ML20126D304
| ML20126D304 | |
| Person / Time | |
|---|---|
| Site: | 05000447 |
| Issue date: | 11/17/1983 |
| From: | GENERAL ELECTRIC CO. |
| To: | |
| Shared Package | |
| ML20126D269 | List: |
| References | |
| FOIA-84-175, FOIA-84-A-66 NUDOCS 8506150016 | |
| Download: ML20126D304 (4) | |
Text
r-11/17/83 Letter to D. G. Eisenhut from J. F. Quirk re: GESSAR-II Internal Event PRA Uncertainity Analysis (64 pages) 16 950215 1
D -66 POR.
A
~
- l" 3 *O F.ETH0DOLOGY A quantitutiva treatment of uncGrtainty may involve, in varying degree,
/
some or all of the following steps:
./
/
/
/
9 1.
Evaluation of input uncertainties.-
2.
Propagation of input. uncertainties.
3.
Combination of the uncertainties to form the output uncertainty.
4.
Interpretation of the unc'ertainties in the PEA results.
A number of methods have been developed to treat and propagate the uncertainty measures.,fhe Monte Carlo simulation method was chosen ]
for the GESSAR II uncertainty analysis.
Ig The Monte Carlo method presents the most direct approach to the problen i
of uncertainty propagation when input uncertainties are represented as E
, distributions on parameters. It involves an evaluation of the output of a computer code, or other analytical model for many sets of combina-tions of the input parameters. These combinations of input values are obtained by a random sampling from the distributions assigned to the f
input variables. Monte Carlo simulation thus constructs an approxi-
/
mation to the output-variable probability distribution.
/
Many codes have been written to perform Monte Carlo computations, including SAMPLE, STADIC, and SPASM. The SPASM code (4)was used for this uncertainty analysis.
i Figure 3-1 illustr' tes the procedure used in the assessment of uncer-a tainty for the GESSAR II PRA. The following paragraphs describe' the t
methodology employed.
i l
i 1
3 u
^
4.0-RF. SUI.TS Tho results of th2 GESSAR II intsrpal event PRA uncertainty analysis m
/
listed in Table 4-1 and described,below
/
/
1.
The error factor for core damage frequency associated with uncer-
~/
tainties in the point value estimates is 3, based on a 5 and 95 percent confidence limit. The distribution curve obtained through Monte Carlo simulation.is shown in Figure 4-1.
2.
The uncertainty factor for the individual release categories varied
'N 1
Y between 5 and 20. The data summaries for the individual release categories are shown in Table 4-2.
A summary of the individual release category distribution parameters are shown in Figure 4-2.
3.
The eight release categories not included in this uncertainty analysis can be shown to have a negligible effect on the esti-mated core damage frequency. For the eight release categories not analyzed, it was assumed that the distribution for each category would be represented by a log w=mel distribution having the mean eae frequency value presented in Reference 3 and the maximum error factor determined in the uncertainty analysis [ Error factor for release category 12 = 18.4]. The plot of cumulative core damage l
frequency shown in Figure 4-3 shows the cumulative contribution of each individual release category to the mean and 95 percentile parameters. The plot shows that the eight release categories not included in the uncertainty analysis do not contribute significantly to the two parameters.
This confirms that the seven release catego-W ries included in the uncertainty analysis represents a goet approxi-nation of the total core damage frequency.
4.
The calenia,ted mesa core damage frequency in.this analysis is lower b\\
than the value reported in Reference 3.
The lower frequency is a result of refinements made to the fault and event trees since the sub-mittal of Reference 3.
Further, the value obtained in this analysis But incorporates all sequences identified during the.534F review of the FRA which were not explicitly addressed in the previous submittal.
'A 14
a
.i Figure 4-2: Summary of Distribution Confidence Limits 7
t i
I I
I Confidence Limits 5%
95%
Release l
I Category
,'e n n h
median M
5 No.5 l
h i
M 3
No.3 1 1
p I
I g
5 4
No.4,I O
E g
I
,I l
!i 1
No.1 j
2 No.2
,1 I
j a
No.12 g
5 12 l
i e
i Q
C 15 No. 15 I I
i 3
8 U
- l l
Total Total r
i
- estimated
-6
-8
~
~
10 10~
10 10 10 10
~
Core Damage Frequency ( Events / Reactor-Year)
-.