ML20116L470
| ML20116L470 | |
| Person / Time | |
|---|---|
| Site: | 05200001 |
| Issue date: | 09/09/1992 |
| From: | Knecht P GENERAL ELECTRIC CO. |
| To: | Palla R NRC |
| References | |
| NUDOCS 9211180322 | |
| Download: ML20116L470 (9) | |
Text
~
y f 5,i,y cz h.k m 3 p,,,
52 -Gl)l September 9, 1992 CC:
JD Duncan CE Buchholz H Careway 5 Visweswaren To:
R. Palla (NRC))
L Tredrick from:
PD Knecht
Subject:
P r el i m i n a ry_Dnfijas to nMLts_NRC_Renv11 Liar _Inletm a t i o n
_,P29T~evn.3n
Reference:
FAX to Duncan from Palla, 8/18/92, ' Request for Additional lafer:ntion Regarding Analysis of LOCAs Outside Containment (0-4)
The basic goal of the Bypass evaluation in Appendix 19E.2.3.3 is to show that Bypass events do not significantly affect the total offsite riski therefore,.
consideration in the PRA is not necessary except for paths which can not be excluded. The exce)tions are addressed in detail in the PRA. The following responses clarify tie information contained in section 19E.2.3.3 and provido a response to the NRC concerns expressed in the referenced FAX.
In the interest of brevity, the NRC concern: have been summarized.
Concern 1.
Please provide a more comprehensive analysis of the frequency of core damage from LOCAs outside containment using event tree and fault tree techniques. The analysis is to include the effects of instrumentation availability and consequential failure of injection tystems.
RESP 0MSE Simplified ovent trees of the containment bypass events are shown in Figures 1 and 2.
Figure 1 represents the LOCA outside containment events.
Figure 2 represents containment bypass situations which occur independently from the initiating event.
For both trees the end result is either equivalent full bypass or no bypass.
Probabilistic values she'n nrn derived frnn Tabic 19E.2 :1 of the 3AP.,
In the treer the following definitions apply:
V)
The frequency of piping breaks in lines which value used (2.6E 4)y with the reactor vessel. The communicate directl is the total estimated frequency for small, medium or largo breaks outside of containment.
The basis for this estimate is provided in Appendix 19E.2.3.3.3 (2)(k).
Q Conditional probability that required makeup for core 3
cooling is affected by the ex containment LOCA.
It is assumed that such failure results from environmental effects (including flooding). The value Q, is discussed in response to concern-3. When isolation is successful core cooling is assumed to be 170021 unaffected by the briaf Period of broak flow and a nine of 1.0 is assumed.
i i
9211100322.920909 PDR ADOCK 05200001' h*
g.
A-PDR
/
~
im e, T mm.m crir r::s p,
q 1
i 0*
conditional probability of core cooling failure by all-sources of cooling not affected by the
'ex containment LOCA.
The values used are derived i
from the event trees in Appendix-190.4. When used with V. the product (0,*V ) is the total core damate frequeIncycalculatedinthIcurrentPRA(1.5E-7).
i 7
X)
Bypass Fraction - Total conditional probabilities of i
all ex containment LOCA paths (from Table 19E.2-21).
Conditional probabilities are weighted by a " flow split fraction" to account for different consequences from individual lines. The result is-therefore the equivalont full bypass fraction.
i The " flow split fraction" compensates 'for the flow split between flow out of the break in comparison with flow into the suppreslion pool. An adjustment is made to each line contributing to the total such that-2:1 equivalent full flow value is obtained.
X includes the probability that operator _ action flils due to a lack of instrumentation availabilityc t
for most bypass paths considered, no credit for-
~
operator action is taken; For ECCS discharge lines tic operator is assumed to act to close an open 1
valvo, if needed.
For these paths, the leak detection system is adequate to alert the operator of a secondary containment break and instrumentation is not considered to provide a strong contribution to the failure probability.
As shown_ on Table 19E.121 and Figure 1,._the bypass fraction from ex-containment LOCAs is equal to 2. 4 E. f,.
V The combined frequency of all initiat'ing events-2 exclusive of ex containment LOCAs.
