ML20071M610
Text
..
henomena in a plant having the passive safety features currently planned for the A%00. Note that this conclusion does not apply to the long-term, low-pressure passive cooling of the AP600 involving the In-Containment Refueling Water Storage Tank (IRWST) and the containment cooling mode. I define reasonable simulation to mean that the key plant components will be present in the LSTF, that these components will be sufficiently representative of the AP600, and that the Maant processes and key phenomena will be reasonably simulated for the initiating events m the planned test matrix. I define a reasonable transient simulation as one in which the major trends are obtamed for key parameters, the events generally occur in the correct order, and the correct end state of pisnt is reached.
940B040301 940629 PDR COP 9fS NRCC CORRESPONDENCE PDR
1 s
i j
.l i
Dr. L. Shotkin June 11,1992
%1192 354
=
3 1
\\
I 2.
I have concluded that the proposed ROSA LSTF test matnx is probably adequate.
However. neither the Reference document nor the discussions in the meeting
-i developed the basis by which this set of expenments was selected. In fact. toward j
the end of the meetmg, it became. clear from a discussion between the INEL and j
Westmghouse that more work should be done m this area, as the discussion focused j
on whether to test with a two or three inch break in the cold leg. In a related manner.
I thmk that it would be wellif the phenomena for the large-break (LB) loss of coolant i
accident (LOCA) were identified. One place to list this informanon would be in the i
Reference document. Although a LBLOCA test cannot be conducted in the ROSA l
LSTF. focusing on this issue may be oflong-term benefit to the NRC in developing a 3
i
{
-logic as to why a LBLOCA test need or need not be conducted by Westinghouse. At a minimum. the basis for not conducting a LBLOCA expenment should be formally
)
documented and reviewed.
I 3.
i I am not convinced that the attempt to extend the LSTF configuration and testing mamx to the low pressure, late-time regime will be successful. As we ccncluded the i
meeting. I felt that the feasibility and workability of the hardware modifications for l
}
this extended testing range were quite uncertain. The stated objective of sucn testing is commendable i.e., to permit an overlapping with other AP600 test facilities such as SPES and the low-pressure facility at Oregon State University (OSU). However,
+
]
I ask the following question: is the potential gain worth the programmatic risk if the LSTF low-pressure simulations involving the IRWST do not, because of facility atypicalities, show the same behavior as that observed in the OSU facility?
i 4.
I remain uncertain about whether adequate instrumentation has been provided. This i
is not an area of my expertise. However, views regarding what instrumentation was 4
needed varied widely among the meeting participants. Decisions were quickly made.
These should be carefully thought through and justified; it is my understanding that the INEL received an action item to do this. I believe that the temperature, pressure, and pressure difference measurements are adequate. The break measurements using JAERI spools designed for this purpose also appear adequate. I am unsure as to 7
whether instrumentation provided for flow measurements of the automatic depressurization system (ADS) will be adequate. I believe that it is of the utmost importance that the facility instrumentation permit mass and energy balances and to be prepared, and the location of fluid in the system to be identified as a function of time.
The inventory comparison plots prepared by the INEL were excellent. When I asked the INEL staff if the ROSA LSTF mstrumentation provide data to generate similar i
plots bar.ed upon the tests, they were uncertain. If possible, it is important that this question be answered in the affirmative for the proposed test program.
5.
The experimental simulations will be most reasonable early in the experiments and will become less reasonable towards the end of each test. This effect will be more pronounced for slow transients, e.g., the 1-in. cold-leg (CL) small-break loss-of-coolant echt (SBLOCA) and less pronounced for fast transients such as the 6 in.
CL SBLOCA. Design modi 5 cations at the meeting such as a more accurate simulation of the upper head bypass draining paths and prototypic pressurizer
')
~
will help to minimize these late-transient distortions but they will not eliminate all such distortions. However, even with these distortions, I conclude that the transient informanon collected will be useful for code validariari.
6.
The efforts conducted so far to ii i'y key AP600 components, processes, and phenomena are worthwhile. '!he tabulations generseed thus far should be revisised and,if necessary, revised based upon the caletdations performed so dase. A second
i 1
Dr. L. Shotxin 4
June i1.1992 N-1192-354
.evision, based upon a review of the test results from the OSU and SPES facilities.
when available. as similarly urged.
