ML20080T538
| ML20080T538 | |
| Person / Time | |
|---|---|
| Site: | Beaver Valley |
| Issue date: | 03/10/1995 |
| From: | George Thomas DUQUESNE LIGHT CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| GL-88-20, TAC-M747378, NUDOCS 9503130375 | |
| Download: ML20080T538 (81) | |
Text
{{#Wiki_filter:N. b' l I
~
Beaver Va ley Power Station Sh pp ng rt FA 15077-0004 (412) 643-8069 FAX l t GEORGE S. THOMAS March 10,'i m
$fk!! .$o U. S. Nuclear Regulatory Commission Attention: Document Control Desk ;
Washington, DC 20555-0001 P
Subject:
Beaver Valley Power Station, Unit No.1 Docket No. 50-334, License No. DPR-66 ; Generic Letter 88-20 (TAC No. M747378)
Reference:
NRC Letter to Duquesne Light Company (DLC), Beaver Valley Unit No.1 - Request for Additional Information Concerning Generic Letter 88-20 Individual Plant Examination for Severe Accident Vulnerabilities (TAC No. M74378) Duquesne Light' Company's responses to the NRC's referenced Request for Additional Information (RAI) are attached. Should you have any questions regarding this submittal, please contact N. R. Tonet at (412) 393-5210. Sincerely, corg S. Thomas Attachment ec: Mr. L. W. Rossbach, Sr. Resident Inspector Mr. T. T. Martin, NRC Region I Administrator Mr. D. S. Brinkman, Sr. Project Manager Mr. R. Maiers, Pennsylvania Depanment of Environmental Resources 140000 . 0A 9503130375 950310 } goa nooexc5oog4 14
r BV1 TPE FRONT-END OUESTIONS
- 1. Reactor Coolant Pump (RCP) seal Loss Of Coolant Accidents (LOCAs) are a significant contributor to the overall Core Damage Frequency (CDF); however, the submittal does not address the RCP seal LOCA model used in the IPE. Please identify and discuss the application -
of the RCP seal LOCA model used in the IPE.
Response
The seal LOCA model is described in Appendix B, Section B.2 of the IPE analysis, which gives the background for the electric power recovery model used in the IPE. This Appendix was not part of the IPE summary report, therefore, a brief synopsis of the RCP seal LOCA model and how it was applied in the IPE is given below. The model for the pump seal leak rates was based on the four-loop RCP seal LOCA study of Reference B.2-4 for the Westinghouse RCPs with the old style 0-rings that existed in the Beaver Valley Unit 1 RCPs at the time of the study, and scaled by the number ofloops at Unit I to reflect the leak rates per pump. These data were then used to develop the probability leak rate model for the electric power recovery analysis. The specific seal LOCA leak rates, used as a function of time after the loss of seal cooling, are provided in Table B.2-1, copy attached. The flow rates listed in gpm define the effective flow area, assuming an RCS pressure of 2250 psig. The time to core uncovery for a given leak rate, which varies with time, was computed accounting for the decrease in RCS pressure as the accident progresses and includes the effects of the operator action to depressurize the Steam Generators. Reference B.2-4 is as follows: NUREG-ll560, Report Reactor Coolant Seal LOCA, "Results of Expert Opinions Elicitation on Internal Event Front-End Issues for NUREG-ll50: Expert Panel", NUREG/CR-5116, Volume 1, Sandia 88-0642, April 1988. PLG developed an engineering code, SEALOC (Reference B.2-5), to calculate the time of core uncovery due to a pump seal LOCA during an SBO with the turbine-driven or dedicated Auxiliary Feedwater pump available. SEALOC was used to calculate the time of core uncovery for the various probabilities of seal leak rates shown in Table B.2-1 and the impact of RCS depressurization. A constant leak rate (initiated immediately when all onsite AC power < fails and RCP seal cooling is lost) of 21 gpm per pump was used in the electric power recovery , analysis as the leak rate for the first hour prior to the severe seal damage. l I I of 69
m f t: t For the worst case, aRer the first hour, a 480'gpm leak rate per RCP is expected, given the - J assumption that the low pressure seal leakoff piping would rupture aRer failure of the #1 RCP - , seal (the assumed rupture of the seal leakoff piping is only made to conservatively bound the i maximum flow rate through the seals). On the basis of the analyses performed with'the SEALOC code, this leak rate would result in core damage (1,200'F) approximately 7 hours , aner the initiation 'of the station blackout, if operators took action to depressurize the_ Steam' Generators in 2 hours or less. If the operators depressurize aRet 2 hours, or completely fail to
~
p' depressurize at all, core damage would occur approximately 2.9 hours aner the station-
- blackout. The results of this analysis was then used in conjunction with other parameters to
L determine a nonrecovery factor used in the electric power recovery event tree. Reference B.2-5 is as follows: Maneke, J. A', D. R. Buttemer, and R. K. Deremer, " Reactor Coolant Pump Seal LOCA ' l Analysis during Station Blackout Events at Seabrook Station", prepared for- New Hampshire Yankee, Pickard, Lowe and Garrick, Inc., PLG-0724, January 1990. ( p r 2 of 69 i
1 ( Birv:r V::lley P;w:r St:ti:n Unit 1 Rivisi:n .0 l Probabilistic Ri;k Assessment Progr:m - f
- Table B.2-1 Seal LOCA Flow Rates (GPM) per Pump with and without Primary '
Depressurization , Probability Cumulative Time after Station Blackout (hours) Probability 0-1.0 1.0 1.5 1.5-2.5 2.5-3.5 4.5-5.5 5.5 + (gpm) (gpm) (gpm) (gpm) ~ (gpm). (gpm) . ,
.0.2712 .2712 21 21 21 21 21 21 0.0151 .2863 21 21 21 61 61 61-0.0161 .3024 21 21 61 61 61 61 0.0181 .3205 21 61 61 61- 61- 61 0.0120 .3325 21 61 108 108 108 108-0.0059 .3384 21 61 108 108 .120 175 0.1120 .4504 21 61 250 250 '250 250' O.0136 .4640 21 120 250 250 250 250 0.5302 .9942 21 250 250 250 250 250 0.0016 .9958 21 308 308 308 308 308
_._Q.0042 1.0000 21 480 480 480 480 480 e i 4 l l l l l l l 3Ofb9 l i 8.2-18 B.2 Electnc Power Recovery Actions.
~' ^
y g~y74 - g ,
- w ,- ; .
1 p' f j.:
. ^j gMy ' L2, ilt is not clear from the' submittal how spray induced failures of equipment were addressed in" q
~~
- !the internal flooding ' analysis. Please ' provide'a discussion of the treatment of spray inducedi: 1
- /7 ' failures'and, if they were screened out, provide the basis for the screemng.. 1
; y Response- ;
~
LThe IPE did focus on modeling submergence-induced failures of equipment 'as a result of - .
~
j
- : internal floods. To assess the effects of spray, design basis event analyses for moderate and - i
^
, high energy line breaks were reviewed prior to the plant-walkdown for internal flooding /
Spray-induced failure modes were then considered during the plant walkdown to identify key? 3 scenariosc As~a result of the walkdown, spray-induced effects were judged to be localized, so' j emphasis was placed on effects beyond design basis; namely, larger floods.- Subsequent consideration of spray effects was then limited to inadvertent actuations, and leaks l and breaks in the fire suppression systems; i.e., the Fire Water System. In generaliinadvertent ' : actuations of the Fire Water System were not considered significant flooding sources because -l of the relatively low capacity of the sprinklers and because alarms would alert the operators. One possible exception to this that was considered, was the actuation ~of the sprinklers above. , the CCR pumps. 'However, the freauency derived for inadvertent actuation of this portion' of ! l the Fire Water System was more than an order of magnitude less than the frequency oflosing; , CCR from all other causes and, therefore, was judged insignificant.- ! The screening approach to the analysis ofinternal flooding used in the Beaver Valley Unit:1 IPE is conservative. ; Spray effects that are localized to a particular compartment are accounted for by the screening assumption that all susceptible items within a location are initially failed. ; Scenarios that survive the screening are then examined on a case-by-case basisi The resulting'- ! flood scenario frequencies that survive the screening have frequencies expected to be as high as e spray events and with greater plant impacts. I u For example at' Beaver Valley Unit 1,' the highest frequency flood scenarios retained'aftsrf l screening for quantification occur in the Intake Structure (9.8 x 10-3 per year) and the Turbine - l Building (7.7.x 104 per year). The Intake Structure flood was modeled as failing all normal
- river water and a raw' water pump. The' Turbine Building flood was modeled as impacting- l main feedwater, turbine plant component cooling water, and station' instrument air. Itis? 4 difficult to see how spray-related failures could result in more severe scenarios than these; l Experience at this and other plants (e.g., Seabrook) suggests that even when spray effects are ;
systematically reviewed, they are not significant when compared to submergence effects. . y
, l
?
y i i 4 of 69 ,
-6
7
- 3. Please address the following item with respect to the need for containment cooling to support :
core cooling. The success criteria indicates that containment cooling is required if energy is released into containment; however, the discussian of the event trees implies that containment cooling was modeled only to establish conditions for the back-end analysis. How did the IPE model the loss of containment cooling as impacting the ability to cool the core?
Response
The inside and outside Recirculation Spray pumps provide the containment heat removal function (via the Recirculation Spray coolers) and part of the containment spray function in conjunction with the Quench Spray System. In addition to these functions, the outside Recirculation Spray pumps can also provide vessel injection by diverting padial flow from the Recirculation Spray header to the suction of the HHSI pumps, when the LHSI pumps are unavailable for the SI recirculation mode. The success of the containment spray and heat removal functions are only required for the containment analysis and have no impact on the calculation of the core damage frequency. ! For non-LOCA cases, core cooling is provided by the Auxiliary Feedwater, Main Feedwater, or dedicated Feedwater Systems. Core cooling for LOCA cases is provided by the HHSI and , LHSI systems which transfer RWST inventory into the core during the SI injection phase and I containment sump water into the core during the SI recirculation phase. The IPE modeled the common cause failure of all four Recirculation Spray coolers, due to their River Water supply check valves failing to open, as failing the containment sump (Top Event SM), since this could l impact the NPSH requirements for the LHSI and RS pumps. The independent failure of all four Recirculation Spray trains was not modeled at impacting the NPSH to these pumps in the IPE submittal. It was expected that the occurrence of this would be extremely small or that such sequences would have already progressed to core melt. To be certain, however, the PRA ; was requantified with this dependency modeled. As expected, there was no increase in the ' CDF. The IPE model also takes credit for a continued SI injection mode if the sump is 3 unavailable, or the sump suction flow path fails during the recirculation phase, by making up to i the RWST (modeled in Top Events MU and WM) to prevent core melt. It is assumed that as long as the flow provided to the core, via the HHSI or LHSI systems, matches the boil-off flow ; rate, core cooling will be provided. ; i If vessel melt-through occurs, a sustained makeup of 140 gpm would be adequate to cool the I core debris once the decay heat level reaches approximately 0.75% of full power, assuming a coolable geometry. Injection of the RWST into the containment provides a source of water for debris cooling if vessel injection during the recirculation mode is available, or if the debris is dispersed from the keyway.
- 4. The contribution to CDF from Anticipated Transient Without Scram (ATWS) for Unit 1 (20%)
is significantly different from Unit 2 (4%). The submittal ir.dicates that top events RT (Reactor Trip) and PA (Primary System Pressure Relief) are significant contributors to CDF from ATWS; however, additional information received by the staff from Duquesne Light Company indicates that the importance of these items is significantly less than previously reported. Please provide a discussion regarding your assessment of this event and its contr:bution, and 5 of 69
J , I s -; the ' contributors :(including- the ' PORV block valves) to' CDF, based on. your current- [
, s understanding.
?
Resnonse- j
-9 . As pan of an effort to determine the impact of a proposed Technical Specification (which will . .!
require two ' operable PORVs)'on plant operations, a review of the ATWS and associated i h pressure relief models in the IPE has been performed. For the IPE, some assumptions had to . l E be made about'the ATWS analysis, which were not documented in WCAP 11993. During this. j recent review, however, additional information about'the WCAP analysis was obtained that' , indicated that the earlier. assumptions' for PORV availability were inconsistent'.with' the- :i reference analysis in WCAP 11993, and that these inconsistencies were ' causing the ~ core - i damage frequencies for ATWS events to be overly conservative. - Therefore, a reevaluation of L : the ATWS model was recently performed to show what effects this new insight had.on core : ~ damage frequency. A discussion of this reevaluation is provided in the following' paragraphs. WCAP 11993 was prepared to demonstrate that Westinghouse plants comply with the ATWS . { target goal listed in SECY-83-293. The results of this analysis are in the form of probabilistic i frequencies of ATWS events which lead to overpressurization of the RCS, The' analysis uses ! plant parameters selected to bound all Westinghouse plants for ATWS considerations. The ; event tree includes system top events for those systems which play an important role'in ! mitigation of an ATWS event. This includes overpressure mitigation capability through the use ; of PORVs and Pressurizer Safety Valves. The overpiessure protection capability is a characterized with respect to the variability of the moderator temperature coefficient over the ~i fuel cycle. The number of relief valves required to prevent overpressurization of the RCS is - 'i calculated as a function of the time in core life and conversely, the time period during a cycle , 1 for which overpressure protection is not adequate for a given number of operable valves is l tabulated. The latter .is defined as unfavorable exposure time (UET). The UET is then j adjusted based on the frequency weighting for transients during a cycle, averaged over the. ! cycle and normalized to a standard period for use in the risk assessment. i
- l From WCAP 11993, for the reference plant, the following UETs were calculated for various l
, PORV availabilities and assumptions for Manual Rod Insertion (MRI) and Auxiliary Feedwater j flows (values are in days for an 18 month cycle): l Condition 2 PORVs 1PORV- 0PORV- i With MRI, all Aux Feed 0.0 0.0 76.3- 'j With MRI, half Aux Feed 0.0 . I8.9 82.6- j No MRI, all Aux Feed 81.7 138.9 192.9 i No MRI, half Aux Feed 110.7- 154.8 209.1 l All auxiliary feedwater for the reference plant is assumed to include flow from all pumps, i.e., 2 ! motor-driven pumps and 1 turbine-driven pump. Half auxiliary feedwater is the capacity of 1 .l turbine-driven pump or 2 motor-driven pumps. In the WCAP, the UETs listed were funher 'i adjusted to account for the frequency of transients as a function of cycle life and normalized to i a one year period. This was done to account for the fact that historical data showed that a - higher frequency of feedwater-related transients occurred during the early part of a cycle. j 6 of 69 .l i n i
. ., . _ , - _. . - . .-. , ~ . . - .. . . _ _
)
~
y;+ - a , _
+
9
+*
y q
.l i
=In reviewing how' this information was applied to the Beaver Valley IPEs, the followingl ;
y' observations were made: j w
~
LOnly the UETs for half Aux Feed were used for both halfind full Aux Feed cases forf each condition of rod insertion.L This has the effect of not taking credit for,the slight J
,j
( improvement in UET for cases in which all Aux Feed is available. In the current PRA j model, success of the Auxiliary Feedwater top event only requires'that either the ) y'4 turbire-dsiven pump or both of the motor-driven pumps be operable.~ >Therefore, no: j additional mitigating credit is taken for the successful operability of all three Auxiliary 1'
. Feedwater pumps; i.e., it is treated the same as a failure of the turbine-driven pump or a ?
failure of one or two motor-driven pumps. This is a conservative approach employed .' ' primarily for modeling efficiency considerations. l l r . _ , ;i In the IPE, the adjusted.UET values from Table B-3 of WCAP 11993 were .used , -l directly without normalization or consideration of the effects of transient weighting i
.with respect to initiating frequencies as used in the IPE In the WCAP, the frequency of overpressure events due to ATWS for an assumed PORV availability condition was lj; calculated by using the adjusted UET values and the failure frequencies for'the'PORV' i p and safety valves. For any period in which the assumed number of PORVs and 3 safetyf i valves was insufficient, the failure probability for the valves is equal to 1, and the total. 1 failure probability is equal to the event frequency for this period. The total failure rate :
for the cycle is determined by adding the failure rate for the different periods (based on j pressure relief requirements) and normalizing to a one year period to arrive at a mean j failure rate. In the IPE, the ATWS event tree is integrated into the overall model and is~- -j entered upon failure of automatic or manual reacior trip. Initiating event frequencies ' l are annualized and, therefore, no weighting of transients during the cycle is required' j since all transients will be included for the period. In order for the UET as presented in? j the WCAP to be consistent with this approach, adjustment of the UET to a one year : period is necessary, but transient weighting should not be included. In the current ' application, the transient weighting was included and no adjustment to a .one year. period (normalization) was performed. The result is that the model becomes.very conservative with respect to the frequency of overpressure failures resulting- from ATWS initiators. A shift of the UET values from the WCAP was performed based on the number of.~ h PORVs at Beaver Valley versus the reference plant assumed in the WCAP. The 1 assumption made here was that the total capacity of all PORVs was equivalent regardless of the number of PORVs installed. Therefore, one PORV at Beaver Valley ; ljl was assumed to have only two-thirds of the relief capacity of one JORV at the: reference plant. Based on this assumption, the UET values as presented in the WCAP ' jl jj were shifted such that where one PORV was assumed operable for the reference plant,
' it was assumed that Beaver Valley required two PORVs. Likewise, the UET values listed for the no PORVs operable case in the WCAP were used to represent the l condition where one PORV was operable at Unit 1. Discussions with Westinghouse l personnel, and a review of the reference documents from WCAP 11993, have indicated !
that the individual PORV capacities at Beaver Valley are equivalent to those assumed i 7 of 69 l 1
%r _ .: .
%.7
~ ?
L 4 i # A 3 l
-I
~ '
;i for the reference plant.- Therefore, the adjustment made for' valve capacities is not
~
i required and myims another conservatism in the analysis. < j
-t
. Direct application of the UET values from WCAP 11993 based on.1 PORV available, half. j
~ Auxiliary Feedwater, without weighting for transient time, and adjusted to a'one year period- >
l resultsin the followmg. l j With manual rod insertion: 5.2% of the time 1 PORV + 3 SRV are insufficient j 17.4% of the time 1 PORV + 3 SRV are required i j 77.4% of the time 3 SRV are required i e .t f Without manual rod insertion:42.4% of the time 1 PORV + 3 SRV are insufficient j 14.9% of the time 1 PORV + 3 SRV .are required ,; 42.7% of the time 3 SRV are required i M Using these values in the Beaver Valley Unit 1 IPE model results in a reduction of the~ l contribution to the~ total core damage frequency from ATWS sequences in general, .and : l specifically those associated with PORV block valve alignment. The previously reported! l contribution for ATWS events for Beaver Valley Unit I was 20.1% of the total CDF. ' Based on the revised fractions as listed above, the contribution is reduced to'6% of the" prior total- -l CDF. The contribution previously reported for ATWS events in which failure resulted from j inadequate overpressure protection capacity (PORV Block Valve Alignment) was 15.6%.- 1 Using the fractions listed above, the contribution is reduced to 2.87% of the previous total d CDF. j i Previous discussions. on resolution of this vulnerability had indicated that . proposed i amendments to the Technical Specifications, which will require two. operable PORV vent paths, would reduce this vulnerability by increasing the number of PORVs available.! While1 j ii increased availability of the PORVs may provide some benefit in terms of ATWS mitigation, . ! this is only true if the block valves remain open. However, the ' existing . Technical y : Specifications and the proposed changes do not require that a block valve remain open in order
~
~i for a PORV to be considered operable. Therefore, it has no direct impact on the PRA ATWS i model because the model takes no credit for the PORVs unless the operable'PORVs' block'. l valves are also open. The revised model will utilize actual plant experience as the basis for j PORV availability. The reevaluation which assumes one PORV available with the block valve 1 open demonstrates that with modeling conservatisms reduced, as stated above, the vulnerability !
is reduced to an acceptable level consistent with the ATWS licensing basis..
- 5. NUREG-1335 requests human reliability data, the time available for operator recovery andL i other generic or plant-specific data for ir..portant equipment or events. The submittal does not 1 identify the data used to model recovery of offsite' power and recovery of failed Diesel 'l Generators (DGs) which are important recovery actions. Please provide this data and the :
source of the data. l
- i f
I 8 of 69 - a'
, s. -- .,. - -. .
~
Q ; o' ,
~ 4 7 , , i 2>,l: - {
- Response: j
'M, The electric power recovery model used in the IPE is described in Appendix B,' Section B.2., l F ,; and as stated in response to Front-End Question 1, was not part of the IPE summary submittal. l
- The loss of offsite power data. at U.S. nuclear plants reported for all years' through 1988' :
'i . (Reference B.2-1) were the basis for developing the mean probability of nonrecovery of offsite - , j power for PLG's generic database and for use in the Beaver Valley Unit 1 FRA' model. d Specific line. data from the-Seabrook site (Reference B.2-2) and South. Texas Project; site Ll (Reference B.2-3) were used. along with. data from' NUREG/CR-5032 l to' develop the :
l
-distributions from the mean for the 10th and 90th percentile cases. These distributions for the. j
< probability of nonrecovery used in the electric power recovery model for. Beaver ValleyLare ; j i
'shown in Figure B.2-1, attached. References B.2-1, B.2-2, and _B.2-3, respectively, are as '
follows:
- I . Nuclear. Safety Analysis Center, Electric Power Research Institute,LInc.,:" Losses of f_
l Offsite Power at U.S. Nuclear Power Plants All. Years through 1988", NSAC-144, April 1989.- ; y Pickard, Lowe,' and Gerrick, Inc., "STADIC4 Model for Frequency of Nonrecovery of_ -! Electric Power at Seabrook Station for Plant at Power and Shutdown", prepared for New l Hampshire Yankee, PLG-0507, May 1988. D j Pickard, Lowe, and Garrick, Inc., " South Texas Project Probabilistic Safety Assessment",L prepared for Houston Lighting & Power Company, PLG-0675, May 1989. j The analysis considers recoveiy factors for cases in which 0,1 or 2 Emergency Diesel l Generators are available for recovery. The discussions which follow first describes the case i when only one Diesel Generator _is recoverable, and then describes the case for when twol 'l Diesel Generators are recoverable.' ! i l 1 n
]
i
-i Li i
i i i 1 t 9 of 69 j
, e
[f]q u a , s.. 4 d t
'f( -
{ x [
}
l
- The;model that wa :used for a single Diesel Generator recovery (i.e., only. one Diesel Generator has DC power available)is:
]
a (Equation B.2.4) j
;4g , .
