ML20045H525
| ML20045H525 | |
| Person / Time | |
|---|---|
| Site: | 05200003 |
| Issue date: | 07/14/1993 |
| From: | Kenyon T Office of Nuclear Reactor Regulation |
| To: | Liparulo N GENERAL ELECTRIC CO. |
| References | |
| NUDOCS 9307200331 | |
| Download: ML20045H525 (14) | |
Text
n Dn.*fsd Y&
pna e
,c
. fh($[%
7 UNITED STATES j
j NUCLEAR REGULATORY COMMISSION
- hypk WASHINGTON, D.C. 20555-0001 July 14, 1993 Docket No.52-003 Mr. Nicholas J. Liparulo Nuclear Safety and Regulatory Activities Westinghouse Electric Corporation P.O. Box 355 Pittsburgh, Pennsylvania 15230
Dear Mr. Liparulo:
SUBJECT:
COMMENTS ON SCALING, INSTRUMENTATION, AND TEST MATRICES FOR THE OSU AND SPES-2 TEST FACILITIES Enclosed are comments on scaling, instrumentation, and test matrices for the OSU and SPES-2 test facilities. consists of comments and recom-mendations related to AP600 testing in the SPES-2 facility at SIET Laborato-ries, based on a review by our contractor. contains comments that have been forwarded to your staff earlier through informal communication to support the test schedules established for the OSU and SPES-2 facilities.
Of the issues raised in Enclosure 1, two are new and the others are similar to concerns on the SPES-2 tests that have already been discussed with Westing-house, either in previous letters or in meetings between the staff and Westinghouse.
More detail regarding some of these concerns is provided in. also provides additional concerns regarding the OSU and SPES-2 facilities.
We are therefore taking this opportunity to summarize these issues'with Westinghouse in these enclosures, and there may be some overlap.
We request that Westinghouse provide a response to the staff either (I) indicating where revisions have been made to the test program or to the facility design in response to staff recommendations, or (2) justifying why specific recommendations have not been implemented. This response should also include consideration of earlier letters and of commitments made at meetings or in telecons.
This information is necessary for the staff to be able to complete its pre-test review of the OSU and SPES-2 testing programs.
You have requested that portions of the information submitted in the June 1992 application for design certification be exempt from mandatory public disclo-sure. While the staff has not completed its review of your request in accordance with the requirements of 10 CFR 2.790, that portion of the submit-ted information is being withheld from public disclosure pending the staff's final determination.
The staff concludes that this request for additional information does not contain those portions of the information for which exemption is sought..However, the staff will withhold this letter from public disclosure for 30 calendar days from the date of this letter to allow Westing-house the opportunity to verify the staff's conclusions.
If, after that time, 0003 i
PDR pg[
th
c Mr. Nicholas J. Liparulo July 14, 1993 you do not request that all or portions of the information in the enclosures be withheld from public disclosure in accordance with 10 CFR 2.790, this letter will be placed in the NRC's Public Document Room.
This request for additional information affects nine or fewer respondents, and therefore is not subject to review by the Office of Management and Budget under Public Law 96-511.
If you have any questions regarding this matter, you can cont &ct me at (301) 504-1120.
Sincerely, (ORIGINAL SIGNED BY)
Thomas J.,'.enyon, Project Manager Standardization Project Directorate Associate Director for Advanced Reactors and License Renewal Office of Nuclear Reactor Regulation
Enclosures:
As stated cc w/ enclosures:
See next page DISTRIBUTION:
Central File PDST R/F TMurley/FMiraglia DCrutchfield PDR RBorchardt TEssig TKenyon FHasselberg GGrant, EDO JMoore, 15B18 MSiemien, 15B18 PShea GSuh (2), 12E4 Alevin, 8E23 MRubin, 8E23 ACRS (ll)w/o enc 1.
