ML20129E530
| ML20129E530 | |
| Person / Time | |
|---|---|
| Site: | 05000447 |
| Issue date: | 01/31/1983 |
| From: | Sherwood G GENERAL ELECTRIC CO. |
| To: | Eisenhut D Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML20127A304 | List: |
| References | |
| FOIA-84-175, FOIA-84-A-66 NUDOCS 8506060529 | |
| Download: ML20129E530 (130) | |
Text
_
GENERAL h ELECTRIC NUCLEAR POWER SYSTEM S ~ DIVISION CENERAL ELECgC g AVE., SAN JOSE. CALIFORNIA 95125 P
7 January 31, 1983 U. S. Nuclear Regulatory Commission Office of Nuclear Reactor Regulation Washington, D.C.
20555 Attention:
Mr. D. G. Eisenhut, Director Division of Licensing Gentlemen:
SUBJECT:
SUBMITTAL OF PROPRIETARY INFORMATION IN RESPONSE TO REQUEST FOR ADDITIONAL INFORMATION REGARDING SEVERE ACCIDENT PORTION OF THE GENERAL ELECTRIC STANDARD SAFETY ANALYSIS REPORT (GESSAR II); DOCKET NO. STN 54 '47
)
Reference:
F. J. Miraglia (NRC) letter to G. G. Shemood (Lin
" Request for Additional Information Regarding Severe Accident Portion of the General Electric Application for an FDA for a Standardized Nuclear Island (GESSAR-II),"
December 27, 1982.
The referenced letter requested additional information regarding the severe accident portion of GE's GESSAR-II submittal.
Attached please find responses to the questions included in the referenced letter.
Also attached are drafts of the proposed modified Appendices C and D to the GESSAR-II Probabilistic Risk Assessment (Section 150.3 of GESSAR-II) which provide additional scrutability to the fault tree and event tree portion of the PRA.
We are requesting this information be withheld from public disclosure and considered as proprietary pursuant to Section 2.790 of 10CFR Part 2.
Very truly yours, Glenn G. Shemood, Manager Nuclear Safety & Licensing Operation GGS: cal /K01315 8506060529 841203 PDR FOIA Attachments CURRAN 84-A-66 PDR cc:
F. J. Miraglia (w/o Att.)
C. O. Thomas (w/o Att.)
D. C. Scaletti (w/o Att.)
L. S. Gifford (w/o Att.)
0-f f
GENERAL ELECTRIC C0MPANY AFFIDAVE I, Glenn G. Sherwood, being duly sworn, depose and state as follows:
1.
I am Manager, Nuclear Safety & Licensing Operation, General Electric Company, and have been delegated the function of reviewing the information described in paragraph 2 which is' sought to be withheld and have been authorized to apply for its withholding.
2.
The information sought to be withheld is contained in the proprietary responses to questions on the Severe Accident position of the 238 Nuclear Island General Electric Standard Safety Analysis Report (GESSAR II).
3.
In designating material as proprietary, General Electric utilizes the definition of proprietary information and trade secrets set forth in the American' Law Institute's Restatement Of Torts, Section 757.
This definition provides:
"A trade secret may consis't of any formula, pattern, device or compilation of information which is used in one's business and which gives him an opportunity to obtain an advantage over competitors who do not know or use it....
A substantial element of secrecy must exist, so that, except by the use of improper means, there would be difficulty in acquiring informa-tion....
Some factors to be considered in determining whether given information is one's trade secret are:
(1) the extent to which the information is known outside of his business; (2) the extent to which it is known by employees and others involved in his business; (3) the extent of measures taken by him to guard the secrecy of the information; (4) the value of the information to him and to his competitors; (5) the amount of effort or money expended by him in developing the information; (6) the ease or difficulty with which the information could be properly acquired or duplicated by others."
4.
Some examples of categories of information which fit into the definition of proprietary information are:
a.
Information that discloses a process, method or apparatus where prevention of its use by General Electric's competitors without license from General Electric constitutes a competitive economic advantage over other companies; b.
Information consisting of supporting data and analyses, includ-ing test data, relative to a process, method or apparatus, the application of which provide a competitive economic advantage, e.g., by optimization or improved marketability;
c.
Information which if used by a competitor, would reduce his expenditure of resources or improve his competitive position in the design, manufacture, shipment, installation, assurance of quality or licensing of a similar product; d.
Information which reveals cost or price information, productio'n capacities, budget levels or commercial strategies of General Electric, its customers or suppliers; Information which reveals aspects of past, present or future e.
General Electric customer-funded development plans and programs of potential commercial value to General Electric; f.
Information which discloses patentable subject matter for which it may be desirable to obtain patent protection; g.
Information which General Electric must treat as proprietary according to agreements with other parties.
5.
In addition to proprietag treatment given to material meeting the standards enumerated above, General Electric customarily maintains in confidence preliminary and draft material which has not been subject to complete proprietary, technical and editorial review.
This practice is based on the fact that draft documents often do not appropriately reflect all aspects of a problem, may contain tentative conclusions and may contain errors that can be corrected during normal review and approval procedures; Also, until the final document is completed it may not be possible to make any definitive determination as to its proprietary nature.
General Electric is not generally willing to release such a document to the general public in such a preliminary form.
Such documents are, however, on occasion furnished to the NRC staff on a confidential basis because it is General Electric's belief that it is in the public interest for the staff to be promptly furnished with significant or potentially significant information.
Furnishing the document on a confidential basis pending completion of General Electric's internal review permits early acquaintance of the staff with the information while protecting General Electric's potential proprietary position and permitting General Electric to insure the public documents are technically accurate and correct.
6.
Initial approval of proprietary treatment of a document is made by the Subsection Manager of the originating component, the man most likely to be acquainted with the value and sensitivity of the information in relation to industry knowledge.
Access to such documents within the Company is limited on a "need to know" basis and such documents at all times are clearly identified as proprietary.
7.
The procedure for approval of external release of such a document is reviewed by the Section Manager, Project Manager, Principal Scientist or other equivalent authority, by the Section Manager of the cognizant Marketing function (or his delegate) and by the Legal Operation for technical content, competitive effect and determination of the accuracy of the proprietary designation in accordance with the
standards enumerated above.
Disclosures outside General Electric are generally limited to regulatory bodies, customers and potential customers and their agents, suppliers and licensees only in accord-ance with appropriate regulatory provisions or proprietary agree-ments.
8.
The information mentioned in paragraph 2 above has been evaluated in accordance with the above criteria and procedures and has been found to contain information which is proprietary and which is customarily held in confidence by General Electric.
9.
The information mentioned in Paragraph 2 provides additional informa-tion on the GESSAR-II Probabilistic Risk Assessment contained in 4
Section 150.3 of the GESSAR-II submittal.
10.
The information to the best of my knowledge and belief, has consis-tently been held in confidence by the General Electric Company, no public disclosure has been made, and it is not available in public sources.
All disclosures to third parties have been made pursuant to regulatory provisions of proprietary agreements which provide for main'tenance of-the information in confidence.
11.
Public disclosure of the information sought to be withheld is likely to cause substantial harm to the competitive position of the General Electric Company and deprive or reduce the availability of profit-
. making opportunities because:
It was developed with the expenditure of resources exceeding a.
$500,000.
b.
Public availability of this information would deprive General Electric of the ability to seek reimbursement, would permit competitors to utilize this information to General Electric's detriment, and would impair General Electric's ability to maintain licensing agreements to the substantial financial and competitive disadvantage of General Electric.
I c.
Public availability of the information would allow foreign competitors, including competiting BWR suppliers, to obtain containment information at no cost which General Electric developed at substantial cost.
Use of this information by l
foreign competitors would give them a competitive advantage I
over General Electric by allowing foreign competitors to l
produce their containments at lower cost than General Electric.
I 6
m
--- -, ~., - -
aq--,,
- - -, - -,, - - - -. ~ - - - -,,, --
,,,-,n,---,w-------
-,-- +-
STATE OF CALIFORNIA
) **.
COUNTY,0F SANTA CLARA
)
Glenn G. Sherwood, being duly sworn, deposes and says:
That he has read the foregoing affidavit and the matters stated therein are true and correct to the best of his knowledge, information, and belief.
Executed at San Jose, California, this g day of January
, 1982 Glenn G. Sherwood General Electric Company Subscribedandswornbeforemethig3 payof January 1983.
eeeeeeeeeeeeeece:ece
[mammmo:apires Dec.
OFFICIAL SEAL KAREN 5. VOGELHUBER SYA AssLT W&01ll JJ)
My comminien t NOTARY PUBLIC, STATE SF CALIFORNIA mmmmmm KH:eal/K01316
'1/31/83 1
e l
l l
t
- I
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.1 War, on-line repair or recovery modeled for all systems considered in the PRA?
If not all systems, then list which systems were considered.in on-line repair or recovery.
How was recovery modeled? What is the difference, if any, between on-line repair and recovery?
.3
RESPONSE
't 4
l l
l l
i l
l l
1 l
GENERAL ELECTRIC COMPANY PROPRIETARY INFONTION QUESTION 720.2 Provide all sources for the failure data in the PRA.
Explain the criteria used for the selection of one data r,ource over the other.
What was the rationale for combining several data sources in some instances and not in others? What were the guidelines used to determine whether or not the data base should be integrated?
RESPONSE
The failure data for the fault tress used in the GESSAR II PRA is provided in the proposed revision to Appendix D to 150.3 (attached). ___
OENERAL RICTRIC COMPANY PROPalETARY INFORMATION QUESTION 720.3 Provide additional information on the treatment of effects of extreme environmental conditions following core melt accidents (beyond DBA conditions) on systems and components and of com' mon manufacturing or design errors of equipment considered in the PRA.
Examples would include but not be limited to the following:
a)
Effects on electrical insulation due to voltage treeing and dielectric loss.
b)
ADS valves ind control logic system.
c)
Use of RCIC when the ambient temperature has exceeded 200'F (this high temperature condition would cause insufficient lube-oil cooling and fail RCIC).
d)
Effects on instrumentation following containment failure and subsequent effects on successful injection (both automatic and manual).
e)
Effects on drywell structural integrity under advarse thermal stress conditions, f)
Effects on safety-related equipment due to prolonged electrical short circuits.
RESPONSE
The question is not clear in defining the period of time of interest.
The point in time where DBA conditions are exceeded does not coincide with the start of core melt.
In all cases, core melt occurs well after A
,g the time when DBA conditions are exceeded.
t 4
i. - -
OgspAL ascTalc COMPANY Y
PROPWRARY INMHtMATION
~
l 4
?
t t
(
b h
B I
l l
l l
- Reference (1):
Hannant, D.
J.,
Effects of Heat on Concrete Strength, l
Engineering (London), V.197 No. 5105, Feb. 1964, I
i
- p. 302.
l l l
GENERAL ELECTRN: COMPANY PROPRLETARY INFORMATION QUESTION 720.4 Provide the GE calculations showing that core damage can be avoided if half of the active fuel remains uncovered during an accident.
How I
sensitive is the a'ssumption of core melt frequency?