4 (t + t) =(($on(t))[@m(t)]dt 3 where; ; a
@i(t + t) = cumulative frequency ofpower recovery from a single Diesel
Generator when'only one diesel is available for recovery ' q a Gon(t)dt = frequency of auxiliary operator response to Diesel Generator; j f-Building between t and t + dt aAer the failure of Diesel < t Generator power j-u , g
@m(t)= cumulative frequency of Diesel Generator hardware recovery' within time (t) aRer operator response <
j! This analysis is performed for conditions when only one diesel is available for recovery; i.e., the : '1 other diesel has failed at t = 0 and cannot be recovered withih 24. hours; In this analysis; ! g approximately 20% of the single diesel unavailability is assumed to be attributed to pre-existing j maintenance scenarios. The 5th percentile model for the single diesel recovery reduces the 1 cumulative frequency of recovery for o' ne diesel by 20%.' For the 95th percentile, however, a' more optimistic view is taken, and it is assumed that this fraction of unavailability .is ' jl L recoverable. This would include, for example, restoring the diesel to service aRer minor 1 l' maintenance or testing. For the 50th percentile, it is assumed that the fraction of unavailability . ! due to maintenance is recoverable aRer 2 hours. l 1 The 5th percentile of the single' diesel recovery model represents a pessimistic model for the c 1 operator response and delsys the auxiliary operator's time by 30. minutes. The 50th percentile. '! of the model represems a delay of the operator's arrival by 10 minutes.- The 95th percentile . j i bound represents a more optimistic model for operator response, and no delay.in the auxiliary , 1 L _ operator's arrival is included. Figure B.2-2 (attached) presents the complementary cumulative - 1 distribution for the Diesel Generator nonrecovery that is derived for these bounding models.'
'i l
For the case where power can be recovered from either Diesel Generator (i.e., two Dieseli
~
j Generators are recoverable), successful recovery has been dermed for this analysis as the 1 restoration of power from at least one of the two Diesel Generators. This nonrecovery model . -l
, for one out of two Diesel Generator recovery is characterized by the expression: !
@m(t + t) = $ i(t + t) + [1 - @j(t + t)][@i(t + t - 0.5)] (Equation B.2.5) j; This model allows recovery of the first of the two Diesel Generators to begin when an auxiliary - j operator arrives at the Diesel Generator Building. Recovery of the second diesel begins 30- j minutes aRer the auxiliary operator arrives, and the repairs of both diesels are modeled as l
]
continuing in parallel thereaRer. 4 10 of 69 s j = x .. .u . - - . . , . .. ..- - , - .-- - : - . , . - . *
, _ ~ .- -
Z . ,} k , , A
._,',Y s
i i
.i Two boundmg scenanos are applied as the 5th and 95th' percentiles for the Diesel Generator 'l L recovery model. For the 5th percentile bound, the single Diesel Generator' recovery model (!
^
(above Equation B.2.4) is used.. This model represents a pessimistic model for ~ operator. !
~ . response, and it allows recovery of power from only one Diesel Generator.; Parallel repairs of; i
^
~
the second Diesel Generator are not considered.nThis bound accounts for possible unidentified . j dependencies in the recovery efforts for both Diesel Generators, which could couple the rc,.'ir j
' time distributions; e.g., limited spare parts availability, limited support personnel availabihtyc j etc. For the 95th percentile bound, the dual Diesel Generator recovery model (Equation B.2.5) L :
is used. The recovery of the second diesel begins 30 minutes after the operator arrives, and the d
. repairs of both diesels are modeled as continuing in parallel thereafter. The 95th percentile j bound thus represents a more optimistic assessment of operator response, and it: includes a 'l
' more realistic model for single and parallel Diesel Generator repairs. The'50th' percentile is : -i estimated from the 5th and 95th percentile. curves. Figure B.2-3 (attached). presents the 1 p ' cumulative distributions for the Diesel Generator nonrecovery derived from these bounding - j
( models. j 9 It should be mentioned that the three curves for each quantity (i.e., the 5th, 50th And 95th) are d weighted in the calculation as 0.1,0.8 and 0.1, respectively. The above terms for recovery of j offsite power and for the recoverable Diesel Generators are used in the electric power recovery l Equation 3.3.3.4 in Section 3.3.3.4.1 of the IPE submittal. !
- l >
I" 1 l 5 Lf i t
~l i
.l
.I I
11 of 69 r [ a, .- . . - . . - - - . , - . , , , . - ~ ,. -
' on
? F.
NOTES: T3 SE '
=<
- 1. 50th PERCENTILE FROM NSAC-144 DATA. EE 2.10th AND 90th PERCENTILE FROM ONE OR THREE LINE DATA FROM SOUTH TEXAS, SEABROOK, h4 AND NUREG/CR-5032.
Fg [g E
. :EE
!l Ec g 1.0 -
?E s g, o
o : , E 5 a m E .8 - z m o O Z b & O .6 g 90th PERCENTILE S g .4 ' 50th PERCENTILE O E m
.2 . . .
10th PERCENTILE g . ~ o r i , : : : 2 : : : : : : j 1 0 2 4 6 8 10- 12 14 16 18 20 22 24 t - m
$ 2 3
! 2-TIME TO RECOVER OFFSITE POWER (HOURS) 'h l 3 5-8 m Figure B.2-1. Probability of Nonrecovery of Offsite Power
- , , - - - - - - , ,,une--an- - ,,,,*s-ws,-w,---,,,,, ,e--~w- +*,,s <w>- <,-,-w-mca s~-. - * - ~ - --,w , rs,-+~ e-w,-- - N - -w we, w e ,w,,- v m - -er---- , , - - ,.,-w , ~v--v .-w-mm1, +- , , ,,w ,se-- ,e ,
+ T..
~~
.+=.
[ i ;; 4,
.=<-
EE' is , , T{1 o, ., g De em
.gg .
- lg i . n:
3* =5'1 -
> or o
.Ec.
5 3 3
.,a a -
g04 , g, c 'a-m
, 05 -
-3 2
2
- a W 3 04 -
1: 'Y % $- a m u O3 - 95e PERCENDLE e-Q2 - 50m PERCENDLE - ~ 8.t .. $st PERCENDLE - i
, e i I I I 'l i
-l I 'l I l' 'I b
0 2 4 -6 8 10 - 12 14 - 16 10 - ' 20 = 22 ' 24 1 4 O DURATION (HOUR $g h .,r a i 1 4 m
- t. g o
'~3'
. O^
- j Figure B.2-2. Nontecovery of Single Diesel Generator g !{
.O.
g . 3 y .o a
-m- w o- ._ _ ew -- v w w e= .-,w en. bv.wmne-- w.oe-w---. + , .waee---.-,<v.-,.wew+'s.+6w.-.----6.- -m, w o - nw -o- w-w ..-ew ~,n. 4e-- e> w w 4 == d ire.--.-i.,. . n ve 4e..& ....-.w,-.= .==w=*euva.-i-y . - . - er b2k AM
e;
.c-Beaver V:lley Pow;{ 2tation Ulit 1 R:vi:lin 0
, tProbabilistic Risk, A:sessment Progr m
.,1 N.
-M lt .
, - 2' F,
r
- 2 "A
- \
g' ._ ,
.s
- w~
{
.,1 *
,e ue g 5 3 en g 8 o=
g - u 3 g .m o
.k o. .
x O E
- = .o T
E E h g - -
.g
& .n' o
- lit
- n Q-ew M
es r i f f f f e s , , E i
-o o $ $ $ ; y ; e ss.
ADN3OO3Wd 3MAv7nans -}
-I L.
i i 14 of 69 ! I F i l r I 8.2-23 B.2 Electric Power Recovery Actions. l
- - ~ ,
r 1 o
- 6. Duquesne Light Company has identified a number of vulnerabilities for BV-1 in the submittal, and their importance.to CDF. However, it is not clear to what extent the potential enhancements have been incorporated in the plant, nor what reduction in CDF is expected for implementation of the enhancements. As requested by NUREG-1335, please identify the status of the implementation of the proposed enhancements and the reduction in CDF for each enhancement, or the total reduction in CDF, for those enhancements expected to be implemented. In addition, please provide a discussion for those identified potential enhancements not being considered for implementation, and the basis for the decision.
Response
The status of the implementation of the proposed enhancements for the vulnerabilities identified in the Beaver Valley Unit 1 IPE submittal is addressed in the following paragraphs. The reduction in CDF from the implementation of these enhancements was not calculated for each item; rather a total reduction in CDF was calculated for all of the changes made including implementation of vulnerability enhancements, slight changes to the top event models to reflect plant modifications performed through 1993, and plant-specific data updates of component failures and maintenance through June 1993. Model changes associated with the vulnerability enhancements only reflect installation of the 4160V station crosstie and revision of the primary pressure relief top event for ATWS events as stated in the response to Front-End Question 4. The new total CDF as a result of all changes to the PRA model is 1.20 x 104 per reactor year (using the same initiating event frequencies as reported in the IPE) This represents an approximate 44% reduction in core damage frequency. I,oss of Emergency Switchgc_ar Room HVAC This specific sequence results from a loss of both normal and emergency cooling to the emergency switchgear area which could lead to equipment damage in these areas and subsequent loss of power to emergency equipment. Although credit was taken for ' restoration, since operators are aware of the potential results oflosing both trains of cooling, the previous alarm response procedures did not provide specific guidance for mitigating the consequences of this event through the use of portable ventilation. More specific response procedures have been developed to provide temporary ventilation for , the emergency switchgear areas through the use of portable fans. However, the human reliability analysis for ventilation restoration was not revised to account for the procedure enhancements and, therefore, the CDF was not affected. Fast 4160V Bus Transfer Failure The specific sequence results in failure of the 4160V fast bus transfer and failures of the diesel generators which would lead to a station blackout condition. Recovery of , electrical power through repair of the fast bus transfer breaker was identified as one . method of mitigating the consequences of this event. A review of the existing procedures indicates that ECA 0.0 provides direction to the operator to transfer power : mr~. ally from offsite sources in the event of a failure and procedure 1/2.36.4A l provides specific direction for racking breakers in and out. It was also noted that the significance of this vulnerability is lessened somewhat by the installation of the 4160V ' 15 of 69 ,
., ,y , - - . . . _ 4 .
~ s y < .i.
^, .
j
-t 1
'i e
^ station crosstic capability l (see 'AC ' Power Generationi Capability below). LThis ' ! procedure enhaaramane was identified a having been implemented at the time ofIPE l submittal and, therefore, no changes to the model were required. i ay :: s Batterv Canacity for Steam Generator Level Instruments - '
- l This item is considered resolved as a result ofinstallation of the 4160V station crosstie l since one train of the emergency battery chargers will be powered from this source. j
/ Therefore, the anhaa=1 procedures on shedding battery. loads _is not necessary,.Th'e reduction in CDF associated with this vulnerability is realized by the installation of the ; - station crosstie. -
n - Reactor Trio Breaker Failure j
~
The specific sequences for this event are those that lead to an ATWS'followed 'by'.. !! opening of PORVs.. Procedures direct the operator to manually drive in the control ' j rods and to manually disconnect the Control Rod M-G sets' following failure of the ! reactor trip breakers, however, this must be done locally .and there may be inadequate l time to prevent overpressurization of the RCS. With the changes made to the_PORV . ! ATWS model (see response to Front-End Question 4, above), the contribution to core : i damage frequency due to ATWS events is significantly reduced. ' Based on this,' the , j installation of capability to remove power from the control rods from the control room . j is not considered warranted.. 1 1; AC Powr Generation Capability
.l For stdion blackout sequences, both onsite.'and offsite recovery actions to reestablish .;
power to 4160V emergency AC electrical buses are important.; Installation' of the . l station crosstie connecting the 4 kV normal buses of Beaver Valley Unit 1 and BeaverL t Valley Unit - 2 is. now complete. The PRA model was revised to reflect this
. q modification, which now takes credit.for the Unit 2 emergency diesel generators,~ if- j j
both are available, given the failure of both Unit 1 emergency diesel generators and the; loss ofoffsite power. ' j RCP Seal Cooling for Station Blackout l
'r No model changes to specifically address reductions in Seal LOCAs were included; l howefer, the CDF associated with Seal LOCAs has been greatly reduced: by- j installation of the 4160V st'ation crosstie. Seal injection can be provided within I hour L d of a station blackout using the 4160V crosstie. Additionally, new RCP seal materials j j
will be installed on a replacement basis as ' stock of current spares is expended. It is . . expected that operating with the new seals will greatly extend the time' available for recovery during station blackouts. Duquesne Light will . evaluate any^ modifications required by any future NRC rulemaking with regard to RCP seal cooling. , l 16 of 69
,, . , , . - ,. . ,>- m -
7 , b y; l f, C
-r i
~
; Pressudzer PORV Stickinn Onen After Loss of Offsite Power
- , i m 7 - -
,. ' Most Westinghouse plants do not experience a PORV challenge following a loss ofI l
offsite power or a loss ofload._ The IPE submittal assumed. Beaver Valley would
&y ' experience a PORV challenge because ofit's 100% load rejection capability, which was : ,
assumed to eventually be' ua=MMI.;This; assumption was;basedion RETRAN ; 4 analyses for the Diablo Canyon plant, which also has a 100% load rejection capability.' 4! 3 1Therefore, the loss of offsite power sequences consist of an unsuccessful'100% load : 1 rejection which results in'a delayed reactor trip and a challenge to a PORV. If the - n1
- PORV sticks open and'a loss of onsite power occurs, a small LOCA will result.which L '
_ , greatly shortens the time available for electric power recovery. This is also true for the 2 y loss ofload sequences if a fast bus transfer failure occurs. An evaluation of the events' > and specific failures which result in PORV challenges with a loss ofisolation capabilityL y was performed. Two initiating events were identified as affecting the 100% loadi
' rejection capability, a loss of both'345 kV and 138 kV lines'(initiating event LOSP),1 .i
( - and . a loss of only.the 345 kV line (initiating event TT).. ;j s Table 6.3-2 " Beaver Valley ' Unit.1_ Potential Enhancements" lists the L pressurizer '! PORV sticking 'open after a loss 'of offsite power as a vulnerability,with; a 2.0% . j contribution to core damage. By taking credit for the stsbn electric power crosstie,'. j further reductions in this vulnerability's contribution to the' core ' amage d frequency? Lj j were gained. This is due to an AC electric power train being available tol enable the?
"i operator to close the stuck open PORVs' block valve. With the revised PRA model,; :
these sequences now account for approximately 0.4% of the new1CDF and, therbfore,; ! are no longer considered a vulnerability.~ j q A pressurizer PORV sticking open after a~ loss of load (i.e [ turbine trip) initiating event contributes less than 0.1% to the core damage frequency,' and is not considered a i
]
p plant vulnerability. Therefore, the design enhancement to eliminate a PORV challenge' j
- by defeating the 100% load rejection capability is not necessary. :!
- s
'!y
.q i
j: N 1! d i I t 1
;{
i 17 of 69 y .
~l
f
. Additionally, it was identified that Table 3.1.1-2, on page 3.1-10 of the IPE submittal - ;
' has some formatting errors associated with the "Offsite Grid" subsystems. .The i following table reflects the correct formatting for the Offsite Grid subsystems. j i
Inspect on Safety System (s) or Initiating : System /Sebsystems Key Plant Equipanent Event . Comment ! Category / Code Designator j Offsite Grid ;
- 345-kV Line TurbineTrip . .
9/IT Resultsin turbine /generatortrip l Reactor Coolant Pumps (RCP) but equipment listed is , ' Main Feedwater(MFW) repowered when fast transfer - [ Condensate. . . to 138-kVline is complete. l Turbine Plant Component Cooling Water ! Reactor Trip 138-kV Line None - Does r ot cause a plant trip. Both 345 and RCPs- 16/LOSP Results in plant trip. Equipment r 138-kV Lines MFW listed is unavailable. Equipment i Condensate normally operating and powered ! Turbine Plant Component Cooling Water from emer. buses must restart. - -) l a Pressurizer PORV Block Valve Alinnment 1 i See the response to Front-End Question 4, above. . 1 Containment i i The containment building issues were discussed and concluded as being potential j actions to be included as part of a severe accident management program. These will i be investigated furtner as the Duquesne Light Company severe accident management : i program progresses. The revised PRA model did not include any changes to the Level l 2 analysis. i Operator Actions for SGTR Events
)
It was originally identified that during a SGTR in which all HHSI fails, the procedures directed the operators to depressurize the RCS but only after the sequence has ! progressed to extreme conditions in which partial core uncovery has- occurred.- l However, earlier depressurization under these conditions may have prevented core' l damage and terminated releases earlier, and therefore would be beneficial. Upon. , further review of the Emergency Operating Procedures (EOPs) and discussions with ; procedure writers, the following was concluded: j ~ Operator responses to a SGTR are governed by EOPs E-0 and E-3. These ; procedures do not provide specific guidance for conditions where a failure of i all HHSI has occurred but rely on the status trees for core cooling to provide j criteria for exit into the function restoration procedures. The criteria in the l status tree rely on core exit thermocouple temperatures and RVLIS indications ! 18 of 69
=
y s b <
, 3 to identify conditions which are indicative.of core cooling de6ciencies. During
~
d a' SGTR it is not anticipated that these conditions would occur- prior.to- 1
~
~
- accomplishment of depressurization as required by E-3. Therefore, as.long as -
the existing procedures were followed in parallel with efforts to restore HHSI,- ; ifit were recognized to be unavailable, the desired actions to;depressurize the - RCS would be accomplished. Therefore, no procedure changes are required;- ' -)
., however, operators should be trained to continue following E-3 even if HHSI - ]
. completely fails. 1 i
Also for SGTRs it was identified that under certain circumstances a stuck.open safetyk
. relief valve on a steam generator may be locally gagged to isolate the mptured steam . .!
generator. It was determined that this' type of action, if needed, would have"to be . ; evaluated based on existing conditions. which ;would likely' be directed from- the - -
' ~
Technical' Support. Center of the Emergency. Response Organization.' Therefore, - ' guidance on this type of action will be included in'the severe accident management
;' guidelines.' ,
1 No change to the human error rate assigned to these actions was made in the revision {
. to the PRA model. ,
i i i
.I L
k i
'i i
'h 19 of 69 .
_ _. .__ _ _ ~ . _
y( , - [ , N l
. 7. - Flooding at the Intake Stmeture is identified as'36% of the CDF due to flooding at Unit 1.
Beaver Valley Units 1 and 2 share a common Intake Structure and, therefore, may be subject-
~
to a dual unit initiating event / Please _ discuss possible flooding events at the Intake Stmeture : lI
.which may impact both units, and the ability to cross-tie or share equipment.which 'would ;j normally be credited in the analysis. ,
j Response- ! d b The Beaver Valley Intake Stmeture houses the Unit'l River Water pumps an'd raw water j pumps, the Unit 2 Service Water pumps, the fire water pumps that supply both units, and some j emergency Motor Control Centers (MCCs). The pumps and the MCCs are contained in I cubicles with normally closed security doors that open into the cubicles and have no gaskets on i the bottom. The following table summarizes the content of the intake structure pump cubicles: ;f
- CUBICLE A CUBICLE C 1 Unit 1 River Water Pump (WR-P-1A) Unit 1 River Water Pump (WR-P-1C)) !
Unit 1 Raw Water Pump (WR-P-6A) Unit 2 Service Water Pump (2SWS-P21B)
Motor Driven Fire Water Pump (FW-P-1) Unit 2 Emergency MCC (MCC-2-E02) !
Unit 1 Emergency MCC (MCC-1-EI)
CUBICLE B CUBICLE D Unit 2 Service Water Pump (2SWS-P21 A) l Unit 1 River Water Pump (WR-P-1B) Unit 1 Raw Water Pump (WR-P-6B) _ ]
Unit 2 Service Water Pump (2SWS-P21C) Diesel Driven Fire Water Pump (F N-P-2) - ]
Unit 1 Emergency MCC (MCC-1-E2) Unit 2 Emergency MCC (MCC-2-E01) 1 1
Each cubicle contains a sump _with a level switch and sump pump. A fire door connects _;
Cubicle A with Cubicle B, and another fire door connects Cubicle C with Cubicle D. These !
doors do not have any gaskets, so water is assumed to propagate only between Cubicles A and -
ll B, or between Cubicles C ud D. No credit.is_ taken for propagation outside a cubicle,x although the main ~open floor area has sufficient grating that would allow drainage back to the river. Additionally, no credit for operator actions to isolate the break prior to pump failures is - ;
taken. j The Unit 1 IPE Intake Structure flood modeled a large floN from the river water, raw water, 1 fire water, or service water piping in Cubicle A or B as hiling the normal river water pumps . -l and associated valves,' and the "A" train of the Raw W2ter System (the Fire Water System is . :'
not modeled in either unit's IPE). No credit was taken
> ; ~s
,y Qf :!; ,
y
, J N l 1
C, 1; In the discussion of Top Event 18, Page 4.6-17, you state that the large majority of hydrogen- [
. burn failures are due to detonations as compared to deflagrations. It is also clear from Page !
