ML20133B880

Resolution of T/H Uncertainty Issues for AP600 Passive Sys Reliability,Dec 1996 Draft (Includes PRA Expanded Event Trees & Definition of Low-Margin,Risk-Significant Cases)

ML20133B880
Person / Time
Site:	05200003
Issue date:	12/31/1996
From:	WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP.
To:
Shared Package
ML20133B864	List: ML20133B880 NSD-NRC-97-4928, Forwards Draft Rept, Resolution of T/H Uncertainty Issues for AP600 Passive Sys Reliability
References
NUDOCS 9701060251
Download: ML20133B880 (76)

Text

r

.c Resolution of T/H Uncertainty Issues for AP600 Passive System Reliability DECEMBER 1996 DRAFT (Includes PRA Expanded Event Trees and Definition of Low-Margin, Risk-Significant Cases) 9701060251 970102 PDR ADOCK 05200003 PDR A

I

_ -.. _ =

P l

Executive Summary of Resolution of T/H Uncertainty Issues, December 1996 Status i

The final effort to resolve passive system reliability issues is to evaluate the potential impact of thermal / hydraulic uncertainties on the PRA. The central question is whether the consideration of uncenainty in success criteria analyses would significantly affect the conclusions of the; PRA. The T/H uncertainty resolution process identifies a set oflow-margin, risk-significant accident scenarios, and shows.

acceptable T/H performance when the uncertainties are bounded.

l De first step of the T/H uncertainty resolution process is to expand the Focused PRA event trees and to quantify the frequency of the success paths. Expanding the event trees is necessary to differentiate between scenarios that are grouped together in the PRA. There are ten expanded event trees developed for T/H uncertainty resolution that encompass all the success paths that require ADS actuation for successful core cooling. He frequency of the success paths are quantified.

The next step is to categorize all the success paths based on similarities in the accident progressions.

Here are 20 categories, which are separated into two types: OK categories and UC categories. The OK categories are ones that are similar enough to design basis that it can be explained why they are not " low margin" scenarios, and they are not further considered within the T/H uncenainty resolution process. The UC categories are considered " low margin," and the frequency of each UC category is further assessed to determine whether it is risk-significant.

The categorization process considers the accident progression through two phases of water injection:

1) short term, when the accumulators and CMTs provide make-up inventory, and 2) 1RWST gravity injection. The final phase of water injection -- long-term sump recirculation - will be treated separately from the OK and UC categorization. The plan to address long-term recirculation for the PRA is outlined, but has not been implemented, pending further discussions between Westinghouse and the NRC.

Each UC category is assessed to determine whether it is risk-significant. This process considers the increase to the Focused PRA Core Damage Frequency (CDF) and Large Release Frequency (LRF) if the success path actually leads to core damage. Risk significance is defined as increasing the Focused PRA j

CDF or LRF by at least 1% if the UC category were counted as core damage. This process identifies five risk-significant categories that are summarized in the following table. More inforn tion on the accident scenarios represented by these cate;ories it in Section 7.0 of the attached report. He ;mpact of using the Focused PRA rather than the Base 3ne PRA as the comparison is also discussed within the repon, but does not alter which categories are designated as risk-significant.

l t

l l

Risk-Significant, Low-Margin Categories (In Order of Risk Significance)

Category Initiating Events Defining Equipment If counted as core damage, Conditions increase to Focused PRA l

ACDF ALRF UC4 LLOCA 1 Accumulator 1.1E-6 6.9E-8 l

UC5 NLOCA 0 Accumulators 7.2E-7 7.6E-8 DVI Line Break SLOCA SGTR Transients UC6 All 2 stage 4 ADS 3.4E-7 7.5E-8 Containment Isolated UCI NLOCA 0 CMTs 1.4E-7 8.2E-9 DVI Line Brek UC2B MLOCA 0 CMTs 1.2E-7 7.5E-9 CMT Line Break From these risk-significant categories, a set of cases is defined for T/H analyses with uncertainties to complete the T/H uncertainty resolution process. A representative case for each category is defined by examining the success paths that dominate the frequency of that category. Table 8-3 within the attached document identifies the cases that will be analyzed. The determination of the limiting break sizes to be analyzed will be made after the MAAP4/NOTRUMP benchmarking is completed.

The final steps in the T/H uncertainty resolution process, that are not completed are:

Identify the risk-significant long-term recirculation cases Perform T/II analyses with uncertainties on low-margin risk-significant cases from the UC categorization and on risk-significant long-term recirculation cases Assess T/H study results on the PRA

=

1 1

t Risk-Significant, Low-Margin Categories (In Order of Risk Significance)

Category Initiating Events Defining Equipment If counted as core damage, Conditions increase to Focused PRA ACDF ALRF UC4 LLOCA 1 Accumulator 1.1E-6 6.9E-8 UCS NLOCA 0 Accumulators 7.2E-7 7.6E-8 DVI Line Break SLOCA SGTR Transients UC6 All 2 stage 4 ADS 3.4E-7 7.5E-8 Containment Isolated j

i UCI NLOCA 0 CMTs 1.4E-7 8.2E-9 j

DVI Line Break t

i UC2B MLOCA 0 CMTs 1.2E-7 7.5E-9 l

CMT Line Break From these risk 9nificant categories, a set of cases is defined for T/H analyses with uncertainties to l

l complete the T/H uncertainty resolution process. A representative case for each cateBory is defined by l

examining the success paths that dominate the frequency of that category. Table 8-3 with! 2.e attached document identifies the cases that will be analyzed. De determination of the limiting break sizes to be analyzed will be made after the MAAP4/NOTRUMP benchmarking is completed.

De final steps in the T/H uncertainty resolution process, that are not completed are:

Identify the risk-significant long-term recirculation cases Perform T/H analyses with uncertainties on low-margin risk-significant cases from the UC j

l categorization and on risk-significant long-term recirculation cases

]

l Assess T/H study results on the PRA l

__.m.

. ~. _ _... _. _..

1 i

b Table of Contents i

W l

1.0

- INTRODUCTION.........................................

1 5

2.0 DEFINITION OF T/H UNCERTAINTY..........................

1 3.0 RESOLUTION PROCESS....................................

2 4.0 EXPANDED EVENT TREES.................................

3 l

\\

i l

l 4.1 Expanded PRA Event Tree Methodology....................

3 4.2 Scope of Expanded Event Trees..........................

7 l

4.3 Impact of Focused PRA vs. Baseline PRA...................

9 l'

4.4 Results of Expanded Event Trees and Frequency Quantification.... 10 5.0 CATEGORIZATION OF SUCCESS SCENARIOS.................. 21 5.1 CMT and Accumulator Injection.......................... 22 5.2 IRWST Gravity Injection............................... 23 5.3 Long-term Recirculation................................ 24 6.0 OK CATEGORIES SIMILAR TO DESIGN B ASIS.................. 26 7.0 UC CATEGORIES OF LOW-MARGIN ACCIDENT SCENARIOS....... 46 8.0 IDENTIFICATION OF LOW-MARGIN, RISK-SIGNIFICANT SCENARIOS 64 8.1 Comparison Method to Focused PRA CDF and LRF,........... 64 8.2 Risk Significant Categories '.............................. 65 8.3 Representative Cases to Address Low-Margin, Risk-Significant Scenarios.......................,...... 68 9.0 IDENTIFICATION OF RISK-SIGNIFICANT LONG-TERM RECIRCULATION CASES...................................

10.0 T/H ANALYSES OF LOW-MARGIN, RISK-SIGNIFICANT SCENARIOS.

10.1 Assumptions for T/H Uncertainty Analyses..................

10.2 NOTRUMP Results...................................

10.3 WCOBRA/ TRAC Results

11.0 ASSESSMENT

OF T/H UNCERTAINTY RESULTS ON PRA..........

12.0 CONCLUSIO N...........................................

13.0 REFERENCES

i n

-.. ~..,

i l

i

1.0 INTRODUCTION

l The AP600 design incorporates passive engineered safety features that perform safety-related functions to mitigate accidents and to establish safe shutdown conditions following an event. An extensive range of activities have been completed as part of the AP600 design cenification process to provide confidence in the design capabilities and reliability of the safety-related, passive systems and components. An overview of these activities, and references to the appropriate documentation, is provided in Ref.1. One of the remaining effons to resolve passive system reliability issues, as identified in Ref.1, is to evaluate the potential impact of thermal / hydraulic uncertainties on the PRA.

Thermal / hydraulic analyses have been performed to support multiple-failure success criteria definitions in the AP600 PRA. To define the cases for analyses, the PRA event trees were reviewed and success paths (i.e., paths that do not lead to core damage) were grouped based on similarities. Each group consists of the same functioning equipment and a range of break sizes and location. Within each group, bounding cases were identified. Bounding cases were chosen to be the most limiting break size, location and set of equipment to bound the group of cases.

Analyses of the bounding cases were perfonned with nominal assumptions, rather than conservatisms that I

are typical of design basis safety analyses. The purpose of using nominal conditions was to preserve plant behavior as it is most likely to occur, so that PRA insights may be gained on the risk importance of different systems. An issue has been raised on whether the consideration of uncertainty in the analyses would significantly affect the conclusions of the PRA. 'Ihis issue is termed "T/H uncertainty resolution" and is the subject of this document. It is the final component to closing the passive system reliability issues for AP600.

2.0 DEFINITION OF THERMAL-HYDRAULIC (TSI) UNCERTAINTY The term "T&I uncertainty" is used in relationship to predicting the behavior of passive systems in AP600.

Because of the passive nature of the safety-related systems in AP600 and the reliance on small AP's, the concern is that uncertainties in predicting the small changes in the system conditions could lead to different conclusions on the success of core cooling. The small changes in system conditions could be due to different accident conditions than modelled, or uncertainty in analytical models. Specific sources of T/H uncertainty that have been identified as potential concerns are:

initial and boundary conditions, code uncenainty (based on testing and scaling uncenainties),

e user-selected inputs and modeling methods.

e l

l If the success criteria are bounding,it must be shown that the consideration of T/H uncertainties does not significantly impact the PRA results. Funhermore, because the concern is passive system reliability, the Focused PRA (that does not include active systems) is the standard for comparison and detemination of cayw,6au_.a.#n Page 1 December 30.1996 l

.m..

i

+

l 9

4 i

impact. Use of the Focused PRA ensures that active systems will not camouflage the importance of passive systems, or the uncertainty in predicting their performance. Section 4.3 provides more information l

on the impact of using the Focused PRA instead of the Baseline PRA as the comparison basis.

As described in the following sections, the T/H uncertainty resolution process does not quantify the l

- sources of uncertainty, nor is it solely a T/H analysis exercise. Rather, the T/H uncertainty resolution process identifies a set of low margin, risk significant accident scenarios, and shows acceptable T/H i

performance when the uncertainties are bounded.

I l~

3.0 RESOLUTION PROCESS

'i j

l The T/H uncertainty resolution process integrates information that can be obtained from the PRA and from j

T/H analyses. PRA methods can direct attention to accident scennios ths are most probable. PRA event trees show a breakdown of the possible equipment successes and failures, and provide a systematic method 3

for assessing the accident configuration. 'The methods used to perform T/H analyses tend to direct i

attention to bounding accident scenarios that most greatly challenge core cooling. However, the T/H l

challenging scenarios may or may not have risk significance to the plant. 'Ihe T/H uncertainty resolution process identifies the accident scenarios for further study that are both significantly high in frequency and 3

consequences and which challenge core cooling. This process concentrates efforts and resources to the most important cases, and is an implementation of risk-informed decision making.

l

'Ihe T/H uncertainty resolution process is briefly outlined below. 'Ihe details of the methods and results are in the following sections of this report.

1 l.

Expand and quantify PRA event trees to further refine the equipment that is available in the accident scenarios that result in successful core cooling. (Section 4.0) 2.

Assign success categories so that all accident scenarios can be systematically. discussed.

(Sections 5.0,6.0 and 7.0) 3.

. Assess category frequency / consequence to determine risk significance oflow-margin scenarios.

(Section 8.0) 4.

Define low margin, risk significant cases for further T/H study. (Sections 8.3 and 9.0) 5.

Define assumptions to bound uncertainties in T/H analyses. (Section 10.1) 6.

Perform T/H analyses. (Sections 10.2 and 10.3) 7.

Assess impact of T/H study results on PRA. (Section 11.0)

I l

e:w.:

._.r..,a, Page 2 December 30,1996

4.0 EXPANDED EVENT TREES 4.1 Expanded PRA Event Tree Methodology l

Re first step of the T/H uncertainty resolution process is to expand the Focused PRA event trees and to quantify the frequency of the success paths. Success paths are not normally quantified in a PRA, since core damage is the focus. The purpose of quantifying the frequency of success paths for T/H uncertainty resolution is to gain perspective on the relative probability of specific success scenarios. This information will ultimately be used to define risk significant scenarios that could be impacted by T/H uncertainty.

l

" Expanding" the event trees is necessary to differentiate between scenarios that are grouped together in the PRA. A single success path in the AP600 PRA represents many combinations of equipment failures and successes. As an example, Figure 4-1 shows the MLOCA event tree as it appears in the Focused PRA. Table 4-1 lists the functioning equipment that are included within the top success path on the MLOCA event tree. Table 41 also identifies the equipment assumptions that are made in the l

corresponding acc! dent analysis that supports the success path.

As shown in Table 4-1, the equipment configuration that is used in the success analysis to justify a specific success path is the most pessimistic set of functioning equipment for that path. Minimum functioning equipment leads to the most limiting accident progression. Even if the bounding scenario analysis shows core uncovery, there are many other accident scenarios (or sets of functioning equipment) l represented by the same success path that may not result in core uncovery. Therefore, the success paths l

on the event trees need to be refined or expanded to show the various equipment success combinations so that differences in accident progressions can be assessed.

There are options of how to expand the success paths on an event tree. Thcre are four key elements to the method that was developed to perform the expansion.

1.

There are many top level events that could be used to ask questions and further refine the success paths. Table 4-2 summarizes the options that were considered, and why they were or were not selected.

2.

The expansion of the event tree does not redefine the definition of success. All success paths on the expanded event tree are represented within an existing success path in the Focused PRA. All core damage paths on the expanded event tree are core damage paths in the Focused PRA.

l Fundamental to the expansion is the necessity to ask additional equipment questions that are not l

explicitly modelled in the PRA. However, each question only differentiates between distinct successful accident progressions that are grouped within a success path in the PRA. The additional questions can better represent reality, but they cannot cause success definitions to become either more or less conservative.

cwpw60cw._ c.sp Page 3 December 30.1996

p.

3.

Success paths containing more than 3 system failures are not further ernanded in the present models. In general, three failures are deemed to decrease the frequency of a path sufficiently.

Imposing the 3 failure limit also helps to restrict the event tree expansion to a manageable size.

The net effect of this restriction is that paths toward the top of the expanded tree are broken into more detail than those toward the bottom.

An alternative approach is to expand an event tree until the success paths reach a cut-off frequency. However, this would require quantification results to be integrated with the construction of the event tree. The 3 system failure expansion method was chosen because it is -

a systematic, understandable method that allows event tree development independent of the quantification results.

4.

Top events were arranged in an order to minimize the number of paths. This changed the location of the injection and recirculation line question from the last top event in the Baseline and Focused PRA event trees to the first top event in the expanded event trees.

e ;_,~.:. :.,:.n Page 4 December 30,1996

i l

f Figure 4-1 MLOCA Event Tree in Focused PRA i

1 or 2 1 OK I

1 or 2 RECIRC 2,3 or 4 IRNST 0

)

2 CD 1 or 2 ADS-4 0

CMT 0 or 1 4 CD 1 or 2 l

5 oK MLOCA 1 or 2 RECIRC 1 or 2 IRWST 0

6 CD 2,3 or 4 Acc 0

7 CD 0

ADS-4 0

)

B CD 0 or 1 9 CD OK = Successful Core Cooling CD = Core Damage l

r I

t Table 4-1 i

Comparison of Equipment on Event Tree Success Path to Equipment Assumptions in Supporting Analysis Equipment That May Function for Bounding Scenario Used for Success Path 1 on MLOCA Event Tree in PRA Accident Analysis Focused PRA 1 or 2 CMTs 1CMT 0,1 or 2 stage 1 ADS

• O stage 1 ADS l

0, I or 2 stage 2 ADS

• O stage 2 ADS l

0,1 or 2 stage 3 ADS

• 0 stage 3 ADS 2,3 or 4 stage 4 ADS 2 stage 4 ADS 0,1 or 2 accumulators 0 accumulators l

1 or 2 IRWST injection lines 1 IRWST line

> 1 recirculation line

> 1 recirculation line Success or failure of containment isolation

• Failure of complete containment isolation t

l Not broken out by a top event question, but implicit within scenano possibilities.

j

)

cs w + m Page 5 December 30,1996

f 8

Table 4-2 Options for Expanding Event Tree Success Paths Option Used?

Reason Break size No Break size and location are already used to define different initiating events. Although within an initiating event there remains some variability in the plant response depending on the Break location size and location of the break, there was no added benefit to further refinement.

Number of CMTs Yes Whether there is I or 2 CMTs does not make a significant difference in the course of the accident progression. However, the CMTs are highly reliable, and make an important contribution to the refinement of the frequency of a given accident scenario. 'Itat is, for a given scenario, the most likely condition is both CMTs available.

Number of stage 1 ADS lines No Stage 1 ADS lines are small, and do not significantly impact the course of the accident progression.

i Number of stage 2/3 ADS Yes Stage 2 and 3 ADS lines can impact the ability to achieve lines IRWST gravity injection.

Number of stage 4 ADS lines Yes Stage 4 ADS lines can impact the ability to achieve IRWST gravity injection.

Number of accumulators Yes The number of accumulators is important to the core uncovery issues discussed in Section 3.1.