RESPONSE
9..
e 0
8-
Wa N ie d kLkbi m s 0 0 n. neul
)
PROPRIETARY INFORMATION QUESTION 720.5 Provide in detail the basis for each of the success criteria used in the PRA (both ATWS and non-ATWS).
If other GE analyses have been referenced,
)
provide each reference or report for our review.
This question refers to both safety and non-safety related systems, and both front-line and support systems considered in the PRA.
RESPONSE
g e
l GENERAL ELECTRIC COMPANY PROiTufTARY fr4 FORMATION j !
J OENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.6 Provide the specifics of how the reductions in transi,ent initiator frequencies were calculated in Table A.1-3.
Discuss the bases used in i '
this evaluation.
Provide the procedures used in applying these reductions in order to arrive at the initiator frequencies in Table A.1-2.
RESPONSE
l
QUESTION 720.8 OENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION I o, d
^\\
3,
/
f Provide the specific method used in order to arrive at the value
/
.l o.-l
' 5 does for the failure to close of one safety relief valve (p. 3-242).
this value compare with experience?
RESPONSE
The operating BWR experience consists mainly of 3-stage Target Rock (TR) valves which have experienced both failure to reclose after opening on demand (50RV event) and inadvertent opening (IORV event).
The BWR/6 valves are made by Crosby.
This valve is of superior design compared to the 3 stage TR valve and has demonstrated high reliability in thousands of qualification tests.
6
).
w.
~e M l i
QUESTION 720.9 G8MF#4L ELECTRIC COMPANY How were mechanical common-mode failures included in the unavailability g i of Scram or Alternate Rod Insertion (ARI) y Provide your bases
!5
~
I for this unavailability estimate (p. 3-241).
RESPONSE
GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION
QUESTION 720.10 GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION
}o.d Provide the basis for using the j
(p. 3-279) i J r, 4.
Jp. 3-242).
Elaborate on why the CI function was not degraded for '
t the small LOCA since it would be difficult to transfer an adequate amount of water to the hotwell.
RESPONSE
1 o
a 6
9
+
d O
E e
i h
'c f
't 4
9 0
t-k b
1 l
s ' '
l 5
QUESTION 720.11 GENERAL FLECTRIC COMPANY
- *'M a,TtON Explain why a failure of the containment spray could be treated as a successful sequence (Figure C.12-1).
What is the relationship between the containment spray and containment vacuum breakers in the contrast of Figure C.12-1?
RESPONSE
This was an error in Figure C.12-1 as originally submitted.
This figure has been corrected in the proposed revision of Appendix C to 150.3 I
GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION.-
I GENERAL ELECTLC COMPANY QUESTION 720.12 PROPRIETARY INFORMATION What does CT7 signify in Figure C.15-1?
RESPONSE
The definition of CT7 is provided on Page 15.D.3-217 of the GESSAR II PRA.
No containment event tree is required for this event.
9 e
GENERAL ELECTRIC CUMPANY PROPRIETARY INFORMATION QUESTION 720.13 In the LPCI fault tree (p. 3-397), the loss of suppression pool (LSP) function was shown transferred to the HPCI fault tree.
Explain how the, LSP function is derived and how it is related to the HPCI system in GESSAR.
Review of the HPCS fault tree did not show any LSP function (p. 3-341).
RESPONSE
Both the HPCS and LPCI systems use water from the suppression pool.
For the HPCS, the primary source of water is from the condensate storage tank, with the suppression pool being available as an alternate or backup source of water.
In the LPCI system, the suppression pool is the only source of water.
The LSP function in the LPCI fault tree (as modeled in the computer code) includes the following two elements:
E The error has been corrected in the proposed revision of Appendices C and D.
9 G6:#GL ELECTRIC COWAM QUESTION 720.14 PROPRIETARY INFORMATION In the GESSAR II P_RA, tne containrrent isolation failure probability was lo -
assumed to be
. / demand.
Provide the bases and the details of the analysis in arriving at the value (p. 3-302).
4.3 /
RESPONSE
The bases and the details for the containment isolation failure probability are provided in Table 0.1.7-2 of'the proposed revision to Appendix D to 150.3.
6.
CENERAL ELECTRIC CEMPANY PREPRIETARY INFORMATION QUESTION 720.15 In the loss of offsite power (LOOP) event tree, an initiator frequency of was used (p. 3-257).
j it,
1.: 3 a)
Provide the basis for the selection of 0.05.
b)
Explain the method used in arriving at such a value and provide the basis of any recovery and the duration of the outage assumed.
c)
If a minimum outage duration was assumed, how were the events with durations shorter than that of the minimum outage included in the analysis?
RESPONSE
a)
The bases and the details for selecting an initiating frequency of is provided in Appendix A, Section 6 of the GESSAR-II PRA.
l
- .', 4 b)
C) i.
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.16 Provide the rationale and the quantitative evaluation for not considering the total or partial loss of DC power as an accident initiator.
RESPONSE
E
, - :f l.,
(
9 f..
GENERAL' ELECTRIC COMPANY QUESTION 720.17 PROPRIETARY INFORMATION Provide all numerical values used in system fault trees in the GESSAR-II PRA.
RESPONSE
Appendix 0 to 15D.3 has been revised to include all numerical input values to the system fault trees.,
1, i
o O
GENERAL ELECTRIC COMPANY QUESTION 720.18 a)
RMREW MNW In addition to the system fault trees that are included in the PRA, functional event trees and functional fault trees were used to calculate branch point probabilities in the event trees.
Provide all functional fault trees and functional event trees used in the analysis; furnish also numerical values used in these trees.
RESPONSE
The functional level fault trees and event trees used in the analysis are documented in the proposed revision to Appendix 0 of 150.3 along with the numerical values used in the trees.
u CENERAL ELECTRIC COMPANY QUESTION 720.18 b)
Provide a detailed discussion on how dependencies were evaluated when the event trees quantified; these dependencies include:
1)
Support system dependency - the sharing of support systems between systems and functions, for example, HPC'S, LPCI, RHR and PCS, etc.,
depending on AC.
RESPONSE
- t..
l e
i
CENERAL RECTitic COMPANY PROPRIETARY INFORMATION QUESTION 720.18 b) ii)
Hardware dependency - common hardware shared between different systems, e.g.,
injection lines, valves, etc.
RESPONSE
i O
O 4'
e I
, j
CENERAL ftECTRIC CCMPANY PROPRIETARY INFORMATION QUESTION 720.18 b) 111)
System dependency between functions - in the event tree, the feedwater and the power conversion system (PCS) were grouped together under the function U ; later on, in the W function, the PCS was also included for p
How would the failure of U affect the success of p
the W function?
(pp. 3-233, 250).
Discuss also the treatment of 2C dependency between low pressure core cooling (LPCC) and W2C (p. 3-236).
RESPONSE
O i s 4 *-
. -m+
, + - -,. -
,,---e--s,-
_m,_,_
1 i
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.18 b) iv)
Initiator faults impacting mitigating systems.
RESPONSE
i e
'l rs 4
1 I. _. -.
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION EXAMPLES OF INITIATING EVENT AND MITIGATING SYSTEM INTERDEPENDENCIES INITIATING EVENT SYSTEM LIMITATIONS O
t O
.L' GENERAL ELECTRIC COMPANY QUESTION 720.19 PROPRIETARY INFORMATION In Section C.3.2, " Turbine Trip Without Scram", (p. 3-237) it is stated that, "In order to quickly reduce reactor power to about of the 2'
4 e, pre-transient power level, a RPT is initiated by the redundant reactor control system (RRCS).
The probability of failure of the RPT includes the RRCS common cause failure probability and is given by R.
This failure may result in containment overpressure in about 10 minutes".
Provide your bases for the value assigned to R.
RESPONSE
In the proposed revision of Appendix C, Section C.3.2 now states that i
I, and is given by R.
The j7 p
basis for the~v~alue assigned to R is provided in Section D.1.1.2 of the
~
p
, proposed revision to Appendix D of 15D.3.
l 33.-
GENERAL ELECTRIC COMPANY OPRIETARY MNTION QUESTION 720.20 What manual action is necessary to accomplish the feedwater runback?
Describe the step by step actions involved and the locations of each action.
RESPONSE
Feedwater runback is automatically initiated by the redundant reactor control system (RRCS).
No manual action is necessary.
4 e. _ _ _..___
b CENERAL ELECTRIC COMPANY QUESTION 720.21 PROPRIETARY INFORMATION Provide documentation for the unavailability of feedwater and PCS values used in all event trees in the GESSAR-II PRA.
RESPONSE
The unavailability of the feedwater and PCS values used in the event trees are documented in Section D.1.3 for the high pressure coolant injection function and Section 0.1.5 for the containment heat removal function.
These sections are contained in the proposed revision of Appendix D to 150.3.
e - _ _.
GENERAL ELECTRIC COMPANY QUESTION 720.22 PROPRIETARY INFORMATION Why was the loop event applicability included only in the LOCA trees but not in any other tree? Provide discussion on how transient induced LOOP was included in the analysis.
RESPONSE
e '
eh O
6 1
0 -.
i QUESTION 720.23 GENERAL ELECTRIC COMPANY i
PROPRIETARY INFORMAT ON During the course of a severe accident, the main steam isolation valves (MSIVs) may be exposed to temperature conditions beyond the design limit resulting in the degra5ation and potential failure of these valves.
Provide an assessment of the integrity of the MSIVs during limiting accident sequences accounting for the various heat inputs to the valves.
Discuss the potential impact and the likelihood of releasing
~
radionuclides through a partially failed MSIV to the environment.
RESPONSE
The MSIVs are designed to automatically close when reactor water level lowers below RPV Water Level 1 on BWR/6.
This level is about one foot above the top of active fuel.
Because adequate core cooling is assured until the, reactor water level is substantially lower (i.e., about the bottom of active fuel), automatic closure of MSIVs is assured prior to the release of severe accident source terms.
j.
l.
Leakage from the RPV through MSIVs to the environment is precluded by the MSIV Positive Leakage Control System discussed in Section 6.7, by the redundant outboard MSIV, and by the main steam line shutoff valve located _
downstream of the outboard MSIV in the steam tunnel.
l f
- Furthermore, significant removal of the radionuclides (equivalent to pool scrubbing) would be expected in the MSIVs due to tortuous paths and the aerosol composition of the severe accident sources.
The potential impact
.of such a release is thus not significant.
GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION
{ <
CENERAL ELECTRIC COMPANY QUESTION 720.24 The RHR system consists of two trains, each of which has two pumps for redundancy of their twin functions, shutdown and suppression pool cooling modes.
However, there is only one suction path'in the shutdown cooling mode.
Failure of any one of three valves would disable the RHR system in the shutdown mode.
In the suppression pool cooling mode, although there are two suction and two discharge paths for th'e pumps, that redundancy can also be negated by failure of any pair of valves in opposite discharge paths or failure of the minimum flow bypass valves.
Discuss the effect of this faiure on the RHR system availability.
RESPONSE
The shutdown cooling and suppressi,on pool cooling modes of the RHR system are described in Sectior) 5.4.7..f I
~
1 it 9
0 s.
