ML20112G671
| ML20112G671 | |
| Person / Time | |
|---|---|
| Site: | 05200004 |
| Issue date: | 06/07/1996 |
| From: | Quinn J GENERAL ELECTRIC CO. |
| To: | Quay T NRC (Affiliation Not Assigned), NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| MFN-063-96, MFN-63-96, NUDOCS 9606120185 | |
| Download: ML20112G671 (9) | |
Text
,
GENuclearEnergy MR nYBW 5kuttn M5165 San Jose, CA 951251014 gams Aven 408 925-1005 (phone) 408 925-3991 (facsimile)
June 7,1996 MFN 063-96 Docket 52-004 Document Control Desk U. S. Nuclear Regulatory Commission Washington DC 20555 Attention: Theodore E. Quay, Director Standardization Project Directorate
Subject:
SBWR - STAFF REVIEW OF GE'S SCALING REPORT NEl>C-32288P, REVISION 1, " SCALING OF SBWR RELATED TESTS" RELATED TO CONTAINMENT AND REACTOR SYSTEMS AREAS
References:
1.
Letter, T. R. Quay (NRC) to J. E. Quinn (GE), StaffReview ofGE's Scaling Report NEDC-32288P, Revision 1, " Scaling ofSBWR Related Tests" related to Containment area, dated April 24,1996.
2.
Letter from T.R. Quay (NRC) to J.E.Quinn (GE), Staff Review of GE Nuclear Energy (GE) Scaling Report NEDC-3228HP, Revision 1, " Scaling of SBWR Related Tests" related to Reactor Systems area, dated May 14,1996.
3.
Letter, J. E. Quinn (GE) to T. R. Quay (NRC), SBWR - Transmittal of Scaling Report Revision, Scaling of SBWR Related Tests, NEDC-32288P, i
Revision 1 (Proprietary), dated October 13, 1995.
We have received the referenced letters 1 and 2 on the NRC Staff's review of the Scaling Report NEDC-32288P (Reference 3). We are pleased that the staff concluded that the GE scaling methodology "is a systematic application of the hierarchial two-tiered scaling method and, therefore, is acceptable"(Reference 1) and "the general scaling approach by GE appears to be acceptable" (Reference 2). The staff has also ideatified a number of comments and concerns.
The concerns raised by the Containment Branch are summarized in the form of fourteen questions in Attachment 2 to Reference 1. The Reactor Systems Branch comments are listed in the Enclosure to Reference 2.
GE has reviewed the concerns raised by the staff (primarily the Containment Branch) and their consultants, and concluded that there are no issues that either impact the validity of the scaling approach orjeopardize the usefulness of the data for its intended purpose. The concerns fall into the following categories: confirmation of the scaling using the test data from the facilities (which was not available when the scaling methodology was developed, and cannot be a basis for a priori facility scaling); relationship to TRACG qualification (which is a pertinent issue, but not related to facility scaling, and will be addressed in the TRACO Qualification Report); additional 9606120185 960607 Ol PDR ADOCK 05200004 v
A PDR
- {
GENuclearEnergy information/ clarifications and closure with the PIRT rankings; and identification and evaluation of scaling distortion 3 resulting from " power / volume scaling". None of these comments /
questions indicate any issues which would invalidate the scaling basis used for facility design and testing.
Some of these issues were also raised in the ACRS meeting in November 1995. Subsequently, GE prepared additional material to respond to three items which are in the staffletters : an executive summary highlighting the main points and conclusions from the report; a discussion on scaling distortions; and a quantitative evaluation of the effects of heat storage in structures. This information will be provided to the staff when the review of the SBWR is resumed. We have attached a response to the fourteen questions in Attachment 2 to Reference 1, indicating how the questions would be addressed. The attachment also addresses the comments in Reference 2.
Thus, GE believes the remaining questions can be successfully resolved if the SBWR review were to be resumed, and that the primary purposes of the Scaling Report have been satisfied.
If you have any questions regarding this issue please contact Bharat Shiralkar of our staff on 408-j 925-6889.