Number of IRWST lines No The ability to achieve IRWST gravity injection and long-term recirculation is most dependent on the number of open ADS lines and whether the conatinment is isolated. The number of Number of recirculation lines lines open, as long as there is a pathway for injection, is not as crucial an element to successful core cooling.

Whether containment is fully Yes The containment back pressure that occurs when the isolated containment is isolated can impact the ability to achieve IRWST gravity injection. Also, containment isolation impacts the large release frequency calculation if the accident scenario is counted as core damage.

I c w e

._c., a p Page 6 December 30,1996

4.2 Scope of Expanded Event Trees I

There are ten expanded event trees developed for T/H uncertainty resolution. They funher define the equipment available for the majority of the success paths modelled in the Focused PRA. The relationship between the expanded event trees and the Focused PRA event trees is shown in Table 4-3 i

The success paths that are not included on the expanded event trees are ones in which successful core cooling can be achieved without ADS actuation. An example of this is a loss of main feedwater event, which is successful without ADS if the PRHR functions. The PRHR is the safety-related method of removing decay heat, and leads to successful core cooling as demonstrated in Chapter 15 of the SSAR.

Primary coolant is not lost, and there is no need for inventory make-up from either the CMTs, j

accumulators, IRWST gravity injection or long-term recirculation. In addition to the PRHR, decay heat j

removal can occur from other active, nonsafety systems. These options are modelled in the Baseline PRA, but are conservatively neglected in -

Focused PRA.

Therefore, the success paths that are expanded for T/H uncertainty resolution are loss of coolant accidents.

The loss of coolant can either be the initiating event, or can be the result of a loss of heat sink wident.

l The loss of coolant is severe enough to require inventory make-up, first from the CMTs and accumulators, I

then from IRWST gravity injection, and finally from long-term recirculation.

The quantification of the success path frequency on an event tree includes the consideration of any events that transition to that event tree. For example, if a pressurizer safety valve sticks open in a transient event j

(e.g., loss of feedwater), the accident progression transitions to the NLOCA event tree (Figure 4-6). The NLOCA success path quantification accounts for the transient events with loss of PRHR and a stuck open pressurizer safety valve. This is just an example of the consequential effects that have been included in i

the expanded event tree quantification.

i s

l enweap60cc._r.wwp Page 7 December $0,1996 l

e c

Table 4-3 Correlation of Expanded Event Trees to Focused PRA Event Trees Initiating Event Break Size Expanded Event Tree Event Trees from Focused PRA Diameter Designator Large LOCA

> 9.0" lloca LLOCA Medium LOCA 6.0" - 9.0" mioca MLOCA CMT Line Break 50" cmtib CMTLB 8

i 5

0" silb SI-LB DVI Line Break 4

Intermediate 2.0" - 6.0" nloca h10CA LOCA Small LOCA with

< 2.0" slocaw SLOCA

• PRHR RCS Leak

• Small LOCA

< 2.0" slocwo SLOCA

• without PRHR Inventory loss can RCS Leak

• also occur through PRHR Tube Rupture pressurizer safety valve SGTRs with Itube sgtrw SGTR m PRHR that Require ADS SGTRs without Itube sgtrwo SGTR*

PRHR that Require ADS Transients that Inventory loss tran Loss of MFW to both SGs (*)

Require ADS through Loss of Offsite Power (')

pressurizer safety Loss of Compressed Air (*)

valves Loss of CCW/SWS (')

Loss of Condenser (*)

Loss of MFW to 1 SG (')

Loss of Reactor Coolant Flow (')

Power Excursion Event Tree (')

SLB Downstream e MSIVs (')

r SLB Upstream of MSIVs (

Stuck-Open Secondary Side SV (')

Transients with HEW (*)

ATWS*

Notes:

1)

Portion of tree with PRHR 2)

Portion of tree without PRHR 3) includes success of PRHR and success of pressurizer safety valves 4)

Includes failure of PRHR cwc__r.#-p Page 8 December 30, 1995 u

4.3 Impact of Focused PRA vs. Baseline PRA The Focused PRA results are the point of comparison for the T/H uncertainty resolution process. The Focused PRA models only the passive, safety-related systems in the AP600 plant. Active, nonsafety systems are not credited in the mitigation of the accident. For this reason, the Focused PRA most clearly demonstrates the irnportance of passive systems, and is the appropriate point of comparison for the T/H uncertainty issue related to passive system reliability.

The choice of the Focused PRA versus the Baseline PRA affects the frequency values that are quantified for the success paths. Because active systems are ignored in the Focused PRA, the passive-only accident I

progressions are often quantified with higher-than-realistic frequencies of occurrence. For example, most LOCA events lead to RCS inventory make-up from the IRWST. The IRWST water can be supplied from eith'.r a pumped system (RNS) or gravity draining of the IRWST. 'Ihe reliability of the RNS is such that it operates approximately 9 out of 10 times needed. Therefore, for a given success scenario with a frequency of IE-7/ year, the passive-only accident progression with IRWST gravity injection would occur approximately 1E-8/ year. However, in the Focused PRA, the IRWST gravity injection success path is the only option considered, and the frequency of this passive-only accident progression is over-estimated at IE-7/ year, j

i The above example illustrates the impact of crediting or not crediting the RNS, assuming that the scenario is one where the RCS pressure is low ecough for either RNS injection or IRWST gravity injection to work. However, if the RNS were credited, there are ad'itional possible success paths with fewer ADS d

lines open than required for IRWST gravity injection. Therefore, even more of the postulated accident progressions would end with the utilization of active systems; passive-only scenarios are much less frequent.

l l

So that the importance and uncertainties of the passive systems can be studied without being skewed by the contributions of the nonsafety active systems, the Focused PRA is chosen for the expanded event tree development and quantification. 'Ihe frequency of a success path that is calculated based on the Focused PRA assumptions cannot be compared to frequencies calculated based on the Baseline PRA conditions.

As illustrated above, the frequency can be an order of magnitude different. 'lhis becomes very important when the frequencies are compared to the core damage frequency and large release frequency to determine risk significance.

The above discussion has been based on the majority of the LOCA accident progressions and event tree structures. However, when considering the impact of using the Focused PRA versus the Baseline PRA, there are some additional effects on some of the initiating events. If the Baseline PRA were used instead of the Focused PRA, the following two effects would be seen.

1)

Transients and SGTRs would decrease in relative importance to other events because there are multiple operator actions and nonsafety systems that can prevent core damage, and are credited e m pur.:._r.,~.m Page 9 December 30,1996

In the Baseline PRA. It is the failure of these other systems that leads to the LOCA-like accident progression that requires ADS for successful mitigation.

2)

Large LOCAs would increase in relative importance to other events. This is because all equipment credited in the Baseline PRA LLOCA event tree are safety systems, and are the same options considered in the Focused PRA. 'Ihe LLOCA quantification does not change, while the frequency of the passive-only success paths for other initiating events decreases in the Baseline PRA. Therefore, the LLOCA relative contribution is larger in the Baseline PRA than in the Focused PRA. 'Ihis aspect will be considered when the LLOCA success paths are examined for risk significance, and when the assessment of T/H uncertainty results on the PRA is made.

4.4 Results of Expanded Event Trees and Frequency Quantification

'Ihe expanded event trees are contained in Figures 4-2 through 4-11. The figures include not only the event tree structure, but quantification results and success path designators. The success path designators are discussed in Sections 5.0, 6.0 and 7.0.

. Ihe quantification method used to calculate the success path frequencies is the same method used to quantify the core damage paths in the Focused PRA. ADS cases are treated in more detail and SLOCA, SGTR and similar events are modeled with or without PRHR to capture the effects of this system.

c%,w.pe,p Page 10 p.e.

30, im l

,u,

,.... ~.

. + - _.

- ~.

.. ~..

..u a

96GURE 44 t

o EXPANDED LLOCA EVENT TREE

___e fWcesc

=

EvuoCA Less O

Caf7 Accutf ADB4 A0m3 8

1 OK1 aswa i it = s 2 CK7 17Em e

e=*

3 OK7 2.5E47 swa A

e 4

CK7 2.6E47 Stes I in e s 5 OK7 4.4E48 23 4 e

owt 4 CK7 S.6E-10 e

7 UC6 9.9E40 t

e i

1912.s 8

UCS 1.9E 08 9

UC7 32E49 e

to UC4 0 9E 07 9et Iswa I If e s 11 004 2JE47 e

ee*

it UC4 2.1E4e see e

13 UC4 11E4e towe s

i,Js te.

14 UO4 S.0E 10 t

l a

15 UC4 11E 10 ew*

to UC4 2.7E 11 e

17 CORE DAMAGE a

is OK7 1.sE47 tswa I~

1**

ges 19 OK7 44E48 e

ewi 20 CK7 3.6E 10 s er e e

21 CK7 8.7E 10 1:e4 I

le 'is 22 CK7 a.0E 11 a

28 UC4 SJE-11 ew*

34 UC7 4.6E 12

• ,t e

25 UC4 IJE49 e

I le'as as UC4 3JE-10

9. t e n was 3,s 27 UC4 SJE12 5

oy*

as UC4 3.3E 14 e

28 CORE DAMAGE o

30 CORE DAMAGE e

31 OKEA 1.8E47 tsee i in e s at OK58 3AE4s e

eet 38 OKBB 3.0E-10 swa e

34 UC8 8.18-10 e

I asee le t ts 35 UCS 4 9E 11 3

3 30 UCS SAE-11 ves

. 00GE 8'

UCS 1.9E-12 eee i

I. esGI as CORE DAMAGE

<et e

as UC4 1.15 48 e

I leik s 40 UC4 3 4E 10 1

e tsee 1

eet 41 UC4 17E.12 EvuoCA

. ssa 42 UC4 -

F.sE 15 e.,

i 1.eens 43 CORE DAMAGE e

44 CORE DAMAGE

• we a

46 OKSA 2 05-10 e

I le 12,s 48 OKBS 4AE 11 i

towe i

two 47 UC8 4.5E tJ f

$ 98GE 48 UCS 2 4E-15 set

[

'ef I.saut 40 CORE DAMAGE ec tsee 50 UC4 SJE13

. ssa

$1 UC4 0.08 00 ewt i

1. sag 1

$2 CORE DAMAGE e

63 COREDAMAGE e

64 CORE DAMAGE 96 CORE DAMAGE eno N EVER 1M

mm..

FK3UT,E 4 3 i

ei, EXPANDED MLCCA EYENT TREE

__.e SWERIIC Wwen.0CA uns G

Obf7 A00taf asse samL$

e 1 OK1 l

3J. 4 i

]

ts e t t OK3 41E48 4

I is e t 3 CK3 s sE47 see e

4 OKI 138.07 na e s I

lie t S

OK4 98840 33, 4 s

se q ee' S OK4 S SE te 7 UCS 99845 t

a l

18 UC8 1M4s l

9e*

9 COM DAaaMIE 7

l e

to OKI 14E46

• et 4

res si CKS S sE47 l

ie w,

in OK3

SM4, see 13 OK2 32E48 3

h s

i le.11A 14 OK4 7.95 10 f

s e

18 UC4 t.6E te ee*

16 CORE DAhiAGE 4

17 UC3 i tE4s

,4 I

hie UC8 18E49

{

th e e l

e I

is e t 19 UC3 178-11 be9 10 CORE 'md8 4

at OK 14847 3

.4 r

'ee hat OK3 SJE46 lee t 83 OKS 6.75-19 see e

to OK2 638 49 h

e i

te *JJ IS CKd 1 AE 10 e

a BB UC4 3.9E 11 i

oei 27 CORE DAldAGE tes e

as

'*P til4e 4

I to.t as 39 OK3 5 eE-16 I

4 ts e e L

19 e s 30 OK3 4 DE t2 le e i 31 ComE DAMAGE sJ e e 32 UC3 1.7E-11 e

i l

le e e 33 COM DAMAGE j

e H UC48 1 aE4e e

I

}

C35 Ur20 48840 uwe e

ar e s N UC30 4H-11 i

eee 37 CORE CAMAGE tot tsee 3B UC2A 14510

+

1 e

lee t 39 COM DAa4 Mat e

40 COME DAMAGE i

i e

41 OK5A 19E47 het OKBS 47E4s ee+

es OK38 43E 10 See e

44 OKs a 38-10 h

s I

in iza e UCe 15E.to e

a e6 Uce

$JE-11 ee+

47 ComE DAMAGE ter a

40 CKEA 10E49 4

Ihe Dies set-ie tae s e

ses 50 One 33513 ee.

St COM DAMAGE agvm tJ e e la UC8 1J811 e

n ree t 53 COM P-a8 ses e

54 OKEA 146 10 4

I l

le 12J

.C OK$8 71811 the e I

i l

t i

is o s to OK6 455-13

lee,

__.7 C= -

we.

i. UC3 m,:

ee, I

je e,

SS CORE DAMAGE e.s e e to UC2A 118-t1 I

• et I

ie ws 81 COM CAM 408 L

e et COM OAMAGE 03 COM DAalAM

]

i I

18 @ 98 am_rira m. - - _7

)

2

..m..

e._--.

...m_

=m._.

_. _ _.. _ _ ~

m.

.m 1.

mum t

e EXPANDED CMTLS EVENT TREE 4

_.._.a IISCEtC Ev4tittS M

O Cef7 Accuef 4086 Aa***

4 1

M1

[

4 ts w a i

it e s t OK3 8.3445 j

e 4

ew*

s OK3 2.1 E 07 sed

1. # 1 4

4 QQ 2.1547 i

s ts w a i 3

If, e S Oke 54648 ts e e s

ow*

4 OK4 4AE10 a

7 UC6 5.5E46 e

f I

lo.9 J.s 8 UC4 72E49 e**

8 CORE DAMAGE 4

to OK2 7SE47 two ts w a i It w s 11 OK3 1.9E 07 e

ew*

12 CK8 1.8E49 se4 1

4 13 OK2 1$$40 ts e e s

i leem.s 14 OK4 42E-10 4

t 18 UC4 7.7E-11 f

ew*

16 CORE DAMAGE a

17 UC3 SJE40 4

1 19*f9 14 UCS 1AE40 ts e4 e

fes

+9 UC3 1.3E 11

=

ew*

30 CORE DAMAGE a

it UC2B 4.0E40 was ts w a i It w s M UC2B t0E46 e

eet at UC2B 1.8 6-10 swa 4

as UCES 1 AE 10 ta w d i

le 12 s as UC2B 4.15-11 t

t as UC28 6.58 12 ewi 27 CORE DAMAGE

• • a a

as UC2A

&7E-10 4

l lesAs as UC2A 1.4E-10 Es w a fas 30 UC2A 1AE 12 o

ew*

31 COREDAMAGE e

N Cof:EDAMAGE a

35 OKSA 14E47 as w a i

It w s 36 OKIS 2.6E48 was e

oee 35 OK58 2.3E 10 see 4

as OKS 4AE 10 ts w e s

I le t2 s 37 UCS 5.3E 11 e

e as Uce 7AE-12 ew' 3e CORE DAMAGE

• wn a

40 ONSA 4.75-10 e

I le 5 2.s 41 ONM 1.0E-10 3.s e d tee 42 OK6 2.0$-12 i

I EvettfLS 43 COREDAMAGE ts w a 44 UC3 4.0E-12 e

I le wi 45 CORE DAMAGE 4

de UC2B 9.26 11 e

1 ND 1092.s 47 UCS 2.0E-11 ts e a e

fas 48 UCES 2.0E 13 ees de CORE DAMAGE

'wt is =

• N UC2A 7AE-13 I

e le ei 51 CORE DAMAGE o

la CORg nau ans

.3 CORE -

no 13

. -~=.

.. ~ _ ~....

i o

FGUAE 4-9 EXPONDEO SILS EVENT TREE F445 Ism yr p I carr carty I acCUu i apse i

apeg,s i

4 1

at l

l ts w o i Is e:

3 OKs 3.esas 4

a OKs asE47 see e

e OK3 2AE47

{

,,3 ts w a i is a s 5

CK4 8.3646 i

i i

to w e s

ew*

6 CK4 53E 10 e

7 UC4 4.4E40 s

I leees 8

UC8 S.0E-09 ew' 9 CORE DAMAGE a

to UCS 44E47 1

ts e e i le e s 11 UCS 1.1547 e

ows 12 UCS 1AE46 swa a

13 UC5 1.0E4e 1s w a s

I E, to UC8 1.5E 10 l

s is uCs 4.1E it ew*

16 CORECAMAGE d

17 Oka 7AE4s ves tswa i le e s it OKs 1.9E4s 4

ew' 10 CKS 1.75 10 Syd j

a 30 OKS 1.8E-10 I

=

as w e e

i le

• s.s 21 OK4 4.0E-11 e

e at UC6 7.15-12 ewe 33 CORE DAMAGE vas i e

34 UCB 825 10 e

l le 's s 35 UCS 736-11 tswa s

SS UC6 7.6E-14 e

i ew, 37 CORE DAMAGE I

e a

N UC1 14E48 a

]

1. e vis as UC1 4JE4e f.s e s 4s e

s 30 UC1 8.95-11 i

e ewe 31 CORE DAMAGE i

No l

e se CoREDAMAGE a

as OK6A 1J64r7 sswa 1 is e e se OKIB 1.0E4s e

eei 36 OsGS 3.75 10 ses 4

as OKS 1.8E 10 ts e s e

i le

• 2.8 37 000 42E 11 9

s at UCS 8.5E 12 ew*

as CORE DAMAGE s

gy464A a

40 UC5 S.15 10 4

I peers 41 UCS 12E 10 ts e e aes 42 UCS 12E 12 e

awe 48 CORE DAMAGE a

44 OSSA SAE 11 I

e i

j eso te e s.s 46 OK58 3.0E 11 ts w a e

sy3 d6 OKs 2.0E-13 i

vas e oei 47 CORE nm*M ts y 4 -

48 UC8 3.7E 13 e

i le we de CORE DAMAGJ ts e a 90 UC1 1DE 11 i

I le e e 51 COREDAMAGE 1

1 e

l en e

52 COELEnmanE l

$3 CORE 04 MAGE no 12,W

$ N Vgg c--

~.