4.2-6 that large amounts of hydrogen can enter the Containment atmosphere. Thus, it is 'not, ! clear why you state that understanding the uncertainties associated with hydrogen mixing,. transport and detonations are "...beyond the scope of this submittal." L (Page 4.6-17) ;Please .; explain. !
-r Resnonse-m The statement on Page 4.6-17 states that,'"the. uncertainties associated .with predicting; j hydrogen mixing and transport and the magnitude of dynamic loads-' associated with
~
y detonations are beyond the scope of this submittal." Because the Beaver Valley and'Surry j plants are essentially " sister" plants, the BV1 Backend analysis relied heavily on the insights l j obtained from the analyses performed for Surry for NUREG-1150 (Ref.1). Accordingly, detailed evaluations of hydrogen mixing and transport in the containment were not performed, j nor was any pressure load analysis (static and dynamic) performed for BVI. The BV1 analysis : j did consider the possibility oflocally high concentrations of hydrogen and the possibility.for l Deflagration to Detonation Transition (DDT). ' Based on the simplified model,.allarge ! containment failure was predicted to occur when the underlying conditions for DDT 'were j predicted to occur. The approach used for DDT is discussed in the paragraphs which follow, j j As noted in the discussion for Top Event 11 - Containment Failure Prior to Vessel Breach on l Page 4.6-14 of the BV1 IPE submittal, the potential for containment failure prior to vessel 1 breach, due to hydrogen burns, was discounted for Surry (Ref.1) on the bases of the ] containment strength and the amounts of hydrogen that are produced prior;to vessel breach.- ; This contention was accepted for BV1 except for the case when a large ainount of hydrogen .is j generated in-vessel, and the hydrogen is suddenly released (i.e., a thermally induced hot leg :l failure) into a non-inerted containment (i.e., containment ~ sprays operating prior to piping j failure). However, to minimize the number of CET top events related to burns, containment- 'l failures related to induced piping failures were addressed at Top Events HE, CE, and LE rather = -j than as a containment failure prior to vessel breach. l q There is no specific statement in the discussion for Top Event 18 that states, 'that the large 1 majority of hydrogen-burn failures are due to detonations as compared to deflagrations.". It is' ; stated that for the case ofinterest (sprays operating, large quantities of hydrogen generated in. 1 vessel, no burns prior to or at vessel breach), the mean probability of containment failure ~of f 0.4_ (given that a burn occurs)is dominated by the assumption that any burn that occurs at a l hydrogen concentration greater than 12% propagates to a detonation. It should be noted that a .! value of 0.38 (see conditional split fraction CEF) was used in the Beaver Valley Unit 1 IPE study. This value corresponds to the probability that the quantity of hydrogen generated in-vessel exceeds that required to' achieve a global hydrogen concentration- of 12% in the containment based on dry air. 22 of 69
. . ~
9 MAAP analyses performed for Beaver Valley indicated that for most severe accident scenarios,- burns would either be precluded by steam inertion or would occur when the hydrogen concentrations reached global flammability levels as determined by the MAAP algorithm. An exception to this observation is when a large amount of hydrogen is suddenly released into a non-inerted containment. As noted on Page 4.6-17 of the BV1 IPE submittal, only thermally induced hot leg failures were assumed ia the evaluation.- As noted in the discussion for conditional split fraction CEC in Table 4.6-4, hydrogen is expected to burn at ' global" concentrations below 12% (i.e., burning in the cavity and/or local burning as the flammable gas leaves the cavity) during pour type vessel failures at RCS pressures greater than 200 psia' and sprays in operation. The burn pressures calculated by MAAP were significantly less than those which would result in a significant probability of containment failure. Hand calculations for adiabatic burns (deflagrations) up to the concentration limits for detonation indicated that the pressure rises associated with 'real"(non-adiabatic) deflagrations were not likely to result in containment failure. The dynamic loads associated with detonations are difficult to calculate and containment , strength criteria for these types of loads were not available. Accordingly, the BV1 IPE adopted what was believed to be a conservative treatment for detonations and consequent containment failure. It should be noted that NUREG/CR-4551 did not address detonations for Surry.
- As noted on Page 4.2-3 of the IPE submittal, the analysis of hydrogen combustion for the Surry plant for NUREG-1150 (Ref.1) assumed that if electrical power were available during the period of hydrogen generation, "the sprays will keep the steam concentration low and sparks from electrical equipment will cause ignition near the lower deflagrable limit", .
preventing significant concentrations of hydrogen. Based on the extent of mixing promoted by spray operation and the relatively low ignition energy levels required for ignition, this argument was assumed to be valid for BV1 as well, except for the sudden release of hydrogen into the containment (e.g., vessel blowdown at high pressure after severe core degradation). The peak pressures associated with detonations are well above the quasi-static pressures associated with deflagrations. However, the energies required for detonation are many orders of magnitude above those required for deflagration. As noted in Reference 2, detonation initiation within a range of hydrogen concentration from 18 to 59 volume percent (the ' approximate range of hydrogen detonability) requires an energetic ignition source, severe confinement, and/or a sufficiently large volume of gas mixture. Reference 2 concluded that, "the energy levels required to directly initiate detonation are orders of magnitude greater than ! those necessary to initiate burning at the same hydrogen concentration", and that "a de facto transition to detonation is highly unlikely in reactor containment buildings particularly when there are high steam concentrations or hydrogen concentrations below about 18 volume percent" Minimum ignition energies of 4100 joules have been reported (Ref. 2) for hydrogen- l air mixtures. According to Reference 2, this energy level is several orders of magnitude higher than would be produced from an electrical spark caused by contact arcing or by electrostatic discharge, and approximately eight orders of magnitude higher than the minimum ignition energy required to initiate deflagration. 23 of 69
, 1' ' -
; y m ,- 1
<4 >
.l J
u j q
-If AEical power is not available, the containment; sprays will not operate, andithe j containment is likely to be inerted by highLconcentrations of steam. .When steam inertion.
Df prevents combustion, the recovery of electrical power and containment ' sprays becomes a
. concern, since operation of the sprays will condense'the steam anl drive.the gas mixture' ,
P - towards the flammability range. The Surry analyses performed for NUREE 1150 assumed that' ~ d g hydrogen would be ignited.and burned as soon a the gas mixture entered the flammability . j
, ~'
range, guaranteeing that the burn would occur at low hydrogen concentration. Operation of ' the containtnent spray would guarantee substantial mixing. The recovery of AC power during or after core degradation was not addressed in the IPE - :j n: submittal. Because of potential deleterious ~ effects (such as containment _ deinerting) the'. j strategy for recovery. of mitigating systemrisuch as containment sprays will be carefully. p examined and fully evaluated in ths context of an accident management program. j> As noted earlier, for scenarios in'which the containment sprays are operating, it is likely that ! hydrogen burns will occur at low concentrations when hydrogen is " slowly" released into the: 1 containment. Only when the hydrogen is suddenly released into the containment (e.g., due' to an induced failure of the hot leg or at vessel breach), will the hydrogen concentrations achieve i significant values. When vessel breach is accompanied by a High Pressure Melt Ejection - -! (HPME), the containment loads discussed for Top Event C2 include the contribution of" j hydrogen burns. However, for " pour" type vessel breaches at high pressure, there could be a - :! sudden release of hydrogen into the reactor csvity and then into the containment. . Pour type: j failures are unlikely for high pressure. Nevertheless, such events were addressed at Top Events. 1 HE, CE, and LE. However, as noted earlier, containment failure due to burns following pour j type melts was deemed to be unlikely.
] i j
For those scenarios in which there was a sudden release of a large quantity of hydrogen into a t non-steam inerted containment atmosphere following an induced hot leg piping failure, it was - j assumed that'if the global concentration exceeded 12%, .a burn would occur which would, in 'j turn, fail the containment. The logic implicit in this assumption is as follows. 3 3
- 1. A deflagration at a 12% hydrogen. concentration is not likely to fail the BV 1 l
containment (based on peak containment pressures determined using the adiabati,
~ burn assumption).
- 2. Although MAAP simulations showed that the containment was relatively well mixed l when sprays were in operation, it was assumed that local concentrations could be 20% j higher than the global concentration. ;
q
- 3. Although the BV-1 contaimnent configuration is not' necessarily ananable to a l deflagration to detonation transition (DDT), it was assumed that a DDT would occur >
iflocal concentrations exceeded a value of 15% (minimum value reported in Ref. 3). j 1 1 L
- 4. It was assumed that a DDT would result in a large containment failure. j
.i 1
24 of 69 fl a f
?
_.7 _ , M& '
~
l : .{ g- 4 y s6 l
' Figure 4.2-1 of the BV1 IPE submittal (based . on the in-vessel ' hydrogen generation ' !
,^~~.
distributions reported in Volume 2 of Ref.1),'was used to determine the probability that'the; l
. amount of hydrogen generated in-vessel would exceed a level necessary to produce a global j concentration of 12%. This probability was estimated to be 0.38. andlwas used'as the ' split . .
'i
; fraction value for Top Events C2 and CE when vessel blowdown occurred at high pressure in -
'the absence of HPME.
] ;
- The references used in this response ~are as follows: 1
. l
.l. ' Breeding, RJ., et al, " Evaluation of Severe Accident Risks: Surry Unit 1", NUREG/CR 1 4551 (SAND 89-1309), Volume 3, Revision 1, Parts 1 and 2, October 1990
]':
- 2. IDCOR, " Hydrogen Combustion in Reactor Containment Buildings", Technical Report 12.3, September 1983 i
- 3. Sherman, M.P., et al, " FLAME Facility...The Effect of Obstacles and Transverse Venting i on Flame Acceleration and Transition to Detonation for Hydrogen-Air Mixtures at Large j
- Scale", NUREG/CR-5275, April 1989 1
,l i
' 2. For split fraction CEC (Page 4.6-38), the failure fraction assigned is 0.0, Why is Containment' ]
integrity ensured for all " pour type" failures when such large amounts of hydrogen are released = 'q into the Containment?
] q Response .l In the analysis of severe accidents performed for Surry (NUREG/CR-4551,'Vol. 3,-Rev.-1, . .l Part 2, Page A.I.1-61) by Sandia National Laboratories, it is concluded for sequences in which j debris pours out of the reactor vessel at breach, 'there may be'some local burns in the cavity !
but a general deflagration is not expected". MAAP analyses for Beaver Valley indicated that 1 burns would either occur in the cavity or as the hydrogen entered the lower compartment (i.e.. . j
' jet" burning). In either case, the associated pressure rise within 4 hours of vessel breach (the !
break point for early failures) is insufficient to challenge the' containment. The Sandia ! conclusion and the MAAP results provided the basis for the Beaver Valley assumpticos. 1 As a project objective, it was decided to take full advantage of the' physical similarities between j
' Beaver Valley Unit I and Surry, and deviate from the NUREG/CR-4551 analyses only when l warranted because of significant physical differences. This approach ~was believed to be. !
entirely consistent with the IPE requirements identified in Generic Letter 88-20 and.the j guidance provided in NUREG-1335. -i
-l 1
- 3. Regarding the type of concrete in the basemat, you state that it will behave like the " siliceous" concrete of Suny, i.e., produce little CO' and H2 during Core-Concrete Interactions (CCI)
(Page 4.5-9), even though the concrete has five (5) times the amount of CO 2 and two (2) times the amount of H 2O that typical " basaltic" concrete contains. How would the results be impacted because of higher gas formation from CCI?
> 25 of 69
' ~ ' ' ~ ^ ' " ~ ^^
qq.
' ~ ~
. n. -[^~1 e71 -
31
,m l
w 1.? Response:
]
i s
.There is no statement on page 4.5-9 or anywhere else in the Beaver; Valley Unit 1 IPE- !
- submittal that suggests that ,the quantities of,CO and H2 Produced during' core-concrete! .i p'
- interactions are small. : In fact, in the discussion for Top Event 21 - Late Burn of Combustible 'l Gases (H3) which begins on page 4.6-18 of the submittal, it is stated that "if the debris remains i 1
$ ' in the reactor cavity, large amounts of hydrogen and carbon monoxide can be generated." It is' - j ~ ' ,! also not clear'why the reviewer makes the comparison to basaltic concrete, since no use was made of any data for that type of concrete. 1 hT " In the discussion for Top Event 21, it is noted that MAAP calculations for a fast stationL blackout (loss of all ac power and'feedwater) sequence with a large RCP seal LOCA (which s j a minimizes the amount of water in the vessel at the time of vessel breach and hence the amount i ex-vessel debris cooling) indicate that large amounts of hydrogen are produced by the CCI and ; the hydrogen bumed as it left the cavity in the form of an annular jet and entered the lowerj l j compartment of the containment. . This combustion was observed in the MAAP. results'to '
. _ l continue until the oxygen in the containment became sufficiently depleted and to' deposit large . O quantities of energy to the containment atmosphere, increasing its temperature and pressure j dramatically. At the time of the evaluation, this burning process appeared to be independent 'of j whether or not the receiver companment (i.e., the lower companment) was inerted. There wasf j
~
uncertainty at the time the analyses were performed relative to the ability of MAAP to model ' ~q
, this phenomena. The Surry analysis assumed that if containment sprays were available, the: j hydrogen emanating from the cavity would burn as soon as flammable concentrations were 1 achieved and many small burns would occur but these would not challenge the containment. j On the other hand, .the Surry analysis assumed that .if; sprays were not available, the-
- i containment atmosphere would be inerted," precluding burns. However, as noted on page 4.6-- 1' 19 of the Beaver Valley Unit 1 IPE submittal, it was assumed that if the debris is_not being-cooled, there is a 50% chance that there will be a hydrogen burn at this time (at Top Event 21), , j regardless ofwhether or not the lower compartment was inerted. Furthermore, it was assumed :
that containment failure was guaranteed, given this burn, if containment heat removal was not lj available. In the analysis of long term containment behavior during severe accidents whic'h- was performed for Surry (NUREG/CR-4551, Vol. 3, Rev.' 1, Part 2, Page A.1;1-85) by Sandia q National Laboratories, it was noted that " eventual OP (overpressure) of the Surry containment 1 due only to the noncondensible gases generated by CCI (core-concrete interaction) is not } '
, credible." This reference'also states that "the concrete forming the Surry containment is[ -
siliceous ........... so the amount of noncondensible gases produced by CCI will be much - smaller than if it were composed of limestone concrete." ~ Furthermore, "the ' amount of concrete that would have to be decomposed to overpressure the Surry containment is such thatl the CCI wouldpenetrate the basemat before enough gas was generated." As noted in'page
'4.5-9 of the IPE submittal, based on an evaluation of the petrographic reports for the plant, the Beaver Valley Unit I concrete was also classified as siliceous and it was " concluded that the observations for Surry regarding the minimal contributions of noncondensible gases to containment overpressure are valid for Beaver Valley Unit I as well." =
26 of 69 a . -. - ... . _ , - - .- - -= . - .- , - -- - .-
AF, , i
# e
-, e s a -
j w < ' j As noted on page 4.6-20 of the IPE submittal, there is somewhat of a race between long term , 5 : overpressure failure modes and the basemat melt-through mode. Sequences involved in either.
- l type of failure were binned to Release Category Group III. (Late Containment Failures) which, f 1
.- ; keeping in mind that no credit for system recovery was taken after the inception of core 1 j damage, amounted to'43% of the CDF. In the Beaver Valley Unit 1 IPE,'it was assumed that basemat molt-through would eventually occur (CET Top Event 26 [BI]) if the debris was not ; 'l
~
cooled ex-vessel. It was also assumed that in the absence ~oflong-term containment heatl removal, containment overpressure failure was assumed to occur first.(CET Top Event 24 3'
'[C4]), independent of basemat attack. Thus, the only Release Category Group III sequences -
which are in question are those in which long term containment heat removal is available but ! H concrete attack could not be prevented and eventual basemat melt-through occurs.nSince a containment - heat L removal ;is 'available, the only mechanism ; for? overpressurizing a the L . containment.is the production and accumulation of noncondensible. gases. As noted in the .. Surry evaluation, overpressurization solely by this mechanism alone is not credible.: !
- 4. In describing _ the high-level simplification of the cavity wet / dry issue in defining r' elease
~
~
categories, the following is noted on Page 4.7-1 of the submittal: ' j "The last case is conditional on the loss of high head ECCS injection, ~or- recirculation with vessel pressure remaining high, or loss of high and low head ECCS injection with the operators - ) failing to go into "early" ECCS recirculation (i.e., with water still in the RWST): in other 1 words, successful Quench and Recirculation Spray operation and lineup : for vessel injection, j but with ECCS failure." j Based on your last phrase, " successful Quench and Recirculation Spray operation and lineup j for vessel injection, but with ECCS failure", it is not clear what system failures and operators actions are involved in this case. Please explain this accident scenario in more detail as it - 1 relates to the source term calculation. : I
Response
Table 4.7-1 notes that a ' Wet'# cavity provides a means of cooling (the debris)or at least - 1 attenuating the fission product release from the core debris. ;It should be notsd, however, that . j for the BV1 source term analysis, the reference to cavity wet or dry was dropped and all~ source term cases were conservatively- treated as if the cavity were dry -(see p.4.7-2). j Therefore, no credit was given for the scenarios in question. ll
. . .. 4i
. As noted in Section 4.3.3.1, because of the BV1 reactor cavity / instrument tunnel ( keyway)L j configuration relative to the containment, it is impossible for water in the containment to 'hpill'. ' l into the keyway. As indicated in Section 4.3.3.1, during operation of the QS, a small amount j of water (140 gpm if both QS pumps are operating) is diverted to the reactor cavity.l If the QS j pumps are the only pumps taking suction from the RWST, the maximum quantity of water - j directed to the reactor cavity prior to emptying the RWST is approximately 12,000 gal. The j spray pattern of the RS does not result in any accumulation ofwater in the reactor cavity. j 27 of 69
)
a
, 0, <
3 The discussion on pdge 4.7-1 attempts toldefine~ the possible methods in which 'a continuous ! p supply of water can be provided to the reactor cavity via the reactor vessel after vessel breach and explain how such sources can be available in the long term when they were apparently l unsuccessful in preventing core damage. -If low pressure systems are available. but high l' pressure injection and/or recirculation were unavailable and the RCS remained at high pressure -
-prior to vessel breach, core damage could not be prevented and the low' pressure _ systems :
would not operate until the RCS pressure dropped below their pump shutoff heads at. the time -l
- of vessel breach. The accident scenarios ofinterest address the situation where normal ECCS ' j
. is unavailable but the QS and RS are successful and the outside RS pumps are manually aligned l .
for vessel injection, thereby providing a' pathway for water to the reactor cavity via the RCS and the hole in the bottom of the vessel. .
?
- 5. Please address the following items related to containment isolation failure: j (a) With respect to the analysis of containment' isolation failure probability, NUREG-1335L (Section 2.2.2.5, Page 2-11) states that "the analyses should address the five (5) areas -
identified in the Generic Letter, i.e., (1) the pathways that could significantly contribute = l to containment isolation failure, (2) the signals required to. automatically isolate the. , penetration, (3) the potential for generating the signals for all initiating events,~ (4) the ; examination of the testing and maintenance procedures, and (5) the quantification of each - , containment isolation failure mode (including common-mode failure)". ~ Please discuss !
.your fin di l dto t eha b ngs re ate ove vefi (5) areas.
Response , 1 Containment isolation failures for the' Level 2 analysis are governed by the Level 1 Top i Event CI model discussed in Section 3.2.1.16 of the IPE submittal. Pre-existing leakage -{ paths via penetrations thought to be closed but due to unforeseen circumstances left open ;I were not modeled because of the subatmospheric containment design; i.e., that any such~ . :i paths would be detected due to the demands on the system maintaining the vacuum.7The .i first two areas identified in' the Generic Letter; i.e., (1) the pathways thaticould; ; significantly contribute to containment isolation failure, and (2) the signals required'to j automatically isolate the penetration, are addressed in Table 3.2.1.16-1, . which is . i explained in Section 3.2.1.16.2. In response to. the examination of the testing and maintenance procedures (item 4), Section 3.2.1.16.7. on page 3.2-106, notes that all- : penetrations whichw ' ere not screened out from Table 3.2.1.16-1 require a Type C leak' j test and quarterly operability verification per OST 1.47.3A. In addition, maintenance is: j performed 'on an as-needed basis and operability checks are performed after feach -l maintenance event. The potential for generating the isolation. signals (item 3), listed in ! Table 3.2.1.16-1, for each initiating event analyzed in the IPE is listed below.
.?
i l 28 of 69 )
,, n 3
1 w '
.-p' f ; :1 CIA: ' SLOCI-isolable small LOCA ' !
SLOCN- nonisolable small LOCA l MLOCA-medium LOCA !
