ML20133A362
| ML20133A362 | |
| Person / Time | |
|---|---|
| Site: | 05000447 |
| Issue date: | 04/09/1985 |
| From: | Hankins D GENERAL ELECTRIC CO. |
| To: | |
| Shared Package | |
| ML20133A243 | List: |
| References | |
| FOIA-84-175, FOIA-84-A-66 NUDOCS 8507200069 | |
| Download: ML20133A362 (9) | |
Text
.. -
(f) l AFFIDAVIT OF DEBORAH A. HANKINS I, Deborah A. Hankins, being first duly sworn under oath, depose and state as follows:
- 1. I am currently a principal engineer in the Nuclear Services Products Department at the General Electric Company.
As the Engineering Program Manager for severe accident issues, I was responsible for the performance and technical content of the GESSAR II PRA and related documents.
- 2. I have reviewed in detail the ten categories of information which The Union of Concerned Scientists contends may be released to the public without causing the General Electric Company any competitive harm. Contrary to the affidavit of Steven C. Sholly, the information in the ten categories would be of great commercial value to GE's competitors and if released to the public it would likely result in substantial competitive injury to General Electric.
- 3. As'I understand Mr. Sholly's Declaration, it is Mr. Sholly's view that the information he denominates as
" quantitative results" would not be of commercial value because it could not be submitted to NRC to satisfy regulatory require-ments without the underlying analyses and data and for similar reasons could not be directly incorporated into a competitor's probabilistic risk assessment.
8507200069 050503 PDR FOIA CURRAN 84-A-66 PDR
- 4. The commercial value of the GESSAR II PRA lies in three areas. First, the entire PRA is, of course, of great commercial value and could be modified by competitors in performing full scale PRA's for operating reactors at little or no cost. Second, the commercial value of the individual analyses in the GESSAR II PRA is in many ways greater than the value of the entire comprehensive PRA. Utilities and industry groups often contract for PRA-type analyses which require only the use of the separate analyses within the PRA. For example, GE has prepared PRA-type analyses for utilities which required only the construction and quantification of dominant accident sequences. Third, the quantitative and qualitative results requested by UCS disclose many of the underlying assumptions, quantitative inputs and modelling techniques used by GE in preparing the GESSAR II PRA. These results would be extremely valuable to competitors allowing them to gain the benefit of GE's design experience, data base, research and assumptions, among other things, at no cost. Moreover, simply knowing the particular results of the analyses prepared by GE would greatly 6
assist a competitor. By comparing GE's results with those obtained in other PRA's, competitors would be able, with little expenditure of time or resources, to determine the basis for the differences and thus understand GE's assumptions and methodology.
- 5. The commerical value of the GESSAR II PRA including the value of the results in the ten categories of information
I is enhanced by the depth of experience of the General Electric Company in the design and manufacture of boiling water reactors.
GE is preeminent in the field of boiling water reactors and is able to bring a more intimate knowledge of the reactor design and capabilities to the preparation of the analyses comprising the GESSAR II PRA than any of its competitors. In addition, GE has developed its own proprietary computer codes, compiled and maintained a proprietary data base and engaged in extensive experiments, all at substantial cost, in order to ensure the accuracy and completeness of its analyses. Com-petitors of GE are, of course, aware of GE's technological achievements and can be expected to have confidence in GE's analyses.
- 6. The first two categories of information requested by UCS can be viewed together to demonstrate the ways in which a competitor would gain a windfall were the data made public.
In category (a), UCS requests the accident sequences which dominate the overall probability of core damage. In category (b), the dominant contributors to accident sequence classes with corresponding probability estimates are requested. */
Defining and describing the dominant accident sequences is a major portion of the effort involved in constructing the event trees. After determining the initiating event, the analysis requires a thorough understanding of the safety functions to
- / Category (b) is simply a refinement of category (a) but includes the quantitative probabilities as well as the qualitative information in category (a).
be performed and the interrelationships of the functions. Once the safety functions are identified, it is necessary to identify the systems provided to perform the functions, the relationship of the functions to the operation of the systems, and the interrelationship of the systems. In performing this analysis, illogical or meaningless combinations of events are eliminated thus narrowing the number of accidents to be considered. The dominant sequences are established after grouping of the remaining accident sequences into classes or categories of sequences having similar release consequences and determining, by assigning probabilities to the sequences in each category, those few high probability sequences within each category which dominate the lower probability sequences that have similar releases.
