<%BANNER%>

Development of an Automated Testing System for Verification and Validation of Nuclear Data

Permanent Link: http://ufdc.ufl.edu/UFE0022098/00001

Material Information

Title: Development of an Automated Testing System for Verification and Validation of Nuclear Data
Physical Description: 1 online resource (46 p.)
Language: english
Creator: Triplett, Brian
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: automation, database, mcnp, nuclear, validation, verification
Nuclear and Radiological Engineering -- Dissertations, Academic -- UF
Genre: Nuclear Engineering Sciences thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Verification and validation of nuclear data is critical to the accuracy of both stochastic and deterministic particle transport codes. In order to effectively test a set of nuclear data, the data must be applied to a wide variety of transport problems. Performing this task in a timely, efficient manner is tedious. The nuclear data team at Los Alamos National Laboratory in collaboration with the University of Florida is developing a methodology to automate the process of nuclear data verification and validation (V & V). This automated V & V process tests a number of data libraries using well defined benchmark experiments, such as those in the International Criticality Safety Benchmark Experiment Project (ICSBEP). The automation also serves to reduce errors and increase efficiency. The process is implemented through an integrated set of Python scripts. Material and geometry data are read from an existing medium or given directly by the user to generate a benchmark experiment template file. The user specifies the choice of benchmark templates, codes, and libraries to form a V & V project. The Python scripts generate input decks for multiple transport codes from the templates, run and monitor individual jobs, and parse the relevant output automatically. The output can then be used to generate reports directly or can be stored into a database for later analysis. This methodology eases the burden on the user by reducing the amount of time and effort required for obtaining and compiling calculation results. The resource savings by using this automated methodology could potentially be an enabling technology for more sophisticated data studies, such as nuclear data uncertainty quantification. Once deployed, this tool will allow the nuclear data community to more thoroughly test data libraries leading to higher fidelity data in the future.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Brian Triplett.
Thesis: Thesis (M.S.)--University of Florida, 2008.
Local: Adviser: Anghaie, Samim.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022098:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022098/00001

Material Information

Title: Development of an Automated Testing System for Verification and Validation of Nuclear Data
Physical Description: 1 online resource (46 p.)
Language: english
Creator: Triplett, Brian
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: automation, database, mcnp, nuclear, validation, verification
Nuclear and Radiological Engineering -- Dissertations, Academic -- UF
Genre: Nuclear Engineering Sciences thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Verification and validation of nuclear data is critical to the accuracy of both stochastic and deterministic particle transport codes. In order to effectively test a set of nuclear data, the data must be applied to a wide variety of transport problems. Performing this task in a timely, efficient manner is tedious. The nuclear data team at Los Alamos National Laboratory in collaboration with the University of Florida is developing a methodology to automate the process of nuclear data verification and validation (V & V). This automated V & V process tests a number of data libraries using well defined benchmark experiments, such as those in the International Criticality Safety Benchmark Experiment Project (ICSBEP). The automation also serves to reduce errors and increase efficiency. The process is implemented through an integrated set of Python scripts. Material and geometry data are read from an existing medium or given directly by the user to generate a benchmark experiment template file. The user specifies the choice of benchmark templates, codes, and libraries to form a V & V project. The Python scripts generate input decks for multiple transport codes from the templates, run and monitor individual jobs, and parse the relevant output automatically. The output can then be used to generate reports directly or can be stored into a database for later analysis. This methodology eases the burden on the user by reducing the amount of time and effort required for obtaining and compiling calculation results. The resource savings by using this automated methodology could potentially be an enabling technology for more sophisticated data studies, such as nuclear data uncertainty quantification. Once deployed, this tool will allow the nuclear data community to more thoroughly test data libraries leading to higher fidelity data in the future.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Brian Triplett.
Thesis: Thesis (M.S.)--University of Florida, 2008.
Local: Adviser: Anghaie, Samim.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022098:00001


This item has the following downloads:


Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20101109_AAAAAT INGEST_TIME 2010-11-09T13:17:29Z PACKAGE UFE0022098_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 1053954 DFID F20101109_AAAPWQ ORIGIN DEPOSITOR PATH triplett_b_Page_14.tif GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
bc32e6e53acfce3058d811a79f326319
SHA-1
027dc4f60af388b3027cca9057ea15cdbad5c892
112108 F20101109_AAAPRS triplett_b_Page_19.jp2
b5382a1423c3f41fcadabc3201341a30
e39ee199ada8adca657d61fd84fbd9c87aa82d8f
F20101109_AAAPWR triplett_b_Page_16.tif
cf5403dc12a3abcfc66d1764ce0a8819
e72a229a1ce596a6cbd5996519ffbe170224f4d5
73731 F20101109_AAAPRT triplett_b_Page_17.jpg
94b90cbabaada160a09aee9d0183c020
7d7bedc6e7568b5782acf7ff0cc9d9f07512704f
F20101109_AAAPWS triplett_b_Page_17.tif
2c72856426e1feb7e0256d6abd035aff
21b23852a720295e50c29765a213420b9f10950b
658829 F20101109_AAAPRU triplett_b_Page_28.jp2
b83b0ae2ec87a33386dbc5e0bd80c350
d1f77aa84397ffaa4a9cdeba2f5197d27f0d00a0
F20101109_AAAPWT triplett_b_Page_18.tif
d3da852cef8e48c743b02f71a388568f
7f2986ee748d4393511d97ac87e4243b07e6e086
370 F20101109_AAAPRV triplett_b_Page_09.txt
c13213143bbde70deb3bbbf57e10220d
1167e19af1aea9036a4cd181dce2093cd4febeb1
45707 F20101109_AAAPRW triplett_b_Page_25.pro
a6d74de5425ee95312eb791e9055a7a7
607646f6f1833d011ca6fa3852f19ecd42a7ebae
80686 F20101109_AAAPRX triplett_b_Page_03.jp2
3af2972ff7b5c41585e7f192a36112d9
e02e85aab6f7463f17149d47254d23bd5c94bb3a
F20101109_AAAPWU triplett_b_Page_20.tif
feaa6a38ed3968f800a4fda34c1804b0
26f322513eb5b3dbee7183fb6fa13138051cae64
47770 F20101109_AAAPRY triplett_b_Page_40.pro
cf53e507d2b217f751751b4632d7e2aa
c76941e77d6433ca6fd8515bbca5b6b7ffbba891
F20101109_AAAPWV triplett_b_Page_21.tif
198d79f33d43746a41451d27e0e9f5c6
c9ca21157b3f8604903608be017cef776fbf0525
3538 F20101109_AAAPRZ triplett_b_Page_26thm.jpg
cf9e16b4d72287432b83a1a3955eb48c
dc0a0006e4cda6698cd6a87131233a775b1a5ebb
F20101109_AAAPWW triplett_b_Page_22.tif
9b55faa70caa7562d6cbb8bb055de18c
f1aca3f3cbcff7e30dcabf76dba782cf70a1a6e2
F20101109_AAAPWX triplett_b_Page_23.tif
4954bdfea8b2ce386c04afd3bd258c58
62feb79f6b29fedd5a02591f86bc5b55f83c8bd6
68491 F20101109_AAAPUA triplett_b_Page_13.jpg
25a91951b4e680aa07ea25be15f1fb9e
aaafdf143bf196437b9fa6f14ebfefd9ca6daa39
25271604 F20101109_AAAPWY triplett_b_Page_26.tif
d6c2dbddd263e4d07698cd94ea4216c8
896a78ceff7b2446499593a63002175b57bc39ff
79220 F20101109_AAAPUB triplett_b_Page_15.jpg
cebe157620b531d1d008e1258f96ed1f
349e3d25453679e139e5563af8ab6039697440b2
F20101109_AAAPWZ triplett_b_Page_27.tif
db4125c523cf3302d646b478ef5ed570
f05536ed49d7cc5e716d7b1a9dc7d3b42b6aff5d
74375 F20101109_AAAPUC triplett_b_Page_16.jpg
5d7d8b5389e12242ec96e1487118f67c
7a393556a1de4cfe6ea773093497aeffb7dec772
2094 F20101109_AAAPZA triplett_b_Page_10.txt
b3fdb9af85739c5344c738422d2beaf7
575fe0fab23569d319ebc6ac327fcd6ce7b2d67b
6689 F20101109_AAAQBA triplett_b_Page_11thm.jpg
96c1b602434bb6e406ef5bdaf958f72e
8e5ce4af88f37f86a1777fbbd4d24091fc168695
74782 F20101109_AAAPUD triplett_b_Page_19.jpg
015d1c33de2efe8465f6a330144a429d
aa79810e3972d43d2a82dff770913f4ef7536bda
2119 F20101109_AAAPZB triplett_b_Page_12.txt
f44e6507c5d90e02b41a60442efaa65e
e460099f4d9159d1b2399a0baade482e0109a0a6
25059 F20101109_AAAQBB triplett_b_Page_12.QC.jpg
6152eed958d1b83423421cfa7503a399
ead5c9ae0f70ec9b030982dc2ce8586c9b0f892c
74313 F20101109_AAAPUE triplett_b_Page_20.jpg
ad578bedee1e8b7334b09182e0954330
fdfe580f9cb0e1fa43e253ca7e4d0435ccabb276
1981 F20101109_AAAPZC triplett_b_Page_14.txt
3ca71f35d9bad8838bbd6f0d0d224fe6
e36be1effa32d77318c4cddf2de2df679d532485
6318 F20101109_AAAQBC triplett_b_Page_13thm.jpg
41f67e55250734b10b1eff35e5c583e6
892323efb69c66bf442de2e09756f53d7b7bde9a
71807 F20101109_AAAPUF triplett_b_Page_21.jpg
2abfcbb52b5e5dba72d9c9665bcba7d9
136f7102046fd90b2bee4466ba4a38e54857eaf0
2193 F20101109_AAAPZD triplett_b_Page_15.txt
3a1057ee6d352fd94e2628be7d708334
a8cbc380810e2534d95c1c6bca626a40be585407
23475 F20101109_AAAQBD triplett_b_Page_14.QC.jpg
8c0248fa817584e5e4c1f8076d6c66a1
8eb3af267237200b117812ed54061cc5b060bf66
72497 F20101109_AAAPUG triplett_b_Page_23.jpg
aadeb9353ff63a9e2870e145badde603
6c6d908c4a043eb7587ddacca8e673d4f30ba9f9
2039 F20101109_AAAPZE triplett_b_Page_16.txt
95e91e0326fc5230b8ba6deadcabc758
4dc894e7188f084fd6f779f114aca3da14be6b95
6328 F20101109_AAAQBE triplett_b_Page_14thm.jpg
5bfec586e5a198e77c7aaf63e3c8feee
9f46d8cbb01f94c73f189ce075c7a1ea912b4cd9
66591 F20101109_AAAPUH triplett_b_Page_25.jpg
74947ea235ea1a8358db08dda4ca2b84
2ee48be76ac22b9e7a081387812ae97565084d32
2068 F20101109_AAAPZF triplett_b_Page_18.txt
770d255356dc714bc88d9e1f45a94709
07b8ea3eb9a7497fa7956ebf42e9fb376727ba48
26080 F20101109_AAAQBF triplett_b_Page_15.QC.jpg
0d3434a23f951d097e43f9b3102e1a6b
3a5f1651a69977b1eaed698509a90be87e0ae2f3
46807 F20101109_AAAPUI triplett_b_Page_26.jpg
90ee5af4bf2ed27b8ce46bfee3c106e7
a0d98ad485581930633bdaf1e2ab52229562e336
2105 F20101109_AAAPZG triplett_b_Page_19.txt
0fa41bb18a5d7808b293d93c63f9245e
bf68149b6e8dc36d6e9a8946ccb5b37c80042554
6860 F20101109_AAAQBG triplett_b_Page_16thm.jpg
b441b14e4c9471e623c5730f7f19cdd8
268c02d4406e3350c0fa8cdc74941c7bb52c8826
46962 F20101109_AAAPUJ triplett_b_Page_27.jpg
41749cd4ee354a0cc46e54b09c60fbef
de2dac6982feb487e230a129c205da261a31aa6c
2098 F20101109_AAAPZH triplett_b_Page_20.txt
c4cb6e86e4eb0bbb83f4a7ee4d1f54dc
23aad96ac646486cd6ce1507e95a8fd046703a2a
24522 F20101109_AAAQBH triplett_b_Page_17.QC.jpg
595f14f524904497292e3cb99c60cff7
92d6173ca03941b32f009d9866bc335d7eea9dc2
42799 F20101109_AAAPUK triplett_b_Page_28.jpg
762639fb38529dfdcb531689620c50e5
6c8c0c9c0419fefba5fd35899a06bb9e092653ad
1975 F20101109_AAAPZI triplett_b_Page_21.txt
2392e9e787851b422442a2cc12a52ea6
f81fe8bdcd9c071870fad419630618d565242f76
24302 F20101109_AAAQBI triplett_b_Page_18.QC.jpg
2b5d40ffa89b89dc963af69d354b0258
6076cd8bd96579540b9384dbe1408a0388eb5503
42111 F20101109_AAAPUL triplett_b_Page_29.jpg
7f08495ca5a49e942842d543877230b0
0d6bba5623e800d4f06b30285c17ffd772d4481e
2021 F20101109_AAAPZJ triplett_b_Page_23.txt
b0c27c0c1b5200ebe32f416a695fc749
f43c67577f9cb7d4cccd4b11afe046236ebcb732
6802 F20101109_AAAQBJ triplett_b_Page_18thm.jpg
6b09deeccad187032382c2ecef466d0c
50ade7c111b182cafdbca2123814beb3f229e9fa
69166 F20101109_AAAPUM triplett_b_Page_31.jpg
b0435e991f98f6e3a95462ad8f6eda13
392e8bbb30804ab3f499e36a22f79d58f4a298aa
2078 F20101109_AAAPZK triplett_b_Page_24.txt
88e5101d1fc389ef9367b2e298be8cb0
74ca57aba033c312202e9fd4457d174f967e1b94
24932 F20101109_AAAQBK triplett_b_Page_19.QC.jpg
e790af698c100c61a49d644abc6e16b0
2b42a49b148c9cdb26f31903245e7d21036b0f71
70277 F20101109_AAAPUN triplett_b_Page_32.jpg
2b048f2bf40b606b5b38e5fba4aa93e9
15331be126273ced89ba50ef972fc71a2cee6356
1828 F20101109_AAAPZL triplett_b_Page_25.txt
d886ca6a6de385bf76105f5b453c04c7
cf121fbcb898c4eef1f9072d5bf5ebc1189a044c
24914 F20101109_AAAQBL triplett_b_Page_20.QC.jpg
7a9e64fb53f4552e9662d23ea5b4cabd
268abe600466a6e02fbe492878a0e2d2738d694c
71273 F20101109_AAAPUO triplett_b_Page_33.jpg
6df52642efcfcab31c721523f39b1043
1cd854b687360a9db9301e0a85bd8a4c4d903be2
534 F20101109_AAAPZM triplett_b_Page_26.txt
df35089582419bd95db3b3e39ff6a6ca
f61ead934bf976766f17cd049b514e9a5c6f7c23
6977 F20101109_AAAQBM triplett_b_Page_20thm.jpg
8df8ff6db048fc5508a1818a1f753233
4d9a8ea1906f8231e80a05aa884e8eda27066b0a
76484 F20101109_AAAPUP triplett_b_Page_34.jpg
f95280f507c85930212700089bc6b154
8486f09500e9d18f0086811699d8e21c63eccccf
1058 F20101109_AAAPZN triplett_b_Page_27.txt
16781f5510955533287e0b171baeb7a1
4977adb4bcc562fb760d7d1c96933cf0f4f645d4
23572 F20101109_AAAQBN triplett_b_Page_21.QC.jpg
a484e38d8e8bad99149b37566c478201
0315bf327a9edb19a7fc4d32c12189eb08d544d7
46342 F20101109_AAAPUQ triplett_b_Page_36.jpg
321157826ae00a740f509630ce513431
c303418acb5abaadf13c973fdd1139c15d35dfbc
574 F20101109_AAAPZO triplett_b_Page_29.txt
f9c5705f1e83d29bd0149a3e584d7a85
0dffaef53802e1ca8206050336f46076b892e72b
25487 F20101109_AAAQBO triplett_b_Page_22.QC.jpg
457c7b08cfd177a65b6253c721de7319
52acb3295f70ad033937d3184174a35ae33691b7
26737 F20101109_AAAPUR triplett_b_Page_37.jpg
27895aa96ffe28362f5f4c278abbe40f
881a006e98fb5591c1028139c3cb9aa09886aed3
1745 F20101109_AAAPZP triplett_b_Page_30.txt
bad440258dce7526b99cace42da42dac
c6d64774dd47ff6f9feedde5ff4340bc55ecce60
24549 F20101109_AAAQBP triplett_b_Page_23.QC.jpg
353eafed54771bf3a53534f1deb5e577
8f75d86dc4189400ce7804465bc4b0c17d792e3c
6622 F20101109_AAAQBQ triplett_b_Page_23thm.jpg
5531a62de854c75975fcb998ff5dcf87
63e3ec3a511ae92876d96d6132c012bd8dfde416
56972 F20101109_AAAPUS triplett_b_Page_38.jpg
122c9cade5abd562c005c24cf0d41e17
0be7f7b7ce10db877fc5004d2333a512033aa25a
3566 F20101109_AAAPZQ triplett_b_Page_31.txt
36d0d3e13a22500314a1976e06f66cf3
5b2c1f5e1f1abfacec13642b4505ce8b9d84f6d3
25828 F20101109_AAAQBR triplett_b_Page_24.QC.jpg
123c22ea424a9e800f31678c0048cffc
7c0c4d385bd3210b6b2b2fb4693f60ef94384123
34991 F20101109_AAAPUT triplett_b_Page_39.jpg
15451ad0118a9dc44a0b159e850b6dc4
ad2d12489d5c38490d442d0e8c46ec1aa46516eb
1902 F20101109_AAAPZR triplett_b_Page_32.txt
7ad289e5d51b6db8f87de6f8988d3ff2
be3ec34ed9a1135be7df415e4ae54e28ada7cc6b
22281 F20101109_AAAQBS triplett_b_Page_25.QC.jpg
a3b5d0c1030e5fd3c42fcf751ad626c6
d6638bd77f3680f0aa30a87c6d433259f146fb55
69444 F20101109_AAAPUU triplett_b_Page_40.jpg
e1461afcc15e07a5589a52497e8ab3e6
a7fcf6f9747d1010947b525f4b516386c1ea2e93
2168 F20101109_AAAPZS triplett_b_Page_34.txt
c255e7a2848f7b861a8dd6f77f93fbc2
7e72adc7705b9364a362da712fdb953470efb808
6253 F20101109_AAAQBT triplett_b_Page_25thm.jpg
2325a36b597baf81c99e10a77cf20d11
76a499fc2bad2a003de4ae7db4ea5fe2ecee2636
74177 F20101109_AAAPUV triplett_b_Page_41.jpg
55c2d3dc19793da468d357eadc1a4cd6
6a384879d27a8daa66ad999b1c720c399f9916d2
1280 F20101109_AAAPZT triplett_b_Page_35.txt
eaa28750592628d0344f2c3897e1f216
edf561415b1e18265202a26bc2dfd99cc86ebe3d
14859 F20101109_AAAQBU triplett_b_Page_27.QC.jpg
af22cba5c49ed7fd8f1a6594b70d0773
4128818f02d60026a84b904aa3ea53876c7e4bd6
66043 F20101109_AAAPUW triplett_b_Page_42.jpg
6c5c77eabf7cb11120ce6c608150dda6
6af72fb71cb61cd499aadc606bf8bb4481463032
862 F20101109_AAAPZU triplett_b_Page_37.txt
02fbae2f6a7ae9bf1e2fc91024bdf390
5456eae3e48e02b9beeea3ab22fe58d9206966ff
13275 F20101109_AAAQBV triplett_b_Page_28.QC.jpg
5fcb66f0a484a828151da98a083ec7e7
357107a85f2f6f3c8c9878fc93a3804aa2cee970
35379 F20101109_AAAPUX triplett_b_Page_43.jpg
d291820262c865de31473505eb1f70cc
68a07cd4d26cdb5d0f10e66c83953cb9dbb76102
1331 F20101109_AAAPZV triplett_b_Page_38.txt
45ee6b9f75d8088ef47d1850b56360a1
5f0dc4823e1f86f812806ecc690ef0c7160dafbc
4095 F20101109_AAAQBW triplett_b_Page_29thm.jpg
af35d1fac7954dacd8fa50a089242484
9cc162f31ac68c965904fedeb9589e1a0c6a3575
F20101109_AAAPSA triplett_b_Page_45.tif
3bb5bd75136905f5d38ddbe5b7b2f1d9
6c4529d2f828fa83da318e14b3e9db01c5731c90
85233 F20101109_AAAPUY triplett_b_Page_44.jpg
dc0f716bf8b13e8f2e4db9c656efc0bd
ba4c11f2b7e04af876bd0cc0ca61bbdd89b27e9d
1273 F20101109_AAAPZW triplett_b_Page_39.txt
c7f2639a309996fe2cd3d74baf631f64
a0e1eb98a435dba7651d0cf9481207c98fed6ac2
F20101109_AAAPSB triplett_b_Page_19.tif
06b292abaaad83ee863801b513eb7fd5
1c24884bd73311073915b8187593cd82ddd32d89
49877 F20101109_AAAPUZ triplett_b_Page_45.jpg
e775facedbeb2a5bfefacbe90d3c27e5
f0c65a5d7af128068f997cb029d99b50376c8626
15869 F20101109_AAAQBX triplett_b_Page_30.QC.jpg
c088eb742bb923f785c107a28b1696c5
d951d8e9783eb4d40f086791d956c89b836e2d46
111759 F20101109_AAAPSC triplett_b_Page_41.jp2
a630e0156df5fec6bb74f807ed5e3981
2e3047a5f32e32fe8cd24453baeb2c7b355f424c
2050 F20101109_AAAPZX triplett_b_Page_41.txt
b966267ba97c2611f40ae05df9283694
3c8b2d8078934be0e88ee1a3f8c67b5cb6936013
4830 F20101109_AAAQBY triplett_b_Page_30thm.jpg
1f72a2698634826dd16de6f7f0e758c2
1d228133470c38b11b57a48628d323b9fa69d7b1
F20101109_AAAPXA triplett_b_Page_28.tif
58ab18ab8063544a8223a90f15313508
c7f5812b6910dc43308d85bdcfb2eb7a9a35a7dc
1639 F20101109_AAAPZY triplett_b_Page_43.txt
59fcec9d3ce005cabba1a68aecb514db
4ccf1012b8ab72173ad59a60dc6fba5d434946fd
19099 F20101109_AAAQBZ triplett_b_Page_31.QC.jpg
ccb0e9d4ef0f6f9a4cb69ce59bc7d03e
934b44df9b693b23b4e00c2213fdb65e79afbb46
13051 F20101109_AAAPSD triplett_b_Page_29.QC.jpg
f9a48f8d1a40e289eb9aedfeaeb260cc
5dce73ec49599a3ef80dd605a034611d8ac67e1a
25265604 F20101109_AAAPXB triplett_b_Page_31.tif
0e21b4614b74acc7d7f8685dc495227d
adf7a45b97a08da5e0efe7c87be4e54950a7ccdb
2382 F20101109_AAAPZZ triplett_b_Page_44.txt
26066888af7d8c1af906b57ba8cbb1e8
678ab9f3c08e2e3b9bac7ef86bb80d56212126e7
1822 F20101109_AAAPSE triplett_b_Page_42.txt
7b5d7b3d358296dac0a5439a1b818e7a
d7bf5f4a017ea0f86c2ee311e3390cd552afe0a1
F20101109_AAAPXC triplett_b_Page_33.tif
4ffd468e27e27884abe66227ef0c267b
0505fd9c234adb6c0dab68666155885184c7c7f2
24643 F20101109_AAAPSF triplett_b_Page_34.QC.jpg
2f014935799814966a797222216c0779
049fb4a0b20005254c7105bc85793d00153aacad
F20101109_AAAPXD triplett_b_Page_34.tif
1ef553231fbedc055e1744708cd26b14
faa016c11c9f779b840f8b3c64c07f4205b43ee6
66580 F20101109_AAAPSG triplett_b_Page_45.jp2
6f5b69ae64b4c5d9dd20eeb1af590598
9e6251941909a4a39ccde2588e6bc5345eb057e0
