLQJ WKHVH EDVLF UHDFWRU SDUDPHWHUV ZHUH FRQn VLGHUHG ZRUWK\ RI IXUWKHU VWXG\ DQG ZHUH LQYHVWLJDWHG GXULQJ WKH PDLQ SDUW RI WKH UHVHDUFK 7KH\ DUH GLVFXVVHG LQ 3DUW &KDSWHU 9 ,Q WKLV VHFWLRQ WKH UHVXOWV SHUWLQHQW WR WKH QHFHVVDU\ FDOLEUDWLRQ RI WKH V\VWHP DUH JLYHQ 7KHRUHWLFDO 1RWHV 7KH ,QYHUVH 0XOWLSOLFDWLRQ 0HWKRG 8QGHU LGHDO FRQGLWLRQV XVXDOO\ PHW RQO\ LQ VPDOOIDVW DVVHPEOLHV WKH UHDFWLYLW\ FDQ EH UHSUHVHQWHG E\

PAGE 86

6Un !UKI rU Nn0 ZKHUH 0 LV WKH QHW QHXWURQ PXOWLSOLFDWLRQ LQ WKH DVVHPEO\ ZLWK D FHQWUDOO\ ORFDWHG VRXUFH ,Q SUDFWLFH WKH PXOWLSOLFDWLRQ LV REWDLQHG IURP WKH UDWLR RI PXOWLSOLHG WR XQPXOWLSOLHG FRXQWV ZLWK D FHQWUDOO\ ORFDWHG VRXUFH 7KH XQPXOWLSOLHG FRXQWV DUH REWDLQHG ZLWK WKH ILVVLOH PDWHULDO UHPRYHG DQG DOO RWKHU FRQGLWLRQV XQGLVWXUEHG ,Q ZDWHUn PRGHUDWHG FRUHV LW LV GLIILFXOW WR PDWFK QHXWURQ VSHFWUD IRU PXOWLSOLHG DQG XQPXOWLSOLHG FRXQWV DQG GHYLDWLRQV IURP WKH LGHDO 0 DUH WR EH H[n SHFWHG ,I SRVVLEOH D VHDUFK IRU GHWHFWRU ORFDWLRQV VKRXOG EH FRQn GXFWHG VR DV WR REWDLQ FXUYHV WKDW IROORZ WKH H[SHFWHG EHKDYLRU RI 0 (YHQ LI N FDQ QRW EH GLUHFWO\ LQIHUUHG IURP WKH 0 GHWHUPLQDWLRQ WKH FXUYH RI UHFLSURFDO FRXQW UDWH YV WKH SDUDPHWHU WKDW FRQWUROV UHDFn WLYLW\ IXHO ORDGLQJ RU PRGHUDWRU KHLJKW RU b FRQWURO URG ZLWKGUDZDOf LV D XVHIXO JXLGH IRU VDIHO\ DSSURDFKLQJ FULWLFDOLW\ LI D ZHOOEHKDYHG FXUYH FDQ EH REWDLQHG 7KH LQYHUVH PXOWLSOLFDWLRQ FXUYH FDQ EH REWDLQHG DV D IXQFWLRQ RI PRGHUDWRU KHLJKW E\ ILUVW REWDLQLQJ D VHULHV RI XQPXOWLSOLHG FRXQWV DW YDULRXV ZDWHU OHYHOV DQG WKH PXOWLSOLHG FRXQWV DV WKH ZDWHU OHYHO LV UDLVHG ZLWK WKH DVVHPEO\ RULJLQDOO\ DLUVSDFHG 6HQVLWLYLW\ WR JHRn PHWULFDO FRQILJXUDWLRQ VRXUFHGHWHFWRUZDWHU OHYHOf UHTXLUHV DQ HPSLUn LFDO GHWHUPLQDWLRQ RI ZHOOEHKDYHG GHWHFWRU SRVLWLRQV 5HDFWLYLW\ 0HDVXUHPHQWV E\ WKH 3XOVLQJ 7HFKQLTXH 7KH SXOVHGQHXWURQ WHFKQLTXH KDV EHHQ XVHG VXFFHVVIXOO\ IRU VHYHUDO \HDUV WR PHDVXUH UHDFWLYLW\ 7KH WUDQVLHQW QHXWURQ GHQVLW\ IROORZLQJ D EXUVW RI QHXWURQV LV XVHG WR GHWHUPLQH WKH UHDFWLYLW\ RI WKH V\VWHP E\ HLWKHU WKH 6LPPRQV DQG .LQJ PHWKRG >@ 6MRVWUDQGnV DUHD UDWLR PHWKRG

PAGE 87

>@ *R]DQLnV H[WUDSRODWHG DUHDUDWLR DQDO\VLV >@ RU WKH *DUHOLV 5XVVHO WHFKQLTXH >@ ,Q DOO WKHVH WHFKQLTXHV LW LV HVVHQWLDO WKDW D IXQGDPHQWDO VSDWLDO GLVWULEXWLRQ RI WKH QHXWURQV EH HVWDEOLVKHG IRU D FRUUHFW GHWHUPLQDWLRQ RI WKH GHFD\ FRQVWDQW DQG WKHUHIRUH WKH UHDFn WLYLW\ RI WKH V\VWHP 7KH 6LPPRQV DQG .LQJ PHWKRG HVWDEOLVKHG WKDW D YDOXH IRU WKH UHDFWLYLW\ FDQ EH REWDLQHG GLUHFWO\ LI D SURPSW IXQGDPHQWDO GHFD\ FRQn VWDQW FDQ EH PHDVXUHG DW GHOD\HG FULWLFDO 7KH YDOXH RI D DW GHOD\HG FULWLFDO GHWHUPLQHV %e DQG LI WKHVH SDUDPHWHUV DUH DVVXPHG FRQVWDQW RYHU WKH UHDFWLYLW\ UDQJH RI LQWHUHVW D YDOXH RI D FDQ EH REWDLQHG 7KH WHFKQLTXH KDV JLYHQ JRRG UHVXOWV XS WR VXEFULWLFDO LQ VPDOO PXOn WLSOLFDWLYH V\VWHPV 7KH PHWKRG VWURQJO\ GHSHQGV RQ EHLQJ DEOH WR HVWDEOLVK WKH SURPSW IXQGDPHQWDO GHFD\ PRGH LW VXIIHUV IURP WKH LQFRQn YHQLHQW QHFHVVLW\ RI D GHOD\HG FULWLFDO PHDVXUHPHQW DQG WKH DVVXPHG FRQVWDQF\ RI e WKURXJKRXW WKH UDQJHV RI UHDFWLYLW\ 7KH 6MRVWUDQG PHWKRG LPSURYHV WKH 6LPPRQV DQG .LQJ PHWKRG LQ WKDW WKH GHOD\HG FULWLFDO PHDVXUHPHQW LV QR ORQJHU QHFHVVDU\ EXW WKH UHVXOWV DUH VKDGRZHG E\ WKH VWURQJ LQIOXHQFH RI KLJKHU VSDWLDO KDUPRQLFV 7KH PHWKRG LV EDVHG RQ WKH SUHPLVH WKDW WKH LPSXOVH UHVSRQVH FXUYH RI WKH V\VWHP LV GRPLQDWHG E\ WKH SURPSW IXQGDPHQWDO PRGH *R]DQLnV WUHDWPHQW LV D VLJQLILFDQW LPSURYHPHQW RYHU 6MRVWUDQGnV PHWKRG *R]DQL SURSRVHG WKH H[WUDFWLRQ RI WKH IXQGDPHQWDO PRGH RI SURPSW QHXWURQ GHFD\ IURP WKH LPSXOVH UHVSRQVH FXUYH DQG WKH H[WUDSRODn WLRQ RI WKLV FXUYH WR ]HUR WLPH 7KH UHDFWLYLW\ LQ GROODUV FDQ EH IRXQG E\ LQWHJUDWLQJ XQGHU WKLV FXUYH WKH PHWKRG LV LQGHSHQGHQW RI WKH SUHVn HQFH RI KLJKHU SURPSW VSDWLDO PRGHV 7KH *DUHOLV5XVVHOO WHFKQLTXH VLPLODU WR *R]DQLnV H[WUDSRODWHG

PAGE 88

DUHDUDWLR PHWKRG LV RI SUDFWLFDO YDOXH EHFDXVH RI LWV LQWULQVLF HOLPLQDWLRQ RI WKH HIIHFW RI SURPSW KLJKHU KDUPRQLFV 7KLV PHWKRG ZKLFK ZDV XVHG LQ WKH SUHVHQW ZRUN ZDV SRVWXODWHG RULJLQDOO\ IRU D UHSHWLWLYHO\ SXOVHG ZLWK D GHOWD IXQFWLRQ VRXUFH LQ WLPHf EDUH PRQRHQHUJHWLF UHDFWRU EXW KDV SURYHQ WR EH RI EURDGHU DSSOLFDWLRQ *DUHOLV DQG 5XVVHOOSRVWXODWH WKDW IRU WKH FRQGLWLRQV VSHFLILHG DERYH 5 5 I1S H[S N_ef WfGW e1SGW 1M5 ZKHUH 1S SURPSW FRQWULEXWLRQ WR WKH QHXWURQ GHQVLW\n GHOD\HG FRQWULEXWLRQ WR WKH QHXWURQ GHQVLW\ 5 SXOVLQJ UDWH 7KH IROORZLQJ FRQGLWLRQV VKRXOG EH VDWLVILHG IRU WKH FRUUHFW DSn SOLFDWLRQ RI WKH PHWKRG Df 5!!; ZKHUH ; LV WKH GHFD\ FRQVWDQW RI WKH VKRUWHVW OLYHG SUHFXUVRU JURXS Ef 5 } D ZKHUH D LV WKH SURPSW IXQGDPHQWDO GHFD\ FRQVWDQW Ff 7KH V\VWHP PXVW EH SXOVHG D VXIILFLHQW QXPEHU RI WLPHV VR WKDW H[S PDJQ5f H[S DJQ5f} ZKHUH P LV WKH WRWDO QXPEHU RI SXOVHV DQG WKH DJQ DUH WKH URRWV RI WKH LQKRXU HTXDWLRQ Gf 7KH SURPSW URRW GRPLQDWHV WKH GHFD\ 7KH *DUHOLV5XVVH8WUHDWPHQW SHUPLWV WKH GHWHUPLQDWLRQ RI S f ZKHQ DOO WKH DERYH FRQGLWLRQV DUH VDWLVILHG E\ WKH UHODWLRQ f N He $Q DEVROXWH YDOXH RI S LV REWDLQHG E\ WKH XVH RI D FDOFXODWHG HIIHFWLYH GHOD\HG QHXWURQ IUDFWLRQ *DUHOLV KDV GLVFXVVHG WKH XVH RI WKH PHWKRG LQ UHIOHFWHG V\VWHPV WKH WHFKQLTXH VHHPV WR EH RI SUDFWLFDO YDOXH LQ

PAGE 89

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n OLVKHG EHIRUH WKH SXOVH KDV GHFD\HG DZD\ VLQFH WKH DV\PSWRWLF PRGH LV RQH WKDW LV XQLIRUP HYHU\ZKHUH LQ VSDFH 7KH\ SURFHHG WR DSSO\ D FRPELQDWLRQ RI ILUVW IOLJKW DJH DQG WLPHGHSHQGHQW GLIIXVLRQ WKHRU\ WR WKH VWXG\ RI SXOVHG PHDVXUHPHQWV LQ ODUJH DTXHRXV PHGLD WKHLU FRQn FOXVLRQ LV WKDW WKHLU PHDVXUHG DSSDUHQW GHFD\ FRQVWDQW LV D JRRG DSn SUR[LPDWLRQ WR WKH DV\PSWRWLF YDOXH DQG WKDW SXOVHG PHDVXUHPHQWV LQ YHU\ ODUJH PXOWLSO\LQJ V\VWHPV PD\ DOVR JLYH JRRG UHVXOWV ,QYHUVH 0XOWLSOLFDWLRQ 0HDVXUHPHQWV 7KH VDIH DSSURDFK WR WKH GHVLJQ YDOXH RI NAA A ZDV XQGHUWDNHQ ZLWK WKH FRQYHQWLRQDO 0 PHDVXUHPHQWV XQWLO N f± a DQG WKHQ E\ ERWK HII WKH 0 DQG WKH SXOVLQJ WHFKQLTXH 7R HVWDEOLVK GHWHFWRU SRVLWLRQV IUHH IURP JHRPHWULFDO HIIHFWV VRXUFHGHWHFWRUZDWHU OHYHOf VL[ GLIIHUHQW ORFDWLRQV ZHUH XVHG XQWLO D ZDWHU OHYHO RI FP N a f ZDV UHDFKHG DQG IRXU ORFDWLRQV DIWHUZDUGV 7ZR RI WKH GHWHFWRU ORFDWLRQV WKH FORVHVW WR WKH QHXWURQ VRXUFH IDLOHG WR GHVFULEH WKH PXOWLSOLFDWLRQ RI WKH V\VWHP 6KRZQ LQ )LJ DUH WKH

PAGE 90

*5,' 326,7,21 :(,56 ),* '(7(&725 326,7,21,1* 6&+(0(

PAGE 91

ORFDWLRQV HPSOR\HG IRU WKHVH PHDVXUHPHQWV DQG IRU WKH DEVROXWH GHWHUPLQDWLRQ E\ SXOVLQJ 7ZR FX FHQWUDOO\ ORFDWHG 3X%H VRXUFHV ZHUH XVHG IRU WKHVH PHDVXUHPHQWV 7KH PHDVXUHPHQWV ZHUH SHUIRUPHG LQ WKH IROORZLQJ VHTXHQFH Df 8QPXOWLSOLHG FRXQWV YV ZDWHU OHYHO ZHUH REWDLQHG IURP D ZDWHU OHYHO RI FP WR FP DERYH WKH ERWWRP RI WKH DFWLYH IXHO LQ VWHSV RI FP Ef 7KH IXHO ZDV ORDGHG LQ WKH DVVHPEO\ LQ WKH SUHVHQFH RI WZR FHQWUDOO\ ORFDWHG 3X%H VRXUFHV ZLWK SURSHU PRQLWRULQJ Ff 0XOWLSOLHG FRXQWV YV ZDWHU OHYHO ZHUH REWDLQHG IURP D ZDWHU OHYHO RI FP LQ FP VWHSV XQWLO DQ HIIHFWLYH PXOWLSOLFDWLRQ IDFWRU B ZDV UHDFKHG 7KLV SURFHGXUH IROORZV WKH FULWHULD HVWDEOLVKHG IRU ,QLWLDO /RDGLQJ RI WKH DVVHPEO\ 3DUW &KDSWHU ,,, RI WKLV ZRUN WKH FP VWHSV ZHUH PRUH FRQVHUYDWLYH LQFUHPHQWV WKDQ WKRVH VSHFLILHG IRU WKH ,QLWLDO /RDGLQJ RI WKH DVVHPEO\ 7KH UHVXOWV RI WKHVH PHDVXUHPHQWV DUH VKRZQ LQ )LJV $ %f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

PAGE 92

,QYHUVH 0XOWLSOLFDWLRQ 0RGHUDWRU /HYHO FPf ),* $ ,19(56( 08/7,3/,&$7,21 YV 02'(5$725 /(9(/

PAGE 93

,QYHUVH 0XOWLSOLFDWLRQ 0RGHUDWRU /HYHO FPf ),* % ,19(56( 08/7,3/,&$7,21 YV 648$5(' ,19(56( +(,*+7

PAGE 94

WKH PXOWLSOLFDWLRQ 6KRZQ LQ 7DEOH ,,, DUH WKH YDOXHV HVWLPDWHG IRU N IRU WKH IRXU HII ORFDWLRQV WKDW VHHPHG WR UHSUHVHQW WKH V\VWHP EHVW 3RVLWLRQ JLYHV D ORZHU OLPLW DQG 3RVLWLRQ DQ XSSHU ERXQG 1R DWWHPSW ZDV PDGH WR HVWDEOLVK WKH HUURU DVVRFLDWHG ZLWK PHDVXUHG N EXW LW LV EHOLHYHG HII WKDW WKH YDOXH REWDLQHG DW WKH ODVW ZDWHU OHYHO LV ZLWKLQ RI WKH WUXH YDOXH $EVROXWH 'HWHUPLQDWLRQ RI NHIe $IWHU DQ HVWLPDWHG YDOXH RI N f± ! ZDV REWDLQHG IURP WKH 0 HIW f§ PHDVXUHPHQWV DQ LQGHSHQGHQW GHWHUPLQDWLRQ RI N ZDV UHTXLUHG E\ WKH RSHUDWLQJ OLFHQVH DW HYHU\ QHZ LQFUHPHQW LQ WKH PRGHUDWRU OHYHO DV GHWHUPLQHG E\ WKH FULWHULD HVWDEOLVKHG LQ 3DUW &KDSWHU f 7KH WHFKQLTXH FKRVHQ IRU WKLV GHWHUPLQDWLRQ ZDV WKH *DUHOLV5XVVHOO PHWKRG RI PHDVXULQJ Nt DQG D VLPXOWDQHRXV GHWHUPLQDWLRQ RI WKH SURPSW IXQGDn PHQWDO GHFD\ FRQVWDQW 'LIIHUHQW GHWHFWRU ORFDWLRQV ZHUH XVHG WR GHWHUPLQH WKH LQIOXHQFH RI WKH VRXUFH DQG RI KLJKHU RUGHU KDUPRQLFV FRQWDPLQDWLRQ 6WURQJ VSDWLDO HIIHFWV ZHUH REVHUYHG LQ ERWK D DQG Ne 7KLV SKHQRPHQD ZLOO EH GLVFXVVHG LQ GHWDLO LQ 3DUW &KDSWHU 9 EHFDXVH RI LWV LPSRUWDQFH $ VHHPLQJO\ WUXH IXQGDPHQWDO GHFD\ FRQVWDQW DQG VSDFHGFRQYHUJHG N6L ZHUH REWDLQHG DW ODUJH GLVWDQFHV IURP WKH VRXUFH DQG ZHUH XVHG WR GHWHUPLQH WKH UHDFWLYLW\ RI WKH V\VWHP $ )RUWUDQ ,9 ,%0 FRPSXWHU SURJUDP QDPHG 81,38/ ZDV FRGHG WR SHUIRUP D XQLILHG DQDO\VLV RI WKH SXOVHG QHXWURQ GDWD VHH $SSHQGL[ %f 7KH SURJUDP FDOFXODWHV WKH GHFD\ FRQVWDQW XVLQJ 3HLHUOnV VWDWLVWLFDO DQDO\VLV >@ DQG Ne XVLQJ WKH *DUHOLV5XVVHOODSSURDFK DIWHU WKH GDWD

PAGE 95

7$%/( ,,, 6800$5< 2) 0 $1' 38/6(' 0($685(0(176 8)6$ 5 &RUH 0: 5DWLR FP ZLGH UHIOHFWHG FRUH ,19(56( 08/7,3/,&$7,21 38/6(' (;3 35(',&7(' 0RGHUDWRU NHII NIIE NHII +HLJKW FPf 3RV 3RV 3RV 3RV $ERYH ERWWRP RI DFWLYH IXHO $YHUDJHG IURP GHWHFWRU SRVLWLRQV

PAGE 96

KDV EHHQ UHVROXWLRQ WLPH FRUUHFWHG DQG EDFNJURXQG VXEWUDFWHG $ SXUH GHOD\HG QHXWURQ EDFNJURXQG LV VWDWLVWLFDOO\ FDOFXODWHG DQG XVHG WR GHWHUPLQH Ne 7KH GDWD FDQ EH QRUPDOL]HG WR D UHIHUHQFH GHWHFWRU SRVLWLRQ IRU WKH DQDO\VLV RI WKH SXOVH SURSDJDWLRQ PHDVXUHPHQWV LQ WKH WLPH DQG LQ WKH IUHTXHQF\ GRPDLQ $ )RXULHU DQDO\VLV RI WKH SXOVH FDQ DOVR EH SHUIRUPHG LI UHTXLUHG 6KRZQ LQ )LJ $ %f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n LQJIXO YDOXHV IRU WKH UHDFWLYLW\ FDQ EH REWDLQHG ZKHQ DSSO\LQJ WKH

PAGE 97

VHF 0RGHUDWRU /HYHO FPf ),* $ '(&$< &2167$17 YV 02'(5$725 /(9(/

PAGE 98

NJe VHF ),* % NJL $1' N YV 02'(5$725 /(9(/

PAGE 99

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e-O LV DIIHFWHG E\ WKH LQSXW SXOVH ZLGWK $V PHQWLRQHG SUHYLn RXVO\ D PRUH GHWDLOHG DQDO\VLV RI WKH SXOVHG QHXWURQ UHDFWLYLW\ PHDVn XUHPHQWV LV FRQGXFWHG ODWHU LQ WKLV WKHVLV

PAGE 100

3$57 63$&(7,0( 5($&725 .,1(7,&6 678',(6 :,7+ 7+( 81,9(56,7< 2) )/25,'$ 63(57 $66(0%/<

PAGE 101

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f 7KH ZRUN DW )ORULGD XVLQJ FRPSDUDWLYHO\ QDUURZ SXOVHV IURP D QHXWURQ JHQHUDWRU DQG WKHUPDOL]LQJ WDQNf ZDV DOVR UHVWULFWHG E\ WKH QHXWURQLF VL]H RI WKH V\VWHP DQG E\ WKH N 1R FOHDQ GHWHUPLQDWLRQ RI FRPSOHWH VSDWLDO HIIHFWV LQ PXOWLSOLFDWLYH PHGLD KDV EHHQ UHSRUWHG 7KH 5HDFWRU 6DIHW\ 'LYLVLRQ RI WKH 8QLWHG 6WDWHV $WRPLF (QHUJ\ &RPPLVVLRQ HVWDEOLVKHG WKH /DUJH &RUH '\QDPLFV ([SHULPHQWDO 3URJUDP DV SURSRVHG E\ WKH $WRPLF (QHUJ\ 'LYLVLRQ RI WKH 3KLOOLSV 3HWUROHXP &RPSDQ\ WR VWXG\ WKH EHKDYLRU RI ODUJH UHDFWRU FRUHV 7KH REMHFWLYH

PAGE 102

LV WKH H[SHULPHQWDO YHULILFDWLRQ RI SUHVHQW VSDFHWLPH NLQHWLFV PRGHOV DQG WR JXLGH WKH GHYHORSPHQW RI QHZ FDOFXODWLRQDO WHFKQLTXHV 3KDVH , RI WKH /DUJH &RUH '\QDPLFV ([SHULPHQWDO 3URJUDP LV EHLQJ FRQGXFWHG DW WKH 8QLYHUVLW\ RI )ORULGD 63(57 $VVHPEO\ XQGHU 6XEFRQn WUDFWV & DQG & ZLWK WKH 3KLOOLSV 3HWUROHXP &R 7KLV SDUW RI WKH SURJUDP ZLOO SURYLGH WKH PRVW IXQGDPHQWDO QHXWURQ SK\VLFV W\SH RI LQn IRUPDWLRQ DV WKH UHGLVWULEXWLRQ FKDUDFWHULVWLFV RI WKH QHXWURQ IOX[ LQ D FORVHWRFULWLFDO DVVHPEO\ DUH VWXGLHG LQ WKH DEVHQFH RI LQKHUHQW IHHGEDFN HIIHFWV 7KH DGHTXDF\ RI EDVLF SDUWV RI WKH DQDO\WLFDO PRGHOV FDQ EH WHVWHG DFFXUDWHO\ E\ FRQFHQWUDWLQJ RQ WKH QHXWURQ SK\VLFV SUREOHP ,Q WKLV ZRUN WKH SURSDJDWLRQ RI D QDUURZ QRQDV\PSWRWLF QHXWURQ EXUVW LV VWXGLHG LQ WKH WLPH DQG LQ WKH IUHTXHQF\ GRPDLQ 7KH H[SHULn PHQWDO UHVXOWV DUH FRPSDUHG ZLWK UHVXOWV REWDLQHG IURP WKH WZRJURXS VSDFHWLPH GHSHQGHQW RQHGLPHQVLRQDO GLIIXVLRQ WKHRU\ FDOFXODWLRQDO VFKHPH NQRZQ DV WKH :,*/( SURJUDP >@ $ VWULQJHQW WHVW RI WKH PRGHO LV SURYLGHG E\ D FRPELQHG DQDO\VLV LQ WKH WLPH DQG IUHTXHQF\ GRPDLQ &OHDQ XQDPELJXRXV H[SHULPHQWDO LQIRUPDWLRQ ZLOO EH VKRZQ GHPRQVWUDWLQJ WKH SUHVHQFH RI VSDWLDO HIIHFWV LQ ODUJH FRUHV 'HVFULSWLRQ RI WKH 6WXG\ 3XOVH SURSDJDWLRQ SKHQRPHQD LQ D ODUJHLQRQHVSDFH GLPHQVLRQ VLGH UHIOHFWHG FRUH ZHUH VWXGLHG E\ LQWURGXFLQJ D IDVW QHXWURQ EXUVW WKURXJK RQH IDFH RI WKH DVVHPEO\ $ &RFNFURIW:DOWRQ 71&f W\SH JHQHUDWRU ZDV HPSOR\HG WR SURGXFH VTXDUH QHXWURQ SXOVHV RI VHOHFWHG ZLGWKV XVLQJ WKH GWf QHXWURQ UHDFWLRQ 7KH DVVHPEO\ ZKLFK ZDV VWXGLHG LV GHVFULEHG LQ GHWDLO LQ 3DUW