-1 I
.h
.I, i
- ' - - - - -.. _ ~ _.
, g-This mode is discussed in the Emergency Procedure g,
Guidelines and in Section 150.2.1.3.
___._--m A
~
mm
_m 39 -
GENERAL ELECTRIC C8MPANY g,3 PROPRIETARY INFORMATION In the event that ' resin within any deminera'lizer is broken up into fragments, provide further details in the likelihood of occurrence and progression of such an event and on how it may result in subsequent degradation or failure of coolant injection or makeup.
RESPONSE
9 e
3L e
+
W e
GENERAL ELECTRIC COMPANY QUESTION 720.25 PROPRIETARY INFORMATION Provide the fault-tree by which the common cause failure of the ARI and standby liquid control systems were evaluated.
Provide also the numerical values used in the fault tree (p. 3-237).
RESPONSE
The calculation basis and fault tree are described in the proposed revision to Appendix D of 150.3.
The following sections of the proposed revision apply to this question:
Appendix 0 Title Section D.1.1.1 Failure of Scram and ARI Section D.1.1.2 Failure of RPT and Feedwater Runback Section 0.1.1.3 Failure of RPT and RRCS Due to Common Cause Figure D.2-12 SLC Fault Tree Figure D.2-13 RPT Fault Tree
. i
\\-
GENERAL ELECTRIC COMPANY OUESTION 720.26 PROPRIETARY INFORMATION
~ Provide all revisions to the PRA per discussion at the BNL meeting on August 26, 1982.
RESPONSE
The. proposed revised C&D Appendices are included in this transmittal.
e h
e GENERAL ELECTRIC COMPANY QUESTION 720.27 PROPRIETARY INFORMATION Provide the basic event importance during each phase of the ECCS operation.for the dominant accident sequences in the GESSAR-II P.RA.
Basic event importance is defined as the probability the basic event is contributing to system failure given the system is failed.
The ECCS operation may be considered to be a three phase mission, i.e.,
initial core cooling, suppression pool cooling, and residual heat removal.
The phase boundary times for the ECCS operation should correspond to each accident sequence.
RESPONSE
e c.*~
.4*
6 i
D('-
GENERAL ELECTRIC COMPANY 1
QUESTION 720.28 PROPRIETARY INFORMATION Discuss any potential impact of the ORNL's precursors study (NUREG/CR-2497) on the GESSAR-II PRA with respect to the core melt probability and overall risk assessment.
RESPONSE
The potential impact of the precursors study (NUREG/CR-2497) on the GESSAR-II PRA with respect to the core melt probability and the overall risk assessment is expected to be negligible for the following reasons:
1)
Precursors have resulted in improvements in plant design and operation.
2)
The BWR/6 Standard Plant design has benefited from the on going review of plant operating experience and Post-TMI investigations.
3)
The BWR/6 solid state design represents a significant improvement from current plants.
4)
The BWR design can be more tolerant of failures.
l l l-
GENERAL ELECTRIC COMPANY QUESTION 720.29:
In the currently available GESSAR-II PRA, the procedural aspects of human errors have been emphasized, i.e., errors of omission and of commission.
It is now recognized that cognitive behavior can potentially have a dominant contribution to risk.
A single wrong decision based on misdiag-nosis or improper prioritization of tasks can lead to a series of incorrect actions.
Discuss the impact of cognitive errors on the logic trees and on the PRA results by providing sensitivity analysis for the dominant accident sequences.
RESPONSE
The impact of cognitive errors on the progression of accident sequences in BWRs is redu:ed due to the relative simplicity of BWR operation and the responses required of operators in off-normal situations, GESSAR-II control room instrumentation and controls, and the clarity of BWR symptom-oriented Emergency Procedure Guidelines (EPGs).
Section 150.2 of the GESSAR-II submittal highlights the diversity and simplicity of BWR accident prevention and mitigation systems.
Also contained in that section is a discussion of Emergency Procedure Guide-lines developed by General Electric and the Boiling Water Reactor Owner's Group.
The EPGs are symptom-based guidelines as opposed to event-based guidelines, and reduce the probability of improper prioritization of tasks.
The operator does not need to identify what event is occurring in the plant in order to decide on what action to take.
Rather, the operator observes the symptoms (utilizing relatively few instruments) which are occurring, and takes immediate actions based on responding tc those symptoms.
There also exists a further benefit of the EPGs in the event that the operator does initially commit an error and attempt to progress along a path of incorrect actions.
The EPGs contain cautions noted of various points throughout the guidelines that clearly remind the operator of
GENERAL ELECTRIC CHMPANY PROPRIETARY INFORMATION important instructions.
They also contain contingencies for use when the symptoms are not being satisfactorily controlled as a result of actions taken from the main guidelines or as a result of equipment malfunction.
The contingencies include level restoration, rapid RPV depressurization, core cooling without injection, core cooling without level restoration, alternate shutdown cooling, and RPV flooding.
Key in identification of many dominant accidents is the assessment of reactor vessel water level.
Section 15D.2.1.2 provides a brief description of BWR water level measurement which gives direct indication of RPV inventory and core cooling adequacy and thus reduces the probability of misdiagnosis of an inventory-threatening event.
Even in the event that the operator is uncertain of the water level, the EPG contains contingencies which direct him to depressurize the RPV and utilize all available water delivery systems.
O
, +.
)
l l -
4 CENERAL ELECTRIC CCMPANY PROPRIETARY INFORMATION QUESTION 720.30:
In the GE suppression pool decontamination factor model, small gas bubbles were assumed.
In view of the fact that in a saturated pool, coalescence of bubbles and bubble growth may become increasingly dominant, provide a discussion on how sensitive the GE model is to the small gas bubble assumption.
How would the assumption on the shape of the bubble influence the result of the model?
RESPONSE
.)
i l' e
o 4
i 4
F
- 48a -
GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.31:
Justify that the suppression pool decontamination factors will not decrease following a core melt accident; (i) when the suppression pool water starts to' boil off, (ii) a large amount of fission products and corium debris are present in the suppression pool, and (iii) fission products, organo-metallic chelate compounds, and small-sized particulates may be evaporated off into the containment air space.
RESPONSE
(i)
(ii)
(iii)
- 48b -
GENER4L ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.32:
In the suppression pool scrubbing model, there are at least three irnportant parameters contained in the decontamination factor in the exponential term, i.e., particle diameter d, bubble diameter D, and bubble rise velocity V, thus an uncertainty in any of these parameters would result in great changes of the decontamination factor.
Provide the expected uncertainty band for each parameter in the exponential term and discuss the sensitivity of the DFs to these uncertainty bands.
RESPONSE
'o e
A_mA.2 h
A A'
L A
a
-m a
y 49 -
CENERAL ELECTRIC CIMPANY QUESTION 720.33 PROPRIETARY INFORMATION Is the in-reactor pressure vessel DF applied to the entire melt and gap release, or to that fraction of the release corresponding to the fraction 3
of core melt in the MARCH calculation at the point of core 3-slumping?
(Section'5.1).
~
-7
RESPONSE
t.
9
CINERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.34:
The decontamination factors (DFs) for pool scrubbing are sensitive to the particle size.
What experimental and/or theoretical evidence is there for choosing the particle size distribution used? Provide clarification regarding the manner in which the model accounts for changes in DF depending on accident sequence and during the course of an accident (i.e., the time dependence of DF due to changing average aerosol particle size as the larger, heavier particles settle out).
Describe the accident progressions from the standpoint of mechanistic aerosol production and transport to the suppression pool, comparing how you envision aerosol production actually happening to the experiments upon which you establish your particle size and particle size distribution.
(Appendix F.1; Table F.3-3)
RESPONSE
The experimental and/or theoretical sources of particle size distribution information are discussed below.
A)
Particle size distributions for core-concrete aerosols:
y.
t f
I
~
i s
,s
?
\\,
1 B)
In-vessel particle size distributions.
^
i l
l L- ~ ~
.........~. - ~.. _
CENERAL~ ELECTRIC CEMPANY PROPRIETARY INFORMATION RESPONSE TO 720.34:
(Continued)
Descriptions of the accident progression from the standpoint of.
aeorsol formation and release are given in the following reports:
1)
NUREG-0772 2)
NUREG-0956 (DRAFT) 3)
References:
1)
Letter, M. R. Kuhlman, Battelle-Columbus Laboratories to D. M.
Rastler', General Electric Company, July 23, 1981.
2)
Telecon, K. Lee, Battelle-Columbus Laboratories to K. W.
Holtzclaw, General Electric Company, February 27, 1983.
3)
NUREG-0956 (DRAFT), "Radionuclide Release under LWR Specific Accident Conditions."
l 4)
J. M. Otter, " Aerosol Transport Analysis of LWR High Consequence i
Accidents using the HAA-4A Code." Thermal Reactor Safety Conference, Chicago, Illinois, August 29 - September 2, 1982.
5)
NUREG/CR-2182 " Iodine and Noble Gas Distribution and Release Following Station Blackout at Browns Ferry Unit 1."
GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.35:
Is a different DF used when the release is through the quenchers as opposed to a release through the first row of port holes in the drywell wall?
(Appendix F.2).
RESPONSE
Yes, differences in submergence, flow rates, initial bubble size and particle size distribution are the reasons why different DFs are ccicu-lated for the quenchers and the horizontal vents.
O
CENERAL ELECTRIC C2MPANY PROPRIETARY INFORMATION QUESTION 720.36:
Four possible combustion processes are defined:
(Appendix I-1)
(a) During any one event (e.g., local combustion) is the containment volume irtvolved assumed to have a uniform composition of all gaseous components?
(b)
Can any of these four possible combustion processes interact?
For example, can a " global deflagration (involving 60 percent of con-tainment volume) be followed by a " local detonation" (involving 40 percent of containment volume).
(c) How do you know that all important/significant combustion sequences (perhaps a very larg'e number of possibilities) are included in your considerations?
RESPONSE
(a)
(b)
Y g
(C)
-4 &
W
SENERAL ELECTRIC COMPANY QUESTION 720.37:
On Page 15.0.3-798, a characteristic time, t, for the decay of'a deton-g ation waves, peak pressure is defined as
$*2 0 a.
Is this expression valid for a closed system?
b.
Is this expression valid for a closed system of any geometry?
c.
Does this expression take account of pressure loadings everywhere in a closed system?
s.
RESPONSE
a.
b.
C.
s.
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATl!N QUESTION 720.38:
After a period'of steam inertion of the atmosphere, condensation may proceed (homogeneously and heterogeneous 1y) to permit combustible /deton-able compositions to exist somewhere.
(Appendix I.3)
'What assumptions are m5de regarding:
1.
Hydrogen homogeneity during steam condensation.
2.
Steam homogeneity during steam condensation.
3.
Post inertion combustion / detonation.
RESPONSE
Steam condensation is a significant factor (relative to. hydrogen combus-tion) only in the drywell of a Mark III containment.
-\\
1 4'
GENERAL RECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.39:
If a detonation is extinguished as it propagates from a detonable mixture into a non-detonable (but combustible / flammable) mixture, how fast does the leading shock wave decay?
Is such a process considered innocuous?
