ML20044E090
| ML20044E090 | |
| Person / Time | |
|---|---|
| Issue date: | 03/04/1993 |
| From: | Rosalyn Jones Office of Nuclear Reactor Regulation |
| To: | Tritch S WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP. |
| References | |
| TAC-M81078, NUDOCS 9305210262 | |
| Download: ML20044E090 (10) | |
Text
-
March 4, 1993 i
t Mr. S. R. Tritch, Manager Nuclear Safety Department Westinghouse Electric Corporation P. O. Box 355 Pittsburgh, Pennsylvania 15230-0355
Dear Mr. Tritch:
3
SUBJECT:
REQUEST FOR ADDITIONAL INFORMATION ON WCAP-12909-P (TAC NO M81078)
In our review of WCAP-12909-P, " Westinghouse ECCS Evaluation Model: Revised l
Large Break LOCA Power Distribution Methodology," and clarifications and -
corrections of November 21, 1991, we have identified a number of items which must be addressed. These items are identified in the enclosed questions for Resolution of these items is necessary for us to complete our your response.
review.
Please contact Frank Orr (301-504-1815) of my staff if you have any questions i
regarding this request for additional information.
l t
i
/s/
1 Robert C. Jones, Chief h
Reactor Systems Branch Division of Systems Technology Office of Nuclear Reactor Regulation l
Enclosure:
Request for Additional Information l
cc:
L. Hochrieter R. Ankney 3
W. Tausche 1
I
Contact:
Frank Orr, SRXB, Ext. 504-1815 DISTRIBUTION i
Central file SRXB:DSSA SRXB:DSSA*
SSA SRXB R/F i
F0rr: Bah
- MCaruso RJ es F0rr 9/02/92 1/08/93 3/ /93 F0rr R/F RJones
- See previous concurrence Doc. Name:
G:\\W12909.RAI MCaruso
'O 170170 9305220262 930304
\\\\s PDR TOPRP ENVWEST-
\\
C PDR u
Attacnment august 4, 1992 CPF-8-92 Page 1 of 9 REQUEST FOR ADDITIONAL INFORMATION WESTINGHOUSE ECCS EVALUATION MODEL:
REVISED LARGE BREAK LOCA POWER DISTRIBUTION METHODOLOGY WCAP-12909-P The following questions concern the information presented in Table 4-2, 1.
Reference 1.
For the statistical terms in Table 4-2, clarify if Westinghouse a.
used any term in a way other than its normal statistical meaning.
Also, clarify Westinghouse's interpretation of what each term implies concerning the ability of the power shape sensitivity model (PSSM) to represent the original database and to correctly predict new situations.
i For the peak cladding temperature (PCT) error in Table 4-2, b.
specifically discuss the following items:
(1)
Clarify how the PCT error was calculated.
In (2)
Explain the basis for claiming 95% confidence.
Figure 4-1, eight points do not lie within the stated confidence band, and, for 95% confidence, at most four points would be expected to be outside the confidence band.
Also, clarify why the Westinghouse model has a constant Most models have a larger confidence band confidence band.
This is at the ends of the data ranges than in the middle.
because there is less data to support the model at the ends than in the middle.
(3)
Clarify why the PCT prediction errer given for 95%
confidence is less than the 2a error normally associated with a 95% confidence level.
l (4)
Clarify how you will handle the situation when the difference between the PSSM PCT and the reference PCT is significantly less than the error given in TabieT2T In
~
1
I k
t this situation, the power shape sensitivity is overwhelmed f
by the prediction error and it is difficult to judge which l
is the worst case power shape.
,i i
Equation 5-1 is used assuming the third ivitegral factors remain 2.
(a)
Clarify the potential effects this assumption could constant.
have on the PSSM because, if the result of Equation 5-1 is different from the power shape being studied, the third integral factors will not remain constant. Clarify how these effects are l
compensated for in the PSSM or show they are conservative.
l Because of the uncertainty associated with the PSSM model, clarify (b) wny a 2a factor is not used in Equation 5-1 to ensure the FOR""
calcul...ed bounds the expected value with 957. probability.
