ML20147C484
| ML20147C484 | |
| Person / Time | |
|---|---|
| Issue date: | 03/13/1978 |
| From: | Johnston W NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| NUDOCS 7810130023 | |
| Download: ML20147C484 (12) | |
Text
^
8 UNITE D STATES lidqtrs. PDR o
- 8 ) g NUCLEAR REGULATORY COMMISSION
- " ~ "?mn c. 20555
% * *"*[* /,THIS a DOCUMENT CONTAINS POOR QUAUTY PAGES gag 1 3 397g MEMORANDUM FOR: Those on Attached List FROM: W. V. Johnston, Chief Fuel Behavior Research Branch
SUBJECT:
TRANSMITTAL OF MONTHLY REPORT BY NRC RESIDENT ENGINEER AT KERNFORSCHUNGZENTRUM, KARSRUHE, GERMANY Enclosed is the fifth of a series of monthly reports to be provided to NRC/RSR by Mr. Walter Murfin, resident scientist from Sandia Laboratories assigned to KfK in Germany. Mr. Murfin's principal responsibility at KfK is active participation in Projekt Kernschmelzen (coremeltdownresearchproject). The monthly report is his vehicle for describing progress in that program.
Mr. Murfin also maintains cognizance to the extent possible of other safety research at German facilities of interest to NRC/RSR including:
- LWR fuel rod behavior
- experimental techniques for measuring two-phase flow
- jointly sponsored (US, FRG, Japan) experiments on upper plenum thermal hydraulics.
Your comments on the contents of the monthly reports are transmitted to Mr. Murfin, or you may contact him directly, b 0i w Y.
W. V. Johnston, Chief Fuel Behavior Research Branch Division of Reactor Safety Research
Enclosure:
January 1978 Monthly Report cc w/o encl.:
W. Murfin, KfK 13(013893
l .
- MAR 131978 ADDRESSEES FOR LETTER DATED R. DiSalvo NRC/RES K. Campe, NRC/DSE G. Chipman, NRR/DSE '
M. Cunningham, NRC/RES R. Fraley, NRC/ACRS S. Hanauer, NRC/0E0 Y. Y. Hsu, NRC/RSR J. LaFleur, NRC/ DIP S. Levine, NRC/RES D. MacPherson, NRC/RSR A. Marchese, NRC/NRR R. O. Meyer, NRC/ DSS J. Murphy, NRC/RES J. Norberg, NRC/EMSB J. Read, NRC/DSE L. Rib, NRC/RSR L. Rubenstein, NRC/NRR g K.IScattolini,:NRC/ADM/PDR2 F" M. Silberberg, NRC/RSR ,
L. S. Tong, NRC/RSR -
R. W. Wright, NRC/RSR A. Malinauskas, ORNL -
M. Osborne, ORNL .
J. Gieseke, BCL R. Denning, BCL F. Kulacki, Ohio State L. Baker, Jr., ANL R. Henry, ANL D. Dahlgren, Sandia L. Kelman, ANL J. Muir, Sandia D. Powers, Sandia '
L. Buxton, Sandia L. S. Nelson, Sandia D. Walker, Offr.hore lower Systems A. Millunzi, DOE /RDD J. MacDonald, GE, Sunnyvale H. Morewitz , Atomics International I. Catton, UCLA D. Swanson, Aerospace Corporation "
R. Ritzman, SAI P. MacDonald, EG&G J. Zane, EG&G T. Pratt, BNL p
I
t s
- 1. Introduction. ,
7his is the fifth report covering the activities of W. B. Murfin, USHRC representative at Kernforschungzentrum Karlsruhe GmbH (KFK). The report covers activities during the month of January,1978. The principal' duties of the representative are tasks assigned in Projekt Kernschmelzen, information gathering, and monitoring Gene -w 'rch results.