I
\\
The informanon generated in the ROSA LSTF facility will, with data from the OSU and SPES facilities, provide a significant data base for code validation. I conclude i
that this data base will be sufficient to demonstrate the applicability of any code for which a thorough assessment matrix is defined and the assessment successfully
]
completed.
4 8.
The NRC's AP600 program is in the initial phases. Therefore,it has been necessary j
to rely on RELAP5A10D2.5 comparative calculations of the same set of selected event sequences for AP600 and the ROSA LSTF facility. I recogmze that this is a necessity given this early stage of the program and the need to make decisions 4
regarding the applicability of the ROSA LSTF. I conclude that this has been a reasonable and proper approach for generating insights that will assist the decision-1 making process relative to the use and modification of the ROSA LSTF. However,it
~
is probable that RELAP5A10D2.5 deficiencies to the extent they exist. affected both l
the ROSA LSTF and AP600 calculations. Thus, there is the possibility that some i
phenomena are either incorrectly observed in the testing program. predicted or not predicted to occur that will b The OSU and SPES test programs should be i
closely monitored so that unexpected phenomena can be identified. and appropriate modifications to the ROSA LSTF testing program made in a timely manner. Some i
progress in this area could also be made by comparing calculations for the same transient by Westinghouse using COBRA / TRAC and by the INEL using RELAP5. I 1
presume that Westinghouse would have commented at the meeting if the RELAPS results had shown important behaviors not predicted by COBRA / TRAC. It would be i
useful to formalize such a comparison and document the results.
1'
]
Basis for Conclusions My conclusions are based upon the materials presented in Idaho Falls, Idaho, on June 2-4, 1992: the discussions during those meeting; and my review of the Reference repon. I note that these conclusions are most strongly based upon my general experience with scaled integral facilities and the associated analysis tools; my familiarity with the AP600 design is j
of more recent vintage and my knowledge of this design is evolving.
Odds and Ends 1.
I recommend that event tabulations such as those on page 4.1-5 of the Reference be modified to include the timing of key events. I generated part of such a table and provided a copy to J. Cozzuol at the meeting.
j 2.
If correct conclusions are to be drawn when heavy reliance is placed on code-calculated results, several reviews must be conducted. Firu, the plant design must be accurately represented and the operation of that plant must be understood. The Reference document reflects the design as it stood in September 1990. During the meeting, Westinghouse identified several recent changes to the design and, on at least one occasion, the plant operating philosophy. An example of the formeris the surge i
line redesign and an example of the latteris the two-second delay bsa.e.. activation i
of successive ADS stages. haA it is important that the input model be understood i
relative to what it does model well, what it models with compromises, and what is not Irwienad Third, it is important so understand what the code can and cannot do by virtue of the models and correlations it contains. INEL appears to have donc a i
1
l I
Dr. L. Shotkin a
June 11.1992 1
N-12-92-354 reasonable job of keeping each of these factors m focus dunng the early stages ofits AP600 efforts; it is important that INEL contmues to ao so.
3.
Westinghouse expressed concern about the existence of a loop seal in the ROSA LSTF. ney did so in anocipation that in one or more tests the loop seal might induce i
phenomena that are impossible in the Westinghouse AP600 design. This does not l
appear to be a technical issue. The data will still be useful for code validation, and technically, the relationship of any adverse phenomena could be explained. The NRC may wish to ensure that this possibility is clearly described in test specifications or other appropriate test planning documents, and that explicit statements are made that.
phenomena associated with loop-scal existence or cleanng that may occur in the l
ROSA LSTF simulation of a plant having AP600-like features cannot occur in the
.I AP600.
i 4
Testing in the ROSA LSTF must be accomplished within a specified and tight w:r.dow. He crder Of testing was examined at the mee:ing. It was decided that j
groups of tests would be ranked by priority and that JAERIinput would be sought i
regarding the most efficient testing order within each group. I concur that this is a j
prudent testing approach. I do note, however, that this approach is not without nsk.
