ML20148H241
| ML20148H241 | |
| Person / Time | |
|---|---|
| Issue date: | 11/30/1987 |
| From: | Catton I Advisory Committee on Reactor Safeguards |
| To: | Boehnert P Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-CT-1909, NUDOCS 8801270177 | |
| Download: ML20148H241 (2) | |
Text
-
eT-/909 7h n atou
- o o
UNITED STATES "g
NUCLEAR REGULATORY COMMISSION g
{
,e ADVlsORY COMMITTEE ON REACTOR SAFEGUARDS WASHIN GT ON, D. C. 20555
.J o,%,******&,g N" A 24 U.SJLR.C.
E Ti1 C?)
November 30, 1987
~
~
@ f,,Gfff MEMORANDUM FOR: Paul Boehnert FROM:
Ivan Catton
SUBJECT:
Technical Program Committee Meeting Code Uncertainty Quantification (CSAU)
Nicolson Lane Bldg 30 October 1987 A number of concerns about nodalization were raised or have been raised by Wulff. They will be addressed at the 3-4 December TPG meeting. I will report on their disposition following the meeting.
Some results of the UPTF studies were reported. The TPG is now getting down to the hard work. All UPTF studies will be done with four azimuthal nodes. Although studies to support four or eight azimuthal nodes have not been done, LANL argues that eight are needed to capture the physics. My observations of the phenomena lead me to the same conclusion. When one considers the errors in differencing the momentum convection tenns found by the English, see my report on the ICAP meeting, one can only conclude that compensating errors and dialing ("tuning") have compensated for the lack of physics. Here one might argue that one should leave well enough alone. There is, however, some experimental evidence indicating that the broken leg location does not seem to impact on the results very much, leading to the argument that four nodes is enough. Even in the face of the arguments given, I would like to see more analysis before accepting the conclusion that four azimuthal nodes can be used without a clear and well defined penalty to the user.
Several blowdown peak clad temperature uncertainties have been obtained by different methods. They are a) varying parameters in the code one by one and adding their individual contributions (a very conservative approach) yielding 1128 deg K b) compare code calculations with experiment yielding *70 deg K 8801270177 871130 DESIGNATED ORIGINAL
-19 PDR Certified By
/
- 1 P. A. Boehnert November 30, 1987 c) upper and lower bounds on all experimental data plotted on the
'same curve yielding 1375 deg K These resuits are very encouraging when one notes that the calculated PCT is very low relative to the Appendix K limit of 2200 deg F.
When calculated blowdown PCT from Thi.1 version 12.2 were compared with version 14.3 it was found that the PCT was reduced 50 deg K. This says a package of data are needed for testing changes ir. the codes.following their implementation.
Determining the uncertainty in reflood calculations will be much more difficult than for blowdown. Not only are there more phenomena to con-sider, but in some cases the physics in the codes is wrong and is tuned to improper data. To get around the problems, code will be fooled into calculating refill correctly for UPTF. The procedure will be used to calculate a reflood PCT for a PWR and then compared with an as found form of the code to establish the impact of refill on reflood PCT. A lot of diddling with the code will be neccesary to ge*: the heat transfer right.
Tmin is still a problem. I sent Zuber results of work done several years ago that might be helpful. I think the bottom line will be comparisons with SCTF and CCTF data. Here I hope all calculations are done with the same nodalization as will be used on full scale PWR calculations. If not, I think all the diddling will come to haunt us.
l
.