ML20116E227
| ML20116E227 | |
| Person / Time | |
|---|---|
| Site: | 05200003 |
| Issue date: | 07/31/1996 |
| From: | WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP. |
| To: | |
| Shared Package | |
| ML20116E214 | List: |
| References | |
| NUDOCS 9608050154 | |
| Download: ML20116E227 (27) | |
Text
-
~
s e
T/H Uncertainty Resolution Process for AP600 Passive System Reliability 9
July 1996 4
t 4
4 e
0 e
9608050154 960729 PDR ADOCK 05200003 A
a Table of Contents P,ge, 1.0 I NTRO D UCTIO N........................................
I 2.0 DEFINITION OF T/H UNCERTAINTY.......................
1 3.0 RESOL UTION PROCESS....................................
2 3.1
- Types of Core Uncovery............................
4 3.2 Expanded PRA Event Tree Methodology....................
10 3.3 Focused PRA impact..................................
15 3.4 Risk Significant Scenarios.............................. 22 3.5 T/H Assumptions for Risk Significant, Core Uncovery Scenarios... 24 3.6 T/H Analyses...................
25 25 4.0
SUMMARY
\\
l I
I 1
i i
1
j 4
i
1.0 INTRODUCTION
i 1
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 were grouped into categories'. Each category consists of a group of functioning equipment and a range of break sizes and location. Within'each category, baseline cases were identified. Baseline cases J
f were chosen to be the most limiting break size, location and set of equipment to bound the group of cases.
i a
Analyses of the baseline cases were performed with nominal assumptions, rather than conservatisms that I
are typical of design basis safety analyses. 'Ihe 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 j
different systems. An issue has been raised on whether the consideration of uncertainty in the analysis l
would significantly affect the conclusions of the PRA.
l
'Ihis document defines the process that will be used to resolve this issue. 'Ihe, process is described with specific examples and numerical values. However, all numerical values are preliminary, and should be
~
l diewed only in context of helping to confirm the feasibility and validity of the process definition.
j j-2.0 DEFINITION OF THERMAL-HYDRAULIC (T/H) UNCERTAINTY
(
l.
The term "TM Uncenainty" is used in relationship to predicting the behavior of passive systems in AP600.
I Because of the passive nature of the safety-related systems in AP600 and the reliance on small AP's, the concern is that uncenainties 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 l
due to different accident conditions, or uncertainty in analytical models. Specific sources of TM uncenainty that have been identified as potential concerns are:
I initial and boundary conditions, e
code. uncenainty (based on testing and scaling uncenainties),
e user selected inputs and modeling J ethods.
=
It must be shown that the consideration of TM uncertainties does not significantly impact the PRA results.
Funhermore, because the concern is passive system reliability, the Focused PRA (that does not include j
active systems) is the standard for comparison and determination of 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.
As described in the following sections, however, the TM uncenainty resolution process does not quantify the sources of uncenainty, nor is it solely a TM analysis exercise. Rather, the TM uncenainty resolution
'l comewr.e a Page I w : im l
process will identify a set of low margin, risk significant accident scenarios, and will show acceptable T/H
& performance when the u'ncertainties are bounded.
3.0 RESOLUTION PROCESS The T/H uncertainty resolution process requires a combination of information that can be obtained from the PRA and from T/H analyses. The PRA directs attention to accident scenarios that are most probable.
T/H analyses direct attention to accident scenarios that most greatly challenge core cooling. The T/H uncertainty resolution process finds the intersection of accident scenarios that are both relatively probable and which challenge core cooling.
This section defines the process that will be implemented to bring the T/H uncertainty issues to closure.
Figure 1 provides a high-level overview of that process. The six parts of the process correspond to the following six sub-sections, where the process is explained.
B 9
9 5
9 6
a luty l9, l#6
FIGURE 1 T/H UNCERTAINTY RESOLUTION PROCESS PART 1 PART 2 IDENTIFY TYPES OF CORE EXPAND EVENT UNCOVERY AND TREES AND WHICH INITIATING QUANTIFY EVENTS THEY SUCCESS PATHS APPLY TO I
I PART 3 IDENTIFY IMPACT ON FOCUSED PRA IF CORE UNCOVERY IS COUNTED AS CORE DAMAGE PART 4 IDENTIFY RISK SIGNIFICANT SCENARIOS PART 5 IDENTIFY KEY T/H ASSUWPTIONS FOR RISK SIGNIFICANT, CORE UNCOVERY SCENARIOS PART 6 FOR EACH RISK SIGNIFICANT SUCCESS CRITERIA CORE UNCOVERY SCENARIO,15 YES NOTRUMP/LOCTA PCT ARE PROYED TO BE
=
< 2200 DEG F RIGOROUS WITH BOUNDING KEY T/H ASSUWPTIONS?