- SEE PREVIOUS CONCURRENCE OFC:
LA:PDST:ADAR PM: P,D,ST: ADARPM:PDST:ADAR (A)SC:PDST:ADAR NAME: PSng g Th?In:sg RHasselberg*
07//hj93 07/j4/93 07/13/93 07/13/93 DATE:
0FFICIAL RECORD COPY:
SRXBRAI.TK e
Mr. Nicholas J. Liparulo Westinghouse Electric Corporation Docket No.52-003 AP600 cc:
Mr. B. A. McIntyre Advanced Plant Safety & Licensing Westinghouse Electric Corporation Energy Systems Business Unit P.O. Box 355 Pittsburgh, Pennsylvania 15230 Mr. John C. Butler Advanced Plant Safety & Licensing Westinghouse Electric Corporation Energy Systems Business Unit Box 355 Pittsburgh, Pennsylvania 15230 Mr. M. D. Beaumont Nuclear and Advanced Technology Division Westinghouse Electric Corporation One Montrose Metro 11921 Rockville Pike Suite 350 Rockville, Maryland 20852 Mr. Sterlir.g Franks U.S. Dep.rtment of Energy NE-12 Washington, D.C.
20585 Mr. S. M. Modro EG&G Idaho Inc.
Post Office. Box 1625 Idaho Falls, Idaho 83415 Mr. Steve Goldberg Budget Examiner 725 17th Street, N.W.
Room 8002 Washingtcn, D.C.
20503 Mr. Frank A..Ross-U.S. Department of Energy, NE-42 Office of LWR Safety and Technology 19901 Germantown Road Germantown,llaryland 20874
.i t
Comments on SPES-2 Test Program 1.
The staff recommends that instrumentation be included to measure void fractions in the cold leg pressure balance lines, and that flow in the pressurizer pressure balance line be w asured directly.
~
2.
The staff is concerned that a potential distortion may occur in SPES-2 due to a large volume of stagnant water in the lower plenum in the test facility. The power cables to the SPES-2 heater bundle pass through this volume, and can heat the stagnant water.
If this fluid flashes during depressurization, it may distort the loop behavior compared to what might be expected in the AP600. The staff believes that Westinghouse has been made aware of this potential problem, and requests Westinghouse to describe what steps will be taken to measure fluid behavior in the lower plenum and to compensate for distortions.
3.
The staff recommends that the pressurizer pressure balance lines (to the CMTs) be heat-traced as well as being insulated.
The extremely small diameter and long length of these lines could otherwise lead to large heat losses, which could cause condensation in the lines and distort system behavior.
4.
Additional temperature and flow instrumentation in the hot leg downstream of the surge line, and additional temperature and void measurements (such as heated thermocouples) in the vessel, especially in the external downcomer, would be useful in interpreting system behavior.
While the current instrumentation appears adequate to understand system response if everything happens as expected, it may not be adequate to allow interpretation of anomalous or unexpected behavior, such as flashing in the external downcomer or bypass of the surge line during depressurization.
5.
The staff has previously raised the following issues with Westinghouse regarding the SPES-2 test program, and have not received acceptable responses to the concerns:
a.
The staff recommends the performance of a very small break LOCA test, due to its long duration and potential systems interactions effects.
In the' meeting at SIET in April 1993, Westinghouse stated that their analyses had shown a 1" break as being the most limiting for AP600, due to the combination of recirculation and heat-up of the CMTs coupled with the inventory loss from the RCS prior to ADS actuation.
However, these analyses have not been provided to the staff. Address this issue in greater detail (see Enclosure 2).
b.
The hot shakedown phase of the test program contains a test involving
" inadvertent" actuation of an ADS valve. The staff believes that this test is an important part of the test program, and recommends that it be performed, to the maximum possible extent, as if it were a part of the matrix program. The staff will review the test in that manner.
4.
c.
The staff has asked Westinghouse to justify the use of a single.4th-stage ADS valve as its most limiting single failure for all such tests.
Westinghouse has not provided any analyses demonstrating that the ADS valve failure is, in fact, limiting for all of the test cases (see Enclosure 2).
d.
During a telephone conversation in April 1993, Westinghouse committed to provide analyses demonstrating that the short SPES-2 pressurizer would have minimal effect on the phenomenology expected if the pressurizer were correctly height-scaled.
Provide that analyses (see ).
e.
At the meeting in April 1993 at SIET, Westinghouse was asked to justify the failure to include any operational transients or other key events, such as ATWS, in the SPES-2-program.
Provide that justification.
f.
The staff has taken the position that multiple steam generator tube ruptures should be tested in SPES-2. Westinghouse has responded by including a simulated 3-tube break in the SPES-2 matrix.