. ~

- ~...

. FIQUTE 4 4 i

e l

e a

j EXPANDED NLOCA EVENT TREE EwsLoos a

carr AccuM ased Asets a

e 1

CK1

,,s a oKa tis 4A 4

we.

3 oK3 t ot4s see 1 -

4 4

QK2 1M48 esee i is e s s oK4 s oE47 IJ e e 1

e<g ee, e cK4 4484e 4

? uce s oE4e e

a 1

le

• 2J 8

UCS 6.8548 ee*

9 C0flE DAMAGE e

19 OK2 69544 set a2es i Le e s 11 CK3 8.7546 I

e le e

• 13 OK3 S M40 See e

13 CKN 16E40 14 OK4 40840

{

.M.,.

)

ee, is CofE oAmaE i

1 a

4 17 UCS SM48 i

e I

IL'OL.,,,

'* UC6 1Mae 1

uwe j

e

i. uC.

i M.,.

f ees 30 CGE DAMAGE 4

31 OK2 1J648 es j

s ses it e s 32 OK3 3.1647 lewi SS OK3 LSE 40 m

see 4

34 OK8 19849 g

s ihas one 74ie e

e SS UCS 1 GE-te i

est 27 COfE LAMAGE e

as oKa t otes t

• !!L l

mas osa smae 8

i

{

use L

lee s 30 oK3 gM.gg v

'g ee' St COIE DAMAGE pas 38 UCS 48 bit

(

e we*

33 COM DAMAGE 34 UC1 SM4e e

l l'.!dd 35 UC1 3.3Eas noe e

s es SS UC1 8.36-18 le e.

., Ca= A E

,e.

uwe 38 UC1

'4546 I

e le e t 38 COM DAMMM s

e 40 COM Ma'aa8

+

41 OKSA 9AE47 E4547 l

42

=

le e '

ASS 22549 see a

en N

fjI49 h

8 1

221_. es uCs s.7E.ie J

t i

9 46 UC8 (JE.10 i

i ees 47 C04 0AMAGE l

• et 48 OIGA totes e

i C *8 OK88 EDE4s ues a

ses to OK8 19811

,e e.

Co. A EnENILOCA aJ e d 58 UCS SM tt e

r

!* e i 53 CofE DAMAGE

• es 84 CKEA 14E4e o

IEE6 OK58 3 es-te uen e

i tes se CK4 14512 le e.

7 Co= oA==

no asee

$s UCS 126-11 ee*

fne*

90 COM DAMAGE tJ e e 80 UC1 110-14 99 1

fe e t 61 C04 DAMAM e

af CM DAMNM e

S3 COBE CAMAGE i

no I

15

=

e.*as e.

1 e

4 EXPANDED SLOCAW EVENT TREE wtTH SUCCESS OF PRHR

_...e alhGIIC IEv4LOCMF M

G CEf7 Acclaf Ages AAALS 1

CKt tae n n l

Is e s 3

OK3 SM48 4

Y

!e e '

S CK3 18E47 see e

4 cK3 3M47 ee ses 8

OK4 33848 ues s

4 le e '

S One 0 0E40 4

7 UCS GJE4e 8

I he UCs 1Jtes ee' 9 COME DAMAGE 1-1--

s es 11 ONS SJG47 l

I le e '

it OK3

$ 0E4e g

see 4

13 CK3 3.0549 I

l ues s

l

!s t23 14 QK4 7.18-10 f

f 16 UCS 1AE44 e..

,. Co E.A 4M 4

17 UCS 10E40 l

4 I

h18 UC8 15648 e

ses is UCs tJs-11 I

ee*

so core passAat l

21 OK2 SJt47 naea i

• • s i

Ires it OK3

$ 3G48 j

4 f

l***

D OK3 52649 1

4 24 OK3

$3644 IJ e

• s I

h35 OK4 1 M-10 a

N UCs 1M41 ees 27 C04 DAMAGE

• et SS CKt t 9549 a

l l

1 39 CK3 44E40 9ts is e s 30 CK3 4Mai

.e w,

Si m oAMAM ues 3R UC8 1 AE-11 e

r roe '

33 C04 DAMAGE 34 - OKS 1.75 46 4

11 35 OKD 43f40 t

i ses SS CMS 4.08-11 lee *

$7 CCfE DAMAGE

'en aJ e e 30 OKS t.4810 9

I e

to e i 30 CCWit DAMAGE e

1. C04 DAMAGE 41 ONEA t.7E47 n3 e 4 l

l Is e s 48 GCSS 44E48

• ta 4

Y is e t e OKSB

$9646 see 4

44 CKS 41849 h

e i1 45 UCe t eE st e

e et UC9 1M41 ees 47 C04 0AMAes

<et e

48 GWA 1M40 e

Ihm OKSB 3,4840 a

tre s 50 QK8 3 4E48 set

$1 C0fE hm*M uwe 53 UC5 10541

,e.,

. =.A A.E o

$4 OKEA 1.06-19 a

I f

le>2 s SS OK50 54541 n2e

• s

.J o s 54 OKS

$ 9643 ees 57 CJuqE DAMAGE e

une SS UCS L3848 er9 I

le ee 50 COM OAMAGE u,.

80 OWS 19E-11 eer i

re e.

51 C04 0AMAGE e

e et CofE DAMMM e COM DAnahet ac atne An eri man men-13r394

. ~, + -

a~

.-... ~ -

FIGUTE 44 EXPANDED SLOCWO EVEldt TREE Wf7H FALURE OF PRHR

..._.e

=-- -

.,7

.E e

1 OK1 9 8,ee I l

h3 CN3 13E45 m

3 0x3 a.sE4r see a

4 OMI SJE47 1 a.ses ih5 CK4 1AE47 6

OM4 tee-te a

f UC4 15E4e s

a i

Es UOs a.eE4e se+

e C0st oaaaaas e

to OK8 81848

'1 CR$

53847 11 CK3 LaE49 see e

18 CM2 40E40 h

e i

1 14 SK4 1JE4e is ua tsa te set te CN W e

17 UCS 1.7540 j

i h18 UCS e 4E40 pes I

e I

If e a UC5 405 11 it

lee,

.. aa e

it CMI SAE47 f.a ses i 4

'et l

h33.000 9AE4e a

s3 ONS 4JE to 4

see e

te CMI 8 eE-10 L

s I

has one tie-to 2

e as UC4 418 11

• e' If CofE 04aamag

'ee e

as Ox3 3JE4e a

I has CMS 7Atte 4

ses 30 ONE 5.0012 g

.e.,

1 CO,4 =

IJ e *

$2 UCS 1.8E-11 e

I

\\***

33 Cops 6 easAGE a

34 OMS tee 4e e

1 Fo us 35 ONS 1.'848 vee a

ses as OKs s.1091 eet 37 COBE DAnaAFA

• es ues 38 ONS tee 10 I

e lee t

$$ CW DM e

40 Cops Dananas e

41 CKEA 19847 9.s see i hof CNBB 7.3540 L

see "e

cr**

sJE-is e

e 44 CKS E 75-19 j

use s

i 1

h48 UCS 1 66-10 t

45 UC9 E9E-91 ee' 47 Copt Dassang l

9et i

e de ONEA 14E46 e

t l

has OKf8 tee-te e

is e s to ONS

$ 18-18 low

• St CopW DansMBE WW4 00WO uee 58 UC5 19E-11 sm o

I le e i 08 C014 Danana8

• ,r I

e s4 0xsa s4E-te s

I

(

m05 OIGE t 0516 ues f

I ses M CME SJE-13 lo w,

7 m e.a.a.

l an use SS LCS 8.4E 18

.e.

i le W9 IO MM

~

eses a0 OKs 3JE-11 4ef I

Le *

• 51 CW Dana 4M e

e et CN 0ansME ta Cost Dananos un 17 m,-

e FaGVAE e 4 gi CXPANDED SGTRW EVENT TREE WITH $UCCESS OF PRHR -

asemic WV4SfstW m

0 tafe Acetas acts ADeLS e

1 OK1 use i y

I Af e t 3 QMS 10844 e

]

le e t 9

OK3 SAE47 ese e

4 OK3 3 3E47 it e s 5 Oud 3.4E47 8

ee' S

OK4 11E46 a

7 UCS 84E48 9

2 ih6 UC4 tages ee' s

a,0fE DaaaaQE l

e 10 OK8 8 4E45 l

'et 3J. 4 ihii CK3 8eE47 I

e~

• ='

12 OKS 7M4e i

See e

13 CKI FM40 i

(

Isee e

i

(

E14 s%4 19849 e

i t

is uce 3.7s.i.

l ee' w 00sE w e

17 UC8 33843 e

I l

his Ucs 7 eses l

uwe 8

i we, is ucs e tt.n e'

39 CofE Dantag OK, 18,

Joe r

• • t tes at OKS 1AE47 e

te e '

33 OK$

14E49 e

les e

34 CMI 14E40 33 e

• 3 i

to.'JJ 25 CK4 118-16 3

2 26 UCS

$ 4511 ee' 37 CCIE 0meMas

'et e

as OK2 5.1549 e

i se <JJ se OK3 1JE4e I

tJ e e

• ns l

in e s 30 OKg 1JE.11 I

i re e '

31 CofE 0anaAGE i

l 23 4 33 UC4 33341 e

i le e '

33 CQsg DananGE e

34 OK9 b847 I

e i

l h35 CKS 1AEW e

in e s 35 OKS 13840 ee' 37 CORE CAa4 AGE

• es RJ o e 30 OK9 63840 i

e

!ee e 99 Co8E OananoE o

a CofE DAasmas e

41 OKEA e sE47 a.s e ihet OKSS 1JE47 h

ic e '

as OKs0 iMas see a

4e OK4 19849 sJ e e s

I UC9 8 4E-10 h48 a

f 48 409 415-15 ee' 47 CofE DAnanGE I

• I e

e OKSA $ 9849 e

i 1

Ed8 CKES 025-10 l

t tes to OKS SJE-12 e'

S1 cofE 0AasMat WV4STIEW tJ e

• SE UC8 18841 19535WMF e

i psLAgIL

'e e '

la CORE DAaaaat j

'et 4

54 OK$A fotot l

e i

re 'JJ E6 CK38 15E40

(

sJoe I

tes 5s one i st.13 l

l e'

57 CORE DAMAGE j

l I

no sJee ss Ucs E st tt 1

I eeg i

te e '

Se CORE DanaheE sJ e e to OK9 728 10 l

see

• 91 CORE DaaanaE e

I is et CORE DanaAGE es CORE DaunGE no 18

m m mo.

s. _..... _... _.

.m_._

,m..

..-,_,m

%w.ma e

EXPANDED SGTRWO EVENT TREE WITH FAILURE OF F RHR

-.e NBCInc I

sw4stugwo Lamb d

Cef7 Acciaf Acto e

a 1

OK1 1.s see i h3 CK3 1.4E47 e

9 OK3 S et-te see e

4 OK3 1M49

+

• 23 5

OK4

$ 4E-10 EJ e e s

• Q le 4

CK4 1 48-13 4

7 W 3.55 11 h6 UC5 4.08 11 e COsa omaemat i

e to OKt 49840

'et

• t see i hit OKS 1JE-00 e

it OK3 508-15 see a

13 OK2 1.0E-11 4

a i

h14 OK4 14E.13 f

15 UCS 175-13 e*'

to CopW DAmeAGE I

a 97 UC5 19511 4

e j

l EL18 UCS 70E 13 mes is uCs s.sE.14 e*'

80 CopW O4ae4GE 4

St OK2 E7519 stseei

'*8 g

C22 OK3 128 10 le IS OK3 BJE-13 see e

84 OK2 13512 h

e I

EN OK4 43813 8

SS UC4 48E 64 I

eei 27 ComE OnasAGE

• ef e

SB Oka 4.5E-13 4

l 1

to. ' J s 39 OK3 13418 nJ o e f

$4s is e s 3o Ogg 1 55-14 e=*

31 ComE DateAGE aJ e e 3B UCS

$ 3514 I

e i

ne' 33 CCAE DateAGE e

34 OKS 158-11 e

I E3S OKS 14411 e

ees 38 OKs 1 48-13 ee' 37 COfE D4444GE ies tJ e e 30 CKS 33813 e

1.

e fee t 3D COBE CAAAAGE e

46 CORE 0AAAAGE e

41 CK5A 4.4E-10 da OKS$

1J51e e

es OKse s55 13 see e

44 OK8 t JE-12 33 4 s

i

, R,45 UC9 116 13

+ - =

s 40 UC9 314-14 ee9 of Cont 04444GE

• es e

48 OREA 60813 e

f g

wsJ 48 OKAS 9 0E-13 SJ e e e

as a t 90 OKS 99818 ee9 51 CORE DAasAGE

)

WV4G7Ruf0 tJ ee EE UC4 175-14 u

e i

ic e +

S3 CORE DAmenet ies a

54 OK5A f at t3 e

i

['

i ne eze 56 OK58 t JE-13 l

32, 4 1

I e

e in e s SS OK8 14E18 lo r,

.7 Co oa.eaat no aJ e e la UC6 44818 set t

i ie e i 50 ComE Danemas

..E.,.

• es I

to we et COME DaaenoE e

e et CORE DAteMaE F

W. DAa44GE e

SLOCT.ELahtGTRuv0 tage6 19

FARE G11 o

e e

EXPANDED TRAN Ett)(T TREE

___.e IWCIRC WW W

i. ales G

Mf7 acclef assa Apets a

1 CK1 RJ e e 1

ta e s 3 ofr3 Les46 I

e to w '

9 CK3 3.1547 see a

4 OK3

&1847 4.t e e i laws 8

CK4 13447 I

'# 4 lee t 6

CK4 4 5E-te e

7 Uce 1.6E44 8

e i

F4J _4 UC4 1.H 44 sei e cong DAa44GE e

10 CK3 8.2E4s

• et e

aes 11 OK3 6.54 47 e

te e t it OK3 3.4848 tee e

18 CK2 44440

$42 * -

9 1

Ele OK4 1.1540 2

18 UCS 128-14 ee' 14 CoflE DAseAGE e

17 UCs tJE4e e

e i

g hit UCS 3.4849 e

in e s is UCs 3.et.t1 ee' 30 CGW DAa4Afst e

81 CK2 3.9847 uwe i

'et j

law s at OK3 t otes j

l 1e e '

23 CK3 4.15 16 i

see g

i e

34 CKS 4JE 19 83

• e I

has OK4 1 85-18 2

SS UCd 3.15-61

• • 1 27 COfE DAa44GE

'et e

as OK3 s eE co

)

e I

hat GC3 4.at te N

la e s 30 0r2 E95 f t 9

1 sei 31 CofE DAAtAGE sJ e e at UC5 1.et 11 e

i

\\

is e 1 33 CofE CandAGE 24 OK9 30840 I

,a i

e E SS CKS telde e

i ses 36 Offs 40E-11 fe e, 27 CCf. oa A.E e2ee 3e OKs a st 11 9

1 e

se e, se C014E DAasAGE e

e CosqE cas44GE e

41 OK5A 3.0547 l

8#H,e.

OK.

v44.

set e OKse t eE te see I

e 44 Ous 8.et-te uee s

i ha6 UCs 9 86-15 e

as VCs 146-11 eet 47 C0fE DeadAGE 4

• et J

e as OfsA 3JEas e

i i

se 5,ea de Oses e 4E-te use I

r ta e s 50 OK4 43E-18 ew*

St C0fs CanaAGE WSTRAse uee la UC8 12E.11 e

i. e,

i se eOfe oa aE i

tee 84 CICIA 3.48-10

]

is iJJ 85 OK58 42519 une f

I if es SS CKS 83E 13 Le, s7 ComoAAeAu F

Jee 68 UC6 106-12 ee, i

le s 9 SS CM M tse s 40 OK9 2.5E-11

• • e i

les t 41 COfW DAntAGE e

e et CofIE DAAAAGE 43 C0flE DAanAGE Afwl.zL9dTW8?ftAfs 12 @ #8

l 5.0 CATEGORIZATION OF SUCCESS SCENARIOS In the expanded event trees, the success paths on the AP600 PRA event trees are further refined to differentiate the functioning equipment in each scenario, he success paths are then binned into categories that distinguish the accident progression. His process of "binning" the end-states is the same concept used in the Level 1/ Level 2 PRA interface. Core damage paths from Level 1 are identified as different accident classes for further study in Level 2. In the expanded event trees for T/H uncertainty resolution, this same concept is applied, but the categorization is made of success paths rather than core damage paths. The categorization of the success paths is a systematic method of defining different types of possible accident progressions that lead to successful core cooling. The categorization enables a thorough assessment and greater understanding of the different successful equipment combinations.

l The nomenclature of the categories defines two main groups of success paths: OK categories and UC categories. OK categories are accident progressions that are similar to design basis accidents. Although l

most OK categories are not identical to design basis, the differences can be defined and the similarities explained. Accident scenarios that are defined within an OK category are g " low margin" and are g further considered within the T/H uncertainty resolution process. Success scenarios that do not fit within OK categories are grouped into UC categories. The c' tegorization as a UC category occurs for two a

reasons: 1) analyses of the accident progression predicts core uncovery, or 2) analyses have not been done to support the accident scenario. The UC categories are accident scenarios that are considered " low margin" and will be further considered in the T/H uncertainty resolution process.

There are 10 OK categories and the same number of UC categories. The number of categories was not pre-defined, rather categories were created based on the need to group similar accident progressions together. The consideration of the accident progression includes two phases of water injection: 1) short term, when the accumulators and CMTs provide make-up inventory, and 2) IRWST gravity injection.