If problems are encountered early in the program, it is possible that tests may only be i
conducted within a single group before the Phase I testing window closes. The NRC may wish to consider if it could live with such an outcome should testing problems anse.
i Sincerely,
, f'[lO $1 t%
Brent E. Bcfack i
Leader Terrestrial Reactor Technology Section l
Reactor Design and Analysis Group N-12 i
BEB:ve Cy: B. Sheron, US NRC D. Bersette US NRC G. Rhee, US NRC A. I.4 vin, US NRC L. Smith, N-DO, MS E561 M. Cappiello, N-12, MS K551 CRM-4(1),MS A150 N-12 File l
l l
L
i i
l APPENDIX 11 l
Westinghouse Electric Corporation View of the NRC ROSA-AP600 Testing Program:
Statement of Dr. Lawrence Hochreiter to the ACRS Thermal Hydraulic Subcommittee, June 24, 1992 l
(
Reference:
Official Transcript of Proceedings, ACRST-1917. pp 513-514
" Integral systems tests, as. we have gotten experience with them in other programs, all these tests have blemishes or warts. There is always going to be some atypicality, but the goal is to try and gather a data base that we can use to validate our computer codes, so what we should be looking for is are we l
getting the right phenomena, the right type of thermal hydraulic behavior that we can test the codes on.
We think that that can be achieved in ROSA.
1 Now the full height, full pressure integral systems. tests we're running at SPES is sufficient to support the AP600 final design approval review.
The ROSA-IV test facility is intended to provide to the NRC confirmatory research at a large scale on passive safety-system performar.ce for computer code verification and i
assessment.
1 The complete representation of the AP600 features in the ROSA-IV facility is not possible. The proposed modifications that we discussed at the meeting in Idaho on June 2nd through.4th will provide sufficient representative behavior of the i
AP600 passive safety' systems such that the key thermal hydraulic phenomena and j
l characteristics of the AP600 can be adequately simulated for computer code l
validation.
l l
While Westinghouse would prefer to see the loop seal totally eliminated in the facility, the proposed modifications will permit a test program that should yield i
l valid confirmatory data for computer code verification and assessment.
The loop seal effect, while not representative of the AP600, can be accounted for in the interpretation of the test data, so Westinghouse supports the proposed ROSA-IV test program for the purpose of providing this confirmatory data for computer code validation.
It is our understanding that you will meet the NRC objectives for providing confirmatory data on passive safety system performance at large scale for validating and assessing their codes."
I
\\
l I
,. Q.-
1 1, - o g,.
\\ -.
v g
\\
ADpilCation Guicelines for Check Valves in Nuc: ear..
Power Plants NP 5479 Researcn Project 2233 20
-i ar:
.ruar, 358 acarea r.
':== :.iSCC;;aE5
'.C
- C :
ec c.t a.e ce 'm
- u- --':n 0 ^^036
- J.L5 E'1GihEE LNG.NC 75 : ark Tvo Drive Sugar ano rexas 7 m 8 1
i i
Prepared for Electre Power Research Institute 3412 Hillvew Avenue Palo Ano, Cahtornia 94304 EPRI Protect Manager B. P Broc*s Nuclear Plant Operations and Maintenance Program Nuclear Power Deveon
\\
._,,,.,.,~,-,..,.,-,,._,,..w,,,,
i e4 1
m.
t r
e.x' J \\M L'
(
DISK POSITION vs FLOW RATE Bseeltra Tert. 204 oo I
i l
l 1
1 1
i l
l i
1 i
i gg g,,gy,
,j j
ro -.-.IHE O RE TIC A L I
,/
j 4
MIN k'
i
' Y'6!
l s
80
' "f l
- Y l
l y
s0 f.
t l6
,?
J' eo c
k 30 i
I 20
/
/
m>
10 y
/
r 0
0 4
a 12 se so u
o I
i new rw tre v,
/
r BASELlHE TEST RESULTS FOR 6* SWING CHECK VALVE I
(
d FIGURE D-7 p.
1 u
'l t
l' r i
J.i
.1 l
j-6
! v.
.).
l APPENDIX Ic l
i l
1 i
l l
1 i
l I
1 l
l 1
1 i
6 8
r-
-