' \~
LLOCA-large LOCA j ELOCA'- excessive LOCA 1 2 VSX -interfacing systems LOCA. . - 1 SGTR- steam generator tube mpture
' SLBI - steam line break in one steam generator SLBC - steam line break in common RHR valve line
~
SLBD -steam line break down stream of the main steam isolation . valves , AMSIV - closure of all main steam isolation valves , IMSIV - closure of one main steam isolation valve MSV - main steam relief or safety valve opening , , ISI - inadvertent safety injection signal ._
.j CRFL - control room HVAC equipment area internal flood :- ;
. . i CIB: SLOCI-isolable smallLOCA-SLOCN- nonisolable small LOCA~
1 MLOCA- medium LOCA LLOCA 2 large LOCA i ELOCA- excessive LOCA j ' SLB1 - steam line break in one steam generator . CRFL - control room HVAC equipment area internal flood '. t The quantification of each containment isolation failure mode including common cause - (Generic Letter item 5) are addressed in Tables 5-1 and 5-2, which are reports generated . ! j from the PRA containment isolation (Top Event CI) model. A brief description for each - of these reports is discussed on the following pages. <j l i i l
, 1
, '1 29 of 69
M' ~, V , Table 5-1. ' This table consists of the containment isolation common cause failure modes, which l
- were developed using the Muhiple Greek Letter (MGL) m. odology. Incorporated l ints this table' are the common cause group identifiers basic events,that are !
affected in the group, the order of the common cause te mode modeled,. the : , failure mode, and the~ database variables that were: used to . quantify the MGL - , equations. .
.?
I I I t _r i i l i i 6
?
s l l
-I
}
l 1 30 of 69
w u p
- m. ,
- h. ,
n ,
+ , ;
f
.p
.)
Table 5-1. Containment Isolation Common Cause Report Page1of3-
.i MODEL Names BV1 l CCF Model Report for Top Event CI 3 13:22:4 -30 JAN 1995 Pete l' -;
T" 7 Gro @ 10 : A- Basic Events . Description j AVFCTV10A100A TV DA 100A FAILS TO CLOBE l
. - AVFCTV1DA1008 TV DA 1000 FAIL 5 TO CLOBE !
- 1 i
- s Algebraic Method MGL - .;
Order = 1 out of 2 > t Falture Mode ID : CLOBE -[ Total Falture Rate = ZTVAOD -l Beta = ZBVAOD ,
.i>
. Group ID 8 Basic Events Description- ..!
[- . AVFCTV1DG1005 TV-DG 1088 FAILS TO CLOSE.' AVFCTV1DG108A 'TV-DG 108A FAILS TO CLOSE r J[ Algebraic Method: MGL [ Order = 1 out of 2 + Failure Mode 10 : CLOSE Total Failure Rate = ZTVAOD Beta = ZBVA00 ,
-t i
Gro @ 10 : C Basic Events Description i AVFCTV10G109A2 TV-DG-109A2 FAILS TO CLOSE- ,
+
AVFCTV1DG109A1. TV-DG 109A1 FAILS TO CLOSE { Algebraic Method: MGL ' order = 1 out of 2 . t falture Mode 10 : CLOSE 5 i Total Failure Rate = ZTVA00 - - Beta = ZBVAOD :[ 1 i t i T 31 of 69 f I
y . v a.n 1 jP y
. Table 5-1. Containment Isolation Common Cause Report '
Page 2 of 3 ~ , i s MODEL Name SV1 j CCF Modet Report for Top Event Cl ; 13:22:51 30 JAN 1995 Page 2 i
~ Group 10 : D. Basic Events . Description 'l:
MVFCMOVICH381 MOV-CH-381 FAILS TO CLOSE { MVFCMOV1CH378 INSIDE CNMT ISOL VALVE MOV CH-378 FAILS TO CLOSE l 1
~
Algebraic Method: MGL. Order = 1 out of 2 ; Falture Mode ID : CLOSE , Total Failure Rate = ZTVNOD. Bets = 2BVMOD i
~.
Group 10 : E Basic Events Description ; FCFCLCV1CH460A LCV-CH 460A FAILS TO CLOSE FCFCLCV1CH4608 LCV E;H 4608 FAILS TO CLOSE AVFCTV1CH204 TV-CH-204 FAILS TO ClosE AVFCTV1CH200A TV-CH 200A FAILS TO CLOSE , AVFCTV1CH200C TV CH-200C FAILS TO CLOSE i AVFCTV1CH2006 TV-CH 2006 FAILS TO CLOSE Algebralc Method MGL' l' order = 3 out of 6 Failure Mode ID : CLOSE t Total Failure Rate = ZTVAOD Beta = 2BVAOD i r Gamma = ZGVAOD Delta = ZDVA00 . i f i Group 10 : F Basic Events Description v AVFCTVICV150A TV-CV 150A FAILS TO CLOSE AVFCTV1CV150s TV-CV 1508 FAILS TO CLOSE ! d
-Algebraic Method: MGL l Order = 1 out of 2
' Failure Mode 10 : CLOSE
' Total Failure Rate a ZTVAOD Beta = 25VAOD .;
i 1 i 32 of 69 1 i 1 l
3
. l.'f t
. o.
Table 5-1. Containment Isolation Common Cause Report .. Page 3 of 3 ! L MODEL Names BV1 i CCF Model Report for. Top Event Cl , 4
'13:22:55 30 JAN 1995 Page 3 ,
Group 10 : G. Basic Events Description E AVFCTV1CV1500- TV-CV 150D FAILS TO CLOSE I AVFCTV1Cv150C: TV-CV 150C FAILS TO CLDSE i Algebraic Method: MGL l Order = 1 out of 2 i Failure Mode ID : CLOSE. f Total Falture Rate = ZTV400 ) Bete = ZBVAOD ;
.i Group ID : H Basic Events Description AVFCTV1CV101A TV-CV-101A FAILS TO CLOSE l i
AVFCTV1CV101B TV CV 1018 FAILS 70 CLOSE ! 5 Algebraic Method MGL Order = 1 out of 2 t Fallu's Mode ID : CLOSE Total f ailure Rate = ZTVA00 l Bets = ZBVAOD j i i s Group ID : I Basic Events Description f AVFCTV1LM103A1 TV-LM-100A1 FAILS 70 CLOSE l AVFCTV1LM1DDA2' TV LM 10DA2 FAILS TO CLOSE l Algebralc Methods MGL l Order = 1 out of 2 i Failure Mode ID : CLOSE
' Total Falture Rate = ZTVAOD Beta = ZBVADD v s
Group 10 : J Basic Events Description # AVFCTV1CV102 TV CV 102 FA!LS TO CLOSE ! AVFCTV1CV1021 TV CV-102 1 Falls 70 CLDSE i i Algebraic Method MGL I order = 1 out of 2 i Falture Mode ID : CLOSE [ Total Failure Rate = ZTVADD i Beta = ZBVAOD i l
, 33 of 69 t
I u
}4 , -'yr 7
g >
+
~
. ... i 1 ," . ,
j p, , l r , . . [ t Table 5-2. This table provides the cause table for each of the contamment isolation split fractions L j [- L that were quantified by using the fault tree for Top Event _CI. The cause tables consist ;
~
of the quantified minimal cutsets for each panicular split fraction. These cutsets are ; a h-ranked in descending order according to their quantified values. Additionally, this table , -l shows the % Importance, or the' percentage that each cutset' contributes to the Monte _
~ Carlo mean' split fraction 'value, and the % Cumulative,"which is the cumulative ~!
summation of the % Imponance. The cause table reports were generated .by using a L ,
- 99.9% mmal=+1ve cutoff for each of the split fractions. The alignment of the system; I when the cutset,was. quantified is also provided. These'are all shown as being in * .
.j normal alignment, since no maintenance or tests are performed on an unisolated ;
I component during plant operation. It should be noted that singleton cutsets, whose ! l l m 4 - basic event identifiers are separated by a _ comma and enclosed in' brackets [ ], are common cause failures of components. Independent failures of common causei 'l components are shown as a single basic event enclosed in brackets. !
.l
.l 8
.]
j fi ff f , :
-i l
l
,j l
l 4
'i i
'\
-[
-2 34 of 69 l
l c i
- . . . . - .. - - . ,r
7 , ,- - - - n L . l l i Table 5-2. Containment Isolation Cause Table Report Page 1 of 6 l 1 w ' U MODEL Names BV1LVL1. '! l' - Cause Table for Top Event CI {
. Split Fraction Cli. ALL SUPPORT -l tt s.
PE Value of CI1 = 6.8151E-03 Date : 25 JUN 1992 19:05-
. MC/LN Vetue of C11 s 6.8701E 03 Date : 03 JUL 1992 02:51.
I 13:21:25 30 JAN 1995
~ Pa0* 1 - l[
No...'Cutsets........... Value..... % leportance % Cuuulative Alignment... j 1- (AVFCTVCV102,AVFCT 6.776E-04 9.8630 9.8630- NORMAL VCV1021)
+
2 (AVFCTVLM100A1,AVF 6.776E 04 9.8630 .19,7259 NORMAL CTYLM10042) i:
. l 3- (AVFCTVCv150A,AVFC 6.776E 04 9.8630 29.5889 NORMAL !
TVCv15083- . i 4 (AVFCTVCv101A,AVFC 6.776E-04 9.8630 39.4519 NORMAL , 1 TVCV1018] t 5 .(AvFCTVCV150D,AVFC 6.776E-04 9.8630 49.3149 NORMAL { TVCV150C3 ? g -6. [AVFCTVDA100A,AVFC 6.776E-04 9.8630 59.1778 NORMAL i TvDA100s] , 7 ' IAVFCTVDG10942, AVF 6.776E 04 9.8630 69.0408 NORMAL I CTVDG109A1] j 8 (AVFCTVDG1085,AVFC 6.776E 04: 9.8630 78.9038 . NORMAL f TVDG108A) :t I 9.4612 88.3650 ; 9 OPRC12 6.500E 04 NORMAL i 10 INVFCMOVCH381,MVFC 1.101E-04 1.6026 89.9676 NORMAL-MOVCH378] i t 11 [AVFCTVDG109A2]
- 8.296E-05 1.2075 91.1751 NORMAL [
[AVFCTVDG109A1) :; 12 (AVFCTVCV102]
- 8.296E 05 1.2075 92.3827 NORMAL {
(AVFCTVCv10213 13 (AVFCTYLM100A13
- 8.296E 05 1.kJTS 93.5902 NORMAL
[AVFCTVLM100A23 )
-i 14 (AVFCTVCV101A1
- 8.296E-05 1.2075 94.7978 NORMAL i' (AVFCTVCV1018) i 15 (AVFCTVCV15003
- 8.296E-05 1.2075 96.0052 NORMAL if (AVFCTVCV150C3
]
- 16 IAVFCTVCV150A]
- 8.296E-05 1.2075 97.2129 NORMAL j
[AVFCTVCV150s) 17 IAVNTVDG1085)
- 8.296E 05 1.2075 98.4205 NORMAL l IAVFCTVDG108A3 {
t 18 (AVFCTVDA100A3
- 8.296E 05 1.2075 99.6281 NORMAL :l
[AVFC1VDA1008) ; l' l i 35 of 69 -l
-t t
L. -
p t u Table 5-2. Containment Isolation Cause Table Report Page 2 of 6 MODEL Name 'BV1LVL1
' Cause Table for Top Event C1 n Split Fraction Cl2 - LOSS OF AC CRANGE
- PE Vetue of CE2 = 8.5235E 03 Date : 25 JUN 1992 19:05 NC/LH Value of C12 = 8.4332E 03 Date : 03 JUL 1992 02:51 13:21:40 30 JAN 1995 Page 1 No... Cussete........... Vetue..... % Importance 1 Cumulative At{gnment...
1 (MVFCMOVCH381) 1.715E 03 20.3364 20.3364 NORMAL 2 [AVFCTVDG1085,AVFC 6.641E 04 7.8749 28.2113 NORMAL TVDG108A] 3 IAVFCTVCV102,AVFCT 6.641E-04 7.8749 36.0861 NORMAL VCV1021) 4 IAVFCTVLM100A1,AVF 6.641E 04 7.8749 43.M10 NORMAL CTVLM100A21 5 (AVFCTVCV150A,AVFC 6.641E 04 7.8749 51.8359 NORMAL TVCV150s) 6' [AVFCTVCV101A,AVFC 6.641E 04 7.8749 59.7107 NORMAL TVCV101B] 7 (AVFCTVCV1500,AVFC 6.641E 04 7.8749 67.5856 NORMAL TVCv150C) 8 [AVFCTVDA100A,AVFC 6.641E 04- 7.8749 75.4605 NORMAL TVDA1008) 9 (AVFCTvDC109A2,AVF 6.641E-04 7.8749 83.3353 NORMAL-CTVDG109A1) 10 OPRCl2 5.%5E-04 7.0733 90.4086 NORMAL 11 [MVFCMOVCH381,N/FC 1.099E 04 1.3032 91.7118 NORMAL MOVCH378) 12 IAVFCTVDG109A2)
- 8.493E-05 1.0071 92.7189 NORMAL
[AVFCTVDG109A1) 13 [AVFCTVCV102)
- 8.493E-05 1.0071 93.7260 NORMAL
[AvFCTVCv1021] 14 [AVFCTVLM100A11
- 8.493E-05 1.0071 94.7331 NORMAL (AVFCTVLM100A21 15 (AVFCTVCV101A1
- 8.493E-05 1.0071 95.7402 NORMAL IAVFCTVCV101B]
16 [AVFCTVCV150D]
- 8.493E 05 1.0071 M.7473 NORMAL
[AVFCTVCV150C1 17 (AVFCTVCV150A]
- 8.493E 05 1.0071 97.7544 NORMAL IAVFCTVCV150B]
18 (AVFCTVDG1085)
- 8.493E 05 1.0071 98.7615 NORMAL (AVFCTVDG108A) 19 IAVFCTVDA100A)
- 8.493E-05 1.0071 99.7686 NORMAL (AVFCTVDA100s) 36 of 69 L-
I r . 8 b: ; u' i. Table 5-2. Containment Isolation Cause Table Report Page 3 of 6 ' MODEL Names BVILVL1 Cause Table for Top Event CI Split Fraction Cl3 - LOSS OF AC PURPLE > PE Value of CI3 = 8.9030E 03 Date : 25 JUN 1992 19:05 MC/LN Value of Cl3 = 8.8773E 03 Date : 03 JUL 1992 02:51.- , 13:21:55 30 JAN 1995 Page 1 ; t No... Cutsets........... Vetue..... 1 Importance % Cumulative Alignment... 1 INVFCMOVCH378) 1.701E 03 19.1612 19.1612 NORMAL 2 [AVFCTVDG1080,AVFC 6.731E-04 7.5823 26.7435 NORMAL TVDG108A) 4 3 (AVFCTVCV102,AVFCT 6.731E 04 7.5823 34.3257 NORMAL VCV1021) ! 4 (AVFCTVLM100A1,AVF6.731E-04 7.5823 41.9080 NORMAL CTVLM100A23 5 IAVFCTVCV150A,AVFC 6.731E 04 7.5823 '49.4902 NORMAL TVCV1508) 6 'IAVFCTVCV101A,AVFC 6.731E 04 7.5823 57.0725 NORMAL TVCV1015) , 7 IAVFCTVCV150D,AVFC 6.731E 04 7.5823 64.6547 NORMAL , TVCV150C) , 8 [AVFCTVDA100A,AVFC 6.731E 04 ~7.5823 72.2370 NORMAL TVDA1008) 9 (AVFCTVDG109A2,AVF 6.731E-04 7.5823 79.8192 NORMAL ; CTVDG109A1) 10 OPRC]2 6.118E 04 6.8917 86.7109 NORMAL , 11 CVFCCH369 3.642E 04 4.1026 90.8135 NORMAL 12 IMVFCMOVCH381,MVFC 1.093E-04 1.2312 ,92.0448 NORMAL MOVCH378) 13 [AVFCTVDG109A21
- 8.569E-05 .9653 93.0100 NORMAL IAVFCTVDG109A1) 14 (AVFCTVCv102)
- 8.569E 05 .9653 93.9753 NORMAL IAVICTVCV1021) 15 [AVFCTVLM100A1)
- 8.569E-05 .9653 94.9406 NORMAL
[AVFCTVLM100 2) i 16 IAVFCTVCV101A]
- 8.569E-05 .9653 95.9058 NORMAL 1 IAVFCTVCv1018) 17 (AvFCTVCv1500)
- 8.569E-05 .9653 96.8712 NORMAL (AVFCTVCV150C]
18 (AVFCT, 150A]
- 8.569E 05 .9653 97.8364 NORMAL ,
IAVFCTVCV150B]
.%53 19 (AVFCTVDG1088)
- 8.569E*05 98.8017 NORMAL (AVFCTVDG108A]
20 IAVFCTVDA100A)
- 8.569E-05 .9653 99.7670 NORMAL
[AVfCTVDA1008) 37 of 69 !
n
.;; w Table 5-2. Containment Isolation Cause T8ble Repott Page 4 of 6 -
H ,, MODEL Names BV1LYL1 Cause Table for Top Event CI Split Fraction Cl4 - LOSS OF $$PS TRAIN A
.l PE Value of Cl4 = 8.0151E-02> Date : 25 JUN 1992 19:05 MC/LN Value of Cl4 = 7.8372E 02= Date : 03 JUL 1992 02:51-13:22:11 30 JAN 1995 i s ,Page 1
- No... Cutsets........... Value..... % Importance . % Cumulative Allgrument...
j 1 (AVFCTVDG1088) 8.829E 03 11.2655. 11.2655 NORMAL. .j 2 LAVFCTVDG109A23 8.829E 03 ~11.2655 22.5310 NORMAL 3 IAVFCTVCv1021)- 8.829E 03 11.2655 33.7966 NORMAL f 4 (AVFCTVLM100A2] 8.829E 03 11.2655 45.0621 . NORMAL e 5 IAVFCTVCV1018] 8.829E-03 11.2655 56.3276 NORMAL 6 (AVFCTVCV150D] 8.829E 03 11.2655 67.5931 NORMAL l 7 8.829E 03 11.2655 78.8586 - NORMAL [AVFCTVCV150A) f 8- [AVFCTVDA1008] 8.829E 03 11.2655 90.1242 NORMAL
-9 [MVFCMOVCH381) -1.722E 03 2.1972 92.3214 NORMAL 10 OPRCl2 6.600E-04 .8421 93.1635 NORMAL-i 11 -[AVFCTVLM100A1,AVF 6.520E 04 .8319 93.9954 NORMAL $
CTVLM100A2] 12 [AVFCTVCv150A, AVFC 6.520E-04 . .8319 94.8274- NORMAL TVCV1508] 13 - [AVFCTVCV101A,AvFC 6.520E 04 .8319 95.6593 NORMAL TVCv101B] j 14 ' (AVFCTVCV1500,AVFC 6.520E 04
.8319 96.4912 NORMAL .;
TVCV150C)
- l 15 (AVFCTVDA100A,AVFC 6.520E 04 .8319 97.3232 NORMAL TVDA1008]
16 [AVFCTVDC109A2, AVF 6.520E 04 .8319 98.1551 NORMAL CTVDG109A1). 17 [AVFCTVDG1088,AvFC 6.520E-04 .8319 98.9870 NORMAL TVDG108A) (AVFCTVCv102,AVFCT 6.520E-04 .8319 99.8189 18
-VCV1021]
WORMAL i r 38 of 69 I i
L' N , t i
^
Table 5-2. Containment Isolation Cause Table Repott - Page 5 of 6 - [ L , .. MODEL Names BV1LVL1 , Cause Table for Top Event Cl
. Split Fraction CIS Lost 0F $$PS TRAIN B j
. . i PE Vetue of CIS = 8.0603E 02 'Date : 25 JUN 1992 19:05 >
MC/LH value of C15 e 8.0354E 02 Date : 03 JUL 1992 02:51 13:22:21 30 JAN 1995 ! Page 1 } l L- No... Cutsets........... Value.....- X Inportance 1 Cumalattve Allyment... 1 (AVFCTVDA100A) 8.966E-03 11.1581 11.1581 NORMAL -
'I 2 (AVFCTVDG109A1) 8.966E 03 11.1581 22.3161 NORMAL L 3 [AVFCTVCV102) 8.966E-03 11.1581 33.4742 NORMAL i 4 .[AVFCTVLM100A1) 8.966E 03 11.1581 44.6322 - NORMAL 5 (AVFCTVCV101A] 8.966E-03 11.1581 55.7903 NORMAL i
6 IAVFCTVCV150C) 8.966E 03 11.1581 66.9484 NORMAL 7 (AVFCTVCV1508) 8.966E-03 11.1581 78.1064 NORMAL { 8 [AVFCTVDG108A] 8.966E 03 11.1581 89.2645 NORMAL -! 9 ' (MVFCMOVCH378) 1.696E-03 2.1106 91.3751 NORMAL ;! 10 -[AVFCTVCV102,AVFCT 7.240E 04 .9010 92.2761 NORMAL l VCV1021) 11 [AVFCTVLM100A1,AVF 7.240E-04 .9010 93.1771 NORMAL F CTVLM100A21 { 12 (AVFCTVCv150A,AVFC 7.240E 04 9010 94.0781 NORMAL .) TVCV150s) .!
. 't 13 IAVFCTVCv101A,AVFC 7.240E 04 .9010 94.9791 NORMAL ;
TVCV1018) ., 14 (AVFCTVCV1500,AVFC 7.240E-04 .9010 95.8802 NORMAL TVCV150C) [ 15 (AVFCTVDG109A2,AVF 7.240E-04 .9010 96.7812 NORMAL l CTVDG109A1] { 16 [AVFCTVDG1088,AVFC 7.240E-04 .9010 97.6822 NORMAL TVDG108A) 17 [AVFCivDA100A,AVFC 7.240E 04 .9010 98.5832 NORMAL' h TVDA100s) ';
-f 18 OPRCl2 5.675E-04 .7062 99.2894 NORMAL j 19 -CVFCCH369 2.056E-04 .2559 99.5453 NORMAL !