(e.g., see p. 15.0.3-797).
RESPONSE
2 *.-
- s e
L l
l l
l l.
i 59 -
l-i cy
+
-s--
y
-y_
_m,_.
m
.m
-yw.
GENERAL ELECTRIC COMPANY I
TlON QUESTION 720.40:
Item by item, provide a detailed justification for each of the conditional probabilities tabulated as Tables I.A-1 and I-5 of the GESSAR-II PRA.
RESPONSE
e e
O 96
CENERAL ELECTRIC COMPANY PRUPRIETARY INFORMATICM QUESTION 720.41 In the GE MARCH input for the TQUV sequence (per letter dated 8/25/82),
it was stated that the initial water level in the core was " adjusted" such that core uncovery would occur _at a time consistent with that predicted by GE's SAFE code.
I
.)
If so, why has this considerable steam source been neglected, and does this missing source result in non-conservative estimates of containment loading?
RESPONSE
9 I
73 -
CENERAL ELECTRIC COMPANY PR"iPRIETARY INFORMATION QUESTION 720.42 Inspection of the passive heat absorbing structures used in the GESSAR MARCH analysis reveals th'at the metal containment shell receives heat from the containment atmosphere on both sides of the wall.
This effectively doubles the heat transfer area of the containment walls and appears to be non-conservative.
Explain how you arrive at these input values.
RESPONSE
}
3 h
1'
/ _-.
OtifRAL ELECTRIC COMPANY PR2PRIETARY INFORMATION QUESTION 720.43 Scaling of the hydrodynamic processes governing gas flow into the suppression pool by way of the SRVs and vent pipes.is reasonably well understood.
Please provide a scaling analysis that demonstrates that to the conditions for the scrubbing experiments are indicative of the hydro-dynamic conditions anticipated in the prototype.
Include therein a discussion of how the effect of surface tension is a.
scaled so that bubble break-up is properly accounted for.
b.
Once the bubble sizes are rationalized, pool depth and terminal velocities of single bubbles and swarms of bubbles must be considered.
Provide a discussion of the scaling considerations employed for the test facility that acc'ount for the pool height to bubble velocities ratio time scale.
Bubbles break through a surface by a complex process that creates c.
small liquid droplets that are thrown upward.
The amount of entrained liquid be a function of the number of bubbles and their sizes.
Scale will play an important role here also; please discuss.
RESPONSE
"r l
l I
l CENERAL ELECTRIC COMPANY PREPRIETARY INFORMAfl3N QUESTION 720.44 The-DF prediction focuses on iodine present as Csl associated with large particles.
What would be the effect of assuming some elemental iodine or organic iodine? What would be the potential for formation of organic iodine in the drywell? To what extent would elemental and organic iodine forms limit decontamination factors?
RESPONSE
The DFs for elemental iodine are comparable to the DF's for iodide in suppression pools (Reference NEDO-22213 " Analytical Models for Fission Product Scrubbing in a Water Pool").
t I
l l
l l
l CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.45:
What shape factor should be used to characterize the Eu,0 in the depletions g
calculation?
Please provide justification for your cont 10sions.
RESPONSE
1 s
S S
4 a '
GENERAL ELECTRIC COM?ANY PRCPRIETARY INFORMATION QUESTION 720.46:
Considering the sensitivity of DF te particle sizes, the determination of an average size of 4.1 p cannot be considered close agreement with the stated "1.87 to 3.1_ p determined by the Quantamet." Which of these values is close to the actual expected value, i.e., a better represent-ation of reality.
Which one did GE use?
How does what GE used compare with either of these values?
RESPONSE
J s
=
1 9,
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.47:
Provide examples of the scanning electron microscopic pictures referred to on Page 49-C33.
RESPONSE
The pictures are attached.
i l
l 6
l 4
l l
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.48:
'What effects do deposition and reentrainment.have on the particles as they actually enter the pool, compared to measurements made at other times or places.
- RESPONSE:
'T' E
W I
l l
l I
i l
l I
84 -
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.49:
Show a copy of Figure 1-2, and discuss, the effect on the experiment of the diluter mentioned on Page 49-C34.
RESPONSE
h e
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.50:
What are the length to diameter ratios for all the sampling lines? What effect or modification will this factor have on the measured size of particles? Will there be any appreciable expected tendency to deposit for lines of large 1/d?
RESPONSE
4 e
3 r
e 6
t 1
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION
\\>
QUESTION 720.51:
Page 49-C33 discusses two impact samplers.
Figure 1-2 shows three (before the pool, above the pool, and after the recirculation line).
Which two are meant to be referenced in the test? What does the third one sample?
RESPONSE
There is an error in the figure 15DA.1-2 and only two impactors were used (see answer to Question 720.49). _
~
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.52:
The last line of Page 49-C35 states that.a "high flow recycle stream" kept particles in suspension.
What was the magnitude of the flow in cfm, and what velocities existed in the recycle circuit?
RESPONSE
4 h
CENERAL FLECTRIC COMPANY i
PRUPRIETARY INFORMATION QUESTION 720.53:
Tables 15DA.1-1 and -2 give what seems to be a calibration for the impactors used.
Is this what they are?
How are the particle diameters in the table defined? Give the equations used in the calculation and a reference therefore.
Which 2 of the 3 impactors are referenced in the tables? Are the calibration conditions typical of the flow rates in the actual experiments?
RESPONSE
A 5
.k
/,
C
4..
CENERAL ELECTRIC COMPANY
'/
PROPRIETARY INFORMATION QUESTION 720.54:
There was in the presentation by a GE representative to the American Chemical Society in Kansas City in September, 1982, a statement that the impactor at the top of the tank may have modified the particle size.
Is this GE's position?
If so, why might this same condition not have occurred on either of the other 2 samplers? How would the comparison of the experiment with the model be changed?
RESPONSE
The statement at the meeting in question was an attempt to explain the low experimental value shown in a slide in relation to the calculated value for large particles.
The examination of a SEM photo of the 1st stage showed only small particles not particles greater than 3 pm as was expected.
The statement was "if small particles were collected, even a
'small' number, it could cause a lower DF than the 'true'" value resulting from only large particles, i.e. >2 pm, being collected.
In other words there was cross-contamination.
4 6
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.55:
The paragraph at the top of Page 49 C36 seems to indicate that all starting and final locations of Eu 0, were sampled.
This should allow a 2
mass balance to be performed.
Did GE do this?
If so, what are the resul ts?, If not, what places remained unaccounted for?
RESPONSE
fa[1 e
9 i
91 -
GENERAL ELECTRIC COMPANY l
PROPRIETARY INFORIAATION QUESTION 720.-56:
On Page 49-C37 an " entrance effect" is discussed.
What is your definition of an entrance effect? How was it calculated?
Is it a function of particle size? Give a reference.
How are values given in Figure 15 DA.1-3 (curves or data) modified for this effect?
l RE3PONSE:
I
- j a
<A
.4 I
l l
l lI.
l !
l l
GENERAL ELECTRIC COMPANY PREPRIETARY INFORMATION QUESTION 720.57:
The same page refers to the particle size distribution in Table 15 DA.1-5.
The teble purports to contain fractions of mass of Eu 0 vs 2
average particle size.
The mass fractions do not add to unity. 3Wnat is the particle size distribution? Considering the extreme sensitivity of DF to particle size, are the bins of particle sizes in that table suffi-ciently small so as not to cause uncertainty in DF assumptions? Give sample calculations.
Since only one size distribution is given, is it correct to assume that all the many experiments had exactly the same size distribution? Were the distributions not measured by the impact samplers in every experiment?
RESPONSE
d o
i 2, J l
e
SENERAL ELECTRIC C2MPANY PROPRIETARY INFORMATION QUESTION 720.58:
How do you get the correct diameter to calculate the Cunningham slip factor 'and the diffusivity, if an assumed value is input for the density?
How much uncertainty can be introduced in the calculated DF as a result?
RESPONSE
The Cunningham slip factor (CSF) depends on the particle size, apparent mean free path, and particle density.
J
. J a
e
- CENERAL ELECTRIC COMPANY PR3PRIETARY INFORMATION
,e QUESTION 720.59:
}L/
In Paragraph (2) on Page 49-C38, the statement was made that the experi-mental results exhibited the trend of DF versus particle size given by the model.
No data are given which would allow this to be reviewed.
Provide the data and the comparison.
RESPONSE
The data are given in the attached figure as well as the theoretical model curve.
\\,
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION 4
5 10 o
e O
O V
0 0
e I
~
THEORETICAL MODEL 0
v D
e O
O 8
3 5 10
- 5 g
E oD 3
v z
v v N
D E
2 o 20 g
Q O
O o
EQUIVALENT
$PHERICAL BU88LE TEMPE RATURE DIAMETER (AIR. WATER)
TEST NO.
(cm)
(*Cl 10 e
9 0.45 AMBIENT C
10 0.86 AMBIENT v
12 1.35 AMBIENT D
14 0.88 AMBIENT C.
15 0.as tomo 18 0.90 AMBIENT e
i i
e g
0 E5 1.0 1.5 2.0 2.5 3.0 3.5 PARTICLE DI AMETER (p)
Figure :
Decontamination Factor for Eu 0 Particles as a 23 Function of Particle Diameter at 168 cm Water
( 0.85 cm bubble diameter is used in model cale :lation)
O CENTRAL ELECTRIC COMPANY QUESTION 720.60:
Paragraph (6) on 49-C40 discusses the water as a perfect sink.
The statement is made that water will absolutely absorb the particle (Emphasis added).
Provide references of supporting data for this absolute statement.
RESPONSE
The aerosol particles released from the core under accident conditions are formed'in a high' temperature steam saturated environment.
As the steam and particles move to colder areas, condensation of fission product vapors, as well as steam may occur on the structural aerosol surfaces (NUREG-0772, Section 7.3).
Thus, the " wet" particles, which may contain solubla fission products, should be easily retained in water (NUREG-0772, Section 6.4).
The agreement between the particle scrubbing test data and the calcula-tional results validate the assumption of particle absorptions by the water used in the model for wettable particles.
This demonstrates that even the dry (but wettable) Eu,03 particles-are nearly quantitatively absorbed.
Even the Fuchs clastical model assumes the liquid is a perfect sink.
D.
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATICN OUESTION 720.61:
Paragraph (7) on Page 49-C41 states that super-heated steam could play an important role in promoting particle growth.
Discuss the mechanism by which this takes place.
Provide references or other supporting information.
RESPONSE
~, ]
a The attached figure is from the book by F.
F.
Cinkotai, "The Behavior of Sodium Chloride Particles in Moist Air," Department of Occupational Health, University of Manchester, England.
~
l l
l
f y./
100.2%
100 1%
100 8% -
11M) 102%
f 00m 105%
10i% P.