Sections 5.1 and 5.2 discuss these phases of injection and some of the considerations that went into the classification process. The final phase of water injection - long-term recirculation - is treated separately from the OK and UC categorization, and is discussed in Sections 5.3 and 9.0.

First, however, there are some general comments about the method of categorization and choices that had to be made.

L Each success path is classified in only one category, although there are some success paths that fit the definition of multiple categories. A choice was made to generally include these success paths in a category based on the loss of CMTs or accumulators. However, success paths with enough failures to fit multiple category definitions are low frequency scenarios, and choice of where to include them does not impact the results of the process.

l l

2.

Expanded event trees do not always separate the success path to differentiate the exact equipment defined by the category. Once again, this only occurs in success paths oflow frequency. The i

l cwc a.gp Page 21 December 30.19%

_ _ _ - _ _ _ _ _. _ _ _ _ _. ~.__ _ _ _ __

i.

l

[

choice of where to categori7? ?is type of success path does not impact the results of the T/H uncertainty resolution proco.

However, generally the success path is categorized with the equipment success / failure that is known to be most probable. For example, a success path that does not distinguish between 2 and 3 stage 4 ADS valves may be included within a category that is defined as having at least 3 stage 4 ADS valves. In all such cases, the frequency of the success path is low, and the fraction that is 2 stage 4 ADS is negligible.

4 3.

De expanded event trees differentiate the number of stage 2 and 3 ADS valves. De fault trees used in the event tree construction can distinguish the number of lines that are open, and this is j

interpreted as:

)

f 4 stage 2,3 All 2 or 3 stage 2,3 At least half j

0 or I stage 2,3 None he number of stage 1 ADS lines is not separated because the valves are much smaller than all the other stages, and by themselves do not impact the course of the accident progression.

I However, the operation of stage 1 is estimated based on information about stages 2 and 3. The interpretations of all, at least half, or none are extended to include stage 1 in addition to stages 2 and 3.

5.1 CMT and Accumulator Injection he first phase, when the accumulators and CMTs provide make-up inventory,is similar to design bash accident conditions as long as there is at least one CMT and one accumulator. CMTs and accumulators 3

are tanks, each containing 2000 ft or approximately 100,000 lbm of water. Accumulators are designed for rapid inventory make-up when the RCS pressure falls below 700 psig. CMTs also play a role in early inventory make-up, starting at higher pressures, but injection rates are not as rapid as accumulators.

Furthermore, CMTs are important because low CMT levels provide the actuation signal for ADS. There are 2 CMTs and 2 accumulators, and the loss of one CMT and/or accumulator leaves the remaining tanks to fulfill the plant fimetions described. Therefore, a scenario with at least one CMT and at least one accumulator experiences a similar accident progression to a scenario with all CMTs and accumulators functioning. His observation is supponed by the MAAP4/NOTRUMP benchmarking effort.

The ability to lose up to 1 CMT and I accumulator without significantly impacting the accident progression is one of the foundation elements in the categorization of Le success paths. De categorization requires that judgements be made on which equipment losses have the largest impact on the accident progression. Although the loss of a CMT and/or accumulator may impact the event and its timing slightly, this impact is less significant than other equipment losses. De loss of I CMT and/o: 1 accumulator does notjeopardize the ability to successfully cool the core. Derefore, categories are defined based on other distinctions, and the following CMT/ accumulator possibilities can be grouped into the same e w wn Page 22 n

December 30,19%

a a

5 l

category:

2 CMTs and 2 accumulators l

2 CMTs and I accumulator i

1 CMT and 2 accumulators i

1 CMT and 1 accumulator i

The exception to this method of grouping is for Large LOCAs..For a LLOCA, the operation of 1

- accumulator versus 2 accumulators can have an impact on the accident progression, and these possibilities 3

are considered separately. Also note that the DBA analysis of the double-ended guillotine DVIline break only includes 1 CMT and 1 accumulator; the other CMT and accumulator spill out the break.

The loss of both CMTs or both accumulators becomes a basis for defining a success category. 'Ihis is l-because the loss of both CMTs or the loss of both accumulators removes a specific function from the plant '

response. Furthumore, the accident progression may be different depending on whether the initiating i

event is a SLOCA, NLOCA, MLOCA, LLOCA or other event.' Therefore, the following success categories are defined to address the accident scenarios with the loss of both CMTs or accumulators:

OK7,' OK8, OK9, UC1, UC2A, UC2B, UC3, UC4. Detailed discussion of each of these categories is given in Sections 6.0 and 7.0. Category UC5 also addresses the loss of accumulators, but relates to the second phase of the accident progression, and is discussed below.

i' 5.2 -

IRWST Gravity Injection The second injection phase of the accident progression, IRWST gravity injection, is generally dominated by the number of ADS lines open and whether containment is isolated. The rt.maining success categones u

(OK1, OK2, OK3, OK4, OKSA, OK5B, OK6, UC6, UC7, UC8, UC9) consider combinations of different

{

ADS failures and containment isolation status. ADS stages 1,2 and 3 vent from the pressurizer to the IRWST, while ADS stage 4 vents from the hot leg directly to containment. Therefore, the plant response to ADS 1-3 is different from the plant response to ADS-4, and this is considered within the categorization.

The plant's response to ADS actuation can also be dependent on whether there is an accumulator available in a high pre;sure (> 700 psig) scenario. Without either accumulator, analyses have shown that core uncovery can occur when a large depressurization is needed, ADS is ac*uated, and there is no make-up i

inventory to offset the inventory loss through the ADS lines. Category UC5 has been defined to address this accident progression possibility.

One of the items that is not differentiated on the expanded event trees is the number of DVI lines that are.

available for IRWST gravity injection. The PRA success criterion is that 1 out of 2 lines is sufficient.

All analyses related to supporting the PRA have been done with 1 line, and have shown this to be a successful option for IRWST gravity injection.

e.w

..,ap Page 23 om.nn n. im

5.3 Long-term Recirculation Long-term recirculation is the safety-related, passive cooling method for LOCA events after the IRWST

. is drained. His mode of cooling occurs only in LOCA events that have lost enough inventory to submerge the reactor vessel cavity with water. This natural circulation method is the back-up to a forced-flow recirculation with the RNS pumps.

The elements that may impact long-term cooling by natural circulation are the height of the water pool, the steam venting capability from the RCS, the resistance in the injection lines, the containment pressure, and the decay heat to be removed. All of these factors are potentially impacted by PRA scenarios when compared to DBA. The T/H uncertainty resolution process addresses the outstanding long-term cooling phase of the accident progression.

Within the T/H uncertainty resolution process, scenarios that are not supported by existing analyses are generally included within the UC categories. If the scenario is risk-significant, it " rises to the top" and further analysis --including the consideration of uncertainties --is done to support the claim of successful core cooling. For long-term cooling, risk-significant cases are defined from all success paths, including both the UC and OK categories. ' All success paths are grouped based on equipment failures that may impact long-term recirculation. Table 5-1 summarizes the potential differences in PRA scenarios when compared to DBA scenari - and identifies the equipment loss that may cause an impact.

From the grouping of the long-term recirculation success paths, the risk significant scenarios can be identified. The most risk-significant scenarios are anticipated to be ones with up to 1 single failure, that are already addressed by DBA analyses. The remaining risk-significant long-term recirculation scenarios are used to define a set of analytical cases to support long-term cooling in the PRA. The results of this process are documented in Section 9.0.

cw.:

a., wp Page 24 December 30.1996

Table 5-1 Summary of Potential PRA Impacts on Long-term Recirculation -

Element Equipment Loss in PRA Height of the water pool The failure of one or more CMTs and/or accumulators to drain may impacts the driving head for result in a lower water level in containment.

natural circulation The failure of a containment isolation line may allow water inventory to be lost.

RCS Steam Venting The failure oflines of ADS causes there to be less venting Capability capability, which may impact the ability to maintain the RCS pressure low enough.

Resistance of injection lines The failure of valves to open in injection / recirculation lines may impact the system flow resistance and influence the recirculation flow rate.

Containment Pressure The failure of a containment isolation line may lower the containment back pressure.

Decay Heat The failure of one or more CMTs and/or accumulators can impact the timing of the accident progression, and cause an earlier transition into long-term recirculation, thereby being at a higher decay heat.

1979 ANS best estimate decay heat is typically used for analyses that support the PRA. Uncertainties on the decay heat need to be y

considered for T/H uncertainty resolution.

ll 1

e w s.p.t.,

Page 25 o-.a*. w. im

l 6.0 OK CATEGORIES SIMILAR TO DESIGN BASIS OK categories are accident progressions that are similar to design basis accidents. Although most OK p

categories are not identical to design basis, the differences can be defined and the similarities further l

explained. Accident scenarios that are defined within an OK category are no. t, " low margin" and are got further considered within the T/H uncertainty resolution process. Generally, the OK categories are similar j

enough to design basis that the conservative SSAR Chapter 15 analyses address the dominant phenomena within the accident progression.

l Table 6-1 provides an overview of the ten OK categories, and the frequencies that have been quantified l

for each category. Following Table 6-1 is a more detailed discussion of each of the OK categories. For each OK category, there is also a table that lists all the applicable success paths from the expanded event trees and the calculated frequency of each path.

)

9 l

i i

?

i I

i l

l 4

1 i

i c m.:._r. s.wp Page 26 December 30,1996 l

i

Table 6-1 Summary of OK Categories Number Description, Relative to Detailed Description Total Frequency Design Basis (per year)

OKI More ADS-4 No Failures Beyond initiating 6.9E-3 Event OK2 Design Basis 2 DBA ADS 2.6E-5 2 1 CMT,1 Acc Containment Isolated OK3 More ADS-4

> DBA ADS-4 5.8E-4 Less ADS 1,2,3

< DBA ADS 1, 2, 3 21CMT,1Acc Containment Isolated OK4 Less ADS 1,2,3 DBA ADS 4 1.4E-6

< DB A ADS 1, 2, 3 2 1 CMT,1 Acc Containment Isolated i

OK5A More ADS-4

> DBA ADS 2.7E-6 CI Fails 21 CMT,1 Acc CI Failure OK5B More ADS-4

> DBA ADS-4 7.0E-7 Less ADS 1,2,3

< DB A ADS 1, 2, 3 CI Fails 2 1CMT,1 Ace CI Failure 1

OK6 CI Fails DBA ADS 5.9E-9 2 1 CMT,1 Acc Cl Failure OK7 2 Accumulators - Design 2 Accumulators 2.7E-5 Basis for LLOCA 2 DBA ADS-4 5 DB A ADS 1, 2, 3 2 1CMT Containment Isolated OK8 DVI Line Break with 0 CMTs 9.6E-8 Automatic ADS Actuation 1 Injecting Accumulator from Faulted CMT 2 DBA ADS-4 5 DBA ADS 1,2. 3 Containment isolated OK9 Loss of CMTs for Smaller 0 CMTs 8.8E-7 Breaks i

Notes-i 1.

"DBA ADS" is all stage 1,2 and 3 ADS and 3 out of 4 stage 4 ADS e w.:.__r...

Page 27 Decemba 41996

-.~ - -..__.

o j

Catenorv OK1 l

1hese accident scenarios are ones in which all equipment functions, except equipment disabled as part of the initiating event. These are the " top paths" on the expanded event trees, and are bounded by the LOCA design basis accident scenarios. They include the actuation of more ADS-4 lines than considered in the design basis analyses. The total frequency of the accident scenarios in this cater.ory is 6.9E-3/ year. This l

category applies to all the initiating events, and the applicable success paths are listed in Table 6-2.

Category OK2 1

These accident scenarios are collectively considered as the design basis accident scenarios. They include all accident scenarios with at least 3 stage 4 ADS, and all stages 1,2 and 3 ADS with successful i

containment isolation. Accident scenarios that meet the design basis ADS conditions are included within -

this category if they have at least 1 functioning CMT and 1 functioning accumulator.

The l

MAAP4/NOTRUMP benchmarking demonstrates that 1 CMT and 1 accumulator provides a similar l

accident progression to 2 CMTs and 2 accumulators.

The total frequency of the accident scenarios in this category is 2.6E-5/ year. The applicable success paths are listed in Table 6-3. Note that although this category can generally be considered as " design basis,"

l many of the highest frequency success paths have more ADS-4 than design basis.

This category applies to all the initiating events except for Large LOCA. LLOCA is excluded because

]

its results are dependent on the number of accumulators, and thus is considered in separate categories.

i 1

Category OK3 i

i Success category OK3 is a minor deviation from design basis. These accident scenarios have more ADS-4 lines (4 rather than 3) but less ADS 1,2 and 3 lines. Containment isolation must be successful, and there l

must be at least 1 functioning CMT and 1 functioning accumulator. The MAAP4/NOTRUMP benchmarking results demonstrate the importance of ADS-4 lines compared to ADS 1,2 and 3 lines, and support this categorization.

The total frequency of the accident scenarios in this category is 5.8E-4 / year. The applicable success paths are listed in Table 6-4. This category applies to all the initiating events except for Large LOCA.

LLOCA is excluded because its results are dependent on the number of accumulators, and thus is

]

considered in separate categories.

~

em.,

Page 28 December 30.1996

I

{

l l

Category OK4 j

Success category OK4 is similar to category OK3, except stage 4 ADS is the same as design basis he

. only difference in category OK4 when compared to design basis is the loss of some ADS 1,2 and 3 lines.

His category definition extemis to the loss of all ADS 1,2 and 3 lines, although the frequency is less than SE-9 for this possibility; the highest frequency success paths in category OK4 have the loss of no more than half of the stage 1,2 and 3 ADS lines. De frequency for the total category is 1.4E-6/ year, and the success scenarios are listed in Table 6-5.

The number of stage 1,2 and 3 ADS lines that actuate has minimal impact on the ability to achieve IRWST gravity injection. - De number of stage 4 ADS lines that actuate determines whether the RCS is i

depressurized fast enough to achieve IRWST injection prior to core uncovery. Stage 4 lines are on the

'l hot legs and vent directly to containment, providing a more effective depressurization than the stage 1, 2 and 3 lines which vent from the top of the pressurizer to the IRWST. De highest frequency success paths in category OK4 also have both accumulators and both CMTs, providing ample short-term water l

supply until IRWST gravity injection is established.

This category applies to all the initiating events except for Large LOCA. LLOCA is excluded because its results are dependent on the number of accumulators, and thus is considered in separate categories.

Categories OKSA. OK5B l

Success categories OKSA and OK5B consider the failure of complete containment isolation. De failure of containment isolation lowers the containment back pressure, which can have an impact on the accident i

progression. The distinction between categories OK5A and OK5B is the number of ADS lines that are assumed. De separation of the categories is done to illustrate that the highest frequency success paths have more successful ADS lines:

Cateeorv Freauency OK5A

- No ADS failure 2.7E-6 OK5B Some ADS 1,2,3 failure 7.0E-7 he failure of containment isolation is offset by the success of more ADS-4 lines than are credited in design basis analyses. All initiating events are included within these categories. De success paths corresponding to these categories are listed in Tables 6-6 and 6-7.

Note that as with other OK categories, a requirement for these categories is that there must be at least one

- functioning CMT and one functioning accumulator. However, there are two exceptions to this.1) The l

LLOCA success paths must have at least 2 accumulators; success paths with only I accumulator are j

classified in category UC4. 2) he DVI line break does not have to have a CMT that injects to the RCS.

His is noted on Tables 6-6 and 6-7 and the details of this possibility are explained in the discussion of category OK8.

i l

em a#wp Page 29 December 30.1996

-m..

~

- - - ~

a

\\

i i

Category OK6 i

i Category OK6 also assumes the failure of containment isolation. While categories OKSA and OK5B had l

a compensating effect with more ADS-4 than design basis, category OK6 does not. Category OK6 is the I

LOCA design basis scenario with the additional failure of containment isolation.

?

Although the design basis scenario includes containment isolation, no credit is taken in most of the DBA 4

analyses for a containment back pressure. The SSAR Chapter 15 small-break LOCA analyses show l

successful core cooling through the IRWST gravity injection phase with no elevated containment back

]

pressure. The Chapter 15 small-break LOCA break sizes correspond to the PRA LOCA initiating events i

smaller than LLOCA. The Chapter 15 large-break LOCA analyses do take credit for a containment back pressure. For this reason, LLOCA is not included in category OK6, while all other initiating events are.

he success paths corresponding to this category are listed in Table 6-8. De total frequency of this success category is 5.9E-9/ year.

i 4

Catenory OK7 l

1 Success category OK7 considers most large LOCA accident scenarios with 2 accumulators. 'Ihe other j

j requirements for classification within this category are successful containment isolation, at least 1 L

functioning CMT, and at least 3 lines of ADS-4 (design basis). There can be failures of stages 1,2 and 3 ADS.

i his category is considered to be design basis for LLOCA. The plant response in the first hundreds of l

seconds is dictated by the plant and fuel design, and the number of accumulators. CMT performance does not impact the limiting portion of the accident progression. However, at least one CMT is needed so that a low-low CMT level actuation signal will open the squib valves to the IRWST. IRWST gravity injection l

l has been demonstrated in design basis analyses supporting SSAR Chapter 15. Thus containment isolation and at least 3 lines of ADS-4 are required for a success path to be included within this category. Stages j

1,2 and 3 ADS have a negligible impact, especially for a large LOCA that provides additional venting 4

capability through the break.

l The total frequency of the accident scenarios in this category is 2.7E-5/ year. The applicable success paths

)

1-are listed in Table 6-9.

l t

l j

Category OK8 i

1

]

Success category OK8 addresses an accident scenario that is unique to a break in the DVI line. If the l

CMT isolation valve on the faulted loop opens, the water inventory from that CMT will be lost through l

i the break.~ If the intact CMT fails, there are no CMTs to provide make-up inventory to the RCS.

l However, the CMT spilling out the break will drain and provide the low level signals for ADS actuation.