F20101109_AAAPXE triplett_b_Page_35.tif
87e2a86571769dbd72b1a53b6c3696fe
a74ce81890ae78f3b9d69fd2b137d1bb844cbbf0
75528 F20101109_AAAPSH triplett_b_Page_24.jpg
341bb3549d5ae587f60d4b417fa3fb24
b483f46f83389f1fd44cbe248beac4c8b8e70433
F20101109_AAAPXF triplett_b_Page_36.tif
155ce0a9118a845fbf6b24df3af1f1ec
22276f0f7e787731ad17979cb0d4053a7e6c7cb6
F20101109_AAAPSI triplett_b_Page_32.tif
c2c3ada01f1d99544093dae820cb635f
d707857b3df9a434ad597ae222b6170d1af32d67
F20101109_AAAPXG triplett_b_Page_37.tif
e0f398a2e37bff3ea17fd7fcce506cef
441a12244132625da15f970c500151a5f3ca4238
2139 F20101109_AAAPSJ triplett_b_Page_22.txt
b39d9eb1310c61e75a4f9d16e0791f3c
901d09e61e870e315e6ff7ca075a12e5fd7259fb
F20101109_AAAPXH triplett_b_Page_38.tif
36c14b6b8c6b16c7f03348e9d277ea94
2d128a7ff25add9dffdf87b8597bb3e3c4b8cf22
100767 F20101109_AAAPSK triplett_b_Page_13.jp2
e6b2f9ebeffed0de95ddb9aa7b930058
9bf2b1fbf36ac4de2b792215107f009cec7c824e
1054428 F20101109_AAAPXI triplett_b_Page_39.tif
b235cf9e426f76a2dd20145ea687158b
c463dba0d2a72bd474c75837715d91d59a02276f
11251 F20101109_AAAPSL triplett_b_Page_39.QC.jpg
82a5d8e5c4db80c9f97c3e5f85ae68d6
c8ffbf224bc35f1f56cc9f397c1410b8a77162d2
F20101109_AAAPXJ triplett_b_Page_41.tif
a07fc535b2cff3b8fe4467d080a188ca
d2dce9805f89c1efd1b28698eaf9010cede389a8
9877 F20101109_AAAPSM triplett_b_Page_43.QC.jpg
4b7ddbcfd80ba291bc7f84c2f91ce6ec
0ad3d03fcf7756c16aaf9dda2daa4117fe7d3352
F20101109_AAAPXK triplett_b_Page_42.tif
d900294231b93592e7ac25c73e2b26d3
7fe39eab929a23d851fc5728b3330805485fe82a
14055 F20101109_AAAPSN triplett_b_Page_36.QC.jpg
cf8cc541db7bbc1811e69bef930c56ed
8381dec537be20350030a5ed07ca74c18eec11e6
F20101109_AAAPXL triplett_b_Page_43.tif
b0613ef4f6fb5ac6977de4e776b478a8
83977e16ac58bdcfbeca6813d2c6c5d0622cd791
F20101109_AAAPSO triplett_b_Page_30.tif
119af2a06d52b3b3257b919a56faf639
d1a2adcccb6828b7e1a1fe443f15d98ae727ac0a
F20101109_AAAPXM triplett_b_Page_44.tif
8ccb6b82eb3d35edd9d0211a5aad1a9b
f8b39bfbcf46a02418ad0e149bb07a4a43440c9d
9276 F20101109_AAAPSP triplett_b_Page_09.pro
bc758057acdb8a40d040b26e3c4d34b5
fa50df63098246995b7d1d9b92648a924e836cdf
35346 F20101109_AAAPXN triplett_b_Page_03.pro
a2416ca178c57c61ee929d48605f23d6
61a0f8fb59bea31379f6ef9d9e7b0f495b5f73f8
82932 F20101109_AAAPXO triplett_b_Page_04.pro
bc43f3b2f0805c90d434f93591078d27
0419af476d8a4c49352373a055931da04d9d5745
72107 F20101109_AAAPSQ triplett_b_Page_14.jpg
f69b4656f5d20911ee2215c0882d6bf5
88072ad9cb8799ad9558a5b96682f45a9ca42585
8656 F20101109_AAAPXP triplett_b_Page_05.pro
6db20eb0d3b4949940a2d660d4882bdc
06d1804ab4ed03c9a9abb135055750b033bf7bee
54537 F20101109_AAAPSR triplett_b_Page_30.jpg
3cfb8ab6fb0b5bfb5105c35dec7f9871
efeb4c34eea16ea554d99f4c33ee2781b1941271
26457 F20101109_AAAPXQ triplett_b_Page_06.pro
dcdd341e7a3dea5b98f704d66f9b8619
41a9bf2eb21ea21c77ffbea470eab7bc07864bd9
54326 F20101109_AAAPSS triplett_b_Page_22.pro
fa26edacb9370b9a9e63d9985ff4794d
592380fe63ae7f36c58ab959803cda431cd1a97a
12386 F20101109_AAAPXR triplett_b_Page_07.pro
33f38e202909a3c7fc406970ea2b4203
8ec634104d5b9138da8a4b9fbeec3ddbe98d5c18
1471 F20101109_AAAPST triplett_b_Page_03.txt
9cc2bcf7d0346379d3b3ff4e8d724fc9
4a579416c3327b0edf87c0d1998b5dad47b05680
49740 F20101109_AAAPXS triplett_b_Page_08.pro
1e718bebc7cf1355272febab76773947
98f9030f1e10029332139591922d7494c32a2d66
F20101109_AAAPSU triplett_b_Page_46.tif
d46e1b8e2ba06da8b050a0fc8b825de4
7a4916972735560dd059941d4b982ca2432db00f
50920 F20101109_AAAPXT triplett_b_Page_10.pro
7b16c268bdfa6645ad9f0729a078892a
b5a2d63d329174269f5209ed1f5a901141cb380f
55694 F20101109_AAAPSV triplett_b_Page_46.jpg
a05cc5df12efdeeeba9f1bc972e9281a
0e41abf57c273937c6297be3be253665291d83a1
52492 F20101109_AAAPXU triplett_b_Page_11.pro
d981b9a7fcfc4c866c2173346fe36d3a
c75d4c82af42f2893bc22829da30524989eb4c2d
6662 F20101109_AAAPSW triplett_b_Page_17thm.jpg
405100bdc714ffa225dcd26cec630ef3
0b5a96582ade5c47bb0242af8f072a09806e537f
24892 F20101109_AAAPSX triplett_b_Page_41.QC.jpg
0221f7f096ab3e2e6d4a5f96d792fc39
ef528f5053a7c60876873e613aa6dd3b534bf7c9
53704 F20101109_AAAPXV triplett_b_Page_12.pro
464e20ba899236671491a9d7de71285a
18825f9463a404f82bb9668b030c0ee30bfd7041
3094 F20101109_AAAPSY triplett_b_Page_02.QC.jpg
70af35e3b933b6ba87a656f31d3a83e7
1dfcb4e3839c90c443aba78c287ad6591ca4896d
46560 F20101109_AAAPXW triplett_b_Page_13.pro
f9a47f7b8e4ebdedf88c9b45132d4575
f3a2eb58e4e04a9328488a68ea2a24665ee640a8
2061 F20101109_AAAPSZ triplett_b_Page_17.txt
567a1f7996b00a0761121e48454c262f
88c5b1e18a811d2b0320c2c1cbde9a8499f2c6c6
55728 F20101109_AAAPXX triplett_b_Page_15.pro
6e0f55db3d59af2fef86ceaba71c6bfc
45eb44d8dba8acc4db5ff9a4f5385cab740bec2c
51754 F20101109_AAAPXY triplett_b_Page_16.pro
017c3d8d1b83ae08597e672bda36a7d5
3265b681d5d3c2fb7c14ca479b3b759b7ed82de7
25165 F20101109_AAAPVA triplett_b_Page_01.jp2
dfe837333f1bdee7927cae37b71fa363
37b612eda6d726225dc791dfd9c4251d2b8779c4
52026 F20101109_AAAPXZ triplett_b_Page_17.pro
474e245a4c2b5408e7633e92ea76de55
0a7365d918585cf8e2fd0cc4b805ad1d7d16c01a
5631 F20101109_AAAPVB triplett_b_Page_02.jp2
92427f3aad001f1783842587316e8924
64fd12c4b0926bb152a872a2d97dcc3ae3325c14
1051986 F20101109_AAAPVC triplett_b_Page_04.jp2
bc98931169690bbd403020316867d42b
69184a55e50faf62c7f20463397558c7cd049552
F20101109_AAAPQG triplett_b_Page_40.tif
0abd848e502c74368b6854884a70e48e
c5d0b51c0447ce912c13b684c25ff09bdbcf0578
5076 F20101109_AAAQCA triplett_b_Page_31thm.jpg
d0c8ebe758b6f52d256ee39264a7523f
6a8311e1601984f2d9392122ec6e6591a28ed018
201592 F20101109_AAAPVD triplett_b_Page_05.jp2
6b0ab6d7df0ce6fcc19d4f3602c785ee
540fcf8dbf879cd0e433b8ee05bca5261164a372
F20101109_AAAPQH triplett_b_Page_29.tif
396c2c3a871ecd52bacab117a864c6b6
d886904a91fbc5fc9be5be1ad273dc477d2d9022
23012 F20101109_AAAQCB triplett_b_Page_32.QC.jpg
35d1b313bd01fb73464c3d191a822f4f
bc3ad2296582c4bc0ac7974c24c6d5b2a18c4606
724437 F20101109_AAAPVE triplett_b_Page_06.jp2
173086ee59360f46307584a9bd6da708
3348cc86c5bd35acb9923b2c6c0d0438b7fd0a63
27856 F20101109_AAAPQI triplett_b_Page_39.pro
14398b124374b99002156dcb252e4fe3
12149a9b2eb8fd858fd70e04b83b628be4c8b8e5
23065 F20101109_AAAQCC triplett_b_Page_33.QC.jpg
2f954aaa162e4ce6dfe3e4498e0c180e
f8d91c48d63aec22ecc77971406e31586c331dad
28929 F20101109_AAAPVF triplett_b_Page_07.jp2
43d1d8853ef5a9c216946ed19db534a1
8246465d34eebd7f4df7d002bf4e6649dde62d45
14400 F20101109_AAAPQJ triplett_b_Page_45.QC.jpg
d033cf95781e40a8177573841362f71b
275cb8e74be80c29825e4e62caa348d7c2122556
6403 F20101109_AAAQCD triplett_b_Page_33thm.jpg
5db07d8f0aca5257723746a887558e1e
ed8e6b76684edac6cbea55780c3fe734df8197ad
107617 F20101109_AAAPVG triplett_b_Page_08.jp2
255aa9e0a6324bd14ac63bc10271211f
75047da44a601a38c12ec3e4f3baab7b9b842fd7
1051983 F20101109_AAAPQK triplett_b_Page_31.jp2
76640cac3f256f477936223831fbf662
3fb4e542aee8f80c86d6890ecc97e73f1104149b
6595 F20101109_AAAQCE triplett_b_Page_34thm.jpg
63a6b3cd83dee7d5aae1d681626ae45b
caca5f787d582bfce2f66851895b1e9e8e200b98
23158 F20101109_AAAPVH triplett_b_Page_09.jp2
b2fb886feb2d17fb2b8bd84275b75cef
ab8f787478f88ced113ee533f4e6ee35d9212989
2046 F20101109_AAAPQL triplett_b_Page_36.txt
5db7ea400aa3b6330ab243731fa1c69a
c479877e474782f0b53ca03cab59c6f75ac232b4
16540 F20101109_AAAQCF triplett_b_Page_35.QC.jpg
2c5cb47e07bf348e70c5016652661e40
0a9c4bcedb83aab83f871ac7bd1058f91a7e4b93
109064 F20101109_AAAPVI triplett_b_Page_10.jp2
e33bc0a9b3fcad82f2006bb33ec0245d
faf7a1a356d9613eea540f36485906af6421765a
104533 F20101109_AAAPQM triplett_b_Page_14.jp2
715038f661e6aa3637579359cc099612
f1dbb53bbd79ecf307479944258391a720347f1c
4650 F20101109_AAAQCG triplett_b_Page_35thm.jpg
1351638d40a790581e30151e49d16265
ee6c39c2193ce81fa31709f564dee6a0920dd008
112039 F20101109_AAAPVJ triplett_b_Page_11.jp2
1106433555057013750780ec529d404c
2d1a56f2d11a2f3852143fc005631c4d699405a9
8045 F20101109_AAAPQN triplett_b_Page_07.QC.jpg
7f024da5f560340b63cafccd3bb3ec47
06d07769cb99a11efd62aaf9fdd245453de3ac37
4453 F20101109_AAAQCH triplett_b_Page_36thm.jpg
03054e665df26cee9ee476e8f9f99df4
7ebc59ced93a2094f1ba0bc5230ba541136cfef1
114851 F20101109_AAAPVK triplett_b_Page_12.jp2
a206cda1b15cb9ad4e54c6b8c0012283
00782dc3c5747a874af8092c5bee7a0bdc2a768b
8812 F20101109_AAAQCI triplett_b_Page_37.QC.jpg
927635d11b00ac4250211bdeb74546c2
eeb54035fd642fa4c57c1e25540ca08ab97d6982
119852 F20101109_AAAPVL triplett_b_Page_15.jp2
2386c0cb8334b635aef3e9d7f8ee3388
1f9660ee80b47d4c32c23109498461bf51d3e20d
10001 F20101109_AAAPQO triplett_b_Page_02.jpg
6e2dfc0e5763e68f9eb1cb6f91ee3710
9073107e79612c88a05e5d39ea74238cafc78b35
18065 F20101109_AAAQCJ triplett_b_Page_38.QC.jpg
d20b3c606da917582216e4c32a0842ca
f549132f83860d76992d756c6cb3d9155762f996
112738 F20101109_AAAPVM triplett_b_Page_16.jp2
92a5e8572bc9c35822978cc2ba16fb9f
c4741e434838039e0b895d4a8e2faff2402e3089
24292 F20101109_AAAPQP triplett_b_Page_10.QC.jpg
8182147bed0f5b96bab852147e99735c
0e786e1b16e84d59e5e3ae401e67918258e57f45
5505 F20101109_AAAQCK triplett_b_Page_38thm.jpg
f80c274ea8c2a93218147152cb721cca
b17f9e9a2c1e26219b399b682f1c160db00158be
112311 F20101109_AAAPVN triplett_b_Page_17.jp2
7f2104cc6d3fa947a652e0cabf6df46d
ec05d8d01be4311c9dd20ce45de30a94e20638c0
76114 F20101109_AAAPQQ triplett_b_Page_31.pro
f6b9aad8ce51d3fc9a4aab5208726c68
7eff8a246e188f91fbb7be9473e03d53b7dfe793
3309 F20101109_AAAQCL triplett_b_Page_39thm.jpg
1c6e006a6001d8043292e343d767d53b
a2c3a8bcf484684e2d55030edd2a36b0cb6ba3fa
113016 F20101109_AAAPVO triplett_b_Page_18.jp2
3e2477e39555b6db681fc80cef187e20
e735c2613c03ea3bc48d860dd471bcee7c7de396
47849 F20101109_AAAPQR triplett_b_Page_33.pro
3cb9e9a45157af895e4a9b2e7cefc0f4
b8c77d388ec290aa512658eee071b879703f359b
23117 F20101109_AAAQCM triplett_b_Page_40.QC.jpg
c1af3459dece6556e67bb460e90c6642
f1c4bbde15c1798a9d97a97db4174ef2dad3e33a
113368 F20101109_AAAPVP triplett_b_Page_20.jp2
6d2db45792ba9c6d084c515d1624c325
acc43de7e1e1ea53bef3cfef3576afaba2329e8d
116156 F20101109_AAAPQS triplett_b_Page_22.jp2
8279dda7d0f62d05bf99280b444cb8c3
0941753d8c8f6d2d281bec9e3ee462ae78a76d5f
6565 F20101109_AAAQCN triplett_b_Page_40thm.jpg
e9f6fd9b9f609c016e1186e718788e8c
4def9ad744d331fae156512938318b144d95fbb9
104918 F20101109_AAAPVQ triplett_b_Page_21.jp2
b2ad628c64af8bb1b8b24629b7576e5e
ea87625efff5e9b3852182dce9afd31bbdbf0e99
2152 F20101109_AAAPQT triplett_b_Page_08.txt
d094a5d72ea898585e2aaff34b779e92
748fad91d59d297db1def97234a7640cace13bac
21979 F20101109_AAAQCO triplett_b_Page_42.QC.jpg
76e2620cb3cbb4d0d9b5fb5326ed6e43
0b1d3b4cf2eb7e748b8be3c65390645e88528997
113920 F20101109_AAAPVR triplett_b_Page_24.jp2
11b3ce46ad9e3838673bd50e462e9335
35174cff87e20169ce065a3daa1df078f14e8b18
F20101109_AAAPQU triplett_b_Page_13.tif
74decc27291cce5396fac150ac2d7f96
77ef0c5e5459ba9b7e11d41c3f4c056b9549471c
5948 F20101109_AAAQCP triplett_b_Page_42thm.jpg
230a34b31372b2e391310d97bf80b142
96012abf2a95597a5700c8841c437c8ec18ddd3f
98729 F20101109_AAAPVS triplett_b_Page_25.jp2
2e6825b1545f9d520d98026f1e58ff79
d98ea70190adc89a6fd08b9138f833cd258a7edb
132322 F20101109_AAAPQV triplett_b_Page_44.jp2
dd520b8fdfd79ee5facbc3a3627336b3
a4b79e6f69bd8b3450ef1485e8a94ea761696a0a
2789 F20101109_AAAQCQ triplett_b_Page_43thm.jpg
acd877371fe24b13a65a2506b466e9e1
cd91bdb7ded41b36ae4b013c7309fa66d94dbe51
638 F20101109_AAAPQW triplett_b_Page_28.txt
421641e3203574378859bf8b0dc0b198
139ab491c1233dca9cd66ed1e169b4e7cbbe9b59
6969 F20101109_AAAQCR triplett_b_Page_44thm.jpg
8a3fea332d7f35b7fe304299411773ee
acca398168ff23531970c75f1baae26daefd019e
1051976 F20101109_AAAPVT triplett_b_Page_26.jp2
73c35ae86d2a9467e8c0719be8a5188a
a893fcb40090e5f0ec0aa9343ba86f57a96d2326
47110 F20101109_AAAPQX triplett_b_Page_14.pro
b661d14869f437f50e8284fe3561202c
1ab5135d79f057de2f072795dafce7445ef6b875
3900 F20101109_AAAQCS triplett_b_Page_45thm.jpg
acc415964d5ffa7d391e1efd5c6c556d
49db66dc1c94220aab888cfaf393f47e10c1a862
635904 F20101109_AAAPVU triplett_b_Page_27.jp2
6819a4917e420299e318f65eaf277afb
83e5183a082062eb7a59387c344d2b016bd0da43
6847 F20101109_AAAPQY triplett_b_Page_41thm.jpg
a86d9de8e68f94cdfbf5312152ba98b6
ec061308200a2e680493da84a4e7218950cd060e
17939 F20101109_AAAQCT triplett_b_Page_46.QC.jpg
a7faf19a398a2a73cee250819daf81a4
695f582bc6d058dc6c90192e5a87cb9a9e9b9ef9
618360 F20101109_AAAPVV triplett_b_Page_30.jp2
4141316d39330501d6275f16a4ab70dc
72c9021e1a8745139cdc9da883a0d73684f39669
1909 F20101109_AAAPQZ triplett_b_Page_33.txt
28865a9a6681e4b1cffb22d669794020
76f51b488aac9e8f8798467c341b092891b0a379
5193 F20101109_AAAQCU triplett_b_Page_46thm.jpg
c1d5ae3151b15e060d9a326f1a5ba1f8
cc5f0b2c0afcabbc28972b05d2f8570b5aba0fa0
101848 F20101109_AAAPVW triplett_b_Page_32.jp2
b6cdf1bf1d2c0e52c0caf31c1556ef22
f9ae3b7f5f7500fdfd10b14a12e4433ad474ba1c
112911 F20101109_AAAPVX triplett_b_Page_34.jp2
b7db201a7d64df4edf717ce5263d92f5
92961c3f67397a4ace708e5d87bf86cb892dfc62
612305 F20101109_AAAPTA triplett_b_Page_29.jp2
eb823dc8f3318e6cd26c471b3352021a
621aaa7d80814531491d50742cdfe5534c450996
70735 F20101109_AAAPVY triplett_b_Page_35.jp2
60cfe16d7eb35dc0ccb866d46706130f
c8633fcb9c6d2490170c307a947bf8e45b58abb1
2002 F20101109_AAAPTB triplett_b_Page_40.txt
0ef6dee0e0eb85c6bcd63b142a40e49e
8c49b1410df75740b1b7ada1816d114c1ae72bc0
