ML19321A013
| ML19321A013 | |
| Person / Time | |
|---|---|
| Site: | Indian Point, Zion  |
| Issue date: | 05/10/1980 |
| From: | Catton I CALIFORNIA, UNIV. OF, LOS ANGELES, CA |
| To: | Quittschreiber Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-CT-1249, NUDOCS 8007220308 | |
| Download: ML19321A013 (6) | |
Text
..
SCHOOL OF ENGINEERING LOS ANGELES: AND APPLIED SCIENCE May 10,1980 N'
4 CJ - /0$
TO:
G. Quittschreiber
, l}
FROM:
Ivan Catton
,,, g
,, O CO l
j Technology Meeting with Zion and Indian Point Utilities on Core f
i Melt Accident Sequences for Zion / Indian Point, 7-8-May.1980.
SUBJECT:
}
. Ju..' ::. Bi5.%
^ '#
' 00 SM y
SUMMARY
The two day meeting was to yield an exchange of information between the The utilities utilitids and the NRC staf f without addressing licensing issues.
demonstrated an apparent lack of initiative by not prese The position taken by own work.
consultants to resolve issues and questions for them.
i l the utilities is that their plants, Z/IP, are better than the 9
This is to be contrasted with the view.of the staff that one mu look at the back end of the various accident sequences to make significant accidents.
reductions in risk because of the large number of contributors, e.g. one mus The IREP program has not focused on Z/IP as yet look to mitigation devices.
and in this respect seems to bc unresponsive to NRR needs.
Themost important technical issue is the MARCH / CORRAL predicted cont The calculated rate of rise of pressure leads to un-reasonable design requirements being placed on mitigation.ievices and itsThe ment pressure spike.
magnitude without them to cor.tainment failure.
At present pressure spike prediction deserves a great dr.1 more attenti Emergency power to the spray sibility of sprays quenching a hydrogen burn.
system, appropriate spray location and flow characteristics may be suffici to eliminate the pressure spike and regulting requirements f d
was made in reaching a common ground regarding knowledge of assumptions ma e containment.
They felt that the steam spike is still a major in tne various analyses.Hence the utilities don't believe that it is clear whether one The heat source for the steam is molten core uncertainty.
wants a dry or wet containment.
material and quantifying it was noted to require a more realistic core slump
- Further, and melting analysis (one shouldn't have to deal with the whole core).
more attention to the dynamics of the reactor cavity itself is needed.
The staff noted that they needed answers in months--not years and that Answers to questions about debris they may have to redirect ARSR programs.
bed coolability, steam pressure spikes, steam explosion induced steam genera-The tion tube failure and hydrogen burning were considered to be primary.
debris bed coolability question leads directly to answering questio the need for a core catcher.
they would arrive at a design basis for mitigation features.
j75ce2.1 X -I. C A7ik 8007220309
s GENERAL DISCUSSION Dr. DiSalvo,NRC Selection of Representative and Envelopine Sequences.
d to. Their pro-staff, stated that the WASH-1400 philosophy would be adhere cedere would be as follows-One order of magnitude Decide what risk reduction should be.aquired.
has been tentatively selected.
1.
Select accident sequences.
2.
Calculate the accident environment.
Determine the mitigation system functional req'uirements.
3 Design the mitigation system to meet the requirements.
4.
5.
Calculate the net effect on risk.
it determines 6.
The accident sequence was stated to be important because Its not to the system.
the physical phenomena that will present challengesIt seems to me that many l
clear that this is the case. accident sequence leading to risk are similar and i
s in risk because power and water are available and where.at the back end of a of the large number of contributors.
initiators Check valve failure and smal'1 breaks were seen to i
(from steam), hydrogen of a sequence.
water, power conversion and containment spray.that lead to burn and steam explosions.*
failure are typically small break or some transient as an initiator a) human error compounding the problem b) hydrogen burn c) containment safety features unavailable d) containment failure above ground.
He said that e)
The utility view in this area was given by Dr. D. Walker.l PWR and tha i
the design goal was to do better than the WASH-1400 typ ca Therefore, no further 0
has been demonstrated that Z/IP are already better.He further'noted Plant specific features, effort is justified.