NO U
OPTIONS q
ASSESS IMPACT OF ASSESS COUNTING SCENARIO AS ALTERNATIVE CORE DAMAGE IN THE OPTIONS BASEUNE & FOCUSED PRA
%3
._..m..
i a
i' JJ Tynes of Core Uncovery Part 1 of the T/H uncertainty resolution process is to identify types of core uncovery and which initiating
)
events they apply to. The purpose of this step is to develop a list of PRA accident scenarios that are
{
closest to the limity of acceptability, and thus would be most susceptible to T/H uncenainty having an l
impact on the conc.usions of success versus core damage.
Core uncovery.ts defin-d ra the predicted coolant two-phase mixture level falling below the top of the i
core. The occurrence of core uncovery is used only as a screenina criterion for an accident scenario to l
be further considered within the T/H uncenainty resolution process. The acceptance criterion for I
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 criterion for LOCAs..
The process of identifying the types of core uncovery extends from the same process that was used to develop the PRA PIRTs to support the MAAP4 / NOTRUMP benchma-king effort. To develop the PIRTs, a spectrum of PRA scenarios were examined by a group of exp with experience in AP600 systems design, small-break LOCA analyses, PRA and PIRTs. Key thermal-hydraulic phenomena which could impact chalkrges to core coolant inventory were identified (with an "H' for high importance).
These same challenges can also be defined in terms of the equipment loss that causes them to occur.+
Figure 2 shows four equipment losses that can lead to core uncovery: 1) loss of both CMTs,2) loss of accumulator (s),3) loss of ADS lines, and 4) loss of containment isolation. Figure 2 also shows the time period of core uncovery (early or later), the problem that defines the type of core uncovery, and the initiating events for which each type of core uncovery applies. Only core uncovery that is. credited as successful core cooling in the PRA is addressed through this process.
To date, core uncovery has been classified into six types, as shown on Figure 2. 'the identification of the occurrence of core uncovery has been based on the preliminary MAAP4 analyses, and preliminary MAAP4 / NOTRUMP benchmarking effons that are in progress. Because the benchmarking is not complete, the types of uncovery and the impacted events may be refined as the benchmarking proceeds.
In addition, this process will be extended to apply to the long term recirculation accident phese, which has not previously been analyzed with MAAP4. There are current efforts to provide additional analytical suppon for the PRA long-term recirculation success criteria. Through these efforts, any instances of core uncovery will be identified, and they will be treated in the same manner as the currently identified six types. Each of the six types of uncovery are discussed belo.w.
cwpupw w ufreg2 e Page 4 My:9tm
FIGURE 2 TYPES OF-CORE UNC VERY
~
~
l
' Time EARLY LATER LONG-TERM Period UNCOVERY UNC0VERY RECIRCULATIDN Loss of Loss of Controttmo Loss of Loss of ADS Contamnent EgApnent Both CMTs Ac w a M s)
Lines Isolation tosa i -
i -
No Make-up Accunulator No Rapid Reduced No More DeFrecult Inventory Depletes Inventory Inventory Make-up to Achieve and Probten +
W RCS Prior to Make-up Make-up Vhen Mainteen Detta P Pressure a Operator During During ADSis For IRVST
> 700 psa0 Interventaon Bloedoen Rettood Actuated Gravtty Injection Initiating i
i I
.I I
I EDCA EDCA MLOCA LLOCA EDCA TED t
+
D'/1 Lane Break CMT Line CMT Lane SLOCA E+. agge StiTR Cora Uncovery Transeents e
l
%.y 5
j I
e J
CORE UNCOVERY TYPE 1:
No Make-up Inventory if RCS Pressure > 700 psig
'Be first type of core uncovery occurs in 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 below this point, there is the potential for core uncovery due to the lack of make-up water.
De potential for this type of core uncovery is also impacted by operator action time. De question to be considered is whether core uncovery occurs prior to the break depressurizing the RCS below 700 psig a!!,
d before the operator manually opens ADS lines. With operator 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 3.
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. De period of core uncovery ends when the operator opens ADS lines, allowing the accumulators to inject rapidly.
The LOCA 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. Larger breaks depressurize so that the accumulator (s) can inject prior to core uncovery.