20 IMVFCMOVCH381,MVFC 6.151E-05 .0765 99.6218 NORMAL l MOVCH378) { i
-t 6
l t i 39 Of 69 I
n , i. t Table 5-2. Containment Isolation Cause Table' Report ' Page 6 of 6 MODEL Names BV1LVL1 Cause Table for Top Event Cl Split traction C16 - LOSS OF ALL AC, SA & SS AVAILABLE PE Vetue cf C16 = 8.9247E*03 Date : 25 JUN 1992 19:05 MC/LN Value of C16 s 8.9281E 03 Date : 03 JUL 1992 02:51 13:22:36 30 JAN 1995
'Page 1 No... Cutsets........... Value..... % !aportance X Cumulative Allgreent...
2 1 OPRC11 2.289E 03 25.6381 25.6381- NORMAL 2 (AvFCTVDG108B,AVFC 6.702E 04 7.5066 33.1448 NORMAL TVDG108A3 3 IAVFCTVCV102,AvFCT 6.702E 04 . 7.5066 40.6514 NORMAL VCV1021) 4 (AVFCTVLM100A1,AVF 6.702E 04 7.5066 48.1580 NORMAL CTVLM100A21 5 IAVFCTVCV150A, AVFC 6.702E 04 7.5066 55.6646 NORMAL TVCv15081 6 IAVFCTVCV101A,AVFC 6.702E-04 7.5066 63.1713 ' NORMAL TVCV10181 7 (AVFCTVCV1500,AVFC 6.702E-04 7.5066 70.6779 NORMAL TVCv150C1 8 (AVFCTVDA100A,AVFC 6.702E 04 7.5066 78.1845 NORMAL TVDA1008) 9 IAVFCTVDG109A2,AVF 6.702E-04 7.5066 85.6912 NORMAL CTVDG109A13 10 OPRCl2 5.810E 04 6.5075 92.1987 NORMAL 11 IAVFCTVDG109A2]
- 8.441E-05 .9454 93.1441 NORMAL IAVFCTVDG109A1) 12 (AVFCTVCV1023
- 8.441E 05 .9454 94.08 % NORMAL IAVFCivCV1021) 13 IAVFCTVLM100A1]
- 8.441E-05 .9454 95.0350 NORMAL IAVFCTVLM100A23 14 [AVFCTVCV101A3
- 8.441E 05 .9454 95.9805 NORMAL IAVFCTVCV1018) 15 [AVFCTVCV1500)
- 8.441E 05 .9454 % .9259 NORMAL IAVFCTVCV150C1 16 IAVFCTVCV150A1
- 8.441E 05 .9454 97.8713 NORMAL
!AVFCTVCV15083 t i
17 (AVFCTVDG108B1
- 8.441E 05 .9454 98.8167 NORMAL l
[AVFCivDG108A3 [ 18 (AVFCTVDA100A1
- 8.441E 05 .9454 99.7621 NORMAL-(AVFCTVDA1008) ;
L I 40 Of 69 l
)
z, , [ f ji ,
!;a .
r i i
+
f- :j
~
~
l(b) Section 7, Page 7-1 of the submittal, notes the following:' W "_ The::operationiof'the containment at subatmospheric conditions andLthe _ continual- . monitoring ofin-leakage'make the likelihood of a pre-existing failure of containment 1 l 3
> isolation at the time of a severe accident negligible; :However, Table-1-3, Page 1.4-7,: ~!
1 S indicates that BV-1 has a containment isolation failure frequency of 3.48E-5 'per reactor year, which is associated with 16.3% of the total CDF./ Please explain this discrepancy.: : Why do the containment isolation failures contribute so high (16.3%) to the CDF?- j Response: ~ l Section 7, page 7-1 of the IPE submittal should read, "The operation of the containment . at- subatmospheric conditions and the; continual monitoring of in-leakage make the _
-l likelihood of a IREEE pre-existing failure of containment isolation at the time of a severe l
- accident negligible." The reasoning behind this assumption 7is that pre-existing) l containment isolation failures greater than 3 inches in diameter.would be obvious to thel j operator since he would be unable' to maintain ~ subatmospheric containment pressure.-
' This is noted on page 4.3-6 " Containment Isolation and Bypass Status" and on page 4.61 'j 14 " Top Event Large Containment Failure Prior to Vessel Breach (L1)". :The containment isolation failure frequency of 3.48 x 10-5 per reactor year (16.3% of the totali .
CDF) reported in Table 1-3, page 1.4-7 only addresses smallLcontainmeat isolation. .
' failures, i.e., smaller than 3 inches in diameter. The reason that the small containment ~- 1
- . isolation failure plant. damage states contribute 'so much to'the CDF. is because the :
majority of the failures (96.2%, based on the saved sequence database) are due to' the ! emergency switchgear ventilation failing (15.5%'of the total CDF), which results in the : guaranteed failure of all emergency power and consequently, containment isolation (Top i Event CI). The normally open RCP seal return line requires AC power to close. Failure; ; to isolate the RCP seal return line was modeled as a failure of containment isolation. l j l y I
]
i; ; 1 l! tl l
,j
'i
.i 1l
.1
\
i. 41 of 69 ,
- ~ . - -
i
.f
_ (c) Since the BV1 plant has a non-negligibly high containment, isolation failure, please explain and give the magnitudes of the contributors to the isolation failures at BV-1. Response-As stated in the response to Back-End Question 5.(b) above, guaranteed failures of u containment isolation due to emergency switchgear ventilation failures, contribute to 96.2% of the total frequency for small containment isolation failure plant damage states. No credit was given for manual isolation of the RCP seal return line for. sequences involving loss of emergency switchgear ventilation.' The loss of this ventilation was assumed to also result in failure of all vital instrumentation which would complicate the l action to affect manual isolation. Another 3.4% of the total frequency is due to guaranteed failures of containment isolation that are related to failures of both SSPS trains or the failure of one SSPS train and the opposite train of AC electrical power. These guaranteed failures of containment isolation account for 99.6% of the total small containment isolation failure plant damage states. Hence, only 0.4% of. the total frequency actually come from probabilistic failures of the valves which must close to effect containment isolvion. (d) As shown in Table 4.8-3, Page 4.8-60 of the submittal, containment isolation failures are involved in 10% of the large, early release group (RCG I). However, isolation failures are excluded from the table on Page 4.8-1, which lists the major contributors to RCG I, including rocket mode failures that contribute less than 0.01% to the group frequency. Please explain.
Response
As noted in Section 4.6 (page 4.6-14), "because the Beaver Valley Unit I containment normally operates at subatmospheric conditions, the existence oflarge preexisting leaks is believed to be negligible." This statement is rooted in the basic understanding of the. manner in which subatmospheric containments are operated and is consistent'with the severe accident analyses performed for Surry (NUREG/CR-4551, Vol. 3, Rev.1, Part 1, Section 2.1.2, Page 2.2) by Sandia Natienal Laboratories This reference states that since containment pressure is normally maintained at approximately 5 psia below atmospheric pressure, ........... "makes the probability of pre-existing leaks negligible." The vacuum pumps that normally maintain containment pressure below. atmospheric are of very limited capacity, and can not maintain such conditions for any significantly sized leak. Deviations from the allowable range of differential pressure between the containment and ambient would be alarmed in the control room. Because of Technical Specification requirements, the size of any hole that would "go unnoticed is so small that it could be ignored." 42 of 69
w- , , _,
<- - - ~~~
- - - +
y - -
),
x , ?x. >
.l a
(v ;l w"
- The containment isolation failure contribution to large, early releases shown in Table 4.8 , . d
. '3 merely reflects'that the fact that laag isolation failures were prevalent in 10% of the i 3, large, early failure frequency. ^ Such failures alone do 'not result in large seleases Small; 1 breaches in the containment boundary will not'necessarily preclude _ the ' possibility of: i
- 4 4 large, early failures, which in fact are predicted to occur. The table shown on page 4.8-1 d ~ indicates the relative importance of the various phenomena which cause alaggg, 'early j ' containment breach.
-(e) The containment isolation failure size for BV-1 was' assumed to be less than 3".:in7 ;;
diameter. What was the lower limit of the opening ' size below which containment was j considered isolated?; 1
Response
l
-i
. q
. n Pre-existing containment isolation failures greater than 3 inches in diameter were ruled -
out in the Level 2 analysis based on the reasoning' stated in the response to Back End, j! Questions 5.(b) and 5.(d), above. Additionally, containment penetrations greater than 21 inches in diameter that connect to the containment environment and are normally isolated - ;i during plant operation were screened out in the Level 1 analysis, as described in Section - 3.2.1.16, on page 3.2-103 of the IPE submittal. The Level I containment isolation l] analysis did not have a lower limit of the opening size in'which the containment was - A considered isolated. It reviewed all containment penetrations, from 42 inches down to 1/8 i l inch diameter in size, as shown on Table 3.2.1.16-1,and did not screen out any~ penetration based on size alone. As can be seen on Table 3.2.1.16-1 and summarized on
~
Table 3.2.1.16-2, containment penetrations of 2",1" and 3/8" in diameter were used in~- . the' containment isolation top event model, however, 3/8 inch lines and smaller were screened out in the Bypass LOCA analysis (see Table 3.1.3-8) as being negligible. l
- 6. You state (on Page 4.6-9)' that the MAAP results indicate that the hot leg will fail first "are- ,
somewhat questionable". Please describe those MAAP results and why they do not affect your - ~ conclusions regarding the hot leg failing before the Steam Generator tubes. .l 1A
Response
The statement on page 4.6-9 states that 'the MAAP resuhs are somewhat questionable" and . not "that the MAAP results indicate that the hot leg willfailfirst are somewhat questionable" .l as stated in the question. MAAP analyses were performed for Beaver Valley at a time when : j there were significant changes in the results when the same-sequence was modeled with J different versions of the program. This was especially true for hydrogen generation , RCS gas. temperatures, and natural circulation flows. In addition, confirmatory runs indicated that there i were even computer to computer differences for the same'run and same ' version' of the- j program. Thus, the analyses ofinduced steam generator tube and hot leg / surge line failures ] which was reported in Section 4.6.2 was tainted by questionable MAAP results. A probabilistic j analysis based on the available MAAP results indicated that the mean value of the conditional 'l (conditioned on RCS pressure at the system setpoint during core degradation) probability of i 43 of 69 l i
, = - ,
3
.l .
5 +r s 1
.j i
- s
' induced steam generator tube rupture was less than 0.001 and the conditional probability of a; l hot leg / surge line failure occurring before vessel breach was 0.9.- l As noted on page 4.6-12, the values for the conditional split fractions actually.used in the .;
Beaver Valley IPE Study for induced steam generator tube failure (0.018) and hot leg /su~rge; ~ line failure (0.72) Lwere based on the evaluations perfonned for Surry by Sandia National Laboratories (NUREG/CR-4551, Vol. 3, Rey,1, Part 2,LAPET Question Nos.19 and 20).' Relative to the predictions based on MAAP, the Sandia values are somewhat conservative. !
' Thus, our concerns with the MAAP results had no impact on our conclusions regarding the:
induced failure of the hot leg prior to failure of the steam generator tubes.
- 7. Please provide the following: i i
(a) . Frequencies of the most significant release categories -l I Response- r See included Table 7-1. ! i
- l.
t 1
)
i 1 l J 44 of 69 , i 1
-l
a I
.i
-)
l Table 7-1 MODEL Name: BV1LVL2 i Release Category End State Totals : Release Frequency Release Cat.
. Category Group I BV01 2.2225E-06 l-BV02 3.8136E-06 i BV03 2.1281E-06 i BV04 9.7823E-07 i BV05 1.8143E-05 11 ,
BV06 0.0000E+00 11 BV07 2.0676E-05 11 i BV08 0.0000E+00 ll BV09 0.0000E+00 111 , BV10 5.0985E-08 111 BV11 0.0000E+00 111 - i BV12 5.1041E-07 111 BV13 7.7928E-05 Ill . BV14- 0.0000E+00 til BV15 4.9134E-06 Ill ! BV16 0.0000E+00 lli ' BV17 8.1426E-06 Ill BV18 3.2414E-07 I BV19 1.0897E-06 l BV20 8.3235E-06 11 BV21 6.1746E-05 IV , 4 1 45 of 69 l
c.- (b) MAAP output curves that are readable. Page 4.6-50 Besponse See included Figure 4.6-7 t 46 of 69
. !i . E p
E-o lE O e R. 3- u. R E d ~ f a- e lPJ-@gl*g.se 1,."1 Be-a o 2R o . 2 I I e d
- one n9 @4 m
vs 4 ,4 e ooz GJ S3eni 10H WO1W3N30 MY31S dOO1 N3M w
.@-, * (
q D
.L ,
.g i I I I l i 1 1
, I I cora ooe oor s cort ooo s oos oos oos col
' UJ WON 31d W3ddn 13SS3A WO10Y3W -3WA1YW3dW31 SYD UJ S39nt 10H WOLW3N30 MV31S dOO1 N I' h
$, D h- 3
'- 1 I I I l3 .Io , I I oces cove cove ooot . cort ,
#8 cos UJ 3BalVW3dW313OON SUOO183110H U.) WON 31d 131Nt WO1YW3N30 WY31S
?2 o
l l l l I I l ooi t oort oos oos one oos t oor i oort God) 3WOS93Hd WO1YW3N3D NV319 G.) D3110H dOO1N3 MOW 8-Suni i y'*
.A ~
~
ocoS "L g l I l l I oove oovi o oors oot i oors a 6sd)BWOSS3Wd W31SAS AWYWIWd U.) 03110H dOO1 N3MOWG -SWf i
. - - . l
69 Jo Lit'
/ O dLEOD2096 OS-9 *lr GAS TEMPERATURE- UNBROKEN LOOP HOT LEO PF) GAS TEMPERA 1 800 soo 1.ono 1.200 s.m t. coo soo eco i
.).
c I I I I I I I l- .l. 2 2 2
= -R 1 4 1 1 I. , I. -
i g,
< e ,
.O
!i,
- SURFACE TEMPERATURE- UNBROKEN HOT LEG PIPE OF) SURFACE TEMPliRd 1p. -
eso 3 soo i 1.cos
.I 1.200 1.m 1.800 soo soo t cos I I I ll 1 -I J. W m u
< v o.
-I m In1 O
.1
) .C @
GAS TEMPERATURE - UNEROMEN LOOP HOT STEAM GENERATOR TUBES OF) GAS TEMPERATURE- BROKE
. eco m eco '
eco no i I I .I I 4
.l N =*
lN 4 4 h 7 q' h i. s , 5 , SURFACE TEMPERATURE-UNBROKEN HOT LEG TUBES OF) GAS TEMPERATURE-BRO 10 eso no soo soo
- -l l I 1-l
.- g 'q _ j_ 4 i
~I_ 1
.s g "
.g.
[ i V .1 - l' , - MASS OF HYDROGEN IN PRIMARY SYSTEM (ann) SURFACE TEMPERATURE-- GRI { .o.' o i l l 'l: i . pjeo aanpedy
- k, 4, uo olq'aiteAV 081V - . >k , d_
t y- .y o -
% w,- -
^b gz i 38 hl% ,
1*A ' ' L L .OBISNV _. e
. .x.
}W b
(c) Drawing.of the vessel 'and reactor cavity showing the'shielditanks,- the lead. shield: p discussed in theIPE submittal.- Resnoose'
; See included UFSAR Figures 5.1-1 through 5.1-7.
F 48 of 69
P h Gh - k f . o
~
r......,...
+.7 ,-._.
7
..q_ _ _ _ ,.
- 7. -
x i , . 3 I j 5 (* f' OU
.. $. IT q**" _
. . . . - s.s l
" "$'** u .
3 cu, i ge uu 7
,/
ND. L'IFF59 [ b
~ .
l l R
\.
t
.L.
l 0.ut'W!?S ,Ca -
- s y s====****~e m -
s gip # %,gggp g %* , pr EIF GL Tbf **6"( &L*190'*4s* g
/ , a IIll*e*IsMm; .
Y E*
. , , , , , . . 'w- < <. 4
- A ' '
,8 . .
v 'y#; [ - a s ,=, =_,: /f: ,/_ ,
,e
.;+.-r e- - -
M c e. .w .
, ,u. i ', -ei
_= .
.o a,o,,, . m e t -l ~
~.
3a N -ff-- = w p: eu..:.%( - m '
- s 2% ;- -
eq!r,.- i n. !
. ~ '*c
~
( A _g_ i
*} . .
n..,**_*. .
.'.5 .k_
~
! ' ?'*/.
I.. $. i-
.a ,,,-6 ...uv . .
t, m-
..L
-. 1 s;.y s >
n.
< . n. , ,,
eF* '# #..- ~, l ,, % 6 mec es. .s.o.a6orco.t RECS C.ep.Av Pusip
.. 4 u #..-
e 5, 4* g. [p
*t.T - ', 'O <-
- r, - A., %
s v'- v ' /
.as . s ,' .
, ."g ff
.. . v - .
e .. Yj *
%- l l :.
- w. w t um
.g , , '. - .
. s , ,
, ,s , , - .. ~
$ s y C M
, sovjrauwt et.
5,. re tu . y
%'s'?l[ irs **
4-P1 4 1 I I
a N I .
. . ~ v= ?4 .
JA N O l' E*' # REV.I2(1-94
.e a '
APERTURE ,g m,r , e%i. = .. ._. i y .'
.,.. .CARD 3,j p l p
.~
w 6. . o n
..ty.._.____._________
_ _ . , ,,,,,, p- =_.-__e_ my,__ _ y m 3 ; nnem
- v. M (c n ,
i e ,t , ;,,7-y ,yky. q . . , .e. n,? mc ui-u,, lp(
~ --
l"9,;t@y - , g.g.,.,uu> q q ' ---- . ,
,.4 e se:g,e.. # :,.m ..u. . . . . _ . __ .
rf 8, / ,.....- ,,,[~l 1 U
.g \ ..-..m.<. -
* - a? ' w. ,, '"
wy.; m. Q ,. vs..A.2r
\ .
a ::'.~. '::' := ni W P : eh w ' , js... L _ , >. b, h m s s x <
,x
,3x v
I r
.c + _.,
ei . @4.x, %. - m
/ ;
ebum$L-~. g l
. s/%
- ,.;;, y s
"L;~*','?., *:.. -,e p u 4 J2n., .- .~ . . . .
a m .. t~. M 4H_'Q fy t h I( - A N*~ ; N,,,3c.s. r [f .v1.ve' 6 p&:_ g...=p
, 1 a
e
% n
. 7 m
_h,-a L b-" > ,t a. s.-g y.ma.,, w,u-. r 7. m.,
, m,.4._ . m-_. ,
_e_
. m._.s. ..
w..,2. o, _ . , at i A C
- my n p'""* *l.turm L a m,p m- t b,,,qg$_eu gv D E T A C be E D PLAN C-it,w.,.
...._ .s..
#g z, 4e r g g q . -.
, u <~~>
CD . r w -_ - -. g/ .,un, . ,7,.,n - s .. s . _ w,,,j,_"'.
" " " . . , . ._ _ . su, _ . _.... . , . ,,_. . _ . _ ..._ ,_ .
h, - g- / ss j ',- 4.
>\ g ,. .
R y, .*' . w r.w4
.s.Lia v.i- 64.avio .am.sp.ca.i.L.as o i. nut.ra.se,n wi , e
; ,. * , e
- j. e4 . ma,eameLE et.e6 Fe. e4.aa. .e eta e. .tpesskeeste mentdal , '-7e .T.
_ . . __ __i . ,# j , m ,,
'* = *=_
. . tiacass
. . , _a'.u=.
. . .aa . m.aann
, . ~ _ eainan. _ _ . *=>'.. tea
. 6 se . r._
- ~ ~._ ._. ma** 4AL asummoLa .Laas ,o os .aLo* ,. r _
.m_ m . . .=,
. _ . _ . _ _ .._ .,~ . ~ ._. .
j .. -
% . .u .e u,w . . ..
.. 1
) ,e, a-.. .s.ct. .*pe.w e.--ia !
g,g . ... -- - ; g)e **k a e [
' .si.1 %E.1 CCC D-t 9448 ;
i l g am., v.._ m
- . o.. _
i QN LL."WIWl h,
/- .. _ .
FIGURE 5.1-1 i ot.tatesED
- 6. vae's
- 4.eovaPLAN (qq,qgg +
REACTOR CONTAINMENT GENERAL ARRANWBENT EL. 767'- 10" BEAVER VAd.EY POWER STATION t.NT NO.1 ! UPDATED FINAL SAFETY ANALYSIS REPORT I
$ I I J
- w ;
hO of fA
f,- -. 1 I -l pe- . - . l
' ) .- , ;
o J,*;-- i ,, ',;; t'r '~' QAn.
/l
. - - . . . s- - - .
1s4 */sowc st.twg as u
. .m e o~ i D q-- -.
)p ,
4
$'h i 4
b a Y '
'4 #
II ,,,e COert.MCamC.W - g; rune asass
- p( * %
-G- ,
a mam_ ~' i o . g , as a ruew '
~.
6, nk et. **
%I
.A-
"' T
, A, 3!
'k l#" h 1 b ,+ g w
E
~
\ #
.c j I y( u. ,,,.
WAWE P8 q 57 i jy / :
/ ac-e.sc q
j j W usas wa
' y[ btYrI I N,'#4 #~
.; aC+K CDest,sacsag. iw owaar as.
l ~7,' j ji 4 ' I N puses"_" ) Ah - inn m -- .
