ML19209C285
| ML19209C285 | |
| Person / Time | |
|---|---|
| Issue date: | 06/27/1979 |
| From: | Catton I Advisory Committee on Reactor Safeguards |
| To: | Plesset M Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-CT-1166, NUDOCS 7910120364 | |
| Download: ML19209C285 (6) | |
Text
" /~/ M c
- e..i
,,, ; * **. ' *J, y
..u....
ADV150RY C0Y.M:T1LE 6 aEACTOR SMEcMRCs U.5 ma^
Boelter Hall 2567 University of California AUG 1 1979 Los Angeles, California 90024 0
N Elt$hl:5:6 June 27, 1979 l
,,3t g0t11t 9
t To:
M. S. Plesset DISTRIBUTED TD ACR$ MEMBERS From:
Ivan Catton
Subject:
ECCS Subcommittee Meeting,19, 20 June 1979, Washington, D.C.
's Suma ry.
Presently available computer codes such as RELAP, CRAFT, RETRAN, and even the B&W Simulation, do reasonably well in representing a power plant. The steam generator modeling and how quickly natural circula-tion might'be lost due to voids need more attention. Boundary conditions for the computer studies don't have enough operator characteristics in them as yet.
Much of the requested TMI-2 supplementary funds are not needed.
The proposed areas are too technical and seem to leave out the human aspects.
A number of the research projects underway in the LOCA area deserve serious attention.
It is not clear to me that cost effectiveness has been a factor.
Reduction in the level of effort on several of the larger programs should more than free up enough funds for areas brought to our attention by TMI-2.
B&W Small Break Analysis. As I view it, there are two parts to a small break analysis. First, one must have a system model that is sufficiently valid to use for accident studies.
Second, one must exercise it properly.
Es,th aspects came up during the meeting and will be addressed separately.
There are two schools of thought regarding the small break. Some believe that a highly sophisticated advanced code is needed whereas others believe a simple code with a lot of thought is better.
A justification for the second approach is Michelson's description of the small break given in in l i'.
1 7910120 b
+
t his January 1978 report. Further confimation of the simulistic view is given by B&W analysts. They found that the process is er.sentially quasi-static.
The simple view does not mean that the componer.ts are easily modeled. Rather, some of the non-linear time advancement prchlems need not be faced.
To hope that an advanced code will be able to follow the course of the a-cident past core damagg is over optimistic and may be even foolish.
Present codes do quite well in following the TMI-2 incident out to the loss of natural cirulation. This means they can be used With some comfort for most accident analysis. Beyond the loss of na'tural circulation points, a BWR type model with subccoled feedwater and superheated steam is needed.
Several aspects still, however, do need attention.
At present the B&W steam generator model is overly simplistic. A single heat transfer coefficient is used.
They argue that for the T?il-2 incident, a plus or minus fifty percent variation in heat transfer will not effect the results singificantly. Based on the RELAP modeling of TMI-2, it is not clear that this is so. More attention needs to be given to the steam generator model.
Natural circulation is lost when sufficient steam and other gases collect 3
at the top of the candy cane (on the order of 80 f t 1 The present models (both B&W and INEL) use a bubble rise model of some sort and a single node at the top. At the low veucities anticipated during natural circulation, a much simpler and, in my opinion, more realistic approach qould be to collect all the rising steam and gases in the upper most node.
Then, when the volume reaches 80 f t, turn off natural circulation.
The codes somehow split the voids leaving the core so that some enter th'O hot 6g and some collect.n the upper plenum.
It's not clear how this is done.
The upper 3
~
plenum is over 1000 ft. As a result, if only a small fraction enters the 4.
3.,
)
2 fI '. _.
hot leg, loss of natural circulation could occur befor the upper plenum is full of steam.
Present calculational tools follow the early course of the TMI-2 accident reasonably well.
The question becomes one of deciding what the boundary conditions on an event should be.
It is my view that the boundary conditions are the result of 'corhplex interactions between the operators, the operators inter 7retation of his plants emergency procedures, the infor-mation available to the operator and the phsycial system he is attempting to control.
In trying to mimic an incident on a computer, one can look at the period of time when certain steps allow recovery without damage or one can try to folicw the course of an event that results in damage. At present the complex operator interaction is not being dealt with properly.
It is my view that it is the analysts interpretation of the operator-reactor inter-action that becomes the boundary condition for the analysis. The operator needs to be brought into the calculational arena if the calculations are to be meaningful.
NRC Water Reactor LOCA/ECCS Research.
The supplement being sought by RSE to strengthen their program in areas brought into prominence by TMI-2 will be discussed first.
Some comments on Dr. Murley's presentation follow:
6 i)
Transients and Small LOCA Events:
$13.8x 10 Most of the money will be spent on code development.