- 7. The process of defining the dominant accident sequences is an iterative process in order to insure accuracy and completeness. It involves, among other things, a detailed knowledge of the reactor design, the complex relationship of functions and systems and the interrelationships between t
l functions and between systems. The public disclosure of the
! accident sequences which dominate the overall probability of core damage would provide GE's competitors with the relative l importance of the various accident sequences in a BWR. In l
l addition, release of the dominant accident sequences would i
provide competitors with the qualitative construction of the i branches on the accident event trees (each sequence defines one branch). Release of the probabilities as requested in
s category (b) would provide competitors with the results obtained from the sy' stem fault trees and from GE's proprietary data base.
- 8. The information in these two categories could be used by competitors in precisely the ways described in the Sholly Declaration. These sequences and probabilities may be valuable to GE's competitors and, of course, potential utility t or industry group customers, to compare PRA analyses, to determine whether plant modifications are necessary or desirable, or to improve operator procedures. GE has in the past performed
~
such analyses for its customers. For example, it has used the GESSAR II accident sequences in an analysis for IDCOR con-cerning operating procedures in the event of station black-out.
- 9. In category (c), UCS requests the listing of the dominant contributors to the release categories. These dominant contributors are developed and used in the definition of the event trees to determine the dominant accident sequences.
As noted above, the grouping of sequences by release character-istics is done in order to screen accident sequences which do not contribute significantly to the overall probability of occurrence of each release category. The grouping is performed by comparing accidents by certain characteristics and placing those with similar characteristica in the same category.
- 10. In grouping or "binning" the various accident sequences, GE, based on experiments, data and its engineering <
judgment, established unique criteria or characteristics.
These criteria or characteristics are different than those used by GE's competitors and result in more realistic assess-ments of the consequences of certain accident classes. A competitor of GE, by reviewing the dominant contributors to each accident class category, could easily deduce the grouping criteria developed by GE. By obtaining the dominant contributors, a competitor would gain the benefits of GE's s
research in the development of the criteria at little or no cost.
- 11. Category (d) is in essence a subset of the material requested in categories (a) and (b) and for the same reasons should not be released.
- 12. In category (e), UCS requests external event hazard curv_es. The seismic hazard curve is already available.
GE created no other hazard curves.
- 13. In category (f) UCS requests the uncertainty analyses results. Those "results" include the ranges of estimates of accident sequence probabilities, accident class
, probabilities, release category probabilities and release category magnitudes. An uncertainty analysis is a study of the variabilities in the frequency of core damage resulting from the propagation of uncertainties in the event and fault e . _ . .- __ -. _. .. , - -. . .
trees. The analysis starts with the actual points in the trees and provides a range or distribution of the probability of occurence of the particular event. Were the information in category (f) made public, a competitor would gain a detailed understanding of the GESSAR II PRA. The competitor would learn what accident sequences were important and those that were unimportant, the complete quantitative listing of the accident sequences and the quantitative listing of the release categories. As Mr. Sholly notes, this information could be used by a competitor or a potential customer in the preparation of studies of proposed plant modifications i.e., to determine which kinds of modifications show the greatest reduction in risk. It should be noted that GE has performed such studies in the past for nuclear plants for precisely this purpose. In addition, this information would be of great assistance to a competitor in preparing a full scale PRA because the data reflects the results of the event tree and fault tree analysis.
- 14. In catogory (g), UCS requests the risk curves, expectation values and peak values. The risk curves and expectation values have already been made available to the public. The GESSAR II PRA and related documents submitted to NRC do not contain peak values.
' 5.
1 In category (h) UCS requests the accident sequence timing data calculated from the MARCH code results.
The timing data was developed by GE using not only the MARCH code, but also proprietary codes developed by GE. These
n- ,
GE codes were specifically designed to accurately model boiling water reactors. The MARCH code on the other hand was developed for pressurized water reactors. These proprietary GE codes are not available to any of GE's competitors.
Release of the timing data would allow competitors to determine how the accident phenomenology is modeled and assess the containment failure modes.
- 16. In category (i), UCS requests the fission product source term estimates for the various release categories.
Release of these estimates would provide competitors with a great deal of information regarding the underlying assumptions and methodologies. The source term estimates in the GESSAR II PRA are significantly different than those found in other PRA's and those differences would reveal a great deal to a competitor about fission product transport modeling and assumptions. The modeling and assumptions leading to these differences are unique to GE and reflect GE's greater under-standing of BWR's. In addition, this data could be used directly by a competitor to perform analyses for utilities of accident mitigative features and in direct comparisons with other PRA's -- a service often requested by utilities.
Thus, a competitor or customer would obtain and use not only the data itself but a great deal of the underlying analysis. ,
- 17. In category (j) , UCS requests the estimates of 1
the quantities and sources of non-condensible gases such as
hydrogen, generated during the accident sequences. Based on its experience in BWR design, GE is able to more accurately model hydrogen generation than any of its competitors.
Because of its efforts in this area, GE has sold, without any underlying data, hydrogen generation rates to an industry group. Were this information made publically available GE would lose its competitive advantage in this area.