However, the staff has recommended that a rupture of up to 5 tubes should be considered, and the ACRS, noting the hexagonal configuration of the AP600 steam generators, has recommended that a 7-tube rupture is-appropriate (center tube and the six surrounding). 'In view of the concerns of the staff and ACRS, the staff recommends that Westinghouse include a simulated rupture of 7 tubes in the SPES-2 test program.
b b
t I
i
Comments on Westinghouse's Integral Test Programs OREGON STATE UNIVERSITY LONG-TERM COOLING TEST Scalina Westinghouse and OSU appear to have performed a scaling analysis, and then proceeded to use a relatively simple scaling methodology for much of the loop:
scaling pressure drops in the piping runs to be one-quarter of the calculated AP600 value.
While the methodology appears on its face to comprise a reasonable approach, there are several questions regarding its application.
1.
The component specification lists numerous piping runs and associated pressure drops.
It does not, however, indicate the conditions assumed for calculation of the pressure drops, e.g.,
single phase or two-phase flow; steady-state or transient; if transient, when during the transient; flow rates; Reynolds numbers; temperatures; pressures; and so forth.
2.
There is no indication whether all of the pressure drops were determined at a single operational point--steady-state or transient--or over a range of different plant conditions. The data on the sheet strongly suggest the latter.
For instance, the pressure drop across the CMT to Pressure Balance Line (PBL) No. 2 is shown as 1593 psi in the plant.
During normal operation, the pressure drop across this line should be almost zero, since the line is there for the express purpose of equalizing the CMT and the pressurizer. The only time that one would expect significant differences in pressure across' this line is during a break of the line. The same argument applies to the "CMT to Pressurizer Balance Line No. 1" pressure drop.
But both lines should not be broken in the same accident.
A brief discussion at the December 9, 1992, meeting indicated that these pressure drops may have Neen chosen on a " worst case" basis. The concerns that the staff has rega,- Ya such an approach include:
a.
The " worst case" accident scenario for one component will clearly not be the " worst case" for all components.
This seems to lead to an inconsistent scaling approach, since everything is being scaled to a different set of criteria, depending on the accident scenario.
Although it is impossible to satisfy all scaling criteria for all cases, and, therefore, compromises must be made, this does not appear to provide a satisfactory means for loop scaling.
b.
There is no indication what models or correlations were used to determine. these pressure drops, nor whether the correlations are applicable over the range applied.
For instance, using the CMT to PBL No. 2 as an example, a 2.624" ID and 5.08 cu. ft. volume gives a line length of about 135.3 ft.
This results in an L/D value for the line of about 619.
For the. scaled line, a 0.43" ID and 0.034 cu. ft.
volume gives a line length of about 33.7 ft, and an L/D of about 941.
Even assuming that the pressure drops listed are based on a " worst case" break that would result in critical flow at the break plane, depending on the location of the break, and how that location is scaled, the critical flow behavior for the OSU pipe may be
i
- 1 considerably different from that for the AP600 pipe, and different correlations could apply.
While frictional effects might dominate for break locations with large L/D values (far from the pipe inlet), this is a concern with shorter pipe runs and/or low L/Ds to the break, since in critical flow, behavior is dominated by inertial effects up to L/Ds of about 40, after whith frictional effects begin to come into pl ay.
Trying to scale the pressure drop using different correlations, or using a correlation beyond its range of applicability, could distort the results.
It is also worth noting in this specific case that the AP600 pressure drop is well below the maximum system pressure, while the scaled OSU value is about the maximum operating pressure for the test loop. This, too, could cause distortions, especially in two-phase flow behavior.
c.
This method also appears to ignore the important phenomena in a particular part of the piping network. The CMT to Cold leg Balance Line pressure drop values also appear to be based on some kind of j
critical flow calculation; it is hard to think of other scenarios where pressure drops of 500-600 psi could exist between the cold leg and the CMT. However, it may be more important, in that piping, to look at the single-or two-phase flow behavior during the recirculation and/or injection modes of CMT operation, not the critical flow behavior of the line during a break.
Taking the Norst case" approach for this line could distort system response.
For lines in which two-phase flow transitions are important, it may be more appropriate to use a Froude number (L/D*) approach, whereas for lines where critical flow behavior is important, it may be more reasonable to conserve L/D.
Looking at the values for'the pressurizer balance line discussed in (b) above, neither L/D (619 vs.- 941) nor L/D* (1002 vs. 617) is conserved, which indicate that considerable distortions in behavior may be seen, irrespective of scenario.