551692 F20101109_AAAPVZ triplett_b_Page_37.jp2
e46c07837a81ca7fa903c57b3105bc09
30425de6a63d6652e8add4da5f37d14e01a561bf
8407 F20101109_AAAPTC triplett_b_Page_01.pro
2bbe1b49663c718264a8b1a05a9d7e77
4b53bf57a0a3fc46b1442d835653d292e2af9175
1199 F20101109_AAAQAA triplett_b_Page_45.txt
3759016fa9bd9977ca0744f2c4bac8c2
a296a4158a19c4ba563afed0d10259dd7937b77b
5262 F20101109_AAAPTD triplett_b_Page_03thm.jpg
33b6bedb0281b2516f614f2a1e4f020a
3ad6724d16c321a928004e350e18a795c9e8b3fc
51925 F20101109_AAAPYA triplett_b_Page_18.pro
a1b9f4cf57b75d3b503eb218d0269b3e
d2a322c24bc7d82ec669904a5c9cead0d14adc87
1434 F20101109_AAAQAB triplett_b_Page_46.txt
2d0d42ca3ed6675ec61ab1bc81e874cc
b03a7e7e9c1f488805645386cfb57e2779c080b0
23990 F20101109_AAAPTE triplett_b_Page_16.QC.jpg
3c9a4c0938323a42d0f06851c7b12931
f9971596fe1a206183e4e850f77d50b1b86c132d
53164 F20101109_AAAPYB triplett_b_Page_19.pro
8019ff800d6fb7e0e81085b4b760d8db
5065d33d820b05aa4d55f3e7ed1320e4e775944e
6988 F20101109_AAAPTF triplett_b_Page_15thm.jpg
59f86665a7b4ad89f01bc25c99e01b85
56d37c12c8efcb57f3611d5ae66c86027bd01573
53216 F20101109_AAAPYC triplett_b_Page_20.pro
1bbdc36f8f8f2baecffee669feba7b45
9cae841b0be46b5263b814a2156e7f99e69b74ff
2323 F20101109_AAAQAC triplett_b_Page_01thm.jpg
70bb0f42513734560886b7ea5307778d
58d71bde41627ec026043d156330d03d0c963a9c
1051960 F20101109_AAAPTG triplett_b_Page_36.jp2
92110e1050bf0087c32752976a851920
8e85104b6eb8d5db8206e224401bcee4aaa898b3
49777 F20101109_AAAPYD triplett_b_Page_21.pro
4c9d7d3b8d15ef603a82352b133eca3d
a73d02b49c34dbfb788049ae5f55cd459a4bc264
1308601 F20101109_AAAQAD triplett_b.pdf
95cd0e8807408c13931acfc95a565a67
f448ccbd2e3b354851a436392727281af9e3afb1
F20101109_AAAPTH triplett_b_Page_24.tif
88c5d0534f3d35b4c7dbc7f1ebc0d187
4af10310aace54361b74573cd516a15c233c8778
52565 F20101109_AAAPYE triplett_b_Page_24.pro
fcd12797917f4667a48f926725a6e56e
f75d49d74cd0cd48709d58f7a250ee3a22d50af4
6799 F20101109_AAAQAE triplett_b_Page_22thm.jpg
f599ac64291a4b8d2d67dfe8413ce441
52270b3928b207bee7e8e6495ce4baf1b29d0fdc
51245 F20101109_AAAPTI triplett_b_Page_23.pro
1fd969eed9bf7acba6c4ef4d06896b7e
34f2b8439a0ccd4c71a045ae0843038ad3615de6
12010 F20101109_AAAPYF triplett_b_Page_26.pro
67fe5c9adee3d8d5918544ad47cf976b
3f437440514591a99b4bf75a2e71548d2dfa7e98
2945 F20101109_AAAQAF triplett_b_Page_37thm.jpg
09984784fb3eb7d363d62cf61fe8f14e
307edc6c3dca30606ea7a037c2ab336cdcfb8755
F20101109_AAAPTJ triplett_b_Page_06.tif
b6f82b0c34154ed8e42675db59669399
2a0b4bb627814d1bf56e994fdb014f4bcc3c1965
22240 F20101109_AAAPYG triplett_b_Page_27.pro
db1f0cdbb8b3e9ab7b8d55345ce22572
78ddeff88a7c6ed4264bb022682a0114c10d62ba
2245 F20101109_AAAQAG triplett_b_Page_09thm.jpg
cea0187404da1affafab016a7973b1a9
d765120da0a326a8da58251efc9cc1b8232e2e64
F20101109_AAAPTK triplett_b_Page_25.tif
8ba5432eaefc38ea38a3fdbf369b8041
b6c18e9debaa61eb4b524683acf4a99c7a89a530
12625 F20101109_AAAPYH triplett_b_Page_28.pro
155911354ab6e261515a4775d6f852c3
52f549bca71bbf29352849e20b7738078a401351
23710 F20101109_AAAQAH triplett_b_Page_08.QC.jpg
f6b2dd5af2642580f394e4009a29a0fb
92238d5e6e55578390bc34a00f26679e6f07a229
F20101109_AAAPTL triplett_b_Page_15.tif
805be50a24b3c109f90f74fd6f260969
207c09b0790d80af012d44b6da0541af7924f90e
33955 F20101109_AAAPYI triplett_b_Page_30.pro
13b527bf4de83413ce64853230fdf132
f93e4b7a12efec38fd85bd2b6737a114a0439318
7056 F20101109_AAAQAI triplett_b_Page_12thm.jpg
f3ec399607a22a4a8f76d4aaa067d0b2
d1a3c60ef4657ef4bb261c6dfb661eabcb9a01c6
12311 F20101109_AAAPTM triplett_b_Page_26.QC.jpg
836f5c6ac8050ee8b2a81a4ad8f3ea17
0aa3ac77eefdb4f62803cc9ca503fbf9206a51df
45843 F20101109_AAAPYJ triplett_b_Page_32.pro
e24ce362ac6a79a14ecbdf7cc28241b0
6e40b881e74fb5c0b0b3a472b42d7641223e38ed
9761 F20101109_AAAQAJ triplett_b_Page_06.QC.jpg
1c277671fed0965c5b37d44886b4e3d2
03b1c7db99b9d4c8064459ee40234b32510f7d7d
1912 F20101109_AAAPTN triplett_b_Page_13.txt
e8107a1e74ced175873db894c2bea19d
28a55d8aa1047bc974263c085a11e06ba78b7a67
52353 F20101109_AAAPYK triplett_b_Page_34.pro
feb901e6b555baaa3b74b0a662b1f30a
96e3fe10a0b80faddcb8f3d90e2ed9388689a916
6452 F20101109_AAAQAK triplett_b_Page_21thm.jpg
55cb750ec8da67b5ed5d4a6a7a261087
a0a512ae31ad36075a6de31344ff2cd3fc651070
2108 F20101109_AAAPTO triplett_b_Page_11.txt
1c57d08f0e83bb6fa25bbf5d57fa79ea
3ff72e7fbf70502a9aeaed7a360cd84caf5b9e6a
31839 F20101109_AAAPYL triplett_b_Page_35.pro
0c7a90ea420ee3b18d6ddd0c2773ba28
25770c572100f8a402ee012770912eb28af7254b
22548 F20101109_AAAQAL triplett_b_Page_13.QC.jpg
f21907da9995631964a1412365e5ca11
234915a8bb5bc29a6e7f5a76b0c327e6c794f77b
56501 F20101109_AAAPTP UFE0022098_00001.mets FULL
616cf1948a1b52b8c0da966b11061fa5
ee2bc52d62d078d5ff8d4caf619e596b15c51f22
37238 F20101109_AAAPYM triplett_b_Page_36.pro
90da131e1881f07c219e690a1626afc5
38f4a9097cd97c573e043a3bf97062bfdac6c9fe
2652 F20101109_AAAQAM triplett_b_Page_07thm.jpg
dd7cbd4180ed0e49768151c45bcc386e
1db01bc91dedf860d0b569445deb08925e42e317
14882 F20101109_AAAPYN triplett_b_Page_37.pro
31f083b10f78b474248dcd245508d8cd
06002b54ebca4ad3e6031d92392cc3cacf0f310a
18653 F20101109_AAAQAN triplett_b_Page_03.QC.jpg
7356fd19d73f6bc97477cd4bd3e15273
d786e1f49d3a6e6e98437be6e1b8362dbc419ab8
32155 F20101109_AAAPYO triplett_b_Page_38.pro
1f2949a6a10acbb487f04a95a0cf596e
d4396b0be50cdf6d7ff28a068d60f17b4c2c30e5
72625 F20101109_AAAQAO UFE0022098_00001.xml
33e17f31bfc587581f8dfa2ca471f1e7
6f3bcf7b5adebd9b9294e252c214ab4e3bf97c14
51855 F20101109_AAAPYP triplett_b_Page_41.pro
02ac9dff48a27a9c08d13db0e273483d
ab2b778335677e708a5f2d3662c886760e53d685
7438 F20101109_AAAQAP triplett_b_Page_01.QC.jpg
728366205365a180d62b6a27164e72f2
2bce9dc28cfdee711e20305426b2072eec3b689a
24138 F20101109_AAAPTS triplett_b_Page_01.jpg
a5b6f41c8621e16cab688c3df36cf12a
64b23f4dadf12c13d0987d5dfa0bc052f624b00f
44709 F20101109_AAAPYQ triplett_b_Page_42.pro
49831646cd1a7e77febee627662f3754
2df67d7729e82051e4e09993d0c425412d65f2b4
1402 F20101109_AAAQAQ triplett_b_Page_02thm.jpg
4a78b724f6c3404e7d6bc7533acbfeef
e074901cf8801b87fd0ead849a69a0f68495a714
56444 F20101109_AAAPTT triplett_b_Page_03.jpg
1f75518e6860ab1f7e2f828d95d9e277
d75f2da7c23431f65a49081de12a21dc8473a33f
26926 F20101109_AAAPYR triplett_b_Page_43.pro
9eae5a5ad75b33cb678e40a4ef8ce77e
9cfd89f46ef1b96ee40bea0979c4a8c113adc3af
20587 F20101109_AAAQAR triplett_b_Page_04.QC.jpg
139453b7ad553850cd9bbcb0f82794f9
93881fe751558e294967faced444cff95aee1ca1
33033 F20101109_AAAPTU triplett_b_Page_06.jpg
b8c675d9989ad4fddc3dd000761e60e0
3aa7b5774b36f990560c8e5cd1452ac2a1b4fad1
29058 F20101109_AAAPYS triplett_b_Page_45.pro
2f76148e3f68b66e2bce6b42ad8c1b1c
7f23b93f2948adccfa5ca5617709684620a8c849
5607 F20101109_AAAQAS triplett_b_Page_04thm.jpg
eb732bf818e4f2e3282d977976dc5dbe
e312b82e01a6dd46da41c7f8d4a0a130b4251908
24102 F20101109_AAAPTV triplett_b_Page_07.jpg
0c2759cfc953b8d8ad31b90ae133993d
4c5eaf2ac57c65bcd593a5694bb97fd28d143430
34859 F20101109_AAAPYT triplett_b_Page_46.pro
4c4f98f8bce3747895274874be8ad6f0
42ca370d18951ab5329de99431740ef7b84151d2
4908 F20101109_AAAQAT triplett_b_Page_05.QC.jpg
8b33948cee8704449573768dc7ccd9f1
ef16167308d3e3334fb2ab9998f9ad8c50eab531
73349 F20101109_AAAPTW triplett_b_Page_08.jpg
46784be1fedeb6c5cd8d7b8e3080c397
97debf7efdcd6b0e45b2a405b9146ef12775ab5b
494 F20101109_AAAPYU triplett_b_Page_01.txt
f5bdbc56d940c490a00ae504a1be5fe5
dcfdde0fcddb6516b9dfbec1704d13e38f9df2aa
1753 F20101109_AAAQAU triplett_b_Page_05thm.jpg
6ff2eb869cf1ce079b3b728fae750f9a
5387a4855140361e889188c133132ade4c92466c
96 F20101109_AAAPYV triplett_b_Page_02.txt
d30b432715f4953578175ae486d95857
b95824d164a5232ac2591ae260d83ffe434a97af
2773 F20101109_AAAQAV triplett_b_Page_06thm.jpg
b93eae3610089fac2c9434c7b765e44f
b5ab7d55cd3963d8a2b8e50f64cef52e799aea22
20186 F20101109_AAAPTX triplett_b_Page_09.jpg
3fb515ee1dcf85ea45d76cc0be2da602
bf557579bc78721cbeb4cf4120de68b119584cd6
6405 F20101109_AAAQAW triplett_b_Page_08thm.jpg
5f5ac8e6d0a04978d8a1973584aa6592
732b886dde222b23b6c3d8f96ee4ac76b4e53866
76575 F20101109_AAAPRA triplett_b_Page_12.jpg
cf08c5ea778fa7438c75f4967d56f299
fc2e96fa37dc273113161cc607f942aaffe6366b
73318 F20101109_AAAPTY triplett_b_Page_10.jpg
3a0bbf0ed957cbcdc5493741c63b1aec
b5fbee83e4681a0b275df3aa38478feb702e42e3
3399 F20101109_AAAPYW triplett_b_Page_04.txt
052d59c6cd9102fc63fedb3e916a007b
8d1d7a295d5e96e40916f6b5483ccda8dff6e6fa
6566 F20101109_AAAQAX triplett_b_Page_09.QC.jpg
3bbc729c8549216d1d88c0b2a2709747
1dbf1da3e4bbbb4ca1767797d8aaa51786085a26
76772 F20101109_AAAPRB triplett_b_Page_22.jpg
a2eb22e4f3722093b876a778d97473ea
b8260e17da4f71ccc996b88d4d7c5764f899afca
73873 F20101109_AAAPTZ triplett_b_Page_11.jpg
8c3c90eaaceca303218330a4d918591b
c449a8b88275f5057eb627cc9ef7e60d60290df4
396 F20101109_AAAPYX triplett_b_Page_05.txt
5fe30cd6eb0147d312d80fb7ac4e7d59
4c453fe74752b9634ad3c7f447fdfe148eb1b79a
6537 F20101109_AAAQAY triplett_b_Page_10thm.jpg
b1f58bff10ba82993c4dff688d7146e6
06c8a929d03856ebe0ac88cfc53bddf36eed0071
58081 F20101109_AAAPRC triplett_b_Page_44.pro
0eb27cb3dd677114c3c0e0c6e0bc1432
c6c4c41831ba0330a574fcb6af95d0c987b7b098
77159 F20101109_AAAPWA triplett_b_Page_38.jp2
47188b687f2f902a2e498897383ba64f
f64169006de23514b23489539535dc63045094f1
1115 F20101109_AAAPYY triplett_b_Page_06.txt
0f42beb6d25233da7ec3f09bee578d68
c6f4d40ca28f72371cf226a55e9bf5c6c05d5afe
24328 F20101109_AAAQAZ triplett_b_Page_11.QC.jpg
4c6ccf0c0d9404db16c30c953d6ca202
e073b9f927cb31f6dd19690fa9d037cf7f4bc2ea
4571 F20101109_AAAPRD triplett_b_Page_27thm.jpg
8369f0e4bc1b5b28b349241446d9e159
9bc73967cba30c5aa21fc41c570ed21f2b36b35f
68357 F20101109_AAAPWB triplett_b_Page_39.jp2
82bbf11cff05eeaa5c344553462bb3e0
d084d615dffa4b526522924e75cbee3469257fed
602 F20101109_AAAPYZ triplett_b_Page_07.txt
99b79a7efd4cb98e533adb4d7dd2239d
1dd13008bf6c330cdf48918578bf00afb0fd8554
6394 F20101109_AAAPRE triplett_b_Page_32thm.jpg
cb75c67b300274ff0fc1af913c31650d
8f9ee84fece5999cb41e310f4dba56cdf861f0af
103839 F20101109_AAAPWC triplett_b_Page_40.jp2
4f49d2f79c7bd6c89f51b131b17789f4
3cc6d58777c89085c4ea2ff324b308abba785c23
6949 F20101109_AAAPRF triplett_b_Page_24thm.jpg
17bb40bb7ca10b9bfb394fe884c84855
45e806a779cec6b77bf6c00cbf0073d7c636483e
949095 F20101109_AAAPWD triplett_b_Page_43.jp2
6e706cb1d6158ddaca4e538ef92d722f
33c48562e26141b146f526c3f1966ce4aa6ba25f
12988 F20101109_AAAPRG triplett_b_Page_29.pro
e85f6c24920a85358e0bca82efcdf4a2
2daecc3ce97c1bb4c45f6671380ce1bdc9430ba9
79976 F20101109_AAAPWE triplett_b_Page_46.jp2
f660a9591bb64f96e4b6a648217bead4
73de4d692b27e0df44038f1781172018333e2e7e
25931 F20101109_AAAPRH triplett_b_Page_44.QC.jpg
9f2bdb9e3f9d7afe1b9cc4491c207e16
6d3dd78a4f0cbf40af688af3e0c90def7ad00cf1
F20101109_AAAPWF triplett_b_Page_01.tif
dd3b9a4ea2fb4c3dd661f88e1c660364
9faea4792f62f277350c3e5ce1fa9b5009cc36dd
6728 F20101109_AAAPRI triplett_b_Page_19thm.jpg
3bbed09bb4134b06771c6ef94421796e
baa6d5d8214c5faa6d000958c9bab346cfa46109
F20101109_AAAPWG triplett_b_Page_02.tif
d83caafe3587fe61ba16dab31a3f07ec
4181426a969eda9150e14a933e25d8d91acb71b2
48545 F20101109_AAAPRJ triplett_b_Page_35.jpg
4286bd0b26ae64cc0c1efc8a140a0287
582b49fff9a41c591015f2e413fb4445fd1d4b53
F20101109_AAAPWH triplett_b_Page_03.tif
d9154d04c59b3d809adb5e85059c8dce
81f5d26dd443c561273b1de4bbfd366c4e10f270
74727 F20101109_AAAPRK triplett_b_Page_18.jpg
31df4332429bd63f30c242fe1c870929
17ba6b696ccc3e25ece404da75c54dd2a352bd4a
F20101109_AAAPWI triplett_b_Page_04.tif
6ed5427fcfddb9e11fdb1251c5e55afa
6fa84d1caf0baf1ea41539b96209f274427a6146
98681 F20101109_AAAPRL triplett_b_Page_42.jp2
5fa7a2a23f56327f6d26be12007572c3
a4b7ae69419aa936f37cd14d97f949fa0e780980
F20101109_AAAPWJ triplett_b_Page_05.tif
3c72fd87320670e1e759ec7a7c508880
5fb231d8538450ef409c609841240b3538e256a5
991 F20101109_AAAPRM triplett_b_Page_02.pro
d9d58b09cb66ea49a76aac279fb797f9
ee2279ba3f77daa9c083299827644b5facec2a40
F20101109_AAAPWK triplett_b_Page_07.tif
38cb5beb147d4c0836d3b7091dd61bf7
f4240109aa3a6d7118de661b1f59c0673f6e827b
76762 F20101109_AAAPRN triplett_b_Page_04.jpg
2a0704e20e3fa31eb5529429d474da36
d1584e7bb4ed6f41fa38dab35a36f8f0fd099b15
F20101109_AAAPWL triplett_b_Page_08.tif
d7a5505f3df014b7c47af2169425105a
b8a15c875a4086554302c77341530631654afe5e
3950 F20101109_AAAPRO triplett_b_Page_28thm.jpg
3c5682e471d236dc9f4580a82f96a3a9
05de9fe14ea47ba03ee6025cad850fc4534da73b
F20101109_AAAPWM triplett_b_Page_09.tif
100ff706d8e17095856d0bc0c5c4d588
a5e3a284eb01cbb95c57c16df2b04c322568ab1f
F20101109_AAAPWN triplett_b_Page_10.tif
34fd1e6166a50ce974a777b03392ab37
42e8f46b41aadf0c1d97a726d6628913467692af
103786 F20101109_AAAPRP triplett_b_Page_33.jp2
65b4635e3bf7e9adadbf426aa9667ab4
7f584b5b68e013b3961c0c2d4d97b74eb498338e
F20101109_AAAPWO triplett_b_Page_11.tif
644874b5aa2a045bab0120522fa63fba
27d9185ebae3567f52e8393952b7f987d5871a2c
110506 F20101109_AAAPRQ triplett_b_Page_23.jp2
5279bb5e0a87f0f07f6390e353d7e36c
a4b731a0fc943fbb01129273b87e38349fcd8215
F20101109_AAAPWP triplett_b_Page_12.tif
fc9a9075c75ba78fa0daf17e3d491133
030c136eec9265a478efe99cd3cc8db60ec474e3
15767 F20101109_AAAPRR triplett_b_Page_05.jpg
b7554a04b117d4ea576f7118fe28ebd0
1ff3c7c67608592302686dd1c4d898b34f5a9fcc