T l
I l
120%
M8 10 E*
g
~5 l
E na t
/
59 7
I s
,,3 lt
)
8E 5
P g
E5 f
/
EE 3
/
n dn lC 2
I nsi :
i i
0 001 0 01 01 1
to Os AMETEm or PumE NaC: PamTICLE tum Tigure The Ratio of the Diameter of Nacl Solution Droplet to the Diameter of the Nacl Particle Trom Which the Droplet Has Been Formed at Various Levels of Relative Humidities (Ref. 8.4)
- 100 -
q GENERAL ELECTRIC COMPANY PR2PRIETARY INFORMATION QUESTION 720.62:
Paragraph (8) on page 49-C41 states that the scrubbing factors are conservative from a temperature standpoint because thermophoresis was
~
neglected.
Thermophoresis would, if calc'ulated, increase the DF.
However, there is an effect in the opposite direction, diffusophoresis.
This effect may be larger than thermophoresis.
Show why the DFs should be considered conservative.
RESPONSE
Diffusophoresis refers to the phenomenon of vaporization at the liquid / gas interface which tends to impede aerosol flow towards the interface.
The discharge of superheated steam into a postulated water pool at saturation conditions could result in bubbles with inward vapor flow from the walls, preventing the scrubbing of aerosols.
~
}
.f
- 101 -
(-
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATiON QUESTION 720.63:
In Table 15DA.1-4, data are given for tests on 12/11, 12/14, and 12/15.
Given GE's model, these tests would all be expected to give the same results.
There is over a factor of 4 difference in the results, however.
Does this represent scatter in the data?
Explain.
RESPONSE
If the data table is examined it is seen tnat all three tests had different particle concentrations and particle size distribution.
These differences and the scatter of the data are probably responsible for the difference observed in the DF.
Data in the table was incorrectly typed, overall DF for 12/11, 12/14, and 12/15 should be 2945, 2270, 928, and for the test on 12/9 the DF should be 4157 not 1260.
See attached table.
- 102 -
i GENERAL ELECTRIC GENERAL ELECTer, enu.puy i
NUCLEAR ENGI ING DIVISION PROPRIETARY INFORMATION l
i' SU W RY OF TEST RESULTS Water Test Bubble Orifice /
Bubble Rate Particle Height Overall Date Diam,cm Cap B/M Cone (g/m3)
(cm)
D.F.+
Eu2 3 - MilIipore 0
9/16 0.47 70/none 1 83 0.18 34.3 10 8 9/29 0.63 70/0.6 145 0.17 34.3 333 9/30 0.63 70/0.6 145-0.44 34.3 214 10/1 0.60 70/0.6 277 0.02 34.3 119 10/8 0.74 130/0.8 254 0.53 34.3 189 10/27
- 0. 85 1 80/0. 8 312 0.48 167.7 1170 10/28
- 0. 85 1 80/0. 8 318 0.91 167.7 1415 10/30 0.45 1 80/0. 2 248
- 0. 87 167.7 1251 11/3 0.45 1 80/0. 2 48 0.48 167.7 719 Eu2 3 - Impactors 0
12/1
- 0. 86 1 80/0.7 260 0.76 167.7 896 12/2
- 0. 86 1 80/0.7 260 5.5 167.7 1260 12/8 1.41 Special++
124 0.30 167.7 534 12/9 1.35 Special "
140
- 1. 8 167.7 MW57 12/10 0.78 1 80/0.7 248
- 4. 95 76.2 910 12/11 0.88 240/0.7 276 4.34 167.7 ef37 MW 12/14 0.88*
240/0.7 272 1.38 167.7 2270 12/15 0.88 240/0.7 304
- N.D.
167.7 928 I
l Elevated temperature test, 60'C Water /60'C Gas
" Not Determined
+ D.F. = total activi+y injected / activity in aerosol l
++ Cap modified to he,e a flat surf ace which allowed bubble to grow in volume before breaking off.
- 104 -
CENERAL ELECTRIC COMPANY j
PROPRIETARY INFORMATION QUESTION 720.64:
p Provide justification for the statement on page 49-C43 that the large bubble shatters within about one bubble radius, especially considering the statements on page 49-C45, 2nd paragraph.
In the justification, consider especially problems of scale.
RESPONSE
The major shattering of a large bubble appears to occur when the bottom surface overtakes the top surface.
Numerical computations show that this occurs in about one initial bubble radius of upward motion.
A simple analysis in which the bottom bubble surface velocity exceeds the top surface velocity by the Bernoulli head indicates the same result.
This phenomenon corresponds to Froude scaling, and therefore will be geometrically similar at any scale before surface tension becomes important.
- 104 -
)
GENERAL ELECTRIC COMPANY QUESTION 720.65:
Justify the submergence of 3 and 5 feet used in the experiment from the point of view of scale.
What are the minimum, maximum, and average submergence values of the horizontal vents in the within plant case?
RESPONSE
The diameter and velocity of rising stable bubbles are essentially the same in small and full scales.
Therefore, to make the tests full scale, the vent submergence would have to correspond to the full size values.
Since the relationship between submergence and scrubbing is well known, the extrapolation of data from 3 or 5 feet to 14 feet is straight-forward.
The range of submergence for the horizontal vents taking into account the water sources appropriate for each accident sequence (upper pool dump, condensate storage tank, primary coolant inventory) is 8.4 to 13.5 feet with the majority of events having 13.5 feet.
6 e
- 105 -
GENERAL ELECTRIC COMPANY PRSPRIETARY INFORMATION QUESTION 720.66:
i For Figure 15DA.2-3, what is the basis for the solid curve?
It does not appear to be a "best fit" to the data.
Was the parameter bubble rise velocity as a function of flow rate used in the analysis?
If so, please present the values used and justify.
Is the "Dubble rise velocity" really the swarm velocity, or is it measured for the first 1-3 bubble radii?
RESPONSE
The solid curve in Figure 15DA.2-3 is simply an " eye balled" curve which had to begin at I ft per second for a single bubble and ends at the data average.
It is drawn to show a trend in the data, nothing more.
- 106 -
n GENERAL ELECTRIC COMPANY s
PROPRIETARY INFORMATION I
QUESTION 720.67:
In Equation (7) on page 49-C50, should there not be a factor for accele' ration due to gravity reflection of Taylor instability theory over the range of wave length possible?
Further, this equation is not appli-cable to determination of a stability threshold as implied in the last sentence of that paragraph; please discuss.
RESPONSE
4 The first unstable wave length from Taylor instablity is:
A = 2ndah u
longer wave lengths are also unstable.
Equation 7 gives the fastest growing wave length as /f A '
u
- 108 -
{'s QUESTION 720.68:
Charge of the particles, due for instance to B decays during the transit of the pool, has not been evaluated as a difference between the tests and actual accidents.
Discuss.
RESPONSE
The Europium oxide used in the scrubbing tests was radioactive as the fission products would be from reactor accidents.
The Eu-152m undergoes beta d*scay with a 9.2 hour2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> half-life.
Therefore, no difference would be expected due to beta decay between accidents and the tests.
GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION k
- 108 -
QUESTION 720.69:
Entrainmeni from the pool has been neglected.
Justify.
RESPONSE
J
.A "l
j-
?
i GENERAL ELECTRIC COfAPANY PROPRIETARY INFORMATlON
- 110 -
1
\\
i QUESTION 720.70:
i We understand that some experiments were performed with Cs!.
Is this true?
If so, provide the data and their evaluation.
RESPONSE
\\ ?
s e
e p
L 9
I GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION f'
- 110 -
.__L QUESTION 720.71 Justify not considering the evolution of iodine from the pool due to such processes as radiolysis.
RESPONSE
4 l
1 e
t-
/
i l
l I
l '
\\
I GENERAL ELECTRIC COMPANY PROPRlHARY INFORMATION 1
111 -
.. _ _ =
\\ s/
QUESTION 720.72 GE's model does not appear to differentiate between bubble rise velocity and swarm rise velocity.
We believe this disti'nction to be an important one, in that it has an effect on calculated DF values.
Please clarify 8
the terms used for diffusion and inertial removal, and justify the velocity used.
RESPONSE
Bubble rise velocity relative to the water is designated by U which largel" determines the inertial scrubbing.
Bubble swarm velocity is desip td by V, which determines how long the bubbles are in the pool befc.,
eaking the surface.
The expressions on pages 230-A38 and 230-A39 are in error and should be replaced by:
k = pV2d2C, and k = 1.8 OU n
d f
D2p V i
R3V2 g
The velocity of individual bubbles relative to the liquid is about 30 cm/sec and the velocity of a large bubble swarm can be as high as 150 cm/sec.
GENERAL ELECTRIC COMPANY PROPRIETARY !NFORMATiON
- 112 -
TABLE 1 COMPARISON OF RISK VALUES FOR VARIOUS STUDIES
' Total Latent-Early Fathlities Fatalities per Per Reactor Reactor Year Year
't GENERA'_ ELECTRIC 00.
PROPRIETARY INFORMATION J
116
1 GENERAL ELECTRIC COMPANY QUESTION 720.73 After the change to the central estimate dose model (letter dated
. July 16, 1982) the comparison with WASH-1400 composite site and GE.
calculation of site 6 show large factors of disparity (see Table 7.2-1).
Does GE still.wish to justify site 6 as an average site?
If so, provide the justification.
If not, state the types of sites for which the PRA will be applicable.
RESPONSE
ed f
,._.-,__-.-.m.
~
. E m a R ei mm 2M NUCLEAR ISLAND Rey,11 CENERAL ELECTRIC COMPANY l
P,.OPRIETARY WFERMATION Class til 4
10 d
10 4
10 x
A E*
8 U<
U 4
10 n
t E
5
=
Y 10*7 4
10 10 0
1 2
3 5
10 10 10 10 10" 10 E ARLY F ATALITIES lxl 22H3 en Figure f.41. Site Comparison of Estly fatality Curves for WASH 1400 BWR 15.D.3 588 w
7
-p-.
w-
..-....~,,,,_,..,_.,----,.-
,.y.-
.-,---,_--^-t-
- * -r-
---*-e-*-
er et
-w--*
23 NU WCISLAN3 M
C ENERAL ELECTRIC COMPANY PROP lETARY INFCRMATION Class ill 10'3 10 m
i i
10-5
-x A
E b
t 2u 4
CT 10 E
0 2
- 6 E
m 10 t
[
l l
I 10 i
a a 1
,g 0
1 2
3 5
10 10 10 10 to" 10 LATENT F ATALITIES PER YE AR (X) 22343 C?
Figure F.4 2. Site Comparison of Latent Fatality Curns for WASH.1400 BWR 15.D.3 589/15.D.3 590
3 17 V
d
.,o u
A to**!
~
h e
vg wAsw 1400 swald AT stTE 6 s
t
-Central Estimate Model-10 %
e i
E g
BWR/6 at Site 6 1o'3
~
.-Linear Model-e E
I I*
BWR/6 at Site 6
-Central Est imate Model-
,n s 9
i
. i i
i 0
7 3
5 10 io' 10 10 to' 10 LATENT F ATALITIES PER YE AR (Xi Figure 3. A Comparison of the Linear and Central Estimate Models for BWR/6 at Site 6.
GEN ERAL ELECTR C CO.