There is also an j
issue, for vertical lines, of conserving scaled elevation differences, i
which does not seem to be accounted for in the Westinghouse i
methodology.
d.
The pressure drops shown are simply point calculations, and they are i
really estimates, since there are no data on a real AP600 against which to compare the calculated pressure drop values.
If the values represent specific points during a transient, it must be recognized that these values change as the transient progresses, and that the worst case for a specific component may not correspond to the worst case for the system as a whole.
For instance, it would seem highly unlikely that the maximum pressure drop across the balance line (worst case for the component) would occur at the same time as a minimum water level in the reactor vessel (worst case for the system) would occur.
Furthermore, the OSU facility is supposed to be devoted primarily to studying the low-pressure portion of transients and accidents.
Choosing a scaling rationale based on conditions existing
-near the beginning of an accident, as would appear to be the case -for the CMT/ Pressurizer balance line, is inconsistent with the concentration on lower pressure, long-term response of the passive i
safety systems.
l
i
- The staff recommends that Westinghouse revisit the scaling methodology and see if there is not a more consistent way to size pipes. A possible alternative i
might be to use Froude number scaling for all pipes--or at least in horizontal and almost-horizontal lines for which elevation is not an issue--in which critical flow phenomena are not crucial or controlling, and to adjust pressure drops as necessary with orifices to meet the "one-quarter of AP600" criterion.
In addition to the considerations discussed above, Westinghouse does not seem to have attempted to scale the containment-related processes that are to be simulated as part of the OSU facility.
Since long-term cooling behavior is the primary objective of these tests, it seems appropriate to include long-term heat removal as part of the scaling analysis. While the containment itself is not being simulated, certain aspects of containment heat removal (condensation) and condensate recirculation are being included as part of the OSU loop.
It appears from the information available that the effluent from the break and ADS is simply being condensed completely and dumped back into the sump for recirculation to the primary system.
There does not appear to be any attempt to bridge the results from the OSU tests back to the one-eighth-scale PCCS tests.
If possible, some means by which to account for containment processes, e.g., heat transfer and condensate distribution should be included in the scaling analysis and in the test program, even if it has to be done parametrically.
Instrumentation The proposed instrumentation for the OSU facility appears to permit adequate data acquisition for characterization of loop response, e.g., flows, temperatures, fluid levels, pressures, during simulated transients and accidents.
Care must be taken, however, in the use of certain instruments.
For instance, measurement of single-phase flow across an orifice is acceptable; however, oscillations that cause pressure pulses through the system can distort such flow measurements to the point where they are meaningless, consideration should also be given to increasing hot leg / surge line instrumentation to acquire data on two-phase flow behavior.in the unique AP600 configuration (see additional discussion below on SPES-2 instrumentation).
The instrument list for the OSU facility contains over 400 instruments.
The scan rate listed in the test specification is a minimum of 70 channels per second.
At this rate, then, it would take approximately 6 seconds to complete a scan of the loop instrumentation.
The staff recognizes that some tests in the OSU facility may require several hours to run, and rapid data acquisition over the entire transient would generate far more data than it is practical to use. However, while it is to be expected that long-term loop response changes less rapidly than conditions near the start of a simulated accident or transient, it still appears that this minimum rate.is too low.
Consideration should be given to increasing the rate to allow a full scan of loop instruments at least every 5 seconds, if not more rapidly.
Ideally, it would be best to have as ' flexible a data acquisition system as possible, to allow rapid data acquisition at the beginning of a test, then shifting to a lower rate once it is clear that the long-term cooling phase of the test has been entered.
e Test Matrix The test matrix for the OSU facility is restricted, with a single exception, to small break LOCAs (the single exception is a large_ break LOCA).
There is no coverage of the other two major design basis events, i.e., steam generator tube rupture and main steam line break. The rationale presented to the staff for exclusion of these two accidents is that they are terminated while the plant is at relatively high pressure, and therefore do not fall within the primary purpose established for the OSU tests--low-pressure, long-term phenomena. These accidents will, however, be simulated in the SPES-2 tests.
As far as the coverage of SBLOCAs is concerned, the matrix is very limited in terms of break size. Most of the breaks are 2" equivalent, with a few double-ended guillotine ruptures and one 1" and one 4" break.
Westinghouse has maintained that a break of the order of 1" equivalent is the most limiting for passive safety system performance; preliminary analyses performed for the staff indicate that a very small break, with an inventory loss just larger than normal makeup capability, is a more limiting event.