CORE UNCOVERY TYPE 2:
Accumulator Depletes Prior to Operator Intervention De second type of core uncovery also occurs in 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, accumulator inventory may deplete prior to operator action to open ADS., his 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 3.
De corresponding initiating events are Medium LOCAs (MLOCAs) and CMTline breaks. Larger breaks do not rely on ADS lines opening to achieve gravity injection since the break will depressurize the RCS to IRWST injection. Derefore, ahhough core uncovery may occur in a larger break scenario, it is not categorized as this type.
Another consideration that impacts whether core uncovery occurs is the number of accumulators that inject. De PRA success criteria for scenarios without ClWTs is at least I accumulator injects. If only I accumulator successfully functions, then the period of core uncovery could be on the order of 10 minutes for the largest MLOCAs. If both accumulators successfully function (which is also a part of the
' success criteria for the impacted events), then the period of core uncovery is significantly reduced.
c w n c.n e Page 6 w :t w
i.
FIGURE 3 LOSS OF CMTs RESULT IN CORE UNCOVERY TYPES 1 AND 2 i
1 j
30 4
i CORE
\\ / UNCOVERS
's (TYPE 2) i i
\\
1 s
j
's
's
's
's 2
o 20 s,
s, 3
's,
's,
.3
's, E
's
- 1:
v
~ ACCUMUIATOR';
's's CORE UNCOVERS o
e g
(TYPE 1) 10 i
t
[.
i l
- SLOCA ~1 :
NLOCA l
MLOCA r
i O
i 0
2 6
9 j
Break Equivalent ID (inches) ia-y
r e
CORE UNCOVERY TYPE 3:
No Rapid Inventory Make-up During Blowdown The third type of core uncovery occurs in scenarios with the failure of both accumulators. The 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 (below 9" diameter), the PRA success paths do not require any accumulators if at least 1 CMT functions. The CMT, although a similarly-sized large tank of water, does not provide the rapid make-up capability. Therefore, 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.
CORE UNCOVERY TYPE 4:
Reduced Inventory Make-up During Re-Flood The fourth type of core uncovery occurs in Large LOCAs (LLOCAs) due to the high rate of inventory j
loss from the break. LLOCA is a design basis accident (DBA) analyzed and documented in Chapter 15 l
of the SSAR. The DBA scenario includes 2 accumulators, and core uncovery occurs dine to the large inventory lose through the break. The success of this accident scenario has been demonstrated, including conservative assumptions,.and is no' subject to further investigation in this T/H uncetainty resolution t
process. Hov aver, the LLOCA success criterion for the PRA only requires i accumulazoi. The LL.OCA scenario with I accumulator has been analyzed using _W_COBRATIRAC with nominal T/H exlitions to support the PRA. The analysis results show that the PCT during reflood, which is the period of time influenced by the accumulators, is less than 2200*F. The plant response to only I accumulator is a type of core uncovery that is to be considered within the T/H uncenainty resolution process.
CORE UNCOVERY TYPE 5:
No Make-up Inventory When ADS is Actuated I
)
The fifth type of core uncovery is also due to the loss of accumulators. The third and fourth types of core uncovery were associated with th'e accumulators and their ability to provide rapid make-up for medium and large breaks. The fifth type of core uncovery. completes the examination of the effect of losing accumulators for the remaining initiating events.
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 pressurizer 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 is opened, and the mass lost through the ADS is high. Accumulators provide rapid inventory make-up for this condition. H ) wever, if both accumulators fail, thermal-hydraulic analyses show that core uncovery l
can occur. This type of core uncovery applies to NLOCA, SLOCA, SGTR and Transients. The applicable
< w w=.nw Page 8 I
waw
SGTR and Transient scenarios include the loss of other mitigation systems.
l CORE UNCOVERY TYPE 6:
More D!mcult to Achieve and Maintain AP for IRWST Gravity Injection The sixth type of core uncovery occurs due to a delay in achieving IRWST gravity injection. He delay in IRWST gravity injection, relative to design-basis scenarios, is due to a loss of ADS lines or a loss of containment isolation.
IRWST gravity injection starts when the AP between the RCS and the containment falls below approximately 15 psi. He loss of ADS lines reduces the capacity for venting steam and makes it more difficult to lower the RCS pressure. He loss of containment isolation results in a lower containment pressure. Either the loss of ADS venting capability or the loss of containment isolation causes a higher AP between the RCS and containment, delaying the start of IRWST injection and potentially leading to core uncovery.