PROPRIETAR.Y INFORMATION 0
'^
GENERAL ELECTRIC COMPANY QUESTION 720.74 PROPRIETARY INFORMATION Since GE' expects that the particle size distribution of a core aerosol will be significantly modified by passage through the pool-(due to orders of. magnitude differences in GE's DF versus particle sizes), provide a review of dose conversion factors and expected consequences, considering that penetrating aero'sols will be preferentially emitted.
RESPONSE
4 96 3.
,,s (s'
k.
ed 117 -
~.,.
1*2
. gi g
>l ig ij il il
\\
/
7 Floor deposition I
x y
//
\\
w.n on.o.iiion
/
\\
/./
j '*',
~
\\
.0m 10
1 1
10
Ceiline dopo.stien\\
7 10'8 7
,,10
,i iI
.I il it i
i 0
10 5 10'3 10*I 10 Particle Diameter (microns)
FIGURE VI B3 Comparison of deposition velocities to smoother floor, wall, and ceiling surfaces for u* = 34.1 cm/sec 0.004 cm; nominal velocity of 670 cm/sec.
and yo =hmel (1973). Reprinted by permission from From Se Pergamon Press.
B-6 e
/
\\
i i
i p g g-
/
s s
\\\\
Nanopa,*gra.es oa p
a
's i
e1S 1'
- v..c.
.ca..
6 l2 i
4 5
l i3 4 eS 1
i l
- . -.ew, es e 02
\\\\
\\ s 01 I
5 10 20 30 SO 70 80 90 95 99 9entras Decos+oa FIGURE VI D 1 Deposition model. The radioactive or mass fraction of an aerosol that is deposited in the nasopharyngeal, tracheobronchial, and pulmonary regions is given in rela-tion to the activity of mass median aerodynamic diameter
-(AMAD or MMAD) of the aerosol distribution. The model is intended for use with aerosol distributions that have an AMAD or MMAD between 0.2 and 10 microns with geometric.
standard devisions of less than 4.5.
Provisional deposi-tion estimates further extending the size range are given by the broken lines. For the unusual distribution having an AMAD or MMAD greater than 20 microns, complete nasopharyngeal deposition can be assumed. The model does not apply to aerosols with AMADs or MMADs below 0.1 micron.
D-4
QUESTION 720.75:
Explain the influence of different event sequences on the estimate of operator reliabilities for similar actions, (e.g., for actions such as
" operator manually opens the ADS upon the failure of control circuits,"
describe the methods evaluating human reliability in both a loss of off-site power sequence and an ATWS sequenc'e).
Also explain which human actions are considered independent within the dominant accident sequences given in GESSAR-II PRA.
Provide, also, the basis for deciding-the degree of dependency between individual sequence, and the basis for dependencies assessed between members of a team for an action.
RESPONSE
',t
\\
I GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION
- 119 -
l
/
QUESTION 720.76:
Provide in a tabular form showing how the human factor analysis was documented in GESSAR-II PRA, e.g., Figure 4-19, page 4-58 of NUREG/CR-2300.
RESPONSE
Additional information on human factor analysis in the GESSAR-II PRA was provided in presentations in Technical Update Meeting #5 on October 19, 1982.
The proposed revisions to Appendices C and D to 150.3 provide all the information on how the human reliability was quantified and incorporated in the PRA.
If, following review of the proposed revised Appendices C and D, additional information on human reliability is required, it can be provided.
i T
4 GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION
- 120 -
QUESTION 720.77:
List the key uncertainties in modeling that were addressed.
Indicate:
(a) How were these identified (e.g., literature survey, sensitivity
^
studies);
s~
(b) Were any uncertainties treated by making conservative assumptions; (c) What quantitative measures were used for the modeling uncertainties and what techniques were,used to determine them (e.g., response surface models, judgment applied to sensitivity studies)?
(d) How are these quantitative measures to be interpreted (e.g.,
stochastic variation of physical processes,' expert opinion about likelihood of various options)?
QUESTION 720.78:
Provide a brief description of how uncertainties were propagated through 7
'the analysis.
Include in the discussion:
i.
(a) For which parts of the PRA (e.g., systems analysis, containment response analysis, in plant consequences and ex plant consequence analyses) uncertainties were propagated.
I
{
(b) How uncertainties in different parts of the analysis were combined.
1 (c) A description of any computer codes used and how uncertainties related to the codes were treated.
(d) A description of the special features of the analysis (e.g.,
correlation between parameter uncertainties for like components).
- 121 -
e QUESTION 720.79:
Unresolved Safety Issues (USI), applicable to BWRs, should be evaluated under GESSAR-II PRA.
Provide a list of USIs not evaluated and the bases for their exclusion from GESSAR-II.
RESPONSE
Appendix IB to GESSAR II provides an assessment of unresolved safety issues applicable to the 238 Nuclear Island.
Appendix 1B was submitted as a portion of Revision 4 to GESSAR II.
r-GENER
- t ELECTRIC COM*/ MY pro?..IARY INFORMai O '
- 124 -
GENERAL ELECTRIC COMPANY QUESTION 720.80 PROPRIETARY INFORIAATION The discussion (p. 15D.3-569) of DF's assumed for plugging of drywell or containment cracks states that the values used ranged from 1.0 to infinity.
Please be a little more precise concerning the values used.
Discuss the crack size and particle size assumed.
Provide a basis for your assumptions and discuss the applicability of the Morewitz model.
(Note that the results of the Marviken containment tests (1974) directly contradict the Morewitz model predictions).
Discuss the significance of other leakage paths bypassing the suppression pool in this context.
RESPONSE
l
+
e l
- 125 -
CENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.81:
The PRA consequence calculations are purported to be realistic, or somewhat conservative.
The evacuation delay assumption for the CRAC analyses, however, is that full-scale mass evacuation.can be accomplished instantaneously.
This is neither realistic nor conservative.
Please discuss the effects of a realistic (non-zero) estimate of evacuation delay times.
RESPONSE
- h f
h I
f l
l i
f
- 126 -
L
CENERAL ELECTRIC COMPANY PROPRIETARY INFORA% TION V
QUESTION 720.82:
In Section 150.4.2, you discuss GESSAR-II relative to rules and propnsed rules which consider severe accidents.
Because conformance to these rules is being considered as necessary requirements in the severe-accident rulemaking (SECY-82-1A), it is important to have a clear understanding of the areas in which GESSAR-II is in conformance as well as the areas in which it is not.
Specific areas where GESSAR-II appears not to conform with these rules are:
(1) GESSAR-II has no provision for hydrogen control (the CP/ML rule
- requires preliminary design information on hydrogen control).
(2) GESSAR-II has no provision for a blanked-off three-foot equivalent containment penetration for possible use in a containment-vent system or containment heat removal system.
(The CP/ML rule requires such a penetration.)
(3) GESSAR-II does not meet the service-level C capability of 45 psig for the primary containment as specified in the CP/ML rule (although on page 150.4-9, it states that the ultimate pressure capability significantly exceeds 45 psig).
Provide confirmation of these apparent non-conformance items or corrections to the staff's interpretation of Section 150.4.
Specifically, provide analyses or appropriate references to the analyses which demonstrate meeting the 45 psig Level C requirement, if this te the case.
RESPONSE
Section 150.4.2 provides a discussion of GESSAR-II relative to rules and proposed rules which consider severe accidents.
One rule that is being
~proposed in SECY-82-1A in which the GESSAR-II design appears to be in nonconformance is the CP/ML rule (FR 1/15/82 and 2/1/82) which was
- 127 -
t
RESPONSE TO QUESTION 720.82: (Continued) considered necessary for issuance of construction permits immediately after TMI.
Given the results from extensive degraded core research and probabilistic risk assessment evaluations that have occurred since the CP/ML Rule was promulgated, it appears that a basis may exist for alternative considerations.
Specifically, the conservative requirements of the CP/ML rule may required reevaluation for application to standard plant licenses.
The GESSAR-II design has the capability for hydrogen control consistent with current requirements.
As noted in Section 15D.4.2-2, it is General Electric's position that additional hydrogen control is not required to..
maintain containment functions nor is it cost beneficial.
The same section of the GESSAR-II submittal discusses the issues of the additional containment penetration and service-level C capability.
Tables G.1-1 and G.1-2 provide level C capabilities for the containment and drywell structures.
The level C capability of the containment ranges from 30 psig for the dome to 63 psig for the cylindrical shell.
The drywell capability ranges from 96 psig (drywell roof slab ender water) to 272 psig for the drywell wall.
The ultimate pressure capability of the GESSAR-II containment has been calculated to be 58 psig.
The analysis is contained in Appendix G of Section 150.3.
The response to Question 720.83 provides additional information on cost-benefit analyses of additional hydrogen control systems, a stronger containment design and venting.
GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION i
- 128 -
l
w GENERAL ELECTRIC COMPANY ONUAW MMM QUESTION 720.83 In Section 150.4, you discuss, in a very general way, addition mitigation features, the small safety benefits (risk reduction) that would result from these features, and why you believe such features are not needed for GESSAR-II.
We believe additional information is required in this area.
In particular:
1)
Provide the analyses which show that risk reduction from hydrogen control is less than 30%.
In your submittal, you should discuss (i) how you considered uncertainties, (ii) the functional requirements that you imposed on the hydrogen control system, and (iii) the risk measures used.
2)
Provide the analyses which allowed you to conclude that a stronger primary containment system will "... not significantly reduce risk due to severe accidents".
Consider as a variation on this, the provision of a primary system vent or filtered vent as an alternative to a stronger containment.
3)
Provide an estimate of the time to basemat penetration, the corium composition at the time of penetration (fission products inc.) and the amount of water (if any) that will be released from the containment together with the fission product inventory in this water.
Why have you not considered core-retention materials as a replacement for portions of the basemat?
If your answer is in terms of risk reduction, provide the analyses that lead to those values.
RESPONSE
r
, t>
\\,., r l
e l
- 129 -
I
o T#
3)
Based on studies done for the Rogovin Report on TMI-2, it was concluded that the basemat ($12-15 feet) would not be penetrated before the corium pool had enough surface area to reject the decay heat and begin solidifying.
Core-retention devices would only allow a longer time to loss of containment integrity for those sequences in which noncondensible gas generation is assumed to overpressurize the primary containment or they could increase accident consequences by forcing h' eat upwards and damaging the drywell upper structures.
Delaying loss of containment integrity 8-10 hours after accident initiation to a later time provides minimal risk reduction.
See response to parts (1) and (2).
9 i
GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION
- 131 -
GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION QUESTION 720.84 The core power used in MARCH calculations for the ATWS sequence has been given as about 15% (p. 3-542).
Prior to ieveling off at this power level, there is a power surge up to 570% Po.
Although this short durath a power rise may not be important in considering long-term effects, it may be important in determining initial SRV discharge, vessel water inventory and vessel water level.
Since MARCH does not model this power rise, it is not clear if MARCH has been used in a manner that._
~
consistently treats SRV discharge and vessel water inventory.
Is the SRV discharge and water level adequately assessed using MARCH?
RESPONSE
9
.e l
l i
I l
\\
(
- 132 -
l
' GENERAL ELECTRIC COMPANY PROPRIETARY INFOR!,MTION QUESTION 720.85(a)
Provide a list of primary and drywell containment electric penetration assemblies (CEPA's) in the GESSAR-II plant, indicating the location and limiting survivable environmental conditions such as pressure, temperature, radiation, hydrogen, and humidity, corresponding to the dominant severe accident sequences.