It appears to be prudent to include a break of this size (approximately 3/8" equivalent, safety system response only) in the test matrix, since it seems unlikely that the apparent discrepancy between the staff's and Westinghouse's calculations will be resc1ved solely by analysis. The break location can be left open in the matrix, to be chosen based on analytical and prior experimental results once some testing experience is gained.
With regard to the decision to omit SGTR and.SLB events from this test matrix, the staff believes that this is not entirely prudent. As a minimum, the capability to perform tests of these types should be. included in the OSU facility. While it may not be necessary to include specific tests in the matrix at this time, a flexible approach should be considered, based on results from the SPES-2 tests.
If testing in SPES-2 and analyses of the AP600 design indicate that MSLBs and single and multiple SGTRs can be accommodated without actuation of the automatic depressurization system, these scenarios could be eliminated from the OSU program.
However, if SPES-2 results indicate that there is a likelihood that the ADS would be actuated for some of these events, it would be valuable to have data from OSU to back up and extend the information gained from SPES-2.
SGTRs are important because actuation of the ADS could potentially result in backflow of unborated water from the secondary side into the primary, which could both affect reactivity and flash to steam, interfering with passive safety system response. The MSLB event is of substantial interest because of resultant rapid cooling of the primary system, causing level shrink and a drop in pressure.
Early staff analysis of this event has indicated that primary system pressure could drop low enough to permit accumulators to inject without benefit of ADS actuation.
- However, should core makeup tank level drop low enough to cause the ADS valves to open, the blowdown process would be substantially different from other types of events due to the relatively low pressure at ADS initiation.
Again, the test in SPES-2 should provide insight into system behavior in the initial stages of the event; should the system not stabilize prior to ADS actuation, a follow-on test at OSU could provide data on the _latter stages of the accident.
1 l
.=
~
. There is also one aspect of AP600 operation that is not covered at all in any of the integral systems tests:
shutdown cooling, including both post-accident transition to shutdown cooling and accidents occurring at shutdown.
The AP600 is incapable of reaching conventional " cold shutdown" conditions using only passive systems.
At some stage after an accident or transient, a transition must be made to use non-safety active systems to help bring the plant to cold shutdown. After an accident, the procedure for making such transitions could be complicated by the nature of the event. Therefore, the staff recommends that Westinghouse revise the OSU test plan to include experiments in which the scenarios proceed to post-accident plant cool down, including bringing on-line any active systems that are required to accomplish this task.
Also, in the event of an accident occurring at shutdown, there will be instances in which passive systems will not be available, in either the short or long term, to replace RCS inventory and remove decay heat. The staff recommends that the OSU test plan also contain tests that involve shutdown accidents, demonstrating the use of necessary non-safety active systems to cope with such events, and also to investigate those scenarios in which there could be a transition to passive system operation, as those systems are enabled following a shutdown event.
These recommendations are in line with the increased regulatory attention to shutdown-related accidents and transients.
As discussed in the above section on scaling, a final area of consideration for the matrix is the simulation of containment processes to be included in the OSU facility.
Although these tests are billed as "long-term cooling" experiments, in reality they do not test the entirety of the long-term core cooling systems.
It appears from information given to the staff that all inventory lost from the reactor cooling system will essentially be dumped back into the simulated sump and returned to the primary loop.
This ensures an abundant supply of recycled cooling water.
However, in an actual event, some of the inventory may not return to the IRWST or the sump, and might therefore be unavailable to the passive cooling systems.
In addition, the temperature of the recycled water will change as a function of time.
These parameters change the gravity driving head available to return water to the core; for instance, a change in water temperature from about 100*F to 212*F results in a drop in density of about 3.5%.
Coupled with a decrease in IRWST water level or redistribution of fluid between the IRWST and the sump, this might cause the gravity-driven flow rate to drop by an even greater percentage. While the containment is not explicitly modeled in this test, the sump and IRWST are.
Accordingly, the return of water to the IRWST and/or the sump could be varied parametrically, to determine if there is some critical rate of return flow needed to maintain passive core cooling, and to provide experimental data over a wide variety of driving heads, water temperatures, and so forth.
RES-2 FULL-HEIGHT. FULL-PRESSURE TEST LOOP Scalino The scaling of SPES-2 is primarily a straightforward exercise in reducing component and pipe cross-sectional area either by the volumetric scale factor or using Froude number scaling, as described above in the discussion of the OSU loop.