PAGE 103

RI WKLV ZRUN 7KH VXEFULWLFDO DVVHPEO\ FRQVLVWV RI D VTXDUH DUUD\ RI b HQULFKHG 8&A 63(57 )O IXHO HOHPHQWV PRGHUDWHG E\ OLJKW ZDWHU DQG ZLWK HVVHQWLDOO\ DQ LQILQLWH OLJKW ZDWHU UHIOHFWRU RQ WKH VLGHV 7KH FRUH LV FP ORQJ FP ZLGH KDV DQ HIIHFWLYH KHLJKW RI FP ZLWK FP ZLGH VLGH UHIOHFWRUV 7KH HIIHFWLYH PXOWLSOLFDWLRQ FRQn VWDQW RI WKH DVVHPEO\ KDV EHHQ GHWHUPLQHG WR EH $ IDVW UHOLDEOH VHW RI HOHFWURQLF LQVWUXPHQWDWLRQ ZDV GHYHORSHG DQG XVHG WR UHFRUG WKH YHU\ KLJK WLPHYDU\LQJ FRXQW UDWH DV D IXQFWLRQ RI SRVLWLRQ LQ WKH FRUH RI WKH DVVHPEO\ $ WHFKQLTXH ZDV GHYHORSHG WR VXEWUDFW WKH FRQWULEXWLRQ RI HSLFDGPLXP QHXWURQV IURP WKH UHFRUGHG WLPH SURILOH RI WKH QHXWURQ IOX[ 7KH VWXG\ RI WKH SXOVH SURSDJDWLRQ SKHQRPHQD ZDV FRQGXFWHG LQ D WZRIROG PDQQHU LQ WKH WLPH DQG LQ WKH IUHTXHQF\ GRPDLQ 7KH VWXG\ LQ WKH WLPH GRPDLQ FRQVLVWHG RI VHYHUDO VHFWLRQV IRFXVLQJ WKH DWWHQWLRQ RQ EDVLF DVSHFWV RI WKH VSDFH DQG WLPH GHSHQGHQFH RI SURSDJDWLQJ GLVn WXUEDQFHV 7KH IROORZLQJ DVSHFWV ZHUH LQYHVWLJDWHG Df 7KH SURSDJDWLRQ RI DQG PVHF ZLGH LQSXW SXOVH LQWURn GXFHG DW RQH HQG RI WKH DVVHPEO\ 7KLV LQYHVWLJDWLRQ FRQVWLWXWHG WKH PDLQ SDUW RI WKH UHVHDUFK LW LQYROYHG D GHWDLOHG FRPSDULVRQ RI WKH H[SHULPHQWDO UHVXOWV ZLWK WKH SUHGLFWLRQV RI WKH :,*/( FDOFXODWLRQDO VFKHPH 7KH VHQVLWLYLW\ RI WKH RQHGLPHQVLRQDO PRGHO WR VPDOO YDULDn WLRQV LQ WKH WUDQVYHUVH OHDNDJH ZDV DOVR LQYHVWLJDWHG Ef 6WDWLF IOX[ WUDYHUVHV ZHUH FRQGXFWHG WR GHWHUPLQH WKH VWHDG\ VWDWH QHXWURQ GLVWULEXWLRQ LQ WKH DV\PSWRWLF UHJLRQ '\QDPLF IOX[ WUDYHUVHV ZHUH SHUIRUPHG WR GHWHUPLQH ZKHWKHU DQ\ SURSDJDWLRQ RFFXUV LQ WKH WUDQVYHUVH GLUHFWLRQ Ff 7KH HIIHFW WKDW WKH LQSXW SXOVH ZLGWK KDV RQ WKH SURSDJDWLRQ

PAGE 104

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n VLGHUDWLRQ ,Q WKLV ZRUN DOO GHOD\ WLPHV DUH UHIHUUHG WR ]HUR WLPH WKLV FRUUHVSRQGV SK\VLFDOO\ WR WKH LQLWLDWLRQ RI WKH SXOVH DW WKH QHXWURQ JHQHUDWRU 3URSDJDWLRQ 7LPH 7KH SURSDJDWLRQ WLPH RI D QDUURZ SXOVH LQ WKH DVVHPEO\ LV GHILQHG DV WKH GHOD\ WLPH EHWZHHQ WZR H[WUHPH SRVLWLRQV LQ WKH DVVHPEO\

PAGE 105

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e LV GHWHUPLQHG IURP WKH SKDVH DQJOHV RI WKH ]HURWK )RXULHU PRPHQW E\ QXPHULFDO WUDQVIRUPDn WLRQ IRU HDFK IUHTXHQF\ RI LQWHUHVW 7KH GDPSLQJ FRHIILFLHQW DQG WKH SKDVH VKLIW SHU XQLW OHQJWK RI SDWK DUH WKH UHDO DQG WKH LPDJLQDU\ FRPSRQHQWV RI WKH FRPSOH[ LQYHUVH UHOD[DWLRQ OHQJWK S UHVSHFWLYHO\ ,W VKRXOG EH QRWHG WKDW LQ WKLV ZRUN SXOVH DQG ZDYH SURSDJDWLRQ IURP WKH FRQYHQWLRQDO YLHZSRLQW LH SURSDJDWLRQ RI D UHVLGXDO GLVWXUEDQFH UDWKHU WKDQ WKH SURSDJDWLRQ RI D SXOVH RU ZDYH IURQW LV EHLQJ VWXGLHG

PAGE 106

&+$37(5 ,, 7+(25(7,&$/ 127(6 ,QWURGXFWLRQ 7KH PDLQ FRQFHUQ RI WKLV ZRUN KDV EHHQ WKH H[SHULPHQWDO LQYHVWLJDn WLRQ RI VSDFHWLPH NLQHWLFV HIIHFWV RQ D ODUJH FRUH VXEMHFWHG WR D SHUWXUEDWLRQ WKH EDVLF DVSHFWV RI WKH QHXWURQ SK\VLFV SKHQRPHQD ZHUH VWXGLHG 7KH DQDO\VLV RI WKH H[SHULPHQW KDV EHHQ FRQGXFWHG LQ ERWK WKH WLPH DQG WKH IUHTXHQF\ GRPDLQ 7KH WKHRUHWLFDO PRGHO XVHG IRU FRPSDULVRQ ZLWK H[SHULPHQW LQ WKH WLPH DQG WKH IUHTXHQF\ GRPDLQ ZDV WKH WZRJURXS VSDFHWLPH GHSHQGHQW GLIIXVLRQ HTXDWLRQV QXPHULFDOO\ VROYHG E\ WKH :,*/( FRGH >@ $ RQHWRRQH FRPSDULVRQ RI WKHRU\ DQG H[SHULPHQW ZDV SHUIRUPHG LQ WKH WLPH GRPDLQ )RU WKH FRUUHVSRQGLQJ VWXG\ LQ WKH IUHTXHQF\ GRPDLQ ERWK WKH :,*/( DQG H[SHULPHQWDO WLPH SURILOHV DV D IXQFWLRQ RI SRVLWLRQ ZHUH DQDO\]HG LQWR LWV ZDYH FRPSRn QHQWV E\ D QXPHULFDO )RXULHU WUDQVIRUPDWLRQ WKLV PHWKRG ZDV VXJJHVWHG E\ 0RRUH >@ DQG FRQILUPHG E\ %RRWK >@ ,Q WKLV ZRUN RQO\ WKH ]HURWK )RXULHU PRPHQW LV DQDO\]HG WKH DPSOLWXGH DQG SKDVH DQJOH RI WKH ]HURWK )RXULHU PRPHQW FRUUHVSRQG WR WKRVH PHDVXUHG E\ WKH FRQYHQn WLRQDO QHXWURQ ZDYH H[SHULPHQW 5HYLHZ RI WKH /LWHUDWXUH $ FRQVLGHUDEOH DPRXQW RI WKHRUHWLFDO HIIRUW KDV EHHQ GHYRWHG WR WKH VSDFHWLPH NLQHWLFV SUREOHP LQ QXFOHDU UHDFWRUV 7KH PRUH LPSRUWDQW

PAGE 107

PHWKRGV SUHVHQWO\ LQ XVH DUH Df 'LUHFW VROXWLRQ RI WKH PXOWLJURXS VSDFHWLPH GLIIXVLRQ HTXDWLRQV UHIHUUHG WR DV WKH H[DFW PHWKRG 7KLV PHWKRG FRQVWLWXWHV D EHQFKPDUN IRU WKH RWKHU WHFKQLTXHV 7KH PRVW SURPLQHQW RI WKHVH FDOFXODWLRQDO VFKHPHV DUH f :,*/( D WZRJURXS RQHGLPHQVLRQDO VSDFHWLPH GLIIXVLRQ WKHRU\ FRPSXWHU SURJUDP > @ f 7:,*/( D WZRGLPHQVLRQDO YHUVLRQ RI :,*/( >@ f )5($. D IDVW UHDFWRU PXOWLJURXS NLQHWLFV FRGH >@ f $ IRXUJURXS RQHGLPHQVLRQDO VSDFHWLPH GLIIXVLRQ WKHRU\ FRPSXWHU SURJUDP >@ GHYHORSHG E\ WKH 3KLOOLSV 3HWUROHXP &R Ef )OX[ 6\QWKHVLV PHWKRGV ZKLFK XVHG WKH LGHD WKDW WKH IOX[ VKDSHV ZKLFK RFFXU GXULQJ D WUDQVLHQW FDQ EH EUDFNHWHG E\ D VHW RI VKDSH IXQFWLRQV > @ Ff &RQYHQWLRQDO 0RGDO PHWKRGV H[HPSOLILHG E\ WKH )RGHUDUR *DUDEHGLDQ WHFKQLTXH >@ LQ ZKLFK WKH IOX[ LV H[SDQGHG LQ WHUPV RI WKH HLJHQIXQFWLRQV RI WKH ZDYH HTXDWLRQ +HOPKROW] PRGHVf Gf 7KH 4XDVLVWDWLF DSSURDFK RI 2WW > @ D PRUH VRSKLVWLFDWHG IDFWRUL]LQJ WHFKQLTXH WKH IOX[ LV IDFWRUL]HG LQWR DQ DPSOLWXGH DQG D VKDSH IXQFWLRQf 8VH LV PDGH RI WKH IDFW WKDW WKH WLPH GHSHQGHQFH RI WKH VKDSH IXQFWLRQ LV OHVV LPSRUWDQW WKDQ WKH WLPH GHSHQGHQFH RI WKH DPSOLWXGH IXQFWLRQ Hf 7KH $GLDEDWLF DSSUR[LPDWLRQ >@ D IDFWRUL]LQJ WHFKQLTXH LQ ZKLFK EHVLGHV LJQRULQJ WKH WLPH GHSHQGHQFH RI WKH VKDSH IXQFWLRQ QR GLVWLQFWLRQ LV PDGH EHWZHHQ WKH SURPSW DQG GHOD\HG QHXWURQ VRXUFHV 2WKHU UHODWHG VWXGLHV RQ D VRPHZKDW GLIIHUHQW FRQWH[WZHUH SHUIRUPHG E\ .\VWUD DQG 8KULJ >@ DQG E\ .DYLSXUDSX >@ 7KH\ VWXGLHG WKH

PAGE 108

VSDFH GHSHQGHQW UHDFWRU WUDQVIHU IXQFWLRQ ZLWKLQ WKH IUDPHZRUN RI D FRPELQDWLRQ RI WLPHGHSHQGHQW )HUPL$JH DQG GLIIXVLRQ WKRULHV $ SDUDOOHO ILHOG RI VWXG\ KDV EHHQ WKH QHXWURQ ZDYH WHFKQLTXH > @ $ PDMRU LPSURYHPHQW LQ WKH QHXWURQ ZDYH ILHOG ZDV DFKLHYHG ZKHQ 0RRUH SURYHG >@ WKDW WKH SXOVHG WHFKQLTXH LV HTXLYDOHQW WR WKH QHXWURQ ZDYH PHWKRG ZKHQ WKH SXOVHG GDWD LV DQDO\]HG LQ WKH )RXULHU WUDQVIRUP SODQH 7KH ZRUN E\ %RRWK >@ FRQILUPHG WKLV VLPSOLn I\LQJ DQG WLPHVDYLQJ DSSURDFK WR WKH SRZHUIXO EXW WLPHFRQVXPLQJ QHXn WURQ ZDYH PHDVXUHPHQWV 7KH VWXG\ RI PXOWLSO\LQJ PHGLD E\ QHXWURQ ZDYH RU SXOVHf SURSDJDn WLRQ KDV QRW EHHQ GHYHORSHG DV H[WHQVLYHO\ DV LQ QRQPXOWLSO\LQJ PHGLD %UHKP PDGH D YHU\ HOHJDQW DQDO\VLV RI WKH SUREOHP VKRZLQJ WKH H[FLWDn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n IXVLRQ SURJUDP ,W FDOFXODWHV WKH VSDFHWLPH EHKDYLRU RI WKH QHXWURQ IOX[ LQ D UHDFWRU GXULQJ D WUDQVLHQW 7KH FDOFXODWLRQ LV UHVWULFWHG WR RQHGLPHQVLRQDO VODE JHRPHWU\ ZLWK ]HUR JUDGLHQW RU ]HUR IOX[

PAGE 109

ERXQGDU\ FRQGLWLRQV 8S WR VL[ GHOD\HG QHXWURQ JURXSV FDQ EH FRQn VLGHUHG LQ WKH FDOFXODWLRQV 7KH IHHGEDFN VXEURXWLQH PD\ EH XVHG WR LQWURGXFH DUELWUDU\ FKDQJHV LQ WKH UHDFWRU SDUDPHWHUV HLWKHU IRU LQKHUHQW IHHGEDFN RU IRU RWKHU WLPHGHSHQGHQW YDULDWLRQV 7KH EDVLF HTXDWLRQV VROYHG E\ WKH :,*/( SURJUDP DUH DV IROORZV f¬' YrO = K ; YLI r Y=I !f f¯ <%f r ( FX &! L VL ALU f 9r' 9I! f¯ (r! [ AY(I A Y=I A \%A L I! 6  &O f &L;L 6 f§ f & ;L &OL < %L Y(IO r OL W & ;L & < %L Y(I A W L Df L ,f Ef 7KH QXPEHU ,f RI GHOD\HG QHXWURQ JURXSV PD\ EH RU FDQ EH RU :KHQ DQG \! WKHVH HTXDWLRQV DUH WKH FRQYHQWLRQDO WZRJURXS WLPHGHSHQGHQW GLIIXVLRQ HTXDWLRQV 7KH WLPHGHSHQGHQW HTXDWLRQV DQG WKRVH HTXDWLRQV QHFHVVDU\ WR UHSUHVHQW WKH LQKHUHQW IHHGEDFN HIIHFWV LI WKH\ DUH FRQVLGHUHG DUH UHGXFHG WR ILQLWH WLPHf GLIIHUHQFH HTXDWLRQV DQG QXPHULFDOO\ VROYHG :,*/( FDQ KDQGOH XS WR UHJLRQV PHVK SRLQWV DQG WLPH VWHSV

PAGE 110

1HXWURQ :DYH $QDO\VLV 7KH QHXWURQ ZDYH WHFKQLTXH KDV SURYHQ WR EH D YHU\ SRZHUIXO PHWKRG RI GHWHUPLQLQJ WKH GHILFLHQFHV D PRGHO KDV LQ SUHGLFWLQJ SURSDJDWLRQ SKHQRPHQD RU LQ HVWDEOLVKLQJ WKH TXDOLW\ RI WKH UHVXOWV IURP D FRUUHn VSRQGLQJ H[SHULPHQW 7KH VWXG\ RI WKH VSDFHWLPH GDWD LQ WKH IUHTXHQF\ GRPDLQ ORRNV DW WKH RYHUDOO EHKDYLRU RI WKH DV\PSWRWLF VSDWLDO UHJLRQ RI WKH DVVHPn EO\ ZLWK SDUDPHWHUV WKDW LQFOXGH LQIRUPDWLRQ RQ DOO VSDWLDO SRLQWV IRU HDFK IUHTXHQF\ RI LQWHUHVW 7KH FRPSDULVRQ RI WKH SUHGLFWHG DQG PHDVn XUHG UHDO DQG LPDJLQDU\ FRPSRQHQWV RI WKH FRPSOH[ LQYHUVH UHOD[DWLRQ OHQJWK LV WKHUHIRUH D PRUH FRPSUHKHQVLYH HYDOXDWLRQ RI WKH GLVFUHSn DQFLHV WKDQ WKH SRLQWZLVH FRPSDULVRQ LQ WKH WLPH GRPDLQ 7KH DQDO\VLV LQ WKH A SODQH LV DQ HVSHFLDOO\ VWULQJHQW WHVW RI WKH PRGHO DQG WKH H[SHULPHQWDO GDWD ,Q WKLV ZRUN WKH FRPSDULVRQV LQ WKH IUHTXHQF\ GRPDLQ ZHUH PDGH E\ QXPHULFDOO\ GHWHUPLQLQJ WKH ]HURWK )RXULHU PRPHQW RI WKH GDWD PHDVn XUHG LQ WKH WLPH GRPDLQ 7KLV PHWKRG ZDV VXJJHVWHG E\ 0RRUHA 0RRUH H[SUHVVHG )RXULHU PRPHQWV RI WKH VSDFHWLPH GDWD DV Q UULIW L_!QU If GW W H LM!U Wf $f ZKHUH L.U Wf LV WKH QHXWURQ SXOVH DW VSDFH SRLQW U DV D IXQFWLRQ RI WLPH W DQG I LV WKH IUHTXHQF\ LQ F\FOHV SHU VHFRQG 7KH IXQGDPHQWDO VSDFH DQG HQHUJ\ PRGH SURSDJDWLQJ WKURXJK WKH PHGLXP LV JLYHQ E\ 7KH DXWKRU DFNQRZOHGJHV 'U 0 1 0RRUH IRU VHYHUDO HQOLJKWHQLQJ GLVFXVVLRQV RQ WKH VXEMHFW RI QHXWURQ ZDYHV SURSDJDWLRQ

PAGE 111

AU If $If 4 U U %Af H[S ^>RA&I %f L I Eef@]` ZKHUH $T VSHFLILHV WKH VRXUFH FRQGLWLRQ 4 U! U! % f VSHFLILHV WKH WUDQVYHUVH VSDFH GHSHQGHQFH DQG DUH UHVSHFWLYHO\ WKH UHDO DQG LPDJLQDU\ FRPSRQHQWV RI WKH FRPSOH[ LQYHUVH UHOD[DWLRQ OHQJWK 7KH )RXULHU WUDQVIRUP U If RI WKH XQLW LPSXOVH UHVSRQVH U Wf LV WKH WUDQVIHU IXQFWLRQ RI WKH V\VWHP 7KHUHIRUH WKH DPSOLn WXGH DQG SKDVH FXUYHV REWDLQHG IURP D QXPHULFDO )RXULHU WUDQVIRUPDWLRQ RI WKH SXOVH SURSDJDWLRQ GDWD FRUUHVSRQGV WR WKRVH PHDVXUHG GLUHFWO\ E\ WKH ZDYH H[SHULPHQW ,W VKRXOG EH PHQWLRQHG WKDW DOWKRXJK PRVW FRPPRQ QRPHQFODWXUH UHIHUV WR WKH SXOVH SURSDJDWLRQ SKHQRPHQD DQGRU WKH QHXWURQ ZDYH SURSDJDWLRQ DV WKHUPDO QHXWURQ SURSDJDWLRQ EHFDXVH RI WKH FORVHWR WKHUPDO VRXUFH HPSOR\HG DQG WKH FKDUDFWHULVWLFV RI WKH PHGLD VWXGLHG WKHUH LV QR VXFK SXUH SKHQRPHQD LQ D PXOWLSO\LQJ PHGLXP 7KH GLVn WXUEDQFH SURSDJDWHV ZKHWKHU LW VWDUWHG DV D IDVW RU WKHUPDO SXOVH DQG TXLFNO\ ORVVHV PHPRU\ RI LWV RULJLQ VSHFLDOO\ LQ ZDWHU PRGHUDWHG PHGLDf 7KH RQO\ GLIIHUHQFH EHWZHHQ LQVHUWLQJ D WKHUPDO RU D IDVW EXUVW RI QHXWURQV OLHV LQ WKH VSDWLDO GLVWULEXWLRQRI VRXUFH QHXWURQV 7KH WKHUPDO VRXUFH ZLOO DSSHDU HVVHQWLDOO\ ORFDOL]HG DW WKH SRLQW RI LQVHUWLRQ ZKLOH WKH IDVW VRXUFH QHXWURQV ZLOO SHQHWUDWH GHHS LQWR WKH DVVHPEO\

PAGE 112

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n PHQWDOO\ REWDLQHG WKHUPDO QHXWURQ SXOVH VKDSHV ZLWK WKH FRUUHVSRQGLQJ UHVXOWV IURP WKH DQDO\WLFDO VFKHPH 7KH VXEFULWLFDO FRQILJXUDWLRQ DQG HOHFWURQLF LQVWUXPHQWDWLRQ XVHG KDV EHHQ GHWDLOHG LQ 3DUW &KDSWHU ,9 8QOHVV VSHFLILHG WKH PHDVn XUHPHQWV ZHUH PDGH DW D PRGHUDWRU KHLJKW RI FP FRUUHVSRQGLQJ WR D NHII 7KH V\VWHP ZDV RSHUDWHG XQGHU FRQWLQXRXV ZDWHU IORZ FRQGLWLRQV 1R KHDWLQJ RI WKH ZDWHU ZDV REVHUYHG E\ HLWKHU WKH HQHUJ\ WUDQVIHUUHG

PAGE 113

E\ WKH SXPS RU WKH ORZ RSHUDWLQJ DYHUDJH SRZHU OHYHO RI WKH V\VWHP ZDWWf 7KH ODUJH YROXPH RI FLUFXODWLQJ ZDWHU DLGHG LQ PDLQn WDLQLQJ WKH WHPSHUDWXUH FRQVWDQW ,Q RQH KRXU SHULRG RI RSHUDWLRQ WKH WHPSHUDWXUH FKDQJH ZDV DERXW r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n LVRQ EHWZHHQ WKH PRGHO DQG H[SHULPHQW WKH FRQWULEXWLRQ RI WKH HSLFDGPLXP QHXWURQV ZDV VXEWUDFWHG IURP WKH WLPH SURILOH UHFRUGHG E\ WKH GHWHFWRU 7KH WKHUPDO DQG HSLFDGPLXP QHXWURQ WLPH SURILOHV PHDVXUHG LQ WKH DVVHPEO\ GR QRW GLIIHU VLJQLILFDQWO\ LQ VKDSH LQ WKH DV\PSWRWLF UHJLRQ EXW VKRZ VLJQLILFDQW GLIIHUHQFHV LQ WKH WLPH VFDOH WKH UHODWLYH DWWHQn XDWLRQ DQG WKH VKDSHV QHDU WKH VRXUFH $OWKRXJK WKH GHWHFWRUV XVHG DUH DERXW WHQ WLPHV PRUH HIILFLHQW IRU GHWHFWLQJ WKHUPDO QHXWURQV WKH HSLFDGPLXP IOX[ LV PXFK ODUJHU WKDQ WKH WKHUPDO IOX[ WKURXJKRXW WKH V\VWHP DQG LQ SDUWLFXODU FORVH WR WKH VRXUFH 7KHUHIRUH WKH

PAGE 114

IROORZLQJ WHFKQLTXH ZDV XVHG LQ PRVW RI WKH PHDVXUHPHQWV Df 0HDVXUHPHQWV ZHUH WDNHQ ZLWK WKH EDUH PRYDEOH GHWHFWRU DW HDFK GHVLJQDWHG SRVLWLRQ DQG D QRUPDOL]DWLRQ IDFWRU REWDLQHG IURP WKH IL[HG GHWHFWRU Ef 0HDVXUHPHQWV ZHUH WDNHQ ZLWK WKH VDPH PRYDEOH GHWHFWRU FRYHUHG ZLWK D LQ WKLFN FDGPLXP VOHHYH WKH &G FXWn RII HQHUJ\ IRU WKLV WKLFNQHVV LV HYf DW WKH VDPH SRVLn WLRQV DQG D QRUPDOL]DWLRQ IDFWRU REWDLQHG IURP WKH IL[HG GHWHFWRU Ff 7KH WLPH SURILOH IURP HDFK PHDVXUHPHQW ZDV UHVROXWLRQ WLPH FRUUHFWHG EDFNJURXQG VXEWUDFWHG DQG QRUPDOL]HG ZLWK UHIHUHQFH WR WKH IL[HG GHWHFWRU UHIHU WR 3DUW &KDSWHU ,9f Gf $ SRLQW E\ SRLQW VXEWUDFWLRQ RI WKH HSLFDGPLXP WLPH SURILOH IURP WKH WRWDO WLPH SURILOH \LHOGHG WKH WKHUPDO QHXWURQ IOX[ SXOVH VKDSH DW HDFK VSDWLDO SRLQW ,W VKRXOG EH PHQWLRQHG WKDW WR EH ULJRURXV D FRUUHFWLRQ VKRXOG EH DSSOLHG WR WKH HSLFDGPLXP IOX[ DW HDFK VSDFH SRLQW WR WDNH LQWR DFFRXQW WKH SHUWXUEDWLRQ RI WKH WKHUPDO IOX[ LQWURGXFHG E\ WKH FDGPLXP VOHHYH WKLV ZDV QRW DWWHPSWHG LQ WKLV VWXG\ ,W VKRXOG QRW EH VLJQLIn LFDQW LQ WKH KLJKO\ DEVRUELQJ PHGLXP EHLQJ FRQVLGHUHG KHUH $V D FRUROODU\ LW ZDV H[SHFWHG WKDW WKH HSLFDGPLXP IOX[ WLPH SURILOH FRXOG EH FRPSDUHG ZLWK WKH WKHRUHWLFDOO\ SUHGLFWHG IDVW IOX[ 7KH PDLQ SUREOHP HQFRXQWHUHG ZLWK WKLV FRPSDULVRQ OLHV LQ WKH IDFW WKDW WKH HQHUJ\GHSHQGHQW HIILFLHQF\ RI WKH GHWHFWRU DIIHFWV WKH FKDUDFWHUn LVWLFV RI WKH SXOVH 7KH HIILFLHQF\ RI GHWHFWLRQ IRU WKH WKHUPDO JURXS HYf LV IDLUO\ FRQVWDQW DQG WKH FRPSDULVRQ ZLWK WKH FDOFXODWHG WKHUPDO IOX[ LV RQ VROLG JURXQGV WKH VDPH DUJXPHQW LV QRW YDOLG

PAGE 115

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f ZDV REVHUYHG 7KH FRUUHFWLRQ YDULHG IURP b RI WKH WRWDO IOX[ YDOXHV 7KH *HRPHWULFDO $UUDQJHPHQW 6KRZQ LQ )LJ LV D SODQ YLHZ RI WKH VRXUFHVXEFULWLFDO DVVHPEO\ DUUDQJHPHQW XVHG WKURXJKRXW WKLV ZRUN 7KH GHWHFWRU SRVLWLRQV XVHG IRU WKH PHDVXUHPHQWV DUH FOHDUO\ LQGLFDWHG 7KH QXPEHUV VKRZQ FRUUHVSRQG WR WKH SRVLWLRQV RI IXHO SLQV ZLWK UHVSHFW WR WKH HQG RI WKH DVVHPEO\ IDFLQJ WKH QHXWURQ JHQHUDWRU 7KHUH ZHUH JULGKROHV EHWZHHQ GHWHFWRU SRVLWLRQV JLYLQJ D GLVWDQFH RI FP EHWZHHQ GDWD SRLQWV 7KURXJKn RXW WKH UHVW RI WKLV UHSRUW WKH SRVLWLRQV RI WKH GDWD SRLQWV ZLOO EH UHIHUUHG WR DV 3;; ZKHUH ;; LV WKH JULGKROH QXPEHU DW ZKLFK WKH GDWD ZDV UHFRUGHG LH 3 UHIHUV WR SRVLWLRQ QXPEHU VWDUWLQJ IURP WKH IURQW IDFH RI WKH DVVHPEO\ $ WRWDO RI URZV RI IXHO SLQV ZHUH LQ WKH DVVHPEO\ 'DWD ZDV WDNHQ DW QLQHWHHQ VSDWLDO SRLQWV IRU WKH PDLQ SDUW RI WKH UHVHDUFK 7KH WDUJHW DVVHPEO\ RI WKH QHXWURQ JHQHUDWRU