RESPONSE
The primary containment electrical penetrations are listed in Table 6.2-29 (pages 6.2-213 through 216 of GESSAR-II) and the locations of these penetrations are given in Figure 3.8-4 (page 3.8-161) and Figure 3.3-12 (page 8.3-165).
3
.\\. '
0
- 133 -
1 1
QUESTION 720.85(b)
Describe the type of termination used on the GESSAR-II plant, such as terminal blocks, crimping lugs, and junction boxes, and indicate the limiting survivable environmental conditions as given above.
RESPONSE
Electrical terminations within the drywell are made by various means including binding-screw terminals, crimped connectors (or lug), bolted lug and multiple pin connectors.
S
-g.
.e GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION 135 -
L
)
GENERAt ELECTRIC COMPANY 1
PROPRIETARY INFORMATION QUESTION 720.85(c)
Provide justification that electrical penetrations through the drywell or the primary containment would be able to maintain their electrical as well as mechanical integrity when subjected to prolonged multi phase short-circuiting and beyond-DBA ambient conditions during the core melt accidents.
In an event of local hydrogen combustion or detonation, provide a discussion on the likelihood of penetration material failure that may result in breaching of the respective compartment.
RESPONSE
The electrical penetrations through the drywell and primary containment would not be subjected to prolonged multi phase short-circuiting because they are protected by redundant over-current devices as required by R.G. 1.63.
The circuit protective devices are located outside of the primary containment and so are not subject to severe environmental consequences.
l 1
Response 720.85(a) above discusses the survival of drywell and containment penetration materials at ambient conditions associated with the postulated severe accident consequences.
The construction of the drywell and containment penetrations is such that gross mechanical integrity will be maintained (except under prolonged hydrogen burning conditions) because of the following design features:
1.
The drywell electrical penetrations are constructed of explosion proof hardware (threaded rigid steel conduit and condulets).
2.
The containment penntrations have standard 150 psi steel flanges drilled for cables and potted with material that is similar to that which has previously been used at temperatures of >280*F and pressures exceeding 60 psid.
The arrangement of floors, equipment and penetrations is such that hydrogen combustion would not occur in the vicinity of most of the
- 136 -
drywell penetration seals.
However, there are a few small diameter penetrations which are in vulnerable locations (Azimuth 200* to 220*)
where H burning may be of sufficient duration to damage the cable 2
insulation within the " CHICO A" compound seal region and might also damage the steel conduit.
It is estimated that the insulation within the
" CHICO A" seal may,not survive a hydrogen burn of 30 minutes.
In the of seal failure, the leakage through the drywell wall would be case limited by the leakage paths through the conduits and fittings enclosing the cables within the drywell and the restrictions afforted by the cables in these conduits through the drywell wall.
GENERAL ELECTRIC COMPANY
(
PROPRIETARY INFORMATION
- 137 -
L
GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION
. QUESTION 720.86 Following a reactor scram in the GESSAR-II plant, the scram discharge volume (SDV) system becomes the reactor coolant retaining boundary outside the primary containment.
In the event of a pipe break downstream of the scram outlet valves and upstream of the SDV system vent or drain valves, any reactor coolant system blowdown will not be terminated by the automatic closure of the vent or drainline isolation valves because these valves are located downstream of the break location.
In such an event, closure of all scram outlet valves would be the only available option to isolate the system and to prevent any release of fission products outside the primary containment.
The successful closure of all scram outlet valves, however, would depend on the operator to manually reset the,
reactor protection system, override any trip signals, the availability of AC power, and to start the motors for closing the valves, in addition to control air supply and the functioning of electric motors and control circuits which are both non-Class 1E, i.e., not have been qualified for DBA conditions.
The non-Class IE equipment incorporates electrical insulating materials which could undergo deterioration under normal plant operating conditions such as the lowering of resistivity.
Provide an analysis to show the probability for the successful closure of all scram outlet valves in such an event.
Provide a list of non-Class 1E equipment which would lead to common-cause failure and result in the unavailability of other safety-related equipment in a severe accident.
RESPONSE
O a
?'
\\
i
- 138 -
TITLE:
Effost af Rydroges Contral ca tho Risk from Ssvoro Assidacts in
}
the BWR/6 Mark III Standard Flaat i
AUTHORS:
D. A. Esakins, I. W. Moltselaw General Electrie Company San Jose, California Introdnetion One of the key issues identified by the NRC related to severe accidents has been hydrogen oomtrol.
In the motice of Proposed Estemaking on Interim Reguiroseats Related to Eydrogen Control in December,1931, the NRC has 4
proposed addittomal hydrogen sontrol systems for the BWR/6 - Mark III design to accommodate hydrogen release from posts 1sted severe accidents.
The proposed sale vos1d require applicants to demonstrate maintensac of containment integrity for events which release sa maomat of hydrogen 7
ogstvalent to 75% metal-water reaction of the active f uel e1 adding.
4 This quantity of hydrogen saa only be generated from a severely degraded sore and most probably as accident with substantial core melt.
Core meltdown sesidents may lead to core-concrete lateraction and loss of sostainment integrity by the generation of non-condensible gases each as earbon monoaide and serbon dioside.
Precinding hydrogen combustion in the aseident segnance will not prevent eventsal loss of containment integrity.
Overpressurisation of containment by concrete decomposition products saa j
oscar la all carrest light water reactor sostalament designs.
This paper discusses the potential risk redaction for preclading hydrogea soebastion la a Mark III sostainment. The potential risk redsstion is minimal. because the Mark III sostainment is designed to mitigate the t
effects of loss of containment lategrity eassed by either hydrogen l
combastion or somerete decomposittoa.
The BTR/6 Standard F1 sat Probabilistic Risk Assessment II(PRA) sonsidered hydrogen generation for severe assidents. The FRA gesatified the son-seguesses of hydroses soebastion events taking assomat of the structural espability of the drywell and esppressloa pool to assare oorshblag of potential fission prodset releases even for cases with loss of primary sostainment latogrity. Fission prodnet retention by suppression pool serabbing means that the sostainment feastion (limiting of fsite doses) is malatelmed even for severe assidents.
Osamtification of the fission prodset sarabblag espability of the suppression poet daring egg 9re assidents was based on GE's Fission Fredsat Serebbing Program A first primelples analytteel model was developed to describe fission prodnet serebbing in the esppression pool.
Raperimental verification of this model was obtained by mass-tr aster and i
hydrodynaale testias. This model predicts that the esppreesine pool vos1d I
rednee parttemiste fisslos prodnet releases by a factor of 10, br1 la the I
salikely event of a severe assident. These resalts confirm that the BWR soppression pool weste ef fectively retain fission prodsets released during
[
severe sealdents.
i l
I l
I i
1 h
I
=
4
.Utilising tho cocid:nt segsocce restits d volsped in the FRA fer full caro
\\
4 moltdown accidents and an assessment of partial core melt accidents (where the melt progression is terminated laside the reactor pressure vessel). the I
maximum potential risk reduction for a hydrogen control system in the BWR/6 Standard Plant was evaluated.
Methodoloar 1
TWo approaches were used to assess the nazimum risk reduction af forded by l
the addition of a hydrogen sontrol system to the BF1/6 Standard Plant design. The first approach used PRA techniques to assess the plant risk in i
terms of man-res/ year. The FRA assumes that all core damage segsences rossit la is11 sore meltdown and loss of containment latogrity.
The ressats from all accident seguences were grouped into fif teen fission j
product release estegories (compared to five used la WASE-1400). These i
fifteeen release categories were inymt to the smalysis which deterstaes i
of fsite consequences.
In seven of the fif teen categories loss of primary
[
l sostainment integrity was postslated to result from hydrogen combustion.
In the remaining eight categories containment overpres surization by steam or noncondensible gases was portulated. The effect of proclading hydrogen sombustion for the first sevos categories is to shif t the loss of j
sontainment integrity from the time of hydrogen combustion to the time of f
ecstainment overpressurisation from noncondensibles genstated primarily by sors soscrete ist.eraction. This delay la the time of fission prodast 3
release could reduce risk by allowiss additional time for fisslos product decay before the loss of contaimaant integrity.
i Nena risk was celes1sted using the CRAC code with the WASE-1400 site 6 meteorology and population withis $0 miles of the plant.
Site 6 was need as a representative site because it has average site characteristles. The population with S0 alles was used to allow sosperison with the proposed NRCsafetygoal(gy Relative to the proposed safety goal. the BTR/6 Standard j
Plant risk is well below the proposed anmeriosi guidelines for sore melt l
probability and risk.
}
l The second method of assessing the maximum ef fect of hydrogen sostrol on 1
severe accident risk considered partial sore melt sequences. Partial core melt is defined as sa accident shore sore cooling is restored in time to
]
arrest the melt progression and establish sootable geometry within the remotor pressare vessel. Partial sore melt was analysed since the NBC has foessed on less them a is11 eore meltdova for sessideration in hydroges
!l sostrol retomaking. TVo esses were smalysed:
tea percent of the sore melted sad fif ty pereent of the sore melted. These two esses were chosen to represent approntmate bonade for partial sore melt conditions.
Events with less than tem percent oore melt do not prodsee embetontist amosats of hydrogen (less than twelve percent metal-water reaction). For events with j
greater than fif ty percent oore melt, termination of the event withis the l
reestor pressure vessel essaot be assared and the event vos1d then fall j
into one of the f all meltdova seguences previonely discussed.
i f
I
( !
I i
Feel rolensos of fissiso pred::sts cero coleclated cs a fanation cf temperature and time at temperature.
Six transient-initiated assident seguences were modeled. These seguonces sostribute 93% of the assessed frogsemey of core damage (see Table 1.0) for the BWR/6 Standard plant.
Results Risk reesits of the fall core melt accident seguences are gives in Table 2.0.
The results are espressed as man-res por reactor year for a population to 50 miles) and the BWE/6 Standard plant PRA rossit (for a population to 500 miles). The BWR/6 FRA result is provided for comparison only. These results combine the probability of the accident and its consegmences.
It can be coactaded that the mesiana risk redsstion achievable is 0.14 man-res / reactor year which is a small risk redsstion when compared to the safety goal asserical guidelines discussed below.
Table 3.0 gives the results of the partial oore-melt analyses in man-rea.
A comparison is made to the natural backgrossd radiation (100 millires) received by the same population la one-year. The partial core melt results are empressed la terms of consegmences only with no regard for the probability of accident occurrence (i.e., accident probability assumed ogsel to one). It osa be seen that for the fifty percent sore melt case the total esposure from the accident is equivalent to 3% (24,360 man-ram) of the assual background radiation esposure. Therefore, in absolute terms, any reduction la this esposare is not significant.
Connerison to pronosed NRC Safete Goal j
In February,1932 the NRC pob11shed for public somment a proposed policy j
statement on safety gosta for aselear power plaats.