The information in Reference 2 provided the correct form of 3.
However, information should also be provided to clarify Equation 5-2.
why Equation 5-2 always results in a power distribution penalty.
l 1
Provide examples of applying Equation 5-2 to representative plant data that cover the range of applicability of the PSSM to clarify the penalty 1
l Also, provide examples of applying nature of the equation.
Ecuations 5-2 to 5-4 to representative plant data that cover all l
applicable ranges to clarify how the penalty affects the plant 6
h surveillance measurements.
1 J
Clarify if Equations 5-3 and 5-4 are shown correctly; some terms appear I
4.
l to be missing.
l The data ranges for the independent variables used in the PSSM are given l
5.
i l
in Table 4-3, Reference 1.
In Reference 2, Westinghouse provided f
information to clarify how they handle the situation when data falls A
outside the ranges given in Table 4-3.
In this case, Westinghouse sets the distribution specific surveillance penalty factor to the value given r
Because the penalty f actor calculation is only one part in Reference 2.
of a PSSM application, justify why the approach described in Reference 2 e
2
ensures the limiting power distribution is identified by the PSSM in situations where the power shape is outside the range of Table 4-3.
l Could the variation in the axial power shapes used in the PSSM f
6.
sensitivity study result in the codes and/or correlations in the codes l
Clarify the being used outside their approved range of applicability?
l basis for your answer.
Provide the ranges of the important reflood parameters such as reflood 7.
j rate, core inlet subcooling, containment pressure, emergency core coolant (ECC) flows, etc. covered in the calculations used to develop Table 3-1.
Compare the ranges of the above parameters in the l
calculations to the ranges expected for all Westinghouse plants and answer the following questions.
at Are there any variables with ranges outside those covered by the l
(a) calculations in Table 3-l?
Are there any variables where insufficient data was included in (b)
Table 3-1 so that applicability of the PSSM is not well supported for all situations expected at Westinghouse plants?
F If the answer to (a) or (b) is affirmative, justify the (c) i applicability of the PSSM model to cases outside the covered f
ranges or without sufficient data to support PSSM application.
i Alternately, clarify how Westinghouse will handle such cases and justify the adequacy and acceptability of this approach.
i The power distributions included in Westinghouse's power shape t
S.
(a) sensitivity study (see Table 3-1) did not include peak power locations that covered the entire core.
Application of the PSSM to power shapes with peak power locations outside the range of data in Table 3-1 would result in its use outside the data range used to develop the model. Therefore, for power shapes with peak power locations outside the database range, justify:
(1) the f
_ applicability, of the FSSM_ to determining the limiting nature o these power shapes; (2) these power shapes are never limiting or 3
4 are not allowed to occur, and the PSSM is not applied to them; or (3) another method that is used to determine the limiting nature of these power shapes.
Because thousands of power shapes are possible during a fuel (b) cycle, clarify why Westinghouse chose the power shapes it did for j
inclusion in the PSSM development (see Table 3-1) and why other power shapes where excluded.
Justify Westinghouse included sufficient calculations in the database to accurately determine power shape effects on PCT. Also, only 6 calculations were included with the axial peak above approximately the 9.7 ft Justify why each plant should not have included a elevation.
axial shape with the peak above 9.7 ft and a high F, in order to Would the PSSM have an appropriate database for PSSM development.
change if additional axial power shapes with peaks above 9.'7 ft and high F, data were included in the database?
In Table 3-1, clarify if any of the axial power snapes used to (c) an axial power-develop the PSSM could be described as follows:
distribution that increases quickly at the bottom of the rod and i
then flattens out; however, the power continues to increase so l
that the axial peak occurs at the 9 to 10 ft elevation on the rod.
f This shape That is, a shape that is flat but skewed to the top.
If the above has the potential to be limiting in some situations.
axial power distribution was included, clarify how many analyses included this type of power distribution and just'.fy the number If this was sufficient for the PSSM to adequately represent it.
i type of power shape was not included, justify why it was not l
j included.
l f
Clarify why the number of cases for each plant differed in (d) i Also, clarify the basis for determining how many cases Table 3-1.
l to run for each plant and the overall total.
l Because of the high PCTs in Table 3-1, it is assumed cladding rupture 9.
was calculated.for the cases an_alyzed.