PAGE 116

)/8; ,5(/$7,9( 81,76f ),*r 7+( 727$/W(3,&$'0,80 $1' 7+(50$/ )/8; &0 )520 7+( 6285&(

PAGE 117

'LPHQVLRQV LQ FP 3,* 8)6$ 6285&(68%&5,7,&$/ $66(0%/< *(20(75,&$/ $55$1*(0(17 3/$1 9,(:

PAGE 118

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b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

PAGE 119

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n LPHQW LV LGHQWLFDO WR WKH RQH GHVFULEHG DERYH H[FHSW WKDW IHZHU GHWHFn WRU ORFDWLRQV ZHUH HPSOR\HG 3URSDJDWLRQ RI D :LGH 3XOVH 7R LOOXVWUDWH WKH SURSDJDWLRQ FKDUDFWHULVWLFV RI D SXOVH YLRODWLQJ

PAGE 120

WKH -RKQVRQ FULWHULRQ >@ D ZLGH PVHF VTXDUH SXOVH ZDV LQWURGXFHG LQWR WKH DVVHPEO\ 7KUHH VHOHFWHG SRVLWLRQV ZHUH XVHG IRU WKHVH PHDVn XUHPHQWV 3XOVH 3URSDJDWLRQ YV 3XOVH :LGWK 'XULQJ WKH H[SHULPHQWDWLRQ GHVFULEHG DERYH WKH REVHUYDWLRQ ZDV PDGH WKDW QDUURZ LQSXW SXOVHV ZLOO \LHOG WKH VDPH SXOVH VKDSH LQ WKH DV\PSWRWLF UHJLRQ ZLWK WKH SHDNV GLVSODFHG LQ WLPH 7R JDLQ DQ LQVLJKW LQWR WKLV SUHGLFWHG RFFXUUHQFH PHDVXUHPHQWV ZHUH WDNHQ DW RQH GHWHFWRU ORFDWLRQ 3f ZLWK LQSXW SXOVH ZLGWKV RI DQG PVHF (IIHFW RI 5RRP 5HWXUQ DW 3HULSKHUDO 'HWHFWRU 3RVLWLRQV 7KH HIIHFW RI QHXWURQV UHIOHFWHG IURP WKH FRQFUHWH ZDOOV RQ WKH PHDVXUHPHQWV ZDV LQYHVWLJDWHG E\ UHFRUGLQJ SXOVH VKDSHV DW SHULSKHUDO GHWHFWRU SRVLWLRQV 7KH PHDVXUHPHQWV ZHUH GRQH ZLWK DQG ZLWKRXW D VKLHOG RI ERUDWHG SDUDIILQ LQ WKH UHJLRQ LQ ZKLFK WKH GHWHFWRU ZDV ORFDWHG

PAGE 121

&+$37(5 ,9 (;3(5,0(17$/ $1' 7+(25(7,&$/ 5(68/76 ,1 7+( 7,0( '20$,1 7KH $QDO\WLFDO 0RGHO 7KH DQDO\WLFDO UHVXOWV ZHUH REWDLQHG XVLQJ DQ ,%0 YHUVLRQ RI WKH :,*/( FRGHA VXSSOLHG E\ WKH 3KLOOLSV 3HWUROHXP &RPSDQ\ DQG PDGH RSHUDWLYH RQ WKH ,%0 DW WKH 8QLYHUVLW\ RI )ORULGD &HUWDLQ PRGLn ILFDWLRQV ZHUH PDGH WR :,*/( WR REWDLQ SXQFKHG FDUG RXWSXW RI WKH WLPH SURILOH DW VHOHFWHG VSDWLDO SRLQWV ,Q WKLV PDQQHU WKH SXOVH VKDSHV DW WKH GHWHFWRU SRVLWLRQV FRUUHVSRQGLQJ WR WKH H[SHULPHQWDO PHDVXUHPHQWV ZHUH UHDGLO\ DYDLODEOH IRU GLUHFW FRPSDULVRQ ZLWK H[SHULPHQW 7KH :,*/( VFKHPH EHLQJ D RQHGLPHQVLRQDO PRGHO UHTXLUHV WKDW WKH WUDQVYHUVH OHDNDJH EH WDNHQ LQWR DFFRXQW E\ WKH SURSHU DGMXVWPHQW RI WKH SDUDPHWHUV 7KH WZRJURXS QXFOHDU SDUDPHWHUV IRU :,*/( ZHUH REWDLQHG E\ WKH SURFHGXUH RXWOLQHG LQ $SSHQGL[ $ 7KH HLJHQYDOXH WKUHHGLn PHQVLRQDO IOX[ VKDSHV DQG WKH EXFNOLQJV ZHUH REWDLQHG IURP WKH &25$ >@ FRPSXWHU SURJUDP 7KH FRUH ZLGWK ZDV VHW DW FP WKH UHIOHFWRU ZLGWK DW FP DQG WKH UHDFWRU KHLJKW DW FP 7KH H[SHULPHQWDO PHDVXUHPHQWV ZHUH PDGH ZLWK FP RI ZDWHU DERYH WKH ERWWRP RI WKH DFn WLYH IXHO EXW WKHUH ZDV D VPDOO ZDWHU UHIOHFWRU DW WKH ERWWRP WKXV WKH 7KH DXWKRU LV LQGHEWHG WR 0UV 0 ( 5DGG 1XFOHDU 6DIHW\ 5HVHDUFK %UDQFK 3KLOOLSV 3HWUROHXP &RPSDQ\ IRU KHU YDOXDEOH FRRSHUDWLRQ LQ IDPLOLDUL]LQJ WKH DXWKRU LQ WKH XVH RI WKH :,*/( FRGH

PAGE 122

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f ,Q )LJ WKH W\SH RI PDWHULDO QXPEHU RI PHVK SRLQWV DQG WKH PHVK VSDFLQJ XVHG LQ WKH FDOFXODWLRQV LV VKRZQ 7DEOH ,9 VKRZV WKH YDOXH RI WKH SDUDPHWHUV IRU 5HJLRQ DQG IRU 5HJLRQV UHVSHFn WLYHO\ %HIRUH WKH ILQDO UXQV ZHUH PDGH WKH LPSRUWDQFH RI WKH PHVK VSDFLQJ ZDV LQYHVWLJDWHG E\ UXQQLQJ WZR LGHQWLFDO FDVHV ZLWK DQG PHVK SRLQWV UHVSHFWLYHO\ 1R GLIIHUHQFH ZDV IRXQG EHWZHHQ UHVXOWV RI WKHVH WZR FDVHV 7KH FKRLFH RI WLPH LQFUHPHQWV ZDV PRUH FULWLFDO :KHQ WLPH VWHSV RI \VHF ZHUH XVHG LQ WKH YLFLQLW\ RI WKH WLPH ZKHUH WKH YDOXH RI WKH IDVW VRXUFH ZDV FKDQJHG IURP WR VHYHUH RVFLOODWLRQV GHYHORSHG DW WKH ILUVW VSDFH SRLQW 7R DYRLG WKLV SUREOHP WLPH VWHSV RI WR \VHF ZHUH HPSOR\HG LQ WKLV WLPH UHJLRQ DQG WKH PDJQLWXGH RI WKH VRXUFH ZDV UHGXFHG LQ WZR VWHSV ILUVW WR RQHKDOI WKH RULJLQDO YDOXH DQG WKHQ

PAGE 123

L U L L L L ."K L L ‘ASVV} LQ FR U? YP P FR L ?R ),* 3/$1 $1' )5217 9,(: 2) 7+( &25( 5(*,21 (1&/26,1* 7+( 1(87521 6285&(

PAGE 124

5(*,21 12 $ 0$7(5,$/ 12 12 2) 0(6+ 32,176 $ $ 0(6+ 6,=( FPf 0(6+ 32,17 12 '(7(&725 326,7,21 $2 $$ $ $ $ 12 $ $ $ $ ),* 21(',0(16,21$/ $55$1*(0(17 2) 7+( 8)6$ &25( 86(' ,1 7+( :,*/( &$/&8/$7,21$/ 6&+(0(

PAGE 125

7$%/( ,9 ,1387 3$5$0(7(56 )25 7+( :,*/( &$/&8/$7,21$/ 6&+(0( 8)6$ 5 &RUH 0: 5DWLR FP ZLGH UHIOHFWHG FRUH &25( 5HJLRQ 5HJLRQV $/80,180 :$7(5 'AFPf (DL&FQIf ,UAFP Y(IOFPf 'FPAf ,DFPf YOI]LFPAf >OYAFP AVHFf fµ( fµ( >OY@FP nf«VHFf ( ( 1RWH 7KHUPDO JURXS HY fµ )DVW JURXS HY 0H9 9

PAGE 126

WR ]HUR ZLWK \VHF WLPH VWHSVf EHWZHHQ WKH WZR VWHSV ,W ZDV DOVR REVHUYHG WKDW D ORQJ WLPH DIWHU WKH LQSXW SXOVH ZDV FXWRII RVFLOn ODWLRQV GHYHORSHG DW VSDWLDO SRLQWV FORVH WR WKH VRXUFH ZKHQ WLPH LPFUHPHQWV ODUJHU WKDQ \VHF ZHUH XVHG 7KH WLPH LQFUHPHQWV XVHG IRU WKH DQG PVHF LQSXW SXOVH SUREOHPV DUH VKRZQ LQ 7DEOH 9 6WDEOH RSHUDWLRQ RI WKH FRGH ZDV REn WDLQHG ZLWK WKHVH WLPH VWHSV 5XQQLQJ WLPHV ZHUH RI WKH RUGHU RI PLQXWHV LQ WKH 8) ,%0 7KH PDLQ SUREOHP LQ HVWDEOLVKLQJ D IDLU FRPSDULVRQ EHWZHHQ WKH FDOFXODWLRQDO DQG H[SHULPHQWDO UHVXOWV OLHV LQ DQ DGHTXDWH SK\VLFDO GHVFULSWLRQ RI WKH VSDWLDO GLVWULEXWLRQ RI WKH H[WHUQDO IDVW VRXUFH )RU H[DPSOH ZKHQ FDOFXODWLRQV ZHUH GRQH ZLWK D VSDWLDOO\ ORFDOL]HG FP ZLGH VRXUFH SODFHG DW WKH SK\VLFDO ORFDWLRQ RI WKH QHXWURQ WDUJHW )LJ f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f§, = JLYHQ E\ D ILUVWIOLJKW NHUQHO RI WKH IRUP H U ZKHUH eU LV WKH UHPRYDO FURVVVHFWLRQ IRU VRXUFH QHXWURQV 7KH GLVWULEXWLRQ XVHG LV VKRZQ LQ )LJ 7KH FKRLFH RI D SODQH VRXUFH NHUQHO LV UHDVRQDEOH HVSHFLDOO\ DW SRLQWV D IHZ FP UHPRYHG IURP WKH VRXUFH GXH WR WKH SK\VLFDO

PAGE 127

7$%/( 9 7,0( 67(36 86(' )25 7+( :,*/( &$/&8/$7,216 8)6$ 5 &RUH 0: 5DWLR FP ZLGH UHIOHFWHG FRUH ,QSXW 3XOVH :LGWK PVHF 7LPH 6WHS 0LFURVHFf 7R 7LPH 6WHS 1R 7LPH 6WHS 0LFURVHFf 7R 7LPH 6WHS 1R ,QSXW 3XOVH :LGWK PVHF

PAGE 128

5HODWLYH $PSOLWXGH ),* 63$7,$/ ',675,%87,21 2) 6285&( 1(875216 ,1&25325$7(' ,172 7+( :,*/( 6&+(0(

PAGE 129

FRQILJXUDWLRQ RI WKH VRXUFH DQG WKH DVVHPEO\ 7KH SODQDU QDWXUH RI WKH VRXUFH ZDV FRQILUPHG H[SHULPHQWDOO\ E\ REWDLQLQJ SXOVH SURSDJDWLRQ GDWD DFURVV WKH ZLGWK RI WKH FRUH FORVH WR DQG IDU IURP WKH VRXUFH 7KH UHPRYDO FURVVVHFWLRQ IRU WKH XQFROOLGHG VRXUFH QHXWURQV LQ UHJLRQ ZDV FKRVHQ WR EH WKH UHPRYDO FURVVVHFWLRQ RI QHXWURQV IURP WKH IDVW JURXS LQ WKDW UHJLRQ DV REWDLQHG IURP WKH WZRJURXS VFKHPH UHI 7DEOH ,9f )RU UHJLRQV DQG WKH FRUUHVSRQGLQJ UHPRYDO FURVVVHFWLRQ ZDV WDNHQ WR EH WKH UHPRYDO FURVVVHFWLRQ RI QHXWURQV IURP WKH IDVW JURXS LQ WKH IRXUJURXS VFKHPH ,W VKRXOG EH QRWHG WKDW WKH UHVXOWV DUH QRW VHQVLWLYH WR WKH YDOXH RI WKH UHPRYDO FURVVVHFWLRQ LQ UHJLRQ 7KH DERYH FKRLFH RI WKH ? UHPRYDO FURVVVHFWLRQ IRU WKH UHPDLQGHU RI WKH DVVHPEO\ LV FRQVLVWHQW ZLWK WKH RYHUDOO FDOFXODWLRQDO VFKHPH XVHG DQG ZLWK WKH YDOXH GHWHUn PLQHG HPSLULFDOO\ IRU 0HY QHXWURQV LQ OLJKW ZDWHU 7KLV FKRLFH RI D ILUVWIOLJKW VRXUFH JDYH FDOFXODWLRQDO UHVXOWV LQ VXUSULQVLQJO\ JRRG DJUHHPHQW ZLWK H[SHULPHQW SDUWLFXODUO\ IRU WKH GHOD\ WLPHV DQG WKH VSDWLDO DWWHQXDWLRQ 7KH SUHGLFWHG SXOVH VKDSHV DUH VRPHZKDW QDUURZHU WKDQ WKH SXOVH VKDSHV REWDLQHG IURP H[SHULPHQW 1R VLJQLILFDQW GLIIHUHQFHV EHWZHHQ WKH UHVXOWV RI WKH ORFDOL]HG VRXUFHV FDVH PHQWLRQHG DERYH DQG D SRLQW NHUQHO VRXUFH GLVWULEXWLRQ ZHUH REn VHUYHG 7KH SRLQW NHUQHO VRXUFH GLVWULEXWLRQ XQGHUSUHGLFWV WKH SHQHn WUDWLRQ RI WKH ILUVW FROOLVLRQ IDVW QHXWURQV LQWR WKH DVVHPEO\ $W WKLV SRLQW D IHZ FRPPHQWV RQ WKH HQHUJ\ GLVWULEXWLRQ RI WKH VRXUFH QHXWURQV DUH SHUWLQHQW $ SXUH WULWLXP WDUJHW ERPEDUGHG ZLWK D EHDP RI PRQRHQHUJHWLF GHXWHURQ LRQV ZLOO SURGXFH HVVHQWLDOO\ PRQR HQHUJHWLF QHXWURQV VLQFH WKH UHODWLRQVKLS EHWZHHQ WKH HQHUJ\ RI WKH LRQV WKH QHXWURQ HQHUJ\ DQG WKH DQJOH RI HPLVVLRQ LV XQLTXH +RZHYHU

PAGE 130

WKH GWf UHDFWLRQ UDSLGO\ EHFRPHV FRQWDPLQDWHG ZLWK WKH GGf UHDFWLRQ GXH WR WKH DFFXPXODWLRQ RI WKH GHXWHURQ EHDP RQ WDUJHW $IWHU a PLFURDPSKU SHU XQLW DUHD RI DFFXPXODWHG EHDP RQ WDUJHW WKH RULJLQDO QHXWURQ \LHOG IURP WKH GWf UHDFWLRQ KDV GURSSHG WR KDOI RI LWV RULJLQDO YDOXH ZKLOH GGf QHXWURQV DFFRXQW IRU b RI WKH QHXWURQ EHDP $IWHU a PLFURDPSKU SHU XQLW DUHD RI DFFXPXODWHG EHDQ RQ WDUJHW WKH QXPEHU RI GGf QHXWURQV KDV LQFUHDVHG WR b WKLV UHSUHn VHQWV D GHWHFWDEOH FRQWDPLQDWLRQ DQG PD\ KDYH VRPH HIIHFW RQ WKH UHPRYDO FURVVVHFWLRQ RI VRXUFH QHXWURQV ,Q WKH SUHVHQW FDVH KRZn HYHU GXH WR WKH FRDUVH HQHUJ\ PHVK XVHG LQ WKH FDOFXODWLRQ VFKHPH WKH HIIHFW LV QRW FRQVLGHUHG WR EH LPSRUWDQW 7\SLFDO UHVXOWV REWDLQHG ZLWK WKH :,*/( FRGH DUH VKRZQ LQ )LJ $ % &f IRU D PVHF LQSXW SXOVH 7KH JHQHUDO FKDUDFWHULVWLFV RI WKH SXOVH SURSDJDWLRQ SKHQRPHQD DUH FOHDUO\ GLVSOD\HG 7KH FDOFXODWHG VSDWLDO GLVWULEXWLRQ RI WKH QHXWURQV DV D IXQFWLRQ RI WLPH LV VKRZQ IRU WLPHV RI DQG PVHF LQ )LJ $ %f 7KH VSDWLDO GLVSHUVLRQ LQFUHDVHV ZLWK WLPH ZKLOH WKH SXOVH DWWHQXDWHV VHYHUHO\ LQ WLPH DQG VSDFH 7KH :,*/( UHVXOWV DUH GLVFXVVHG VLPXOWDQHRXVO\ ZLWK WKH UHVXOWV RI WKH FOHDQ FRUH PHDVXUHPHQWV )OX[ 7UDYHUVHV 6WDWLF DQG G\QDPLF IOX[ WUDYHUVHV ZHUH REWDLQHG LQ WKH FOHDQ FRUH WR HVWDEOLVK WKH IROORZLQJ Df 7KH DV\PSWRWLF VWHDG\VWDWH IOX[ VKDSHV LQ WKH WUDQVYHUVH GLUHFWLRQ Ef 7KH SODQDU SURSDJDWLRQ RI WKH QHXWURQ SXOVH 7KLV LV H[WUHPHO\

PAGE 131

)/8; 5(/$7,9( 81,76f ),* $ 38/6( 6+$3(6 35(',&7(' %< :,*/( $7 ',))(5(17 326,7,216 ,1 7+( 8)6$ 5 &25(

PAGE 132

)/8; 5(/$7,9( 81,76f ),* % 38/6( 6+$3(6 35(',&7(' %< :,*/( $7 ',))(5(17 326,7,216 ,1 7+( 8)6$ 5 &25(

PAGE 133

U/8; 5(/$7,9( 81,76f ),* & 38/6( 6+$3(6 35(',&7(' %< :,*/( $7 ',))(5(17 326,7,216 ,1 7+( 8)6$ 5 &25(

PAGE 134

),* $ 7+( &$/&8/$7(' 63$7,$/ ',675,%87,21 2) 7+( 7+(50$/ )/8; $7 ',))(5(17 7,0(6 $)7(5 7+( 38/6(

PAGE 135

),* % 7+( &$/&8/$7(' 63$7,$/ ',675,%87,21 2) 7+( 7+(50$/ )/8; $7 ',))(5(17 7,0(6 $)7(5 7+( 38/6(

PAGE 136

LPSRUWDQW VLQFH WKH RQHGLPHQVLRQDO :,*/( FDOFXODWLRQV FDQQRW DFFRXQW IRU WUDQVYHUVH SURSDJDWLRQ RI D GLVWXUEDQFH 6WDWLF )OX[ 7UDYHUVHV 6KRZQ LQ )LJV DQG DUH WKH PDSSHG VWHDG\VWDWH IOX[HV DFURVV WKH KHLJKW DQG WKH ZLGWK RI WKH FRUH UHVSHFWLYHO\ DW D GLVWDQFH RI FP IURP WKH QHXWURQ VRXUFH 7KLV ORFDWLRQ ZDV FKRVHQ EHFDXVHnWKH GHOD\ WLPH DQG DWWHQXDWLRQ RI WKH SXOVHV VKRZHG LW WR EH DW WKH EHJLQn QLQJ RI WKH DV\PSWRWLF UHJLRQ YL] WKH UHJLRQ LQ ZKLFK ERWK WKH YHORFLW\ RI SURSDJDWLRQ DQG WKH UHOD[DWLRQ OHQJWK KDYH UHDFKHG WKHLU DV\PSWRWLF YDOXHV %RWK IOX[ PDSV ZHUH GRQH ZLWK ,QGLXP IRLOV DQG DSSURSULDWH FRUUHFWLRQV ZHUH DSSOLHG WR DFFRXQW IRU VHOIVKLHOGLQJ DQG LQWHUIHUHQFH VKDGRZLQJf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n IRUPHG :KHQ DOO WKLV LV WDNHQ LQWR DFFRXQW WKH UHVXOWV ORRN UHDVRQn DEOH ,W ZDV IRXQG WKDW WKH ERZLQJ RI WKH IXHO URGV DIIHFWHG WKHVH PHDVXUHPHQWV VLJQLILFDQWO\ 7KLV ZDV GHWHUPLQHG E\ SRVLWLRQLQJ WKH

PAGE 137

)OX[ 5HODWLYH 8QLWVf ),* 7+( $6<03727,& 67($'<67$7( 9(57,&$/ )/8;

PAGE 138

)OX[ 5HODWLYH 8QLWVf 5 &OHDQ &RUH FP IURP WKH VRXUFH ,Q )RLO $FWLYDWLRQ ),* 7+( $6<03727,& 67($'<67$7( +25,=217$/ )/8;

PAGE 139

ORQJ GHWHFWRU LQ RQH SRVLWLRQ DQG URWDWLQJ WKH IXHO HOHPHQWV WKDW VXUURXQG LW 7KH QDUURZ FRUH LQ ZLGHf LV REYLRXVO\ YHU\ VXVFHSn WLEOH WR WKLV HIIHFW DQG WR WKH H[DFW FHQWHULQJ RI WKH QHXWURQ VRXUFH ZKRVH ORFDWLRQ FRXOG QRW EH GHWHUPLQHG WR EHWWHU WKDQ FP DFFXUDF\ '\QDPLF )OX[n7UDYHUVHV '\QDPLF IOX[ WUDYHUVHV DFURVV WKH ZLGWK RI WKH FRUH DQG UHIOHFWRU ZHUH FDUULHG RXW DW GLVWDQFHV RI DQG FP IURP WKH VRXUFH $V PHQWLRQHG DERYH WKHVH PHDVXUHPHQWV DUH LPSRUWDQW VLQFH WKH WKHRUHWn LFDO PRGHO LV RQHGLPHQVLRQDO DQG SODQDU SURSDJDWLRQ LV FRQFHSWXDOO\ GHVLUDEOH 6KRZQ LQ 7DEOH 9, DUH WKH SHDNLQJ WLPHV REWDLQHG IRU VHYHUDO SRVLWLRQV DFURVV WKH ZLGWK RI WKH FRUH DQG VLGH UHIOHFWRUV 7KH H[SHUn LPHQW ZDV SHUIRUPHG ZLWK ERWK DQG PVHF LQSXW SXOVHV 2QO\ WKH UHVXOWV IRU WKH PVHF LQSXW SXOVH DUH VKRZQ 7KH UHVXOWV IRU WKH PVHF FDVH GLVSOD\ WKH VDPH EHKDYLRU $OO WKH SHDNV RFFXUUHG ZHOO ZLWKLQ WKH H[SHULPHQWDO DFFXUDF\ &RPSDULVRQ RI SXOVH VKDSHV DW GLIn IHUHQW SRVLWLRQV DFURVV WKH FRUH DQG UHIOHFWRU UHYHDOHG QR GLIIHUHQFHV LQ WKH EDVLF VKDSHV 1HDU WKH RXWHU ZDOO RI WKH UHIOHFWRU D VHFRQGDU\ SXOVH ZKLFK SHDNHG DW D WLPH FRUUHVSRQGLQJ WR WKH LQSXW SXOVH ZLGWK ZDV REVHUYHG ,W ZDV FRQILUPHG H[SHULPHQWDOO\ WKDW WKLV ZDV GXH WR QHXWURQ UHIOHFWLRQV IURP WKH ZDOOV RI WKH IDFLOLW\ URRP )URP WKH SXOVH SURSDJDWLRQ PHDVXUHPHQWV UHDFWLYLW\ GDWD ZDV DOVR REWDLQHG 7KHVH UHVXOWV DUH DQDO\]HG LQ 3DUW &KDSWHU 9, &OHDQ &RU« 3XOVH 3URSDJDWLRQ 0HDVXUHPHQWV 7KH PDLQ SRUWLRQ RI WKH UHVHDUFK SURSDJDWLRQ PHDVXUHPHQWV LQ WKH FOHDQ FROG VLGHUHIOHFWHG DVVHPEO\ 7KHVH PRVW GHWDLOHG H[SHULPHQWV

PAGE 140

7$%/( 9, '(/$< 7,0(6 0($685(' $&5266 7+( :,'7+ 2) 7+( &25( 06(& ,1387 38/6( 8)6$ 5 FRUH 0: UDWLR FP ZLGH UHIOHFWHG FRUH 'HOD\ 7LPHV PVHFf 3RVLWLRQ 'LVWDQFH WR 'LVWDQFH WR WKH VRXUFH 1R &RUH &HQWHU FPf FP FP $ % & ' ( ) * + , 5( 5( 5( f§ 5( 5( f§ 5( f§ 5( f§f§

