ML20084R092

From kanterella
Jump to navigation Jump to search
Draft AP600 Thermal/Hydraulic Uncertainty Assessment, Assessing Impact of Thermal/Hydraulic Uncertainty on AP600 PRA Success Criteria
ML20084R092
Person / Time
Site: 05200003
Issue date: 05/31/1995
From:
WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP.
To:
Shared Package
ML20084R068 List:
References
NUDOCS 9506090180
Download: ML20084R092 (18)


Text

'

DRAFT e

AP600 Thermal /Hydrmlic Une rtainty Assessnwnt ASSESSING THE IMPACT OF THERMAIJHYDRAULIC UNCERTAINTY ON AP600 PRA SUCCESS CRITERIA Introduction and Background The NRC has asked that Westinghouse perform an assessment of the potential that the AP600 probabilistic risk assessment (PRA) success criteria could be significantly affected by uncenainty in the prediction of the thermal / hydraulic (TM) performance of passive systems.

Specifically, the NRC has requested that Westinghouse demonstrate that any potential impact on core damage due to such TM uncenainty is significantly smaller than the core damage frequency due to hardware failures and human error.

The NRC is asking that Westinghouse demonstrate that sequences which have been labeled as success, as a result of the extensive success criteria modeling effons performed for the PRA.

remain as success sequen:es even when reasonable TM uncertainty is accounted for. An approach has been formulated to address this issue which will demonstrate that, for all success sequences, one of the following is true: the success criteria are robust with respect td TM uncenainties such that there is sufficient margin to com damage with the existing success criteria (the preferred outcome); or, if inclusion of TM uncertainty does affect the success of a given sequence, the effect on the results of the PRA is not significant. This approach complements, not replaces, the systematic approach to definition of success criteria which Westinghouse has already undenaken for the PRA,in suppon of which over 400 computer analyses have been performed. This effon is being perfonned for AP600 in order to address an NRC concern that TM uncertainty may be more important for passive plants than for evolutionary plants since the driving heads for core cooling flow are not provided by pumps.

This evaluation is new and unique in its application to PRA, so it is possible that the approach that is outlined in the following paragraphs will require some modification as the assessment proceeds. To avoid unnecessary delays or rework that could result if a portion of the process requires modification, the tasks have been defined to include several progress review and test points, so that the NRC has an opportunity to review and comment on results at intermediate stages, and so that the approach can be tried first on a limited set of sequences with a subset of TM parameters. Then the rest of the evaluation can proceed, with any necessary modifications identified from the test case.

The list of tasks to be performed for the TM uncertainty assessment was provided to the NRC in reference 1. The assessment generally involves: selection of the sequence success criteria to be evaluated including the effects of TM uncenainty; identification of important TM parameters and selection of bounding values; analysis of the sequence success criteria i

V,'estmpbouse Electne Corporauon i

May IWS t

9506090180 950531 PDR ADOCK 05200003 A

PDR

e DRAFT P

AP600 Thermal / Hydraulic L.'ncertainty Assessment l

o i

with TM uncertainty included: and an evaluation of any resulting effect of TM uncenainty on the PRA results. His document is the deliverable for assessment Task 2, " Prepare Program Definition and Binning Criteria with Examples." Task 2 is defined as follows.

i The objecuve of Task 2 is to document the approach Wesungbouse presented on Apnl 20,1995 to the NRC for

)

addressing T/H uncertainty and to provide a desenpuon and an example of the bmning process that will be used to reduce tbc number of sequences that must be examined. De entena Wesungbouse plans to use for bmning of sequences and for Jusufymg the bounding sequence selecuon will be provided to the NRC for review and l

comment with an example. This will demonstrate bow the Wesunghouse approach relates to the NRC Margms i

Approach to resolvmg tbc issue of passive system rehabihty.