Clarify if this assumption is 4
l
9 If rupture was not calculated, clarify how the PSSM will be j
correct.
applied to situations where rupture is calculated.
Clarify if a rod-to-rod radiation model is included in the approved Westinghouse reflood models.
If a rod-to-rod radiation model is l
10.
l included, clarify how the model accounts for the effects of different axial power shapes.
Westinghouse used the blowdown phase results of the reference analyses 11.
with the cosine power shape as the bases for all reflood calculations in Therefore, you did not directly calculate the effect of their study.
the different axial power shapes on the possibility of a blowdown Because different axial power shapes affect the possibility of rtrue.
a blowdown rupture and, therefore, stored energy at the st:rt of reflood, justify how the PSSM accounts for this effect.
j' x
i
/
i J
Section 6.0, WCAP-12909-P, is entitled PSSM Uncertainty Evaluation.
12.
However, the information presented in Section 6.0 is confusing.
Although the section starts out discussing uncertainty, the last It is the INEL's sentence makes a conclusion regarding conservatism.
Uncertainty relates to the understanding that the two are not the same.
possible range within which the actual answer is expected to be found relative to the calculated result. For example, a PCT me' hod with an uncertainty of 1150*F that calculates a PCT of 1500*F indicates the Conservatism, however, actual PCT would be between 1350 and 1650*F.
f relates to whether a method consistently gives results that are above or below (depending on which direction is considered a worse result) a known or expected reference value. For example, a method that calculates cladding temperatures above expected cladding temperatures is
[
Therefore, clarify the purpose of Section 6.0 considered conservative.
(that is, Is Westinghouse discussing uncertainty or conservatism?), and provide sufficient information to justify your conclusions (see question 13).
In Section 6.0, Westinghouse attempted to demonstrate the integrated 13.
uncertainty / conservatism of the PSSM approach by comparing the
' ~ ~
Evaluation Model (EM), PSSM, and 95th percentil'e cilculated PCTs'.~At t
5 i
I
=
i a
the end of Section 6, Westinghouse stated, " based on these results [the PCT comparison discussed abovei, it is concluded that sufficient t
i conservatism exists in the PSSM Application Methodology described in f
Based on the information in Reference 2, the PCT this report."
prediction error is r at included in the PSSM calculated PCT because of I
The staff does not agree with the conclusion of j
this conservatism.
Section 6 and the information provided in Reference 2 is insufficient to justify not including the PCT prediction error in the PSSM calculated j
The ~ staff does not agree with Westinghouse's conclusion / approach j
(I) Because f
PCT.
in Section 6/ Reference 2 for the following reasons:
Westinghouse compared the PCT results for different power shapes, the Rather it l
comparison does not really show the conservatism of the PSSM.
shows different power shapes give different PCTs, and a power shape pushed to the technical specification limits will give a higher PCT than If conservative is defined to mean a higher l
a nominal power shape.
(2) i PSSM PCT relative to the EM PCT for the same power shape, Figure 4-1 l
This is because the PSSM PCT shows the PSSM is not always conservative.
f se,.c power shape.
(3) in some cases is lower than the EH PCT for the T
Only one data comparison is provided, and one comparison is not I
sufficient to establish either the PSSM uncertainty or conservatism.
l The process discussed on page 29 should be applied a number of times to Therefore, provide demonstrate PSSM uncertainty or conservatism.
sufficient information and/or additional appropriate comparisons (that
[
J is, comparisons based on the same power shape) to support Westinghouse's l conclusions in Section 6.0 regarding either the PSSM uncertainty or f
Also, provide additional information to justify why it is conservatism.
not necessary to account for the PCT prediction error in the PSSM in l
3 order to determine the worst case power shape as stated in Reference 2.
f In Section 6.0 (page 28, third paragraph, first sentence), Westinghouse 14.
You made a statement regarding the composite uncertainty of the PSSM.