PAGE 141

ZHUH FRQGXFWHG WR LQYHVWLJDWH WKH SURSDJDWLRQ WKURXJKRXW WKH FRUH RI DQG PVHF LQSXW SXOVHV $V PHQWLRQHG LQ WKH SUHYLRXV VHFWLRQ SDUWLFXODU DVSHFWV RI WKH SURSDJDWLRQ SKHQRPHQD LQYHVWLJDWHG DUH GHDOW ZLWK VHSDUDWHO\ 7KH H[SHULPHQWDO WHFKQLTXH XVHG KDV DOUHDG\ EHHQ GHVFULEHG %ULHIO\ WKH WLPH SURILOH RI WKH QRUPDOL]HG WKHUPDO QHXWURQ IOX[ LV REWDLQHG DW VHOHFWHG SRVLWLRQV E\ WKH LQWHJUDO FRXQW PHWKRG RI QRUPDOn L]DWLRQ DQG WKH HSLFDGPLXP VXEWUDFWLRQ WHFKQLTXH 1LQHWHHQ GHWHFWRU SRVLWLRQV ZHUH XVHG IRU WKHVH PHDVXUHPHQWV )RU HDFK SXOVH ZLGWK WZR UXQV ZHUH PDGH DW HDFK SRVLWLRQ H[FHSW DW WKH IDU HQG RI WKH FRUH ZKHUH GDWD DFTXLVLWLRQ WLPHV EHFDPH VLJQLILFDQWO\ ORQJHU 0RVW RI WKH H[n SHULPHQWDO HUURUV DVVLJQHG WR WKH GDWD ZHUH REWDLQHG E\ DQDO\]LQJ WKH UHVXOWV RI WKHVH WZR PHDVXUHPHQWV )ROORZLQJ WKH QRPHQFODWXUH GHYHORSHG E\ 'RVKL DQG 0LOH\ >@ DW WKH 8QLYHUVLW\ RI ,OOLQRLV WKH GHOD\ WLPHV WA WKH G\QDPLF LQYHUVH UHOD[DWLRQ OHQJWK NA WKH DV\PSWRWLF YHORFLW\ RI SURSDJDWLRQ Y WKH IXOOZLGWK DW KDOIPD[LPXP ):+0f RI WKH SURSDJDWLQJ SXOVH DQG WKH SXOVH VKDSHV ZLOO EH DQDO\]HG WR GHVFULEH WKH SURSDJDWLRQ FKDUDFWHUn LVWLFV RI D QHXWURQ EXUVW LQ WKH YHU\ FORVH WR FULWLFDO DVVHPEO\ 7KH H[SHULPHQWDO UHVXOWV DUH SUHVHQWHG WRJHWKHU ZLWK WKH FDOFXODWHG YDOXHV ,W VKRXOG EH QRWHG WKDW SUDFWLFDOO\ DOO WKH ILJXUHV VKRZLQJ WKH SXOVH VKDSHV ZHUH SORWWHG XVLQJ D &DOFRPS SORWWHU ZKLFK LV SDUW RI WKH ,%0 &RPSXWHU IDFLOLW\ LQ WKH 1XFOHDU (QJLQHHULQJ 6FLHQFHV 'HSDUWn PHQW DW WKH 8QLYHUVLW\ RI )ORULGD 7KH GDWD LV SORWWHG LQ D FRQWLQXRXV PDQQHU DQG WKH SRLQWV DUH VXSHULPSRVHG ODWHU IRU FRQYHQLHQFHn 7\SLn FDOO\ WKH H[SHULPHQWDO SXOVH VKDSHV ZHUH SORWWHG IURP WLPH VWHSV WKH FRUUHVSRQGLQJ FDOFXODWLRQDO UHVXOWV KDYH WLPH VWHSV

PAGE 142

6KRZQ LQ )LJ $ % &f DUH D VHW RI WKH H[SHULPHQWDOO\ GHWHUPLQHG WKHUPDO QHXWURQ WLPH SURILOHV IRU WKH PVHF LQSXW SXOVH FDVH $ UHSUHVHQWDWLYH VHW RI WKH H[SHULPHQWDOO\ GHWHUPLQHG VSDWLDO GLVWULEXWLRQ RI QHXWURQV LQ WKH DVVHPEO\ IRU LQFUHDVLQJ WLPHV DIWHU WKH VRXUFH SXOVH LV VKRZQ LQ )LJ $ %f 7KH VSDFH DQG WLPH GHn SHQGHQW UHGLVWULEXWLRQ RI WKH WKHUPDO IOX[ DV WKH GLVWXUEDQFH SURSDn JDWHV WKURXJK WKH DVVHPEO\ LV FOHDUO\ GLVSOD\HG 7DEOHV 9,, DQG 9,,, VKRZ WKH H[SHULPHQWDOO\ GHWHUPLQHG DQG WKHRn UHWLFDOO\ FDOFXODWHG GHOD\ WLPHV ZLWK UHIHUHQFH WR ]HUR WLPH WKLV WLPH FRUUHVSRQGV WR WKH WLPH DW ZKLFK WKH QHXWURQ JHQHUDWRU LQLWLDWHV WKH EXUVW DQG WKH PXOWLFKDQQHO DQDO\]HU VWDUWV LWV VZHHSf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n WLRQ XVHG 7KH 3URSDJDWLRQ 7LPH DQG WKH $V\PSWRWLF 9HORFLW\ RI 3URSDJDWLRQ 7KH SURSDJDWLRQ WLPH ZKLFK LV GHILQHG DV WKH WRWDO GHOD\ WLPH EHWZHHQ SHDNV DW WKH WZR H[WUHPH SRVLWLRQV RI WKH DVVHPEO\ LV PVHF 7KH DV\PSWRWLF YHORFLW\ RI SURSDJDWLRQ LV GHILQHG DV WKH LQYHUVH RI WKH VORSH RI WKH FXUYH RI WKH GHOD\ WLPH YV GLVWDQFH LQ WKH

PAGE 143

)/8; 5(/$7,9( 81,76f 1250$/,=(' 7' 7+(50} 38/6( :,'7+ 06(&6 7,0( &0 (f ),* $ (;3(5,0(17$/ 38/6( 6+$3(6 $7 ',))(5(17 326,7,216 ,1 7+( 8)6$ 5 &25(

PAGE 144

)/8; 5(/$7,9( 81,7(f ),* (;3(5,0(17$/ 38/6( 6+$3(6 $7 ',))(5(17 326,7,216 ,1 7+( 7-)6$ 5 &25(

PAGE 145

)/8; 5(/$7,9( 81,76f % % ‘ ‘ 1250$/,=(' ,' 7+(50 38/6( :,'7+ 06(&6 3 g r r p : Br IW f±} m m" D Dr f§ IW IW IW ¯¯IWpm¯ADp ppHB r ppf±LFF !L6 R nn3 n‘ -N RW9QA Hr }m :f±f± f± Y/7 nnA!f§ Z/Y/ 6 7,0( 06(&f ),* & (;3(5,0(17$/ 38/6( 6+$3(6 $7 ',))(5(17 326,7,216 ,1 7+( 8)6$ 5 &25(

PAGE 146

H ! f ! WH W+ 8 3L ; G L+ WQ 3RVLWLRQ 1XPEHU ),* $ (;3(5,0(17$//< '(7(50,1(' 63$7,$/ ',675,%87,21 2) 7+( 1(875216 $7 ',))(5(17 7,0(6 $)7(5 7+( 38/6(

PAGE 147

)OX[ 5HODWLYH 8QLWVf 8)6$ 5 &OHDQ &RUH FP EHWZHHQ SRLQWV 3RVLWLRQ 1XPEHU ),* % (;3(5,0(17$//< '(7(50,1(' 63$7,$/ ',675,%87,21 2) 7+( 1(875216 $7 ',))(5(17 7,0(6 $)7(5 7+( 38/6(

PAGE 148

7$%/( 9,, &/($1 &25( 38/6( 3523$*$7,21 678',(6 (;3(5,0(17$/ $1' 7+(25(7,&$/ 5(68/76 06(& ,1387 38/6( :,'7+ 8)6$ 5 &RUH 0: 5DWLRQ FP ZLGH UHIOHFWHG FRUH 326,7,21 ',67$1&( 72 3($. $7 PVHFf ):+0 PVHFf 3($. &28176 12 6285&( FPf 7KHRU\ ([S 7KHRU\ ([S 7KHRU\ ([SE fµ D 1RUPDOL]HG WR 3RVLWLRQ E (UURUV DVVLJQHG IURP WKH GHYLDWLRQ IURP WKH PHDQ RI WZR PHDVXUHPHQWV H[FHSW IRU WKH ODVW VSDFH SRLQWV 7KLV HUURU LV XVXDOO\ ODUJHU WKDQ WKH FRXQWLQJ VWDWLVWLFV HUURU 7KH HUURU RI WKH ODVW SRLQWV ZDV FDOFXODWHG IURP FRXQWLQJ VWDWLVWLFV

PAGE 149

7$%/( 9,,, &/($1 &25( 38/6( 3523$*$7,21 678',(6 (;3(5,0(17$/ $1' 7+(25(7,&$/ 5(68/76 06(& ,1387 38/6( :,'7+ 8)6$ 5 &RUH 0: 5DWLRQ FP ZLGH UHIOHFWHG FRUH 326,7,21 ',67$1&( 72 3($. $7 PVHFf ):+0 PVHFf 3($. &28176 12 6285&( FPf 7KHRU\ ([S 7KHRU\ ([S 7KHRU\ fµ D 1RUPDOL]HG WR 3RVLWLRQ N (UURUV DVVLJQHG IURP WKH GHYLDWLRQ IURP WKH PHDQ RI WZR PHDVXUHPHQWV H[FHSW IRU WKH ODVW VSDFH SRLQWV 7KLV HUURU LV XVXDOO\ ODUJHU WKDQ WKH FRXQWLQJ VWDWLVWLFV HUURU 7KH HUURU RI WKH ODVW SRLQWV ZDV FDOFXODWHG IURP FRXQWLQJ VWDWLVWLFV

PAGE 150

'HOD\ 7LPH PVHFf } ),* &$/&8/$7(' $1' (;3(5,0(17$/ '(/$< 7,0(6 06(& ,1387 38/6(

PAGE 151

'HOD\ 7LPH PVHFf IW IW ),* &$/&8/$7(' $1' (;3(5,0(17$/ '(/$< 7,0(6 06(& ,1387 38/6(

PAGE 152

)OX[ 5HODWLYH 8QLWVf ),* $03/,78'( $77(18$7,21 2) 7+( 7+(50$/ 1(87521 )/8; 06(& ,1387 38/6(

PAGE 153

)OX[ 5HODWLYH 8QLWVf ),* $03/,78'( $77(18$7,21 2) 7+( 7+(50$/ 1(87521 )/8; 06(& ,1387 38/6(

PAGE 154

DV\PSWRWLFf° UHJLRQ RI WKH DVVHPEO\ ,Q WKLV UHJLRQ D OLQHDU UHODn WLRQVKLS EHWZHHQ GHOD\ WLPHV DQG GLVWDQFH IURP WKH VRXUFH H[LVWV 7KH DV\PSWRWLF YHORFLW\ RI SURSDJDWLRQ IRU WKH FOHDQ 8)6$ FRUH LV JLYHQ LQ 7DEOH ,; 7$%/( ,; $6<03727,& 3523$*$7,21 9(/2&,7< Y 3 8)6$ 5 &OHDQ &RUH 0: 5DWLR FP UHIOHFWHG FRUH ,QSXW 3XOVH :LGWK PVHFf 7KHRU\ YS PVHFf ([SHULPHQW ,W VKRXOG EH PHQWLRQHG WKDW WKH DV\PSWRWLF YHORFLW\ RI SURSDJDn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

PAGE 155

7KH DWWHQXDWLRQ RI WKH DPSOLWXGH RI WKH SHDNV IRU WKH 8)6$ DVVHPn EO\ ZHUH VKRZQ LQ )LJV DQG WKH QRUPDOL]HG SHDN YDOXHV ZHUH WDEXODWHG LQ 7DEOHV 9,, DQG 9,,, $JDLQ H[FHOOHQW DJUHHPHQW ZDV REn WDLQHG EHWZHHQ WKHRU\ DQG H[SHULPHQW QRW RQO\ LQ WKH DV\PSWRWLF UHJLRQ EXW WKURXJKRXW PRVW RI WKH DVVHPEO\ 7KH ILUVW WZR GDWD SRLQWV 3 DQG 3f DUH FRQVLGHUDEO\ ODUJHU LQ PDJQLWXGH WKDQ WKH YDOXHV SUHGLFWHG E\ :,*/( 7KLV LV SUREDEO\ FDXVHG E\ DQ XQGHUFRUUHFWLRQ IRU WKH HSLFDGPLXP IOX[ GHWHFWHG DW WKHVH SRVLWLRQV DQG E\ QRW DFFRXQWLQJ IRU WKH VLJQLILn FDQW QHXWURQ VWUHDPLQJ LQ WKH ODUJH YRLG FUHDWHG E\ WKH DOXPLQXP SRUW LQ ZKLFK WKH WDUJHW LV ORFDWHG 3RVLWLRQ LV SUREDEO\ WKH RQO\ EDG GDWD SRLQW DQG QR H[SODQDWLRQ KDV EHHQ IRXQG IRU LWV EHKDYLRU VLQFH IRXU FRQVHFXWLYH UXQV ZHUH PDGH DW WKDW SRVLWLRQ DQG DOO IRXU VKRZHG LGHQWLFDO EHKDYLRU 7KH :,*/( UHVXOWV UHIOHFW WKH LQIOXHQFH RI HQG HIIHFWV PRUH PDUNHGO\ WKDQ H[SHULPHQW ,W VKRXOG EH QRWHG WKDW UHIOHFn WLRQV IURP WKH FRQFUHWH ZDOOV LQ WKH DVVHPEO\ URRP DUH VLJQLILFDQW DW SHULSKHUDO GHWHFWRU SRVLWLRQV 7$%/( ; '<1$0,& ,19(56( 5(/$;$7,21 /(1*7+ G 8)6$ 5 &OHDQ &RUH 0: 5DWLR FP ZLGH UHIOHFWHG FRUH ,QSXW 3XOVH G FP f :LGWK PVHFf 7KHRU\ ([SHULPHQW

PAGE 156

7KH )XOO :LGWK DW +DOI0D[LP¼P ):+0f 7KH ):+0 LV GHILQHG DV WKH WLPHZLGWK RI WKH SXOVH DW KDOI WKH SHDN DPSOLWXGH 7KH ):+0 LV RQH RI WKH SDUDPHWHUV FRQYHQWLRQDOO\ XVHG WR FRPSDUH SXOVH VKDSHV DQG LV WKHUHIRUH LQFOXGHG LQ WKLV VWXG\ $Q DQDO\VLV RI WKH H[SHULPHQWDO DQG FDOFXODWHG YDOXHV RI WKH ):+0 WDEXODWHG LQ 7DEOHV 9,, DQG 9,,, VKRZHG WKH IROORZLQJ Df &ORVH WR WKH VRXUFH WKHRU\ SUHGLFWV VOLJKWO\ ODUJHU ):+0 YDOXHV WKDQ H[SHULPHQW WKH GLIIHUHQFH GHFUHDVLQJ ZLWK GLVWDQFH IURP WKH VRXUFH $V GLVFXVVHG DERYH WKH GLVFUHSDQF\ QHDU WKH VRXUFH LV GXH WR GLIILFXOWLHV LQ DGHTXDWHO\ GHVFULELQJ WKH YRLG UHJLRQ EHKLQG WKH VRXUFH Ef %HWZHHQ 3 DQG 3 WKH ):+0 DUH SUDFWLFDOO\ LGHQWLFDO Ff 3DVW 3 WKH H[SHULPHQWDOO\ GHWHUPLQHG ):+0 EHFRPHV LQFUHDVLQJn O\ ODUJHU WKDQ WKH :,*/( SUHGLFWLRQ 7KH PD[LPXP GLIIHUHQFH LV DERXW PLFURVHF ZKLFK FRUUHVSRQGV WR a b RI WKH H[SHULPHQWDO YDOXH 7KH DJUHHPHQW LV VWLOO FRQVLGHUHG WR EH YHU\ JRRG 7KH )OX[ 6KDSHV $ GHWDLOHG FRPSDULVRQ RI WKH WKHRUHWLFDO DQG H[SHULPHQWDO IOX[ VKDSHV REWDLQHG IRU WKH 8)6$ DVVHPEO\ ZDV DOVR PDGH 3ORWV RI WKH WLPH SURILOHV RI WKH WKHUPDO IOX[ DW HDFK GHWHFWRU SRVLWLRQ DUH LQFOXGHG LQ $SSHQGL[ & IRU WKH DQG WKH PVHF ZLGH VRXUFH SXOVHV 7KH SHDNV DUH QRUPDOL]HG WR XQLW\ 7KH IROORZLQJ LQIRUPDWLRQ LV VXSSOLHG LQ HDFK ILJXUH Df ,GHQWLI\LQJ 5XQ 1XPEHU Ef /RFDWLRQ RI WKH ([SHULPHQWDO 3HDN LQ 7LPH 7KLV SDUDPHWHU LV GHWHUPLQHG GLUHFWO\ E\ D FRPSXWHU URXWLQH ZKLFK LV SDUW RI WKH SORWWHU SURJUDP 7KH URXWLQH VLPSO\ ILQGV WKH WLPH FKDQQHO

PAGE 157

LQ ZKLFK WKH ODUJHVW QXPEHU RI FRXQWV ZDV UHFRUGHG DQG LQ PDQ\ LQn VWDQFHV LV QRW WKH EHVW YDOXH IRU WKH ORFDWLRQ RI WKH SHDN 7KH YDOXHV TXRWHG LQ 7DEOHV 9,, DQG 9,,, UHSUHVHQW UHILQHG YDOXHV REWDLQHG E\ LQVSHFWLRQ RI WKH H[SHULPHQWDO GDWD Ff 7KH ,QSXW 3XOVH :LGWK Gf 7KH ([SHULPHQWDO DQG 7KHRUHWLFDO ):+0 Hf 7KH /RFDWLRQ ZLWK 5HVSHFW WR WKH 6RXUFH LV JLYHQ LQ WKH ILJXUH WLWOH $ FRPSDULVRQ EHWZHHQ WKH WKHRUHWLFDO DQG H[SHULPHQWDO SXOVH VKDSHV UHYHDOV WKH IROORZLQJ Df 6KDSHV DJUHH YHU\ FORVHO\ LQ WKH VSDWLDO UHJLRQ GRPLQDWHG E\ WKH QHXWURQ VRXUFH 6PDOO GLVFUHSDQFLHV VWDUW WR VKRZ XS QHDU 3 Ef 3DVW 3 WKH WKHRUHWLFDO SXOVHV EHFRPH FRQVLVWHQWO\ QDUURZHU WKDQ WKH H[SHULPHQWDO RQH $V GLVFXVVHG DERYH WKH RWKHU PDLQ FKDUDFn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f D VHULHV RI

PAGE 158

FDOFXODWLRQV ZHUH SHUIRUPHG LQ ZKLFK WKH KHLJKW RI WKH DVVHPEO\ ZDV YDULHG IURP WR FP 7KH UHVSHFWLYH FRUH KHLJKWV ZLWK WKHLU FRUUHVSRQGLQJ YHUWLFDO EXFNOLQJV N A FDOFXODWHG XVLQJ $,0f DQG DEVRUSWLRQ FURVVVHFWLRQV XVHG LQ WKH :,*/( FRGH DUH OLVWHG LQ 7DEOH ;, 7KH UHVXOWV RI WKH FDOFXODWLRQV VKRZHG D PDUNHG VHQVLWLYLW\ WR FKDQJHV LQ WKH WUDQVYHUVH EXFNOLQJ 7KH GHOD\ WLPHV DQG WKH SHDN FRXQWV YV GLVWDQFH WR WKH VRXUFH DUH VKRZQ LQ )LJ DQG UHVSHFn WLYHO\ 7KH H[SHULPHQWDO UHVXOWV IRU FP DUH DOVR VKRZQ 7KH GDWD IRU WKH DWWHQXDWLRQ FXUYHV ZDV QRUPDOL]HG WR WKH DWWHQXDWLRQ DW 3 RI WKH FP :,*/( FDOFXODWLRQV 3XOVH VKDSHV IRU WKH GLIIHUHQW KHLJKWV DUH VKRZQ WRJHWKHU ZLWK WKH H[SHULPHQWDO UHVXOWV DW FP LQ )LJ $ % &f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

PAGE 159

7$%/( ;, &+$1*(6 ,1 7+( 18&/($5 3$5$0(7(56 '8( 72 &+$1*(6 ,1 7+( 75$169(56( %8&./,1* 8)6$ 5 &RUH 0: 5DWLR FP ZLGH UHIOHFWHG FRUH :,*/( ;6HFWLRQ FP f 5HDFWRU +HLJKW FPf (VWLPDWHG N HII 9HUWLFDO %XFNOLQJ FPf 6 ( ( ( ( ( (

PAGE 160

'HOD\ 7LPHV PVHFf ),* '(/$< 7,0(6 2) 7+( 7+(50$/ )/8; &$/&8/$7(' %< 7+( :,*/( &$/&8/$,21$/ 6&+(0( )25 ',))(5(17 &25( +(,*+76

PAGE 161

)OX[ 5HODWLYH 8QLWVf ),* $03/,78'( $77(18$7,21 2) 7+( 7+(50$/ )/8; &$/&8/$7(' %< 7+( :,&/( &$/&8/$7,21$/ 6&+(0( )25 ',))(5(17 &25( +(,*+76

PAGE 162

)/8; 5(/$7,9( 81,76f ),* $ 7+( 6(16,7,9,7< 2) 7+( 21(',0(16,21$/ 7:2 *5283 63$&(7,0( .,1(7,&6 6&+(0( 72 &+$1*(6 ,1 7+( 75$169(56( %8&./,1*

PAGE 163

)/8; 5(/$7,9( 81,76f ),* % 7+( 6(16,7,9,7< 2) 7+( 21(',0(16,21$/ 7:2 *5283 63$&(7,0( .,1(7,&6 6&+(0( 72 &+$1*(6 ,1 7+( 75$169(56( %8&./,1*

PAGE 164

)/8; 5(/$7,9( 81,76f ),* & 7+( 6(16,7,9,7< 2) 7+( 21(',0(16,&1$/ 7:2 *5283 63$&(7,0( .,1(7,&6 6&+(0( 72 &+$1*(6 ,1 7+( 75$169(56( %8&./,1*

PAGE 165

)/8; 5(/$7,9( 81,7(f ),* (;3(5,0(17$/ 38/6( 6+$3(6 $6 $ )81&7,21 2) &25( +(,*+7

PAGE 166

HYDOXDWLRQ RI WKH QXFOHDU SDUDPHWHUV 7KH DV\PSWRWLF SURSDJDWLRQ YHORFLW\ DQG G\QDPLF LQYHUVH UHOD[DWLRQ OHQJWK DV DIXQFWLRQ RI FRUH KHLJKW DU« VKRZQ LQ 7DEOH ;,, $V H[n SHFWHG WKH YHORFLW\ RI SURSDJDWLRQ DQG WKH DWWHQXDWLRQ GHFUHDVH DV FULWLFDOLW\ LV DSSURDFKHG 7KH DERYH DQDO\VLV SXWV WKH JRRG DJUHHPHQW EHWZHHQ H[SHULPHQW DQG WKHRU\ RQ D ILUPHU IRRWLQJ :KHQ WKH REVHUYHG SKHQRPHQRQ LV YHU\ VHQn VLWLYH WR VPDOO YDULDWLRQV LQ V\VWHP FKDUDFWHULVWLFV WKH SUREDELOLW\ WKDW IRUWXLWRXV DJUHHPHQW PD\ EH REWDLQHG EHWZHHQ WKHRU\ DQG H[SHULn PHQW GLPLQLVKHV ZLWK WKH GHJUHH RI VHQVLWLYLW\ &RPPHQWV RQ WKH )DVW *URXS 5HVXOWV 7KH IROORZLQJ FRPPHQWV DUH SHUWLQHQW UHJDUGLQJ WKH IDVW JURXS UHVXOWV DV FDOFXODWHG E\ WKH WZRJURXS :,*/( PRGHO 7KH FXUYH RI GHOD\ WLPHV YV GLVWDQFH UHPDLQV IODWWHU WKDQ WKH FRUUHVSRQGLQJ FXUYH IRU WKH WKHUPDO JURXS QHDU WKH VRXUFH UHJLRQ LW WKHQ EHQGV UDSLGO\ WR \LHOG WKH VDPH DV\PSWRWLF YHORFLW\ RI SURSDJDn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

PAGE 167

7$%/( ;,, 7+( &$/&8/$7(' $6<03727,& 9(/2&,7< 2) 3523$*$7,21 $1' '<1$0,& ,19(56( 5(/$;$7,21 /(1*7+ 96 &25( +(,*+7 06(& ,1387 38/6( :,'7+ 8)6$ 5 &RUH 0: 5DWLR FP ZLGH UHIOHFWHG FRUH &RUH .G 9 +HLJKW FPf 4 FP ;f 3 PVHF f

PAGE 168

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

PAGE 169

7$%/( ;,,, '(/$< 7,0(6 $1' ):+0 )25 $ 1$552: ,1387 38/6( ',67$1&( 72 6285&( FPf 3($. $7 7KHRU\ PVHFf ([S ):+0 7KHRU\ PVHFf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£ :LGH 3XOVH $ VHULHV RI PHDVXUHPHQWV ZHUH DOVR PDGH LQ WKH FOHDQ FRUH ZLWK D PVHF ZLGH LQSXW SXOVH 7KLV SXOVH LV PXFK ZLGHU WKDQ WKH