PWR at the Z/Ip sites is not the proper approach.
d Apparently a number such as those of Z/IP, must be incorporated into the stu y.
d te the high popu-of special features were incorporated into Z/IP to accommo aCeorge K s might dominate lation density during the licensing stage. expressed concer sed just because they are ir.ce estingA further consideration, as expre happens).
operator actions and possible blunders.
'h t
It is possible that sprays could eliminate the first two as a med in WASH-1400.
and steam explosions are not considered as dangerous as assu
- NOTE:
.~
3 Evolution to Meltdown.
Dr. J. Rivard of SANDIA gave his view of how an accident evolves to a meltdown and what the uncertainties in our understanding The process is a complicated, not well understood heat transfer problem.
are.
He noted that the areas of uncertainty were Early in the sequence Late in the sequence clad failure crusting fission product distribution steam explosion melt-water interacti'ns early melting internal structure failure debris coolability modes vessel failure mode and timing
~
blockage effects melt motion It is not The picture presented was one of a hundred year research program.
clear how well one has to tie down the uncertainties before coming to con-It is clear that NRR cannot wait for RSR to address the above un-clusions.
certainties before making decisions regarding the need for mitigation devices.
It is my opinion that we will not be able to improve our knowledge sufficiently One should, however, be able to identify in the near future to do any goo'd.
the sensitivity of the uncertainties to risk and bound them enabling the decision making process to procede.
MARCH Code. The MARCH code developed at BCL plays an important part in With it, the infamous pressure spike is pre-studies of class 9 accidents.
dicted.
Dr. P. Cybulkis gave an overview of tbe nodeling, assumptions and limi-tations of the MARCH code. The " ode is made up of components each dealing with part of the sequence. They are Mel tdown-BOIL code Assumes fuel melts, moves, refreezes according to one of the following model s a) melt runs down, refreezes then melts again to maximize downward pene-tration. Melt does not dribble into lower plenum until core support plate fails.
b) high upward heat transfer to maximize upward penetration (not sure of details of the model).
c) as soon as a node melts it falls into the lower plenum.
The models contain a great deal of uncertainty and which is most likely is not known.
Dr. Cybulkis believes, as do I, that model 'a' is most likely.
~
Vessel Failure-HEAD code Evaporates the water in the lower plenum, heats up the head and A number of calculations lead one to believe that water w head when the core support plate fails with some degree of certainty.
it.
Entry of Molten Debris into the Reactor Cavity-HOTDROP code When the vessel fails, the fuel pool is assumed to fragment into some d heat assumed radius particles and to transfer energy to water with an assumeThe g
This gives the steam spike.
ilure.
of accumulator dumping or it can be in the cavity.before vessel fa transfer coefficient.
Concrete-Corium Interaction-INTER code -
A code written at SANDIA to account for the.omplex chemical reactio An assumed and heat transfer taking place when hot corium decomposes concrete.
The amount of concrete heat transfer coefficient couples the two materials.The code is only valid during penetration and gas production are predicted.the very early stag It is not clear what is done past the first several hours.
Containment Thennal Hydraulics-MACE code The MACE code is a quasi-steady multi-compartment trede terms in the momentum equation.
as well as gases resulting from decomposed concrete.
Some uncertainty studies have been carried out an d
key parameters are fission product plate out, fission product source te underway.
Little direct dependence on the core melt models Uncertainty in containment f ailure modes.
This conclusion bears on Dr. Rivard's concerns.
i containment pressure depends on how the molten pool enters the r was found.
and the timing of the accumulators.
In coming to conclusions based or not did not appear to be too important.
d ling on these uncertainty studies, one must not loose sight of the primitive mo that is the backbone of the MARCH package.
The pressure spike calculated to occur immediately following failu (1/5) the vessel results from release of primary system high pressure steam Another and interaction of the molten materials with accumulator w sequence might have resulted in lower primary system pressu actuation before vessel failure. The slow pressure rise allows mitigation measures pressure spike would not occur.An ADS would allow operater action to accomplish th to be taken.
Dr. Lipinski of SANDIA presented the results He assumed that the entire In-Vessel Debris Bed Coolinq.
of his calculations showing that dryout could occur.