His type of core uncovery can potentially occur for any in'uatin;; nent that relies on ADS actuation to achieve IRWST gravity injection. The loss of ADS lines, especially if coupled with loss of containment, isolation, can lead to core uncovery or even core damage. The ADS success criteria that specify the -
number of lines required to consider an accident scenario as success have evolved over the past few years.
The current revision of the PRA considers that at least 2 stage 4 ADS must work. Further investigations must be done to determine which break sizes result in core uncovery for this success criterion. In addition, it must be determined how many more ADS lines are needed to prevent core unco.very. His investigation will be completed after NOTRUMP validation and the MAAP4 / NOTRUMP benchmarking tasks are done.
9 e
$\\
YQ Y M y 29.t*6
a I
M Expanded PRA Event 'lree Methodolony Part 2 of the T/H uncertainty resolution process is to expand the Focused PRA event trees and to quantify the frequency of the success paths. Quantification of success paths is not normally done in a PRA, since core damage is the focus. The purpose in quantifying the frequency of success paths for T/H uncertainty resolution is to gain perspective on the relative probability of specific success scenarios. His information will ultimately be used to define risk significant scenarios that could be impacted by T/H uncertainty, as will be discussed in Sections 3.3 and 3.4.
I
" 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 4
j and successes. As an example, Table A identifies the functioning equipment that are included within a j
single success path (MLOCA, with CMTs) in the Focused PRA. Table A also identifies the equipment j
assumptions that are made in the corresponding accident analysis that supports the success path.
j i
4 Table A
}
Comparison of Equipment on Event Tree Success Path to
{
Equipment Assumptions in Supporting Analysis Equipment Rat May Function for Baseline Scenario Used for j
Success Path I on MLOCA Event Tree in PRA Accident Analysis Focused PRA l
j 1 or 2 CMTs 1CMT 0, l' or 2 stage 1 ADS.*
O stage 1 ADS j
0, I or 2 stage 2 ADS
- 0 stage 2 ADS j
0.1 or 2 stage 3 ADS
- O stage 3 ADS l
2,3 or 4 stage 4 ADS 2 stage 4 ADS
)
0,1 or 2 accumulators O accumulators j
1 or 2 IRWST injection lines 11RWST line 1 recirculation line g I recirculation line
~
2 Success or failure of containment isolation
- Failure of complete containment isolation j
f Not broken out by a top event question, but implicit within scenario possibilities.
f As shown in Table A, the scenario that is used in the accident an'alysis to represent a specific success path is the most pessimistic set of functioning equipment for that path. Even if the baseline scenario accident analysis shows core uncovery, there are many other accident scenarios (or sets of functioning equipment)
{
represented by the same success path that may not result in core uncovery. Therefore, the success paths i
ww.n%: a Page 10 n a na
)
a on the event trees need to be expanded to show the various equipment success combinations to obtain an estimate of frequency for core uncovery scenarios, and to find those scenarios for which T/H uncertainty
'could have an impact.
3 There are many options of exactly how to expand the success paths on an event tree. 'Ihere are four key elements to the method that was developed to perform the expansion.
1.
There are many top level events that could 'oe used to ask questions and further refine the success paths. Table B summaraes 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 arc c presented within an existing success path in the Focused PRA Sensitivity Study. All core damage paths on the expanded event tree are core damage paths in the Focused PRA.
8.
No success path was expanded beyond 3 system failures. In general, each failure decreases the 2
frequency of a path by a factor of 10, so that a path with at least 3 failures has a frequency on the order of 10* per year. There is no need to further decompose paths to define frequencies into the 10" versus 10" range. 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.
4.
Top events were arranged in an order to minimize the number of paths.
An example of an event tree (MLOCA) in the Focused PRA is given in Figure 4, and an example of how it was expanded is in Figure 5. The same process will be used to expand success paths for the following Focused PRA event trees:
LLOCA-MLOCA (done as example)
CMT Line Break
=
DVI Line Break
=
NLOCA SLOCA SGTR that require ADS Transients that require ADS rwima 2 Page 11 w a im
1 i
Table B l
Options for Expanding Event Tree Success Paths Option Used?
Reason l
l Break size No Break size and location are already used to define different i
I initiating events. Although within an initiating event there remains some variability in the plant response depending on the i
Break location s ze and location of the break, there was no added value to iunher refinement.