PAGE 170

FKDUDFWHULVWLF SURSDJDWLRQ WLPH A PVHFf DFURVV WKH DVVHPEO\ RI D QDUURZ PVHFf SXOVH ,W VKRXOG EH UHFDOOHG WKDW WKH -RKQVRQ FULWHULRQ >@SRVWXODWHV WKDW WR EH DEOH WR REVHUYH VSDWLDOO\ GHSHQGHQW HIIHFWV WKH +:+0 RI WKH LQSXW SXOVH VKRXOG EH VPDOOHU WKDQ WKH FKDUDFn WHULVWLF SURSDJDWLRQ WLPH DFURVV WKH DVVHPEO\ )RU H[DPSOH WKH H[SHULn PHQWV RI 0LOH\ DQG FRZRUNHUV DW WKH 8QLYHUVLW\ RI ,OOLQRLV > @ VKRZHG WKDW IRU DQ LQLWLDOO\ DV\PSWRWLF YHU\ ZLGH LQSXW SXOVH QR VKDSH FKDQJHV ZHUH REVHUYHG DV WKH SXOVH SURSDJDWHG DFURVV WKH DVVHPEO\ 7KH SXOVH VKDSHV UHVXOWLQJ IURP WKHVH PHDVXUHPHQWV DW WKUHH SRVLn WLRQV LQ WKH DVVHPEO\ DUH VKRZQ LQ $SSHQGL[ ) 7KH GHOD\ WLPHV WKH ):+0 DQG DUHD XQGHU WKH SXOVHV QRUPDOL]HG WR XQLW\ DW WKH ILUVW SRVLn WLRQ DUH JLYHQ EHORZ LQ 7DEOH ;,9 7$%/( ;,9 '(/$< 7,0(6 $1' ):+0 )25 $ :,'( ,1387 38/6( ',67$1&( 72 6285&( FPf 3($. $7 PVHFf ):+0 PVHFf $5($ 81'(5 7+( &859( )URP WKHVH UHVXOWV LW VHHPV WKDW DOWKRXJK WKH GHOD\ WLPHV EHWZHHQ WKH UHFRUGHG SHDNV DUH TXLWH VPDOO D FHUWDLQ UHDUUDQJHPHQW RI WKH LQLWLDOO\ QRQDV\PSWRWLF SXOVH WDNHV SODFH &ORVH WR WKH VRXUFH WKH SXOVH LV SUDFWLFDOO\ VTXDUH DW D GLVWDQFH RI FP LW KDV ZLGHQHG VRPHn ZKDW DQG DW ODUJH GLVWDQFHV IURP WKH VRXUFH WKH SXOVH VHHPV WR EH WU\LQJ WR DFKLHYH WKH QRUPDO GXPEEHOO VKDSH -RKQVRQnV FULWHULRQ DSSHDUV WR

PAGE 171

EH WRR UHVWULFWLYH VLQFH VRPH VSDWLDO UHGLVWULEXWLRQ RI WKH LQSXW SXOVH GRHV RFFXU DOWKRXJK WKH HIIHFW LV FHUWDLQO\ PXFK OHVV SURQRXQFHG WKDQ LQ WKH FDVH RI D QDUURZ SXOVH ,Q WKH FDVH RI WKH ,OOLQRLV H[SHUn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

PAGE 172

7$%/( ;9 38/6( 6+$3(6 96 ,1387 38/6( :,'7+ 8)6$ 5O &RUH +: 5DWLR FP ZLGH UHIOHFWHG FRUH ,1387 38/6( 3($. $7 ):+0 :,'7+ PVHFf PVHFf PVHF

PAGE 173

FKDUDFWHULVWLF DV\PSWRWLF SXOVH VKDSH DOUHDG\ H[LVWV &KDUDFWHULVWLF GXPEEHOO VKDSHV DUH REVHUYHG KRZHYHU XQWLO WKH +:+0 A SURSDJDWLRQ WLPH )RU ZLGHU LQSXW SXOVHV WKH SURSDJDWLQJ SXOVHV DUH XQDEOH WR DFKLHYH WKH VPRRWK VKDSH REVHUYHG IRU QDUURZHU SXOVHV 6WLOO D FHUWDLQ VSDWLDO UHDUUDQJHPHQW RI WKH VKDSH RFFXUV DV ZDV GLVFXVVHG DERYH IRU WKH PVHF LQSXW SXOVH fµ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f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n WRU WDQN ZDOO ,W ZDV IRXQG WKDW WKH HIIHFW ZDV QHJOLJLEOH DW D

PAGE 174

IOX[ UHODWLYH XQLWVf * 9 ³ 816+,(/'(' IW r IW r IW IW I IW ƒ IW fµ , ? fµrrrrnr"!! $ f¬ $rr ILWnr \" ?DA I r \ t D iU 6+,(/'(' U b ‘ rK " UZ L Q U t } I? L L D LI I 8r+?IW 9$ 9QL Q? 1250$/,=(' 72 ‘ 55()5()* SXOVH ZLGWK PVHFV (;3 ):+0 06(&6 9rrf¬ m * % 7,0( 06(&f ),* $ ())(&7 2) 5220 5(7851 $7 3(5,3+(5$/ '(7(&725 326,7,216

PAGE 175

a/8; 5(/$7,9( 81,76f 55()5()* 3($. $7 6 06(&6 38/6( :,'7+ 06(&6 (;3 ):+9 06(&6 816+,(/'(' 6+,(/'(' 0 8L &a • 6 D &7,0( 06(&f ),* % ())(&7 ') 5220 5(7851 $7 3(5,3+(5$/ '(7(&725 326,7,216

PAGE 176

GLVWDQFH RI FP IURP WKH RXWHU ZDOO $W WKH IDU HQG RI WKH DVVHPEO\ WKH UHIOHFWHG SXOVH SURSDJDWHG LQWR WKH DVVHPEO\ DWWHQXDWLQJ DQG GLVSHUVLQJ ZLWK GLVWDQFH WR WKH HQG ZDOO 7KLV SXOVH TXLFNO\ GLVVLn SDWHG LQWR WKHDVVHPEO\ ZLWK QR VLJQLILFDQW HIIHFW RQ WKH PHDVXUHPHQWV

PAGE 177

&KDSWHU Y (;3(5,0(17$/ $1' 7+(25(7,&$/ 5(68/76 ,1 7+( )5(48(1&< '20$,1 ,QWURGXFWLRQ 7KH UHVXOWV RI WKH SXOVH SURSDJDWLRQ PHDVXUHPHQWV DQG RI WKH VSDFHn WLPHGHSHQGHQW :,*/( FDOFXODWLRQV ZHUH DQDO\]HG LQ WKH VHQVLWLYH )RXULHU WUDQVIRUP SODQH 7KLV QHXWURQ ZDYH W\SH DQDO\VLV SUREHV LQWR WKH GHIL FLHQFHV RI WKH PRGHO DQGRU WKH H[SHULPHQWDO GDWD WKDW WKH DQDO\VLV LQ WKH WLPH GRPDLQ PLJKW IDLO WR UHYHDO ,Q WKLV VWXG\ WKH DQDO\VLV LV SHUIRUPHG ZLWK WKH VROH LQWHQWLRQ RI WHVWLQJ WKH PRGHO DQG QRW WR H[n WUDFW SDUDPHWHUV 7KH :,*/( SUHGLFWHG WLPH SURILOHV RI WKH SURSDJDWLQJ GLVWXUEDQFH DV D IXQFWLRQ RI VSDFH DQG WKH SXOVH SURSDJDWLRQ GDWD ZHUH )RXULHU DQDO\]HG LQ DQ LGHQWLFDO PDQQHU WKH DPSOLWXGH DQG SKDVH DQJOH RI WKH ]HURWK )RXULHU PRPHQW WKXV REWDLQHG ZHUH OHDVWVTXDUHV ILWWHG ORJ DPSOLWXGH YV ] DQG SKDVH DQJOH YV ] IRU HDFK IUHTXHQF\ RI LQWHUHVWf \LHOGLQJ WKH GDPSLQJ FRHIILFLHQW D DQG WKH SKDVH VKLIW SHU XQLW OHQJWK e 7KH SUHGLFWHG DQG PHDVXUHG GLVSHUVLYH FKDUDFWHULVWLFV RI WKH V\VWHP DUH WKHQ FRPSDUHG LQ WKH S DQG S SODQH 7KH QHXWURQ ZDYH DQDO\VLV ZDV DSSOLHG WR ERWK WKH DQG PVHF LQSXW SXOVH VSDFHWLPH PHDVXUHPHQWV DQG FDOFXODWLRQV

PAGE 178

0HWKRG RI $QDO\VLV 7KH IROORZLQJ PHWKRG RI DQDO\VLV ZDV DSSOLHG WR WKH FDOFXODWHG DQG PHDVXUHG UHVXOWV RI WKH 8)6$ 5 FOHDQ FRUH VSDFHWLPH VWXGLHV $ FRQYHQWLRQDO ]HURWK PRPHQW )RXULHU WUDQVIRUPDWLRQ ZDV SHUIRUPHG RQ WKH H[SHULPHQWDO GDWD 7KH QXPHULFDO LQWHJUDWLRQ HPSOR\HG 6LPSVRQnV UXOH RI LQWHJUDWLRQ (TXDO WLPH VWHSV PLFURVHFf ZHUH XVHG WKURXJKRXW WKLV DQDO\VLV 7KH QXPEHU RI GDWD SRLQWV YDULHG IURP WR GHSHQGLQJ RQ ZKHUH D VWDEOH QHXWURQ EDFNJURXQG ZDV UHDFKHG DV GHWHUPLQHG E\ DQ VWDWLVWLFDO DQDO\VLV SHUIRUPHG LQ WKH 81,38/ SURn JUDP $ FRPSXWHU FRGH 025( ZDV FRGHG IRU WKLV SXUSRVHA $ FRQYHQWLRQDO ]HURWK PRPHQW )RXULHU WUDQVIRUPDWLRQ ZDV SHUIRUPHG RQ WKH VSDFHWLPH UHVXOWV REWDLQHG IURP WKH :,*/( FDOFXODn WLRQV ,W VKRXOG EH UHFDOOHG WKDW WKH :,*/( FDOFXODWLRQ HPSOR\V D VHULHV RI WLPH LQFUHPHQWV LQ GLIIHUHQW WLPH LQWHUYDOV 7R FRQIRUP WR WKLV VFKHPH 6LPSVRQnV UXOH RI LQWHJUDWLRQ ZLWK GLIIHUHQW WLPH LQFUHn PHQWV ZDV XVHG IRU HDFK WLPH LQWHUYDO 7UDSH]RLGDO LQWHJUDWLRQ ZDV XVHG WR EULGJH WKH JDS EHWZHHQ WKH GLIIHUHQW WLPH LQWHUYDOV $ FRPSXWHU SURJUDP 025:,* ZDV FRGHG IRU WKLV SXUSRVH 7KH DPSOLWXGHV DQG SKDVHV RI WKH ]HURWK )RXULHU PRPHQW DV D IXQFWLRQ RI IUHTXHQF\ REWDLQHG IURP WKH 025( DQG 025:,* QXPHULFDO WUDQVIRUPDWLRQ ZHUH OHDVWVTXDUHV ILWWHG WR REWDLQ WKH GDPSLQJ FRHIn ILFLHQW DQG WKH SKDVH VKLIW SHU XQLW OHQJWK IRU WKH IUHTXHQFLHV RI LQWHUHVW ,Q WKLV PDQQHU WKH WKHRUHWLFDO DQG H[SHULPHQWDO UHVXOWV LQ WKH 7KH 025( DQG 025:,* SURJUDPV ZHUH FRGHG E\ 'U 0 2KDQLDQ 8QLYHUVLW\ RI )ORULGD

PAGE 179

WLPH GRPDLQ ZHUH WUDQVIRUPHG WR WKH IUHTXHQF\ GRPDLQ ZLWK LGHQWLFDO QXPHULFDO DQG ILWWLQJ SURFHGXUHV VHWWLQJ D ILUP EDVLV IRU D RQHWR RQH FRPSDULVRQ 7KH IROORZLQJ LQWHUHVWLQJ QXPHULFDO SUREOHPV DUH ZRUWK\ RI QRWH Df 7R GHWHUPLQH WR ZKDW H[WHQW WKH QXPHULFDO WUDQVIRUPDWLRQ LV DIIHFWHG E\ WKH DFFXUDF\ RI WKH RSHUDWLRQV RI WKH ,%0 FRPSXWHU VDPSOH SUREOHPV ZHUH UXQ LQ VLQJOH SUHFLVLRQ E\WH ZRUGVf DQG GRXEOH SUHFLVLRQ E\WH ZRUGVf XVLQJ WKH 025:,* SURJUDP 1R GLIIHUHQFHV ZHUH IRXQG LQ WKH UHVXOWV XS WR WKH VL[WK VLJQLILFDQW ILJXUH EXW WKH GRXEOH SUHFLVLRQ FRPSXWDWLRQV FRQVXPHG DOPRVW WZLFH WKH DPRXQW RI WLPH DV WKH VLQJOH SUHFLVLRQ RQHV Ef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n FUHPHQWV FDUULHG IDUWKHU LQ WLPH DQG HYHU\ WLPH VWHS SXQFKHG IRU LQSXW WR WKH 025:,* SURJUDP $ WRWDO RI DQG SRLQWV UHVSHFWLYHO\ ZHUH QRZ DYDLODEOH IRU WKH )RXULHU WUDQVIRUPDWLRQ IRU WKH DQG PVHF LQSXW SXOVH FDVHV 7KH ODUJHVW WLPH LQFUHPHQW ZDV PLFURVHF 7KH FDOFXODWLRQ ZDV FDUULHG WR DERXW PVHF DIWHU WKH LQLWLDWLRQ RI WKH SXOVH $ VLJQLILFDQW LPSURYHPHQW ZDV QRWHG LQ WKH UHVXOWV RI WKH QXPHUn LFDO WUDQVIRUPDWLRQ 7KH UHVXOWV ZHUH VPRRWK WKURXJKRXW XS WR a FSV ZKHUH D FHUWDLQ DPRXQW RI VFDWWHULQJ ZDV REVHUYHG

PAGE 180

&RPSDULVRQ RI WKH 7KHRUHWLFDO DQG WKH 0HDVXUHG 5HVXOWV RI WKH 1HXWURQ :DYH $QDO\VLV 7R GHWHUPLQH WKH HIIHFW WKDW WKH LQSXW SXOVH ZLGWK FRXOG KDYH RQ WKH UHVXOWV RI WKH ZDYH DQDO\VLV WKH VWXG\ ZDV SHUIRUPHG IRU ERWK WKH DQG PVHF LQSXW SXOVH FDVHV 7KHVH H[SHULPHQWDO DQG WKHRUHWLFDO UHVXOWV LQ WKH WLPH GRPDLQ ZHUH SUHVHQWHG LQ 3DUW &KDSWHU ,9 RI WKLV ZRUN 6KRZQ LQ )LJV WKURXJK DUH WKH DPSOLWXGH DQG SKDVH RI WKH H[n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a FSV >@ 8QGRXEWHGO\ WKH IUHTXHQF\ FRQWHQW RI WKH QDUn URZ IDVW EXUVW LV VXSHULRU WR WKH VPHDUHG SXOVH WKDW LV QRUPDOO\ REn WDLQHG LI D WKHUPDOL]LQJ WDQN LV XVHG 7KH ZDYH DQDO\VLV RQ WKH :,*/( VSDFHWLPH UHVXOWV VKRZHG DQ LGHQWLFDO EHKDYLRU WKURXJKRXW WKH IUHn TXHQF\ UDQJH 7KH UHDO DQG WKH LPDJLQDU\ FRPSRQHQWV RI WKH FRPSOH[ LQYHUVH UHOD[DWLRQ OHQJWK DUH VKRZQ IRU ERWK WKHRU\ DQG H[SHULPHQW LQ 7DEOHV ;9, DQG ;9,, IRU WKH DQG WKH PVHF LQSXW SXOVHV UHVSHFWLYHO\ $ JUDSKLFDO FRPSDULVRQ RI WKHVH UHVXOWV LV VKRZQ LQ )LJV DQG 7KH IROORZLQJ FRPPHQWV DUH SHUWLQHQW Df $ GLVWLQFW EXW QRW VLJQLILFDQW GLIIHUHQFH LV IRXQG EHWZHHQ WKH UHVXOWV RI WKH DQG WKH PVHF FDVHV LQ ERWK WKHRU\ DQG H[SHULPHQW

PAGE 181

$PSOLWXGH 5HODWLYH 8QLWVf ),* $03/,78'( 2) =(527+ )285,(5 020(17 YV ',67$1&( )25 6(9(5$/ )5(48(1&,(6 06(& ,1387 38/6(

PAGE 182

$PSOLWXGH 5HODWLYH 8QLWVf ),* $03/,78'( 2) =(527+ )285,(5 020(17 YV ',67$1&( )25 6(9(5$/ )5(48(1&,(6 06(& ,1387 38/6(

PAGE 183

2 5 &RUH FP EHWZHHQ SRLQWV 3RVLWLRQ 1XPEHU 3+$6( 2) =(527+ )285,(5 020(17 YV ',67$1&( )25 6(9(5$/ )5(48(1&,(6 06(& ,1387 38/6(

PAGE 184

2 3RVLWLRQ 1XPEHU 3+$6( 2) =(527+ )285,(5 020(17 YV ',67$1&( )25 6(9(5$/ )5(48(1&,(6 06(& ,1387 38/6(

PAGE 185

7$%/( ;9, 7+( 5($/ $1' 7+( ,0$*,1$5< &20321(176 2) 7+( &203/(; ,19(56( 5(/$;$7,21 /(1*7+ 06(& ,1387 38/6( )5(48(1&< D FP nf UDGFPf FSVf 7KHRU\ ([S 7KHRU\ ([S

PAGE 186

7$%/( ;9,, 7+( 5($/ $1' 7+( ,0$*,1$5< &20321(176 2) 7+( &203/(; ,19(56( 5(/$;$7,21 /(1*7+ 06(& ,1387 38/6( )5(48(1&< D FPf " UDGFPf FSVf 7KHRU\ ([S 7KHRU\ ([S

PAGE 187

D R RV &? ),* &203$5,621 2) 7+( 7+(25(7,&$//< 35(',&7(' $1' 7+( 0($685(' '$03,1* &2()),&,(17 D

PAGE 188

UDGFPf 3 )UHTXHQF\ ),* &203$5,621 2) 7+( 7+(25(7,&$//< 35(',&7(' $1' 7+( 0($685(' 3+$6( 6+,)7 3(5 81,7 /(1*7+ e

PAGE 189

7KH SUHGLFWHG DQG PHDVXUHG YDOXHV DUH LQ H[FHOOHQW DJUHHPHQW XS WR FSV WKHRU\ IROORZV FORVHO\ WKH PHDVXUHG GLIIHUHQFH EHWZHHQ WKH DQG WKH PVHF FDVHV 7KH GLIIHUHQFHV LQ WKH GDPSLQJ FRHIILFLHQW DW ]HUR IUHTXHQF\ DUH VRPHZKDW SX]]OLQJ $ VWHDG\VWDWH PHDVXUHPHQW RI WKH LQYHUVH UHOD[DWLRQ OHQJWK \LHOGHG D YDOXH RI FP ? LQ H[FHOOHQW DJUHHPHQW ZLWK WKH H[SHULPHQWDO PVHF FDVH DQG a b ORZHU WKDQ WKH PVHF FDVH Ef 7KH WKHRUHWLFDO DQG H[SHULPHQWDO W[nV DQG enV VWDUW WR GLYHUJH VLJQLILFDQWO\ SDVW FSV 7KH GHYLDWLRQ LV ODUJHU IRU WKH PVHF LQSXW FDVH 7KH GHYLDWLRQ LV a b DW FSV IRU D DQG PVHF FDVHf DQG DERXW b DW FSV 6WLOO ERWK WKH WKHRUHWLFDO DQG H[SHULn PHQWDO UHVXOWV ZHUH VPRRWK IXQFWLRQV RI IUHTXHQF\ WKURXJKRXW WKH UDQJH ERWK FROODSVH GUDPDWLFDOO\ DW a FSV Ff $Q DQDO\VLV RI WKH OHDVWVTXDUHV ILWWLQJ SHUIRUPHG RQ ERWK WKH SUHGLFWHG DQG WKH PHDVXUHG YDOXHV UHYHDOV WKDW D SRRU ILW LV REWDLQHG DW FSV 7KH ILW JHWV LQFUHDVLQJO\ EHWWHU ZLWK IUHTXHQF\ XS WR FSV ZHUH IOXFWXDWLRQV DUH REVHUYHG LQ WKH YDOXHV 7KH FULWHULD XVHG WR VHOHFW WKH YDOXH RI D DQG e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n RXVO\ UHSRUWHG UHVXOWV >@ $ VPRRWK WUHQG LV IROORZHG XQWLO WKHRU\

PAGE 190

UDGFPf ),* 7+( 8)6$ 5 &25( S ',63(56,21 /$:

PAGE 191

DQG H[SHULPHQW DSSURDFK FSV fµ 7KH GLIIHUHQFHV REVHUYHG UHIOHFW WKRVH QRWLFHG ZKHQ WKH FRPSDULVRQ LQ WKH WLPH GRPDLQ ZDV FRQGXFWHG 7KH IDFW WKDW WKH :,*/( SXOVHV DUH FRQVLVWHQWO\ QDUURZHU WKDQ WKH H[SHULPHQWDO RQHV LV UHIOHFWHG LQ WKH VPDOOHU GDPSLQJ FRHIILFLHQWV DW ODUJH IUHTXHQFLHV 7KH GLIIHUHQFHV VKRZHG XS PXFK PRUH PDUNHGO\ LQ WKH XOWUDVHQVLWLYH S SODQH 6LQFH D D DQG ( ! WKH UHDO SDUW RI S D e f A WK H[S WK H[S U VHHPV WR LQGLFDWH D GLVFUHSDQF\ WKDW LV QRW DV ODUJH DV LW DSSHDUV 6KRZQ LQ )LJ LV WKH UHDO SDUW RI S IRU ERWK WKHRU\ DQG H[SHUn LPHQW DV D IXQFWLRQ RI IUHTXHQF\ $OWKRXJK WKH\ VWDUW WR GLYHUJH VLJn QLILFDQWO\ DW I FSV LW VKRXOG EH UHPDUNHG WKDW DQ\ VPDOO GLI IHUHQFHV LQ D DQG e DUH PDJQLILHG RXW RI SURSRUWLRQ LQ WKH S SODQH ,Q FRQWUDVW WKH LPDJLQDU\ SDUW RI S D DSSHDUV WR VLJQLI\ SHUIHFW DJUHHPHQW LQ WKH UHVXOWV VLQFH WKH GHYLDWLRQV LQ D DQG e FDQFHO HDFK RWKHU VHH )LJ f 7KH S GLVSHUVLRQ ODZ RI WKH 8)6$ 5 FOHDQ FRUH LV VKRZQ LQ )LJ 7KLV LV D PRVW VHQVLWLYH LQGH[ $Q\ VPDOO HUURU LQ WKH GDWD RU IDLOXUH RI WKH PRGHO ZLOO LPPHGLDWHO\ DSSHDU EORZQXS LQ S $OWKRXJK WKH WKHRU\ DQG H[SHULPHQW GLYHUJH LQ WKH S SODQH DIWHU FSV ERWK EHKDYH VPRRWKO\ WKH GLIIHUHQFHV VKRXOG EH FRQVLGHUHG ZLWKLQ WKH FRQn WH[W RI WKH RYHU DOO DQDO\VLV &HUWDLQO\ LW FDQ EH VDLG WKDW WKH :,*/( UHVXOWV GR QRW IROORZ WKH H[SHULPHQWDO UHVXOWV WKURXJKRXW WKH IUHTXHQF\ UDQJH EXW WKDW LV WR H[SHFWHG WKH DFWXDO GLVFUHSDQFLHV REVHUYHG LQ WKH WLPH GRPDLQ DQG WKH QXPHULFDO XQFHUWDLQWLHV SURSDJDWH LQWR WKH S SODQH VR VLJQLILFDQWO\ WKDW SHUIHFW DJUHHPHQW VHHPV WR EH XQUHDFKDEOH ZLWK WKH SUHVHQW PRGHOV DQG NQRZOHGJH RI SDUDPHWHUV 7KH VWHDG\VWDWH LQYHUVH UHOD[DWLRQ OHQJWK DQG WKH FRUUHVSRQGLQJ FWIBRf LV VLJQLILFDQWO\ VPDOOHU WKDQ WKH G\QDPLF LQYHUVH UHOD[DWLRQ

PAGE 192

([S 2 Â’ V ),* )UHTXHQF\ &203$5,621 2) 7+( 7+(25(7,&$//< 35(',&7(' $1' 7+( 0($685(' F[ef

PAGE 193

)UHTXHQF\ ),* &203$5,621 2) 7+( 7+(25(7,&$//< 35(',&7(' $1' 7+( 0($685(' D .