The remainder of this document is organized as follows. First, a discussion of the technical basis involved in the assessment of TM uncertainties on the PRA success sequences is presented, along with a set of definitions of terms appropriate to this assessment. Next, a discussion of the preliminary criteria to be used for any binning of sequences, and for justifying that bounding of sequences has been properly performed,is presented. Examples of binning and bounding are provided within this discussion. Finally, the steps in the NRC l

" Risk-Based Margins Approach for Passive System Performance Reliability Analysis," as presented by the NRC to Westinghouse on April 20,1995, are discussed to demonstrate how' l

this approach coincides with the NRC Margins approach. His document does not address issues specific to the identification and selection of TM parameters or values, or to the analysis of the sources and effects of uncenainties; these are the subjects of later tasks.

l 1

Technical Bases The AP600 PRA provides a prediction of the ability of the plant and operators to prevent and/

or mitigate severe accidents. This prediction is a function of the uncertainties in assessing the l

probabilities ci success and failure for the various systems and actions. Uncenainty in hardware and human reliability is addressed explicitly in the PRA. There is also uncenainty l

in the analyses performed to define the success criteria for the various event sequences. This analytical, or thermal / hydraulic, uncertainty is not readily addressed explicitly within the l

PRA.

For the AP600, the NRC and Westinghouse have agreed to employ a " margins" approach to assessing the impact of TM uncenainty on the PRA results. In this approach, the effect of TM uncertainty on the various PRA success criteria is examined to show that the success criteria are conservative, even when this uncenainty is included. Defining parameters and criteria conservatively is not normally done in a PRA, in order to provide insights into the more likely plant and operator responses to different accident scenarios The use of a bounding method precludes any need to quantify TM reliability and simplifies the overall process.

Westingbouse Electnc Corporauon 2

May IWS i

DRAFT AP600 Thermal / Hydraulic l'ncertainty Assessment Thermal / hydraulic reliability has been defined (reference 2) as "the probability that a sequence involving one or more passive systems will not lead to core damage due to failures associated with uncettainties in the TM performance of passive systems." This assumes that an assessment of the effect of TM uncertainty on the success criteria can be used to address TM reliability concerns.

Obiective of this Assessment. The objective of this assessment is to show that the success critena for sequences involving passive systems, as defined through a systematic and conservative process for the PRA, are not sensitive to uncertainties in/he prediction of plant t

TM performance Definitions. In the discussions that follow, a number of terms are used to denote specific meanings. These are defined as follows A modeled sequence is any combination of equipment operation or operator action successes or failures shown explicitly on one of the event trees (Figures 4-1 through 4-26) of the AP600 PRA (reference 3).

A bounded sequence is any combination of equipment operation or operator action success or failures not explicitly shown on the event trees of the AP600 PRA, but with consequences which are less severe than, and therefore, bounded by, one of the modeled sequences.

Bounding, in general for this assessment, refers to the process of determining that the effects of a parameter of interest (e.g., a sequence, or a system success criterion, or a sequence success criterion) for a s-t of similar scenarios may be measured by examining the effects of a scenario shown to have the greatest effect within the set.

A success criterion (for a given sequence) is defined as the set of criteria of each of the various systems required to function for success of the sequence (i.e., no com damage). The following is an example of the success criterionfor path 2 of the medium Loss of Coolant Accident (LOCA) event tree, Figure 4-2 of the PRA (repeated here as Figure 1): injection from I out of 2 core makeup tanks (CMTs) AND opening of 2 out of 4 automatic depressurization system (ADS) Stage 4 salves AND opening of check valves in 1 out of 2 in-containment refueling water storage tank (iRWST) injection lines AND opening of valves in 1 out of 4 recirculation lines. There is one sequence success criterion for each success path in the PRA: these are summarized in Table 6-2 of he PRA.

t A success criteria baseline case is one for which the TM modeling includes the effects of the following:

minimum injection (typically 1 CMT or 1 accumulator, and I line of IRWST injection):

o o the worst break location and break size for the event category; failure of containment isolation (which is limiting for IRWST injection); and o

the longest credited operator action time (where applicable).

o Wesunghouse Electnc Corporanon 3

May IWS

DRAFT AP600 Thermal / Hydraulic Uncertainty Assessment Any baseline case is conservative, since it bases modeled sequence success on a bounded sequence which includes multiple worst-case assumptions, and whose probability of i

occurrence thus falls within the " tail" of the distribution of probabilities of all sequences represented by the modeled sequence.

Binning is a process in which initiating events or event sequences are grouped together and analyzed as if they were a single event. These events must be similar with respect to characterstics such as plant reactivity and thermal / hydraulic response, system or operator

~

response, sequence timing, or end results. Each bin is represented by-a success criteria baseline case.