1 This is the only initiating event that can have "no CMTs," and yet automatic ADS actuation occurs j

i l

eyew

---*.pm.,

Page 30 December 30,1996 h

i

without operator intervention.

The success paths in this category have successful containment isolation,1 accumulator, and DBA ADS (failure of I line of ADS 4) or all ADS 4 with the possible failure of some stages 1,2 and 3 ADS. The ADS conditions are the same as categories OK2 and OK3, which is no wowe than design basis. The only other distinction from the design basis DVI line break scenario is the failure of the CMT on the intact loop. As can be seen in Chapter 15 of the SSAR, the role of the intact CMT is minimal. It is not responsible for the ADS actuation signals, and provides very little make-up inventory to the RCS. The failure of the intact CMT does not have a significant impact on the accident progression.

Table 6-10 lists the accident scenarios in category OK8. The total frequency of the success paths in this category is 9.6E-8/ year.

Category OK9 Success category OK9 consists of scenarios that require manual ADS actuation because both CMTs fail.

However, only initiating events with relatively small breaks are included within this category. The significance of the small break area is that inventory loss is relatively slow, ar.d the operator has sufficient time to open the ADS lines before much RCS inventory is lost. The initiating events within category OK9 are transients, SLOCA, and SGTR. Larger breaks, with the same conditions of both CMTs failing, are classified within UC categories.

]

The additional requirements for this category are intended to be DBA ADS (failure of I line of ADS 4) or all ADS 4 with the possible failure of some stages 1,2 and 3 ADS. However, when 2 CMTs fail, the expanded event trees only differentiate one more failure. Therefore, some of the success paths listed on Table 6-11 include the possibility of 1 more stage 4 ADS line failure. The frequency of these paths are small, and the effect ofincluding them within this category is negligible, and do not impact the definition of this category. The total frequency of this category is 8.8E-7/ year.

It is also worth noting that this category includes success scenarios with and without PRHR. Itis questionable that some of the very small break scenarios with PRHR actually need ADS to achieve successful core cooling. However, the need for ADS has been conservatively included within the PRA modelling (i.e., if ADS fails, core damage is assumed), and thus this assumption is maintained in the expanded event trees for T/H uncertainty resolution.

mm.,

Page 31 Deesmber 30.1996

a r.

.-m s

Table 6-2 Success category OKI (Sorud by Descending Frequency)

Success Path Equipment Assumptions Frequency C1 CMT Acc ADS-4 ADS 2.3 sgtrw01 Yes 2

2 4

4 5.5E-3 nloca01 Yes 2

2 4

4 5.9E-4 tran01 Yes 2

2 4

4 1.9E-4 slocwool Yes 2

2 4

4 1.8E-4 mioca01 Yes 2

2 4

4 1.2E-4 slocaw01 Yes 2

2 4

4 1.1E-4 11oca01 Yes 2

2 4

4 7.6E-5 silbOI Yes 1

1 4

4 7.6E-5 cmttb01 Yes 1

2 4

4 6.5E-5 sgtrwo01 Yes 2

2 4

4 4.2E 7

)

TOTAL 6.9E-3 Notes:

j c m c6r.:

Page 32 c

o.ceae. 30. im

1 j

Table 6-3 Success category OK2 i

(Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequewy l

(per ear)

C1 CMT Acc ADS-4 ADS 2.3 i

nloca10 Yes 2

1 4

4 6.9E-6 sgtrwl0 Yes 2

1 4

4 3 4E-6 1

tran10 Yes 2

1 4

4 2.2E-6 sloewol0 Yes 2

1 4

4 2.1E-6 nloca04 Yes 2

2 3

4 1.9E-6 mioca10 Yes 2

1 4

4 1 AE-6 slocawl0 Yes 2

1 4

4 1.3E-6 l

nioca21 Yes 1

2 4

4 1.2E-6 sgtrw04 Yes 2

2 3

4 9.3E-7 cmtib10 Yes 1

1 4

4 7.6E-7 tran04 Yes 2

2 3

4 6.1E-7 sgtrw21 Yes 1

2 4

4 6.1E-7 I

stocwoM Yes 2

2 3

4 5.8E-7 tran21 Yes 1

2 4

4 3.9E-7 slocwo21 Yes 1

2 4

4 3.8E-7 mioca04 Yes 2

2 3

4 3.8E-7 slocaw04 Yes 2

2 3

4 3.5 E-7 mloca21 Yes 1

2 4

4 2.5E-7 1

silbo4 Yes 1

1 3

4 2.4E-7 slocaw21 Yes 1

2 4

4 23E-7 cmtibM Yes 1

2 3

4 2.1E-7 niocal3 Yes 2

1 3

4 1.6E-8 nloca28 Yes 1

1 4

4 1.0E-8 sgtrwl3 Yes 2

1 3

4 7.9E-9 sgtrw28 Yes 1

1 4

4 5.1E-9 slocwol3 Yes 2

1 5

4 4.9E-9 sgtrwo10 Yes 2

1 4

4 4.9E-9 i

tran13 Yes 2

1 3

4 4.8E-9 slocwo28 Yes 1

1 4

4 3.2E-9 enwaxmiam.p.1, Page 33 o

..b. 30. im

l Table 6-3 Success category OK2

_(Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency C1 CMT Acc ADS-4 ADS 2,3 mloca13 Yes 2

1 3

4 3.2ti-9 tran28 Yes 1

1 4

4 3.0E-9 1

slocaw13 Yes 2

1 3

4 3.0E-9 nloca24 Yes 1

2 3

4 2.9E-9 mloca28 Yes 1

1 4

4 2.1E-9 slocaw28 Yes 1

1 4

4 1.9E-9 j

i cmtib13 Yes 1

1 3

4 1.8E-9 sgtrw24 Yes 1

2 3

4 1.4E-9 1

sgtrwo04 Yes 2

2 3

4 13E-9 slocwo24 Yes 1

2 3

4 8.8E-10 sgtrwo21 Yes 1

2 4

4 8.7E-10 tran24 Yes 1

2 3

4 8.2E-10 mioca24 Yes 1

2 3

4 5.7E-10 slocaw24 Yes 1

2 3

4 53E-10 nloca30 Yes 1

1 2.3

• O-4*

2.5E-11 sgtrw30 Yes 1

1 2,3

• O 4*

1.2E 11 sgtrwo13 Yes 2

1 3

4 1.1E-11 tran30 Yes 1

1 2,3

• 0-4*

7.0E-12 sgtrwo28 Yes 1

1 4

4 6.6E-12 sloewo30 Yes 1

1 2,3

• O-4*

5.9E-12 mioca30 Yes 1

1 2,3

• O-4*

4.9E-12 slocaw30 Yes 1

1 2,3

• O-4*

4.5E-12 sgtrwo24 Yes 1

2 3

4 1.BE-12 sgtrwo30 Yes 1

1 2.3

• O-4*

1.5E-14 TOTAL 2.6E-5 Notes:

These success paths include accident scenarios with more failures than defined by category OK2. The inclusion of additional equipment failures in these paths is of negligible importance because of the low frequency of the paths.

ews.p6c._c.#m Page 34 Decemte30.1996 i

Table 6-4 Success category OK3 (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency C1

-CMT Acc ADS-4 ADS 23 a

oloca02 Yes 2

2 4

23 2.1E-4 sgtrw02 Yes 2

2 4

2.3 1.0E-4 tran02 Yes 2

2 4

23 6.6E-5 slocwo02 Yes 2

2 4

23 63E-5 mioca02 Yes 2

2 4

23 4.1E-5 slocaw02 Yes 2

2 4

23 3.8E-5 silbO2 Yes I 1

1 4

23 2.6E-5 cmtlbO2 Yes 1

2 4

23 23E-5 nloca03 Yes 2

2 4

0,1 1.9E-6 nlocall Yes 2

1 4

2,3 1.7E-6 sgtrw03 Yes 2

2 4

0,1 9.5E-7 sgtrwil Yes 2

1 4

2,3 8.6E-7 traall Yes 2

1 4

2,3 5.5E-7 slocwo11 Yes 2

1 4

23 53E-7 mioca03 Yes 2

2 4

0,1 3.9E-7 slocawo3 Yes 2

2 4

0,1 3.6E-7 miocall Yes 2

1 4

23 35E-7 slocaw11 Yes 2

1 4

23 3.2E-7 nloca22 Yes 1

2 4

2,3 3.1E-7 tran03 Yes 2

2 4

0,1 3.1E-7 slocwo03 Yes 2

2 4

0,1 2.8E-7 silt 03 Yes 1

1 4

0,1 2.5E-7 cmtlbO3 Yes 1

2 4

0,1 2.1E-7 cmtibli Yes 1

1 4

23 1.9E-7 j

sgtrw22 Yes 1

2 4

2,3 1.5E-7 agtrwo02 Yes 2

2 4

23 1.5E-7 tran22 Yes 1

2 4

23 9.8E-8 sloewo2?

Yes 1

2 4

23 9.5E-8 mloce -

Yes 1

2 4

23 6.2E-8 l

)

cwu r.#,wp Page 35 December 30,1996

I Table 64 Success category OK3 (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency C1 CMT Acc ADS 4 ADS 2,3 slocaw22 Yes 1

2 4

2,3 5.8E-8 nlocal2 Yes 2

1 4

0,1 1.6E-8 j

sgtrw12 Yes 2

1 4

0,1 7.9E-9 miocal2 Yes 2

1 4

0,1 3.2E-9 slocaw12 Yes 2

1 4

0,1 3.0E-9 nloca23 Yes 1

2 4

0,1 2.9E-9 nloca29 Yes 1

1 4

0-3 2.6E-9 tran12 fes 2

1 4

0,1 2.4E-9 slocwol2 Yes 2

1 4

0,1 2.4E-9 cmtib12 Yes 1

1 4

0,1 1.8E-9 sgtrw23 Yes 1

2 4

0,1 1.4E-9 sgtrwo11 Yes 2

1 4

2,3 1.2E-9 estrw29 Yes 1

1 4

0-3 1.2E-9 sloewo29 Yes 1

1 4

0-3 7.8E-10 tran29 Yes 1

1 4

0-3 6.8E-10 sgtrwo03 Yes 2

2 4

0,1 6.4E-10 mioca23 Yes 1

2 4

0,1 5.7E-10 stocaw23 Yes 1

2 4

0,1 5.2E-10 mioca29 Yes 1

1 4

0-3 5.0E-10 stocaw29 Yes 1

1 4

0-3 4.6E-10 stocwo23 Yes 1

2 4

0,1 4.2E-10 tran23 Yes 1

2 4

0.1 4.1E-10 sgtrwo22 Yes 1

2 4

2,3 2.2E-10 sgtrwol2 Yes 2

1 4

0,1 5.0E-12 sgtrwo29 Yes 1

1 4

03 1.5E-12 sgtrwo23 Yes 1

2 4

0,1 8.5E-13 l

l TOTAL 5.8E-4 Notes:

l ewc_._ r..,_wp Page 36 December 30,1996

Table 6-5 Sucesss category OK4 (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency CI CMT Acc ADS-4 ADS 2,3 l

nloca05 Yes 2

2 3

23 5.0E 7 satrw05 Yes 2

2 3

23 2AE-7 trao05

• a 2

2 3

23 1.5E-7 sloewo05 Yes 2

2 3

2,3 1.5E-7 mioca05 Yes 2

2 3

2,3 9.8E-8 slocaw05 Yes 2

2 3

23 9.2E-8 silbo5 Yes 1

1 3

2,3 63E-8 cmtib05 Yes 1

2 3

23 5.4E-8 nloca06 Yes 2

2 3

0,1 4.4E-9 nloca14 Yes 2

1 3

0-3 4.0E-9 sgtrw06 Yes 2

2 3

0,1 2.1E-9 sgtrwl4 Yes 2

1 3

03 1.93-9 stocwol4 Yes 2

1 3

0-3 1.2E-9 tran14 Yes 2

1 3

03 1.1E-9 mioca06 Yes 2

2 3

0,1 8.8E-10 slocawo6 Yes 2

2 3

0,1 8.1E-10 miocal4 Yes 2

1 3

03 7.9E-10 uloca25 Yes 1

2 3

03 7AE-10 slocawl4 Yes 2

1 3

0-3 7.2E-10 tran06 Yes 2

2 3

0,1 6.5E-10 sloewo06 Yes 2

2 3

0,1 6.5E-10 silbo6 Yes 1

1 3

0,1 5.5E-10 cmtib%

Yes 1

2 3

0,1 4.8E-10 cmtibl4 Yes 1

1 3

0-3 4.2E-10 agtrwo05 Yes 2

2 3

2,3 3.4E-10 sgtrw25 Yes 1

2 3

0-3 3.lE-10 slocwo25 Yes 1

2 3

0-3 2.1E-10 tran25 Yes 1

2 3

0-3 1.8E-10 mioca25 Yes 1

2 3

0-3 1.5E-10 eww=-tve p Page 37 December 30,1996

1 l

l Table 6-5 l

Success category OK4 (Soned by Descending Frequency)

Success Path Equipment Assumptions Frequency CI CMT Acc ADS-4 ADS 2,3 slocaw25 Yes 1

2 3

0-3 1.2E-10 silb21 Yes 0 0) 1 3

0-3 4.0E-11 sgtrwoJ 4 Yes 2

1 3

0-3 2.4E-12 i

i sgtrwo06 Yes 2

2 3

0,1 1.4E-12 sgtruo25 Yes 1

2 3

0-3 4.1E-13 i

niimaner a 3::

j TrJrAL 1 AE-6 I

NoteM (1#

Although no CMT injection to the RCS is credited, ADS actuation occurs from the faulted CMT blowing down through the break.

==,

l l

l e.pp-nw.,

Page 38 l

o.c.a*. n. im l

l l

.=

l j

l Table 6-6 Success category OK5A (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency Cl CMT Acc ADS-4 ADS 2,3 Lloca41 No 2

2 4

4 9.5E-7 sgtrw41 No 2

2 4

4 4.6E-7 tran41 No 2

2 4

4 3.0E-7 slocwo41 No 2

2 4

4 2.9E-7 mioca41 No 2

2 4

4 1.9E-7 j

slocaw41 No 2

2 4

4 1.8E-7 j

11oca31 No 2

2 4

4 13E-7 silb33 No 1

1 4

4 1.2E-7 cmtlb33 No 1

2 4

4 1.0E-7 nloca48 No 2

1 4

4 8.0E-9 sgtrw48 No 2

1 4

4 3.9E 9 slocwo48 No 2

1 4

4 2.5E-9 tran48 No 2

1 4

4 23E-9 mloca48 No 2

1 4

4 1.6E-9 slocaw48 No 2

1 4

4 1.5E-9 aloca54 No 1

2 4

4 1.4E-9 i

i cmtib40 No 1

1 4

4 8.7E-10 sgtrwS4 No 1

2 4

4 7.0E-10 sgtrwo41 No 2

2 4

4 6.6E-10 l

sloewoS4 No 1

2 4

4 4.4E-10 tran54 No 1

2 4

4 3.5E-10 mioca54 No 1

2 4

4 2.8E 10 slocawS4 No 1

2 4

4 2.6E-10 lloca45 No 1

2 4

4 2.0E 10 silb44 No Om 1

4 4

8.8E-11 I

l sgtrwo48 No 2

1 4

4 5.0E-12 l

l sgtrwo54 No 1

2 4

4 7.6E-13 TOTAL 2.7E-6

c. w w _;. # wp Page 39 o.aoe. m im

i i

Table 6-6 Su: cess category OKSA r

(Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency l

CI CMT Acc ADS-4 ADS 2,3 Notes:

(1)

Although no CMT injection to the RCS is credited. ADS actuation occurs from the faulted CMT blowing down through the break.