Sine rely, a ' p/)
f>0
. Quinn, ects Manager Attachments:
cc:
P. A. Boehnert (NRC/ACRS) (2 paper copies plus E-Mail)
I. Catton (ACRS)
(1 paper copy plus E-Mail)
S. Q. Ninh (NRC)
(1 paper copy plus E-Mail)
D. C. Scaletti (NRC)
(1 paper copy plus E-Mail) l
GENuclearEnergy MFN 063-96 bec: (E-Mail only except as noted)
R. Asamoto P. F. Billig R. H. Buchholz 1
T. Cook (doe)
(1 paper copy plus E-Mail)
J. D. Duncan R. T. Fernandez (EPRI)
J. R. Fitch J. N. Fox R. E. Gamble P. C. Hecht J. E. Leatherman J. E. Quinn i
T. J. Mulford (EPRI)
(2 paper copies plus E-Mail)
F. A. Ross (doe)
(1 paper copy plus E-Mail)
B. Shiralkar G. Sliter (EPRI)
R. Srinivasan (EPRI)
J. L. Thompson J. E. Torbeck GE Master File M/C 747 (1 paper copy plus E-Mail)
SBWR Project File (1 paper copy plus E-Mail) 1 l
l t
GENuclearEnergy l
i Responses to Questions in Attachment 2 of Reference 1 General Comment l
GE has reviewed the "Roadmap for Closure" outlined by the staff, and believes that closure could be reached as suggested. However, we take strong exception to the characterization of the large full-height test facilities, which have been carefully scaled, and constructed at a cost of millions of dollars, as " scale model pilot experiments".
1.
In SBWR, the upper and lower drywell are connected by a number of vents. Breaks can i
occur in the lower drywell as well as the upper drywell. The PANDA facility has two drywells separated by vents, but these do not seem to be representing the upper and lower drywell. Please explain how the multi-chamber drywell is simulated in the scaled tests and what the two drywells in PANDA are simulating. A better explanation of the rationale for
" tube and tank' test facilities is needed.
L
Response
Any breaks in the lower drywell will be small liquid breaks. The steam discharge to the containment in these cases also is dominated by the flow through the DPVs to the upper drywell. The only difference in this situation is the potential for earlier release of noncondenaibles from the lower drywell. This issue is discussed in response to Question 10.
A multi-chamber drywell (i.e. separated lower and upper drywell regions) is not simulated in the PANDA tests. The two drywells in PANDA simulate regions of the drywell that are diametrically opposite and allow the creation of multi-dimensional mixing and distribution effects. The diameter of each drywell vessel is approximately equal to the hydraulic diameter of the drywell annulus in the SBWR The distance between them represents j
approximately the full scale diametral separation between the opposite sides of the drywell annular region. This allows for the simulation of three-dimensional effects in drywell temperatures and noncondensible gas concentration corresponding to maximum azimuthal asymmetry of steam discharge with respect to PCC location. See the response to Question 6 for specifics.
I 2.
None of the tests simulate thermal stratification in the suppression pool. In fact, a portion of the lower suppression pool is not included in the scaled facilities. GE should explain how this will be addressed in using the test data to validate containment design computer codes.
I J
.. 3 GENuclearEnergy Response -
The SBWR Test and Analysis Program Description (NEDC-32391P, Revision C) lists the PSTF tests being used for validation of the thermal stratification model. The TRACG Model Description (NEDE-32176, Rev.1) Section 7.11 shows a comparison of the model against thermal stratification data. Both these reports are under review by NRC. The PSTF tests capture thermal stratification in the suppression pool during the short-term or blowdown phase of a LOCA.
The SBWR tests, PANDA and GIRAFFE, can in fact simulate, if picsent, thermal stratification during the long-term phase. Both facilities model the ptential flow of hot gases from the drywell to the wetwell via the top LOCA vent or through the PCCS vent.
Condensation of steam at the exit of these vents can result in thermal stratification. The lower part of the suppression pool is not necessary to simulate this phenomenon in the long term since the water in that region is not expected to mix with the upper layers.
3.
Equation 2.5-6 on page 2-40 of NEDC-32288P, Rev.1 appears to be incomplete. Please verify and provide the complete equation.
Response
An Errata has been compiled which includes the correction to this equation.
4.
For report NEDC-32288P, Rev.1, provide a complete list of the input parameters for the PCCS phase (in addition to Table 4.1-22) sufficient to allow an independent calculation of the value of each pi-group. Where ajudgment is made in selecting an input value provide a justification for the value selected.
Response
GE will provide such a list when the review is resumed.
5.
Provide an explanation of how the test facilities are " preconditioned" to achieve the correct energy partition in the various parts of the system at the start of the test simulation. Address initial conditions of the pools, drywell, vapor spaces - pressure, temperatures, stored energy.