*t*
/
DETACMED PL AN E5' .t er*y N* Q ' EL.TSR'4*
/ 3 D
Y ENEtihib M f v& > es l Tuts seemees er usua team Arirm gauerseest es assrve 1 4 I I I
l a l l REV. 5 (1-SM
)
c.. s.- D e )n
-.' 2, a g, j wr ea q-'- ~ - - - < ;Q ^n./- ANSTEC
$f "I '
p R.1 #.,,,, APERTURE
- p. ~l7 7
~~ $il -@$ 4 3 -
kiY 7 _j ~
=
.-jh 'l +
~i
'1 8 t
E L %.
$2W, .=f c3 L___.
l CARD Also Available on 4 ~ Aperture Card 3; n C.m,,,.Is E s e
' # _h, r r 3 ,
m , s j 5 L.
- 1. :,.
EL ?se'e* . =4 ****Es(usa 4 a.)
~ Qg,., -.n, LgT s'
~, - G--
s s.
. . mm.
=
4=,ggo ,,ln ; g y , a.ur e - 's -
.~ . ,
-g
. 'e .
"W#
cue- ar . m
/ e E"'== #*'
\ d 's I-
... _ M
\ ,t , , ,- f , g"' G t ,. - m. a " l 1 I4 P Ls M -
.j - . q ~= c'* w- = ___
= _ - ,
m . r_ q_ w_ _. } i ,. ,,3 Ba g,,g n' s . [EE.5" ' 5 I . ~Y
%eim.61mn.swa n
.. ., ,,,~ e.=.,
~ .
..=
... g
; su -
1 . p-w - f ?_ r ..=.. -4yA - -
~
y ~ ,.. - . . 7 k- _ , , , , e3, s
. =.s .. _ ...._.
l_.. $._ 8. m,(A= = - - 9503/30 37sc o FIGURE 5.1-2 (RM-18. REY. 8) R.AN EL 738'-10' REACTOR CONTAINMENT GENERAL A'RRANGEMENT E'L. 7'38*-10" BEAVER VALLEY POWER STAT 10NINT NO.1
~
UPDATED FNAL SAFETY ANALYSG REPORT t I i _ 50 of 69
F l I c-a l t r
- y. .. ..
, c
/ I4
.,,, , se et... .t . vtr4 - k, rw m n.tres z
~
y .a a:
'1*C -' ,I'. ) -
g::.=flt
/ , ; a-- p I#' ,"Ti". * '
_ . ~, ,,.y ama n=rr ,
.,- 3 A, * """ a r- y 2 um m.e a.op(-~
g es
'i., .
g 2. J
. ! =% L'1"' $,. a%.== n =. == .1
/ , """
=N a
-a..a== Q ',, % .
. .s..y <,7. x !
73 ,, ij_ . %. ' e.n e. g
- t. , M[
.c ,9 ,_h'*O(
. *,1...n, 3.. . . .
s, 2 "; , .v < f
- - , (
,,7, "5'* h"/ m .
i
* . . . Si
. m Tv
/ -p I~
J
] 36 9
,/ .
. /, ue_2. > .; ,,
c Le
' g- _.4,,,
on.e,,. , n=-nu- e .9 , _ r . i uto y ,, s ,o ' ..,"_" . .- r' - ( ' 4, g
- s. _ _~
f , ._ .. -
, ma abijh/ y 3 ~ c..g . y Ib I, p d (, l . II$ p!
is..- . -n .. r.. \
+i_ _. ..'
... ., ',,w; h 7.s..d C.n or,w ULTM" s,,
, g a,;g
~1,pra n r' .
h'a u, - g , gi ' , p , "\$M " nw.,,
$P , 1 N !
J. =na g:" . . gyg,ga.e
. LPl!
*tT2:%%7 ':d
, v;.
'ft
- k*
- o-w to n.1.r.</
nwr ,
t'. ..
y emmenERG - w.g y >
.m
;",""""' nws -
s 3. ., .
.e 1
_I 3,_o R.AN EL. 7026* i l i m Osm*. 4mm we
-7 REV. 0 (1/82)
! lil y.n w.r '
I
- s._
.p' ;
- i
~ '
I'**
- 7 sas s a ELas-s-IL ANSTEC gr
" ~ ~ - :
i APERTURE g LP CARD L -'
, - t::m'**
Also Available on
= ../b .t.n r. .- .,,,,,,
Aperture Card s gy .i f""':::"l0. e- ~ s+ ~n , - ll' s .
, y
,. Asj '
y\h .--
- . , n.,.
s i,4 \ h4f :
=IM[s *"'$Jy},
" v.,_ . .' J ' -
- f. J .
J~e f_"_. ,. _w_ . X .' g ",.'_.'. _ _ _ ,u: .v l '1g m, g, o1 i - !! A U i
- u. ', ? -.
-- a j'~
iikI*'r
' 1; r
, ,f n x .4
'.,.j .
L, _ , U L-r
- ' * - * - - 9503/30 375 O FIGURE 5 1-3 (RM-10, REV. 7)
REACTOR CONTAINMENT GENERAL ARRANGEMENT EL. 718'-6" BEAVER VALLEY POWER STATION UNIT NO. 1 UPDATED FINAL SAFETY ANALYSIS REPORT 51 of 60 -- -
4;
. ~. W I I l-t 1
,b 1
e
'I s
int 1
.. h g 5 F
, me.a . no
, --~~ t _I .. f
; _g_ -
(m vN .,~jc if _u , + ewv.. s ma
.,%- i .. , .f a y x; !/
ww 4 a -. t j[ E aramuneveps gEMont mit N '- F geo sp.es / M N
.r bW .s.y _ ,,
.
- tar.6ly.
us meJes.t.eene
.=.
he tas , M 4A / __ .._ a
,(\, . . . ...
,/..
4
- s. .--
g g\ ,
~ -
,,/
.6
.e
__.t W W) DETACHED PLAN EL.107ts' I I l
~-
i l l FEV.2 (1-84 ANSTEC y' APERTUR$ CARD ) . 4 Also Available on p Aperture Card
, . - j __ j
'bp ( - - m ..,eu n_ / s}nYS?A~ '
. A .,,~, N ,N . ,
p t
* ,~ ; , ,
,g . s..s /,2qh ** ,, ,! , . . . ,,
- m. ,
i
/
( .
'+ ' .'_3:, :=.~
l ' t M4i ' ;,..!. e. ,. IY'
. a. .
, Y,1; lYCq. 'S{,
9 ,,g !' e-_ !j , ,
/
\
,,,, J y ,s
- /,., . y
= ~ " -
g, ,,,,r x - yg \' ,1,,; ),_.e ,1 ', e-m?p ,r . s;r'%.~. A. . - < r,m. ,,( .#
+
. e i -.
+~w ug;gc. . - =.=/r -.\ .
b _. .
. .<d Z. ba riu.cj""6
,a9_. d%. m. r
. u ;o.,.
ut [ C,X m.a, -~.
~ * "
r
' ' 'i.
. T n< . - 9We i '
m.
,_,@ l't e g ;& E "i g ,j ys. ,
o 1-. ', / , . ** ~
. y'(c, rM./ q c_p-
/> sac. 'n. ~ .,
e,,:- - p *9~ ~w.. g737 l - ' *t. I , , ,.
, \ %*p, u
~ ~ ~ff ., - N -*
,,,',,,,','a,;,
llf , s,
.f f*.,llll*g
+* _
.* r g .) '?.
- +* M
\
,s f s.
em i.a7 .-
- . .=-
c_ . y ), g. i i. 4 7 f ' .y *
\s \'g s =.r.~.-ps .\
wpg= .,
~ w'= am i+ ' v w y 2 f / , ..,,--
g ' \ .' ' #+ N 'e s cm.].n. ,g} N
/ c7,,'1"Q"*a,,
/
[ _
-s ,, & /
pc.x
' ~
.s x/{'<a"-l"'"jn.o
-.. . -,g. . _ . gj, .;X, 1
~. , . ~ 4_
' ~
N. y' 4 A
~
_] w
*-of .I N5c q PLAN EL 692'-ll' m , g,,,
9503/30 37.F 03 fMUFE 5.1-4 Oue-to, fry. s)
#EACTOR CONTANWENT GDERAL APRANGBENT EL 692* - 11" EEAVER VALIEY POWER STATON (NT NO.1 LPDATED FNAL SAFETY ANALYS0 FEPORT C I l 52 of 69 __
' .dmuume anyAg> ,
... ~ .. ,. % %,
'J
, : , -- %g j/
,/ ~
r
./
,/
o r
/
/
4 u /[**,_s.,[...-
!f / 'l. g_ jg 4. lu, -n
/
- //
I -
*[- ,
d
- Yg . gI"
/ : b= P
.k ..
~ ~
,t ii.eWa{ .
se% vsna (LWif at E et e CL79T 8 _. ... 30,:.... esactua emnesem
. c.a.1 E" l
%q % .
, .8-l w! ,
er wieu evan cowissa, ano . 6 I . - [9tivtSatso
, {- -s- , :. , . - - ! n ';L,.<,,
~. . .sJ. ~ .& -
m .
~-
lw** . N -- ---
+
\ .
i G c.w mano -l *s
] a m3 gy ggggg .
d *
- ennawinM ae --
ca 5 - .
;=:
l l awan == -i m
.7 g
L m m, '
%e.r
,, atacion savny
.g.s .
-l - .
y p=< **st ***6 t
.. I ,
i ,3g h [w"1 EL.T RS-h* - 2-lorn + g
. . . . . . . . . tt== , .4
. t g ,. a , * .
.gpffges SRACTIM 1_
g' RL 791(o" 1 '- .
$UDPORT MO .J
=..3 ,
l
,j -
} .i s
'l RL-a* % f,d tat +se -.-
j
, , , , . \ N, id j .,
,3 asum,- ?'at* \ d l u& -
,] ., M.M*' 1
.?': i I
i - . G_ett . t.T._, ._ .__. , 1 A
~
_M. .u .i y,'"" . et 6p(Aq 3 ._' j
,,yy ;
p., gg g,Se ,_ N,.,,/ ..y,,,.
="
-- .t.,,, .
e . ,u . 8**N"h . ,. y
) .c$. e.e \ y uuaa " h a..~ .
g 1 -I
, i=+i4.i c un f
I I l i I l i i
r r d l l REV.12(1-94) t . . ,. . - ~ ANSTEC
.~. ..... >= ,re,r APERTURE
--s v u.......
. . 4. 2 , =
j# .... CARD x\ ' ,n.....,-
- L..., 7 0' . 0" N
1.- Also AVallab!e on
/ .
.. Aperture Card f],---@p .. .
1 u ,w w 6 =P O. f'+ -e a f o .
-u
^"'""
. .--. x <
.# \. 24 .,o c
74 .,.-.t. - paer- a m - av't e' c
\, ,~ -
,/
\s ~ .~ n
\,
a.,a., .s. .y. .y_t.- ./ -
-Ih.
rgy,u2 : E c " **' '.
+ y .
l - a*r_.am si. ..c,r l u n<. ,
.i '
j
- j.g%%
- \R n.0&st g p, gg
- '0.
- b. - '* ed-# gg
<w ic.
u-46**T g et t92 o-W, ?T -- a
. - / - - .. - ....
a .(" m ce,a ee g*= , ps 6.,ro.
,,x - 4 ,.L
.t var s'
,s, . - .t ut
, . a.v.eATio.
&(> m e
- v tuate sam.t
- - .eneror - .su .e . o. .
- m.m,. s . . .
i , . . _ . .. r aaa' ==
- 3. . {n.~ +
.m....
6U*(pie 4G g j,,,, S v.(
.o e
. b =ga _
4. _,_,. m. ' .- % l b..
' - "M b,, ' '_7 i
, ,, to st% ,, q
,--- t -4 g , , ,. ;, t.>. . it.rve
,.caam
"***y, y,
(i.r.._ . ***sLLL2,53f.j
~~
i~~ , N , , _ s.u.i.t t.a.ae
. .e.a 1.
1
.I ' .
Lgg W.-. s.e i
,,,,f e u.f te d.
N L. a s , ,,, ,.
, ,- -nppxd 4" V,j7., '
~
i wesciic ne.w _ a. c, I ' ' pt r] *.' M* === .spe, ,ul,ln
"- - e m ,.
k.D W_1
-
- su'.I,
,u u ,e -
o* "' W
.uaru U i
<f w_p=_
p y - s
.a . /- n+= **
h .3:. 1 c"' ~ 1 wwg '=. t
* ~ y"~=""". _
N't* 55 J"" 7
-j-1 ' r d * **
.svt.necxn -
qiutu- Q 4. ~~.* s, e- n
.a..., ......-. ..
-.c... i .
a r, uoin s t e.' *a ama i.m ev..e a. p.. ., ..s.c 4 py
.w g- =- wn. .
93o3 3c 37 FIGURE 5.1-5 O(o (RM-1E) REACTOR CONTAINMENT SECTIONS 1-1, 6-6, 22-22 AND 24-24 BEAVER VALLEY POWER STATION (NT NO.1 UPDATED FNAL SAFETY ANALYSTS REPORT
$ l I
~ ~'
53 of @
- f. .
N j .
.. .., ~
* .4 g g
7 _ w ' . ., h _ / -[
/
/
.v.
N c
/
4 i
- tootou Cwsus
- CR.s .
. :q g .
.: r" r=
- .n eeu
. m, -
m ' i
,.I
===
- c. -,
u.
"'.,,,1 r
s
+ .- r%1, ~ . }[ _ .l l
..- _- 1 i i .- 9 .
w .. a y . g ~c - ra -a ani
- -- m . .;, ,u .
-l
-.9 JF' ' l'/ P-.Miak c - .. ',
_ . L.l.yh 7 7"y ..,.[. ..; ,
,(
A .p _, L %,k y%. : iW ' * " ' -
.. ,: .;. s-s. .
-w ,
.. ,
- C,. f.;.&. ----- p, ~m~ ::.,.c m
.. l 3
=. ..
c
. ... ..o w . . .
..,.c>
.l .. .. . '..
r- ,
.- ., u ,.,.,a1
'l s a. . . ,
g r .
%as
.,t., ,
. g,a.* .messe i
- i <
,", *** **1 aC
.J
%.. . 4,* ls, ad
=a- .,..
4e s. ,
.-,. .y v.7. _ s .. q u, m. ,. r, <
t . .. , -
. r* , . . >
,s e
nc -
~~=v -a;y
(*:aam mt m
- .., .:,: e m.j.
3
; 7 .
u - jj f t..-.m -*++"
'j ' *a = = ~r ,. '4.
i .
- t .;~' o ui m ..,. . ij , ..-
7,
,w
. s ...
r- .r .s._' ! = =3,3
?
$. 11
- - M .
. 'b (. .)
~,1 e ii s :r-
.j m - ,.<. ,
^
;::,7A, -
w ' . /- i
- - + >wb meau'.o= ami if
{
.)
3, "[ g d ,it $h# = m
#i2 ; ,
7.~.. W. h I#. 4' Eh ,
- f. ~.:
l N tee.es'.
?2
(* * ** t aC & es
ju , REV. 1 (1/83) L t
- ^
I-- ,M - N h..m ww .
, a, 31W1") g %.
., .s.
- t=: .- .
0
\
g ,
*n
;,'u ,, -,..esa32
- r. ,,2 ,.w.. ,,. y ..., 7. n.a.4,.
"'T*'
-Q L - c~ , ..au,.
esln;, mYe
.,,9 Y e.n..m _ .,.,= :: ,:
._. v 4-t~;.,:: : .,,, p. . n j' A b o a.,< aiia'o !a o;t-
- u.--g,4. A e :s6 m*:*::; :~.r::
l'a -a-a NN g,, ::= = ,g,, f - Accr: ara C:cd
-I\-
y N s-,- % lN-.ll .
.\ a,
15;i'UJ2
- d. 14-'.a=
} -
- g. - - - ~
R;
., i. .p 'r WN.Y--r , . . . , -
, , \..i
~
.. % ) . .
g i i. 4.in. - g l-l i
. 11.. ~t '.*r.f,.
a'.. ,- i
..m...
. ,1hu:m.
.L w.
. n .
- .= s n .. '
* . , y
, unc uni n w,
%J . -
~~..a s. .'.
;,s yd
.l.
g&
. . 9 Lc-n .n . ,..-.;. r A,
.(n.9m) 5'-
n, - re .a e n.= g ,
. . . } .
g g..-- ,.
- N, /,
,.. e
+ t hi - r 3 / :'r::t!.= 3.,;.
- - 3 .-
.e!, .
, : p4. 3,. / .. ; ;.
. ! ;. ,y
-o.;..-
#g.-MT
- ri W .i, U .f .... _,;;. n.2, y .
, .c2 jp- j c.
c U. s.g r. s nq j vt v .-
--r l- <
w i:, ~ - m, ., v
? ;
,.. c., .e.: .: .~ 3,
}
. c....;. r ."i.b . -
.m . . <
-.wm ' . --" . E '
i
- a. l.. . . l- ;n?@@
- l . .%- .
.% a ,.'c::t. .
,: =
1 R, .- L'
- dr.rt.1 lt
'lI m-ms._
. ~ , .
. +
,4.s.~
,. -- w
- l. ,
p.~. n. a
*[
p
.... i .. . n
,,,7 J,,,.,c. ]%. ,, , .
, , t.4, M.
.y . .p'=\..
c
,:, S.m'c '
?
i - o, , t m.
,a m . .....
....'.., %:.-.-c. . ,
- u. :... . :.p.
4. 4: ma. .: m..,.c..,....
. . ,,. ; . ~.. 3 y.
n .- -
. .c.; .
c + . .. : w . .,; t. .
.. ly uv. -
.- i I
,a 1
, i * ';n,J i t
< \. - -,.
g,
,g
.d__.,, .... .-'
.,od . ...... .
x_ _. .. , . . . ~ _g,- ; .. ~.
-1 -.. r.:. w ,- .
FIGURE 5 1-fd (RM-1F, REV. 7) NEACTOR CONTAINMENT ; SECTIONS 2-2, 5-5 AND 23-23
- i
; BE AVER VALLEY POWER ST ATION UNIT NO. 1 ,
UPDATED FINAL SAFETY ANALYSIS REPORT f J l
% of fo '
l
REV. 0 (1/82) ;
.- - r*-r: . -- ~<' .
p.yb..,-c.. L%v Q.. pyBRTONC (j cA11D N
\', g;sema@ef y, g m ure
\'
r I.
~ ;. ,=~
~
m'r, i??: w.~.hl pf, r
. . . . .. i 1
su e,w e1-
-s-.
.. f..
s.
/N , Ay b, .rd -
p, ,- _ . . . ~ . _ p.. .-... .. l ,
- q. ,
1 u - i r. . ;, ,. c ~ _ . , J,-
~,_ ,_
- l. . . . ~.. __,_
, md , , , r,--"--- ! _s,
/ i i ____< m ) r s .
v ... . . . . . , , , . p,- : t. ,,,,,,,y4 *gy 21-- gg 1y [ ,,.,..
- .2 ,e L -
3 i ;, ,ss a lg.c .. ZL.m:: t . ..
., - ,.:. .:. w w --
%-._., _.a ,.m;e.s_,
e
,c. f s}. --,. J.L n ,
n p ? G. --, J.. "~"l .,. . y,a.- y...g
- . m.,v m .., --
.,_ m/; ,e
_. , , , ... :I / .-~ ,, ,,,5 c ,/
- .i
-'h g ".,
# ' N.- ,%/ .'! j,; }
e' ~
, ") .
M m:... 2 ~~=W -+
. v ma ,,
,- - t.
} u. u _
- . . um.m ,L'. - 1
- w L
.. a ,-- P5 -
m syp k l , . '
=> <=
=uw , . .,
p o ; th.= . ' . -sm, ,
- w.- .y: ,. .
. n H,-- . t
.I ._ , s x
.. l
~
- -o 1 \ ,- ,
, w,>f. -.t <: --t. . c, = =-
- '-~
u 1
, T::=
.v
. , _ ,._4.,,-. - ...Q7 !
. 4.. / - ..
l I
'C #*!- A- '
h Ll~.j.... : ' :d - .&.. ., =I. i ,
@b-sg.e--~s..-brH m' % --,- d i c. ' ! I._j.
4
, u 1 3..
_,. .L:.m ,,._
.d m
.1
, ,s
- u. - .
- i
.- i
.u .a.. .
4-4
,..s.... .. . -
~'~."~
i ..
' FIGURE 5 1-7 (RM-16, REV. 7)
REACTOR CONTAINMENT
- SECTIONS 3-3 AND.4-4 BE AVER VALL' E Y POW ER STATION UNIT NO.1- '
UPDATED FINAL SAFETY ANALYSIS REPORT 9503 30 <; 63 -
'55 cf /9 i
4i l 1 t t
., N.
. i ;
- f. ,'~.
2 , P N g %.. s .
'N s N ,
Y d# \
.\
e e*.+.. .- i
/
c'lI."= T1 - too 5e= + c e*=t
/
l a-- I IT .'~s g
+
/ .~ .
#/ *** ,'3 i ,S p easse, J;,,;
.e. 4 I *. ,
1 e l i -e
,.. +
- s .
. : L t 4
- I
*"' i. a=*
e
! l l t I
t i. 3t t' .:.. C
- d
- y; ,, ,
. ,* i i * . .
I t, f- coate.cm. ano Ineswt sankca % w I l , i M** ma'l .*: es m,.ie
'E i
+
.,] gg n a '
. d
, .w
$) l'
~~
h ri e u'O'
- f. ( ~ ~ '* * ~ V 'o G
i v .