Preliminary Criteria for Sequence Selection, Bounding and Binning Because of the large number of scenarios modeled in a PRA, it is not practical to consider each individually. However, as noted in the definitions for binning and bounding, similar scenarios can be examined together, and this has been done in constructing the AP600 PRA event tree models. The following paragraphs discuss the manner and extent to which success criteria selecuons will be made for the assessment of TM uncertainty.

Success Criteria Based on Design Basis Analyses. Cenain success criteria are defined from analyses performed using design basis analysis codes and modeling assumptions. Such analyses already incorporate allowances to account for uncenainties, including TM uncertainties, sufficient to suppon the bounding natum of the results. As a result, sequence success criteria based on design basis analyses need not be evaluated in this assessment of the impact of T/H uncenainty on the PRA results, because of the conservative natum of those analyses.

Success Criteria Involvine Operation of Active Systems. Some sequences involve operation of active systems (either in addition to or instead of passive systems) for success. Examples of active systems modeled in the PRA include main and stanup feedwater, and normal residual heat removal. The primary issue being addressed in this assessment is T/H uncenainty related to passive systems, and it is not necessary to address uncenainty in sequences involving active systems, for the following reasons.

Thermal / hydraulic uncertainty has not been an issue relative to success criteria for active systems. The nature of active systems is that success criteria for preventing core damage depend strongly on whether.or not the hardware operates, and less on the specific conditions under which the hardware operates during a sequence. Thus, the thermal / hydraulic uncertainty associated with predicting the performance of active systems, which typically involve the use of pumps, turbines, and so fonh for motive power, can be reasonably estimated and bounded by analyses.

Wesunghoue Elecme Cor rauon 4

May IWS

DRAFT AP600 Thermal / Hydraulic Uncertainty Assessment For sequences in which passive systems and active systems are modeled, the passive system would be operating under less challenging conditions than in passive-only sequences.

Therefore, the effect of T/H uncertainty for active / passive sequences would be less than for passive-only sequences.

Also, for any success path in the PRA in which an active system is modeled, there is an equivalent success path involving only passive systems. Thus, if the active / passive system success path were determined to result in failure rather than success, the sequence would continue with the backup passive system, potentially resulting in success. An example of this can be seen in medium LOCA sequences 1 and 2 (see Figure 1). Sequence 1 involves success of CMT injection (a passive system) and normal RHR (an active system). If this sequence were to somehow be determined to be unsuccessful as a result of failure of the normal residual heat removal system (RNS), the result would not be core damage, but rather a different scenario (in this case sequence 2), in which, if RNS fails, success can be achieved if ADS operates to depressurize the RCS and IRWST gravity injection and recirculation (passive functions) operate to provide RCS inventory control and core cooling. For these reasons, it is only necessary to examine the passive-only sequences in order to evaluate the impact of T/H uncertainty.

Potential Adverse Passive / Active System Interactions. The above justification for examining only sequences without active systems relies in part on the absence of any significant adverse active / passive system interactions. The operation or failure of active systems will not adversely affect the operation of passive systems. Further, the operation of additional equipment beyond the minimum required for success of a sequence (i.e.,2 CMTs instead of 1, or 1 CMT and I accumulator instead of I CMT only) will not lead to a worse (i.e., higher peak core temperature) transient. Such potential system interactions have been and are being examined. Documentation of the resuhs of these assessments will be documented. separately from the T/H uncertainty assessment. No significant adverse interactions have been identified. Should these ongoing examinations identify an adverse interaction with potential impact on this T/H uncertainty assessment, it will be addressed Binning of Success Criteria. To the extent possible, binning of success criteria for the pumoses of evaluating the effects of T/H uncertainty will utilize the binning already performed in setting up the PRA event tree models. Bins defined in the PRA event trees include the various LOCA categories and transients. From the standpoint of minimizing the number of success criteria to be evaluated, it may be appropriate to combine some of the LOCA and/or transient criteria, making use of the extensive set of success criteria analyses performed to date to provide the necessary justifications for the combined bins. Any further binning will be done on the basis of the success criteria already defined in the PRA.