Number of CMTs Yes Whether there is 1 or 2 CMTs does not make a significant
' difference in the course of the accident progression. However, l
l the CMTs are highly reliable, and make an important contribution to the refinement of the frequency of a given 3
accident scensaio. That 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 so all, and do.not significantly impact the course of the accident progression.
j 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.
i Number of IRWST lines No The ability'to achieve IRWST gravity injection and longterm rectreulation is most dependent on the number of open ADS lines and whether the containment is isolated. The number of Number of recirculation lines lines open, as long as there is a pathway for injection, is pot 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 calcul& tion if the accident scenario is counted as core damage.
c.wame: un Page 12 w :9. im
l
_ 1 i
,\\
l I
Figure 4 MLOCA Event Tree in Focused PRA l
l 1 or 2 1 OK j
l 1 or 2 RECIRC l
l 2,3 or 4 IRWST 0
2 CD 1 or 2 ADS-4 0
)
3 CD 4 CD 1 or 2 5 OK I
MLOCA 1 or 2 RECIRC 1
)
1 or 2 IRNST 0
2,3 or 4 Acc 0
7 CD l
0 ADS-4 0
8 CD 1
0 or 1 9 CD l
l l
l OK = Successful Core Cooling CD =, Core Damage 1
i i
i j
i l
1 comim,w 2.
Page 13 l
w :9. i m
1 Figure 5 Expanded MLOCA Event Tree (With Frequencies of Success Paths)
(
CT, W aq-astem At24 --
A362.3 M 4
4 1
114LL 3 35-9
- l'?, Pi c :,an y #
H... L *"-n;:1 60':,'
, t.
so i f-4*
4 i
Liat L. se at-7 1
i to a-
,a sen 3
is )E-1
,J.LL2. it st-eo 1'
j l#
2 es 9t-iI l0 w L tend DamA6T g
4 4195-1 i
I WJAL-es af *1 le i
ia 5
-,,.3e..
t newL yp g as SE-1 4
I M 33 Ss-S
- 23 4 t-80 l
34 bG*tO LaLLL as i t-*o
. ag, 4t -il
- ^ -
- w a m ase 4
- at af-i I
Eq qr-so l'
Io., i se as -45,,,,ma
,, e,,, t
,.1E
,"."1.
w 3E-tt isn
,, p
$4 tt-6 4
% S SC-1 1
S is,s 1
I 3 q,,,g t a ae L
_ g, gy,g m
[,
,1 5 -*
- 3g i g-10 0
n a., t gitees W n
no ther Da*****
9 46 2E-7 t'M 4:4 r-s 4
43 3e.lO 3
44 $8-s0 h45 tr 'O 44, as-el o*i
_ 4 y w tun **A88 i
M 49 2E4 I
h43 Se-io a
.1 L'"*
so se-ia j
_gi ends M 8 L
~o g.a.s 93 se.ie a'
93 come omana*8 t
4
$4 gg.40 "W
f
[
LLIA1-99 9E-88 a
[
'aw n
-56 18-12 Mie L 51CA8f D"d**8' l elo L#
nioa g g.g
_ a_, a-99 g,,, A 60 M *ll t., a i'**i-s,l tear TWa84f lo 63 tear smaa**f c **v s thncern
- del esas M J wly
.f., I M
%l
g M
Focused PRA Impact i
Part 3 of the T/H uncenainty resolution process is to identify the impact on the Focused PRA, if core uncovery scenarios are counted as core damage. he purpose of this step is to tie together the following information:
accident scenarios that could be impacted (success versus core damage) by the consideration of
=
T/H uncertainty (from Part 1), with the frequency of those accident scenarios (from Pan 2).
j lt should be noted that the scenarios being discussed are counted as success in the Baseline and Focused PRAs. Part 3 of the T/H uncenainty resolution process introduces the concept of considering the PRA l
impact if the scenarios were not success. Because these are core unco ~very scenarios (based on nominal conditions), it is theorized that the consideration of T/H uncenainty could lead to core damage. Whether or not this is of concern is dependent on the probability of the accident scenario. For higher frequency scenarios we must be more cenain that a success scenario does not result in core' damage, regardless of the range of possible T/H conditions. Pan 3 starts to bring together the necessary information that will provide the confidence of successful core cooling. Yet Part 3 is still only an information gathering stage.
The results of this process are summarized and assessed in Pan 4 to determine risk significance.
Tne work in Part 3 will be documented through a set of tables. Examples are provided of Tables 1 j
through 5, which correspond to core uncovery types I through 5. Thes: tables are incomplete and the numerical values are all preliminary estimates, but the examples establish the format and illustrate how the process will be implemented.