. .p p'f
.Tl '/
i ' f caste smas F il
*,t ** .::n. g.auc #,. " . . **ta
( ss.1 sew s i 1 . onwt" sam (e . j1' I us t.D a * . , {p ; b,y g Y'
...e w- eset 4;
jg, .wrar
..--m. veiswervarmas
.l
, me -
M ;_._d u,u " ".~~ E \ pag a w m_ ..% ;
,[ ..
Ah ._, a..s.u.n.cu.une.m. nu m , i , jh i _ e.1ss'.ar*7 . d: 4, -e E D. 36 r- {,-- so m?m*
,,,,,.,)
;7 1.... g], u,7 .u rests-.wi th c.qiyeattasservaavfg....T,.,fi acman s .ato r
* > == w.n n y'W
- q. a , , ,
1,
..: gJ,.e m tf _.. - -.L .
- - a cge' g fi- 7 i
r= * ==
,r8 o.a
- a. acu, $., e C' *.t.
l .
, *b .*d !;. * ' *** "* 'a - - -
, )' , 3. - .g-- :
L . . ,< ,.
"' i ,
Y EE? 1- . .i
; n . J' l'~_ ;^t i
! h '
ma -! t d
=
ah N J i l y , batataf otAT _ , Il 8**IP' ' A "N 2 1 Q... +,..ce i i m t l l .<, ..
] '
s . o t '. .. .~o p
.oi ' - . . '
l
**$,$ j l t .>
wgten
- i -a=== j j
,;,-'
- am .w * ' ""
g . .- -
- l g[**** are g
I
- i. . 4 e-,tF i( ,*-
y .
,,, j ,s- . w.,,.
3
. , , , . fr + .L 1. * *
-.- gasans ./ esce*6 e. sus g. g J ,*
~.
LW *** su g.a..r l t
* ** + S j
.* g memu o L=as. ,,
j
. = e L serf.ti. !
33 p..in is it aiso : h
?
i i 1
- g. - -
- q t
V i
' (d) Descriptions of the containment failure sizes and locations.
Resnonse: I
- As noted in Section 4.4-1, the Beaver Valley containments were believed to be very -
similar to ' the Surry containment which was analyzed in NUREG-1150.- Upon confirmation of. this similarity, it. was determined that the. probability distributions ' (containment failure and conditional probability. of large vs.f small failure) for ressure; capacity developed for Surry in Reference 7.d could be used for Beaver Valle c : As a .. ; result, no plant specific evaluation of containment pressure capacity was performed.for.. Beaver Valley. In Reference 7.d, it was noted that a large hole or rupture is one for. which the containment would depressurize in less than approximately 2 hours.- It was also noted that large, dry containments would depressurize in 2 hours for hole sizes on- i the order of 0.3 to 0.5 fl2 It was then sted that a small hole or leak should be of the ; order of 0.1 fl2. . A failure area of 1.0 fl2 was~ identified as.' definitely a large hole ori 3 rupture." Thus, in the CET 'quantification process, small and large failures are typified . by failure areas of 0.1 fl2 and greater than approximately 1.0 ft2, respectively. I The Surry pressure capacity distribution is a composite of four expert analyses and 'a: l number of failure modes including hoop failure in the cylinder and dome, shear at the cylinder-basemat junction, liner tearing, and failure at.the penetrations. ..The composite nature of the curve makes it difficult to identify specific failure locations at any given failure pressure. In Reference 7.d, it is noted that ' failure location did not turn out to be important since any failure location except shear at the basemat-cylinder junction would - result in a direct path to the outside." As :.md in Section 4.7 (see Table 4.7-7), all large containment failures were assumed to release fission products directly to the environment. Some ex-containment retention was' - credited only for small containment failures. Reference 7.d is as follows: 7.d Breeding, R.J., et al, ' Evaluation of Severe Accid' ent Risks: Quantification of Major Input Parameters, Experts' Determination of Structural Response Issues," , NUREG/CR-4551, Vol. 2, Rev.1, Part 3, March 1992 i (e) A quantification of contributors to small, early containment failure. f
Response
=!
Table 4.8-4 on pages 4.8-62 and 4.8-63 of the IPE submittal lists the non-guaranteed ' j failure split fraction importance for the major contributors to the small, early containment ! failure and bypass release category group. The Level 2 top event failures and respective l e RCG II contributions tnat lead directly to a small, early containment failure or bypass are listed below; Top Event BY (Containment bypassed prior to core damage) = 16.3% Top Event C1 (Containment failure prior to vessel breach) = 72.1% Top Event C2 (Containment failure at vessel breach) = 13.0% ! Top Event CE (Containment failure due to early H2 burn) =0.0% 56 of 69 ! 4
.\
, I t
p The reason that the total exceeds 100% is that there is in oserlap of sequences which J involve both C1 and C2 failures. This is due to the containment event tree structure that I stillinquires if a containment failure at vessel breach occurs as a result of a HPME, since these types of failures can lead to a large, early containment f.tilure even though small l
' containment isolation failures occurred prior to vessel breach. The majority of failures -
to this release category group (88.4%) come from small ccatainment isolation failures and bypasses from the Level I analysis. Early hydrogen burnt below a 12% concentration are not expected to fail the containment structure within 4 hwm of vessel breach, and bums above a 12% concentration result in global detonations that lead directly to a large. containment failure. (f) The size assumed for a "large" bypass.
Response
All large containment bypasses are governed by the Level 1 interfacing systems LOCA - events or by a Level 2 induced steam generator tube rupture. The interfacing systems LOCA (i.e., containment bypass LOCA) is addressed in Section 3.1.3.6 of the IPE submittal. As discussed in this section, the LHSI/RCS pathway includes a single 6 in. diameter header that penetrates the containment at Penetration No. 61. The piping from the RCS connection to the normally open motor operated containment isolation valve located outside the containment is designed to withstand normal RCS pressure, as is the valve itself. The 10 in. diameter piping upstream of the isolation valve is not designed for high pressure and is predicted to fail when pressurized to RCS conditions. Break flow through a mpture in the 10 in. piping would be restricted by the flow areas associated with failed check valves, which could be almost as small as the flow areas associated with the LHSI relief valves (105 gpm total choked flow), but not more than that associated with the single 6 in. line through the containment penetration. For the induced SGTR, it was assumed that, if such an event were to occur, the primary system pressure would be high enough to lift the secondary side safety valves, which have a 10 in. diameter outlet, thus creating a containment bypass route during the time of core overheating and fission product release from the fuel. It should be noted that for the induced SGTR the associated minimum bypass area could be restricted by the number of tubes failed. t (g) The contribution of global detonation to ccnditional containment failure probability (large, early; small, early; and late). l
Response
The reference to " global detonation"is somewhat confusing since no such terminology is used in the submittal. As indicated in the response to Back-End Question 1, a global hydrogen concentration of 12% was used as a " benchmark" for the occurrence of a deflagration-to-detonation transition (DDT). In efTect, it was assumed that localized conditions for DDT could be achieved if global concentrations reached 12%. 57 of 69
E p
'U D ;
~
~
, The percentage contributions of hydrogen bum /DDT to. the various release category - -[ C groups are as follows: 6 0.2% of RCG I - Large, Earl'y Containment Failures and Bypasses -
. (based on importance of conditional split fraction LEF) 0.0% of RCG II - Small, Early Containment Failures and Bypasses (DDTs are assumed to cause large containment failures)
; 11.3% of RCG III - Late Containment Failures _ ;
(based on combined importance of conditional split fractions C3 A L , and C3C) i It should also be noted that containment failures at vessel breach (CET Top Events C2 - and L2) due to HPME are due in part to hydrogen combustion. As indicated on page 4.6-15 of the IPE submittal, the' distributions used for. the pressure rise at vessel breach- I represent the combined effects of blowdown, hydrogen burning, direct. containment . heating, and ex-vessel steam explosion. J 8. Because of the high contribution of Direct Containment Heating (DCH) : to- early overpressurization failures, it is not clear what role the induced hot leg failure depressurization plays in reducing potential DCH failures. Please explain.
Response
As indicated in Table 4.6-4, induced hot leg failure is assumed to occur with a conditional I probability of 0.72 when RCS pressure is at or near the setpoint pressure of the relief valves (' (i.e., PDSs beginning with the letter S). ' No induced hot leg failures are predicted at lower pressures because the natural circulation flow at lower pressures is insufficient to heat these -i components significantly. Table 4.8-6 shows the combined impact on Release Category Group I (large, early ; containment failures and bypasses) of eliminating induced steam generator tube and hot leg. 1 failures. Since induced steam generator tube ruptures are predicted to occur at a much lower - conditional probability than hot leg failures, nearly all of the 8.7% increase (9.0 x 10-7 per
' reactor year) in Release Category Group I frequency to 1.12 x 10-5 per reactor year results from elimination ~of the induced hot leg failures. On a conditional basis, eliminating induced failures increases the percentage of large, early containment failures and. bypasses to ' y approximately 5.3%. !
.]
l l i 1 58 of 69 .I l
'L
O
~
i This impact can also be estimated by ' Walking-through"the containment event tree (CET) quantification process. As noted 'on page 1.4-6 and in Table 1-3, the contribution to core-damage frequency for PDSs where the pressure at UTAF is at the system setpoint is 3.77 x 105 per reactor year (or 17.6% of the CDF). Subtracting the containment bypass component
- (4.0 x 10-7 per reactor year) from this total, a core damage frequency of 3.73 x 105 per reactor year is processed by CET Top Event LS (Induced PORV Failure). For RCS pressure -
at the system setpoint, PORV failures are expected 50% of the time (conditional split fraction , LS3 = 0.5). All PORV failures are expected to drop the RCS pressure out of the system setpoint range (see conditional split fraction RPV). Thus, only 1.87 x 10 5 per reactor year (0.5
- 3.73 x 10 5/ reactor-year) remains at the system setpoint pressure beyond Top Even; LS.
Neglecting the minor impact ofinduced steam generator tube failures, a CDF of 1.34 x 10-5 per reactor year (0.72 x 1.87 x 10 5/ reactor-year) is converted to low pressure prior to vessel
, brea:h because ofinduced hot leg failures, eliminating this frequency from the potential for-containment failure due to DCH effects caused by high pressure melt ejection. Given system setpoint pressure at the time of vessel breach, an upper bound estimate of the conditional probability of a large, early containment failure can be derived from ti,e product (0.1) of ,
conditional split fractions ME3 (0.92), C2S (0.1875), and L2S (0.589). Thus, ifinduced hot leg failures are eliminated, an upper bound increase in the frequency of large, .early containment failures of 1.3 x 104 per reactor year (0.1
- 1.34 x 10-5/ reactor-year) could be estimate'd. This result compares favorably with the actual sensitivity case which was summarized in Table 4.8-6.
Considering the expected similarities between Beaver Valley Units I and 2, had any differences 9. been identified that had an impact on either of the Unit's Level 2 findings? If so, please discuss.
Response
Beaver Valley Units 1 and 2 have containment buildings that are very similar in design and function, as can be seen in Tables 4.1-1,4.1-2, and the Table shown on page 4.1-2. Therefore, the Unit 1 Level 2 back-end model made use of the logic and split fraction values from the back-end model used for Unit 2. As expected, there were no major differences identified that ' had an impact on either Units' Level 2 findings. The only difTerence worth mentioning is that the Unit I large containment bypass contribution was 11% of the RCG I total, while Unit 2 had only a 4% contribution. However, this difference was expected since the Level 1 interfacing . systems LOCA initiating event frequencies for each unit were so different.~ .
- 10. With respect to the hydrogen burn issues, please address the following:
(a) Have plant walkdowns been performed to determine the probable locations of hydrogen released into the Containment? Including the use of walkdowns, discuss the process used to assure that: (i) local deflagrations would not translate to detonations given an unfavorable nearby geometry, and (ii) the containment boundary, including penetrations, would not be challenged by hydrogen burns. 59 of 69
y
~
", d i (-
.f.
Responsei Section 4.1.2 discusses the walkdown that was performed for Beaver Valley Unit 1. l As
. noted, relative to the hydrogen issue, the walkdown consisted of a general' visual
- inspection of the containment geometry and 'bpenness,"and a more detailed inspection of the reactor cavity.1 As noted on page'4.1-5, 'the configuration of structures and-equipment inside the containment appears to be conducive to good air circulation. The steam generator and pressurizer cubicles 'and most coinpe.Uwents within the containment ; "
are open at their tops to the general containment atmosphere. The. reactor vessel head : s*' W
- laydown area on the bottom floor of the. containment is completely: open to the-containment dome."
It should again be emphasized that because of the similarity between the Beaver Valley , and Surry plants, the approach adopted for the BV1 IPE was to take full advantage' of the severe accident analysis that had been performed for Surry in support of the NUREG-' 1150 study. Accordingly, many of the insights, conclusions, and numerical values for' failure probabilities were taken from the Surry evaluation. : As noted in the response to-Back-End Question 1, detonations were not addressed for Suny. As indicated in the response to Back-End Question 1, the possibility of deflagration to' detonation transitions was addressed in the BV1 IPE submittal. This possibility _was inferred from global conditions and a large containment failure was assumed given a-DDT. This latter assumption is very conservative. In NUREG-1150, failure of the: Sequoyah containment was predicted to be relatively unlikely even if DDT occurred in-the ice condenser. ' A value of 0.1 was used for containment failure, given DDT.' The ice - condenser geometry and function (condens%g steam) is much more conducive to DDTJ than the configurations oflarge, dry contain sents. (b) Please identify potential reactor hydrogen rel' ease points and vent paths. . Estimates oft compartment free volumes and vent path flow areas should also.be provided. Please specifically address how this information is used in; your assessment of hydrogen pocketing and detonation. Your discussion (including important. assumptions) should '
. cover likelihoods oflocal detonation and potentials for missile generation as a result of local detonations.
Response
There are four likely release ' points"for the release of hydrogen from the RCS into the containment for the range of accident sequences that was considered in the IPE. e Hot leg piping
~
. Cold leg piping
- Pressurizer Relief Tank (PRT) mpture disc
. Bottom head of the reactor vessel 60 of 69
my.3 , #
# =- e g3,
+
;l y o 4
,; 3 _
j a m
~ . ,
'w- ~ g ~ b
.N
+ . 1 ,
- Hot and cold leg piping failures release hydrogen mto the lower compartment of thel .
i N,< <
- containment. For LOCA initiators,- this release would be relatively slow. For thermally. l a induced hotLleg failures, the':poter.tial. exists for .a'large quantity of hydrogen'to be? 1 i suddenly released into the. containment. This phenomena is discussed in detail'in the f K response provided for Back-End Question 1.
N ' The PRT and its rupture disc are located in the lower co...y .... , which are located: , 1 approximately 40 feet above the containment floor. The bottom' head of the reactori ! vessel is located in the reactor cavity. <>
, l
>d The following table provides estimates of the compartment free volumes. j
;)
Containment Compartment Compartment Free Volume (A3) - , j J' Reactor Cavity / Instrument Tunnel 4 8,661 : Lower Compartment Upper Compartment 430,000-1,020,000
]j
. Annular Compartment - 255,000- ,
;i H
- In the above table, the reactor cavity / instrument tunnel is specific for Beaver Valley Unit . J L- 1 (see Table 4.1-1). The remaining compartment volumes were taken from the Beaver ;
Valley Unit 2 MAAP parameter file, as are the flowpath areas listed in the following; y table. m Pathway Cross-sectional Flow Area (A2) )j Flow area en=eing cavity to lower .52 1; compartment via instrument tunnel' j Bypass flow area coupling cavity to 10 j lower compartment .l Lower companment to upper . 2,448-l compartment .; Lower compartment to annular 4,360 )i compartment '! Upper compartment to annular 3,583 ! compartment l
- i r
As noted in the response to Back-End Question 1,1 detonations per se are extremely - j unlikely since the source of energy required to directly' initiate a detonation is extremely - 4 large. The transition of deflagrations to detonations (DDTs) has been addressed in the - }l BV1 IPE and was discussed in the response to Back-End Question 1. The treatment of. j DDT in the BV1 is believed to be very conservative, .Whenever conditions for DDT . j
-were predicted, a large failure of the containment was assumed,' therefore, the potential : ;
for missile generation is irrelevant. The most likely region for a DDT to occur, ifindeed j any could cccur, would be in the lower compartment region. It is not clear that the such . .! events would fail the containment boundary. Hence, we are convinced that a more- l detailed treatment of detonations would show that the BV1 IPE results are conservative l in its treatment. ! i
'}
61 of 69 ! q i _ . . _ _ _...;. d
11.: : How much are the contributions to the CDF from Steam Generator Tube Rupture (SGTR), , and interfacing systems Loss Of Coolant Accidents (LOCAs)? . Resnonse-l The contribution to the total CDF from steam generator tube ruptures is 3.45%. This is shown' > on Figure 1-3, located on page 1.4-5 of the IPE submittal. It should also be noted that after - ;
" core damage has occurred with the RCS' at system pressures (> 2,000 psia) and the steam l generators dry, the possibility ofinduced SGTRs becomes a concern. Therefore, based on'the - ;
Level 2 Top Event IS, another 0.17% of the total CDF results in induced SG11s. Interfacing ' , l systems LOCAs contributed 0.52% to the total CDF, and is it.cluded with the "Other" initiating : I events on Figure 1-3. : i t 1 i
-I t
P i l
)
i
)
62 of 69 ! q i z - ,,
s 1 3 7 E,VI IPE HUMAN RELIABILITY ANALYSIS (HRA) OUESTIONS s 1. ~ Pre-initiator human errors are stated as being evaluated separately and incorporated into each .
- Systems analysis'as a specific cause for equipment inoperability. However, the submittal'does .~
not identify specific pre-initiator ' actions considered, nor does it discuss the plant-specific analysis conducted'to support the quantification estimates of the Human Error Probabilities "
~
'(HEP) for pre-initiator actions, nor their impact on the system unavailability. It is not clear, for -
example, whether HEPs were calculated for specific human errors,'or whether. component failure data was intended to include human-related failures. Please address the following: (a) Other PRAs'have found pre-initiator human errors to be important and non-negligible contributors to core damage frequency; therefore, if pre-initiator' (including restoration and miscalibration) human errors were not specifically addressed, please provide the basis
~ for' not including them as part of your analysis. If they were addressed as part 'of .
component failure data, please address the basis.for your assumption that. the data actually captures all the pre-initiator events at BV-1, and that it accurately reflects the impact of the pre-initiators.
"(b) If pre-initiators were specifically addressed, please discuss the following:
i) If the actions were screened out, what was done to assure that the actions that = were screened out actually did have a low contribution to system unavailability c and, therefore, their contribution to CDF was indeed negligible? ii) Reviews of maintenance, test and calibration procedures for'the systems and components modeled that were performed by the Systems Analysts.. iii) Discussions that were held with appropriate plant personnel (e.g., Maintenance, Training, Operations) on the interpretation and impirentation of the plant's test, maintenance and calibration procedures to identify and understand the ' specific actions, and the specific actions and'the specific components manipulated when performing the maintenance, test or calibration task. iv) Consideration of plant-specific information such as: . plant conditions (e.g., poor lighting), human engineering (e.g., labels, accessibility, etc.), performance by same - crew, same time, adequacy of training, and adequacy of procedures in the quantification of pre-initiator events. v) How dependencies associated with pre-initiator human errors were addressed and treated.- These dependencies could, for example, affect all of the human events simultaneously, or could only affect certain human events such that only a series of human events are determined to fail simultaneously (e.g., complete dependence may be assumed for miscalibration of all reactor water level sensors). Please provide examples demonstrating how the dependencies were treated. l 63 of 69 I r
e . : o . , ,
/ 3
( . _,.j- < p : p. - {w ' l1[ a. 3 Resnonse: y
~ The IPE submittal did not systematically addreas pre-initiator human errors in the PRA (
models or component failure data. The only one that was felt to be important in the Surry
~j analysis (NUREG-IISO)'ws: mis-calibration; errors associated.with the RWST level ~ ;
transmitters. Therefore, for the IPE submittal the PRA staff carefully considered this j
* ; situation at Beaver Valley Unit 1. Upon review of past operating experience at Beaver j
Valley Unit I from January 1980 through December 1988 (i.e., the freeze dates for the' data . collection used. in the IPE),' no such errors occurred. for these instruments. : Moreover, it was felt by the PRA staff that with the staggered instrument' calibrations l j and independent verification and restoration checks performed after maintenance, test,L d and calibration procedures,'the occurrence of any such errors.would belsmall:when : compared to the total failure of a single train. Therefore, while these type of errors were i specifically looked at, they were not included in the PRA model (Top Event OR) since l they were not expected to have a significant numerical impact on the system model. -The - .: PRA model for Top Event OR did, however, account for out-of-calibration ~ errors of the ,j individual RWST level transmitters by including an exposure time equal to half of their j monthly channel functional test surveillance period in with the mission time for the basic.' event equation . .j Another type of pre-initiator human error looked at in the Surry analysis was the failure ~ to restore valves to their proper alignment after a pump test. The Beaver Valley Unit.1
] -
IPE did not explicitly' include these types of errors in the submittal. A recent review of . past operating experience in the plant problem report database did not discover any such ~ :; errors for the time period ofinterest (i.e., January 1980 through Decerwer '988) while j the plant was operating in Mode 1. Additionally, it is felt tha. since the cpus of key safety-related components are independently _ verified to be in their par nal. system-aligranent position every shift, the occurrence of such errors would be small, certainly 1
~'
less than the frequency of hardware failures due to all other causes. Likewise, important ' standby systems; e.g., the HHSI, LHSI,~ quench spray, and recirculation spray systems, j are verified every. 31 days to insure that each valve (manual, power operated, or-automatic) in the flow path that is not locked, sealed, or othenvise secured in position is in its correct position. It can be argued that the MGL method fo' rquantifying common- . cause failures implicitly models potential pre-initiator errors that could impact multiple ' 1 trains of systems. The PRA staff concluded that.it was not necessary 'to model pre-initiator errors that could impact multiple trains because of the thorough treatment of 4 common cause failures already considered. d I Furthermore, it should be noted that system . mis-alignments and human errorsLwhich resulted in 'a reactor trip are included in with the appropriate initiating event frequencie's for such transients. These types of plant trips are included in with the Unit I reactor trip ' -; events presented in Table 3.3.1-6 of the IPE submittal. An example of this would be - Event No. 31, listed on page 3.3-26, in which the main steam trip valves closed due to a - test engineer inadvertently isolating the air supply. This event was included with the closure of all main steam isolation valves (AMSIV) initiating event frequency. 64 of 69 _. . m _ - - _ .. . _. . . . _ . . - . - _ _ _. _ a . . _ . . _
e
; Q .
b y [ l ^ ~ 12! While the Success Likelihood Inder Methodology (SLIM)-based analysis inherently provides'-
- ; a means for systematic im444, .Gon of subjective evaluation of plant-specific performance .