In addition, a i
separate report (NUREG-0300) discusstas the development of the proposed I
policy statement was published. Althossh the NRC safety goal policy is
)
only la draft fore, it provides a usefal comparisos is assesslag the results of the BWE/6 Standard Flaat FRA.
The proposed NRC policy statement proposes the followlag numerical l
l guidelines:
i 1.
Individual and Societal Nortality Risks l
l The risk to na individsal or to the population la the vielsity of a aselear power plant site of prompt fatalities that might i
rossit from reestor aseidents shesid not esseed one-tenth of one percent (0.1%) of the sua of prompt fatality riska rossittaa j
from other aseidents to which members of the U.S. population are 1
generally esposed.
}
The risk to na ladividsat or to the popsistion la the area seer 1
a aselear power plant site of esseer fatalities might resmit i
from reactor sealdents shos!d not escoed one-tenth of one percent (0.1%) of the saa of senser fatality risks ress1 ting i
from all other causes.
I l
1 1
} I 4
i
. = _
- - ~ _ -
4 2.
Ber.ofit-Cost Onidalico The benefit of as lacramental redsstion of risk below the I,
1, asserical guidlines for societal mortality risks should be soepered with the associated costs on the basis of $1,000 per man-res averted.
i
'3. Plant Performance Guideline Large-Scale Cor e-Melt: The likelihood of a nuclear reactor accident that roamits la a large-scale core sc1t should normally be less than one la 10,000 per year of reactor operation.
A comparison of the FRA results to the NRC proposed guidelines is provided t
la Table 4.0.
Comparison is made to all the numerical guide 11 ass dealing i
with mortality risks and plaat performasse.
The celos1sted core melt probability of ~$ z 10-6 per reactor year for i
the BWR/6 Standard Flaat is a factor of 20 below the proposed guideline.
There were no esiculated early (prompt) fatalities for the segmences sonsidered.
Conseguently, the BTR/6 results are well below the NRC 3 sidelines for individsal and societal prompt fetality risks. Tie NRC numerical guidelinegjyr individual latest fatality risk is based on 0.1% of i
national statistics and is equivalent to ~2.0 a 10-6 The FRA result of 1.7 a 10-5 latest fatalities (the mesa of the risk earve for fatalities withia 500 miles) when divided by the population within 1 mile of the site (168 people) yields as individual latest fatality risk of 1.0 x 10-7 The societal latent fatality risk is sales 1sted to be 1.7 a 10-5 and is five orders of magnitude below the guideline valse of 3.2.
I Comparison of the rossits la Table 2.0 to the proposed cost-benefit guide-line of $1000 per man-rom averted shows that sa estimated $10 million j
hydrogen control system falls the proposed sost-benefit oosparison by
)
orders of masaltade.
If a systes sos 1d avert all man-res from the accident (which is elearly impossible), the system should eost less than $140 per i
year (0.14 man-res a $103) to be sost-beneficial.
Considerlag partial l
sore melts without regard to accident probability, the maximum risk reduction afforded by a hydrogea control system is insigniftsamt (24,000 l
maa-res) sompared to backgronad radiation (800,000 man-res).
Coas1astoms Therefore, relative to natural backgronad radiation, the addition of a i
l hydrogen sostrol system provides minimal risk redsetton.
Farther, the BTR/6 Standard Flaat risk is already low cospered to the proposed NRC Safety Goal and thus the provision of an addittomal hydrogen control system is not oost effective.
l a
)
i i 1
y-i TABLE 1.0 g/
BTR/6 FRA RESULTS:
BREAKDOWN OF THE ASSESSED FREQUENCY OF CORE DAMAGE PER REACTOR YEAR Fregnency of Core Damage Percent of Core Event Descrintion Per Resctor Year Demano Probability 98.
o Transients 4.1 x 10 '
(38)
~
~
e All others 5 x 10 (10)
~
o l ess of Beat Removal 2 x 10 0.4
~
o ATTS 6 x 10 1.3
~#
o LOCA 2 x 10 0.04 4.7 x 10 '
~
Total TABLE 2.0 BWR/6 PIA RISK RESULTS FOR FULL CORE MELTDOWN SEQUI!NGS Risk (Man-rom ser reactor-year) o Standard Plant PRA Result, i
All accident seguences 0.26 I
Population withia 500 miles i
o All assident sogneuses 0.14 l
Population withia 50 miles l
l l
-S~
l I
l TABLE 3.0 g'
PARTIAL CORE lELT RESULT 5 NAN-REM PER EVI!NT*
o 100 percent core'elt 30.120 m
o 50 percent core melt 24,360 o 10 percent oore melt 19,430 MAN-REM PER TEAR o Backgronad Radiation **
820,000
- It shon1d be noted that the probability of this event is only ~5 x 10-6 per year whereas the background radiation occurs every year.
- Annual exposure to 8.2 million" people within 50 mile radius of plant.
l l
e t
9
-g
~, _ _... _ _ _. _. _ _. _. _ _
/'
TABLE 4.0 y/
COMPARISON OF BWR/6 STANDARD PLANT PRA RESULTS M PROPOSED NEC SAFETY GOALS Criteria Proposed NRC BWR/6 Standard Plant Per Reactor Year Guideline Rossit Core Melt Probability 1.0 x 10-4
~5.0 x 10-6 Individual Frampt Fatality Risk S.0 x 10-7 (1) 0 (4)
Individual Latest Fatality Risk 2.0 x 10-6 (1) 2 x 10-10 Societal Frampt Fatality Risk 1 x 10 4 (2) 0 (4)
Societal Latest Fatality Risk 3.2 (3) 1.7 a 10-3 N0'I15 :
III 0.1% of National Fatality Statistics.
2)
Assuming 1-mile average population of 168 people.
( I l
Assuming 50-mile average population of 1.7 millica people.
(4I l
No prompt f atalities were calculated la any of the 238 Nuclear Island I
FRA accident sequences.
\\ l
w.
RFJF2EN GS (1)
GESSAR II, BTR/6 Nuclear Island Design, General Electric Company, March 1982 (22A7007 Rev 2) Section 15D.3 (2)
GESSAR II, BTR/6 Nuclear Island Desisa, General Electric Company, May 1982 (22A7007 P.ev 2) Section 15D.2 (3)
U.S. Nuclear Regulatory Commission, Saf ety Goals for Nuclear Power Plants: A Disenssion Paper, USNRC Report NUREG-0880 February 1982.
l l
i OENERAL ELECTRIC COMPANY
' ^
'^
QUESTION 820.87 For a complete station blackout with the less of offsite and onsite AC power, RCIC system would be the only avaible source of makeup flows in the GESSAR-II plant.
In the event of a break in the RCIC steamline between the outer containment isolation valves and the turbine stop valves, this source of makeup flows would also be depleted.
a)
Discuss the impact on the RCIC system availability as given in GESSAR-II PRA.
b)
Discuss the impact on the core-melt probability in sequences which have assumed the availability of RCIC.
RESPONSE
A 6
K f
e d
- 140 -
GENERAL ELECTRIC COMPANY QUESTION A1:
PROPRIETARY INFORM AT ON Table 7.2-1 gives risk values for early and latent fatalities for RSS BWR/4' Mark I composite site as calculated with "GE CRAC" codes.
No discussion of this calculation is included in the text.
Provide infor-mation on this calculation including, at least, population, meterology, accident sequence characterization (release fractions and timing para-meters), evacuation assumptions, shielding factors and breathing rates, radionuclide inventories, and specification of any dose-consequence relationship that is different from WASH-1400.
RESPONSE
h e0
- 141 -
GENERAL ELECTRIC COMPANY QUESTION A2:
PROPRIETARY INFO:'M ^ v'^"
Table 7.2-1 gives Risk values for Early and Latent fatalities for RSS BWR/4 Mark I at site 6 as calculated with "GE CRAC" code.
These numbers are different from the previous revision.
There is no discussion in the revision of why the numbers have changed.
Provide information on the causes for the change including input changes and justification for them; coding changes, justification for them, and quality assurance; and option changes and justification for them.
Highlight any differences in input, coding or options compared to the " composite site" calculation referred to in question 1.
Discuss the expected direction and magnitude of the change in consequences with each change in the calculation.
RESPONSE
O t
142 -
GESSAR II 22A7007
~
238 NUCLEAR ISLAND R3v. 11 GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION Class III Table 7.2-1 ESTIMATED CORE DAMAGE AND RISK COMPARISON
' Assessed ofEven$
Risk (Per Year) t Per Reactor Early Latent Event Year Fatalities Fatalitiesb I.
CORE DAMAGE RSS BWR/4 Mark I
-5
-5"
-2" 9 composite site
%4x10 1x10 c
45x10 c
-5
%4x10 2.4x10-5 2.5x10~2 RSS BWR/4 Mark I
-5
-6
-2 9 site #6c N4x10 7.8x10 2.1x10 BWR/6 P. ark III
-6
-5 9 site #Ge 5x10 40 1.7x10 II.
U.S. NATURAL BACKGROUND Continuous 0
814 RADIATION "With WASH-1400 Methods (calculated from the reported curves).
bThe total accident-caused fatalities over the lifetime of the exposed population or the calculated excess cancers in the same population from one year of background radiation.
cComputed with the GE CRAC Code.
4 6
0 15.D.3-143
GENERAL RECTRIC COMPANY PROPRIETARY INF7RM47'n*'
QUESTION A3:
Relative to the "GE CRAC" code, including any recompilations performed at execution:
(a)
Is the version used exactly the same as used for the Limerick PRA?
If not, enumerate any differences with appropriate equations.
(b) How was this version of the CRAC code benchmarked? Discuss runs relative to any commonly accepted codes and how the decision was made that the chosen benchmark program was adequate.
Give numerical comparisons for the program.
RESPONSE
(a) The code used for the final Limerick PRA results and the BWR/6 PRA results are the same revision of CRAC.
(b) The models used in the 1975 version of CRAC, available from Argonne Code Center, are reasonable conservative radiological consequence models.
The models in the code were verified by hand calculations while the overall results were compared to the results given in WASH-1400.
e f
- 8 e
a, y
b 0
- 143 -
CESSAR li 22A7007 23B NUCLEAR ISLAND Rev.33 CENERAL ELECTCIC COMPANY PROPRIETARY INFORMATION Cims lil 10'3 10 10-5 s
x A
E*
8 U<
E 4
10 5s U
8 E
10*7 4
4 10 i
4 t0 I
3 4
5 8
10' 10 10 10 10 E ARLY FATALITIES (x) 22343 on Figure F.41. Site Comparison of Esely Fatality Curves for WASH 1400 BWR 15.D.iS88
~'
CESSAR 11 22AiOO7 22 NUCLEAR ISLAND p.,, y CENERAL ELECTRIC COMPANY PROPRIETARY INFCRMATION Clau til 10'3 b
10 10-5 x
A E*
8 0
2s E 10 v
5 a
m 10 10 4
10 3
d 5
0 2
10 10' 10 10 10 10 LATENT F ATALITIES PER YEAR (x) 22H347 fi vre f.4 2. Site Comparison of Latent Fatality Curws for WASH 1400 BWR 9
3 5.D.3-589/15.D.3 590 l