PAGE 194

ORRR ! ),* 7+( 8)6$ 5 &25( S ',63(56,21 /$:

PAGE 195

&+$37(5 9, 63$7,$/ '(3(1'(1&( 2) 38/6('1(87521 5($&7,9,7< 0($685(0(176 ,QWURGXFWLRQ $ UHVXPH RI SXOVHGQHXWURQ UHDFWLYLW\ PHDVXUHPHQWV ZDV SUHVHQWHG LQ 3DUW &KDSWHU 9 RI WKLV ZRUN 'XULQJ WKH QXFOHDU FDOLEUDWLRQ RI WKH 8)6$ DVVHPEO\ FHUWDLQ VSDWLDO HIIHFWV ZHUH REVHUYHG LQ WKH GHWHUn PLQDWLRQ RI WKH UHDFWLYLW\ RI WKH ORQJ DVVHPEO\ ,Q WKLV FKDSWHU WKHVH HIIHFWV DUH LQYHVWLJDWHG 7KH SXOVH SURSDJDWLRQ PHDVXUHPHQWV SHUIRUPHG LQ WKH 8)6$ 5 FOHDQ FRUH ZHUH DQDO\]HG XVLQJ WKH *DUHOLV5XVVH8WHFKQLTXH >@ WR GHWHUPLQH WKH UDWLR N% WKH IXQGDPHQWDO GHFD\ FRQVWDQW ZDV REWDLQHG E\ DSn SO\LQJ 3HLHUOf¬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

PAGE 196

\LHOGHG WKH GHFD\ FRQVWDQWV DW GLIIHUHQW WLPHV DIWHU WKH SXOVH DQG FRPSXWHG D YDOXH RI S f DV ZHOO DV DEVROXWH S DQG NAA XVLQJ D FDOFXODWHG IF!HIe IRU HDFK GHFD\ FRQVWDQW 7KH 'HFD\ &RQVWDQW $ VHULHV RI GHFD\ FRQVWDQWV ZHUH REWDLQHG E\ DQDO\]LQJ WKH WDLO RI WKH SXOVH DW LQFUHDVLQJ WLPHV IURP WKH SHDN 7KH FRQYHUJHQFH RI WKH GHFD\ FRQVWDQW LQ WLPH ZDV VRXJKW EHIRUH EDFNJURXQG LV UHDFKHG DQG WKH YDOXH WKDW JDYH WKH OHDVW GHYLDWLRQ DFFRUGLQJ WR 3HLHUOf¬V VWDWLVn WLFDO DQDO\VLV ZDV FRQVLGHUHG WR JLYH WKH GHFD\ FRQVWDQW DW WKDW SRVLn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f UHYHDOV WKDW LQGHHG PRUH WKDQ PVHF DUH UHTXLUHG IRU WKH DVVHPEO\ WR DFKLHYH DQ DV\PSWRWLF QHXWURQ GLVWULEXWLRQ 7KHUHIRUH RQO\ DW ODUJH GLVWDQFHV IURP WKH VRXUFH FDQ D WUXH IXQGDPHQWDO GHFD\ FRQVWDQW EH PHDVXUHG SULRU WR 3 FPf EDFNJURXQG LV UHDFKHG EHIRUH D XQLIRUP VSDWLDO GLVWULEXWLRQ LV HVWDEOLVKHG

PAGE 197

6KRZQ LQ )LJ DUH WKH H[SHULPHQWDOO\ GHWHUPLQHG GHFD\ FRQVWDQWV DV D IXQFWLRQ RI SRVLWLRQ LQ WKH DVVHPEO\ IRU D DQG PVHF LQSXW SXOVH $V GLVWDQFH WR WKH VRXUFH LQFUHDVHG PRUH ZDLWLQJ WLPH ZDV DYDLODEOH WR H[WUDFW WKH GHFD\ FRQVWDQW 7KH YDOXHV IRXQG SDVW 3 DUH EHOLHYHG WR UHSUHVHQW WKH IXQGDPHQWDO GHFD\ FRQVWDQW RI WKH DVVHPEO\ ,W VKRXOG EH QRWHG WKDW WKH LQVWDQWDQHRXV GHFD\ FRQVWDQW DW GLIn IHUHQW WLPHV DIWHU WKH SXOVH ZDV FDOFXODWHG IRU DOO WKH GHWHFWRU SRVLn WLRQV LQ WKH DVVHPEO\ XVLQJ WKH :,*/( FRGH 7KH UHVXOWV DJUHH FORVHO\ ZLWK WKRVH PHDVXUHG H[SHULPHQWDOO\ DW GLIIHUHQW WLPHV DIWHU WKH SHDN 7KH WZRJURXS VSDFHWLPH GHSHQGHQW PRGHO GHVFULEHV ZHOO WKH REVHUYHG SKHQRPHQD 7KH 5DWLR N 7KH *DUHOLV5XVVH8WHFKQLTXH ZDV DSSOLHG WR WKH WZRPHGLXP V\VWHP UHIHUUHG WR DV WKH 8)6$ 5 FRUH 7KH PHDVXUHG NJ YDOXHV GHWHUPLQHG DV D IXQFWLRQ RI GLVWDQFH IURP WKH VRXUFH VKRZHG D VSDWLDO GHSHQGn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

PAGE 198

'HFD\ &RQVWDQW VHF ),* '(&$< &2167$17 96 $;,$/ 326,7,21

PAGE 199

f§ 2 PVHF LQSXW SXOVH ZLGWK 2 PVHF LQSXW SXOVH ZLGWK [f n2 F ; n S S LI n ‘ 0 0 f§2r f§ &" n‘ ‘ BB 4 R R R 4 8)6$ 5L &OHDQ &RUH FP EHWZHHQ GDWD SRLQWV " L LOO , O , , L W , , , 3RVLWLRQ 1XPEHU YV $;,$/ 326,7,21

PAGE 200

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n QLTXH DUH SHUWLQHQW $ GRPLQDQW IDFWRU LQ WKH GHWHUPLQDWLRQ RI NJ LV WKH GHOD\HG QHXWURQ EDFNJURXQG ZKLFK LV VWURQJO\ GHSHQGHQW RQ WKH UHSHWLWLRQ UDWH RI WKH QHXWURQ JHQHUDWRU ZKLFK VKRXOG EH GHWHUPLQHG YHU\ DFFXUDWHO\f 0HDQLQJIXO YDOXHV RI NJ DUH REWDLQHG IDU IURP WKH VRXUFH IRU QRW WRR ZLGH LQSXW SXOVHV $ FRUUHFWLRQ RQ WKH GHWHUPLQHG NJ YDOXHV WR DFFRXQW IRU WKH ILQLWH ZLGWK RI WKH LQSXW SXOVHV ZDV GHYHORSHG E\ * $ 0RUWHQVHQ 1XFOHDU 6DIHW\ 'LYLVLRQ 3KLOOLSV 3HWUROHXP &R >@ DIWHU WKH H[SHULPHQWDO REVHUYDWLRQ RI WKH SKHQRPHQD 7KH VXJJHVWHG PHWKRG ZDV DSSOLHG WR VHYHUDO RI WKH FDVHV XQGHU FRQVLGn HUDWLRQ EXW DQ RYHUFRUUHFWLRQ VHHPV WR UHVXOW WKH FRUUHFWHG YDOXHV NHHS FKDQJLQJ GUDVWLFDOO\ DW ORQJ GLVWDQFHV IURP WKH VRXUFH ZKHUH WKH LQIOXHQFH RI WKH LQSXW SXOVH LV QHJOLJLEOH )XUWKHU ZRUN LQ WKLV FRUn UHFWLRQ VKRXOG LPSURYH WKH UHVXOWV 7KH 0HDVXUHG 5HDFWLYLW\ DQG NHIe 9DOXHV 7KH UHDFWLYLW\ LQ GROODUV LV REWDLQHG IROORZLQJ WKH *DUHOLV 5XVVHOO PHWKRG E\ WKH H[SUHVVLRQ 3f D NJ

PAGE 201

7$%/( ;9,,, 7+( '(&$< &2167$17 $1' NJ$ 9$/8(6 0($685(' $6 $ )81&7,21 2) ,1387 38/6( :,'7+ 06(& ,1387 38/6( :,'7+ 8)6$ 5 &OHDQ &RUH 0: 5DWLRQ FP ZLGH UHIOHFWHG FRUH 38/6( :,'7+ PVHFf D VHF Ne VHF

PAGE 202

8VLQJ D FDOFXODWHG YDOXH RI IRU WKLV DVVHPEO\ FRQILJXUDn WLRQ WKH DEVROXWH YDOXH RI WKH UHDFWLYLW\ S LV REWDLQHG 7KH FRUn UHVSRQGLQJ YDOXH RI NA LV WKHQ IRXQG E\ WKH UHODWLRQ N f§f§ HII S 6KRZQ LQ )LJ DUH WKH PHDVXUHG DEVROXWH UHDFWLYLW\ DQG NHee DV D IXQFWLRQ RI SRVLWLRQ LQ WKH ODWWLFH ,W LV REYLRXV IURP ZKDW KDV EHHQ VDLG DERYH UHJDUGLQJ WKH EHKDYLRU RI D DQG NJe WKDW D PHDQLQJIXO YDOXH IRU HLWKHU N RU S FDQQRW EH REWDLQHG LQ WKLV FDVH XQWLO r FP VHSDUDWH WKH VRXUFH DQG GHWHFWRU 7KH DVVLJQHG YDOXH RI NAA ZDV 7KLV YDOXH FRPSDUHV YHU\ IDYRUDEO\ ZLWK D YDOXH RI IRXQG E\ LQYHUVH PXOWLSOLFDWLRQ PHDVXUHPHQWV DQG D FDOFXn ODWHG YDOXH RI $OWKRXJK WKH DJUHHPHQW REWDLQHG FRXOG EH VRPHn ZKDW IRUWXLWRXV LW FDQ EH VDLG WKDW D UHSUHVHQWDWLYH YDOXH RI WKH UHDFWLYLW\ ZDV IRXQG ,W VHHPV WKDW LI SURSHU FDUH LV WDNHQ WKH SXOVHG UHDFWLYLW\ PHDVXUHPHQWV ZLOO \LHOG JRRG UHVXOWV LQ ODUJH PXOWLn SO\LQJ PHGLD 5HJLRQZLVH 'HSHQGHQFH RI WKH 5HDFWLYLW\ 0HDVXUHPHQWV %HFNHU DQG 4XLVHQEHUU\ DV ZHOO DV :DOWDU DQG 5XE\ >@ KDYH SRLQWHG RXW WKH VSDWLDO GHSHQGHQFH RI SXOVHG VRXUFH UHDFWLYLW\ PHDVXUHn PHQWV LQ WZRPHGLD V\VWHPV 'LIIHUHQFHV LQ WKH PHDVXUHG UHDFWLYLW\ FDQ EH SUHVHQW ZKHWKHU WKH GHWHFWRU LV ORFDWHG LQ WKH FRUH RU LQ WKH UHIOHFn WRU $ VHULHV RI SXOVHG VRXUFH WUDYHUVHV ZHUH FRQGXFWHG DFURVV WKH 8)6$ 5 FOHDQ FRUH DQG UHIOHFWRUV DW WKUHH GLIIHUHQW GLVWDQFHV IURP WKH VRXUFH 1R GLIIHUHQFHV ZHUH IRXQG IRU WKLV FRUH LQ WKH UHDFWLYLW\ PHDVXUHG LQ WKH FRUH RU LQ WKH UHIOHFWRU 6KRZQ LQ 7DEOH ;,; LV D

PAGE 203

5HDFWLYLW\ f ),* 5($&7,9,7< f $1' N 96 $;,$/ 326,7,21 HII (IIHFWLYH 0XOWLSOLFDWLRQ )DFWRU

PAGE 204

7$%/( ;,; 5(*,21:,6( '(3(1'(1&( 2) 7+( 5($&7,9,7< 0($685(0(176 0HDVXUHPHQW DFURVV WKH FRUH ZLGWK 8)6$ 5 &OHDQ &RUH 0: 5DWLR FP ZLGH UHIOHFWHG FRUH 326,7,21 12 ',67$1&( 72 &25( &(17(5 FPf D ,V YVHF f N %L VHF 3 N HII $ % & & ' ( ) * + , 5( 5( 5(

PAGE 205

OLVW RI WKH GHFD\ FRQVWDQWV NJe S DQG NHA PHDVXUHG DFURVV WKH FRUH DQG VLGH UHIOHFWRU DW D GLVWDQFH RI FP IURP WKH VRXUFH 7KH FRUHUHIOHFWRU FRQILJXUDWLRQ LQYHVWLJDWHG LV WRR QDUURZ WR REVHUYH DQ\ WUDQVYHUVH SURSDJDWLRQ RI WKH SXOVH FRQVHTXHQWO\ QR GLIn IHUHQFH LQ WKH YDOXH RI WKH UHDFWLYLW\ ZDV PHDVXUHG DFURVV WKH V\VWHP 7KH YDOXH RI D DQG RI NJe REWDLQHG QHDU WKH RXWHU HGJH RI WKH UHIOHFWRU ZDV ORZ FRPSDUHG WR WKH RWKHU YDOXHV EXW VRPHKRZ FRPSHQVDWHG HDFK RWKHU WR PDLQWDLQ S DSSUR[LPDWHO\ FRQVWDQW

PAGE 206

&+$37(5 9,, &21&/86,216 7KH VSDFHWLPH NLQHWLFV EHKDYLRU RI D ODUJHLQRQHVSDFH GLPHQVLRQ VLGHUHIOHFWHG KLJKO\ PXOWLSOLFDWLYH NAA a f VXEFULWLFDO DVVHPEO\ KDV EHHQ VWXGLHG E\ WKH XVH RI WKH SXOVH SURSDJDWLRQ WHFKQLTXH 7KH H[SHULPHQWDO UHVXOWV ZHUH XVHG WR WHVW WKH SUHGLFWLRQV RI WKH RQHn GLPHQVLRQDO WZRJURXS VSDFHWLPH GLIIXVLRQ WKHRU\ FDOFXODWLRQDO VFKHPH :,*/(f 7KH DQDO\VLV RI ERWK WKH WKHRUHWLFDOO\ SUHGLFWHG DQG H[SHULn PHQWDOO\ PHDVXUHG UHVXOWV ZDV SHUIRUPHG LQ WKH WLPH DV ZHOO DV WKH IUHn TXHQF\ GRPDLQ 7KH FRPSDULVRQ RI WKHRU\ DQG H[SHULPHQW LQ WKH WLPH GRPDLQ VKRZV JRRG DJUHHPHQW :,*/( DFFXUDWHO\ SUHGLFWHG WKH GHOD\ WLPHV DQG WKH DWWHQXDWLRQ RI WKH SHDN RI WKH SXOVH ZKHQ D ILUVW IOLJKW NHUQHO ZDV XVHG WR GHVFULEH WKH VSDWLDO GLVWULEXWLRQ RI WKH IDVW VRXUFH $W ODUJH GLVn WDQFHV IURP WKH VRXUFH WKH WLPH SURILOHV SUHGLFWHG E\ :,*/( DUH FRQn VLVWHQWO\ QDUURZHU WKDQ WKH PHDVXUHG SXOVH VKDSHV 7KH VHQVLWLYLW\ RI WKH RQHGLPHQVLRQDO PRGHO WR VPDOO FKDQJHV LQ WKH WUDQVYHUVH EXFNOLQJ ZDV VWXGLHG :,*/( DQG H[SHULPHQWf VKRZHG D ODUJH VHQVLWLYLW\ WR WKHVH FKDQJHV 7KHVH UHVXOWV SRLQW WR WKH QHFHVVLW\ RI REWDLQLQJ WKH EHVW SRV VLEOH HVWLPDWH RI WKH QXFOHDU SDUDPHWHUV WR EH XVHG LQ VSDFHWLPH NLn QHWLFV FDOFXODWLRQV 7KH FRPSDULVRQ RI WKHRU\ DQG H[SHULPHQW LQ WKH VHQVLWLYH IUHTXHQF\ GRPDLQ FRQILUPV WKH JRRG DJUHHPHQW IRXQG LQ WKH WLPH GRPDLQ 7KH

PAGE 207

FRPSDULVRQ RI WKH :,*/( UHVXOWV DQG WKH SXOVH SURSDJDWLRQ GDWD ZHUH SXW RQ WKH VDPH EDVLV E\ SHUIRUPLQJ LGHQWLFDO )RXULHU WUDQVIRUPDWLRQV DQG ILWWLQJ SURFHGXUHV RQ ERWK 7KH SUHGLFWHG DQG PHDVXUHG UHDO SDUW RI S D e f GLYHUJH SDVW FSV 7KLV LV DOPRVW D QDWXUDO RFFXUUHQFH LQ WKH XOWUDVHQVLWLYH S SODQH LQ ZKLFK WKH RSSRVLQJ GLVFUHSDQFLHV RI D DQG e DUH EORZQ RXW RI SURSRUWLRQ 7KH DJUHHPHQW LQ WKH S GLVSHUn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n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

PAGE 208

$33(1',; $ &$/&8/$7,21$/ 352&('85(6 86(' ,1 7+( '(7(50,1$7,21 2) 7+( 18&/($5 3$5$0(7(56 $1' 7+( N 9$/8(6 HII

PAGE 209

'HWHUPLQDWLRQ RI WKH )RXU *URXS 3DUDPHWHUV $ IRXUJURXS GLIIXVLRQ DSSURDFK ZDV XVHG WR FDOFXODWH WKH HIIHFWLYH PXOWLSOLFDWLRQ IDFWRU RI WKH VL[ 8)6$ FRQILJXUDWLRQV WR EH VWXGLHG DV 3KDVH , RI WKH ODUJH FRUH G\QDPLFV H[SHULPHQWDO SURJUDP 7KH FDOFXODn WLRQ RI WKH SDUDPHWHUV ZDV SHUIRUPHG E\ WKH 1XFOHDU 6DIHW\ 5HVHDUFK %UDQFK RI WKH 3KLOOLSV 3HWUROHXP &RPSDQ\ >@ 7KH IRXU HQHUJ\ JURXSV ZHUH GLYLGHG DV IROORZV *URXS ; (9 *URXS ; ; (9 *URXS ; (9 *URXS (9 7R REWDLQ WKH IRXUJURXS FRQVWDQWV WKH IROORZLQJ VHTXHQFH RI FDOn FXODWLRQV ZDV SHUIRUPHG IRU HDFK RI WKH VL[ FDVHV PHQWLRQHG )DVW JURXS JURXSV f FRQVWDQWV ZHUH FDOFXODWHG E\ WKH 3+52* >@ FRPSXWHU SURJUDP 7KHVH FDOFXODWLRQV LQFOXGH UHVRQDQFH LQWHJUDO DQG 'DQFRII FRUUHFWLRQ FDOFXODWLRQV DV SURYLGHG E\ WKH 5$9(1 >@ FDO FXODWLRQDO VFKHPH 7KHUPDO FRQVWDQWV ZHUH FDOFXODWHG E\ WKH 727(0 >@ FRPSXWHU SURJUDP 7KH FRQVWDQWV XVHG IRU WKH FRUH REMHFW RI WKLV ZRUN DUH JLYHQ EHORZ ,W VKRXOG EH PHQWLRQHG WKDW WKH SURFHGXUH GHVFULEHG ZDV DOVR HPSOR\HG WR GHWHUPLQH WKH WZRJURXS SDUDPHWHUV XVHG IRU WKH :,*/( FDOFXODWLRQV

PAGE 210

8)6$ 5 &25( 3$5$0(7(56 0: UDWLR FP ZLGH UHIOHFWHG FRUH *URXS ' AD =U B YOI Y 2[ [ [ a [ B 9RLG &RHIILFLHQW RI 5HDFWLYLW\ [ $Nb YRLG 7HPSHUDWXUH &RHIILFLHQW RI 5HDFWLYLW\ [ A $Nr& 'HWHUPLQDWLRQ RI N HII 8VLQJ WKH FRQVWDQWV WDEXODWHG DERYH HLJHQYDOXH FDOFXODWLRQV ZHUH SHUIRUPHG ZLWK WKH $,0 FRPSXWHU FRGH >@ n $,0 LV D PXOWLJURXS PXOWLUHJLRQ RQHGLPHQVLRQDO GLIIXVLRQ WKHRU\ FRGH WKDW KDV EHHQ ZLGHO\ XVHG IRU FULWLFDOLW\ FDOFXODWLRQV 7KH FRGH ZLOO KDQGOH DV PDQ\ DV HQHUJ\ JURXSV VSDFH SRLQWV DQG UHJLRQV %DVLFDOO\ WKH FRGH VROYHV WKH HTXDWLRQ 'A9LUf Uf [ALUf A ( ML I!MUf L -T V L 12* 12* QXPEHU RI JURXSV 7KH V\PEROV IROORZ FRQYHQWLRQDO QRWDWLRQ DQG ;r WKH LQWHJUDO RI WKH ILVVLRQ VSHFWUXP RYHU WKH OHWKDUJ\ UDQJH UHSUHVHQWHG E\ JURXS L

PAGE 211

1* (IfLc!LUf VUf ] f§/ L O ; ZKHUH ; LV WKH HLJHQYDOXH 7KH FRGH WKHQ VHWV DQG VROYHV D VHW RI ILQLWH GLIIHUHQFH HTXDWLRQV

PAGE 212

$33(1',; % '(6&5,37,21 2) 7+( &20387(5 352*5$06 81,38/ 025( 025:,* $/;,/6 08/7,3/27

PAGE 213

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nV VWDWLVWLFDO PHWKRG >@ 7KH GHFD\ FRQVWDQWV FDQ EH FDOFXODWHG DW GLIn IHUHQW ZDLWLQJ WLPHV IURP WKH SHDN RI WKH SXOVH WR GHWHUPLQH WKH WLPH FRQYHUJHQFH 7KH DnV DUH VWRUHG LQ &20021 IRU WKH UHDFWLYLW\ GHWHUPLQDn WLRQ 7KH EDFNJURXQG LV VXEWUDFWHG IURP WKH GDWD $ WHVW LV PDGH WR ILQG LI DQ\ FKDQQHO KDV D QHJDWLYH QXPEHU RI FRXQWV ,I D IHZ FKDQQHOV KDYH QHJDWLYH FRXQWV WKHLU YDOXH LV VHW WR ]HUR 7KH QXPEHU RI 7KH FRRSHUDWLRQ RI 'U 0 2KDQLDQ LQ WKH FRGLQJ RI WKHVH SURJUDPV LV JUDWHIXOO\ DFNQRZOHGJHG

PAGE 214

FKDQQHOV LV QRZ UHGXFHG WR DQ RGG QXPEHU RI SRLQWV WHUPLQDWLQJ ZKHUH WKH EDFNJURXQG KDV EHHQ GHWHUPLQHG WR VWDUW 'HWHUPLQDWLRQ RI WKH UDWLR NIe IROORZLQJ WKH *DUHOLV5XVVHO PHWKRG 6LPSVRQnV LQWHJUDWLRQ LV XVHG IRU WKH SXOVH DQG WKH 5HJXOD )DOVL LWHUDWLYH PHWKRG LV XVHG WR ILQG WKH URRW 7KH GDWD LV QRUPDOL]HG DFFRUGLQJ WR D VHOHFWHG QRUPDOL]DWLRQ VFKHPH VHH 3DUW , &KDSWHU ,9f 7KH QRUPDOL]HG GDWD FDQ EH SXQFKHG LQ FDUGV IRU IXUWKHU DQDO\VLV ,I D QHXWURQ ZDYH W\SH DQDO\VLV LV GHVLUHG D QXPHULFDO )RXULHU WUDQVIRUPDWLRQ LV SHUIRUPHG DQG WKH DPSOLWXGH DQG SKDVH RI WKH SXOVH IRU HDFK IUHTXHQF\ VHOHFWHG LV SULQWHG DQG SXQFKHG IRU IXUWKHU SURFHVVLQJ 81,38/ ZDV XVHG WR SURFHVV WKH SXOVH GDWD SXQFKHG LQ FDUGV IURP RULJLQDO SHUIRUDWHG SDSHU WDSH DW WKH ,%0 FRPSXWHU RI WKH 'HSDUWPHQW RI 1XFOHDU (QJLQHHULQJ 6FLHQFHV 8QLYHUVLW\ RI )ORULGDf WKH SDSHU WDSH LV WKH RXWSXW RI WKH PXOWLFKDQQHO DQDO\]HU XVHG LQ WKH H[SHULPHQWDWLRQ 025( 7KH )2575$1 ,9 ,%0 FRPSXWHU SURJUDP 025( SHUIRUPV D FRQYHQn WLRQDO QXPHULFDO )RXULHU WUDQVIRUPDWLRQ RI WKH SXOVH QHXWURQ GDWD ]HURWK PRPHQW RQO\f IRU WKH DQDO\VLV RI WKH H[SHULPHQW LQ WKH IUHn TXHQF\ GRPDLQ 7KH WUDQVIRUPDWLRQ XVHG HTXDO WLPH VWHSV D PD[LPXP RI WLPH SRLQWV DQG 6LPSVRQnV LQWHJUDWLRQ VFKHPH $Q RGG QXPEHU RI SRLQWV DUH UHTXLUHG E\ WKH SURJUDP 7KHUPDO QHXWURQ GDWD FDQ EH HQWHUHG GLUHFWO\ DV LQSXW RU WKH WRWDO DQG HSLFDGPLXP QHXWURQ IOX[ DUH UHTXLUHG WR REWDLQ WKH WKHUPDO IOX[ E\ VXEWUDFWLRQ

PAGE 215

025( SXQFKHV LQ FDUGV WKH DPSOLWXGH DQG SKDVH RI WKH ]HURWK )RXULHU PRPHQW RI WKH SXOVH IRU LQSXW WR WKH $/;,/6 SURJUDP $ FRPn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n WLRQ SHUIRUPHG E\ HLWKHU WKH 025( RI WKH 025:,* SURJUDP $ OLQHDU OHDVWVTXDUHV VFKHPH LV XVHG 7KH YDOXHV RI $/3+$ DUH REWDLQHG E\ ILW WLQJ WKH ORJ DPSOLWXGH YV GLVWDQFH IRU HDFK IUHTXHQF\ 7KH YDOXHV RI ;, DUH REWDLQHG E\ D OLQHDU ILW RI WKH SKDVH DQJOH YV GLVWDQFH 7KH FKDQQHO FKRSSLQJ WHFKQLTXH LV XVHG WR GHWHUPLQH WKH FRQYHUJHQFH RI D DQG e LQ VSDFH 7KH ZHLJKWLQJ RI WKH SRLQWV FDQ EH VHW WR REVHUYHG YDOXHf RU WR HTXDO ZHLJKWV 6WDWLVWLFDO TXDQWLWLHV GHWHUPLQLQJ WKH JRRGQHVV RI WKH ILW DUH FDOFXODWHG IRU HDFK ILW $ FRPSOHWH GHVFULS WLRQ FDQ EH IRXQG LQ 5HJ 08/7,3/27 7KH )2575$1 ,9 ,%0 FRPSXWHU SURJUDP 08/7,3/27 XVHG WKH RXWn SXW RI WKH 81,38/ DQGRU WKH :,*/( FRGHV WR SORW WKH WLPH SURILOHV RI

PAGE 216

WKH SXOVHV $ &$/&203 SORWWHU LV XVHG LQ FRQQHFWLRQ ZLWK WKH FRPSXWHU IRU WKH DFWXDO SORWWLQJ RI WKH SXOVH 7KH SURJUDP LV FDSDEOH RI QRUPDOL]LQJ WR WKH SHDN FRXQWV RI D UHIHUHQFH SXOVH IRU WKH VXSHUn SRVLWLRQ RI UHODWHG WLPH SURILOHV

PAGE 217

$33(1',; & 8)6$ 5O&/($1 &25( 7,0( 352),/(6 2) 7+(50$/ 1(87521 )/8; $7 1,1(7((1 &25( 326,7,216 )25 ,1387 38/6(6 2) $1' 06(&

PAGE 218

)/8; 5(/$7,9( 81,76f ),* &O 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 219

IOX[ UHODWLYH XQLWVf 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; r (0 )520 7+( 6285(n( ),* &

PAGE 220

5(/$7,9( 81,76f ,' 7+(50 62 3($. $7 06(('6 38/6( :,'7+ 06(&6 7+(&5< ):+9W 06((&6 (,;3 fµ ):+0 06(&6 % 92 92 7,0( 06(Lf & 7,0( 35&)n,/( ') 7+(50$/ 1(87521 )/8; 6% (0 )520 7+( 6285&(

PAGE 221

8; 5(/$7,9( 81,76f ),*} & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; } (0 )520 7+( 6285&(

PAGE 222

)/8; 5(/$7,9( 81,76f r , r} r n ? 7+(50 n 3($. $7 06(n&6 38/6( :,'7+ 06(&6 7+(&5< ):+0 06(&6 (;3 ):+0 06(&6 * , 6 V r m r b b K2 2 ^f§ 9 rrrrrr rrrmm L / rrrrrmfµ}rf§ }}>‘ 7,0( 06(&' ),* & 7,0( 35'),/( *) 7+(50$/ 1(875*1 )/8; r %6 *0 )5'0 7+( 6285&(

PAGE 223

)/8; 5(/$7,9( 81,7(f 7,0( 06(&f 7,0( )5&),/( &) 7+(50$/ 1(87521 )/8; ), &/ & %6 &0 )520 7+( 6'-5&( PXf