An example can be seen by referring to Table 1, which summarizes the initial set of success criteria to be considered for bins for the T/H uncertainty assessment. (These criteria are taken from PRA Table 6-2.) The success criteria for success sequence number 2 (for medium LOCA) and sequence number 4 (for intermediate LOCA) are the same. Thus, it may be beneficial to consider these two in a single bin. The analyses for this combined bin would then need to cover a range of conditions appropriate to both the medium LOCA and the Westmgbouse Electnc Corporauon 5

Ma lW5

DRAFT AP600 Thermal / Hydraulic Uncertainty Assessment intermediate LOCA break sizes, locations, and so fonh. Similarly, considering the success criteria for success sequence number 6 (small LOCA) and sequence number 8 (transicnt whn failure of decay heat removal) in a single bin may be appropdate, based on similar plant behavior in these sequences.

As another example, within the transient event categories, the passive success criteria for most of the transient initiating events include sequences involving CMT case CM2AB, ADS case ADA, gravity injection case IW2AB, and recirculation case RECIRC.,All such sequences could be considered to be a single bin for the purpose of assessing the impact of T/H uncertainty, since, once decay heat removal is lost following a transient, the plant response for any sequence with the same set of success criteria is essentially independent of the initiating event.

These binning examples are, at this stage of the assessment, preliminary, and for illustrative purposes only. In general, the initial set of success criteria bins to be considered for this assessment is the set of passive-only baseline success criteria for success criteria other than those based on design basis analyses, that wem developed for the PRA. Tnis initial set is presented in Table 1. It is an initial set at this time because it is possible that, as the assessment progresses, cases may be identified that are similar enough such that one can be bounded by another, as discussed in the preceding paragraphs. For such cases in which success criteria are to be combined into a single bin, justification for the binning would be provided as part of the final assessment documentation. A more clearly defined set of bins will be defined in conjunction with Tasks 4 and 5 of this assessment.

Success Criteria Bounding. As stated in the definitions provided for modeled and bounded sequences, each modeled sequence represents several bounded sequences. The success criterion assigned to a modeled sequence is therefore required to be appropriate for all associated bounded sequences. The selection process for basel ne case success criteria (per Task 3 of this assessment) has been defined in a manner such that the baseline case success critena for a modeled sequence bounds the success criteria for each bounded sequence. This is achieved by using conservative assumptions such as minimum injection sources (e.g.,1 of 2 CMTs or 1 of 2 accumulators, I of 2 IRWST injection paths. I of 2 recirculation paths), as well as longest credited operator msponse time, and worst break size and location in the baseline success criteria. Thus, any given success criterion is based on the " bottom-most" bounded sequence, which, in the absence of adverse system interactions, produces the highest peak core temperature for the modeled sequence.

Since the initial set of success criteria bins includes one bin for each sequence (considering the various transients together), the amount of bounding to be done is limited to the individual bins. For example, for medium LOCA, there would be bins for the passive-only success paths on the medium LOCA event tree. These are the paths labeled MLO-OK2 (path

2) and MLO-OK4 (path 8) on Figure 1. If these paths were expanded to show the bounded sequences, the result would be as shown in Figure 2. (For the pugoses of T/H Uncertainty Task 2. Figure 2 is illustrative only.) All paths on Figure 2 labeled MLO-OK2 represent bounded sequences, that is, sequences bounded by gath 2 on Figure 1. Similarly, all paths on Figure 2 labeled MLO-OK4 represent sequences bounded by path 8 on Figure 1.

Wesungbouse Electnc Corporauon 6

MavIWS I_

DRAFT AP600 Thermal /Hydrartic Uncertainty Assessment Table I lists the initial set of sequences that have been identified as being bounded by the various baseline success criteria. This listing is preliminary at this time for the same reasons as discussed under binning.

Assessment of the impact of T/H Uncertainty on Bounded Seouences. If a modeled success criterior. can be shown to be successful by analyzing its baseline case with the effects of T/H uncertainty included, then the analysis will have demonstrated, for that particular modeled sequence and for its associated bounded sequences, that T/H uncertainty does not affect the PRA results. If a modeled success criterion cannot be shown to be successful by analyzing the baseline case with the effects of T/H uncertainty included, then additional analysis is required.

For the latter case, the approach calls for finding a less restrictive set of hardware failures for which the criterion succeeds. As an example, if the baseline success criterion for modeled sequence MLO-OK4 (see Figure 1) could not be shown to be successful with T/H uncertainties included, an approach similar to the following would be followed. (Note that modeled sequence MLO-OK4 includes the 16 bounded success sequences. which are also labeled MLO-OK4, between sequences 88 and 118 on Figure 2.)

