ML20237L066
| ML20237L066 | |
| Person / Time | |
|---|---|
| Issue date: | 07/03/1987 |
| From: | Catton I NRC |
| To: | Houston D NRC |
| References | |
| ACRS-CT-1894, NUDOCS 8708270405 | |
| Download: ML20237L066 (3) | |
Text
_
e TO:
Dean Houstop 3 Jsly 1987
~
g FROM:
Ivan Cattorr
SUBJECT:
Direct Containment Heating Research Review Group Meeting [O 18-19 June 1987, Bethesda M)
O '8 b-The meeting focused on energy. transfer to the containment atmosphere given that the core melt resulting from a severe accident has breached the vessel. Sandia (SNL), Brookhaven (BNL), and UKAEA gave presentations summarizing their progress in understanding the phenomena and their future plans. The presentations were seperated into 1) analysis, and 11) experimental efforts.
ANALYSIS SNL--The CONTAIN code, developed by SNL, is being used with a containment heating model. The code is a lumped parameter code that has been ubder development for some time. It is probably the best code of its type available. The present " Direct Heating" model is simple yet probably as good as our understanding of the phenomena can justify.
Some improvement in modeling of the thermophysical processes may be justified to properly account for cavity mass and energy losses.
A code called KIVA is being proposed to carry out 3-D turbulent flow cniculations in the cavity. In my view this is not a justified approach to obtaining meaningful estimates of the response of the containment to direct heating.
BNL--Support for the SNL approach was given with emphasis on the need for a 3-D finite difference model of the turbulent multiphase process resulting from vessel blowdown into the cavity and transport to the containment. Arguements were given as to why one should not use lumped parameter modeling such as is available in CONTAIN.
UKAEA--Simple models are being developed for use in a 1-D gas flow code. Parametric treatment of most of the complex phenomena will be incorporated into the modeling. The main models are an isentropic vessel, corium-gas or two phase flow, hole ablation, gas side limited metal oxidation, radiation and convective. heat transfer, structural erosion by a corium jet, crusting on cavity surfaces, and 1-D gas flow.
First principal analysis of the DCH is a difficult if not impossible task. The experiments, see below, are difficult and measurement of the parameters needed to describe the micro physics that detailed modeling is based on will require extensive instrumentation development. In my view, proper balance is obtained by using a code like CONTAIN with a sensible use of the results of present experimental programs. The important observations are the following:
- 1) more than 75% of the ejected mass gets out of the cavity under all circumstances, and
- 11) most of the available thermal and chemical energy is converted to containment atmosphere thermal energy.
The reduction in uncertainty associated with DCH will not come from more complex modeling. Rather, more complex modeling will probably 8708270405 B70703 D EIGNATED ORIGlBAL PDR ACRS CT-1894 PDR N t m py
increase the uncertainty because of the large number of assumptions i
needed te obtain the various details of the physics needed for such l
modeling.
l interesting comment can be found in the minutes of a meeting sent An to the Severe Accidents Sbete. Ken Bergeron said a code like the one i
being considered, 3-D with turbulence, recommended above will cost millions to develop, and that he would like to bid on it.
EXPERIMENTAL STUDIES SNL--The analytical models under development will be based on the SNL DCH and SURTSEY series of experimental runs. With our understanding of scaling severe accident phenomena from the SNL small scale experiments to full scale being weak, it is not clear f
that the results to be used for code assessment are adequate.
Dispersal measurements cannot be reliably scaled until we understand the phenomena ve must predict. The experiments planned do not yield details of what happens in the cavity. The amount of mass in the melt is equivalent to a rather large fraction of the core.
Experiments with lesser amounts need to be considered as well. At
(
present the melt temperature is unknown and thought by some to be
{
too high. Disagreements about the chemistry of the melt interactions hydrogen need to be settled for the test results from with steam or the DCH or SURTSEY test series to be meaningful. Even with such questions answered, it is not clear that the expense of the experiments can be justified.
BNL--A series of experiments are being carried out with a PWR cavity i
simulated by a small scale plastic model and the molten core materials represented with molten woods metal. There are lots of reasons to question the results but they will be relatively inexpensive to obtain and will give a good qualitative picture of
)
the process.
1
SUMMARY
l BNL are proposing a finite difference 3-D code be developed SNL and with a two equation model for turbulence. To accomplish such modeling one must understand the physics at a micro level. The present To obtain the experimental programs will only yield macro date.
needed data, will require an experimental program of massive proportion along with instrumentation development of unknown cost.
The The code development program will be lengthy and expensive.
CONTAIN code is argued to be inadequate yet, in my view, is the loading.
only sensible method of calculation of containment With present studies showing that between 75% and 100% of the available thermal and chemical energy transfers from the core melt materials to the containment atmosphere, the question of how much melt must be dealt with and what its temperature is become Both Paul Shevmon and I have preached of paramount importance.
on this subject at length. The research focus should be changed to l
l 2
address questions where the answers will be helpful to the decision making process.
HRR has made an appeal for help. It seems to se that RES should be responsive by focusing their program on areas where it will do some good. A very simple model vill be satisfactory for analysis of the ex-vessel behavior of the core following a high pressure
- j. g.
8
- E ejection into the reactor cavity.
g*
v.
u i
l 3
4
____- ___ ________-______ -