The two areas in which questions have arisen involve the pressurizer, which is shorter in SPES-2 than in the AP600 design by about 4
a
. meters, and the cold leg configuration, which for SPES-2 involves a split cold leg with one pump, rather than the AP600 design of a separate. pump for each cold leg.
The pressurizer is still an area of major concern.
At the meeting.on December 10, 1992, Westinghouse presented a RELAP analysis of SPES-2 behavior with and without a properly scaled pressurizer.
The results indicated that the effect of the pressurizer scale on overall loop behavior was minimal. The staff is concerned that the analysis:
a.
was performed for only one break scenario, probably the most straightforward of the SBLOCAs that could occur in the AP600 design, and b.
was performed with a code that does not have validated models for the passive safety systems in the AP600.
The first of these issues is of considerable importance.
Pressurizer level and pressurizer response play very important roles in the AP600.
Pressurizer level is an initiator for CMT injection and provides a manometric balance against CMT level to control injection in the early stages of an accident; in addition, the ADS is attached to the pressurizer, so that all effluent passing -
through the first three ADS stages must flow through the pressurizer..The short SPES-2 pressurizer does not have the same level vs. volume characteristics as a properly scaled AP600 component.
CMT actuation and.
drain-down behavior may therefore be considerably different from that in the AP600, especially during more complex scenarios with,. for instance, loop-to-loop asymmetries. When the ADS is actuated, the flow passing through the ADS valves may also be quite different from that in the AP600 due to the differences in flow path between the plant and the test loop.
Some of these differences might be' accommodated if the shorter SPES-2 pressurizer were filled to the same level (in absolute terms) as that in the AP600. -'However, this is probably not practical, since this would leave an insufficient vapor-space in the pressurizer and would run the risk of overpressurizing the system during a level swell.
It would also put far too much water. volume into the pressurizer relative to the rest of the loop, which would create further _
distortions in response.
In order for Westinghouse to demonstrate that the-current pressurizer is acceptable, further analyses should be performed, with-the most current RELAP code and plant model, using a range of accident scenarios, including SBLOCAs in various locations _ and of several sizes, SGTRs, and MSLBs.
If these analyses continue to show minimal distortions in loop response, the use of the shorter pressurizer would be acceptable.
The split cold leg configuration is a considerable departure from AP600 geometry. There are a number of ways in which the cold leg configuration will impact system response during accidents. The most obvious of these, perhaps,
-1 is the resistance to flow from an unbroken inlet line to a break in the other i
inlet line on the same side of the plant.
Other such concerns include:
i a.
resistance to flow returning from the PRHR heat exchanger to the steam generator channel head, and from there to the cold legs.
. b.
interactions between the two non-pressurizer-side cold legs during CMT operation, particularly during accidents involving asymmetric loop behavior.
c.
behavior during transients, since the presence of additional piping in the cold legs adds inertance to the loop, which contributes to pressure losses during changes in flow rates.
It appears unlikely that this non-prototypicality in SPES-2 can be corrected.
Therefore, to the maximum possible degree, the pre-operational " shakedown" tests should have as a primary objective the characterization of the flow behavior in the cold legs.
This will permit the best choice of the compensating resistances placed in the cold legs to match the calculated behavior of the AP600.
Instrumentation In general, the instrumentation proposed for the SPES-2 loop looks appropriate, and appears adequate to be able to determine thermal-hydraulic conditions throughout the system during the experiments. The most difficult aspect of data acquisition is likely to be in the interpretation of the data, since two-phase flow behavior is notoriously tricky to measure accurately.
Since Westinghouse is using condensers to measure effluent flows, great care will have to be taken to determine the energy removed from these streams over the course of the tests, in order to give satisfactory mass-energy balances.
Many of the concerns related to AP600 response in tne early phases of-an t
accident have involved fluid distribution in the system; the proposed instrumentation lay-out should be capable of providing valuable data in this regard.
One area that should be examined closely in the instrumentation plan is the-hot leg / pressurizer surge line piping. The unique configuration of the AP600 in this area has led to concerns that there is not an adequate data base to be i
able to calculate two-phase flow behavior during transients and accidents.
If the SPES-2 hot leg and surge line are properly scaled, they could serve as a source of additional data.