DEVELOPMENT OF AN AUTOMATED TESTING SYSTEM FOR VERIFICATION AND
VALIDATION OF NUCLEAR DATA




















By

BRIAN SCOTT TRIPLETT


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2008


































2008 Brian Scott Triplett









ACKNOWLEDGMENTS

I would like to acknowledge my advisor, Dr. Samim Anghaie, for his mentoring and

coordination of my graduate career to date.

I would also like to acknowledge Dr. Morgan White of Los Alamos National Lab for his

ideas and insights into the nuclear data world. I would also like to thank the other members of

the nuclear data team at Los Alamos including Robert Little, Angela Herring, David Pimental

and Doug Coombs.

I was not the only contributor to this computer program. Many lines of code were also

written by Christopher Sommer, John Hopkins, Kristen Triplett, Angela Herring, David

Pimental, Morgan White and Doug Coombs.

Finally, I would like to thank my wife, Kristen, not only for her direct contributions in the

form of computer code during both summers at Los Alamos but also for the support and love she

provided during my entire graduate work.

I gratefully acknowledge the Los Alamos National Laboratory, in particular the Nuclear

Data Team within X-1-NAD, for funding much of the work presented herein.

The research at the University of Florida was performed under appointment of the Office

of Civilian Radioactive Waste Management Fellowship Program administered by Oak Ridge

Institute for Science and Education under a contract between the U.S. Department of Energy and

the Oak Ridge Associated Universities.









TABLE OF CONTENTS

page

A CK N O W LED G M EN TS ................................................................. ........... ............. 3

L IST O F T A B L E S ...... .. ................ ............ ...... ...... .. ........................ ..............

LIST OF FIGU RE S ................................................................. 6

LIST OF ABBREVIATION S ........... ..... ............. ................... ........................ 7

ABSTRAC T ...................................................................

CHAPTER

1 NUCLEAR DATA VERIFICATION AND VALIDATION OVERVIEW .........................10

In tro d u c tio n ....................................................................................................................... 1 0
M motivation ............... ............................................... ................ 11
C current B enchm parking E efforts ....................................................................... ..................13
Scope of this Thesis ................................ ......................... ...... .... ........ 13

2 DEVELOPMENT AND IMPLEMENTATION OF THE NDVV SYSTEM ......................14

H isto ry .................................................................................................................................... 1 4
Overview of the NDVV System ........................................................................ 14
Input Generation .............. ....................................... 17
C o d e E x e cu tio n ............................................................................................................... 2 0
Output Processing .............. ........................ ..................23
Plotting w ith gnuplot.................................................................. ........ 23
D database storage ................................................... .. .............. ... 24

3 DEMONSTRATION OF THE NDVV SYSTEM......................................................32

E xam ple Problem Specification ..................................................................... ..................32
Execution of the NDVV system ........................................................................ 34
Storage and D isplaying of R esults......... ...................................................... ............... 34

4 FUTURE WORK AND CONCLUSIONS................................................................... 40

F u tu re W o rk ...................................................................................................................... 4 0
C onclu sions.......... ..........................................................42

APPENDIX: USER INPUT FILE FOR EXAMPLE PROBLEM.................. .... ............... 43

L IST O F R E F E R E N C E S .................................................................................... .....................44

B IO G R A PH IC A L SK E T C H .............................................................................. .....................46


4









LIST OF TABLES

Table page

3-1 ICSBEP experiments used in example NDVV run.........................................................38

3-2 E xam ple results sum m ary ............................ .............................................. ..................39









LIST OF FIGURES

Figure page

2-1 Example XML template file used to specify benchmark problems.............................. 26

2-2 Input generation process for the NDVV system..........................................................27

2-3 Job execution in the N D V V system ........................................................ ............... 28

2-4 Output processing in the N DVV system ............................................... .. ... .......... 29

2-5 N D V V database design layout............................................... ............................. 30

2-6 E x am ple B IR T report.............................................................................. .....................3 1

3-1 BIRT report generated from example run............................................................ 36

3-2 Exam ple problem results from gnuplot........................................ .......................... 37









LIST OF ABBREVIATIONS

QA Quality Assurance

NDVV Nuclear Data Verification and Validation

V&V Verification and Validation

LANL Los Alamos National Lab

GUI Graphical User Interface

ICSBEP International Criticality Benchmark Experiment Project

CSEWG Cross-Section Evaluation Working Group

XML Extensible Markup Language

BIRT Business Intelligence and Reporting Tools

00 Object Oriented









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

DEVELOPMENT OF AN AUTOMATED TESTING SYSTEM FOR VERIFICATION AND
VALIDATION OF NUCLEAR DATA

By

Brian Scott Triplett

August 2008

Chair: Samim Anghaie
Major: Nuclear Engineering Sciences

Verification and validation of nuclear data is critical to the accuracy of both stochastic and

deterministic particle transport codes. In order to effectively test a set of nuclear data, the data

must be applied to a wide variety of transport problems. Performing this task in a timely,

efficient manner is tedious. The nuclear data team at Los Alamos National Laboratory in

collaboration with the University of Florida is developing a methodology to automate the process

of nuclear data verification and validation (V&V). This automated V&V process tests a number

of data libraries using well defined benchmark experiments, such as those in the International

Criticality Safety Benchmark Experiment Project (ICSBEP). The automation also serves to

reduce errors and increase efficiency. The process is implemented through an integrated set of

Python scripts. Material and geometry data are read from an existing medium or given directly

by the user to generate a benchmark experiment template file. The user specifies the choice of

benchmark templates, codes, and libraries to form a V&V project. The Python scripts generate

input decks for multiple transport codes from the templates, run and monitor individual jobs, and

parse the relevant output automatically. The output can then be used to generate reports directly

or can be stored into a database for later analysis. This methodology eases the burden on the user

by reducing the amount of time and effort required for obtaining and compiling calculation









results. The resource savings by using this automated methodology could potentially be an

enabling technology for more sophisticated data studies, such as nuclear data uncertainty

quantification. Once deployed, this tool will allow the nuclear data community to more

thoroughly test data libraries leading to higher fidelity data in the future.