PAGE 224

)/8; 5(/$7,9( 81,76f ),*r & 7,0( 352),/( ') 7+(50$/ 1(875*1 )/8; r &0 )520 7+( 6*85&(

PAGE 225

8; 5(/$7,9( 81,76f ),* & 7,0( 352),/( *) 7+(50$/ 1(875*1 )/8; &0 )520 7+( 6285&(

PAGE 226

)/8; 5(/$7,9( 81,76f ; ),*r & WLPH 352),/( &) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 227

)/8; 5(/$7,9( 81,7(f ),* &,2 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 228

)/8; 5(/$7,9( 81,7(' I ),* &OO WLPH 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6'85&(

PAGE 229

3n/ž; 5(/$7,9( 81,76f ),*r & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; %% &0 )520 7+( 6285&(

PAGE 230

)/8; 5(/$7,9( 81,76f ),*r & 7,0( 352),/( ') 7+(50$/ 1(875'1 )/8; &0 )520 7+( 6285&(

PAGE 231

)/8; 5(/$7,9( 81,76f ),*} & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; ,%6 &0 )520 7+( 6285&(

PAGE 232

)/8; 5(/$7,9( 81,76f ),* & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 233

)/8; 5(/$7,9( 81,76f ),* & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 234

)/8; 5(/$7,9( 81,76f ),* & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; 3&6 &0 )520 7+( 6285&(

PAGE 235

)/8; 5(/$7,9( 81,76f ),* & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 236

)/8; 5(/$7,9( 81,76f ),*r & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; 44 &0 )520 7+( 6285&(

PAGE 237

)/8; 5(/$7,9( 81,76f ),* & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 238

)/8; 5(/$7,9( 81,76f ),*r & 7,0( 352),/( *) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6&85&(

PAGE 239

)/8; 5(/$7,9( 81,76f ),* & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; % &0 )520 7+( 6285&(

PAGE 240

)/8; 5(/$7,9( 81,76f ),*r & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; r &0 )520 7+( 6285&(

PAGE 241

)/8; 5(/$7,9( 81,76f ),* & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; * &0 )520 7+( 6285&(

PAGE 242

)/8; 5(/$7,9( 81,76f ),* & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; (* &0 )520 7+( 6285&(

PAGE 243

)/8; 5(/$7,9( 81,76f ),' & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 244

)/8; 5(/$7,9( 81,76f ),*r & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 245

)/8; 5(/$7,9( 81,76f ),* & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 246

)/8; 5(/$7,9( 81,76f ),* & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 247

)/8; 5(/$7,9( 81,76f ),'r & WLPH 352),/( 2) 7+(50$/ 1(87521 )/8; ,62r,% &0 )520 7+( 6285&(

PAGE 248

)/8; 5(/$7,9( 81,76f ),*} & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 249

)/8; 5(/$7,9( 81,76f ),*} & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 250

)/8; 5(/$7,9( 81,76f ),*} & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; ,%%r &0 )520 7+( 6285&(

PAGE 251

)/8; 5(/$7,9( 81,76f ),* & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 252

)/8; 5(/$7,9( 81,76f ),* & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 253

)/8; 5(/$7,9( 81,76f ),*r & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; (*6 &0 )520 7+( 6285&(

PAGE 254

)/8; 5(/$7,9( 81,76f ),*r & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; r &0 )520 7+( 6285&(

PAGE 255

)/8; 5(/$7,9( 81,76f ),* & 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; (% &0 )520 7+( 6285&(

PAGE 256

$33(1',; ' 8)6$ 5 &/($1 &25( 7,0( 352),/(6 2) )$67 1(87521 )/8; $7 6(9(5$/ &25( 326,7,216 )25 ,1387 38/6(6 2) $1' 06(&

PAGE 257

)/8; 5(/$7,9( 81,76f ),*} ' 7,0( 352),/( 2) )$67 1(87521 )/8; } &0 )520 7+( 6285&(

PAGE 258

)/8; 5(/$7,9( 81,76f r ( * r IW , 7 ) 5,)6 3($. $7 06(&6 SXOVH ZLGWK RV PVHFV 7+(25< ):+0 06(& (;3 ):+0 06(& WR 2f¯I 7,0( 06(,2 ),* ' 7,0( 352),/( 2) )$67 1(87521 )/8; &0 )520 7+( 6285&( 89-

PAGE 259

)/8; 5(/$7,9( 81,76f ),* ' 7,0( 352),/( 2) )$67 1(87521 )/8; %* (0 )520 7+( 6285&(

PAGE 260

)/8; 5(/$7,9( 81,76f r ( r 6 r r 5,)6 3($. $7 06(&6 38/6( :,'7+ 06(&6 7+(*5< ):+0 %6 06(&6 (;3 ):+0 r 06(&6 WR X! 02 7,0( 06(&f 7,0( 352),/( 2) )$67 1(875'1 )/8; % &0 )5'0 7+( 6285&( ),* '

PAGE 261

)/8; 5(/$7,9( 81,76f ),* ' 7,0( 352),/( 2) )$67 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 262

)/8; 5(/$7,9( 81,76f ),* ' 7,0( 352),/( 2) )$67 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 263

)/8; 5(/$7,9( 81,76f ),*r ' 7,0( 352),/( 2) )$67 1(87521 )/8; &0 )520 7+( 6285&( LQ LQ

PAGE 264

)/8; 5(/$7,9( 81,76f ),*r ' 7,0( 352),/( 2) )$67 1(87521 )/8; r (0 )520 7+( 6285&(

PAGE 265

)/8; 5(/$7,9( 81,76f 7,0( 06(&f 7,0( 352),/( 2) )$67 1(87521 )/8; 6% &0 )520 7+( 6285&( ),*r ' -,/2

PAGE 266

)/8; 5(/$7,9( 81,76f ),* ',2 7,0( 352),/( 2) )$67 1(87521 )/8; (* &0 )520 7+( 6285&(

PAGE 267

)/8; 5(/$7,9( 81,76f ),*r 'OO 7,0( 352),/( 2) )$67 1(87521 )/8; ,%%r 2n0 )520 7+( 6285&(

PAGE 268

)/8; 5(/$7,9( 81,76f ),* ' 7,0( 35'),/( 2) )$67 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 269

$33(1',; ( 8)6$ 5 &/($1 &25( 7,0( 352),/(6 2) 7+(50$/ 1(87521 )/8; $7 )285 326,7,216 ,1 7+( &25( )25 $ 06(& ,1387 38/6(

PAGE 270

)/8; 5(/$7,9( 81,76f 7,0( 06(&f ),*r (O 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; r%6 &0 )520 7+( 6285&(

PAGE 271

)/8; 5(/$7,9( 81,76f ),*n ( 7,0( 35'),/( 2) 7+(50$/ 1(87521 )/8; &0 )520 7+( 6285&(

PAGE 272

)/8; 5(/$7,9( 81,76f 7,0( 352),/( 2) 7+(50$/ 1(87521 )/8; m,6 &0 )520 7+( 6285&( ),' (

PAGE 273

)/8; 5(/$7,9( 81,76 7,0( 06(&f ),* ( 7,0( 352),/(n 2) 7+(50$/ 1(87521 )/8; ( &0 )520 7+( 6285&(

PAGE 274

$33(1',; ) 8)6$ 5 &/($1 &25( 7,0( 352),/(6 2) 7+(50$/ 1(87521 )/8; $7 7+5(( 326,7,216 ,1 7+( &25( )25 $ :,'( 06(&f ,1387 38/6(

PAGE 275

)/8; 5(/$7,9( 81,76f ),*r ), 3523$*$7,21 2) $ 38/6( :,'(5 7+$1 7+( 6<67(0 3523$*$7,21 7,0( (;3(5,0(17$/

PAGE 276

$33(1',; * 8)6$ 5 &/($1 &25( 6+$3( 2) 7+( 3523$*$7,1* 38/6( $6 $ )81&7,21 2) ,1387 38/6( :,'7+ 3XOVH :LGWK PVHFf

PAGE 277

)/8; 5(/$7,9( 81,76f ),* * 6+$3( 2) 7+( 3523$*$7,1* 38/6( $6 $ )81&7,21 2) 38/6( :,'7+ (;3(5,0(17$/

PAGE 278

)/8; 5(/$7,9( 81,76f ),*r * 6+$3( ') 7+( 3523$*$7,1* 38/6( $6 $ )81&7,21 2) 38/6( :,'7+ (;3(5,0(17$/

PAGE 279

/,67 2) 5()(5(1&(6 6 2 -RKQVRQ DQG 5 : *DUQHU SULYDWH FRPPXQLFDWLRQ f 5 ( 8KULJ eW DO 'HVLJQ DQG 6XPPDU\ +D]DUGV 5HSRUW RI WKH 8QLYHUVLW\ RI )ORULGD $FFHOHUDWRU 3XOVHG 6XEFULWLFDO $VVHPEO\ f 0 2KDQLDQ 1 'LD] DQG 5 % 3HUH] 'HVLJQ DQG +D]DUGV 5HSRUW RI WKH 8QLYHUVLW\ RI )ORULGD 63(57 $VVHPEO\ f 0 2KDQLDQ DQG 1 'LD] 5HSO\ WR 5HTXHVW IRU $GGLWLRQDO ,QIRUPDWLRQ LQ &RQQHFWLRQ ZLWK 5HYLHZ RI $SSOLFDWLRQ IRU 610 /LFHQVH5HI 1R '0/5// f f 0 2KDQLDQ DQG 1 'LD] $GGHQGXP WR WKH 'HVLJQ DQG +D]DUGV 5HSRUW RI WKH 8QLYHUVLW\ RI )ORULGD 63(57 $VVHPEO\ FRPPXQLFDWLRQ WR 86$(& 0DWHULDOV /LFHQVLQJ 'LYLVLRQ f / % 0\HUV 8)6$ 6DIHW\ ,QVWUXPHQW DQG &RQWURO 6\VWHPV 5HSRUW ,QWHUQDO 5HSRUW 'HSW RI 1XFOHDU (QJLQHHULQJ 6FLHQFHV 8QLYHUVLW\ RI )ORULGD f 6 -RKQVRQ DQG 5 : *DUQHU SULYDWH FRPPXQLFDWLRQ f ( *DUHOLV DQG / 5XVVHOO -U 7KHRU\ RI 3XOVHG 1HXWURQ 6RXUFH 0HDVXUHPHQWV 1XF 6FL (QJ f 5 / &RDWHV DQG 1 5 +RUWRQ 56$& $ 5DGLRORJLFDO 6DIHW\ $QDO\VLV &RPSXWHU 3URJUDP ,'2 f 5 :DJQHU ,5(.,1* 3URJUDP IRU WKH 1XPHULFDO 6ROXWLRQ RI WKH 5HDFWRU .LQHWLFV (TXDWLRQV ,'2 f 6 5 %LHUPDQ . / *DUOLG DQG 5 &ODUN 5HVROYLQJ 7LPH RI D 3XOVHG1HXWURQ 6RXUFH 'DWD$FTXLVLWLRQ 6\VWHP 1XF $SSOLF f % ( 6LPPRQV DQG 6 .LQJ $ 3XOVHG 1HXWURQ 7HFKQLTXH IRU 5HDFWLYLW\ 'HWHUPLQDWLRQ 1¾F 6FL (QJ B f 1 * 6MRVWUDQG 0HDVXUHPHQWV RQ D 6XEFULWLFDO 5HDFWRU 8VLQJ D 3XOVHG 1HXWURQ 6RXUFH $UNLY )\VLN f 7 *R]DQL $ 0RGLILHG 3URFHGXUH IRU WKH (YDOXDWLRQ RI 3XOVHG 6RXUFH ([SHULPHQWV LQ 6XEFULWLFDO 5HDFWRUV 1XNOHRQLN f

PAGE 280

( *DUHOLV $ &DOFXODWLRQ RI WKH 3URPSW 'HFD\ &RQVWDQW IRU D 5HIOHFWHG $VVHPEO\ 7UDQV $PHU 1XF 6RF f 0 %HFNHU DQG . 4XLVHQEHUU\ 6SDWLDO 'HSHQGHQFH RI 3XOVHG 1HXWURQ 5HDFWLYLW\ 0HDVXUHPHQWV $(& 6\PSRVLXP 6HULHV 1R f . / *DUOLG DQG 6 5 %LHUPDQ $SSOLFDWLRQV RI 3XOVHG 1HXWURQ 0HDVXUHPHQWV LQ 9HU\ /DUJH 6\VWHPV 1XF $SSOLF B f 5 3HLHUO 3URFHHGLQJV RI WKH 5R\DO 6RFLHW\ /RQGRQf $ f 0 ( 5DGG SULYDWH FRPPXQLFDWLRQ f 3 . 'RVKL DQG * + 0LOH\ 5HDFWRU 1HXWURQ 3XOVH 3URSDJDWLRQ LQ 0XOWLSO\LQJ 0HGLD 7UDQV $PHU 1XF 6RF f 3 . 'RVKL 5HDFWRU 1HXWURQ 3XOVH 3URSDJDWLRQ 7KURXJK 0XOWLSO\LQJ 0HGLD 8QSXEOLVKHG 3K' 'LVVHUWDWLRQ 8QLYHUVLW\ RI ,OOLQRLV f + $ + +DVVRQ ([SHULPHQWDO DQG 7KHRUHWLFDO 6WXGLHV RI 6SDFH 7LPH 1XFOHDU 5HDFWRU .LQHWLFV 8QSXEOLVKHG 3K' 'LVVHUWDWLRQ 8QLYHUVLW\ RI ,OOLQRLV f + 'XQODS 1HXWURQ:DYH 3URSDJDWLRQ LQ D +HDY\ :DWHU 1DWXUDO 8UDQLXP 6XEFULWLFDO $VVHPEO\ 8QSXEOLVKHG 3K' 'LVVHUWDWLRQ 8QLYHUVLW\ RI )ORULGD f : 5 &DGZHOO $ ) +HQU\ DQG $ 9LJLORWWL :,*/(f§$ 3URJUDP IRU WKH 6ROXWLRQ RI WKH 7ZR*URXS 6SDFH7LPH 'LIIXVLRQ (TXDWLRQV LQ 6ODE *HRPHWU\ :$3'70 f 0 1 0RRUH 7KH 'LVSHUVLRQ RI 7KHUPDO1HXWURQ 3XOVHV LQ 1HXWURQLF 6\VWHPV 1XF 6FL (QJ f 5 6 %RRWK $ 7KHRUHWLFDO DQG ([SHULPHQWDO 6WXG\ RI 1HXWURQ :DYH 3URSDJDWLRQ LQ 0RGHUDWLQJ 0HGLD 8QSXEOLVKHG 3K' 'LVVHUWDWLRQ 8QLYHUVLW\ RI )ORULGD f 0 ( 5DGG:,*/( $ 7ZR*URXS 7LPH 'HSHQGHQW 'LIIXVLRQ 7KHRU\ 3URJUDP IRU WKH ,%0 &RPSXWHU ,'2 f $ ) +HQU\ DQG $ 9 9RWD :,*/ $ 3URJUDP IRU WKH 6ROXWLRQ RI WKH fµ 2QH'LPHQVLRQDO 7ZR*URXS 6SDFH7LPH 'LIIXVLRQ (TXDWLRQV $FFRXQWLQJ IRU 7HPSHUDWXUH ;HQRQ DQG &RQWURO )HHGEDFN :$3'70 f %
PAGE 281

0 $ 3HUNV )5($. $ )DVW 5HDFWRU 0XOWLJURXS .LQHWLFV 3URJUDP $1/ f 5 : *DUQHU SULYDWH FRPPXQLFDWLRQ f 6 .DSODQ 0DUORZHU DQG %HZLFN $SSOLFDWLRQ RI 6\QWKHVLV 7HFKQLTXHV WR 3UREOHPV ,QYROYLQJ 7LPH 'HSHQGDQFH 1XF 6FL (QJ f %
PAGE 282

+ 'XQODS DQG 5 % 3HUH] 'LVSHUVLRQ /DZ IRU D 6XEFULWLFDO $VVHPEO\ 3URFHHGLQJV RI WKH 6\PSRVLXP RQ 1HXWURQ 1RLVH :DYHV DQG 3XOVH 3URSDJDWLRQ $(& 6\PSRVLXP 6HULHV 1R f &25$ DQ LPSURYHG YHUVLRQ RI WKH ,03 FRPSXWHU SURJUDP ,03 LV GHVFULEHG LQ ,'2 6 -RKQVRQ 63(57 352-(&7 4XDUWHUO\ 3URJUHVV 5HSRUW 3KLOLSV 3HWUROHXP &R ,'2 f * 0RUWHQVHQ SULYDWH FRPPXQLFDWLRQ f $ ( :DOWDU DQG / 5XE\ ,QWHUSUHWDWLRQ RI 3XOVHG6RXUFH ([SHULn PHQWV LQ D 5HIOHFWHG 5HDFWRU ,QWHUQDO 5HSRUW 'HSDUWPHQW RI 1XFOHDU (QJLQHHULQJ 8QLYHUVLW\ RI &DOLIRUQLD f 3+52* LV D 3KLOOLSV+DQIRUG UHYLVLRQ RI *$0 * ' -RDQRX DQG 6 'XGHN *$0 *$ f ) :KHHOHU 5$9(1 $ &RPSXWHU 3DFNDJH IRU (YDOXDWLQJ 5HVROYHG DQG 8QUHVROYHG 5HVRQDQFH $EVRUSWLRQ ,QFOXGLQJ 3LQ 6KDGRZLQJ ,'2 f 727(0 OLQNV WKH 723,& DQG 7(03(67 FRGHV IRU FDOFXODWLQJ WKHUPDO FRQVWDQWV * ( 3XWQDP 723,& ,'2 f DQG 5 + 6KXGGH fµ DQG '\HU 7(03(67 ,, 7,' ff + 3 )ODWW DQG ' & %DLOHU $,0 $ 0XOWLJURXS 2QH'LPHQVLRQDO 'LIIXVLRQ (TXDWLRQ &RGH f 1 'LD] DQG 0 2KDQLDQ 81,38/ $ 8QLILHG 'DWD 3URFHVVLQJ 3URJUDP IRU 3XOVHG 0HDVXUHPHQWV ,QWHUQDO 5HSRUW 8QLYHUVLW\ RI )ORULGD f 0 2KDQLDQ 025( $ &RQYHQWLRQDO )RXULHU 7UDQVIRUP 3URJUDP IRU 3XOVHG 0HDVXUHPHQWV ,QWHUQDO 5HSRUW 8QLYHUVLW\ RI )ORULGD f 0 2KDQLDQ 025:,* $ &RQYHQWLRQDO )RXULHU 7UDQVIRUP 3URJUDP ZLWK 9DULDEOH 7LPH ,QFUHPHQWV ,QWHUQDO 5HSRUW 8QLYHUVLW\ RI )ORULGD f 1 'LD] $/;,/6 $ /LQHDU /HDVW6TXDUHV 3URJUDP IRU WKH )LWWLQJ RI $OSKD DQG ;L ,QWHUQDO 5HSRUW 8QLYHUVLW\ RI )ORULGD f

PAGE 283

%,2*5$3+,&$/ 6.(7&+ 1LOV 'LD] ZDV ERUQ LQ 0RURQ &XED RQ $SULO +H ZDV JUDGXDWHG LQ -XQH IURP /D6DOOH +LJK 6FKRRO LQ +DYDQD &XED ,Q -XO\ , KH REWDLQHG WKH GHJUHH RI 3URIHVVLRQDO 0HFKDQLFDO (QJLQHn HULQJ IURP WKH 8QLYHUVLW\ RI 9LOODQRYD +DYDQD &XED ZLWK KRQRUVr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

PAGE 284

7KLV GLVVHUWDWLRQ ZDV SUHSDUHG XQGHU WKH GLUHFWLRQ RI WKH FKDLUPDQ RI WKH FDQGLGDWHnV VXSHUYLVRU\ FRPPLWWHH DQG KDV EHHQ DSSURYHG E\ DOO PHPEHUV RI WKDW FRPPLWWHH ,W ZDV VXEPLWWHG WR WKH 'HDQ RI WKH &ROOHJH RI (QJLQHHULQJ DQG WR WKH *UDGXDWH &RXQFLO DQG ZDV DSSURYHG DV SDUWLDO IXOILOOPHQW RI WKH UHTXLUHPHQWV IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ 0DUFK 'HDQ *UDGXDWH 6FKRRO 6XSHUYLVRU\ &RPPLWWHH &KDLUPDQ


FLUX (RELATIVE: UNITS)
FIG* C7
TIME PROFILE DF THERMAL NEUTRGN FLUX
73*30 CM FROM THE SGURCE


259
15. E. Garelis, "A Calculation of the Prompt Decay Constant for a
Reflected Assembly," Trans. Amer. Nuc. Soc., 10, 2, 594 (1967).
16. M. Becker and K. Quisenberry, "Spatial Dependence of Pulsed
Neutron Reactivity Measurements," AEC Symposium Series No. 7
(1966)
17. K. L. Garlid and S. R. Bierman, "Applications of Pulsed Neutron
Measurements in Very Large Systems," Nuc. Applic., 2_, (1966).
18. R. Peierl, Proceedings of the Royal Society (London), A 149, 467
(1935).
19. M. E. Radd, private communication (1968).
20. P. K. Doshi and G. H. Miley, "Reactor Neutron Pulse Propagation in
Multiplying Media," Trans. Amer. Nuc. Soc., 11, 191 (1968).
21. P. K. Doshi, "Reactor Neutron Pulse Propagation Through Multiplying
Media," Unpublished Ph.D. Dissertation, University of Illinois
(1968).
22. H. A. H. Hasson, "Experimental and Theoretical Studies of Space-
Time Nuclear Reactor Kinetics," Unpublished Ph.D. Dissertation,
University of Illinois (1968).
23. J. H. Dunlap, "Neutron.Wave Propagation in a Heavy Water, Natural
Uranium, Subcritical Assembly," Unpublished Ph.D. Dissertation,
University of Florida (1967).
24. W. R. Cadwell, A. F. Henry and A. J. Vigilotti, "WIGLEA Program
for the Solution of the Two-Group Space-Time Diffusion Equations
in Slab Geometry," WAPD-TM-416 (1964).
25. M. N. Moore, "The Dispersion of Thermal-Neutron Pulses in Neutronic
Systems," Nuc. Sci. Eng., 25, 422 (1966).
26. R. S. Booth, "A Theoretical and Experimental Study of Neutron Wave
Propagation in Moderating Media," Unpublished Ph.D. Dissertation,
University of Florida (1965).
27. M. E. Radd,"WIGLE-40, A Two-Group, Time Dependent Diffusion Theory
Program for the IBM-7040 Computer," IDO-17125 (1965).
28. A. F. Henry and A. V. Vota, "WIGL 2, A Program for the Solution of the
One-Dimensional, Two-Group, Space-Time Diffusion Equations Accounting
for Temperature, Xenon, and Control Feedback," WAPD-TM-532 (1965).
29. J. B. Yasinsky, M. Natelson and L. A. Hageman, "TWIGL A Program
to Solve the Two-Dimensional, Two-Group, Space-Time Neutron Dif
fusion Equations with Temperature Feedback," WAPD-TM-743 (1968).


a
o
os
C\
FIG. 42 COMPARISON OF THE THEORETICALLY PREDICTED AND THE MEASURED DAMPING
COEFFICIENT a


168
The predicted and measured values are in excellent agreement up to 200
cps; theory follows closely the measured difference between the 0.5 and
the 1.0 msec cases. The differences in the damping coefficient at zero
frequency are somewhat puzzling. A steady-state measurement of the
inverse relaxation length yielded a value of 0.0221 cm \ in excellent
agreement with the experimental 0.5 msec case and ~ 5.% lower than the
1.0 msec case.
b) The theoretical and experimental tx's and £'s start to diverge
"significantly" past 200 cps. The deviation is larger for the 1.0 msec
input case. The deviation is ~ 6% at 400 cps for a and 5 (0.5 msec
case), and about 8% at 800 cps. Still both the theoretical and experi
mental results were smooth functions of frequency throughout the range;
both collapse dramatically at ~ 1000 cps.
c) An analysis of the least-squares fitting performed on both the
predicted and the measured values reveals that a "poor" fit is obtained
at 0 cps.. The fit gets increasingly better with frequency up to 400
cps were fluctuations are observed in the values. The criteria used to
select the value of a and £ quoted were that of spatial convergence, i.e.,
the chopping technique was employed to determine a spatially converged
value. This was not always encountered, especially past 500 cps in both
theory and experiment. A judgement was made to select an appropriate
value in each case where convergence was not obtained. This made the
theoretical and experimental results look worse at high frequencies
because they had different trends and the same criteria had to be used
for both.
The system p dispersion law is shown in Fig. 44. The agreement can
be termed good, considering the percentage differences and other previ
ously reported results [23]. A smooth trend is followed until theory


150
be too restrictive, since some spatial redistribution of the input
pulse does occur, although the effect is certainly much less pronounced
than in the case of a narrow pulse. In the case of the Illinois exper
iments the input pulse was ^30 msec wide and already had an asymptotic
shape, with the characteristic propagation time being of the order of
a few msec.
In the next section the propagation of pulses as a function of the
input pulse width will be studied qualitatively. These measurements
were made to complement the above results.
Pulse Shape vs. Input Pulse Width
The analysis of the measurements with input pulse widths of 0.1,
0.5 and 1.0 msec revealed that these propagating pulses achieved
identical shapes, within the experimental accuracy, at distances larger
than 130 cm from the source. This fact was believed to hold for input
pulses much narrower than the characteristic propagation time in the
assembly. A series of pulse shapes were measured as a function of
input pulse width at a distance of 142.9 cm from the source; these
measurements should provide information on the qualitative behavior of
propagating pulses arising from widely differing input pulse widths.
The results of these measurements are shown in Appendix G; the
peak times and the FWHM for each input pulse recorded at P83 are given
in Table XV.
As previously observed, the peaks are increasingly displaced in
time as the input pulse widens. This behavior has been well predicted
by the two-group, space-time dependent diffusion theory model. It
appears that for the 2 msec wide input pulse deviations from the