2 Projekt Kernsei ._. zen.
2.1 Core / Concrete Modeling (PNS4331) 7here ara currently four models in use in Germany for heat transfer from a molten pool to concrete. A brief discussion of each of the models follows.
2.1.1 Envirical Model This model (described in SANIf/7-0370) makes use of empirical coefficients which have been derived from melt-simulant /concreta experiments. Concrete ablation is based on the decomposition of calcium hydroxide. At the time this model was formulated, little experimental information was available on the ablation of concrete, and it was presumed that decomposition of the cement matrix would cause the concrete to disintegrate. Concrete constituents rush of liberated gases. However, later experiments at KFK and}.Sandia have- caseous, lio shown that the matrix can be decomposed to a considerable depth without losing its integrity, even thobFh mechanical strength is seriously degraded.
If the empirical coefficients are correctly chosen, this model can still give reasonable results down to temperatures below which tnie ablation does not occur. In effect, the lower temperature boundary for use of this model is about 1600 K. Because this is lower than the freezing temperature of the metallic phase, the model can be used for all conditions in which the metallie phase is molten.
Stationarity of the melt front is not assumed. In principle, this makes the model useful for the initial transient period of a core melt. However, ex-periments (SAND 77-1467) suCgest that this transient period is less than one-half minutes calculations by M. Redmann show that the transient period could be even shorter for the very high heat transfer rates expected in a core l r.elt. Non-stationarity of the melt front velocity may thus be an unnecessary I
complication except for the sinulation of very short duration tests.
This model also assumes an initial spall of 5 mm. thickness. Spalled frag- .
ments up to this thickness were observed in the earliest tests. However, more recent data suggest that spall may not be a universal phenomenon. Spalls of this small thickness could only be significant for simulation of very short duration or low temperature tests, in which erosion is comparable with the assumed spall thickness.
t
-2 The erpirical model considers a therral boundary layer at the pol inter-face (Figure 1.). It is recognised that heat transfer across the boundary layer is variable, probably being quite high at some locations and relative-ly low elsewhere. An 'eqvivalent* boundary layer thickness of 5 mm. is as-suced. he actual boundary layer is probably much thinner than 5 mm. at the locatiens of hie hest heat transfer. Le assumed ' equivalent" thickness is an attempt to estir. ate the average heat transfer rate across the entire in-terface. If the erpirical heat transfer coefficients are correctly estimated, any discrepancy in the boundary layer thickness will be absorbed in the en-pirical heat transfer coefficients.
It is recognised that heat transfer on the sidewalls of the cavity will be lower than at the botten, because of the relatively thick gas film at the sides. he experimentally observed reduction in boat transfer is of the order of 50 percent.
2.1.2 Gas Filn Model.
A gas file r.edel has been developed by Dr. M. Reir. ann of KFK. Bis medcl assures that a continuous file of gas exists between the melt and the een-crete. Classical Tayler instability leads to gas bubbles being released frem the file.1he nedel uses either theoretical or experimental concrete erosien te peratures and enthalpies.
N r.edel is based en experir.ents in which the erosion of dry ico under vater was observed. In these experiments a regular bubble spacing was ob-s e rved : the spacing agreed closely with that predicted by the model. Eowever, experiments with an inhomegeneeus caterial (CO2 snow with frosen xylol in-clusiens) shew a very irregular bubble spacing. The redel implicitly assumes that scre averaEe bubble spacing exists for inhocegeneous r.aterials like concrete s this average spacing is expected to be equivalent to the regular spacing observed - and well predicted - in homogeneous materials. This last assurption has not been quantitatively eenfirmed.
The existence of a gas film has not been unequivocally confirmed for the case of noiten core caterials over concrete. Ecwever, the existence of the file is definitely theoretically predicted for higher temperatures. A pos-sible interpretation of sore data Eathered under high te=perature conditions is that a film is present at least part of the ti=e. At lever temperatures, the file. is predicted to collapse, so that direct centact occurs. The tez-perature at which this happens has not been experimentally confirmed.