Response
The approach for specification of the initial conditions is described in the Test and Analysis Program Description (NEDC-32391P, Revision C). The initial conditions and the details of the procedures for achieving the initial conditions are available in the Test Plans and Test Procedures for each facility. These documents have been previously sent to the NRC.
6.
GE scaling is based upon the lumped parameter approach which eliminates all spatial variations within the nodes. The test design itself eliminates multidimensional effects with
GENuclearEnergy its large length to diameter ratio. Also, the instrumentation is not designed to measure multidimensional effects, ao the design, the scaling and the instrumentation are all one-dimensional. Therefore, the data will be useful only for validating computer codes based upon one-dimensional representations. Multi-dimensional effects must be treated in a bounding manner. Additional explanation is neededjustifying that multi-dimensional effects can be bounded when extrapolating from the scaled tests to the prototype.
Response
GE scaling is based on the hierarchical two-tiered scaling methodology, which has been found acceptable by the NRC staff as part of this review. In the application of this methodology to SBWR test scaling, parameters of a global nature are scaled ' top down',
while the detailed multidimensional effects are picked up from ' bottom up' considerations.
This cannot be called " lumped parameter scaling". Multidimensional effects are not precluded in the PANDA test facility; on the contrary, the facility was designed to potentially produce such effects. The aspect ratio compared to the prototype is different by a factor of 5, not a huge difference. Moreover, as explained in response to Question 1, the two drywells can be configured to provide an estimate of the effect of the maximum asymmetry in blowdown location versus steam removal. Three tests (M2, M10a and M10b) were specifically run to study this effect. This data can be used to bound multidimensional effects related to the mixing of the noncondensibles and steam in the drywell and their effects on the heat removal by the PCC condensers. The two PANDA wetwells, which have connections in both the gas and liquid spaces, also allow the creation of multidimensional effects. PANDA has extensive instrumentation such as thermocouple strings at central and off center locations which provide multidimensional information. NRC staff have been involved from the beginning in reviewing the facility instrumentation.
7.
Please provide a copy of the report on vacuum breaker scaling.
Response
GE will provide this report when the review is resumed. The report is an example of the very detailed bottom-up scaling considerations made in the design of the PANDA facility components.
l 8.
The current scaling for the long term (Phase 4) has a single node, the drywell. Why are the suppression pool and wetwell not included. It would seem that important parameters, including the wetwell pressure and liquid vapor space temperature, etc. should be included.
How do you justify not including these in the scaling? Please provide a long term (Phase 4) scaling analysis which addresses the wetwell and suppression pool regions.
i GENuclearEnergy
Response
For the long term phase, the pressure rate scaling was based on treating the suppression chamber (pool and gas space) together. For certain situations, for example, with bypass leakage from the drywell, the assumptions of thermal equilibrium may not be accurate. GE will refine this part of the analysis when the review is resumed.
9.
There is essentially no closure with the PIRT. A large number of pi groups are included in the final approach which makes it difficult to identify the important variables, and to focus on a figure (s) of merit. GE should provide closure with the PIRT by using the information generated by the experiments to confirm or modify the importance of the phenomena identified in the PIRT Key non-dimensional parameters (based upon the important pi groups) should be clearly identified and plotted for each test facility and compared to calculations for the prototype as further discussed in item 11 below.
Response
GE will highlight the most important conclusions and scaling groups in an executive summary. The top-down scaling of the facilities addresses all relevant parameters and this necessarily results in a multiplicity of pi numbers. Although the scaling of the phenomena identified by PIRT has been addressed by the GE scaling effort, there has not been a display of one-to-one correspondence. The closure on the PIRT as requested can be achieved with additional effort. GE believes that this effort is not cost effective and will not result in any consequences relevant to plant safety. GE will include nondimensional comparisons of data with calculated SBWR transients in the Summary Test Report.
i 10.
All of the GE SBWR test facilities are scaled in a similar manner (full height, power to volume approach). Important processes, such as losses to ambient and condensation on structural heat sinks will be distorted in the same manner in all of the facilities, making it impossible to determine the effect of scaling distortions on these processes. Therefore, a bounding treatment will be necessary. GE should present the bounding manner of treating these phenomena based upon the experimental data.
Results While it is true that all GE SBWR test facilities are scaled in a similar manner, there is a considerable difference in size between the GIRAFFE (1:400) and PANDA (1:25) i facilities. Because of these differences, the distortions will be present to different degrees, and information can be obtained about their relative importance. (The prototype is also a full height, power to volume scaled (1:1) facility in the ultimate extrapolation). The effects oflosses to the ambient and condensation on structural heat sinks are fairly straight forward to calculate both in the prototype and the test facilities, and this is the approach taken in considering such effects. Furthermore, the PANDA facility has been deshned to keep the heat losses very low.