The baseline success criterion for MLO-OK4 is based on conditions represented by the bottom-most, and therefore, most restrictive, success path (sequence 114), i.e., no CMT, ADS case ADQ, I accumulator,1 IRWST injection line, and I recirculation line. Several options exist regarding specification of alternative hardware configurations for success with T/H uncertainty included. These include (and are not limited to) the following: requiring 2 1RWST injection paths ra:her than 1 (but retaining the existing criteria for ADS, accumulators, and recirculation), which would result in paths 106,107,113, and 114 being failure paths: requiring 2 accumulators rather than 1 accumulato; (but retaining the existing criteria for ADS, IRWST, and recirculation), which would result in paths 110, Ill,113, and 114 being failure paths; or requiring more ADS valves than specified in case ADQ (but retaining the epsting criteria for accumulators, IRWST, and rucirculation), which would keep paths 103,104,106,107,110, Ill,113, and 114 as success paths, but at a lower success frequency. Another possibility would be to credit containment isolation, which would increase injection performance. (Note that, on Figure 2, paths with successful containment isolation are not shown explicitly, since Figure 2 assumes failed isolation at the first branch point. Sequence 1 on Figure 2 in fact represents 117 success and failure sequences with success criteria generally less restrictive than those for sequences 2 through 118. Note also that paths with less T/H uncertainty are not shown, since Figure 2 for this assessment, assumes worst T/H uncertaipties.)

Any of these options results in the availability of additional injection flow with increased likelihood of success (which would be confirmed with additional analysis). The options could be tried one at a time, working up from the bottom of the tree, until success was achieved.

However, it is expected the selection can be made based on knowledge of plant performance with various configurations for the particular initiating event.

Once it is determined which bounded sequences are affected by modifying the hardware Wesunghouse Electne Corporauon 7

Mav IWS

DRAFT AP600 Thermal / Hydraulic Uncertainty Assessment availability assumptions, an assessment of the significance of the impact on PRA core damage frequency can be made by examining the bounded sequences that changed from success to failure or for which the success frequency changed. For example, suppose that it were shown that changing from ADS critedon ADQ (which is 2 of 4 ADS Stage 4 valves) to 3 of 4 ADS Stage 4 valves resulted in success. Then if there is an effect on core damage frequency, it would be from the difference between the contribution of bounded sequences 93,94,97,100.

102,105,108,109,112.115,116,117, and 118 (i.e., the failure paths) with the 3 out of 4 i

ADS Stage 4 cdtedon and with the 2 out of 4 ADS Stage 4 criterion, for the conditions of most limiting T/H uncertainty. If the difference in the failure probabilities for both of these ADS criteria is sufficiently small, then it can be reasonably stated that the effect of this changc on the PRA results is not significant, and the analysis will have demonstrated that, for that particular modeled sequence and for its associated bounded sequences, the effect of TM uncertainty does not affect the PRA. If the assessment requires moving up through the bounded sequences so far that it cannot be reasonably stated that the effect of the change on the PRA results is not significant, then it may be necessary to revise that PRA success criterion and determine the impact of the change on the PRA, Correspondence to NRC Margins Approach The approach to assessing the impact of TM uncenainty on the PRA results as outlined in the preceding paragraphs generally follows the NRC Margins approach.

l A binning process will be used as described in this Task 2 deliverable, based on the existing l

PRA event sequences. The bounding sequence (i.e., the sequence for which the highest peak l

core temperature will occur) for each bin will be used; sequence frequency cutoffs will not be used. The binning will be finalized in conjunction with Task 5 and documented in Task 10 of the TM Uncertainty Assessment.

Identification of sources of uncertainty associated with the TM performance of passive systems will be done through Tasks 4,5,6, and 7 of the TM Uncertainty Assessment.

Identification of "large impact" variables is not anticipated, since all identified sources of TM uncertainty will be addmssed. A systematic process will be used to identify sources of i

uncertainty.

Analysis of available margin to core damage will be performed for the bounding sequence for each bin, in Tasks 9 and 10 of the TM Uncertainty Assessment. Bounding values will be used for the identified uncertainty vadables.

If, for any bin, the analyses fail to show margin te com damage, a different set of hardware success criteria will be selected for the bin (correspoNing to a different bounded sequence) such that margin to core damage can be shown. h shis can be shown to result in an insignificant change to the PRA results, then there is no impact of TM uncenainty on the success criterion for the bin. If there is a significant impact on sequence core damage frequency as a result of selecting a different bounded sequence to represent the bin, this will be addressed for the PRA.