CHAPTER 1
NUCLEAR DATA VERIFICATION AND VALIDATION OVERVIEW

Introduction

Nuclear and radiological engineering applications require a large degree of quality

assurance (QA) in order to guarantee safety and reliability. These applications may include

nuclear reactors, radiation diagnosis/treatment devices, and radiation detector systems. The

complexity of nuclear physics and the underlying nuclear interaction data can make the QA

process difficult. The interaction data for neutrons alone encompass thousands of isotopes over

many decades of energy and the nuclear physics involved makes generalizing particle behavior

extremely difficult. For this reason, the nuclear interaction data are often stored in tabular form,

typically in formatted files of particle energy versus the nuclear cross section. Several

organizations have produced such evaluated data libraries, (e.g., ENDF/B in the United States,1

JEFF in Europe, JENDL in Japan).3 These evaluated data libraries must be further processed

into an application specific library, (e.g., into ACE format for use in MCNP4 or NDI format for

PARTISAN. 5

In the course of producing an application library, it is desirable to perform detailed QA on

the data itself. This is done to verify that data were translated and processed correctly and to

validate that the library performs within expectations for a problem set of interest. Given the

immense phase space that nuclear interaction data encompass and the limited models available,

this can be an exceptionally burdensome task. Often the problem set against which the library is

validated is rather small due to time and/or financial constraints. It is for these reasons that the

nuclear data team at Los Alamos National Laboratory, in collaboration with University of

Florida, is developing a robust testing harness that will automate many of the validation tasks

associated with nuclear data libraries.









The goal of this project is to create a generalized modular testing harness that is publicly

available to test both nuclear transport applications and their underlying data libraries. One of

the initial aims of this project is the creation of a suite of benchmark problems for testing data

using both MCNP (a continuous energy Monte Carlo code)4 and PARTISN (a deterministic SN

code).5 A key feature of the test harness is a modular design that will allow the application to

other codes in the future. Additionally, such a system may enable other types of studies in the

future, (e.g., integration in regression testing for automated code V&V and automation of

uncertainty quantification).

Motivation

The inspiration for this project stems from a proprietary data testing system developed by

Steven van der Marck at the NRG laboratory in the Netherlands used for testing the ENDF/B-

VII.O library during its development.6 Similar to van der Marck's work, the nuclear data

verification and validation (NDVV) project aims to increase efficiency and accuracy when

testing nuclear data.

To demonstrate the necessity of an NDVV tool, we can consider the latest release of the

US data library, ENDF/B-VII.O. This library has undergone significant benchmarking to date

and is expected to be scrutinized further as it continues to be used for more diverse

applications.6'7 This latest release contains 14 sublibraries. The neutron interaction sublibrary

contains data for some 393 nuclides, each for typically dozens of reaction types and spanning

many orders of magnitude in particle energy. Performing V&V for this sublibrary alone can be

an immense task without taking into account this library's inter-relationship with the other

sublibraries such as the thermal scattering sublibrary, the photonuclear sublibrary, the radioactive

decay library and so on. It would take many man-hours to perform quality assurance on these

data libraries with little or no automation.









Another significant motivation for the NDVV project is the testing of new/improved

transport methods. Upon the development or deployment of a nuclear transport application, it is

common to perform some type of regression test in which the results of many well-defined

problems are compared to benchmarks. This is necessary to ensure that the application is

performing as intended and no developer errors (bugs) or platform specific issues are present. It

is the burden of the application developers to implement their own regression testing harness (if

any is created at all). However, if the NDVV system is used, it is possible that the developers

will need only to develop a way to "plug-in" their application to the NDVV system in order to

perform regression testing. If a system has already been incorporated into the NDVV system,

such as MCNP or PARTISN, the regression testing using the NDVV system is already present

and can be used by developers.

The growing need for nuclear data uncertainty quantification is another motivation for this

development. Tabular data libraries, such as ENDF/B, are usually the conglomeration of both

experiment and nuclear physics models, both of which have a degree of uncertainty. Many in the

nuclear data community desire to quantify how nuclear data uncertainty affects their complicated

transport application and how that uncertainty will propagate into further calculations.

A recent study at LANL utilized a Monte Carlo method to perturb the fission cross section

of 239Pu in order to evaluate the uncertainties in the fission cross section.12 This type of

undertaking involves a number of similar calculations to determine how slight perturbations in a

parameter of interest (e.g., a cross section) affect another parameter (e.g., kef). In this case, a one

dimensional PARTISN calculations were performed. If driven by an automated process such as

the NDVV system, the resulting time savings could potentially allow broader and more detailed

types of uncertainty studies.









Current Benchmarking Efforts

Benchmarking nuclear data libraries is not a novel concept. Upon every release of a new

data library, many institutions set out to perform some degree of V&V on the new data set.

Similarly whenever an institution or individual releases a new calculation methodology

(typically in the form of a computer program), the burden is on the developers to validate their

program's results versus experiments or at least verify that their results are consistent with

similar calculation methods. As already mentioned, the ENDF/B-VII.0 release was tested

extensively by S.C. van der Marck upon its release. Similar studies have also been performed

such as the shielding validation performed on the European data library release, JEFF 3.1 by W.

Haeck.8 Inter-library comparisons are also performed regularly (see Reference 9 for one such

example). Many transport applications will pick-and-choose different isotopic evaluations from

different libraries in order to obtain the best data possible. However, these studies tend to be

focused on a single goal and are the result of many hours of dedication by a number of

individuals. The intent of the NDVV automation system is that these studies will become more

commonplace.

Scope of this Thesis

It is important to point out that this work is not focused on the nuclear data analysis itself.

Those topics can be found in the many volumes of journal articles and reports published

regularly. This paper will put forth some example calculations performed using the NDVV

system which involve MCNP and a number of different libraries. It is not within the scope of

this paper to analyze the accuracy of these codes and libraries but merely to present a typical

application of the NDVV system.









CHAPTER 2
DEVELOPMENT AND IMPLEMENTATION OF THE NDVV SYSTEM

History

The current NDVV system is the result of over two years of development by a number of

students and staff at Los Alamos National Lab (LANL) and the University of Florida. The

automated verification and validation of nuclear data concept was pioneered by Dr. Morgan

White at LANL. As a member of the LANL nuclear data team and the Cross-Section Evaluation

Working Group (CSEWG), he believed that this type of automated testing harness would be a

valuable asset to the nuclear data community.

The software development was first started in the summer of 2006 by a team of four

students at LANL with backgrounds in nuclear engineering and computer software. The initial

focus was the generation of problem inputs and conversion to and from generalized problem

descriptions to an application specific format. During the summer of 2007, the application

control/execution and output processing was further developed. The development continued into

early 2008 to complete the prototype system. In its current state the NDVV process represents a

complete prototype and can be used for a fairly wide array of data studies.

Overview of the NDVV System

The NDVV testing harness automates the creation of input, handles transport code

execution, post-processes results, and generates standardized reports. All of these tasks are

accomplished using the Python programming language. Python was chosen because its object-

oriented capabilities make for robust, flexible coding. Python also features built-in capabilities

to facilitate using a number of key file formats selected for the NDVV system. Python also

offers tools for interacting with important computer resources such as databases and the system

shell.









Python was also chosen because of its portability. A key NDVV goal is to maximize

system independence whenever possible. The ability to run on many different types of systems

maximizes the computing resources available, enables more transport codes to be used, and more

easily facilitates new uses of the system. Python is an "interpreted" language and is generally

system independent." Provided a user has installed the appropriate version of the Python virtual

environment most Python applications will run on any type of system (e.g., Windows,

Macintosh, Linux), without need for modification or recompilation. In the NDVV system's

current implementation, Python is only required on one master system (typically the user's local

computer). Any computation-only nodes simply require command line interfaces to the transport

codes. The reporting tools are also only required on a single system. Minimizing the resource

requirements maximizes the systems flexibility while reducing the installation burden.

The first priority of the NDVV developers was to provide a full system demonstration

before widening the scope of the NDVV system. The NDVV prototype system currently

supports only the transport codes MCNP and one-dimensional (i.e., spherical geometry)

PARTISN. These two codes are different in their calculation approach and in the type of data

they require. MCNP is a stochastic, Monte Carlo based code that utilizes continuous energy

cross section data. PARTISN is a deterministic code that uses a number of discrete energy

groups for its cross section data. It is the view of the developers that these two codes provided an

appropriate demonstration the flexibility of the NDVV system for the initial prototype.

At this current point, the development for deterministic codes, such as PARTISN, requires

further study; specifically in regard to automating the discretization of both space and energy

without the introduction of error or the wasting of computer resources. There is a delicate,

sometimes artful, balance required when determining the balance between accuracy and speed









when performing deterministic calculations. Therefore, automating these types of decisions can

be challenging. Some studies have been performed in the development of the NDVV system but

have not progressed past the scoping phase yet. Due to this current limitation, PARTISN only

functions in ID geometries. Even a severely over-meshed ID problem does not overwhelm

computer resources on most modern systems. This allows the system to err on the side of finer

spatial meshing for ID cases without the worry of wasting computer resources.

Data and results are stored in many different forms throughout the NDVV process

including: Extensible Markup Language (XML), relational databases, and formatted text. The

templates used to store benchmark descriptions are formatted in XML because of the self-

documenting nature and the ease of Python to parse XML. The input and configuration files to

the NDVV system are formatted in XML for the same reasons. Code input and output formats

are dictated by the transport codes themselves but usually are given as semi-structured ASCII

text or binary files. Creation of these inputs and parsing of the subsequent output must be

programmed into the system on a code-by-code basis. Any addition of a new transport code will

require development of a new code module.

Analyzing the simulation outputs is possibly the most daunting and time consuming task in

the V&V process. If any automation exists in traditional methods, it is typically employed

during this stage. The transport code outputs are typically text or binary files that have evolved

over time and rarely present the information in all the ways desired. Often users must find a

means to conglomerate the results into another desirable medium, such as a spreadsheet. The

NDVV automated output parsing offers faster access to the parameters of interest (e.g., keff,

fluxes, run times, memory usage, etc). Once stored in the NDVV system they can then be









recombined and presented in a number of formats. Possible formats include formatted text,

tables, plots and relational databases.

One of the most important means of long-term storage is a relational database. By

combining results from multiple studies into an integrated database, the data may be reused and

mined in multiple ways. Using a relational database such as MySQL or PostgreSQL offers some

of the best capabilities of modem data storage techniques and tools. When properly

implemented, a database can store large amounts of data while eliminating redundancies and

allowing versatile searching and presentation of data.

The NDVV process has the ability to connect to a database and insert results from the

NDVV runs. Currently the NDVV system includes an interface to store results in a PostgreSQL

database. As the project expands it is expected that more types of databases will be included.

Another benefit of relational databases is the third party tools that support them. These

tools are not tied to any specific types of data and are available to any users of the database.

Once data are in a database, an existing tool can be used to search and present data without the

need for specific development by the NDVV developers or the user.

The following discussion provides greater detail regarding the three major steps in the

NDVV process: Input Generation, Code Execution, and Output Processing.

Input Generation

In order to validate a code/library one must have some a-priori results for comparison.

This can be an analytical model or an actual experiment. Analytical models tend to be limited in

scope for typical nuclear physics applications. Therefore, for most validation problems, an

experiment forms the basis by which accuracy is determined. For a code-to-code or library-to-

library comparison an experimental or analytical basis is not needed but the accuracy can only be

inferred based on the agreement between results.









A common example of problems of interest for data benchmarking are those experiments

included in the International Criticality Safety Benchmark Experiment Project (ICSBEP)

Handbookl2 and those in the ENDF-202 report by the Cross Section Evaluation Working Group

(CSEWG).13 These experiments are well-defined criticality and shielding experiments that cover

a wide range of scenarios and offer an excellent base with which to test the NDVV system. For

this reason, the examples selected for this paper are from the ICSBEP Handbook.

Using the benchmark experimental data as a starting point, the user will generate a

template that describes each problem of interest for data benchmarking. These template files

store the necessary information needed to create transport code inputs for the various codes used

by the NDVV system.

The templates are formatted in XML. XML is a portable, hierarchical markup language

that is used to store data. It is commonly found in many web-based applications because of its

system independent nature. XML also has strong software support from the commercial sector

since it aides many businesses in making their data portable.

In the current system, it is possible to "initialize" a template file from a pre-existing MCNP

input file. One reason for this is because many benchmark experiments in the ICSBEP have an

MCNP input attached as an appendix. Although the editors of the ICSBEP do not guarantee the

accuracy of those inputs, they can often serve as a good starting point. This feature has been

added as a convenience so that not every template file must be generated from scratch.

A full template file is demonstrated in Figure 2-1. This template describes the Jezebel

critical assembly used at LANL for nuclear physics research.

The template generation phase typically occurs once for each problem. Once a template

has been created and its quality been assured, it will typically be stored in a secure location









and/or placed under revision control. If a repository of templates is already available, then the

user only needs to specify the location of those templates in the input file prior to the NDVV run.

Once a template file is complete, the NDVV conversion scripts read the template and

generate inputs for the transport codes of interest. For every transport code a geometry and

material specification is required as a minimum. There is also typically code specific

information required, (e.g., KCODE cards in MCNP). All of this information is currently stored

in code specific sections of the XML template file that describes the experiment. A goal of this

system is to specify as much of this information as possible in a code independent manner to

eliminate redundancies and ensure input consistency.

There may also be computer dependent information required in the transport code inputs

(file locations, specific system directives, etc). This is not stored in the problem template but

given at run time by the user (either directly or stored in machine description files). This is to

ensure that a suite of experimental benchmark templates will remain portable and can be easily

transmitted among the nuclear data community.

When the inputs are first generated, any computer specific input requirements (e.g., file

locations) are filled with placeholders. These inputs are labeled as "code input template" in the

Figure 2-2. When the run phase begins, these placeholders are filled in just before the input is

sent to a specific machine. This ensures that the input has been properly customized for the

given computer. Figure 2-2 describes this process and how transport code inputs are generated

by the NDVV system.

This input generation approach yields a number of benefits in regard to quality assurance.

The first is the reduction of input files to be reviewed. The XML templates are tied to the

problem of interest and not to the code/library to be used. Therefore, this approach only requires









one file per benchmark experiment instead of one for each code/library combination. This

reduction of files simplifies the benchmark data review process.

This input model also increases the consistency by which inputs are generated. Often user

preferences can translate into small differences between results. This typically manifests itself in

a few common ways: specifying of a certain variable versus using the code default, differing

floating point precision, breaking up isotopes versus using natural evaluations. These user

decisions do not necessarily lead to inaccurate results but can sometimes lead to confusion when

results differ slightly. The input files generated by the NDVV process are linked to the

benchmark templates and therefore maintain consistency with the original template from run to

run.

Arguably the most probable error during problem setup is a transcription error. An user-

introduced typographical error in the input file can change the material or geometry specification

and lead to erroneous results. Often the typographical errors have a small effect and are not

immediately noticed until detailed analysis is performed. Given that the initial template has been

fully vetted, automated generation of input decks are immune to typographical errors (or at least

errors are propagated consistently).