GENuclentEnergy It is more difficult to scale the effects ofinternal flow distributions on the noncondensible distribution, and GE agrees that this effect needs to be treated in a bounding manner. The following discussion outlines a strategy to achieve this end.
During the blowdown following a large pipe break (or following DPV actuation for a small break), most of the noncondensibles in the upper drywell and annular region are swept to the wetwell. The flows are large enough to guarantee good mixing in these regions. Some noncondensibles will be drawn back from the wetwell during the vacuum breaker actuations that occur following GDCS injection. After the first hour, the noncondensibles in the drywell are continuously diluted to a very small value by the continuing discharge of steam. Bottom up scaling (Section 3 of the Scaling Report) indicates good mixing in the upper drywell. This is also confirmed by the PANDA tests, even for the extreme bounding case of three tests with asymmetric injection of steam in one drywell. This will lead to an exponentially decreasing noncondensible concentration in the upper drywell, decreasing rapidly to negligible values. As seen in the PANDA tests, the noncondensibles were quickly purged to the wetwell and subsequently the PCCs had a substantial overcapacity for decay heat removal. The effects of vacumn breaker openings have already been demonstrated by the PANDA tests to be transient blips with no long term impact on the PCC heat removal. The only multidimensional, multicompartment effect of consequence which is not present in the PANDA tests, is the bleeding of noncondensibles from the lower drywell and other restricted regions (space above GDCS tanks) and its effect on the PCC perfomlance. This can be bounded in the following way : The pertinent variable is the history of this noncondensible bleeding through the PCC condenser units. As the initial volume of noncondensibles in the drywell compartments is bounded, different trajectories of noncondensibles through the PCC can be postulated and the worst case obtained, with respect to the integrated heat removal and wetwell pressure. With this bounding approach, internal complexities such as recirculating flows, etc. are of no consequence.
- 11. The presentation of the results by GE does not address some key issues regarding suitability of the experimental data for computer code validation. In particular, the most important dimensionless groups need to be identified and transient plots presented which show comparisons of the variation in these key dimensionless parameters (based upon the pi groups) for the various experimental facilities (data) and prototype (calculations).
Response
There appe trs to be a fundamental disagreement between different groups of consultants on how much information from a computer code should be used in the scaling process. In accordance with admonitions from the ACRS consultants, GE has adopted the view (that GE considers correct) that scaling should not be based on code-calculated comparisons between prototype and model behavior, and has limited the use of TRACG to the definition of certain initial test conditions only. Rather than compare transient plots calculated from
+
4 GENuclearEnergy TRACG for the SBWR and the test facilities, we have used scaling comparisons at selected points for different phases of the transient, using well defined initial conditions for each phase. We believe this is adequate. Confirmation of the correctness and relevance of the l
scaling, in that the test facilities do not produce non-prototypical behavior unexpected in l
the SBWR, can now be obtained by comparison of the data from scaled test facilities with the TRACG calculated SBWR response.
12.
The impact of distortions also needs to be related to the degree ofimportance of the phenomena distorted. GE should provide a list of all known distortions introduced by the scaling of experimental facilities and provide an evaluation of how each distortion impacts i
safety of the design.
l
Response
An evaluation of the distortions has been performed. This information will be provided when the review is resumed. The only item of significance is the potential efTect of the.
absence of noncondensible gas " hideout" volumes, which can be addressed as outlined in 1
response to Question 10.
l i
l 13.
GE should provide a "roadmap" of TRACG qualification and validation efforts as the present "roadmap" relies only upon facilities scaled similarly and hampered by the same distortions.
Response
Please refer to the Test and Analysis Program Description (NEDC-32391P, Revision C),
which provides a detailed listing of the TRACG validation plan. This document is currently under review by NRC. Please also refer to the Responses to Questions 6 and 10.
- 14. As there is little effort concerning bottom-up scaling, GE is requested to provide a -
"roadmap' of TRACG qualification and validation against separate effects tests featuring i
phenomena not scaled in the integral test facilities.
Response
l Please refer to the Test and Analysis Progran: Description (NEDC-32391P, Revision C),
which provides a detailed listing of the TRACG validation plan. This document is currently under review by NRC.
l l
I i
. -