FLUX (RELATIVE UNITE)
FIG- CIO TIME PROFILE OF THERMAL NEUTRON FLUX 117-44 CM FROM THE SOURCE


121
Shown in Fig. 27 (A, B, C) are a set of the experimentally
determined thermal neutron time profiles for the 0.5 msec input pulse
case. A representative set of the experimentally determined spatial
distribution of neutrons in the assembly for increasing times after
the source pulse is shown in Fig. 28 (A, B). The space and time de
pendent redistribution of the thermal flux as the disturbance propa
gates through the assembly is clearly displayed.
Tables VII and VIII show the experimentally determined and theo
retically calculated delay times with reference to zero time (this
time corresponds to the time at which the neutron generator initiates
the burst and the multichannel analyzer starts its sweep), the FWHM
and the counts at the peak of the pulse normalized to position 83
which is located in the asymptotic region. In Figs. 29 through 32 the
delay times and the spatial attenuation for both pulse widths is
shown.
Very good agreement is obtained between the theoretical and
experimental results, throughout the length of the assembly, including
the region near the source and the region farthest removed from the
source where end effects are significant. This apparent agreement is
encouraging and vouches for the calculational scheme and source descrip
tion used.
The Propagation Time and the Asymptotic Velocity of Propagation
The propagation time, which is defined as the total delay time
between peaks at the two extreme positions of the assembly, is 3.65 +
0.1 msec.
The asymptotic velocity of propagation is defined as the inverse
of the slope of the curve of the delay time vs. distance in the


FLUX (RELATIVE UNITS)
1-0 +
0-3
0-B +
0-7
0-B
J3
0-4
0-3-
0-2
0-1-
NORMALIZED ID
THERM--5- 76
PULSE WIDTH = 0-5 MSECS
P104
Â¥ * 9 W
9 _* ft
? a .a*
9 ft ft -ft9
ft0-^a0
9e_
0,i0c-c5
>iS o '.-'P ' Jk
-o0.t.Vn^
e W vLT ','^->
wLvJ
0
-L
4
7
S
10
TIME (MSEC)
FIG. 27C EXPERIMENTAL PULSE SHAPES AT DIFFERENT POSITIONS IN THE UFSA R1 CORE
124


93
following technique was used in most of the measurements:
a) Measurements were taken with the bare movable detector
at each designated position and a normalization factor
obtained from the fixed detector.
b) Measurements were taken with the same movable detector
covered with a 0.018 in thick cadmium sleeve (the Cd cut
off energy for this thickness is -.53 ev) at the same posi
tions and a normalization factor obtained from the fixed
detector.
c) The time profile from each measurement was resolution time
corrected, background subtracted and normalized with
reference to the fixed detector (refer to Part 1, Chapter IV).
d) A point by point subtraction of the epicadmium time profile
from the total time profile yielded the thermal neutron flux
pulse shape at each spatial point.
It should be mentioned that, to be rigorous, a correction should
be applied to the epicadmium flux at each space point to take into
account the perturbation of the thermal flux introduced by the cadmium
sleeve; this was not attempted in this study. It should not be signif
icant in the highly absorbing medium being considered here.
As a corollary, it was expected that the epicadmium flux time
profile could be compared with the theoretically predicted fast flux.
The main problem encountered with this comparison lies in the fact that
the energy-dependent efficiency of the detector affects the character
istics of the pulse. The efficiency of detection for the thermal group
(0-.5 ev) is fairly constant and the comparison with the calculated
thermal flux is on solid grounds; the same argument is not valid


LIST OF REFERENCES
1. S. O. Johnson and R. W. Garner, private communication (1966).
2. R. E. Uhrig, £t al, "Design and Summary Hazards Report of the
University of Florida Accelerator Pulsed Subcritical Assembly"
(1964).
3. M. J. Ohanian, N. J. Diaz and R. B. Perez, "Design and Hazards
Report of the University of Florida SPERT Assembly" (1967).
4. M. J. Ohanian and N. J. Diaz, "Reply to Request for Additional
Information in Connection with Review of Application for SNM
License"(Ref. No. D.M.L.:RLL 70-1068) (1967).
5. M. J. Ohanian and N. J. Diaz, "Addendum to the Design and Hazards
Report of the University of Florida SPERT Assembly," communication
to USAEC Materials Licensing Division (1968).
6. L. B. Myers, "UFSA Safety Instrument and Control Systems Report,"
Internal Report, Dept, of Nuclear Engineering Sciences, University
of Florida (1968).
7. S. 0. Johnson and R. W. Garner, private communication (1967).
8. E. Garelis and J. L. Russell, Jr., "Theory of Pulsed Neutron
Source Measurements," Nuc. Sci. Eng., 16, 263 (1963).
9. R. L. Coates and N. R. Horton, "RSAC A Radiological Safety
Analysis Computer Program," IDO-17151 (1966).
10. R. J. Wagner, "IREKING Program for the Numerical Solution of the
Reactor Kinetics Equations," IDO-17114 (1966).
11. S. R. Bierman, K. L. Garlid and J. R. Clark, "Resolving Time of a
Pulsed-Neutron Source Data-Acquisition System," Nuc. Applic., 2
(1964).
12. B. E. Simmons and J. S. King, "A Pulsed Neutron Technique for
Reactivity Determination," Nc. Sci. Eng., _3, 595 (1958).
13. N. G. Sjostrand, "Measurements on a Subcritical Reactor Using a
Pulsed Neutron Source," Arkiv Fysik, 11, 233 (1956).
14. T. Gozani, "A Modified Procedure for the Evaluation of Pulsed Source
Experiments in Subcritical Reactors," Nukleonik, 4, 348 (1962).
258


LIST OF FIGURES (contd)
FIGURE Page
17A DECAY CONSTANT vs. MODERATOR LEVEL 76
17B kS/£ AND k vs. MODERATOR LEVEL 77
18 THE TOTAL, EPICADMIUM AND THERMAL FLUX 117.44 CM
FROM THE SOURCE 95
19 UFSA SOURCE-SUBCRITICAL ASSEMBLY GEOMETRICAL
ARRANGEMENT PLAN VIEW 96
20 PLAN AND FRONT VIEW OF THE CORE REGION ENCLOSING THE
NEUTRON SOURCE 102
21 ONE-DIMENSIONAL ARRANGEMENT OF THE UFSA CORE USED IN THE
WIGLE CALCULATIONAL SCHEME 103
22 SPATIAL DISTRIBUTION OF SOURCE NEUTRONS INCORPORATED
INTO THE WIGLE SCHEME 107
23A PULSE SHAPES PREDICTED BY WIGLE AT DIFFERENT
POSITIONS IN THE UFSA R1 CORE 110
23B PULSE SHAPES PREDICTED BY WIGLE AT DIFFERENT
POSITIONS IN THE UFSA R1 CORE Ill
23C PULSE SHAPES PREDICTED BY WIGLE AT DIFFERENT
POSITIONS IN THE UFSA R1 CORE 112
24A THE CALCULATED SPATIAL DISTRIBUTION OF THE THERMAL
FLUX AT DIFFERENT TIMES AFTER THE PULSE 113
24B THE CALCULATED SPATIAL DISTRIBUTION OF THE THERMAL
FLUX AT DIFFERENT TIMES AFTER THE PULSE 114
25 THE ASYMPTOTIC STEADY-STATE VERTICAL FLUX 116
26 THE ASYMPTOTIC STEADY-STATE HORIZONTAL FLUX 117
21k EXPERIMENTAL PULSE SHAPES AT DIFFERENT POSITIONS
IN THE UFSA R1 CORE 122
xiv


(sec
Moderator Level (cm)
FIG. 17A DECAY CONSTANT vs. MODERATOR LEVEL


TABLE II (Continued)
B. NON-NUCLEAR
Measured Parameter
a. Reactor water temper
ature
Method of Measurement
Fenwall temperature switch in
inlet line
b. Reactor water level
c. Reactor door and per
sonnel
Barksdale pressure switch mounted
on weir"box"with sensing line
connected to core
Limit switches on doors and push
buttons inside reactor room
d. Reactor core width
Limit switches on reflector tanks
e. Reactor water level Barnstead pressure switch senses
level in weir box
f. Reflector tank water Float switches in reflector tanks
level (low)
g. Flow control valve Limit switch on valve
shut
h. Reactor flow Differential pressure cell and
pneumatic control
i. Reactor water level D/P cell and pneumatic system
Application
Scram reactor on low reactor
inlet water temp. (60F)
Scram reactor if water level in
core exceeds top of weir height
Scram system and shut down neutron
gun if reactor doors are opened
or interior switches are acti
vated
Prohibit filling reflector tanks
when distance between tanks
exceeds widest reflected core
width
Stops pump when water reaches 11cm
below weir apex
Indicates water is filling reflec
tor tanks
Requires closing valve before
starting pump
Control flow rate. Indicate and
record flow rate
Indicate and record water level


FLUX (RELATIVE UNITS)
FIG. 35A THE SENSITIVITY OF THE ONE-DIMENSIONAL, TWO GROUP, SPACE-TIME KINETICS SCHEME
TO CHANGES IN THE TRANSVERSE BUCKLING


FLUX (RELATIVE UNITS)
FIG* Dll TIME PROFILE OF FAST NEUTRON FLUX IBB*31 O'M FROM THE SOURCE
246


preferred over the BF^ variety at the University of Florida because of
the larger neutron absorption cross section and operational reliability.
3
He undergoes the following reaction
He^ + n H^ + p + 0.764 Mev.
The cross section for this reaction is 5327 + 10 barn at v = 2200
o
10 3
m/sec compared to 3840 +.11 barn for B He follows a 1/v law in the
3
energy range from 0 to 200 kev. The pulse heights yielded by a He
filled counter are proportional to the energy of the neutrons plus 764
kev. The reaction has been used for neutron spectroscopy from the 100
kev to 2 Mev energy range. Gamma discrimination can be easily accom
plished by the use of a biased integral discriminator.
The detectors were long and thin and each took the place of a fuel
element in the core. The active length of the counter is slightly less
than the active length of the fuel. Two one atmosphere (predominantly
thermal detection) and one ten atmosphere (more responsive to higher
energy neutrons) detectors were used in the experimental program. A
sketch of the physical characteristics of the counters is shown in
Fig. 10. The thin, long cylindrical shape enhances the time character-
3
istics of the counter. Although normally a 1 atmosphere He detector
operates with a bias voltage of 1000 volt, the minimum input pulse
voltage requirement of the pulse transformer was such that the operating
voltage of the counters had to be raised to 1200 volt. At this
voltage the slope of the counts vs. high voltage curve was about 8% per
100 volt; therefore very stable, low ripple high voltage supplies were
used to insure reproducibility of the detector response.


186
comparison of the WIGLE results and the pulse propagation data were put
on the same basis by performing identical Fourier transformations and
fitting procedures on both. The predicted and measured real part of
2 2 2
p =(a -£ ) diverge past 100 cps. This is almost a "natural" occurrence
2
in the ultrasensitive p plane in which the opposing discrepancies of
2 2
a and £ are blown out of proportion. The agreement in the p disper
sion law of the system is good and within the experimental accuracy and
capabilities of the model.
It is believed that WIGLE did predict accurately the dynamic
behavior of the narrow, long core studied. It is also believed that the
sensitivity of the one-dimensional calculational scheme to small changes
in the transverse buckling should be born in mind when feedback effects
and/or two-dimensional effects are to be considered.
The followings topics deserve future investigation:
1. The sensitivity of the dynamic behavior of the system to small
changes in the transverse buckling. This can be accomplished by a
detailed pulse propagation study at two core heights or at two core
widths.
2. Impose a more severe test of the model by conducting measure
ments with a narrower input pulse.
3. Further interpretation of reactivity measurements in large
systems is needed. A correction for the influence that the input pulse
width has on the determination of kg/S, by the Garelis-Rus sell method is
necessary. The extrapolated area-ratio method should provide a basis
for comparison.
4. A comparison should be made between the calculational results
of two-group diffusion solved directly in the frequency domain and those
obtained by numerical transformation of the WIGLE results.


Relative Amplitude
FIG. 22 SPATIAL DISTRIBUTION OF SOURCE NEUTRONS INCORPORATED INTO THE WIGLE SCHEME
107


170
and experiment approach 1000 cps.
The differences observed reflect those noticed when the comparison
in the time domain was conducted. The fact that the WIGLE pulses are
consistently narrower than the experimental ones is reflected in the
smaller damping coefficients at large frequencies.
The differences showed up much more markedly in the ultrasensitive
2 2 2 2
p plane. Since a < a and E > 5 the real part of p =(a -£ )
^ th exp th exp r
seems to indicate a discrepancy that is not as large as it appears.
2
Shown in Fig. 45 is the real part of p for both theory and exper
iment as a function of frequency. Although they start to diverge sig
nificantly at f = 100 cps, it should be remarked that any small dif-
2
ferences in a and £ are magnified out of proportion in the p plane.
2
In contrast, the imaginary part of p =(2a0 appears to signify perfect
agreement in the results, since the deviations in a and £ cancel each
other (see Fig. 46).
2
The p dispersion law of the UFSA R1 clean core is shown in Fig.
47. This is a most sensitive index. Any small error in the data or
2
failure of the model will immediately appear "blown-up" in p Although
2
the theory and experiment diverge in the p plane, after 100 cps, both
behave smoothly; the differences should be considered within the con
text of the over all analysis. Certainly it can be said that the WIGLE
results do not follow the experimental results throughout the frequency
range but that is to expected; the actual discrepancies observed in the
2
time domain and the numerical uncertainties propagate into the p plane
so significantly that perfect agreement seems to be unreachable with
the present models and knowledge of parameters.
The steady-state inverse relaxation length =.023 (and the corresponding
ctf_o) is significantly smaller than the dynamic inverse relaxation =.030.


long, 39 inches high, and of variable width. In this study, the core
was 6.5 inches wide and 30 inches high. The effective multiplication
constant of the assembly was determined to be 0.990jK003. The assembly
is equipped with nuclear instrumentation capable of automatic scram
action.
For the kinetics studies, a fast data acquisition system was
developed to handle accurately the very high, time-changing count rate
encountered in the measurements. It essentially consists of a trans
former-coupled pulse amplifier to produce a fast logic signal at the
input of a multichannel analyzer from the input signal originating in
3
a long, thin He counter. The instrumentation adequately handled count
rates up to 3 x 10^ counts/sec at the peak of the pulses. A high degree
of reproducibility and fidelity in following the pulse profiles was
obtained with this instrumentation.
The space-time kinetics studies were performed by analyzing the
propagation of a fast neutron burst introduced at one end of the assem
bly, in the absence of inherent feedback effects. The experimental
results are compared with the results obtained from the two-group,
space-time dependent, one-dimensional diffusion theory scheme known as
the WIGLE program. A stringent test of the model is provided by a
combined analysis in the time and the frequency domain.
Tiie WIGLE calculational scheme accurately predicts the delay times
and the attenuation of the pulses when a first-flight spatial distribu
tion is assumed for the fast source. At large distances from the source
WIGLE underpredicts (~ 8% in the FWHW) the spreading of the pulse. A
marked sensitivity to small changes in the transverse buckling was
'found for the model, as well as the experiment.
xx


Dimensions in cm
PIG. 19 UFSA SOURCE-SUBCRITICAL ASSEMBLY GEOMETRICAL ARRANGEMENT
- PLAN VIEW -


CHAPTER I
INTRODUCTION
Statement of the Problem
The dynamic behavior of large reactor cores is recognized as one
of the areas of reactor analysis in which practical, reliable calcula-
tional methods are needed. Although a significant amount of theoretical
work has been done in this area, the experimentation has been restricted
to the fundamental work done by Miley and co-workers at the University
of Illinois [20, 21, 22] and the dispersion law studies in a heavy
water-moderated, natural uranium assembly performed at the University
of Florida [23]. The work at Illinois was hampered by the very broad
input pulse obtained at the thermal column of the TRIGA reactor, by
the neutronic size of the assembly and by the low value of the effective
multiplication (ke^ j^.92). The work at Florida, using comparatively
narrow pulses from a neutron generator (and thermalizing tank), was
also restricted by the neutronic size of the system and by the k
No clean determination of "complete" spatial effects in multiplicative
media has been reported.
The Reactor Safety Division of the United States Atomic Energy
Commission established the Large Core Dynamics Experimental Program,
as proposed by the Atomic Energy Division of the Phillips Petroleum
Company, to study the behavior of large reactor cores. The objective
80


(RELATIVE UNITS)
ID
-J
THERM--3- SO
PEAK AT 0-320 M,SEEDS
PULSE WIDTH = 0-3 MSECS
THECRY FWHVt = 0-773 MSEECS
EIXP FWHM = 0-773 MSECS
0-3
B 7
VO
VO
10
TIME (MSEi;)
C3
TIME PRCF'ILE DF THERMAL NEUTRON FLUX SB 43 EM FROM THE SOURCE


9
8
7
6
5
4
3
2
1
O
162
R1 Core
12.716 cm between points
48 62 76 90
Position Number
104 118 132
. 40 PHASE OF ZEROTH FOURIER MOMENT vs. DISTANCE FOR
SEVERAL FREQUENCIES
- 0.5 MSEC INPUT PULSE -


51
FIG. 11 MOVABLE DETECTOR DATA ACQUISITION SYSTEM


19
FIG. 6 WEIR "BOX"


Inverse Multiplication
Moderator Level (cm)
FIG. 16B INVERSE MULTIPLICATION vs. SQUARED INVERSE HEIGHT


5 (rad/cm)
P
Frequency
FIG. 43 COMPARISON OF THE THEORETICALLY PREDICTED AND THE MEASURED PHASE SHIFT PER
UNIT LENGTH £
167


11
The end walls are permanently covered with Cadmium on the outside
surface while the side walls have movable Cadmium covers to define the
boundaries for the bare and reflected cases. To optimize the number of '
neutrons inserted into the assembly by the neutron generator, the accel
erator target penetrates about 4 inches into the core. A water-tight
port is provided for this purpose. The port can be removed and a blind
flange inserted in its place. Several fuel rods must be taken out, the
number depending on how deep the target goes into the assembly and on the.
lattice pitch.
The core section of the assembly consists of an interchangeable
fuel rod spacing system made of 3/4 x 1/2 x 1/8 inch channels, 5/8 x 1/4
inch bars and aluminum shims mounted on the base plate of the tank. The
bars have milled slots to accommodate the .25 inch end tip of the fuel
rod and to set the pitch along the core width. The shims are placed
between the channel bar units to set the pitch along the core length
(96 inches) (see Fig. 2). The top fuel rod spacing system consists of
an aluminum grating. The mesh is determined by the lattice pitch under
study. The grating is made of aluminum bars and spacers, as shown in
Fig. 3. Thus, fuel rod removal along the length of the core is pos
sible to locate the detector for the experimental measurements.
The one-half inch long rod tip is fully surrounded by aluminum,
with practically no reflecting characteristic, but there is a 7/8 inch
length of rod between the end of the active fuel and the tip which is
t
surrounded by water. This bottom reflector is unavoidable and will be
considered in the calculations.


9
is interlocked with the neutron generator and the subcritical assembly
scram system, as is the door on the only other entrance to the shielded
room from the fuel storage area. Across the front face of the assembly,
a screened wire cage with a lockable door controls access to the core.
While not in use in the assembly the fuel is stored in a room
adjacent to the facility, built entirely for this purpose.
A more complete description of the. facility and its characteristics
can be found in the Design and Hazards of the UFSA and its addenda
[3, 4, 5],
The system has been licensed under Atomic Energy Commission SPECIAL
NUCLEAR MATERIAL LICENSE SNM 1050, March 1968. The license allows for
the possesion of 5400 fuel rods with a total U235 inventory of 190 kgs.
Fuel Characteristics
The UFSA is fueled with Spert F-l type fuel elements provided by
the Phillips Petroleum Company.
The fuel characteristics are:
Fuel Composition: UO^ in pellet form
U235 enr:*-c^linent: 4.81 + .15%
Active Fuel Length: 36" + .062"
Active Fuel Diameter: .42" + .0005"
Fuel Tube Material: stainless steel
Fuel Tube Length: 41.625"
Fuel Tube o. d.: .4655" + .0025"


CHAPTER III
DESCRIPTION OF THE MEASUREMENTS
Introduction
In most pulse propagation and/or neutron wave measurements the
experimental technique employed to obtain the thermal neutron pulse
shape is the so-called cadmium difference method. In this technique
the contribution of the epicadmium source neutrons is subtracted by
making measurements with and without a cadmium shutter between the
source and the system. No attempt is usually made to subtract the
epicadmium neutron contribution due to fissions if the medium being
studied is multiplicative.
But, most neutron detectors interact with epicadmium neutrons with
varying degrees of efficiency; therefore to isolate the thermal neutron
pulse shapes a different experimental technique was used in this work.
Thus, it was possible to make a one-to-one comparison of the experi
mentally obtained thermal neutron pulse shapes with the corresponding
results from the analytical scheme.
The subcritical configuration and electronic instrumentation used
has been detailed in Part 1, Chapter IV. Unless specified, the meas
urements were made at a moderator height of 75 cm, corresponding to a
keff = 0.99+.003.
The system was operated under continuous water flow conditions.
No heating of the water was observed by either the energy transferred
91


Abstract of Dissertation Presented to the Graduate Council
in Partial Fulfillment of the Requirements for
the Degree of Doctor of Philosophy
SPACE-TIME REACTOR KINETICS STUDIES WITH
THE UNIVERSITY OF FLORIDA SPERT ASSEMBLY
By
Nils J. Diaz
March 1969
Chairman: Dr. M. J. Ohanian
Major Department: Nuclear Engineering Sciences
A large-in-one-space dimension, side reflected, highly multiplica
tive (kgff ~ 0.99) subcritical assembly was designed and calibrated.
The sole purpose of the facility is the experimental investigation of
the dynamic behavior of large reactor cores and to provide a test for
space-time kinetics models presently in use. With this facility the
linear aspects of neutron physics phenomena can be investigated in the
absence of inherent feedback effects. This work was conducted under a
subcontract with the Nuclear Safety Research Branch, Atomic Energy
Division, Phillips Petroleum Company, under a prime contract with the
United States Atomic Energy Commission.
The University of Florida SPERT Assembly (UFSA) is a light-water
moderated subcritical facility fueled by 4.81% enriched UC^ pellets
encased in stainless steel tubes of 0.465 inch outside diameter (SPERT
F-l Fuel). The fuel arrays are contained in a rectangular tank, 8 feet
xix


20
perimeter of the weir "box" into the drain line effectively preventing
any further increase of the moderator level in the assembly.
The measurement of the water height in the core is accomplished by
fixing a reference mark on the slide block at the same level as the
bottom of the weir within the "box". An accurate scale is provided to
read off the distance between the bottom of the core and the apex of the
weirs. Continuous indication of the moderator level in the core is
provided on the console by means of a recorder calibrated between the
bottom of the active fuel and the maximum moderator height and by a
manometer, connected directly to the core, for precise measurement of
the water level. These two measuring systems insure reproducibility of
the moderator height for the experiments.
The water is pumped out of the storage tank by a constant speed
centrifugal pump which has a "no load" capacity of 20 gal/min. The
control valve is designed to restrict the flow to the maximum design
value of 12 gal/min. A deionization system is provided to keep the
water as pure as possible at all times. The pneumatic flow control
system consists of two differential pressure cells, transmitters, control
valve, and recorder-controller. The strip type chart recorder-con
troller records both flow rate and moderator level in the core. The
control valve is of the air-to-open type which will close in the case of
air supply loss, stopping the flow into the assembly. The flow diagram
for the air system is shown in Fig. 7. A pressure differential from the
pressure transmitters applied to the recorder-controller allows both
manual and automatic control' of the flow rate through the valve operated
by the controller.


137
calculations were performed in which the height of the assembly was
varied from 70 to 85 cm. The respective core heights with their
corresponding vertical bucklings, k ^ (calculated using AIM-6) and
"absorption" cross-sections used in the WIGLE code are listed in Table
XI.
The results of the calculations showed a marked sensitivity to
changes in the transverse buckling. The delay times and the peak
counts vs. distance to the source are shown in Fig. 33 and 34 respec
tively. The experimental results for 76 cm are also shown. The data
for the attenuation curves was normalized to the attenuation at P83
of the 76 cm WIGLE calculations. Pulse shapes for the different heights
are shown together with the experimental results at 76 cm in Fig. 35
(A, B, C) for three representative positions in the assembly.
Some additional experimental measurements were also made at a
given position for different core heights to supplement the calcula-
tional predictions. It should be pointed out that these measurements
are of a preliminary nature, made over a short period of time simply
to obtain confirmation of the theoretical trends. The results of these
measurements are shown in Fig. 36.
It is not surprising that the results for the core showed as much
sensitivity to changes in the transverse buckling since the width and
height of the core are rather small in comparison to the length of the
assembly. It is to be noted that the changes in the "absorption"
cross-section due to changes in the transverse buckling are rather
small and are reflected in the third significant figure in the fast
group cross-section and the fourth significant figure in the thermal
group cross-section. This emphasizes the importance of an accurate


FLUX (RELATIVE UNITED
f
FIG- Cll time PROFILE OF THERMAL NEUTRON FLUX 130-16 CM FROM THE SDURCE
207


NO
00
FIG. 8
UFSA SAFETY SYSTEM LOGIC FLOW DIAGRAM


under a prime contract with the United States Atomic Energy Commission.
The technical aid of the University of Florida Computing Center in the
development of the computer programs and their financial assistance is
gratefully acknowledged.
Special mention is due Messrs. S. 0. Johnson, R. W. Garner,
G. A. Mortensen and Mrs. M. E. Radd of the Nuclear Safety Research
Branch, Atomic Energy Division, Phillips Petroleum Company for their
continuous assistance and invaluable suggestions throughout the
research program.
To my sister, Miss Lydia Gonzalez, my sincere appreciation for.
typing an elegant manuscript from my unintelligible characters.
To the fellow students, who struggled with me to reach the un
reachable star, my space and time independent friendship.
iv


33
criterion.
d) Operating limits have been set to delineate the normal oper
ating ranges of the assembly.
e) Initial loading procedures have been established to determine
the safe operating multiplication factor of each configuration.
A series of administrative controls are necessarily applied to all
segments of the experimental program and strictly enforced.
Initial Loading
A series of calculations were done to determine which of the two
following schemes should be employed for the initial loading of UFSA:
I) Step loading of the fuel from the center out, accompanied by
step increases in water level with the usual inverse multiplication
determination.
II) Loading all the fuel into the dry tank and proceeding with a
careful evaluation of the multiplication as a function of water level.
Since the UFSA core is very loosely coupled as far as the lumped
reactivity parameter is concerned, the second method was selected due to
the fact that a better determination of the multiplication was possible
from a basic moderator height-zero loading inverse counts determination.
The slope of the kg^^ vs. water height curve has a slope substantially
smoother than the kg^^ vs. per cent fuel loading (full water height)
curve.
The moderator level control system in operation at the facility
provides a extremely reliable and safe mode of adjusting the water level
without safety compromises.
The following regulations were followed for the initial fuel