ML20028G327
| ML20028G327 | |
| Person / Time | |
|---|---|
| Issue date: | 01/11/1983 |
| From: | Catton I Advisory Committee on Reactor Safeguards |
| To: | Boehnert P Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-CT-1548, NUDOCS 8302080416 | |
| Download: ML20028G327 (3) | |
Text
.,
U* / sty f/M DAW V3 MEMORANDUM 11 Jan 83 TO:
Paul Boehnert FROM:
Ivan Catton
SUBJECT:
Advanced Code Review Group Meeting, 14-15 December 1982, Bethesda, Maryland Lou Shotkin, RSR, raised a number of concerns for the revaw group to consider. He noted that void fractions being reported by the Japanese (SCTF and CCTF) were much lower than those reported by the FLECHT project. Here, I say reported because the void fractions are calculated from pressure drop data. A more careful look at run conditions should resolve this question.
The only direct measurements of void fraction have been at ORNL and at UCLA on their small scale bundle.
It seems to be that RSR should be able to re-solve these differences by a closer look at the experiments.
The COBRA code was to be the glue that put blockage experiments and the i
development of correlations for use in licensing together to address the clad ballooning issue.
It was found, only recently, that the COBRA code takes from 20 to 40 hours4.62963e-4 days <br />0.0111 hours <br />6.613757e-5 weeks <br />1.522e-5 months <br /> to run a reflood problem. The long running time makes its planned use unreasonably expensive. One can't help but wonder whether or not COBRA is needed and why so late in the game is it discovered that the running time is out of hand. Somebody was not paying attention.
Flow blockage studies may not be cost effective. When one looks at SCTF or CCTF results one cannot see blockage or grid spacer effects on the clad tem-perature.
From the ORNL experiments one can see that the grid spacer effect is local. One could calculate the core thermal behavior without considering grid spacers or blockages then correct the clad temperatures locally. Further, the j
chances of ballooning occurring are much less if one does a best estimate analy-sis. Ballooning may be a rare event. It seems strange that after years of re-search and millions of dollars that we can't answer the ballooning question in a more positive way.
Chapman, ORNL, discussed their studies of clad ballooning. He noted that azimuthal variations of temperature on a pin cause a great deal of scatter in the final strain-temperature results. Very small temperature differences; l
as small as 5'C, were noted as being enough. No thermal hydraulic code will ever be able to deal with this.
It seems to me that we probably have more bal-l looning data then we need.
The Zire deformation process is very fast compared to other processes (thermal hydrauli-).
This means that if it is to be calculated, it can be l
treated as a quasi-steady process. Codes that do more than this are probably l
gu!1 ding the lily. Trying to do such calculations is, in part, the problem with COBRA.
On seeing how the balloning data was interpreted to obtain flow blockage, I cannot help but conclude that NUREG-630 is overly pessimistic. This work should be reviewed and more realistic flow blockage estimates should be ob.
tained. The review should be by somebody who understands fluid mechanics--not a materials expert.
I was surprised to hear that blockage pressure drop measure-DESIONU D ORIGIM L B302000416 830111 i
1 PDR 7
u
/$
=
t,,;
/
Catton to Boehnert 2 of 3 11 January 1983 ments had been made at ORNL last summer but that the data had not been reduced.
This is a _ key piece of data. With it, one can estimate how much flow there is through the blockage and make estimates of the heat transfer.
Hochreiter, W, argues that there is a need for more work and testing.
It is my feeling that the present FLECHT, SCTF and CCTF data are enough. There were also arguments about the need for more detail. There may already be more-than we can handle. His stated goal is Appendix K relief on blockage. Without hearing the full story, I think the problem is the non-real world of Appendix K, not a need for inore data. Whatever further results are obtained, they will be sufficiently different from what is expected in-core that we cannot resolve the issue. The present FLECHT progran should devote more effort to obtaining flow blockage pressure drop--bypass flow relationships.
Kelley, pNL, described COBRA-TF flow blockage model development.
PNL has decided that detailed deterministic analysis is impossible.
(I would agree).
One simply cannot predict temperatures accurately enough when one considers the sensitivity of zircaloy to temperature differences as small as 5'C.
Even so, the code takes up to 40 hours4.62963e-4 days <br />0.0111 hours <br />6.613757e-5 weeks <br />1.522e-5 months <br /> to run. They are, therefore, proposing a "new approach (?)." They will use lumped channel local fluid conditions and an em-pirical fuel rod blockage model.
I really don't understand what the purpose of the new approach is.
If the blockage can be described, it seems to me that one can judge whether it is cool-able based on existing experiments given the boundary conditions. There is ex-perimental data available that demonstrates that if one knows the area ratio change caused by the blockage and the local flow through it, one can estimate the heat transfer very satisfactorily. This says we need the blockage K factors!
The "new approach" to be true should focus on demonstrating this.
It is a ques-tion of bypass.
Williams, LASL, carried out an interesting study that in many respects confirmed observations others have made about high void fraction heat transfer.
His calculations, in my opinion, show that droplet enhanced turbulence is the key to a good prediction rather than fancy non-equilibrium considerations.
At the conclusion of the two-day meeting, thr group was asked for sugges-tions as to how the flow blockage question should be addressed. My comments are mostly made in the above text but are re-iterated here for emphasis. They are as follows:
- 1) Measure blockage K-factors (pressure drop-velocity proportional.ity re-lationship).
- 2) Test the hypothesis of Dhir that knowing the flow through the blockage and its area ration one can predict the heat transfer coefficient to within twenty percent.
- 3) Use the COBRA code in its new form to test 1 & 2 against data from the eight facilities studying reflood.
- 5) Resolve the apparent differences in void fraction measurements between FLECHT, SCTF and CCTF by more careful data reduction.
- 6) Do not try to satisfy or give relief from Appendix K in its present form.
- /
Catton to Boehnert 3 of 3
(
11 January 1983 As a postscript I would like to note that, as I understand, certain committments were made to Westinghouse in return for various proprietary models that are now in COBRA and as such in the public domain.
It seems that Westing-house was interested in Busing the COBRA code with their models so that they would get proper verification for later use in the licensing arena. A number of questions arose at the meeting regarding the usefulness of the program as present-ly planned.
I certainly raised some myself. Rather, however, than terminating the FLECHT-COBRA plans at Westinghouse, as recommended by some, re-direction should be sought so that resources are not wasted and unnecessary bad feelings are avoided.
-..