The gas film eodel by itself would not predict the very low observed erosion in a test of a pure oxide relt over concrete. Accordingly, this r.edel has been cozhined with a thernal boundary layer medel, in which it is assured that heat transfer from the pool interior to the pool boundary is concen-trated at the stagnation points of the liquid return flev (Figure 2.). ne ve'y thin thernal beundary layer at these points of high heat transfer is averaged over the entire interf ace area. This ' equivalent" boundary layer
3- _
is of the order of 1-2 mm. thick for molten steel over concrete. The actual boundary layer at the points of maximum heat transfer is of course much thinner.
Model experiments show that the film model as described above is only valid for horizontal or slightly incline.1 surfaces. At steeper angles of incli-nation the bubbles will not break loose from the film. There is thus a pro-grossively thicker film going up the side walls of the cavity. Beat trans-fer across this thicker filn is reduced from that at the horizontal lower surface. At some point on the side walls the gas in the film can make the transition from laminar to turbulent flow. Above this point the heat trans-fer is again ir.areased. However, because of the greater film thickness, the heat transfer can still be lower than on the horizontel bottom surface.
This sidewall film model has not yet been applied to melt / concrete inter-action simulation, but gives promise of being able to predict cavity shapes well. At the present time, the film model is used only for the bottom of the pool, and the empirical model is used for the sides.
2.1.3 Discrete Bubble Model.
A discrete bubble model, based on model experiments, is due to Dr. E. Bei-neke of the Technical University of Hannover. In this model, bhbbles are assumed to grow at isolated sites on the concrete surface. 7be bubbles are fed by gas flow through a porous layer of partially decomposed concrete.
Direct liquid-solid contact is assumed between bubbles.
The angular dependence of the heat transfer around the periphery of the melt is given by this model, so that in prir ciple the cavity shape can be predicted.
Experiments with an inhomogeneous mixture of CO2 and frozen xylol under water showed no liquid-colid contact. In this case the water and molten xylol are mutually insoluble, corresponding to a metallic melt over concrete.
Mutually soluble materials, e. g. an oxide melt over concrete, might behave differently. Furthermore, the CO /2 xylol experiments were conducted,at a tenperature at which gas production was vigorous, which would have enhanced the formation of a continuous film.
At some lower limiting temperature a gas film cannot be maintained, and direct contact should be observed. The valid temperature regime for the discreto bubble model has not been experimentally investigated for realis-tie materials. Because the heat transfer rates predicted by each of the theoretical models are decidedly different, it is important to determine their regimes of validity.
2.1.4 New Empirical Model.
In this model, concrete decompesition is bat,ed on melt, rather than on calcium hydroxide decomposition. Theoretical or experimental concrete de- ,
l l ___ _______-__ _ .
t .
' composition temperaturas end enthalpies can be used. As in the film modelqu rather than calcium hydroxide decomposition as an ablation criterion is closer to observed fact. However, it has been observed in melt / concrete simulation experiments that whole pebbles can be dislodged from the concrete without melting. In fact, in experiments at KFK, these pebbles have been thrown right out of the molt by the force of the liberated gases. Melt tem-perature is thus perhaps a slightly too high criterion on the other hand, the calcium hydroxide liberation temperature is decidedly too low.
Obviously, the empirical coefficients using the melt criterion will be dif-ferent from those using calcium hydroxide decomposition as the ablation c riterion.
2.1.5 Discussion.
7he model of 2.1.1 is embodied in the American code INTER. That model, and the gas film model of 2.1.2 are used in the WECHSL code at KFK either model may be chosen at the user's option. The gas film model and the discrete bubble model are used in the EETZ code at KWU, again the choice of model is at the user's option. The EE10N code at the Technical University of Han-never uses the discrete bubble model. The new empirical model of section 2.1.4 has been installed in an experimental version of the WECESL code at KFK, but is not yet ready for production.