I don't believe this is where the effort is needed.
Present codes can predict the early stages of the event reasonably well.
The planned code develop-ment efforts should not need augmentation.
Some of the weaknesses of present approaches were discussed earlier.
6 ii) Enhanced Operator Capability:
$3.8 x 10 The title is one that could describe an area that deserves a great deal 4 n 3
f 3
I
'S-
of attention. RSR, cowever will be devoting a large fraction of this task to instrumentation.
It is my opinion that much of the instrumen-tation is available and only needs to be properly diagnosed for use.
More effort it needed on the psychology of operations.
If this task is not redirected, it will not be needed as present 20/30 tasks cover instrumentation.
/,
6 iii) Plant Response Under Accident Conditions: $5.1 x 10 This area deserves augmentation. At present situations that could lead to core melt have been lef t to our fo' reign partners.
The German program appears to be very well thought out.
Some more attention to it and some of our own efforts would be worthwhile.
0 iv) Post Mortem Examination of TMI-2:
$2.7 x 10 TMI-2 presents us with data for a full scale experiment. Whatever effort is necessary should be made to insure that it is obtained, processes, and made available. The dollars allocated may be too few for this task.
6 v)
Improved Risk Assessment:
$3.1 x 10 Part of this task appears to be a WASH 1400 type study.
It is not clear that the operator and his intereaction with system as effected by his interpretation of procedures and instruments is a part of the 6
study.
If it is not utilization of the $3.1 x 10, it will be much less effective. The second part of the task is human engineering.
Combining item iii) with this task makes much more sense.
6 vi)
Improved Reactor Safety:
$1.7 x 10 Improved reactor safety has always been a goal of safety research.
It is not clear to me why, augmentation is needed.
A number of experimental programs that initiated as separate effects experiments are being modified to include more and more aspects of the whole fI((
11 [
o system. The purpose fer doing this is not clear.
For example, the Japanese 2000 rod reflood test supposedly augments FLECHT by providing additional information on system and multi-dimensional flow effects. Why then is FLECHT continuing.' Programs should be terminated when they cease to be productive.
If FLECHT is to be pursued, then maybe the U.S. participation with the Japanese is not cost. effective.
Semiscale has been very productive in the past.
Its early use was as a separate effects facility. Many of the systems effects experiments on Semiscale,have beci questionalbe. Certain scaling questions resulting from comparisons with LOFT should be answered before Semiscale is modified for further systems effects experiments.
In particular, before modifying Semiscale for studies of small breaks, a study of its further usefulness and cost / effectiveness should be carried out. Further, its one-dimensional characteristics make its utilization for anything other than confirmation of one-dimensional code results questionable. The BDHT experiments conduc-ted by GE and ORNL should be reviewed to see if more data is needed. My own view is that enough data from RLTA type facilities nas been gathered.
With a number of unanwered questions about how representative electrically heated rods are of fuel rods, it seems as if a pause is in order. Unpub-lished EPRI data, presented at the Denver Workshop on Renet Behavior, show that there are differences and that they should be accounted for.
Repeated use of rods, results in non-prototypic oxide layer build up which, along with no gap conductance, leads to different renet behavior. The BWR spray cone contraction experiments at Lynn, M iachusetts are not as comp?ete as the Japanese experiments are thougt e to be. GE has used its licensee's data in the past when it was to their advantage to do so. Two questions come to mind. First, why aren't the Japanese experiments part of our information exchange agreement? Second, why can't the U.S. just buy the 3
,3,
.n 5
145-
Japanese results? These questions should be answered before a great deal of money is spent.
The LOFT program is the largest program supported by RSR. The nuclear tests that have been completed have demonstrated that ECC systems work.
The data resulting from the two tests have given the nuclear industry an opportunity to confirm some a,spqcts cf their ability to predict the course of a LOCA. Unfortunately the LOFT core uses external T/C's to measure clad temperature. The external T/C's have been shown to cause early quenching by Professor Chen of Lehigh, Professor Dhir of ijCLA, Dr. Hewitt of the United Kingdom, and by some German work.
The early quenching modifies the in-core flow behavior to an unknown degree and the clad temperature a great deal.
It is my opinion that the LOFT program amanger should be advised to address the questions immediately.
It was a surprise to hear that a part of the LOFT program was a $300 K study of electrically heated fuel pins.
It seems as if such a
.ady beongs elsewhere. There is a small effort at LOFT to reproduce the UCLA results using inductively heated simulated fuel pins.
Code development is winding down whereas code assessment is building up. The result is a constant level of spending in this area over the next several years.
Most of the effort will be devoted to TRAC althcugh some activity versions of RELAP still persist.
It is not clear why further work on RELAP is not imeciately terminated.
It is my opinion that the small break emphasis could be accomplished within the present budget.
l a
6