1 i

i c W e T..,

twp Page 40 December 30,1995 l

Table 6-7 Success category OK5B (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequeon C1 CMT Acc ADS-4 ADS 2,3 nloca42 No 2

2 4

2,3 2AE-7 sgtrw42 No 2

2 4

2,3 1.2E-7 tran42 No 2

2 4

2,3 7.4E-8 slocwo42 No 2

2 4

2,3 73E-8 mioca42 No 2

2 4

2,3 4.7E-8 stocaw42 No 2

2 4

23 4AE-8 11oca32 No 2

2 4

2,3 3.4E-8 silb34 No 1

1 4

2,3 3.0E-8 cmtlb34 No 1

2 4

2,3 2.6E-8 nloca43 No 2

2 4

0,1 2.2E-9 oloca49 No 2

1 4

0-3 2.0E-9 sgtrw43 No 2

2 4

0,1 1.1E-9 sgtrw49 No 2

1 4

0-3 9.2E-10 stocwo49 No 2

1 4

0-3 6.0E-10 trao49 No 2

1 4

0-3 4.4E-10 mloca43 No 2

2 4

0,1 43E-10 j

stocaw43 No 2

2 4

0,1 3.9E-10 mioca49 No 2

1 4

0-3 3.8E-10 nloca55 No 1

2 4

0-4 3.6E-10 61ocaw49 No 2

1 4

0-3 3.5E-10 stocwo43 No 2

2 4

0,1 3.2E-10 11oca33 No 2

2 4

0,1 3.03-10 tran43 No 2

2 4

0,1 2.8E-10 silb35 No 1

1 4

0,1 2.7E-10 cmtib35 No 1

2 4

0,1 23E-10 cmt!b41 No 1

1 4

0-3 2.0E-10 l

sgtrwo42 No 2

2 4

2,3 1.6E-10 sgtrwS5 No 1

2 4

0-4 1.6E-10 sloewoS5 No 1

2 4

0-4 1.0E-10

(

ews

--.r..,~.wp Page 41 December 30,1996

?

i l

Table 6-7 Success category OK5B (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency Cl CMT Acc ADS 4 ADS 2,3 i

mioca55 No 1

2 4

0-4 7.2E-11 tran55 No 1

2 4

0-4 6.2E-11 slocaw5$

No 1

2 4

0-4 5.8E-11 I

lloca46 No 1

2 4

0-3 4.4E-11 silb45 No 0 0) 1 4

0-3 2.0E-11 sgtrwo49 No 2

1 4

0-3 9.6E-13 sg:ewo43 No 2

2 4

0,1 5.8E-13 sgtrwoS5 No 1

2 4

0-4 13E-13 TOTAL 7.0E-7 Notes:

I (1)

Although no CMT injection to the RCS is credited ADS actuation occurs from the faulted CMT blowing down through the break.

l l

l l

l 1

l MM S5 Deceba M.1996

+

e.

l l

Table 6-8 Success category OK6 (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency CI CMT Acc ADS-4 ADS 2,3 nloca44 No 2

2 3

4 2.2E-9 sgtrw44 No 2

2 3

4 1.lE-9 slocwo44 No 2

2 3

4 6.7E 10 tran44 No 2

2 3

4 5.6E-10 mioca44 No 2

2 3

4 4.4E 10 slocaw44 No 2

2 3

4 4.1E 10 silb36 No 1

1 3

4 2.8E-10 cmtib36 No 1

2 3

4 2.4E-10 nloca50 No 2

1 23 0-4 1.9E-11 sgtrwSO No 2

1 2.3 0-4 9.2E-12 tran50 No 2

1 2,3 0-4 4.5E-12 mloca50 No 2

1 2,3 0-4 3.8E-12 slocawSO No 2

1 2,3 0-4 3.4E-12 oloca56 No 1

2 2,3 0-4 3.4E-12 sloewoSO No 2

1 2,3 0-4 3.1E 12 cmtlb42 No 1

1 2,3 0-4 2.0E-12 sgtrw56 No 1

2 2,3 0-4 1.6E-12 sgtrwo44 No 2

2 3

4 1.2E-12 sloewoS6 No 1

2 2,3 0-4 83E-13 l

\\

mloca56 No 1

2 2,3 0-4 6.5E-13 1

tran56 No 1

2 2,3 0-4 6.3E-13 1

I slocaw56 No 1

2 2,3 0-4 5.9E-13 silb46 No 09 1

2,3 0-4 2.0E-13 sgtrwoSO No 2

1 2,3 0-4 9.9E 15 sgtrwoS6 No 1

2 2,3 0-4 1.4E-15 TOTAL 5.9E-9 l

Notes:

(2)

Although no CMT injection to the RCS is credited, ADS actuation occurs from the faulted DMT blowing down through the break.

r Page 43 e.pwe __r.#,

p December 30,1996

9 1

1 i

l Table 6-9 Success category OK7 (Sorted by Descending Frequency) l Success Path Equipment Assumptions Frequency Cl CMT Acc ADS-4 ADS 2,3 lloca02 Yes 2

2 4

2,3 2.7E-5 lloca03 Yes 2

2 4

0.1 2.5E-7 Iloca04 Yes 2

2 3

4 2.5 F-7 l

11oca18 Yes 1

2 4

4 1.6E-7 Iloca05 Yes 2

2 3

2,3 6.4E-B Iloca19 Yes 1

2 4

2,3 4.0E-8 Iloca06 Yes 2

2 3

0,1 5.6E-10 lloca21 Yes 1

2 3

4 3.7E-10 lloca20 Yes 1

2 4

0,1 3.6E-10 lloca22 Yes 1

2 3

0-3 8.0E-11 TOTAL 2.7E-5 Notes:

Table 6-10

)

Success category OK8 (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency Cl CMT Acc ADS-4 ADS 2,3 silb17 Yes 0 (U 1

4 4

7.6E-8 silbl8 Yes 0 (0 1

4 2,3 1.9E-8 silb20 Yes 0 (D 3

4 1.8E-10 silbl9 Yes O ("

1 4

0,1 1.7E-10 l

TOTAL 9.6E-8 1

I Notes:

1 l

j (1)

Although no CMT injection to the RCS is credited. ADS actuation l

occurs from the faulted CMT blowing down through the break.

l en+p6xe._ _c.,~.wp Page 44 December 30,1996 I

l i

Table 6-11 Success category OK9 (Sorted by Descending Frequency)

Success Path Equipment Assurnptions Frequency CI CMT Acc ADS 4 ADS 2,3 sgtrw34 Yes 0

2 4

4 63E-7 sgtrw35 Yes 0

2 4

0-3 1.6E 7 slocwo34 Yes 0

2 4

4 2.8E-8 tran34 Yes 0

2 4

4 2.8E-8 slocaw34 Yes 0

2 4

4 1.7E-8 slocwo35 Yes 0

2 4

0-3 7.1E-9 tran35 Yes 0

2 4

03 6.0E-9 sgtrw38 Yes 0

1 2-4 0-4 53E-9 slocaw35 Yes 0

2 4

0-3 43E 9 sgtrw36 Yes 0

2 2,3 0-4 1.5E-9 sgtrw60 No 0

1 2-4 0-4 7.2E-10 sloewo38 Yes 0

1 2-4 0-4 2.4E-10 slocaw38 Yes 0

1 2-4 0-4 1AE-10 tran38 Yes 0

1 2-4 0-4 9.8E-1I stocwo36 Yes 0

2 2,3 0-4 6.7E-11 sgtrwo34 Yes 0

2 4

4 6.5E-11 tran36 Yes 0

2 2,3 0-4 6.0E-11 j

stocaw36 Yes 0

2 2,3 04 4.0E-11 sloewo60 No 0

1 2-4 0-4 3.2E-11 tran60 No 0

1 2-4 04 2.5E-Il slocaw60 No 0

1 2-4 0-4 1.9E-11 sgtrwo35 Yes 0

2 4

0-3 1.4E-11 sgtrwo38 Yes 0

1 2-4 0-4 23E-13 sgtrwo36 Yes 0

2 2,3 04 1.4E-13 sgtrwo60 No 0

1 2-4 0-4 5.9E-14 TOTAL 8.BE-7 Notes:

l i

l i

ewe

...,~..wp Page 45 December 30.1996

i l

i 7.0 UC CATEGORIES OF LOW-MARGIN ACCIDENT SCENARIOS De categorization method of the success paths in the expanded event.res started with the concept of needing to define low-margin accident scenarios. As the process evolved, the low-margin scenarios were grouped into "UC" categories. De purpose of defining UC categories is to develop a list of PRA accident j

scenarios that are closest to the limits of acceptability, and thus would be most susceptible to T/H uncertainty having an impact on the conclusions of successful core cooling versus core damage.

Low-nargin is defined as a scenario that experiences core uncovery. Core uncovery is defined as the predi;ted coolant two-phase mixture level falling below the top of the active fuel. The occurrence of core uncovery is used only as a screenine criterion for an accident scenario to be further considered within the T/H uncertalaty resolution process. The acceptance criterion for considering an accident scenario as successful core cooling in the PRA is that the PCT remains below 2200*F, which is consistent with the Appendix K cr.terion for LOCAs.

i ne process of identifying the types of core uncovery extends from the same process that was used to develop the PRA Phenomena identification and Ranking Tables (PIRTs) to support the MAAP4/NOTRUMP benchmarking effort. To develop the PIRTs, a spectmm of PRA scenarios were examined by a group of experts with experience in AP600 systems design, small-break LOCA analyses, PRA and PIRTs. Key thermal-hydraulic phenomena which could impact challenges to core coolant inventory were identified (with an "H" for high importance). Dese same challenges can also be defined in terms of the equipment loss that causes them to occur. This process lead to the definition of categories UCI through UC5.

Categories UC6 through UC9 are developed slightly differently. These UC categories include accident scenarios that cannot be directly supported by existing analyses, and are therefore assumed to result in core uncovery in the categorization process. Rather than perform additional analyses to determine whether the core remains covered, the information from the expanded event trees permits a risk-inforTned decision to be made on whether additional analyses are needed.

Table 7-1 provides an overview of the ten UC categories, and the impact on the Focused PRA if these categories were counted as core damage rather than successful core cooling. The impact is provided in terms of the change in the Focused PRA Core Damage Frequency (CDF) and Large Release Frequency (LRF), if the accident were core damage rather than successful core cooling. The method for determing the CDF and LRF impact is explained in Section 8.1. Following Table 7-1 is a more detailed discussion of each of the UC categories. For each Uc category, there is also a table that lists all the applicable succes:, paths from the expanded event trees and the calculated frequency of each path. Summaries and conclusions on the risk significance of each category can be found in Section 8.0.

i ewpww.#wp Page 46 December 30.1996

Table 7-1 Summary of UC Categories Number Description Initiating Defining If counted as core Event Equipment damage, increase to Conditions Focused PRA ACDF ALRF UCI No Make-up Inventory if NLOCA 0 CMTs 1.4E-7 8.2E-9 RCS Pressure is Greater DVILB than 700 psig UC2A 1 Accumulator Depletes MLOCA 0 CMTs 1.0E-9 8.1E-Il Prior to Operator CMT LB 1 Accumulator intervention UC2B 2 Accumulators Deplete MLOCA 0 CMTs 1.2E-7 7.5E-9 Prior to Operator CMT LB 2 Accumulators Intervention UC3 No Rapid Inventory MLOCA 0 Accumulators 2.2E-8 1.3E-9 Make-up During CMTLB Blowdown J

UC4 Reduced Inventory Make-LLOCA I Accumulator 1.lE-6 6.9E-8 up During LLOCA

)

Reflood UC5 No Make-up When ADS NLOCA 0 Accumulators 7.2E-7 7.6E-8 is Actuated at Higher DVI LB Pressure SLOCA SGTR i

Transients UC6 Reduced ADS-4 All 2 stage 4 ADS 3.4E-7 7.5E-8 Cont Isolation UC7 No ADS-4 LLOCA 0 stage 4 ADS 3.2E-9 1.9E-10 Cont Isolation UC8 No Containment Isolation LLOCA CI Failure 3.lE-10 3.1E-10 UC9 No Containment Isolation All 1.7E-9 1.7E-9 Reduced ADS cwpw,60x w.r.

twp Page 47 December 30.1996

Category UCl Category UCI contains scenarios with the failure of both CMTs. Without CMTs, operator action is the only means of opening ADS lines to depressurize the RCS to achieve IRWST gravity injection. Prior to operator intervention, the only source of make-up water is the accumulators. However, accumulators can inject only after the RCS pressure falls below 700 psig. For LOCA break sizes that do not depressurize i

below this point, there is the potential for core uncovery due to the lack of make-up water.

The potential for this type of core uncovery is also impacted by operator action time. The question to be considered is whether core uncovery occurs prior to the break depressurizing the RCS below 700 psig and before the operator manually opens ADS lines. With operetor action times of 20 or 30 minutes credited in the PRA success scenarios, the core may uncover prior to accumulator injection, as shown in Figure 7-1. Accumulator injection starts shortly after the core uncovers, but the RCS depressurization rate is not sufficient to provide rapid accumulator injection to recover the core. The period of core uncovery ends when the operator opens ADS lines, allowing the accumulators to inject rapidly.

The I OCA break sizes that lead to this type of core uncovery are approximately 2" to 4" in diameter.

The corresponding initiating events are Intermediate LOCAs (NLOCAs) and DVI Line breaks. Smaller break sizes lose inventory at a slow enough rate that the coolant inventory is not challenged prior to operator action; they are classified in category OK9. Larger breaks depressurize so that the accumulator (s) can inject prior to core uncovery, and are classified in categories UC2A and UC2B, Table 7-2 shows the applicable success paths, and the impact on the Focused PRA CDF and LRF if the scenario were counted as core damage.

Catecory UC2 A, UC2B Like category UCl, categories UC2A and UC2B address the failure of both CMT::. Without CMTs, operator action is the only means of opening ADS lines to depressurize the RCS to achieve IRWST gravity injection. Prior to operator intervention, the only source of make-up water is the accumulators.

However, for relatively large breaks, accumulator inventory may deplete prior to operator action to open ADS. This can create a period of core uncovery after accumulators empty and prior to operator intervention. This type of core uncovery applies to breaks from approximately 7" to 9" diameter, as shown in Figure 7-1. 'Ihe corresponding initiating events are Medium LOCAs (MLOCAs) and CMT line breaks Larger breaks do not rely on ADS lines opening to achieve gravity injection since the break will depressurize the RCS to IRWST injection. Furthermore, larger breaks count failure of both CMTs as core l

damage.

The distinction between category UC2A and category UC2B is the number of accumulators available for injection to the RCS. The depth and duration of core uncovery is greater when there is only one accumulator (category UC2A). With two accumulators, the operator has more time to take action to open e w.:._.r.#twp Page 48 December 30.1996

i ADS before core uncovery would occur. However, for the largest breaks in category UC2B, core uncovery may still occur.

Table 7-3 and Table 7-4 show the applicable success paths and the impact on the Focused PRA CDF and LRF if the scenarios were counted as core damage.

Category UC3 Category UC3 is a type of core uncovery that occurs in scenarios with the failure of both accumulators.

He rapid make-up capability of the accumulators is essential for large breaks, and the failure of both accumulators is counted as core damage in the PRA large loss-of-coolant accident (LLOCA) event tree.

However, for breaks smaller than a LLOCA, the PRA success paths do not require any accumulators if j

at least I CMT functions. De CMT, although a similarly-sized large tank of water, does not provide the rapid make-up capability. Herefore, core uncovery can occur for breaks a little smaller than LLOCA.

De corresponding initiating events are MLOCA and CMT Line Break. For smaller break sizes, inventory loss through the break is at a slower rate, and the CMT can perform an inventory make-up function in time to prevent this type of core uncovery.

Table 7 5 shows the applicable success paths, and the impact on the Focused PRA CDF and LRF if the scenarios were counted as core damage.

Cateeory UC4 The fourth type of core uncovery occurs in Large LOCAs (LLOCAs) due to the high rate of inventory

]

loss from the break. LLOCA is a design basis accident (DBA) analyzed and documented in Chapter 15 of the SSAR. The DBA scenario includes 2 accumulators, and core uncovery occurs due to the large inventory loss through the break. The success of this accident scenario has been demonstrated, including conservative assumptions, and is not subject to further investigation in this T/H uncertainty resolution process. However, the LLOCA success criterion for the PRA only requires 1 accumulator. The failure of an accumulator could impact the PCT during reflood.

Table 7-6 shows the applicable success paths, and the impact on the Focused PRA CDF and LRF if the scenarios were counted as core damage.

Cateeorv UCS l

Category UC5 is a type of core uncovery also due to the loss of accumulators. Categories UC3 and UC4 were associated with the accumulators and their ability to provide rapid make-up for medium and large breaks. Category UC5 completes the examiraion of the effect oflosing accumulators for the remaining initiating events.

I c w,6m.:

._c..,s..p Page 49 December 30,1996 l

l

The initiating events to be considered are all those with breaks smaller than MLOCA (6"), including Transients with loss of heat removal that can result in loss of inventory through the pressunzer safety valve. The accumulator cannot function until the RCS pressure is less than 700 psig, which happens when ADS lines are opened. The RCS pressure is relatively high (between 700 psig and 2500 psig) when ADS l

1s opened, and the mass lost through the ADS is high. Accumulators provide rapid inventory make-up l

for this condition. However,if both accumulators fail, thermal-hydraulic analyses show that core uncovery j.

can occur. This type of core uncovery applies to NLOCA, SLOCA, SGTR and Transients.

l l

Table 7-7 shows the applicable success paths, and the impact on the Focused PRA CDF and LRF if the scenarios were counted as core damage.

1 1

Cateeorv UC6 1

I Category UC6 contains accident scenarios from all initiating events with 2 stage 4 ADS and successful i

containment isolation. The concern for this category is whether the reduced ADS capacity influences the ability to achieve and maintain IRWST gravity injection with the increased injection capability afforded l

by containment isolation.

There are currently no analyses that support this accident scenario. Preliminary MAAP4 analyses were performed with 2 stage 4 ADS. However, the MAAP4/NOTRUMP benchmarking effort determined that l

the ADS stage 4 model implemented in MAAP4 had not ad~equately accounted for the line resistances.

Subsequently, benchmarking cases were modified to model the more probable condition of 3 stage 4 ADS, i

although the pessimism of no containment isolation was maintained.

I Because of the lack of analytical suppon for the 2 stage 4 ADS scenario, it is conservatively assumed to j

result in core uncovery and the possibility of core damage is entertained through this T/H uncenainty resolution process. When comparing this category to other analyzed scenarios, the main issue becomes whether the positive effect of the containment back pressure compensates for the loss of ADS venting I

capability.

Table 7 8 shows the applicable success paths, and the impact on the Focused PRA CDF and LRF if the scenarios were counted as core damage.

Category UC7 Category UC7 addresses the special scenario of a large LOCA without any ADS, but with the success of containment isolation. Large LOCA is the only PRA initiating event that credits IRWST gravity injection without the actuation of any ADS. The size of the LOCA break is believed to be large enough to provide i

the needed venting for IRWST gravity injection. However, analyses to support this have not been performed.

i l

mep6mmm=nwp wp Page 50 p.ae. 30. im i

. _ _ _ _ ~.

l Table 7-9 shows the applicable success paths, and the impact on the Focused PRA CDF and LRF if the scenarios were counted as core damage. Note that although the desire is to separately consider the impact of no ADS, the expansion of the LLOCA event tree is not refined to the isolation of this option. De result is that the estimated numerical values listed for the frequency of this category are high. However, this still results in a non-risk-significant frequency.

Category UC8 Category UC8 is defined as the loss of containment isolation for the large LOCA initiating event. Another defining criterion of this category is design basis ADS assumptions. With the additional failure of containment isolation, no analyses have been done for large LOCA to show either the short term or long term effects. All other initiating events with smaller break sizes have been analyzed, and are within category OK6.

Table 7-10 shows the applicable success paths, and the impact on the Focused PRA CDF and LRF if the scenarios were counted as core damage.