Wesunghouse Electnc Corporauon 8

May 1995 g

n e-m g

-m-e

o DRAFT AP600 Thermal / Hydraulic Uncertainty Assessment Table 1 initial Set of Success criteda to be Considered for Bins for T/H Uncertainty Assessment initiating Event or Success Suc.

Success Criteria Bounded Category Sequence ID cess Success Seq.#

Sequences Medium LOCA MLO-OK2 1

1 of 2 CMTs, ADS case hEO4K1 CMT Line Break CMT-OK2 ADM, I of 2 IRWST CMT-OK1 51 Line Break SIL-OKI injecuon hnes, I of 4 IRWST recirculauon hnes MLO-OK4 2

ADS case ADQ, hEO-OK3 CMT-OK4 1 of 2 accumulators, I of 2 CMT-OK3 SIL-OK2 IRWST injection hoes,1 of 4 IRWST rectreulation lines Intennedsate LOCA S10-OK2 3

RCP tnp, I of 2 CMTs, N10-OKI (including consequenual ADS case ADM,1 of 2 LOCA due to stuck open IRWST injecuon hoes, pressunzer safety valve) 1 of 4 IRWST recirculation unes S10-OK4 4

ADS case ADQ.

N10-OK3 1 of 2 accumulators, I of 2 IRWST injecuon hnes, I of 4 IRWST rectreulation lines

)

l l

Westinghouse Electnc Corporauon 9

May 1995

DRAFT AP600 Thermal / Hydraulic Uncertainty Assessment I Table 1 initial Set of Success criteria to be Considered for Bins for T/H Uncertainty Assessment Initiating Esent or Success Suc-Success Criteria Bounded Category Sequence ID cess Success Seq.#

Sequences Small LOCA SLO-OK2 5

RCP tnp, I of 2 CMTs,1 SLO-OKI (includmg consequennal SGR-OK4 PRHR beat exchanger, SGR-OK3 small LOCAs due to RCS ACS case ADS, Leak. SGTR. PRHR tube 1 of 2 IRWST injection rupture, transients)

Imes, I of 4 IRWST recirculauon hnes SLO-OK2' 6

RCP tnp,1 of 2 CMTs. O SLO-OKl' SGR-OK2' PRHR beat exchanger, SGR-OKl*

ADS case ADA.

1 of 2 IRWST injecuon lines. I of 4 IRWST recirculabon hnes SLO-OK4 7

0 or i PRHR beat SLO-OK3 SGR-OK6 exchanger, ADS case SGR-OK5 ADT. I of 2 accumulators.

I of 2 IRWST injection lines, I of 4 IRWST recirculabon hnes WesunFbouse Electnc Corporauon 10 hty!W5

DRAFT AP600 Thermal / Hydraulic Uncertainty Assessment Table 1 initial Set of Success criteria to be Considered for Bins for T/H Uncertainty Assessrnent initiating Esent or Success Suc.

Success Criteria Bounded Category Sequence ID cess Success Seq.#

Sequences Steam Generator Tube SGR-OK2 8

Mam steam flow isofation, Rupture steam generator overfill protecuan, RCP tnp, I of 2 CMTs,1 PRHR beat exchanger Transients with failure of TRA-OK5 8

RCP cnp I of 2 CMTs, O TRA-OK4 decay beat removal PRHR beat exchanger, ADS case ADA, 1 of 2 IRWST injecuan lines, I of 4 IRWST rectreulanon lmes j

TRA-OK7 9

0 PRHR beat exchanger, TRA-OK6 ADS case ADT, I of 2 accumulators, I of 2 IRWST injecuon lines,1 of 4 IRWST recirculadon hnes

]

Loss of Offsite Power LSP-OK5 10 1 of 2 CMTs,0 PRHR LSP-OK4 beat excbanger, ADS case

ADAL, 1 of 2 IRWST injecdon lines, I of 4 IRWST rectreuladon hnes LSP-OK7 11 0 PRHR beat excbanger, LSP-OK6 ADS case ADT, I of 2 accumulators, I of 2 IRWST injecuon lmes, I of 4 IRWST rectreuladon bnes l

'Nesungbouse Electnc Corporauon 11 May IWS

DRAFT a

AP600 Thermal / Hydraulic Uncertainty Assesstnent j

Table 1 Initial Set of Success criteria to be Considered for Bins for T/H Uncertainty Assessment Initiating Esent or Success Suc-Success Criteria Bounded Category Sequence ID cess Success Seq 3 Sequences Stauon Blackxt SBO-OK2 12 1 of 2 CMTs, O PRRR beat exchanger, ADS case

ADAB, I of 2 IRWST injection lines, I of 4 IRWST recirculation lines SBO-OK3 13 0 PRHR beat exchanger.