The fragmentation core melted and as a result was faced with very deep beds.
ll under-process leading to particle sizes that must be dealt with i
)
large.
a consultant to the utilities, as to the expected particle sizes were stood.
l
Henry refers to French work resulting i.n sizes on the order of centimeters whereas Lipinski refers to SANDIA work resulting in sizes on the order of fractions of a millimeter. The latter result in dryout and remelting.
Mechanism for Introducing Water into the Vessel for Various Sequences.
It was noted that HPI, LPI and a number of non-ECCS sources such as seal water Further all ESF operate were available for introducing water into the vessel.
of f diesels (with few exceptions). The question will become one of operator action.
Progression of the Radiological Source Term.
Radiological materials that result in risk are in the water, vapor phase and tore debris and they can plate out on various surfaces. Ptst are found in the water.
Fission products had This to penetrate subcooled water in the' pressurizer at T?iI-2 to get out.
says that plate out is not too important and that water is the important trans-port path.
The most MARCH Prediction of Ex-V'essel Seouence--The Pressure Spike.
important aspect of tne results is the pressure spike.
The initial rapid rise at vessel failure is followed by a pressure decrease due to condensation.
This is followed by a slow rise as the structure heats up and becomes less The second peak is as high for some sequences as effective as a heat sink.
Without some means of cooling or venting, containment failure the first.
No credit was given for containment cooling or sprays.
The pos-will occur.
Zion sibie quenching of a hydrogen ' burn by sprays has not been considered.
has diesel power to the sprays whereas Indian Point does not.
Heat transfer from the particles (and their characteristics) to the water are key factors in detennining the pressure spike. The dynamics of the pro-For example how can enough steam be generated in cess are also important.
such a short period of time to pressurize the entire containment without blowing The small particles necessary to get such rapid heat trans-the water away.
fer result from a steam explosion that will probably blow the water out of the cavity. A rapid pressure rise does not allow coolers to be effective 3
6 ft / min (sprays could attenuate it). The present calculations require 10 x 10 3 vent rate whereas a five minute pressure rise only requires a 350000 ft / min vent rate. This area deserves a great deal more attention.
It does not appear to be a part of the RSR class 9 program.
Water in the Reactor Cavity. At present water that gets into the cavity is by accident. A curb between the sump and the cavity in one of the plants will limit the water to that from the sprays.
It was estimated by the utilities that about 25,000 gallons will be in the cavity before vessel failure. More would be in the cavity if the curbs w re drilled.
In that Indian Point loses the sprays on loss of power, it is not clear where the 25,000 gallons come Water in the cavity helps with cooling and decreases the potential f rom.
for penetration of the base mat but feeds the pressure spike. On the basis of the MARCH calculations, the utility position is one of reluctance to deliberately flood the cavity.
Ex-Vessel Debris Bed Coolability.
The sat!DIA view is that even with a water supply one cannot guarantee debris bed coolability when the entire core melts and is in a pile underneath the vessel.
It is not clear that the total core will melt and if it does it will probably spread. A need to guarantee a water supply to the cavity seems to be appropriate.
4
~.
6
~
Lipinski (SANDIA) has extended 'his w0rk to deep beds where the simple Darcy law approach is not valid.
He has not, however, fully accounted for particle size distribution in that he weights fines too heavily.
This aspect with the piling up of the debris may be too conservative. As noted by Henry, The small particles are the result of a steam explosion that spreads the bed.
deep bed and fines are therefore too conservative.
I agree with this view.
It is not clear, however, that Henry's conclusion (the utility position) that dryout will not occur and that a retention device is not needed is correct.
SANDIA fragmentation studies will be very important in answering some of these questions.
Steam Exolosions. A quote of a statement by O'iSalvo seems to sum up the steam explosion area- "Best available.information indicates that failure due to a steam explosion induced missile is much less probable than previously assumed to be--if not 1,mpossible". Steam explosion sequences are probably not as important as they once were and it is time to re-direct our resources.
There are still questions about the effect of molten material-water interaction effects on rapid heat transfer and fission product purging.
e f
.