Category UC9 Category UC9 is defined as the loss of containment isolation along with ADS losses that reduce the ADS venting capacity below that assumed in design basis conditions. His category is defined to encompass all initiating events. It includes the most limiting success paths (i.e., ones with the most failures) on all the event trees.

Although preliminary MAAP4 analyses had been done to support most of the success paths applicable to this category, no analyses have been done since the MAAP4 code was benchmarked. Herefore, no attempt is made to draw distinctions between which of the initiating events and break sizes would result in core uncovery. Rey are all pessimistically assumed to result in core uncovery. Table 7-11 lists the success paths, and the impact on the Focused PRA CDF and LRF if the scenarios were counted as core damage.

l i

T c.p.pera_r.s-p Page 51 Deceaker30.1996

i Figure 7-1 PRA Scenarios Without CMTs 1 Accumulator, No ADS 30 CORE

/ UNCOVERS

[20 8

E v

' ACCUMULATORS i q)

CORE UNCOVERS

.. gg

~

f5 to

.p 4

j

+-- SLOCA ~1 :

NLOCA lc MLOCA i

=

0 0

2 6

8.75 Break Equivalent ID (inches) eet r..

.p Page 52 December 30.1996

Table 7-2 Success category UCI (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency If counted as core damage, CI CMT Acc ADS-4 ADS 2,3 A CDF A LRF "'

oloca34 Yes 0

2 4

4 9.2E-8 9.2E-8 5.5E-9 nloca35 Yes 0

2 4

03 23E-8 23E-8 1.4E-9 silb28 Yes 0

1 4

4 1.6E-8 1.6E-8 9.8E-10 silb29 Yes 0

1 4

0-3 4.2E-9 4.2E-9 2.5E-10 nloca38 Yes 0

1 2-4 0-4 7.8E 10 7.8E-10 4.7E 11 nloca36 Yes 0

2 2,3 0-4 23E-10 23E-10 1.4E-Il nloca60 No 0

1 2-4 0-4 1.lE-10 1.1E-10 6.4E-12 silb30 Yes 0

1 2,3 0-4 3.9E-Il 3.9E-11 23E 12 silb50 No 0

1 2-4 0-4 1.9E-11 1.9E-11 1.lE-12 TOTAL I AE-7 1.4E-7 8.2E-9 Notes:

(1)

LRF for scenarios with containment isolation is estimated at 6% of core damage. Scenarios without containment isolation increase the LRF by 100% of the core damage frequency.

l e w wp.t.,

Page 53 December 30,1996 l

Table 7-3 Success category UC2A (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency if counted as core damage, C1 CMT Acc ADS-4 ADS 2,3 A CDF A LRF

• cmtib28 Yes 0

1 4

4 6.7E-10 6.7E-10 4.0E-11 cmtib29 Yes 0

1 4

0-3 1.6E-10 1.6E-10 9.5E-12 mioca38 Yes 0

1 2-4 0-4 1.5E-10 1.5E-10 9.2E-12 mioca60 No 0

1 2-4 0-4 2.1E 11 2.1E-11 2.1E-11 cmtib30 Yes 0

1 2,3 0-4 1.6E-12 1.6E-12 9.5E-14 cmtib50 No 0

1 2-4 0-4 7.6E-13 7.6E-13 7.6E-13 TOTAL 1.0E-9 1.0E-9 8.1E-11 Notes:

(1)

LRF for scenarios with containment isolation is estimated at 6% of core damage. Scenarios without containment isolation increase the LRF by 100% of the core damage frequency.

l i

es e pocow =nve m.,

Page 54 December 30,1996

(

Table 7 4 Success category UC2B (Sorted by Descending Frequency)

Sucress Path Equipment Assumptions Frequency If counted as core damage, C1 CMT Acc ADS-4 ADS 2,3 A CDF A LRF

• cmt1b21 Yes 0

2 4

4 8.0E-8 8.0E-8 4.8E-9 cmtib22 Yes 0

2 4

2,3 2.0E-8 2.0E-8 1.2E-9 mioca34 Yes 0

2 4

4 1.8E-8 1.8E-8 1.1E-9 mioca35 Yes 0

2 4

0-3 4.6E-9 4.6E-9 2.8E-10 cmtlb24 Yes 0

2 3

4 1.9E-10 1.9E-10 1.1E 11 l

cmtib23 Yes 0

2 2,3 0-4 1.8E-10 1.8E-10 1.1E-11 cmtib46 No 0

2 2,3 0-4 9.2E-11 9.2E-11 9.2E-11 mioca36 Yes 0

2 2.3 0-4 4.5E-11 4.5E-11 2.7E-12 cmtib25 Yes 0

2 2,3 0-4 4.1E-11 4.1E-11 2.5E-12 cmtib47 No 0

2 2.3 04 2.0E-11 2.0E-11 2.0E-11 cmtib26 Yes 0

2 2,3 0-4 6.5E-12 6.5E-12 3.9E-13 cmtib48 No 0

2 2-4 0-4 2.0E 13 2.0E-13 2.0E-13 TOTAL 1.2E-7 1.2E-7 7.5E-9 Notes:

(1)

LRF for scenarios with containment isolation is estimated at 6% of core damage. Scenarios without containment isolation increase the LRF by 100% of the core damage fiequency.

1 1

l l

e w p p6m m - = w.p m.,

Page 55 December 30,19%

l

Table 7-5 Success category UC3 (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency if counted as core damage, (per year) increase to Focused PRA g

gg 4CDF A LRF 0) miocal7 Yes 2

0 4

4 1.lE-8 1.1E-8 6.7E-10 cmt!bl7 Yes 1

0 4

4 6.2E-9 6.2E-9 3.7E 10 miocals Yes 2

0 4

0-3 2.8E-9 2.8E-9 1.7E-10 cmtibl8 Yes 1

0 4

0-3 1.6E-9 1.6E-9 93E-Il mioca19 Yes 2

0 2,3 0-4 2.7E-11 2.7E-Il 1.6E-12 mioca32 Yes 1

0 2-4 0-4 1.7E-Il 1.7E-11 1.0E-12 cmtlbl9 Yes 1

0 2,3 0-4 IJE-Il 13E-11 8.0E-13 mloca52 No 2

0 24 0-4 13E-11 13E-Il 1JE-Il nloca58 No 1

0 24 0-4 1.2E-Il 1.2E-Il 1.2E-Il cmtib44 No 1

0 2-4 0-4 6.0E-12 6.0E-12 6.0E-12 mioca58 No 1

0 2-4 0-4 23E-12 23E-12 23E-12 TOTAL 2.2E-8 2.2E-8 13F-9 Notes:

(1)

LRF for scenarios with containment isolation is estimated at 6'A of core damage. Scenarios without containment isolation increase the LRF by 100% of the core damage frequency.

ew:

,,n Page56 December 30,19%

l i

t i

Table 7-6 l

Success category UC4 (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency if counted as core damage, C1 CMT Acc ADS-4 ADS 2,3 A CDF A LRF

• i lloca10 Yes 2

1 4

4 8.9E-7 8.9E-7 53E-8 Ilocall Yes 2

1 4

2.3 2.2E-7 2.2E-7 13E-8 l

Ilocal3 Yes 2

1 3

4 2.lE-9 2.lE-9 1.2E-10 Ilocal2 Yes 2

1 4

0,1 2.lE-9 2.1E 9 1.2E-10 lloca25 Yes 1

1 4

4 13E-9 13E-9 8.0E-11 floca39 No 2

1 4

4 1.lE-9 1.1E-9 1.1E-9 llocal4 Yes 2

1 3

0-3 5.0E-10 5.0E-10 3.0E-11 lloca26 Yes 1

1 4

0-3 3.2E-10 3.2E-10 1.9E-11 I

lloca40 No 2

1 4

0-3 2.6E-10 2.6E-10 2.6E-10 Ilocal5 Yes 2

1 2

0-4 2.0E-10 2.0E-10 1.2E-11 11ocal6 Yes 2

1 0.1 04 2.7E-11 2.7E-11 1.6E-12 lloca27 Yes 1

1 2.3 04 3.2E-13 3.2E-13 1.9E-13 lloca4i No 2

1 2,3 0-4 2.7E-13 2.7E 13 2.7E-12

(

Iloca50 No 1

I 2-4 0-4 8.2E-13 8.2E-13 8.2E-13 11oca28 Yes 1

1 0,1 04 33E-14 33E-14 2.0E-15 l

Iloca42 No 2

1 0,1 04 7.6E-15 7.6E-15 7.6E-15 11oca51 No 1

1 0,1 0-4 0.0 0.0 0.0 TOTAL 1.lE-6 1.lE-6 6.9E-8 l

l Notes:

1 (1)

LRF for scenarios with containment isolation is estimated at 6% of core damage. Scenarios without containment isolation increase the LRF by 100% of the core damage frequency.

l l

l f-ew.pm..,

Page 57 December 30.19%

l

l Table 7 7 Success category UC5 (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency If counted as core damage, CI CMT Acc ADS-4 ADS 2,3 A CDF A LRF m silbl0 Yes 1

0 4

4 4.4E-7 4.4E-7 2.6E-8 silbil Yes 1

0 4

2.3 1.lE-7 1.l E-7 6.7E-9 nlocal7 Yes 2

0 4

4 5.6E-8 5.6E-8 3.4E-9 sgtrwl7 Yes 2

0 4

4 2.8E-8 2.8E-8 2.8E-8 tran17 Yes 2

0 4

4 1.dE-8 1.8E-8 1.l E-9 le slocwol7 Yes 2

0 4

4 1.7E-8 1.7E-8 1.0E-9 oloca18 Yes 2

0 4

0-3 1.4E-8 1.4E 8 8.6E-10 slocawl7 Yes 2

0 4

4 1.0E-8 1.0E-8 63E-10 sgtrwl8 Yes 2

0 4

0-3 7.0E-9 7.0E-9 7.0F sloewo18 Yes 2

0 4

0-3 4.4E-9 4.4E-9 2.6E-It, tran18 Yes 2

0 4

0-3 3.6E-9 3.6E-9 2.2E-10 slocawl8 Yes 2

0 4

0-3 2.6E-9 2.6E-9 1.6E-10 silbl3 Yes 1

0 3

4 1.0E-9 1.0E-9 6.1E-11 silbl2 Yes 1

0 4

0,1 1.0E-9 1.0E-9 6.1E-11 silb40 No 1

0 4

4 5.lE-10 5.lE-10 5.lE-10 silb24 Yes 0*

O 4

4 3.2E-10 3.2E-10 1.9E-11 silbl4 Yes 1

0 3

0-3 2.5E-10 2.5E-10 1.5E-Il nloca19 Yes 2

0 23 0-4 1.4E 10 1.4E-10 83E-12 silb41 No 1

0 4

0-3 1.2E-10 1.2E-10 1.2E-10 nloca32 Yes 1

0 2-4 0-4 8.5E-11 8.5E-11 5.lE-12 silb25 Yes 0*

O 4

0-3 7.8E-11 7.8E-11 4.7E-12 nloca52 No 2

0 2-4 0-4 6.5E-11 6.5E-11 6.5E-1I sgtrwl9 Yes 2

0 23 0-4 6.lE-11 6.lE-!!

6.1E-Il silbl5 Yes 1

0 2

0-4 4.lE-11 4.1E-11 2.5E-12 slocwo19 Yes 2

0 23 0-4 4.0E-11 4.OE-11 2.4E 12 sgtrwol7 Yes 2

0 4

4 3.9E-11 3.9E-11 3.9E-11 sgtrw32 Yes 1

0 2-4 0-4 3.7E-Il 3.7E-11 3.7E-11 l

tran19 Yes 2

0 23 0-4 2.9E-11 2.9E-11 1.7E-12 i

l ca e.no.nv.pe,

Page 58 December 30,1996

/

f Table 7-7 i

Success category UC5 (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency if counted as core damage, i

C1 CMT Acc ADS-4 ADS 2,3 A CDF A LRF S sgtrwS2 No 2

0 24 0-4 2.8E-11 2.8E-11 2.8E-11 l

tran32 Yes 1

0 24 0-4 2.6E-11 2.6E-11 1.6E-12 slocwo32 Yes 1

0 2-4 0-4 2.5E-11 2.5E-11 1.5E-12 1

I l

slocaw19 Yes 2

0 2,3 0-4 23E-11 23E-11 1 AE-12 l

sloewoS2 No 2

0 2-4 0-4 1.9E-11 1.9E-11 1.9E-11 slocaw32 Yes 1

0 2-4 0-4 1.4E-11 1 AE-11 8AE-13 tran52 No 2

0 2-4 0-4 1.2E 11 1.2E-11 1.2E-11 slocawS2 No 2

0 2-4 0-4 1.0E-11 1.0E-11 1.0E-11 sgtrwo18 Yes 2

0 4

0-3 7.8E-12 7.8E-12 7.8E-12 sgtrw58 No 1

0 24 0-4 5.8E-12 5.8E-12 5.8E-12 tran58 No 1

0 24 0-4 2.9E-12 2.9E-12 2.9E-12 sloewoS8 No 1

0 2-4 0-4 2AE-12 2AE-12 2.4E-12 slocaw58 No 1

0 2-4 0-4 2.2E-12 2.2E-12 2.2E-12 silb42 No 1

0 23 0-4 1.2E-12 1.2E-12 1.2E-12 P

silb26 Yes 02 0

23 0-4 7.6E-13 7.6E-13 4.6E-14 silb48 No 02 0

2-4 0-4 3.7E-13 3.7E-13 3.7E-13 sgtrwo19 Yes 2

0 23 0-4 63E-14 63E-14 63E-14 sgtrwo32 Yes 1

0 2-4 0-4 5.8E-14 5.8E-14 5.8E-14 sgtrwoS2 No 2

0 2-4 0-4 2.7E-14 17E-14 2.7E-14 sgtrwoS8 No 1

0 2-4 0-4 6.5E-15 6.5E-13 6.5E-15 TOTAL 7.2E-7 7.2E-7 7.6E-8 Notes:

(1)

LRF for scenarios with containment isolation is estimated at 6% of core damage. SGTRs and scenarios without containment isolation increase the LRF by 100% of the core damage frequency.

(2)

Although no CMT injection to the RCS is credited, ADS actuation occurs from the faulted CMT blowing

[

down through the break.

I i

l I

t t

i i

l l

e w.:

r. s..,

Page 59 December 30.1996 I

T6ble 7-8 Success category UC6 (Sorted by Descending Frequency)

Success Path Equipment Assanptions Frequency If counted as core damage, (per year) increase to Focused PRA g

gg gg ACDF A LRF

• nioca08 Yes 2

2 2

0-3 6.6E-8 6.6E-8 4.0E-9 nloca07 Yes 2

2 2

4 5.0E-8 5.0E-8 3.0E-9 sgtrwo8 Yes 2

2 2

0-3 33E-8 33E-8 33E-8 satrw07 Yes 2

2 2

4 2.5E-8 2.5E-8 2.5E-8 sier.wo08 Yes 2

2 2

0-3 2.0E-8 2.0E-8 1.2E-9 Iloca08 Yes 2

2 2

0-3 1.9E-8 1.9E-8 1.1E-9 tran08 Yes 2

2 2

03 1.9E-8 1.95-8 1.1E-9 l

tran07 Yes 2

2 2

4 1.6E 8 1.6E-8 9.6E-10 sloewo07 Yes 2

2 2

4 1.5E-8 1.5E-8 9.2E 10 1

mioca08 Yes 2

2 2

0-3 13E-8 13E-8 7.8E-10 slocawo8 Yes 2

2 2

0-3 1.2E-8 1.2E-B 73E-10 1

mioca07 Yes 2

2 2

4 9.9E-9 9.9E-9 5.9E-10 11oca07 Yes 2

2 2

4 9.9E-9 9.9E-9 5.9E-10 slocaw07 Yes 2

2 2

4 9.2E-9 9.2E-9 5.5E-10 i

cmtibO8 Yes 1

2 2

03 7.2E4' 7.2E-9 43E-10 I

silbO7 Yes 1

1 2

4 6.4E-9 6AE-9 3.8E-10 l

cmtibO7 Yes 1

2 2

4 5.5E-9 5.5E-9 33E10 silbO8 Yes 1

1 2

03 5.0E-9 5.0E-9 3.0E-10 i

cloca15 Yes 2

1 2

0-4 8.6E-10 8.6E 10 5.1E-11 sgtrw15 Yes 2

1 2

04 3.7E 10 3.7E-10 3.7E-10 l

sloewo15 Yes 2

1 2

0-4 2.5E-10 2.5E-10 1.5E-11 l

mloca15 Yes 2

1 2

04 1.6E-10 1.6E-10 93E-12 nioca26 Yes 1

2 2

0-4 1.6E-10 1.6E-10 93E-12 slocaw15 Yes 2

1 2

04 1.4E-10 1.4E-10 83E-12 tran15 Yes 2

1 2

04 1.2E-10 1.2E-10 73E-12 cmtibl5 Yes 1

1 2

0-4 7.7E-11 7.7E-11 4.6E-12 sgtrw26 Yes 1

2 2

0-4 5.4E-11 5.4E-11 5.4E-11 sloewo26 Yes 1

2 2

0-4 4.1E-11 4.1E-11 2.5E-12 I

ew:

_r..,~ a Page 60 Decernbar 30,1996

l l

Table 7-8 Success category UC6 (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency if counted as core damage, (p r year) increase to Focused PRA d

WT h

@S4 MSU A CDF A LRF

• sgtrwo08 Yes 2

2 2

0-3 4.0E-11 4.0E-11 4.0E-11 sgtrwo07 Yes 2

2 2

4 3.5E-11 3.5E-11 3.5E-11 Iloca23 Yes 1

2 2

0-4 3.2E-11 3.2E-11 2.0E-12 mioca26 Yes 1

2 2

0-4 3.1E-11 3.1E-11 1.8E-12 tran26 Yes 1

2 2

04 2.1E-11 2.lE-11 13E-12 slocaw26 Yes 1

2 2

0-4 2.0E-11 2.0E 11 1.2E-12 silb22 Yes 0*

1 2

0-4 7.1E-12 7.1E-12 43E-13 sgtrwo15 Yes 2

1 2

0-4 2.7E-13 2.7E-13 2.7E-13 sgtrwo26 Yes 1

2 2

0-4 4.6E-14 4.6E 14 4.6E-14 TOTAL 3.4E-7 3.4E-7 7.5E-8 Notes:

(1)

LRF for scenarios with containment isolation is estimated at 6% of core damage. SGTRs increase the LRF by 100% of the core damage frequency.

l (2)

Although no CMT injection to the RCS is credited, ADS actuation occurs from the faulted CMT blowing down through the break.

i l

i i

ew r:

. -,am Page 61 December 30.1996 l

^

t a

Table 7-9 Success category UC7 (Sorted by Descending Frequency) l Success Path Equipment Assumptions Frequency If counted as are damage, C1 CMT Acc ADS-4 ADS 2,3 A CDF A LRF ")

Ik>ca09 Yes 2

2 0,1 0-4 3.2E 9 3.2E-9 1.9E 10 11oca24 Yes 1

2 0,1 0-4 4.6E-12 4.6E-12 2.7E-13 TOTAL 3.2E-9 3.2E-9 1.9E 10 Notes:

(1)

LRF for scenarios with containment isolation is estimated at 6% of core damage.