From Appendix 190.4 this is about 2.0 events per year.-
X Bypass Fraction Total conditional probability from p
all potential Suppression-Pool-Bypas's paths resulting in complete bypass of.the suppression pool. This represents the failure of an isolation barrier inia' lino communicating with the drywell-(excluding vacuum; breakers, purge and vent lines). = These failures do not initiate transients or core damage sequences..NoL cperator action was assumed for.theso paths, w-J
wrT C M i m C D.t : U 4 r.1 To compensate for the flow split between flos cut of
,the bypass path in comparison with flow into the suppression pool, an adjustment is made to each line contributing to the total such that an equivalent full flow value is obtained. As shown on Table 19E.2-21, the bypass fraction from suppression pool hypat paths is equal to 1.5E 5.
Conclusion Although the consequence from bypass events is greater, the frequency of such events concurrent with core damage is extremely small (2.3E 12/yr).
This is well below the criteria of exclusion for consideration in the PRA, regardless of consequence.
The total bypass fraction of 3.9E 5 represents approximately an 18% increase in the offsite exposure risk within 50 miles (between see response to Concern Sc).
Since significant margin exists the current PRA results and the safety goals, it can be concluded that the bypass events do not significantly contribute to the effsito oyposure risk.
Concern 2.
It is not cicar that all potential bypass lines were included in accordance with the current ABUR design. Phy was the RWCU Line cmiticd?
RESPONSE
The bypass lines identified in Table 19E.2 1 were derived from a systematic review of the ABWR palDs available about June, 1988.
Significant additions or revisions are not believed to have occurred since that time which would affect the results of this study. A confirmation is planned to verify that.o un-evaluated bypass paths have been introduced into the design.
The RWCU Return line to Feedwater was included in Table 19E.2-1 and excluded from further analysis because the bypass path is protected.by the feedwater check valves which are already included in the analysis. A break in the discharge line also could result in a bypass pathway through the.RWCU suction. This line is included in the analysis.
Concern 3.
An systematic evaluation of the environmental and thermal hydraulic effects of ex containmont LOCAs on the availability of core injection capability should be provided. The effect of unspecified line layout should be addressed.
RESPONSE
This concern addresses the assumption in the analysis that the value if Qi in the event trees.
It is apparent from Figure 1 that these events can not be significant unless the amount of increase in the coolant makeup frequency (Q is greater than the probabilitythatthecoolantmakeu$)frequenr.yisaffected(0).
3
rtr to
- m e m rn m t r.:.E r,r A systematic evaluation of potential flooding due to ex containment breaks was summarized in Appendix 19R, Probabilistic flooding Analysis.
flooding in the reactor building is noted to disable the system affected and potentially flood the Reactor Building corridor, but not disable other makeup equipment due to the water tight doors contained in the design. A value of Q for flooding events affecting one division was estimated to be a8out 1.3E 6 wh.ch is similar to the value of Q without consideration of bypass effects. TheconditionS1makeupfailure probability with two systen:s out of service is estimted in be 2.7E 5.
It can be concluded that although the makeup probability ((
likely to be slightly higher due tr '.he loss of the floodec') is system, the value of Q would need to be greater than.?. for thiscontributiontoaffecttheexcontainmentLOCAoffsite exposure risk.
Based on the flooding analysis, it is judged that Q is likely much less.
The net affect on the analysis is that t.keriskfronsuchbypassinitiatedeventsisnegligib1c, Although line routing are nei specified, the analysis assumes a break in the reactor building anywhere outside of the containment isolation valves. This is either inside secondary containment or outside secondary containment and is addressed by the flooding analysis of Appendix 19R.
Therefore tF results are not sensitive to the final layout.
Centern 4.
COL applicant actions to confirm the assumptions in the PRA should be provided. A reassessment of potential bypass paths should be included.
PESPONSE ITAACs were provided for the major assumptions included in Appendix 19E.2.3.3 (See Table 19.8 2).
In addition, system interactions which could cause a consequential system failure due to ex containment LOCAs are aircady a requirement for the COL applicant (see Appendix 198.2.3). A systematic evaluation of potential pool bypasses will resylt from the system interaction review.
Concern 5a The contribution of bypass events which lead to cor,e damage appears to be ignored in the methodology.
RESP 0fl5E Pool bypass cycnts which lead to core damage are encompassed by the ex containment LOCA event tree (Figure 1). This tree is discussed in more detail in responses to concerns 1 and 2.
Concern 5b The methodology assumes that the bypass event frequency is much less than the non-bypass frequency. This assumption may not.be valid if some bypass events have a relatively high probability of ccusing cnre d rage.
crece. e rp e r cr v r.rc r..s l
ptSPONSE Bypass events do not have a relatively high probability of causing core damage.es discussed in response to conc.ern 3.
Concern Sc The SAMDA analysis appears to indicate that the ratio of consequences is non~ conservative.