- 18. In summary, the information requested in the ten categories would be commercially valuable to GE's com-petitors in at least three ways. First, by obtaining the results of GE's analyses, the time and expenditure necessary to substantiate those results would be lessened substantially.
Second, the information denominated by UCS as "results" would provide competitors with a great deal of information regarding GE's modeling assumptions, methodology and criteria.
Third, in certain instances the information could be directly used by competitors or potential customers. Given that NRC has already reviewed the underlying data, potential customers of GE, for example, could be expected to present this data to NRC in its present form.
ffboynhv G.
Deborah A. Hankins, Ph.D.
N" S
,ubscribed 1985. and sworn to before me this fN day of
//cm d"[AR Notary Pbblic My Commission Expires Mfu 4/v // /98 7 OFFICIAL SEAL '
MARY L K1ENDALL '
NOTAfrr PusDC - CAUFORNIA
! SANTA CtAPA COUNTY {
L
CESSAR II 22A7007 q 238 NUCLEAR ISLAND Rev. 2 CENERAL ELECTRIC COMPANY .
PROPRIETARY INFORMATION '
i-i F.3.1.3 Vaporization Release (Continued) '
Fygy = vaporization release in the RPV
{
Fyy =
vaporization fraction limit which is based on the following discussion In WASH-1400, which assumed no in-vessel vaporization or attenua-tion, the vaporization release fraction during the core-concrete interaction in the drywell was 100% of the remaining Kr, I, Cs, <
and Te, 1% of the Sr, Ba, and La, and 5% of the Ru. For this assessment, the above release fractions for core-concrete inter-action are used except 0.5% of the Sr, Ba, and La and 3% of the '
Ru were released (Reference F.3-1). j Since this assessment has limited the total vaporization release 5 to the above values, the fission products released to the drywell a from vaporization in the RPV and in the drywell are limited to the following condition: ;
i F
vid Fvid + Fygy / RPVdf 1Tyy where RPV In vessel decontamination factors which are tabulated df in Table F.3-2 The fission product source terms of the in-vessel vaporization contributions and their corresponding RPV attenuation factors are illustrated in Table F.3-1 and Table F.3-2 along with the source terms fron other release periofa.
15.D.3 67 KUL b TJ S jff - I
.o. ,
GESSAR II 22A7007 238 CUCLEAR ISLAND R,v. 11 GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION Class III F.3.2 Fission Product Transport and Deposition (Continued) ,'
For the leakage paths which are drywell or containment cracks, the DF values of the cracks are based on the crack dimensions and the aerosol concentration available for plugging the cracks. The DFs are estimated from the model cited by Morewitz (Reference F.3-4). {
For natural deposition in the reactor pressure vessel during the {
gap and melt periods, a particulate DF value of 10 is assumed for '
I
? )
transient and small break cases. This DF value is due to the long residence time of aerosols in the vessel after the fission products p are released from the melting fuel (Reference F.3-3) . The CORRAL ,
code calculates natural deposition in each compartment based on I flow rates and surface to volume ratios. i F.3.5 ilmsion Product Release Fractions The fractions of the core inventory released to the environment are presented in Table F.3-4. The data in the table are used by the CRAC code for assessing the consequences of release for each acci-dent scenario.
F.3.4 References F.3-1 NUREG-0772, Technical Bases for estimating Fission Product Behavier during LWR Accidents (1981).
F.3-2 NUREG-75/014 WASH-1400, Reactor Safety Study (1975).
F.3-3 NRC Memo, Response to Commissioner's Requests during brief-ing of SECY 81-240 Draf t NUREG-0771 and NUREG 0772, '
June 1981.
l F.3-4 H. A. Morewitz, Submitted to Health Physics in 1981, also cited in NUREG 0772, Section 7-6.
F.3-5 NUREG/CR-0324/ SAND 78-1511, " LWR Safety Research Program, Quarterly Report, January - March, 1978" P17-24.
F.3-6 NUREG-0183-5/ SAND 78-0076, " LWR Safety Research Program, Quarterly Report, July - September, 1977" P8-12.