As an example, consider the MLOCA entries on Table 3. Table 3 addresses core uncovery type 3, which is discussed in Section 3.1. This type of core uncovery occurs in scenarios with at least 1 CMT, but no, accumulators. The MLOCA expanded event tree (Figure 5) was examined to find all success paths with this set of equipment. There are six success paths, and the equipment assamptions of each path are summarized on the table, along with'the frequency of the event. The final two columns on the table identify the impact on the Focused PRA if the scenario were counted as core damage. The Focused PRA impact is done by adding the frequency of each path to determine the potential impact on the Core Damage Frequency (CDF). The method for determining the impact on the Large Release Frequency (LRF) is explained below.
If there is no containment isolation (i.e., containment isolation failed), all core damage would resu.It in a large release to the environment. Thus, for no containment isolation scenarios, the increase to the LRF is the same as the increase to the CDF. If the containment is isolated. however, only a fraction of the core womac.n%2m Page 15 w : im
damage accidents result in a large release to the environment. The determination of this fraction is done
,by binning core damage accidents into an appropriate accident class, and the sequence frequency is multiplied by the containment matrix for the accident class to determine the contribution to the large release frequency. The accidents being considered in this T/H uncertainty process, if they 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. This is a conservative estimate.
overestimating the threat to containment integrity for many of the scenarios. Thus, the impact on LRF on Tables 1 - 6 for scenarios with containment isolation is determined by multiplying the CDF by 0.06.
I 1
i
, wwwm2.
Page 16 w :um
i l
Table 1 No Make-up Inventory if RCS Pressure is Greater than 700 psig Event Equipment Assumptions Success Frequency If counted as core Path (per year) damage, increase to CI CMT Acc ADS-4 ADS Focused PRA 2,3 A CDF A LRF l
NLOCA Yes 0
2 4
J 0
2 4
<4 IE-8 gMd 0
2
.23 54 2E-10 0
1 2.3.4 14 7E-10 No 0
1.2 2.3.4 14 IE-10 IE 10 IE-10
)
DVI Line '"
Ies 0
1 4
0 1
4
<4 IE-9 O
I 3
$4 2E-Il j
0 1
2
$4 4E-13 l
No 0
1 23.4 54 IE 11 IE-11 1E-11 Notes:
(1)
LRF for scenanos with containment isolation is estimated at 6% of core damage.
(2)
The expacded event tree for DVI Line Break includes other paths without CMTs that are not listed on this table. These other paths consider the scenanos where the CMT on the broken DVI line generates an automatic ADS actuatton signal. The automatic ADS actuarmo occurs quickly, and there is not a penod of time waiung for operator action that results in the core uncovery addressed on this table.
i i
c.w. o-nwm2..
Page 17 wy :n. im
I Table 2A Accumulator Inventory Depletes Prior to Operator Intervention (1 Accumulator)
Event Equipment Assumptions Success Frequency If counted as core Path (per year) damage. increase to Cl CMT Acc ADS-4 ADS Focused PRA 2.3 A CDF A LRF MLOCA Yes 0
1 2.3.4
No 0
1 2.3,4 54 60 '"
< 3E.11
< 3E-11
< 3E-l1 CMT Line
%N Notes:
(1)
LRF for scenarios with containment isolation is estimated at 6% of core damage.
(2)
Event tree is not decomposed to separate i versus 2 accumulator on this success path. The scenario with I accumulator is a small part of the path frequency.
1 Table 28 Accumulator Inventory Depletas Prior to Operator latervention (2 Accumulators)
Event Equipment Assumptions Success Frequency If counted as core Path (per year) damage. increase to Cl CMT Acc ADS 4 ADS Focused PRA I
2.3 A CDF ALRF MLOCA Yes 0
2 4
4 34 2E-8 2E-8 IE-9
- i 0
2 4
<4 35 3 E-9 O
2 2.3
$4 36 4E-Il No 0
2 2.3.4 54 60
- 3E 11 3E-Il 3E Il CMT Lin'e
_ _ a agN yytw ~
Notes:
(1)
LRF for scenario <, with containment isolation is estunated at 6% of core damage.