Rf!PONSE The SAMDA analysis (section 19P) provides the results of the offtite exposures (over a 50 mile radius) for four separate soquences.
The consequence ratio in Appendix 19E.2.3.3 was based on the ratio of the site boundary doses for a full release of noble gases 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> after the event (sup)ression pool scrubbing assumed) to that with full bypass and no acid up.
The closest comparison of consequences is provided by comparing
" Case 7", a bypass event through the wetwell drywell vacuum breakers, with " Case 1", coro melt with release through the rupture disc. Accounting for the assumed flow split of.26 in the vacuum breaker line yields an approximate consequence ratio of 1.1E-3.
This comparcs with 1.2E.3 assumed in Appendix 19E.2.3.3.
It should be noted that as a result of ACR$ cuestions on the appropriateness of this ratio, GE has concluced that the proper ratio should have been based on the dose from normal containment leakage folloving core damage rather than the 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> rupture disc release.
The ratto of doses indicated in Table 19P.21 may be used to estimate the consequence ratio appropriate for the evaluation of suppression pool bypass.
" Case 1", adjusted for the assumed flow split, may be used as an approximation of the full suppression pool bypass consequence: "NCL" is the consequence from normal
~
containment leakage.
The corresponding ratio is 2.2E 4 which compares with the value of 1.2E-3 indicated in App 1ndix 19E.2.3.3.
This new consequence ratio ca') be used in the evaluation of pool bypass significance.
, Concern 5d Split fractions appear t,0,bc non-conservative, for,.large lines and for. lines which ini,titte core damage.
Resp 0NSE The flow split fractions were determined from first principles calculations of the anticipated pressure drops and flows expected during a severe accident. Two types of flow splits were censidered: For ex. containment LOCAs, the pressure drop associated with bypass flow was compared with the pressure drop associated with four open SRVs; for suppression pool bypass paths the bypass pressure drop was compared with that for ten upper drywell vents uncovered about one fourth of the available opening,. As shown on Table 19E.219, the calculations indicated that even-for-the larcest lines complete bypass dnos not. occur.
=
FIGURE 1 l.0#. A OUT910F CONT *.f HMEt'T Line Break Line' Coolant Hakeup Cool ant Outside_.,
11pittipn Not Affected Hakeup (V;)
(X )
(0 )
(0,)
3 3
OK OK 2.3E 6 2,fE 4
- 6.0E-10 NON DYPA55 OK 2.3E 6 2.4E 5 1.4E-15 BYPASS N,11 OK tlegl BYPASS-
- e t
e
str. 6 '!2 Ot f71 So u: r. tts' r.0 i
FIGURE 2 w'PPRCSSION POOL BYPASS (NON INITIATING f. VENT)
No Pool Coolant Other Core -
fLvf 8s i l X 1 lltkellp {0 l E10111LJ3 L.
7 y
,q OK 7.5E 8 2.0 1.5E 7 NON BYPASS 1.5E-5 OK 7.$E 8 2.3f 12 BYPASS
$4 h
..k,-...
^.,.,.!-.-v
,+-no.
,,r,,,
4'~,
w v
m.
e'
<-TRANSACTION REPORT >
09-09-1992(WED) 17i54 C
RECEIVE 3
NO.
DATE Tit 1E DESTINATION STATIOf4 PO.
DURATION NODE R2SULT.
6848 9-09 17:5C 408 925 1289 8
0'03'41' NORt1. E OK 8
0'03'41' 0
4*
e g
9
\\
Wee a Sr s
=;.=.aw:p+
w
- r 44-_s-4-'e-m.s
09 *st c.:-:rm: 7 e r u :.-
r.. t a
.1 i;U ts bf$f!C.l90f[09fgi t
SYSTEMS INTEGRATION-ENGINEERING SAN JOSE, CALIFORNIA (408) 925-1289 RAPIFAX a
OR Otti COMM S-425-1289 TE'.ECOPY TRANSMITTAL Df.TF :
7/9./F2.-
,_fd sN qf_ ~$_Ay.&c TD:
C /ca fr//y*
pg_c
_sh h Lt A w a c.
A FTt:
..jf t....h.dgff_tr~
sCdW L
TFIS TP.AN5"ITTAL INCt.llDES COVER SHEET +
7 PAGES CO*tE!!TS:
AtL:
RY.W' Ac 4
. C'" * [GLL ["6
.?~' 2 ES.".'1 r-w__.
-O gamme _ _
__