Code Execution

When the user is ready to execute a V&V calculation, a NDVV queuing system controls

the execution of the transport codes. The user specifies the computers to use and the transport

codes and data libraries available on each computer. The NDVV control process determines if

computers are available and sends the input file(s) and execution instructions as appropriate.

The specification of the V&V job to run and the resources available are given in a NDVV

user input file. This file tells the NDVV process a number of specific parameters required to

control and execute the V&V calculations. Like the benchmark templates, this input file is in









XML format. Although XML offers a moderate amount of readability to the user, it is the intent

of the developers to eventually drive the user input setup process with a graphical user interface

(GUI). This not only facilitates quick straightforward input creation but also can be used as a

means to control/limit the choices a user can make and reduce errors. For example, objects such

as drop-down menus and radio buttons can limit user's choices to a few safe options; whereas in

a XML file the user is free to type anything he or she wants leading to potential typographical

errors.

The first requirement in the user input file is the choice of benchmark templates and their

location. It is the developers' expectation that a template repository or collection of templates

will be located on a system available to the user (either locally or remotely). In the input file,

the user will specify their desired subset of templates (or the entire repository if applicable) and

the templates' location.

The user must also furnish his or her choice of transport codes. This may be one or more

codes to be included in the run. As previously noted, the current implementation of the NDVV

system only includes support for MCNP and PARTISN (ID) with more code modules to be

added in the future. The location of the code executables and any specific command line

requirements must also be included in the user input file.

For each code specified, one or more libraries must be specified. The most pertinent

library information required is the location of these libraries on the computer systems to be used.

Finally, the computer systems themselves may require some additional information from

the user. This typically includes connection parameters (if it is a remote machine), queuing

limits, and system type. The underlying assumption in this user input preparation methodology









is that the user knows his or her systems) best and it the most qualified to determine how and

where system resources should be used.

Once all the necessary information has been input, the NDVV control process handles the

sending of NDVV jobs to each of the available computers via its queuing system. It is possible

to have different codes available on different computers. For example the user may specify two

computers and one is only able to run code X and has library A but the other computer has code

X and Y and only library B. The NDVV control process will handle the distribution of tasks

based on the capability of each computer to perform a calculation.

A NDVV job consists of a unique combination of template, code, and library. NDVV runs

can consist of few or many jobs. Once the control process determines that a given job is ready to

run on a specific computer, it sends the required input(s) and a set of system commands to that

computer. These commands are the instructions to execute the code on that computer and are

contained in a system-dependent shell script. A command is issued to run the shell script and the

connection is closed with that computer. The NDVV process periodically reconnects with each

computer and assesses the status of each job. When a job is completed, any relevant outputs are

copied back to the user's computer and placed in the NDVV project directory. The user can

provide the specifics regarding the maximum number of jobs that can run on each computer and

the frequency at which the jobs' status is checked in the NDVV user input file.

It is important to note that all of the NDVV computations can be run on a single local

computer or make use of simple or complex cluster computing. In striving for maximizing

efficiency while maintaining as much system independence as possible, the simple queuing

presented here provides minimal facilities that will work anywhere. If an alternate queuing

system is available and desired, job control can be ceded to that system simply by submitting all









jobs to that system. Examples of these alternative queuing systems include the LSF and MOAB

systems at various national labs. It is anticipated that for large verification and validation (V&V)

jobs more powerful computer resources will be used in this manner. Figure 2-3 gives an

overview of the execution phase of the NDVV system.

Output Processing

Upon completion of each code execution, the NDVV control system collects the relevant

transport code output files back to the master computer. After the output files are collected, post-

processing begins. This is done on the master host because of the availability of more advanced

tools, (e.g., Python) which are needed to perform the post processing. The files are parsed for

pertinent information (keff, flux distribution, reaction or leakage rates, etc) via the NDVV post-

processing system. The data may be stored in a database, written to a formatted file, or used to

generate plots and reports.

The NDVV scripts search the outputs from the various transport codes and gather various

important parameters. The data from each code are stored in a custom NDVV data structure so

that it can be accessed in a standardized manner. Once these results have been gathered and

stored in the NDVV structure, the user may choose any of the supported NDVV formats for

displaying/storing results. The NDVV process always produces some limited textual summary

information from each run but also contains more elegant post-processing options with more to

come as development continues.

Plotting with gnuplot

The first post-processing option is the gnuplot option. This option searches through the

NDVV result data and creates the files necessary to plot those results using the open source

plotting software, gnuplot.15 The gnuplot program uses these files as the basis to create

postscript plots based of the NDVV results. This saves the user the time and effort required to









create plots of their results. The user is also free to modify the gnuplot script files in order to

customize the plot appearance if the defaults are insufficient or undesirable. Figure 2-4 give an

overview of the output processing phase of the NDVV system.

Database storage

In addition to plotting results with gnuplot, NDVV results can be stored into a PostgreSQL

database for later querying and retrieval by the user. The user is required to have access to a

PostgreSQL database with the proper table structure. Once the tables are set up, the automated

NDVV system inserts results as rows in the appropriate tables.

The algorithm that inserts NDVV results is based on a specific database design. The

design of this database is based on the selection of variables the NDVV developers felt would be

important for V&V. This design will continue to evolve and improve with the NDVV system.

The NDVV table layout in the NDVV database follows the principles of database

normalization. By following these guidelines, the database will be free of redundancies and less

prone to data anomalies. An overview of this NDVV database design, including table names,

keys, and columns, is shown in Figure 2-6.

The NDVV database design, like the rest of the current NDVV system, is a working

prototype that will continue to evolve with time. As more input and requirements are included

the database design will mature and become more extensive. Since the SQL directives generated

by the NDVV code are intimately tied to the design of the database, it is expected that a database

creation tool will also be packaged with the NDVV system. This will ensure the compatibility of

the NDVV database storage mechanisms and the database structure. This tool will allow users to

create the required tables for storing their own NDVV results. Currently, a simple SQL script

exists for the initial creation of the database tables. This method is the one presently used by the

NDVV developers.









Once data are in a database, a wealth of tools exists to query and present data from the

database. On such tool is the Business Intelligence and Reporting Tools (BIRT) program. BIRT

is an open source reporting system that can generate standardized reports based on a database

source.14 BIRT uses a "What-You-See-Is-What-You-Get" interface so that users know exactly

how their reports will appear. BIRT is able to be deployed on a web server so that anyone with

internet access can view the reports created. BIRT also allows reports to be exported to

common formats such as DOC and PDF. BIRT also includes built in methods for creating plots

and tables and is able to perform calculations using JavaScript.

An example BIRT report is shown in Figure 2-6. This report displays the benchmark keff

along with the keff results from MCNP and PARTISN. The calculated/experimental (C/E) values

are performed internally by BIRT.

The results given in this example are from previous calculations performed at LANL and

were not obtained via an NDVV run. However, once full PARTISN support is implemented in

the NDVV system, they could easily have been the result of the NDVV process.

It is important to point out that BIRT is not part of the NDVV system. It is merely an

example of how a third-party tool can be used to present results once they are contained in a

relational database. Users are free to use the display mechanism of their choice (or none at all)

when presenting their own results. The important point from this demonstration is the benefits of

using a relational database. Once data are stored in a database, such as PostgreSQL, an entire

world of software tools exist to search and display those results as the user requires.





































yards>
defir


L definite
definitic
Ll name>
Ll number;
aerial>
ometry>
portance>
L definite
yards>

D cards>
ace define
face nam,
face numn
face typi
face parE
face defir
ce cards>

al cards>
trial defir


Example XML template file used to specify benchmark problems. This file

describes PU-MET-FAST-001 from the ICSBEP handbook, otherwise known as the

Jezebel critical assembly. XML was truncated for brevity.


Figure 2-1.












MCNP
Input File


p python

- r
C ^



Template File

---~----

OR python
User Input:
Codes and Data to Use C In' erri n 'T.ft,:
I ^ n


Graphical User
Interface


Code Input Code Input
Template 1 Template 2


Template
Generation


Setup


000


Code Input
Template N


- r -



User Input:
Machine Specifics


0 python
C., r.. ', it


Run-time


Code Input 1 Code Input 2


Code Input N
Code Input N


Figure 2-2. Input generation process for the NDVV system. Divided up into template
generation phase, setup phase, and run-time phase. The template generation phase
typically occurs once, in order to generate a template repository. The setup and run-
time phases occur for every NDVV run.


DecnmacTK
Experimental Data


User Input


_j_ -
























culput Fi]e{s)


ji. r,[it Iir i i
^i-----^ '


Master
Computer


OuLput File(s)


Run Script


Inpul File(s)


Remote Computer



















Cluster Computers
















Remote Computer


Figure 2-3. Job execution in the NDVV system. A master computer sends input file(s) and
scripted instructions to each computer. The results are copied back to the master
computer when execution completes. The master computer can also act as a
computational node.













Code Output 2


python'
C'it


PostgreSQL


MySQL


'
AdC,


Business Intelligrece
Reporting Tools (BIRT)

Figure 2-4. Output processing in the NDVV system. Output files are parsed for information and
the resulting data are stored in the desired format. Databases such as PostgreSQL
and MySQL can be used with third party tools such as BIRT for further data
processing/presentation.


Formatted
Text File


Code Output N


Code Output 1





mncrp_keffresults

PK,FK1 lob id SHORT

run_type TEXT(256)
run date DATETIME
runtime DATETIME
total_cp _time DOUBLE
keff DOUBLE
std dev DOUBLE
promptremoval_lifetime DOUBLE
avg_fission_energy DOUBLE
avgfission_ lhsr,.' DOUBLE
pct_thermalfissions DOUBLE
pct_intermediatefissions DOUBLE
pct_fast_fissions DOUBLE


Figure 2-5. NDVV database design layout. Tables are shown with their linking primary and
foreign keys. Variable types are specified with PostgreSQL types.


code

code id COUNTER

code TEXT(256)
version TEXTi2tfi]
load date DATETIME
platform TEXT(256)


benchmark

PK problem name TEXT(256) PK

description LONGTEXT
category TEXT(256)


1.*

benchmark keff results

PK,FK1 problem name TEXT(256)

source TEXT(256)
version TEXT(256)
date DATETIME
keff SINGLE
uncertainty SINGLE


job
PK job id COUNTER

FK1 problem_name TEXT(256)
FK2 code id SHORT
FK3 library_name TEXT(256)


paisn_kefftresults

PK,FK1 job id SHORT

run_type TEXT(256)
rundate DATETIME
run time DATETIME
total_cpu_time DOUBLE
keff DOUBLE
lambda DOUBLE
flux_change DOUBLE
fission_change DOUBLE


PK,FK1 library name TEXT{256)
PK zaid SHORT













BIRT Report Viewer


Showing page 1 of 1


1 tVt Coto page: [


Alumilnum-Reflected Assemblies



HEU-MET-FAST-022 D mn-eflectedHEU Sphere, VNEF 1000000021 0.99540.0006 0.99540.0022 1.0033000001 1.00330.0021
IEU-MET-FAST-006 Duralumin-Reflecled IEU Sphere 36 wl.'). VNIIEF 1.000000023 0.9931 00006 0.9931 0024 0.9962: 000001 0.996200023
PU-MET-FAST-009 Aluminum-ReflectedPu.94.;' Sphere. Conl A niblv 1000000027 1.004200006 1.00420.0028 1.0161000001 1.01610.0027

Bare Metal Assemblies

C p(0

HEU-MET-FAST-001a Oodiva, Umeflected Sphere of HEU, Simple Sphere 1.0000*00010 0.9988 00006 0.99880.0012 10009 0.00001 1000900010
representation
HEU-MET-FAST-001 b Godiva, Unreflected Sphere of HEU, Nested Spherical Shell 1 00000.0010 0.99930.0006 0.99930.0011 1.0009 000001 10 009 00010
repren.ilUtion
HEU-MET-FAST-008 Bare HEU Sphere, VNfITF, 3D model 0.99890016 0.995300006 0.996400017 0.9915 0.0001 099260.0016
HEU-MET-FAST-015 BareHEU Cylinder, VNIIT 0.99960.0017 0.9937O.0006 0.99410.0018 0.98960.00001 0.990000017
HEU-MET-FAST-018 Simplified Bare HEU Sphere, VNIIEF 1000000016 0.9995 00006 0.99950.0017 1.008000001 10008.0016
IEU-MET-FAST-003 Bare IEU Sphere (36 wt.%), VNIIEF 1l000000017 1 0032 00006 1.00320.0018 1006100000X1 1l00610.0017
PU-MET-FAST-001 Jezbe-Pu(4 5%). Bare Sphere o Pu-239with 45% 1.00M 0.0020 10004 00006 1.0040.0021 0.9989000001 0.998900020
Pu-240
PU-MET-FAST-002 Jezebel-Pu (20%), Bare Sphere of Pu-239 with 20% Pu-240 1 000 0.0020 1.0018 00006 1.00180.0021 0.9993 000001 0.9993 00020
PU-MET-FAST-022 Simplified Plulonium i98 .) Bare Sphere, VNIIEF 1.00000021 1.000100006 100010.0022 0.9978000001 0.99780.0021
U233-MET-FAST-001 Jezebel-23,BareSphereofU-233 10000.0010 0.998000006 0.99800.0012 1IJ002000001 1020002 10

Beryllhm and Berylllum-Oxkldde-Reflected Assemblies



233U-ET-FAST-005.I 0.805" Be-Reflecled Sphere of U-233. Panel Asembl 1 0000= 00030 0.9968 0.0006 0 9968 0 03 0.9998=000001 0.9998= 00030
233U-MET-FAST-0052 1.652" Be-Reflected Sphere of U-233, Planet Assembly 1.00000030 0.996800007 0.99680.0031 1.0015 0O0001 1.00150.0030
HEU-MET-FAST-009.1 Be-Reflected HEU (-89.6) Sphere, VNITF 0.99920.0015 0.9995 00006 1.0003 0.0016 0.9933 0001 0.9941 00015
HEU-IET-FAST-009 BeO-Reflc-lcd HEU '-89.6' Spbhcr. VNTITF 0.9992 0.0015 0.9962 r 0X006 -- i-1 16 i 0.9923 OfX0001 0.9931 0001
PU-MET-FAST-018 Be-Reflected Pu i94.79) Sphere. Planel Aeiembl I o00o000030 I 0016 0jX006 0I 016J00031 I o034 000001 I 0034 000 0 ,
Done


Figure 2-6. Example BIRT report. Table can be customized to show whatever the user desires. This example shows keffresults from
MCNP and PARTISN with the keff values color coded based on their agreement with the benchmark









CHAPTER 3
DEMONSTRATION OF THE NDVV SYSTEM

Example Problem Specification

To demonstrate the NDVV system, a set of thirteen experiments from the ICSBEP

Handbook have been selected for evaluation using MCNP5-1.40 with three different data

libraries: JEFF31, ACTI, and ENDF70. In all, 39 simulations were performed using the

automated NDVV system. More benchmarks could have easily been included; however, the

author felt this subset provided an adequate demonstration without overwhelming readers with

superfluous results.