(2)
Event tree is not decomposed to separate i versus 2 accumulator on this success path. The scenario with 2 accumulators is the dommant part of the path frequency.
nom,%m.w,y: un Page 18 w :4 im
j a
i i
e l
l i
Table 3 No Rapid Make-up Insentory During Blowdown Event Equipment Assumptions Success Frequency If counted as core Path (per year) damage, increase to CI CMT Acc ADS-4 ADS Focused PRA 2,3 A CDF aLRF MLOCA Yes 2
0 4
4 17 9E-9 IE-8 7E.10 "'
2 0
4
<4 18 2E-9 2
0 2,3 14 19 3E-Il 1
0 2.3.4
$4 32 2E-Il No 2
0 2,3,4
0 2.3,4 54 58'"
< 4E 12 gi CMT Line
,g
.ft,W Notes:
(1)
LRF for scenarios with contamment isolation is estunated at 6% of core damage.
(2)
Event tree is not decomposed to separate 0 versus I accumulator on this success path. The scenario with no accumulators is a small part of the path frequency, l
l l
l l
l I
com.%-.nem2 e Page 19 w :n im 1
o Table 4 Reduced Inventory Make-up During Re-flood Event Equipment Assumptions Success Frequency If counted as core Path (per year) damage.~ increase to CI CMT Acc ADS-4 ADS Focused PRA 2,3 aCDF ALRF i
LLOCA Yes 2
1 4
4 8E.7 9E-7 6E 8 'h i
2 1
4 2.3 IE 7 1
2 1
4 0.1 IE-9 2
1 3
4 2E-9 2
1 3
<4 3E 10 l
2 1
2
<4
' 6E-11 1
1 4
4 2E-9 1
1 4
<4 3 E-10 1
1 2,3.4 54 IE 13
(
No 2
1 4
54 IE-9 IE-9 IE-9 2
1 4
<4 2E 10 i
l 2
1 2.3 54 3E-12 l
1 1
2.3,4 54 4E-12 l
Notes:
(1)
LRF for scenanos with containment isolation is estimated at 6% of core damage.
i i
I l
l l
i i
l o.riw.,w,w2.
Page 20 sa:ai*.
l l
4
~
]
0 Table 5 No Make-up Water When ADS is Actuated Event Equipment Assumptions Success Frequency If counted as core Paths (per year) damage, increase to Cl CMT Acc ADS-4 ADS Focused PRA 2.3 aCDF ALRF NLOCA Yes 2
0 4
l 2
0 4
<4 SE-9 F
egow,t#8
-<4 IE-10 2
0 23 1
0 2.3.4
<4 1E.10 No 2.
0 2.3.4
0 2.3.4
<4
< 2E-Il
. SLOCA SGTR Transients Notes:
(1)
LRF for scenarios with containment isolation is estimated at 6% of core damage.
(2)
Event tree is not decomposed to separate 0 versus I accumulator on this success path. The scenario with no f
accumulators is a small part of the path frequency.
i 1
1 com.% w.2.
Page 21 July 26. IW6
- '. o J.4 Risk Sinnificant Scenarios Pan 4 of the T/H uncertainty resolution process is to identify the risk significant low margin scenarios.
This is done by summarizing and sorting the information collected in Part 3.
Table X is provided as an example of the format that will be used to collect the information. The scenarios are now " summarized" so that they no longer refer to one specific path on an expanded event tree. The equipment assumptions in the summarized scenario include all the options from paths that had a high enough frequency to impact the CDF listed on the table. He scenarios are then sorted by their potential contribution to CDF, if they were counted as core damage. The LRF is also tracked to ensure that the highest contributors to LRF are not ignored.
De scenarios with the highest frequency will be defined as risk significant and are chosen for further T/H analyses to address T/H uncertainties. His will be done by comparing the top scenarios to the Focused PRA CDF and LRF. De scenarios that influence the results by more than 1% will be selected for gnalysis in Part 6 of the T/H uncertainty resolution process. Although the Focused PRA is being updated 4
at this time, the CDF is expected to be on the order of IE-5.