4 indicated the composite uncertainty was related to PSSM uncertainties and measurement uncertainties. Clarify how measurement uncertainties l
[
affect the composite uncertainty of the PSSM, an analytical model that j
does not use measured data as input.
?
it 6 -
_ m-m_.
m y-,,.w.
,m.
..,e i
I
?
._m
i 1
In Section 6.0, Westinghouse completed an integrated uncertainty j
l 15.
evaluation. You argued the PSSM was conservative because you based it on conservative EM results and analysis of conservative power shapes.
Another factor was the PSSM result was higher than the 95% probability result. While this may lead to conservatism in a general sense (the l
PSSM calculates higher than expected results}, it does not mean the PSSM l
is conservative for its intended use. The PSSM needs to predict conservative results in order for the EM comparison to be meaningful.
)
If the PSSM was BE and the EM conservative, or if the PSSM was
)
conservative and compared to BE results, conclusions on the PCT I
This is because the sensitivity to power shape could not be made.
l results compared would be based on different assumptions.
Therefore, i
clarify Westinghouse's meaning regarding the use of conservative in i
Provide sufficient information to support your definition.
l Section 6.0.
In the power distribution report, Westinghouse's meaning in using 16.
(a)
I the following phrases is not clear:
95/95 upper bound, page 28.
i a.
b.
95% probability PCT, page 29.
Statistical 95/95 PCT, Table 6-2.
1 c.
Therefore, clarify your meaning in using the above phrases and l
I l
provide sufficient information to justify their use in the report.
~
i Also, Westinghouse seems to use b and c above to say the same l
(b) thing in Section 6.0.
From a statistical viewpoint, they are not l
It is also not clear from the report why l
i the same thing.
Westinghouse applied either b or c to the result discussed on f
page 29 or in Table 6-2.
Therefore, provide suf icient information to support Westinghouse's nie of the above phrases in Sectic. 6.0.
The title of Figure 6-1 refers to a probability density function, 17.
(a) i but the figure itself appears to be what is usually called a cumulative distribution function. Therefore, clarify the contents of Figure 6-1.
7
i i
i Based on Westinghouse's description on page 29, it is not clear f
(b)
Provide a more detailed how Westinghouse developed Figure 6-1.
i explanation en how you developed Figure 6-1.
Westinghouse developed Table 3-1 using two versions of the Westinghouse P
18.
- However, EM; therefore, you concluded the PSSM applied to both ems.
i Table 3-1 shows Westinghouse completed most of the calculations with one l
Therefore, provide the following information to support of the ems.
In Westinghouse's conclusion regarding PSSM applicability to both ems.
I J
each case provide sufficient information to support your conclusions.
Are there any differences in the data trends from the two ems?
(a) t If there are differences, do the differences impact the PSSM as it (b) is applied to one or both ems?
i 3
l i
If there are differences, could Westinghouse represent the data
, (c) j better 5y asin{ two different models?
7 i
i The following questions relate to PSSM applicability.
19.
i On page 11, Westinghouse stated the PSSM was based on all current (a)
However, Table 3-1 shows only three of 80 l
plant types.
calculations were for the two-loop plants. Because of the small number of data points for two-loop plants, justify the l
applicability of the PSSM to two-loop plants.
i On page 11, Westinghouse also listed the fuel designs included in r
(b)
Clarify how many calculations were performed for j
the database.
each fuel type, and justify the number was sufficient to l
Also, the PSSM does j
accurately model PCT effects for that design.
not directly include any bundle geometry effects which can have a i
significant effect on PCT. Therefore, clarify how the PSSM i
adequately accounts for fuel design geometry effects on PCT in f
order to justify PSSM applicability to all fuel types.
~
O i
4 REFERENCES D. L. Stucker, et al., Westinchouse ECCS Evaluation Model:
Revised 1.
Larae Break LOCA Power Distribution Methodoloay, WCAP-12909-P, May 1991.
S. R. Tritch letter to USNRC Document Control Desk, Methodology 2.
clarifications to WCAP-12909-P, ET-NRC-91-3633, November 21, 1991.
(
v _
9