ADS case ADT,1 of 2 accumulators, 1 of 2 IRWST injection lines, I of 4 IRWST recirculauon hnes Mam Steamhne Break SLB-OK6 14 RCP tnp. I of 2 CMTs, O SLB-OKS PRHR beat exchanger, ADS case ADA, 1 of 2 IRWST injecuon l

lines,1 of 4 IRWST rectreulation lines SLB-OK4 15 ADS case ADT, SLB-OK3 SLB-OK8 I of 2 accumulators, I of 2 SLB-OK2 IRWST injection hnes,1 SLB-OKI of 4 IRWST rectreutation SLB-OK7 lines Anucipated Transients ATW-OK3 16 RCP tnp,1 of 2 CMTs, ATW-OK2 Witbout Scram (with ATW-OK6 ADS case ADW, I of 2 ATW-OKS failure of boration or stuck IRWST injecuon hnes, open pressunzer safety 1 of 4 IRWST valve decay beat removal) recirculauon hoes 1

Wesunghouse Electne Corporauon 12 hbyIWS

o DRAFT f

Fegure 1 Medum LOCA Event Tree from AP600 PRA Page 1 of 1 MLOCA CMT NRHR ADS.F ACC 1RWST RECIR 1

MLO-OK1 2 MLO-OK2 IOL llW2AB 4 3BE RNR ADM 5 30 l

6 MLO-OK3 lAC2AB 3BA f fBL CM2L l RECIRC iilW2AB 10 3BE RNR AC2AB 11 3BR ADQ 12 30 i

i f

I vs.

w

b e

DRAFT 1

Fig.1 Medium LOCA Event Tree from AP600 PRA List of top events Event Description MLOCA MEDtVM LOCA EVENT OCCURS CMT RCP TRIP. AND ONE OR BOTH CORE MAKEUP TANKS INJECT NRHR 1 OR 2 TRAINS OF NORMAL RHR IN INJECTION MODE ADS-F FULL DEPRESSURIZATION WITH ADS (CASE ADM/ ADO OR BETTER)

ACC 1 OR 2 ACCUMULATORS INJECT IRWST 1 OR 2 UNES OF IRWST INJECTION REClR 1 OR 2 UNES OF WATER RECIRCULATION i

i l

k k

i I

i 1

i l

DRAFT Fig.2 Exemple AP600 Modum LOCA Event Tree with Passive Paths Expanded Page 1 of 2 MLOCA Cl CMT2 CMT1 ADS-A ADS-F ACC2 ACC1 IRWS2 IRWS1 RECR2 RECR1

, (nna 2'-us')

Cearmwww 2 MLO-OK2 isdATO

'1 M*8 3 MLO-OK2 1 M*5 l'

4 3BL 5 MLO-OK2 2 AEW4-11 FM 6 MLO-OK2 IP I'

7 3BL o

8 3BE 9 MLO-OK2

'E 10 MLO-OK2 AMM l'

11 3BL i

7ADM i

12 MLO-OK2 144 4 11 13 MLO-OK2 IN*

l' 14 3BL o

15 3BE 16 MLO-OK2

't 17 MLO-OK2 1 Paws lI 18 3BL 0

19 MLO-OK2 20 MLO-OK2 21 3BL o

22 3BE 23 MLO-OK2 24 MLO-OK2 E **5 I'

25 3BL

2. CMTs 26 MLO-OK2 1 ^"*4-

'1 27 MLO-OK2 8

l8 28 3BL 0

29 3BE 30 MLO-OK2

'E 31 MLO-CK2 1 Mm5 l4 32 3BL ADM 33 MLCr OK2 I MM-11 34 MLO CK2 3

l' 35 3BL 36 3BE 37 MLO-Of2

'l 38 MLO-OK2 2W la 39 3BL 0

e 40 MLO-OK2

'1 41 MLO-OK2 I

l8 42 3BL o

43 3BE 4 APM 44 30 45 MLO-OK2

'I 46 MLO-OK2 l

1 l'

47 3BL I

48 MLO-OK2 14M-

't 49 MLO-OK2

)

l' 50 3BL o

51 3BE 52 MLO-OK2

'L 53 MLO-OK2 l

1 l'

54 3BL PApM 55 MLO-OK2 i

C rm=*==tf I AC'*4.