9 i
i a
1 I
i i
i
}
a f
l c.ww%2 a Page 22 w :o.s m r
a o
Table X (PARTIAL)
SUMMARY
OF CORE UNCOVERY SCENARIOS THAT ARE SUCCESS IN FOCUSED PRA (Sorted by Contnbution to CDF)
Event Equipment Assumptions Type of Core If counted as core Uncovery damage, increase to CI CMT Acc ADS-4 ADS 2,3 Focused PRA ACDF A LRF LLOCA Yes 2
1 4
2.3.4 4
9E-7 6E-8 1
NLOCA Yes 0
2 4
54 1
0 4
$4 5
5E-8 3E-9 P
MLOCA Yes 0
2 2.3.4
$4 28 2E-8 IE-9 MLOCA Yes 2
0 4
$4 3
IE 8 7E.10 DVI Line Yes 0
1 4
s4 1
6E-9 4E 10 LLOCA No 2
1 4
$4 4
IE-9 IE-9 MLOCA Yes 0
1 2.3.4 14 2A 2E-10 9E-12 i
NLOCA No 0
1.2 2.3.4 54 1
NLOCA No 2
0 2.3.4 54 5
9E-Il 9E-ll j
MLOCA No 0
2 2,3,4 14 2B 3E-Il 3E-Il MLOCA No 0
1 2.3.4 54 2A
< 3E 11
< 3E-il MLOCA No 2
.0 2,3,4 s4 3
2E 11 2E-11 l
DVI Line No 0
1 2.3.4
$4 1
PRE,dU e
e.m s4nuncenwreo2 u.
Page 23 luiv :A im
1
)
4
,\\,
0 o
5 T/H Assumptions for Risk Sinnificant. Core Uncovery Scenarios J25 Part 5 of the T/H uncertainty resolution process is to identify key T/H assumptions for risk significant, core uncovery scenarios. He goal is to define "DB A-like" analyses, except the equipment assumptions will be from the results of Part 4 of the T/H uncertainty resolution process.
To assess the impact of the uncertainties, the analysis will be performed using the approved Appendix K version of the NOTRUMP code. Using the Appendix K requirement is a conservative method of bounding the uncertainties for LOCA analysis, which has been confirmed by the 1988 Appendix K rule
]
change and the recent approval of WCOBRA/ TRAC for best estimate large break analyses. NOTRUMP calculations will be performed for the limiting scenarios with the Appendix K assumptions to determine if an acceptable PCT of less than 2200*F occurs. He DBA analysis type of plant modeling and assumptions will be used for the calculation to be consistent with the Appendix K approach used for AP600 DBA analysis. The difference will be the assumption of increased component or system failures which make the accident beyond design basis.
For small break LOCA, scenarios with multiple failures reduce injection whichi can then result in a predicted core uncovery and associated heat-up. De application of the NOTRUMP code for this situation is expected to be straight forward since the acceptance criterion is the same as the design basis transients ~
(PCT s 2200*F). In other words,'if the PCT exceeds 2200*F, the scenario is considered to result in cor'e i
damage; NOTRUMP is not being used as a basis for claiming successful core cooling beyond a PCT of 2200*F. Berefore, the phenomena ofinterest for the multiple-failure scenarios is the same as that of the.
l design basis transients, and NOTRUMP is applicable.
The key thermal-hydraulic phenomena is the resulting two-phase mixture level in the core. The NOTRUMP two-phase level swell models have been validated over a range of pressures and indicate that,.
in general, NOTRUMP predicts a conservatively low mixture level relative to the test data. Herefore, the calculation of the level swell with NOTRUMP.will be conservative, resulting in a conservative calculated PCT.
w mn a.o.
Page 24 seya:=
je, o a 1
16 TM Analyses 1
Part 6 of the TM uncenairty resolution process is to perform TM analyses for each risk significant, core uncovery scenario using bounding T/H assumptions.
l For each accident that is identified as a risk significant core uncovery scenario, supporting analyses will be performed with a detailed computer code, such as NOTRUMP / LOCTA or,WCOBRAfrRAC. For each accident scenario, one bounding analysis will be done in which the methodology identified in Part 5 is implemented. If the analysis results show that the PCT remains below 2200*F, the success criteria will have been proven to be " robust." No further TM uncertainty analyses will be performed.
j
' However, if the PCT exceeds the acceptance limit of 2200*F for any of the risk significant core uncovery
)
scenarios, there are options of how to address this outcome. Westinghouse will choose between these options based on the details of the accident scenario and the TM analysis results.
Assess the impact of counting the scenarios as core damage in the Baseline and Focused PRAs.
Tae PRA results will be modified if the impact is significant.
Identify and assess alternative options.
4.0
SUMMARY
This document has presented the details of a process plan to resolve outstanding TM uncertainty issues that are a part of passive system reliability. The plan identifies scenarios based on their impo'rtance to the Focused PRA and the potential that the consideration of TM uncertainty could impact the conclusion of success versus core damage. It is expected that only a few scenarios will be identified that meet these requirements. For the scenarios that are identified, a conservative analysis will be performed to bound the TM uncertainties.
cw1wega Page 25 u,, a im
_.