Prior to the selection of the thirteen benchmarks for this demonstration, a much larger

repository of benchmark templates were constructed. The MCNP team at LANL retains a set of

approximately 90 MCNP input files based on ICSBEP experiments. NDVV initialization tools

were used to convert those 90 MCNP inputs into the beginnings of a template library. Those

templates were then reviewed and compared with the official ICSBEP specifications. Any errors

were corrected and additional benchmark data was added. Once the review was complete, each

benchmark was placed into an internal LANL document control system so that changes require

official review. This is expected to be the model most users will follow when creating their own

template library.

The thirteen ICSBEP criticality experiments for this example were chosen to demonstrate

the breadth of the problem type. They feature a number of fuel and reflector types and span the

full spectrum of neutron energy. They were also chosen because they were included in the

ENDF/B-VII.0 benchmarking performed by van der Marck. Therefore, the MCNP results have

some basis for comparison. These ICSBEP benchmarks are described in detail in Table 3-1.









For this example MCNP5-1.40 was used for all calculations. It is typical for high-fidelity

library comparisons to prefer the Monte Carlo method due to its "exact" treatment of geometry

and cross sections.9'20 Also the limited support for PARTISN in the current NDVV

implementation would constrain the experiments to spherical geometry only. It was determined

that performing MCNP calculations with a wide variety of problems and libraries would

represent a reasonable example of how the NDVV system would be applied by end-users.

The three libraries included are common evaluations used in many nuclear calculations.

The JEFF31 MCNP library is based on the JEFF-3.1 nuclear data library currently used in

Europe. This library was created by using cross section data from the JEFF-3.1 nuclear data

library and processing it with the NJOY code into the ACE format used by MCNP.16

The ACTI library is a combination of 41 newer evaluations and the ENDF66 MCNP data

library. Most of the ACTI evaluations were eventually included in the US library ENDF/B-

VI.8.18 This library was used with the thermal scattering data from a 2002 evaluation as detailed

in Reference 19.

The ENDF70 MCNP library is based on the latest release of the US ENDF/B series and

represents the current standard for nuclear data in the US. This MCNP library was created using

ENDF/B-VII.0 data and includes the latest 2007 thermal scattering data. The processing of

ENDF70 is detailed in Reference 7.

It is important to emphasize that the intent of this example is to demonstrate the capability

of the NDVV system and not for the purposes of validating the libraries included in the example.

Many works have already been published with the focus of data library accuracy. Such

validation studies can be found in References 9 and 20.









Execution of the NDVV system

A simple NDVV input file was built to specify the inclusion of the aforementioned

templates. This file also included the location of the MCNP5-1.40 executable and the locations

of the three libraries on the available computer resources. Using this input specification and the

high-performance computer resources at LANL, this NDVV project was executed to demonstrate

the prototype system.

The time required to complete this run is not of first importance. Users' computing

resources can vary greatly. In this case, the high-performance cluster computers at LANL could

execute all 39 simulations in parallel and greatly reduce the time required. The important point

from this exercise was the limited need for user involvement offered from the NDVV system.

The creation of transport inputs was handled entirely by the NDVV system as well as the

execution of the 39 MCNP calculations. None of these steps required any user intervention.

Following the NDVV run phase, the NDVV output parsing capability was executed. This

placed the MCNP results into a special NDVV data structure and produced a summary table

similar to that in Table 3-2. As an added confirmation, the results for the ENDF70 libraries were

compared to those obtained by van der Marck in Reference 6 (outside the NDVV system). The

values of keff obtained by the NDVV calculations were found to be consistent with van der

Marck's results for ENDF70 (ENDF/B-VII.0).

Storage and Displaying of Results

The database storage option was executed to place the example problem results into a

PostgreSQL database. This database was created on a server at LANL prior to the NDVV run

using an SQL script. Following the storage in the database, a custom BIRT view was created to

present the keffresults from each library and the benchmark keff. The resulting report is shown in

Figure 3-1. This view is entirely customizable and can display any parameters the user desires.

34









An important benefit of this report generation methodology is its ability to automatically

update based on the results in the NDVV database. If more runs are performed in the future and

the results are inserted into the same database, then the BIRT view will automatically update to

show those new results. For example, a NDVV run could be performed using these same

templates with another data library, such as the Japanese library, JENDL. After the results are

stored in the database the BIRT view will automatically update to show those new results. This

feature is not found with static files such as Excel plots or even the gnuplot files created by

NDVV.

This example BIRT report is hosted on a server at LANL and was viewed through a web-

browser. This demonstrates how a user could execute NDVV runs, store their results to a

database server, and then create a custom view of the results to be accessed via a website. This

would allow members of the nuclear data community to collaborate and compare results easily.

The gnuplot option was also run to create the plot featured in Figure 3-2. All the necessary

files for gnuplot were automatically created by the NDVV system based on the MCNP results.












NDW Results Table


Font color C/E within 1 std deviation from unity
key:
C/E between 1 and 2 std deviations away from unity
/E between 2 and 3 std deviations away from unity
C/E more than 3 std deviations away from unity


S Nae B r e


HEU-MET-FAST-022


Duralumin Reflected HEU Sphere (90% U-235)
1.0000 0.0019 Simplified Model
VNIIEF Facilitv

ENDF70 0.9972 0.0006 0.9972 0.0020
ACTI 0.9921 0.0006 0.9921 0.0020
JEFF31 0.9939 0.0006 0.9939 0.0020


[CSBE1 Nm Benchmark,1.. I ption.


HEU-MET-FAST-028


Natural Uranium Reflected High-Enriched Uranium Sphere
1.0000 0.0030 Flattop Assembly
See also CSEWG-F22

ENDF70 1.0028 -0.0007 1.0028 -0.0031
ACTI 1.00264 0.0006 1.00264 0.0031
JEFF31 1.0004 0.0006 1.0004 0.0031


* Nm Benc r D


HEU-SOL-THERM-032


1.0015 0.0026 48" Unreflected Sphere of Uranyl Nitrate (93.2 wt% U-235) Solution
1.15 See also ORNL-10 and CSEWG T-5


ENDF70 0.9996 0.0003 0.9981 0.0026
ACTI 0.9988 + 0.0003 0.9973 + 0.0026
JEFF31 0.9994 + 0.0003 0.9979 + 0.0026


Figure 3-1. BIRT report generated from example run. Only first three benchmarks are shown for brevity. Data in BIRT view is
obtained from PostgreSQL database were result data is stored. Once a view has been defined the data is automatically
updated as results change.










1.010

1.008

1.006

1.004

1.002

1.000

0.998

0.996
0.994

0.994
0.992


8 ,
.
k^


& 000-
'0


Li
W X






I I F I I I I I II


'%A

;, ,,
*'^^ Cr


x~
4?


Benchmark +/- o JEFF31 o
ACTI K ENDF70

Figure 3-2. Example problem results from gnuplot. The Monte Carlo uncertainty is not shown on the plot for readability and because
it is small relative to the benchmark uncertainty.


Z
X\:









Table 3-1. ICSBEP experiments used in example NDVV run


ICSBEP Name


HEU-MET-FAST-022


HEU-MET-FAST-028


HEU-SOL-THERM-032


IEU-MET-FAST-002


PU-MET-FAST-002

PU-MET-FAST-005


PU-MET-FAST-006

U233-MET-FAST-001

U233-SOL-THERM-001.1


U233-SOL-THERM-001.2


U233-SOL-THERM-001.3


U233-SOL-THERM-001.4


U233-SOL-THERM-001.5


Description


Duralumin Reflected HEU Sphere (90% U-235)
Simplified Model

Natural Uranium Reflected High-Enriched Uranium Sphere
Flattop Assembly

48" Unreflected Sphere of Uranyl Nitrate (93.2 wt% U-235)
Solution

Jemima Idealized Natural Uranium reflected stack of Natural
Uranium and HEU plates

Jezebel Bare sphere of Pu-239 with 20% Pu-240

Tungsten Reflected Plutonium Sphere with 4.9% Pu-240
Planet Assembly

Natural Uranium Reflected Pu Sphere Flattop Assembly

Jezebel-233 Bare Sphere of U-233

Unreflected 27.24" Sphere ofU-233 Nitrate Solution
Experiment 1 1.0226 g/1

Unreflected 27.24" Sphere ofU-233 Nitrate Solution
Experiment 2 1.0253 g/1

Unreflected 27.24" Sphere ofU-233 Nitrate Solution
Experiment 3 1.0274 g/1

Unreflected 27.24" Sphere ofU-233 Nitrate Solution
Experiment 4 1.0275 g/1

Unreflected 27.24" Sphere ofU-233 Nitrate Solution
Experiment 5 1.0286 g/1









Table 3-2. Example results summary. Monte Carlo uncertainty not shown for clarity since it is significantly smaller than the
benchmark uncertainty.
Prnhlpm Ikeff


U233-SOL-THERM-001.1

U233-SOL-THERM-001.2

U233-SOL-THERM-001.3

U233-SOL-THERM-001.4

U233-SOL-THERM-001.5

HEU-MET-FAST-022

HEU-MET-FAST-028

HEU-SOL-THERM-032

IEU-MET-FAST-002

PU-MET-FAST-002

PU-MET-FAST-005

PU-MET-FAST-006

U233-MET-FAST-001


Benchmark
1.0000 +/- 0.0031

1.0005 +/- 0.0033

1.0006 +/- 0.0033

0.9998 +/- 0.0033

0.9999 +/- 0.0033

1.0000 +/- 0.0019

1.0000 +/- 0.0030

1.0015 +/- 0.0026

1.0000 +/- 0.0030

1.0000 +/- 0.0020

1.0000 +/- 0.0013

1.0000 +/- 0.0030

1.0000 +/- 0.0010


JEFF31 ACTI ENDF70


0.99875

0.99881

0.99765

0.99812

0.99744

0.99390

1.00043

0.99936

0.99215

1.00351

1.00361

1.00273

1.00414


0.99810

0.99929

0.99794

0.99749

0.99809

0.99206

1.00257

0.99879

1.00347

0.99805

1.00804

1.00355

0.99406


1.00080

1.00164

1.00127

1.00067

1.00002

0.99724

1.00280

0.99956

0.99914

1.00146

1.00907

1.00145

0.99854









CHAPTER 4
FUTURE WORK AND CONCLUSIONS

Future Work

Although the current NDVV system is complete and functional, most of the NDVV code is

in the prototype phase. The development started from scratch and the software design has

evolved as goals and requirements were still being determined. In order to ensure that the

system remains modular, expandable, and maintainable, the software design should be

reevaluated in a number of ways.

The first area that warrants revisiting is the XML benchmark template. The original

specification was based mainly on MCNP because it allows for very flexible geometry and

material definitions. Geometry issues arise with deterministic codes because they typically only

allow for fixed coordinate systems (Cartesian, cylindrical, etc). Special geometry processing

tools will be required to translate the MCNP surface based geometry into the mesh structure

required for many deterministic codes.

The automatic determination of mesh fineness is another significant issue still to be solved.

It is difficult for an automated system to determine the proper mesh-sizing for every region of a

problem. It is possible to set rules based on particle mean free path in each cell; however, this

would require an intimate knowledge of the problem prior to the NDVV run. This data would

then have to be included in some manner in the benchmark template.

Until these issues are resolved, there will be significant limitations on the NDVV systems

ability to utilize deterministic transport codes.

Another critical aspect of the NDVV system that warrants review is the controller/machine

interface, particularly the manner in which commands are sent to remote machines. The need to

send more complex commands will likely become more apparent as the system involves. Up to









this point the developers have relied on a minimal set of scripted system commands existing on

each type of system. This can greatly limit the types of command that can be sent to each system.

Given that Python is a system-independent language, it is possible that keeping the commands

internal to Python may help meet the requirements of the NDVV system in a more robust,

straightforward way. By using a wholly Python client-server system to send commands to

remote machines, the developers can easily guarantee the same behavior on a wide variety of

systems. The only drawback to this approach is that the minimum version of Python will be

required on each system the user intends to use; however, for most users, this would not be a

prohibitive requirement.

Following these critical maintenance issues, there are a number of prospective modules

that could be plugged into the NDVV system. A full PARTISN module is likely to be completed

first due to the already existing development. When the deterministic code issues are solved, full

support for PARTISN could be implemented. This would allow the NDVV system to run three-

dimensional stochastic and deterministic transport calculations.

Another major expansion could be the inclusion of more post-processing modules. Since

the NDVV results data are stored in a standardized data structure, it would be possible to display

theses results in more formats that those currently available. An example of this could be an

auto-generated HTML webpage for quick display of the results on a web server. There are very

few limitations to how the data could be displayed from this system. The developers expect the

nuclear data community will serve to offer insights into the most convenient ways to display

V&V results.

A GUI to drive the user input creation process is another potential addition to the NDVV

system. A GUI would better facilitate the selection of templates, codes, and system resources.









Finally, more database modules could be developed for the NDVV system. It not expected

of every user to use exclusively PostgreSQL. The current availability of Python modules for

interfacing with PostgreSQL was the main factor for its selection during this prototype phase. It

is likely that Python modules for interfacing with other databases, such as MySQL will become

available soon. A structure already exists in the NDVV software design to allow for more

database formats should they be desired.

Conclusions

The goal of the NDVV project is to produce a tool that allows members of the nuclear data

community to more easily and quickly test data libraries and assess their applicability for a

specific problem. Such a tool may be used to help justify use of a specific code and library for a

specific application. Eventually, it may be also be used for formal uncertainty quantification.

This should lead to a better understanding of data sensitivities, which can then feed back into

improved evaluated data libraries.

The current implementation of the NDVV system represents a full prototype version of the

aforementioned V&V tool. A format has been defined to describe benchmark experiments for

use in the NDVV system. The processes are in place to convert those templates into full input

files for MCNP and one-dimensional PARTISN. A control system is able to automatically

manage and monitor the execution of those transport codes on a wide variety of systems. The

NDVV system can parse both PARTISN and MCNP outputs for storage into standardized

NDVV data structures. NDVV tools can then insert those results in a PostgreSQL database or

create gnuplot plots based on the results.













APPENDIX

USER INPUT FILE FOR EXAMPLE PROBLEM




rersion="1.0" encoding="UTF-8"?>


ieck frequency>0.5

emplates>