11 56 MLO-OK2

)

NcT '50N I

l' 57 3BL a

58 3BE 59 MLO-OK2

'l 60 MLO-OK2 2

l#

61 3BL 62 MLO-OK2 Ik 63 MLO-OK2 l

I' l[

64 3BL i

o DRAFT Fig.2 Exemple AP600 Medium LOCA Event Tree with Passive Deths Expanded Page 2 of 2 MLOCA Cl CMT2 CMT1 ADS A ADS-F ACC2 ACC1 IRWS2 IRWS1 RECR2 RECR1 le 65 3BE 66 MLO-OK2 67 MLO-OK2 fcgy 1 4 #85 l'

68 3BL 69 MLO-OK2 2 Accua4 70 MLO-OK2 IM '

l8 71 3BL o

72 3BE 73 MLO-OK2 74 MLO-OK2 2-l' 75 3BL 4pg 76 MLO-OK2 l 48CW4 77 MLO-OK2 l'

78 3BL o

79 3BE 80 MLO-OK2 81 MLO-OK2 2.

l#

82 3BL 0

a,3 MLO.OK2 84 MLO-OK2 I

I' 85 3SL a

86 3SE 4ADH 87 30 88 MLO-OK4 89 MLO-OK4 L

l8 90 3BL 91 MLO-OK4 1 4CCadM.

't 92 MLO-OK4

/

l8 93 3BL o

94 3BE 95 MLO-OK4 96 MLO-OK4

> 4 PCs l8 97 3BL

^

98 MLO-OK4 leccu+

99 MLO-OK4 i

100 3BL o

101 3BE o

102 3BR 103 MLO-OK4 81 104 MLO-OK4 O uf g

l*

105 3BL 106 MLO-OK4 L Accv.4 l&

107 MLO-OK4 I

108 3BL 109 3BE 110 MLO-OK4 I&

111 MLO-OK4 AN 1

l8 112 3BL 113 MLO-OK4 l ACC A 114 MLO-OK4

/

3 l8 115 3BL o

116 3BE o

117 3BR 4 ADQ 118 3D

DRAFT Fig.2 Exemp4 AP600 Medium LOCA Even: Tree wnh Possive Pethe Expanded List of top events

)

1 Event Desenption MLOCA MEDIUM LOCA EVENT OCCURS Cl CONTAINMENT ISOLATION CMT2 RCP TRIP AND SOTH CORE MAKEUP TANKS CMT1 RCP TRIP AND ONE CORE MAKEUP TANK ADS A FULL DEPRESSURIZATION, ALL ADS VALVES OPERATE i

ADS-F FULL DEPRESSURIZATION BY ADS ACC2 BOTH ACCUMULATORS ACC1 ONE ACCUMULATOR IRWS2 GRAVITY INJECTION FROM 6%TH UNES IRWS1 GRAVITY INJECTION FROM ONE UNE RECR2 TWO PATHS OF WATER GIRCULATION TO RPV RECR1 ONE PATH OF WATER. HECIRCULATION TO RPV W

l f

i n

i 5

I 9

h h

l 4

e

)

DRAFT l

AP600 ThermaFHydraulic Uncertainty Assessment References 1.

NTD-NRC-95-4455, DCPINRC0320, Proposed Schedule and Milestones for Resolution of AP600 Passive System Reliability, N.J. Liparulo (Westinghouse) to T.R. Quay (NRC),

May 8.1995 2.

Passive System Performance.aeliability Analysis in Advanced Reactor Designs --- Approach and Implementation Issues, NRC Probabilistic Safety Assessment Branch presentation at March 30,1995 NRC/ Westinghouse meeting on AP600 Thermal-Hydraulic, Performance.

3.

AP600 Probabilistic Risk Assessment, Revision 3. February 28,1995.

4 I

I I

s l

Wesungbouse Electnc Corporauon 18 Mrv IWS

_--