15.D.3-570
CESSAR II 22A7007 238 NUCLEAR ISLAND Rev. 2 GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION Class III Table F.3-2 PARTICULATE DECONTAMINATION FACTORS
, WITHIN RPV Decontamination Factor During Gap, After Accident Class Melt Period Core Slump CLASS I - transient 10 20 CIASS I - small break 10 10 CLASS I - large break 1 10 CLASF II - large break 1 100 CLASS II - transient 10 100 CLASS IV - ATWS 10 20 l
15.D.3-5,73 t- b
~
.. <' , ;, . CESSAR II 22A7007 238 NUCLEAR ISIAND Rev. 2 GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION
..a . . _ _ . . _ . . . .
class III Table F.4-3 CORE INVENTORY NUPSER NAME OROUP PARENT I NI TI ALI CURIES )
\
1 CO SS 7 3.000E*04 2 CO-80 7 2.900E*04 3 KR-SS 1 7.700E*05 4 KR SSM 1 2. 990E
- 07
. 8 KR-87 1 5.SSDE*07 8 KR-SS 1 7.980E*07 7 RB-46 4 8.300C+04 8 SR-49 8 1. 060E
- 08 9 SR-90 S 6.660E+06 10 SR-91 S 1. 353E + 08 11 Y-90 4 SR-90 S.940E*06 12 Y-91 4 SR-91 1.370E+08 13 ZR-95 4 1. 630E + 08 14 2R-97 4 1. 880E + 08 15 pe-95 8 2R-95 1.550E+08 15 MO-99 7 1. 800E + 08 17 TC-99M 7 MO 99 1. 550E + 08 18 RU 103 7 1. 310E + 08 19 RU 105 7 7.980E+07 to RU-106 7 2. 930E + 07 21 RH-105 7 RU-105 7. 090E + 07 22 TE-127 5 SB-127 6. 660E + 0G 23 TE-127M 5 9.0E*E+05 24 TE-129 5 SB-129 2.45*E+07 25 TE-1299 5 6. 660E + 06 26 TE-131M 5 1. 263E + 07 27 TE-132 5 1. 330E + 08 28 58-127 5 6.870E+06 29 SB-129 5 2. 610E + 07 30 1-131 3 TE-131M 9.130E + 07 31 1-132 3 TE-132 1.340E+08 32 I-133 3 1. 980E + 0 8 33 1-134 3 2.190E+08 34 1-135 3 1. 860E + 08 35 XE-133 1 1-133 2. 070E + 0 6 30 XE-13* 1 1-13* 3. 530E = 07 37 CS-134 4 8 . 800E + 06 38 CS-136 4 2. 4 30E + 06 39 CS-137 4 7. 970E + 06 40 BA-140 6 1.750E+06 41 LA-140 S BA-140 1.790E+08 42 CE-141 4 1. 630E + 08 43 CE-143 8 1. 500E + 08 44 CE-144 4 9.840E+07 i 45 PR-143 8 CE-143 1. 570E + 08 I
46 NO-147 8 6. 550E + 07 47 NP-239 8 1. 090E + 09 48 PU-238 8 CM-242 7.640E+04 49 PU-239 8 NP-239 2.470E*04
- 50 PU-240 6 CM-244 2.800E+04 51 PU-241 4 S.490E+06
{ . 52 AM-241 8 PU-241 3.700E+03
- 53 CM-242 8 1.200E+06 54 CM-244 8 1.410E+04 l
15.D.3-584
.- . . ' - CESSAR II 22A7007 238 NUCLEAR ISLAND Rev. 2 GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION Class III Table F.4-6 EVACUATION AND SHIELDING DATA MAXIMUM DISTANCE OF EVACUATIONtM) 1.624E+04 EVAC VEL (M/S) - ACCEL INEG) (M/S/S) 5.360E-Ol TIME LAG BEFORE EVACUATIONtOAYS) O.
TRAVEL DISTANCE WHILE EVACUATING S.000E*03 ANGLE OF EVACUATED DOWNWIND SECTORS 4.500E*01 EVACUATION DIRECT COSTtS/ EVACUEE / DAY) 9.500E*01 CtlTERIA OF DURATION OF RELEASE FOR EVAC 3.000E*00 CLOUD SHIELDING WITH EVACUATION 1.000E*00 CLSUD SHIELDING WITHOUT EVACUATION 7.500E-01 GROUND SHIELDING WITH EVACUATION 5.000E-01 GROUND SHIELDING WITHOUT EVACUATION 3.300E-01 BREATHING RATE 2.660E-04 1
Table F.4-7 PIANT DIMENSIONS DATA REACTCF B U I L DI NG LEN3TM (M) 7.500E-01 REACTOR BUILDIN3 HEIGHT (M) 5.900E-01 e er INTERVALS FOR SPECIAL WAv.E EFFECTS 0
, n 15.D.3-587
. GESSAR II 238 NUCLEAR ISLAND '
)
GENERAL ELECTRIC COMPANY l PROPRIETARY INFORMATION l Class III ;
F.3-5 NUREG/CR-0324/ SAND 78-1511, " LWR Safety R'esearch Progrim Quarterly R'epo' r t, January - March; 1978" P17-24.
~e F.3-6 NUREG-0183-5/ SAND 78-0076, " LWR Safety Research Program, Quarterly Report, July - September,1977" P8-12
(
15.D.3-570
. W