It appears that the instrumentation proposed for.
these lines may be adequate from a systems standpoint, but not for acquisition of detailed data on flow dynamics, especially during ADS operation.
Additional thermocouple pairs to measure flow stratification and/or density measurement along the surge line should be considered, if practical.
The minimum scan rate of 2 samples per second per channel listed in the test specification for the SPES-2 loop should also be examined again.
Oscillatory behavior, especially during the initial phase of a simulated accident and during ADS actuation is likely to require a higher resolution than one sample every 0.5 seconds.
As was suggested for the OSU loop, a flexible system that could allow variable or programmable scan rates would be very. helpful.
However, it is expected that tests in SPES-2 would not continue for the durations expected in OSU, and a rapid rate of data acquisition (perhaps up.to 10 samples per second per channel) would not be unreasonable in terms of data production and subsequent processing.
. Test Matrix The SPES-2 test matrix includes a reasonably good selection of proposed experiments, covering as range of three of the four major design basis events:
SBLOCA, SGTR (single and multiple), and MSLB.
No LBLOCA tests are planned, and, based on analyses prepared for the staff, this accident does not appear to be most limiting for the AP600 design.
The staff has provided comments to Westinghouse on an earlier version of the test matrix, and some changes were made in response to the staff's recommendations.
However, there are some areas in which consideration should be given either to changing the test conditions or to adding further tests.
The main focus of these proposed revisions is the SBLOCA and SGTR tests.
However, there are additional subordinate issues, as well.
Eight SBLOCA tests are included in the matrix. Only three different break types are examined in SPES-2:
1", 2", and double-ended guillotine breaks (although the DEG breaks may be of various sizes, depending on the line that is broken). As noted in the discussion of the OSU tests, the staff and Westinghouse are not in agreement at this time as to the most limiting beak size for the AP600.
The staff has suggested previously that Westinghouse include a very small break, of the order of 3/8", in the SPES-2 matrix, and continue to recommend such a test, to allow examination of system response over what would be expected to be a very long transient at relatively high pressures.
There are three SGTR simulations planned for SPES-2.
Two involve a single SG tube break, and the third simulates a 3-tube rupture.
While the matrix refers to the 3-tube case as "beyond design basis," the design basis SGTR for the AP600 is still being studied by the staff, and the Westinghouse designation may be inappropriate.
The three SGTRs in SPES-2. all involve some. degree of non-safety system operation.
In the first, operator action using non-safety CVC and SWF systems is permitted to bring the loop to a stable post-rupture condition.
In the two other SGTRs, operator action _is not permitted, but both the CVCS and SFWS are actuated until high-water signals from the affected steam generator or low RCS T shut them off.
While this may be the " normal" ig plantresponsetoanSGTR,a,lossofoffsitepowerandfailureofthenon-safety-grade diesels to start would deprive these systems of power, and the plant would be forced to respond using passive safety systems only.
It is, therefore, strongly suggested that one of the single SGTR events and the multiple SGTR accident test be run without either CVCS or SFWS flow from the initiation of the event.
The subordinate issues in the test matrix involve break sizes and locations for the SBLOCAs.
Westinghouse's position is that a small break and a DEG beak of a small or intermediate line can bracket loop response, and that system behavior can be interpolated between those two conditions for other break sizes.
However, LOCA behavior is not necessarily a linear function of break size.
It is conceivable that, due to unforeseen systems interactions, a. 4" break of a particular line could be more limiting than either a 2" break of a DEG break.
This may be an area that is amenable to investigation by analysis, once some experimental data are available. ' Break location can also be a complicating factor in system behavior. As pointed out in earlier comments,
4 9,
4 4.
_g_
the location of the double-ended guillotine rupture of a direct vessel injection line can have a substantial impact on the response of the passive safety systems, since the loss of fluid from'the IRWST depends upon DVI line pressure as a function of time.
Consideration should be given to including tests to examine this aspect of loop response.
A final note on the test matrix was also included in the first set of comments: where a single failure is listed, the single failure assumed for all events except the MSLB is that of.a 4th stage ADS valve. The. staff requested Westinghouse.to (1) provide an analysis showing that the chosen single failure is the most limiting or (2) for those events without a single.
failure, justify the omission of this condition.
To date this has not been-done. The staff recommends that at least a scoping analysis be provided to address this question, particularly for the SGTR events for which no single failure is listed.
I I