. . shaping factors,'thejudgment of the assessment teams is influenced by.the number and . type ofs
" personnel in'the: group, and information that is prepared and presented to them.1Please -
. identify the number of groups '(teams) which participated in the ratings, and the number and 4
* -type 'of personnel (i.e., PRA/HRA analyst,' opera 3ons, maintenance or training ~ personnel)' '
within each group. Response. The BV1 personnel which evaluated the human reliability actions consisted of one group with .
~
N' the following team members:
' 1 - PRA/HRA analyst _
1 - PRA/HRA analyst (former SRO license trainee and STA) 1 - Beaver Valley Unit 1 senior nuclear operations instructor (former SRO)
- 1 - Beaver Valley Unit I licensed reactor operator -
Because of scheduling conflicts and time limitations, it was not practical for the entire group ti L evaluate all of the human actions. Therefore, only the most important actions' deemed by.the . HRA analysts were rated by the entire group. These actions _which were evaluated by this group are shown in Table 3.3.3-6 as having a "Yes". in the " Action Rated by Operations &' Training" column. This group was given a brief description of the action to be performed, the scenario and prior conoitions for the action, available procedures. and time frame re,_ ired to
- perform the action, for each human reliability action rated by this group.
Actions not rated by operations ud training personnel (i.e., shown in Table 3.3.3-6 as having a "No" in the "Actior. rated by Qerations & Training", column) were analyzed by the HRA analysts using similar previously evaluated actions by the group as n' guide, or using the Beaver -
-Valley, Unit 2 HRA as s' guide for similar actions, or both. The Beaver Valley; Emergency-
' Operating Procedures are based on Westinghouse Owners Group Emergency' Response Guidelines, and have similar actions between the Units in response to'an accident.4Therefore, similar actions which were evaluated by Beaver Valley Unit 2 personnel were slightly modified
' by the PRA/HRA analysts for Beaver Valley. Unit 1 actions. The Beaver Valley Unit 2 HRA team used the same PRA/HRA analysts used for. Unit 1,- different operations' and training personnel, and an additional PLG HRA analyst.
- 3. The submittal ' emphasizes'.the strength of the SLIM-based methodology for. addressing
' dependencies among , post-initiator response (dynamic) actions through . the. subjective evaluation process which considers actions in the context of the scenarios in which they are imbedded. The submittal states (in Table 3.3.3-2, Sheet 2 of 7) that, "if necessary, some strongly dependent failures may be accounted for by spec _ific split fractions in event trees".
No information is provided as to what constitutes a strongly dependent failure, or the criteria used to identify one. Please identify the criteria used to identify "strongly dependent" failures, and identify if any were considered in the HRA analysis. Provide examples of how 65 of 69
g - .. , . , _. . .. I 1 m
' ' fw ,
, j s
r
~f q
~
~
. . . . 1 4
dW% and the level of dependency, were factored into the HEP quantification and.. )
, . addressed between 6perator actiom in separate top events in the event trees; j s i Response-
]
The SLIM approach to human error rate quantification considers dependencies between I actions occurring in the same. accident sequence via the second performance shaping factor;; j
. i.e., the one for Signi6 cant Preceding and Concurrent Actions. This factor, and the others, are j evaluated by the plant operating staff and the PRA analysts for every dynamic action. The 1 influence on the final human error rate of this action is then determined by Equation 3.3.3.1 ofs i the submittal for the failure likelihood index and the calibration curves.
~ '
When dynamic actions are dependent, the human error rate assigned to the second action : . should be made dap-%t on the outcome of the first action. For a particular sequence, the j success or failure of the first action can be inferred by the status of the top event in which it is . l modeled. Then the error rate for the second action in the sequence can be made dependent on; ; the status of the top event in which the first action appears. The split fraction assignment logic i used by the event tree quantification code (RISKMAN) is structured so that the split fraction .
"for the top event which contains the second operator action then reflects this dependency. ll a '
For example, in the event of a *w of auxiliary feedwater two actions were identified in the 1
- event sequence. diagrams'and modeled for restoring core cooling; i.e., restoration ~of main?
l feedwater and initiation of bleed-and-feed cooling. These actions are directed by the same .i emergency procedure. Therefore, they were judged dependent; The first action is modeled via Top Event OF. The second action is modeled via Top Event ' 1 OB. In this case, bleed and feed cooling is only required if restoration of main feedwater is j unsuccessful. However, there.are two categories of reasons for failing to restore main ! feedwater. The necessary hardware may be unavailable (i.e., Top Event MF fails), or. the operating staff may have failed to perform the restoration (i.e., Top Event OF fails). l Since the act?ons in Top Events OF and OB are dependent, the PRA analysts concluded that separate error rates for Top Event OB should be calculated depending on the status of Top. :} Event OF. The different error rates are reported in Table 3.3.3-5 as ZHEOBl (1.22 x.103) .
= and ZHEOB2 (1.39 x 10-2). These different error rates are then used in the quantification of. - .!
the different split fractions for Top Event OB. The split fraction assignment rss for Top q Event OB are then made dependent on the status of Top Event OF. If Top Event,OF fails,; ., only Top Event OB split fractions which use ZHEOB2 are assigned. Conversely, if Top Event J OF succeeds but Top Event MF fails so that bleed and _t' e ed is still required, only split fractions - 'I which use ZHEOB1 are assigned. Numerous other event tree top events which consider multiple dynamic actions are considered - in the Beaver Valley Unit 1 IPE models. These can be seen in Table 3.3.3-5. For _ example, seven different operator action error rates are used for Top Event CD (i.e., for cooldown' and . depressurization) depending on the specific sequence conditions that exist. The different error rates are used in different split fractions for Top Event CD, and the appropriate split fractions are then selected during event tree quantification based on the split fraction assignment logic. 66 of 69
4 m ' ' ( & q r . j
. . l1 Similarly, five error rates are used for the~ actions in Top Event MU and three are used for_ g
.those in Top Event SL. !
I
- Actions appearing in the same accident sequence are identified as strongly dependent if they j are directed at the same goal, guidance is provided by the same procedure, and time period in : j
-which the actions are to occur are roughly the same time frame. Actions directed at the same 3 :
goal but separated by several hours in time, are not said to be strongly dependent. No explicit j numerical criteria were used for assigning the dependence between two actions as strong. 7
- Rather, the PRA analysts used judgment on a case-by-case basis to determine wiiether the' .
A [> . above stated crit 6ria are met. For strongly dependent actions'in the same sequence, it is recognized that the impact' of the j dependency between actions on the human error' rate for the second action may be more- -i pronounced than can be realized through the linear equation for combining performance. 1 shaping factors. For such dependencies, the PRA analysts, in many cases, decide to take no credit for the second action; i.e., assign an error rate of 1.0 to it. In such cases, the SLIMf , quantification_ model is omitted as 'the limitations of the model, in extreme _. cases, are . j acknowledged. Then during sequence quantification, the split fraction assigned to the top . event, which accounts for the second human action, is set to 1.0 for that sequence. l i One example of strongly dependent actions modeled for Beaver Valley-Unit 1 involves the j actions to initiate recirculation from the sump following a small LOCA (Top Event OR) and the action to align for long-term makeup to the RWST (Top Event MU) given recirculation _ from the containment sump is unavailable. In the split fraction assignment logic for Top Event i "MU, when Top Event OR fails earlier in the sequence, no credit was taken for Top Event' . MU; i.e., effectively the operator error rate was set to 1.0 by assigning a split fraction with a l value of 1.0. l t 1 A second example is that for initiation of manual control rod insertion (Top Event RI) and l emergency boration (Top Event OA) during an ATWS following attempts by the operators to : ] manually initiate a reactor trip (Top Event OT). If Top Event OT fails, error rate ZHERI2, !! which has a value of 1.0, was used for Top Event RI. Also, if Top Event OT fails, no credit ; l was taken for emergency boration via Top Event OA. These dependencies between the three t , actions were accounted for during event tree quantification by the split fraction assignment ; logic. ! J A final point is that the thought process used to apply the event tree linking methodology of RISKMAN is that all split fractions for both hardware failures and human errors are presumed to be dependent until proven otherwise. The identification of dependencies between split , fractions is a central task in constructing a RISKMAN PRA model. When completed, all of: ; the dependencies modeled are evident in the split fraction assignment rules files. j
- 4. Timing of operator actions is specifically addressed in the qualitative and quantitative analysis ;
herformed in the evaluation of post-initiator actions. The submittal notes that there is a !
"relatively well-defined time _ window available for successful operator response". It also notes i that timing determines important factors that influence the operators' ability to diagnose the problem, decide what actions are appropriate, and complete those actions within the required ;
67 of 69 . t
, 4 . .- __ ,, _ . , , - .,
w: . ym o lA - g4 , n L
' time window. However, little detail as .to-how time required to' perform was actually evaluated, or whether additional factors may have been important for out-of-Control Roon ~
; actions. Thus, it is not clear thatlwalkdowns were performed for HRA. purposes and walkdown-time measurements were taken for time-critical actions or arrived at by simulator.
runs; whether assumpons about accessibility, availability of tools, etc., were verified by walk- .
~'
throughs'or " simulations" of operator actions in the plant, and how environmental factors and "other physiological or psychological " stressors" were accounted. (a) Discuss by way of examples how such factors were addressed for operator, actions,
' especially for the out-of-Control Room actions.-
Response
- Walkdowns and walkdown-time measurements;were not performed specifically.for'the.-
. HRA. The specific time available for operator actions to bel performed are located on Table 3.3.3-5. These times were either calculated by thermal-hydraulic analyses, ~past?
plant experience, other reference analyses, or by simulator runs. In addition to the time window available, the SLIM process also accounts for psychological and cognitive; conditions of the operators based on existing procedures, training, and stress factors. It is ; in these ratings of procedures and training-that assumptions about accessibility and; availability . of tools are accounted for-- by actual plant procedure implementationi experience. This knowledge also provides a sense of how much time is required ~to - perform actions carried out in past experiences, which can be compared to the time frame available. The feasibility of performing each action within the time frame available was - discussed among the HRA group before ' performing the: qualitative rankings. Environmental factors and other physiological or psychological " stressors" .were: accounted for in the stress performance shaping factor rankings. Examples can be seen in Table 3.3.3-6, e.g., ZHEAFI and ZHECD4.' ZHEAFI was rated as a 5 in its procedure PSF,' and also as 5's in training and stress. These ratings were evaluated by the HRA group. The time frame available was calculated by thermal-hydraulic analysis based on the time for a steam generator to dryout to 10% : < on the wide range level indicator with no feedwater. flow available.' Since this human action is the same action used during surveillance testing or is part of normal training, the operators have a good basis on which *a judge accessibility and availability of tools. They/ also know how much time is required to' complete the action and any stress associated with the action through past experience. Another example, ZHECD4 was rated as being 8's in both procedures and training, and ? as'a 10 in the stress rating. Once again these actions were rated by the' HRA group,- however the time frame available was based on simulator runs to see how' fast the - operators could cooldown and depressurize the RCS to below 212 *F before the RWST - empties, given an initial 800 gpm SGTR with a stuck-open SG atmospheric steam dump L valve. Only vague guidance exists in procedures to perform this action, since it requires y local manipulation of steam dump valves to cooldown during the loss of AC orange L power. Also this is a non-routine action, but is an option in annual or biannual training. 68 of 69
~'
i_.l_._._______'_4. - _ _ _ _ m__r- . _ . .u.. . . e , ---o n -
n . f, - , ww.+m ~ ~ ' -- '~ - - - o x .
-_ . Even though there is sufficient time available to complete the action at normal speed and ~ !
. to verify results, the-' pre-existing conditions:and local environment in which ' to' l accomplish the action, puts tremendous physiological and psychological stresses on the -
operator. These undue stresses are reflected in the stress rating of 10. l j (b) Identify which recovery' actions are out-of-Control Room actions. i
~
p The following human actions are performed outside of the control room. A brief-D desciiption of the action to be performed is provided in Table 3.3.3-5 and Table 3.3.3-10.-
.}
ZHEAFI ZHECD7~ ZHEIA3 ' ZHEOF5 ZHERE8. ZHEBV1 . ZHECII ZHEIA4 ZHERE1 -ZHERE9 . ZHEBV3 ZHECTI. ZHEIC1 ZHERE2- ZHEREA- l ' ZHECCI ZHEDF1 ZHEIC2 ZHERE3 ZHERED ZHECC2' ZHEFL4 ZHEIC3 ZHERE4 ZHEREE .i ZHECD2. ZHEHH1- ZHEMA1 ZHERES 'ZHEREH ZHECD4- ZHEIAl ZHEMA2 ZHERE6 ZHESL2 l
~
ZHECDS ' ZHEIA2 ZHEOF3 ZHERE7 ZHESL3 It shculd be noted that, although the description in Table 3.3.3-5 for action ZHEIA2 j states that it is performed from the Control Room, the action was actually evaluated for starting the diesel driven air compressor outside the Control Room. j l
'l l
t i r l 69 of 69
- j
~ - ^ '
n g1 ",4~
,, , e -
u l a u a l y > ,
< (d)" Descriptions of the containment failure sizes and locations. q ws.
] q* __
q As'noted~in Section'4.4-1,1the Beaver Valley containments were believed to be very . i similar to the Surry containment which was 1 analyzed lin NUREG-1150. Upon : 1 4 confirmation of this similarity, it . was'. determined that the probability Ldistributions '
.(containment failure and conditional probability oflarge vsc small failure) for' pressure D -
- . capacity developed for Surry in Reference 7.d could be used for Beaver Valley.1 As a' i result, no plant specific evaluation'of containment pressure capacity was performed for? ,;
Beaver Valley. In Reference .7.d, it .was noted that n'large hole or rupture is one fori j which the containment would depressurize in less than approximately 2 hours. -'It was: ' 4i also noted that large, dry containments would depressurize in 2 hours for. hole sizes on! the order of 0.3 to 0.5 ft2. 'It was then stated that a small hole or leak should be of the order of 0.1 ft2 A failure area of.1.0 ft2 was identified as 'tlefinitely a large hole or j rupture." Thus, in the CET quantification process,L small and large failures are typified . j by failure areas of 0.1 ft2 and greater than approximately 1.0 ft2, respectively. j. The Surry pressure capacity distribution is a composite of four expert analyses.and"a - l number of failure modes including hoop failure in the cylinder and' dome, shear at thei ! cylinder-basemat junction, liner tearing, and failure at the penetrations. The composite - , nature of the curve makes it" difficult to' identify specific failure locations at any given j failure pressure. In Reference 7,d, it is noted that ' failure location did not turn out to be j important since any failure location except shear at the basemat-cylinder junction would ' 1 result in a direct path to the outside."- 1
.i As noted in Section 4.7 (see Table 4.7-7), all large containment failures.were assumed to - 1 release fission products directly to the environment. Some ex-containment retention was s j credited only for small containment failures. Reference 7.d is as follows: 4
. l 7.d Breeding, R.J., et al, ' Evaluation of Severe Accident Riskst Quantification of- 1 Major Input Parameters, Experts'. Determination of Structural Response Issues," ' 1 NUREG/CR-4551,- Vol. 2, Rev.1, Part 3, March 1992 'l (e) A quantification of contributors to small, early containment failure. .
j I Response-1
, Table 4.8-4 on pages 4.8-62 and 4.8-63 of the IPE submittal lists the non-guaranteed' ';
^
failure split fraction importance for the major contributors to the small, early containment j failure and bypass release category group. The Level 2 top event failures and respective l RCG II contributions that lead directly to a small, early containment failure or bypass are l listed below: J l Top Event BY (Containment bypassed prior to core damage) = 16.3% .. Top Event C1 (Containment failure prior to vessel breach) = 72.1% j ' Top Event C2 (Containment failure at vessel breach) = 13.0% Top Event CE (Containment failure due to early H2 burn) =0.0%
$6 of 69
.~ 'j
+ w -. - -- _ - - . . - -.. - I
The reason that the total exceeds 100% is that there is an overlap of sequences which involve both C1 and C2 failures. This is due to the containment event tree structure that still inquires if a containment failure at vessel breach occurs as a result of a HPME, since these types of failures can lead to a large, early containment failure even though smail containment isolation failures occurred prior to vessel breach. The majority of failures to this release category group (88.4%) come from small containment isolation failures and bypasses from the Level 1 analysis. Early hydrogen burns below a 12% concentration are not expected to fail the containment structure within 4 hours of vessel breach, and burns above a 12% concentration result in global detonations that lead directly to a large containment failure. (f) The size assumed for a "large" bypass.
Response
All large containment bypasses are governed by the Level 1 interfacing systems LOCA events or by a Level 2 induced steam generator tube rupture. The interfacing systems. LOCA (i.e., containment bypass LOCA) is addressed in Section 3.1.3.6 of the IPE submittal. As discussed in this section, the LHSI/RCS pathway includes a single 6 in. diameter header that penetrates the containment at Penetration No. 61. The piping from the RCS connection to the normally open motor operated containment isolation valve located outside the containment is designed to withstand normal RCS pressure, as is the valve itself. The 10 in. diameter piping upstream of the isolation valve is not designed for high pressure and is predicted to fail when pressurized to RCS conditions.-Break flow through a rupture in the 10 in. piping would be restricted by the flow areas associated with failed check valves, which could be almost as small as the flow areas associated with the LHSI relief valves (105 gpm total choked flow), but not more than that associated with the single 6 in. line through the containment penetration. For the induced SGTR, it was assumed that, if such an event were to occur, the primary system pressure would be high enough to lift the secondary side safety valves, which have a 10 in. diameter outlet, thus creating a containment bypass route during the time of core overheating and fissian product release from the fuel. It should be noted that for the induced SGTR the associated minimum bypass area could be restricted by the number of tubes failed. (g) The contribution of global detonation to conditional containment failure probability (large, early; small, early; and late).
Response
The reference to " global detonation" is somewhat confusing since no such terminology is used in the submittal. As indicated in the response to Back-End Question 1, a global > hydrogen concentration of 12% was used as a " benchmark" for the occurrence of a deflagration-to-detonation transition (DDT). In effect, it was assumed that localized conditions for DDT could be achieved if global concentrations reached 12%. 57 of 69
m ,
,t- ,
s 3
" , (
l
- 3 y The percentage contributions of hydrogen burn /DDT to the various release category
^
groups are as follows: jl N ,
' O.2% of RCG I - Large, Early. Containment Failures and Bypasses . j j
y (based on importance of conditional split fraction LEF) 0.0% of RCG II - Small, Early Containment Failures and Bypasses ; (DDTs are assumed to cause large containment failures) 11.3% of RCG III - Lite Containment Failures (based on combined importance of conditional split fractions C3A and C3C) g It should also be noted that containment failures at vessel beach'(CET Top Events C2.. .i and L2) due to HPME are due in part to hydrogen combustion. As indicated on page 4.6-- j 15 of the IPE submittal, the distributions used for the pressure rise at vessel breach ~ l represent the combined eff'ects .of blowdown, hydrogen burning, direct ' containment 1 . heating, and ex-vessel steam explosion. ]:; l
- 8. Because . of the high contribution of Direct Containment ~ Heating- (DCH) to early. :f
- overpressurization failures, it is not clear what role the induced hot leg failure depressurization j plays in reducing potential DCH failures. Please explain. !
ResD0nse'
.l As indicated in Table 4.6-4, induced hot leg failure is assumed to occur with a conditional -
i;
probability of 0.72 when RCS pressure is at or near the setpoint pressure.of the relief valves ;
(i.e., PDSs beginning with the letter S). No induced hot leg failures are predicted at lower j pressures because the natural circulktion flow at lower pressures is insufficient to heat _these - j components significantly. 1 Table 4.8-6 shows the- combined impact on' Release Category Group I (large, _early . j containment failures and bypasses) of eliminating induced steam generator tube 'and hot leg - ' failures. Since induced steam generator tube ruptures are predicted to occur at a much lower conditional probability than' hot leg failures, nearly.all,of the 8.7% increase (9.0 x 10-7 per reactor year) in Release Category Group I frequency to 1.12 x 10t per reactor year results i from elimination of the induced hot leg failures. On 'a c'onditional basis, eliminating induced! j failures increases. the percentage of large, early containment- failures and bypasses to j approximately 5.3%. f i 58 of 69 l l __ _ _. . _ . _ , .. _ -_ 4
'f
_.,}}