Table 7-10 Success category UC8 (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency if munted as core damage.

C1 CMT Acc ADS-4 ADS 2.3 A CDF A LRF ")

lloca34 No 2

2 3

4 3.1E-10 3.lE-10 3.lE 10 lloca47 No 1

2 2,3 04 4.5E-13 4.5E-13 4.5E-13 TOTAL 3.lE-10 3.lE-10 3.1E-10 Notes:

(1)

Scenarios without containment isolation increase the LRF by 100% of the core damage frequency.

i l

i l

l l

c:we-r.,a.,

Page 62 December 30.1996

i I

Table 711 Success category UC9 (Sorted by Descending Frequency)

Success Path Equipment Assumptions Frequency If counted as core damage, Cl CMT Acc ADS-4 ADS 2,3 A CDF A LRF

• nloca45 No 2

2 3

0-3 5.7E-10 5.7E-10 5.7E-10

)

l sgtrw45 No 2

2 3

0-3 2.4E 10 2.4E-10 2.4E-10 l

slocwo45 No 2

2 3

0-3 1.6E-10 1.6E-10 1.6E-10 nloca46 No 2

2 2

0-4 1.2E-10 1.2E-10 1.2E-10 mioca45 No 2

2 3

0-3 1.1E-10 1.lE-10 1.lE-10 tran45 No 2

2 3

0-3 9.8E-11 9.8E-11

,.8E-11 slocaw45 No 2

2 3

0-3 9.0E-11 9.0E-11 9.0E-11 11oca35 No 2

2 3

0-3 6.9E-11 6.9E-11 6.9E-11 silb37 No 1

1 3

0-3 6.2E-Il 6.2E-11 6.2E-11 emttb37 No 1

2 3

0-3 53E-11 53E-11 53E-11 sgtrw46 No 2

2 2

0-4 4.lE-Il 4.1E-11 4.lE-11 stocwo46 No 2

2 2

0-4 2.9E-11 2.9E-11 2.9E-11 lloca36 No 2

2 2

0-4 2.4E 11 2.4E-11 2.4E-Il mioca46 No 2

2 2

0-4 23E.11 23E-11 23E-11 stocaw46 No 2

2 2

0-4 15E Ii 1.5E-11 1.5E-11 tran46 No 2

2 2

0-4 1.4E-11 1.4E-11 1.4E-11 silb38 No 1

I 2

0-4 8.5E-12 8.5E-12 8.5E-12 cmtib38 No 1

2 2

0-4 7.9E 12 7.9E-12 7.9E-12 lloca37 No 2

2 0.1 0-4 1.9E-12 1.9E-12 1.9E-12 sgtrwo45 No 2

2 3

0-3 2.2E-13 2.2E-13 2.2E-13 sgtrwo46 No 2

2 2

0-4 3.1E-14 3.1E-14 3.1E-14 lloca48 No 1

2 0,1 0-4 2.4E-15 2.4E-15 2.4E-15 TOTAL 1.7E-9 1.7E-9 1.7E-9 Notes:

(1)

Scenarios without containment isolation increase the LRF by 100% of the core damage frequency.

l cw.-

.,- m Page 63 December 30,1996

.. _ _.. _ _ ~

l l

8.0 IDENTIFICATION OF LOW. MARGIN, RISK-SIGNIFICANT SCENARIOS i

he climax of the T/H uncertainty resolution process is identifying the risk-significant, low-margin scenarios that will be further defended with T/H analyses including bounding uncertainties. His section l

documents this process, which sttrts with summarizing results from the UC categories in Section 7.0, and concludes with the definition of the cases for further T/H analysis.

i 8.1 Comparison Method to Focused PRA CDF and LRF Section 7.0 contains discussion of the low-margin categories of success paths from the expanded event trees.' Within Section 7.0, tables of the suxess paths contained information on the increase to the Focused l

PRA core damage frequency (CDF) and large release frequency (LRF) if the path were counted as core damage, it should be emphasized th$ these are sucpgps paths in the Baseline and Focused PRAs.

However, this process considers the pm.sibility that the path is incorrectly categorized as success, and l

should actually be counted as core damage. This allows a determination of the impact that would be seen on the Focused PRA CDF and LRF.

If a success path is counted as core damage, the increase to the CDF is simply the addition of the frequency of that path to the Focused PRA CDF. To determine the impact on the LRF, some estimates had to be made. De cases of no containment isolation and SGTR scenarios are straight-forward, since l

all core damage are assumed to result in a large release to the environment. Thus, the increase to the LRF -

l is the same as the increase to the CDF. If the containment is isolated, lowever, only a fraction of the core damage accidents result in a large release to the environment. The determination of this fraction is done l

by binning core damage accidents into an appropriate PRA accident class, and the sequence frequency is i

multiplied by the containment matrix for the accident class to determine the contribution to the large j

release frequency. The accidents being considered in this T/H uncertainty resolution process, if they l

resulted in core damage, would have minimal core damage which would neither relocate debris to the lower head nor generate significant hydrogen. Based on Level 2 PRA work,it was estimated that 6% of the core damage scenarios with containment isolation could lead to a large release. His is a conservative estimate, overestimating the threat to containment integrity for many of the scenarios.

The impact of counting success paths as core damage was considered for each category. Individual success paths were treated as just described with respect to the determination of LRF, but the entire i

category is considered as a unit when determining risk significance. This is because the UC categcrier, are defined around a specific issue that is common to all the success paths that fit that '.ategory, Therefore,ifit were incorrect to credit success in one success path, this would likewise apply to the other l-success paths with the same conditions defined by the category. Although there are probably exceptions to this rule, it is a conservative ilmitation to apply to the definition of risk significance.

Risk significance for the T/H uncertainty resolution process is defined as increasing the Focused PRA CDF or LRF by at least 1% if the success category were counted as core damage. De at-power, Focused PRA t

c ww e, Page 64 r

Decembado.1996 L

l

/

l i

CDF is 7.7E-6/ year and the LRF is 5.5E-7/ year. 'Iherefore, the cut-off frequency of a success category to determine risk significance is 7.7E-8/ year for CDF and 5.5E-9/ year for LRF.

8.2 Risk Significant Categories The results of the UC categories from Section 7.0 are summarized in Table 8-1, and a determination of whether the category is risk significant is made. The five categories that are risk significant are briefly j

discussed below,in order of their risk significance. As committed to in Section 4.3, LLOCA success paths l-are compared not only to the Focused PRA, but also to the Baseline PRA.

I l

1.

Category UC4 1'

If this category is counted as core damage, the impact on the Focused PRA corresponds to a 14%

l increase in CDF and a 13% increase in LRF. This category consists of the LLOCA initiating l

event with only 1 accumulator. The impact on the Baseline PRA would be approximately an -

order of magnitude larger than the Focused PRA impact. However, since the category is already l

defined as risk-significant, further T/H analyses will be performed, and the magnitude of risk

[

significance is only a concern if acceptable results are not obtained.

i 2.

Category UCS If this category is counted as core damage, the impact on the Focused PRA corresponds to a 9%

l increase in CDF and a 14% increase in LRF. This category applies to initiating events with j

breaks no larger than the NLOCA (6" diameter) with the loss of both accumulators.

3.

Category UC6 l

If this category is counted as core damage, the impact on the Focused PRA corresponds to a 4%

i increase in CDF and 14% increase in LRF. This category applies to all initiating events with the actuation of 2 stage 4 ADS to achieve IRWST gravity injection. The LLOCA success paths,if counted as core damage, would result in an increase of 2.9E-8 to the CDF and 1.7E-9 to the LRF.

l The impact of this change on the Baseline PRA is a 17% increase in CDF and a 9% increase in LRF.

4.

Category UCI If this category is counted as core damage, the impact on the Focused PRA corresponds to a 2%

increase in CDF and 2% increase in LRF. This category applies to NLOCA and DVI line breaks with the failure of both CMTs.

evarsmm=n%w Page 65 December 30.1996

~ -

. ~

~..

~~

j t

I 5.

Category UC2B

?

If this category is counted as core damage, the impact on the Focused PRA corresponds to a 2%

l Increase in CDF and 1% increase in LRF. This category applies to MLOCA and CMTline breaks with the failure of both CMTs.

1 To complete the assessment of the LLOCA impact on the Baseline PRA, other UC categories that are l

applicable to LLOCA need to be examined. The non-risk-significant categories that include LLOCA are l

UC7, UC8 and UC9. With the Baseline At-Power PRA CDF of 1.7E-7 and the LRF of 1.8E-8, the

~

following summary shows the LLOCA Baseline PRA impact for these categories.

j 4

Impact if counted as core damage l

Category Description Baseline CDF Baseline LRF l

i j

UC7 LLOCA 3.2E-9 1.9E-10 O or 1 ADS-4 2%

1%

j Containment Isolated j

UC8 LLOCA 3.1E-10 3.1E-10 DBA ADS

<1%

2%

Containment Unisolated I

i UC9 LLOCA 9.5E-11 9.5E-11

< DBA ADS

<1%

<1%

Containment Unisolated i

I l

Although some of the impacts are 1% or 2% of the Baseline PRA, these LLOCA scenarios are not j

classified as risk-significant. The impact of considering these scenarios as core damage in the Baseline j

PRA will be further discussed in Section 11.0.

t 4

i a

e en-f -"

.#w, Page 66 December 30,1996

f~

Table 8-1 i

Risk Significance of UC Categories Number Initiating Event Defining Equipment If counted as core damage, Risk Conditions increase to Focused PRA Significant?

l ACDF ALRF i

l UCI NLOCA 0 CMTs 1.4E-7 8.2E-9 Yes l

DVI Line Break l

UC2A MLOCA 0 CMTs 1.0E-9 8.lE-Il No CMT Line Break UC2B MLOCA 0 CMTs 1.2E 7 7.5E-9 Yes l

CMT Line Break UC3 MLOCA 0 Accumulators 2.2E-8 1.3E-9 No CMTLB UC4 LLOCA 1 Accumulator 1.1E-6 6.9E-8 Yes UC5 NLOCA 0 Accumulators 7.2E-7 7.6E-8 Yes DVI Line Break SLOCA SGTR Transients UC6 All 2 stage 4 ADS 3.4E-7 7.5E-8 Yes Cont Isolation UC7 LIDCA 0 stage 4 ADS 3.2E-9 1.9E-10 No Cont isolation UC8 LLOCA Cl Failure 3.1E-10 3.1E-10 No UC9 All Cl Failure 1.7E-9 1.7E-9 No

< DBA ADS Notes:

The bold numbers indicate values that are greater than 1% of the Focused PRA CDF or LRF.

T l

l i

l ews _.r.#w, Page 67 December 30.1996

e 2

1 l

8.3 Representative Cases to Address Low-Margin, Risk Significant Scenarios From the five risk significant categories that are defined, a set of cases is defined for TM analyses with uncertainties to complete the TM uncertainty resolution process. De list of risk significant cases is augmented by long-term recirculation considerations discussed in Section 9.0.

First, the risk significant categories are further examined to define representative cases for analysis. His was done by looking at the dominant scenarios in each of the categories. For this purpose, dominant is i

l

' defined as one that contributes to the category CDF or LRF exceeding 1% of the Focused PRA CDF or

]

LRF. He residual effect of all scenarios not identified as dominant for a given category adds up to less l

than 1% of the focused PRA CDF or LRF. The dominant scenarios are listed in Table 8-2.

)

i l

I j

For most categories, the information in Table 8-2 provides a clear definition of the equipment assumptions l

{

for each analysis case. There are two exceptions.

l For categories UC5 and UC6, there are several initiating events in the dominant scenarios and a

{

decision was made to choose the path with the highest frequency, having the largest impact on the risk significance. However,in category UC6, the Baseline PRA impact of the LLOCA event did not cause it to be selected. His is because venting area to achieve IRWST gravity injection is not as challenging for a LLOCA due to the venting capability through the break.

i

)

Most of the categories include dominant scenarios with the failure of some ADS stage 1,2 and 3 lines, yet the expanded event trees are not refined to define the exact number. (In some cases, all possible combinations of stage 1,2 and 3 failures are included.) To balance the desire to be q

conservative from the TM viewpoint with the desire to consider risk significance, it was decided to assume that half of the ADS stage 1,2 and 3 lines function.

The resulting cases for TM analyses with uncertainties are listed in the top pertion of Table 8-3.

ewwpe.,

Page 68 December 30.1996

O s

i I

l r

Table 8-2 Dominant Scenarios in Risk Significant Categories Category Success Path Equipment Assumptions if counted as core damage, increase to Focused PRA C1 CMT Acc ADS-4 ADS 2.3 A CDF ALRF

, sn.;-

c.; - n.,

UCI aloca34 Yes 0

2 4

4 if9.2ETT,,

V5.5E-9f-

$1hh$1 4.869 UC2B cmtib21 Yes 0

2 4

4

.n..

.. n

<. -. ~ -a.;g l

Uc4 Iloca10.11 Yes 2

1 4

2-4 M1.154 s li 6.6E4 W" l

.,..,.0

/ :..l y -.-A

v..:,,.

UC5 silbl0 ll Yes 1

0 4

2-4

$$SS.5L7'ry.9 S: 33L83 E

1 nlocal7 Yes 2

0 4

4 5.6L8 3.459 sgtrwl7.18 Yes 2

0 4

0-4 3.558

$$$853 tran17 Yes 2

0 4

4 1.8L8 1.159 i

w,,...

x....

UC6 aloca07,08 Yes 2

2 2

04 E13.L7,[:!..

%74E4h m:...,, :4 sgtrw07.08 Yes 2

2 2

0 4 5.8L8

1. 5.85.8%

stocwo07 08 Yes 2

2 2

03 3.558 2.159 2

2 2

0-4 3.5L8 2.159 tran07.08 Yes Iloca08 Yes 2

2 2

03 1.968*

1.169

• mioca08 Yes 2

2 2

0-3 1.3L8 7.8L10 l

1 Notes:

Dominant scenarios are defined as ones that contribute to the category CDF or LRF exceeding 1% of the Focused PRA CDF or LRF. The residual effect of aD scenarios eidentified as dominant for a given category adds up to less than 1%

of the Focused PRA CDF or LRF.

Shaded blocks indicate accident scenarios that individuaDy exceed 1% of the Focused PRA CDF or LRF.

(a) Other LLOCA success paths, which are not dominant scenarios based on the Focused PRA impact, increase these values tv 2.968 CDF and 1.7&9 LRF. This is a 17% CDF and 9% LRF increase to the Baseline PRA if they are counted as ccre damage rather than sua:essful core cooling.

l i

enwwepen Page 69 December 30,1996

Table 8-3 Cases for T/K imalysis with Uncertainties Applied Case Break Equipment Assumptions Code lajection Phase Cl CMT Acc ADS ADS

-4 1,2,3 Case UC1 NLOCA

• Yes 0

2 4

all NOTRUMP /

Accumulator LOCTA Case UC2B Largest Yes 0

2 4

all NOTRUMP/

Accumulator CMT LB LOCTA Case UC4 LBLOCA Yes 2

1 4

balf WCOBRAffRAC Accumulator Case UC5 DVI LB

• Yes 1

0 4

balf NOTRUMP/

Accumulator /

LOCTA IRWST Inject Case UC6 NLOCA

• Yes 2

2 2

balf NOTRUMP/

IRWST Inject LOCTA Limiting break size to be determined by MAAP4 analyses when MAAP4 benchmarking is completed.

l 1

ewwwwp.t-p Page 70 Decemtur 30,1996

i 9.0 IDENTIFICATION OF RISK SIGNIFICANT LONG-TERM RECIRCULATION CASES To be done.

16 0 T/H ANALYSES OF LOW-MARGIN, RISK-SIGNIFICANT SCENARIOS 10.8 Assumptions for T/H Uncertainty Analyses 1

i To be done 10.2 NOTRUMP Results i

To be done.

10.3 WCOBRAffRAC Results i

To be done.

11.0 ASSESSMENT

OF T/H UNCERTAINTY RESULTS ON PRA To be done.

12.0 CONCLUSION

To be done.

13.0 REFERENCES

'l.

NSD-NRC-96-4796/DCP/NRC0576, Docket Number STN-52-00?., Letter from Brian McIntyre (Westinghouse) to T. R. Quay (NRC) on "AP600 Passive System Reliability Roadmap," 8/9/96.

l I

)

ewr.p6memoc nv. pew, Page 71 December 30,1996 i

l 1