An empirical model - by correct choice of the coefficients - can give results very close to those of either of the theoretically based models. However, the empirical model will not exactly match the other models over the entire temperature range, because of the linear dependence of melt velocity on i temperature difference which is assumed in the empirical model. Because l the temperature range over which the models will be used is restricted to l
those expected in a core melt, this poses no real practical problem.
None of the models correctly addresses the decomposition of concrete after solidification of the metallic phase of the melt. The temperature of the metallic phase could remain below its freezing point - but above the decom-positi6n temperature of concrete for long times. It has been assumed in all the models that the liquid decomposition products are rapidly removed from l the interface. 7his removal could be accomplished easily when the pool is molten. However, the removal mechanisms after the pool has frozen are un-clear. It is cenceivable that the outflow of gases from the decomposing con-crate could carry off the liquid decomposition products, but this point has not been investigated.
The composition of the oxide phase is eventually almost the same as that of concrete. Therefore, no further melting of the concrete coul;$ occur after solidification of the oxide. At early times the oxide has a hisher melting point than the concrete. However, the low thermal conductivity of the oxide ensures a rather low interface temperature, so that even at early times the solidification of the oxide phase effectively stops concrete decomposition.
I l .
-5~
2.2 Fission Product Release _.
Uork on the ir: proved SACHA facility has been delayed due to slow deli-veries of equipment, and a final test has not yet been made. Work is pro-gressing on enlarging the tube and filter diameters. An entirely new transport system will be tested with masses in the order of 100-500 grams to determine if the capacity will be adequate for samples in the Kg range.
Work has also been carried out on computer programs so that analysis of the results of future experiments can be carried out in-house. Detector efficiencies are being recalibrated for the larger filters and changed geometries.
2.3 Steam Explosion Pesearch.
Dr. H. Kottowski has reported on steam explosion research at Ispra.
Table I.
Summary of Results.
Type Mat'Is Sys tem Explosion? Remarks Expe r. Pressure Channel Steel 1 bar No No fragmentation.
HO 2
Reproducible.
Channel "
5 bar Yes Peak press. 80 bars.
Reproducible.
Channel
- 25 bar Yes Peak press. 175 bars.
t Reproducible, Exceptionally great fragmentation.
Channel UO2 -gran i bar No Not known'if reproducible.
Eo g
l Tank Steel i bar No Helt:H2O - 1:1000, 1:30, ~
l B0 2 1:8.
' Slight fragmentation.
~
- Peak press. 2 bar.
Reproducible, Tank UO2 -gran i bar No Melt:H2O - 1:500 i
HO 2
Peak press. 1.6 bar.
l Reproducible.
l Tank UO2 1 bar No Melt:Hp 0 - 1:1000.
HO 2
Facility nearly ready for tests at 1:5, 1:2
R. Benz and H. Unger have reported on steam explosion research at IKE, Stuttgart.
- 1. No failure of the stressed geometry is to be expected during the quasi-static pressure phase of the steam explosion.
2 The integrity of the RFV head could not be assured for the case of some tor.s of Corium reacting coherently with the available water.
- 3. Containment failure by shock waves from the steam explosion can probably be ruled out. The estimated strains are less than the dynamic failure limits, although exceeding static failure limits for some milliseconds.
These results have the character of estimates. Strains due tolprojectiles hurled by the steam explosion were not investigated.
2.4 Kernschmelzen Sachverstaendigenkreis (SE).
The SK met on February 1,1978, at GRS, Cologne. An item of interest for Anerican core melt reseamh activities was the presentation of the proposal "large Core Melt Simulation" by Dr. J. P. Hosemann of KFK (FNS). The SK did not take action on the presentation at this meeting. Presentations on steam explosions are surnarized in section 2.3 of this report.
- 3. Fuel Rod Behavior Reseamh a_t KFK.
No report in this period. -
4 Two-Phase Flow Measurer.ents g KFK.
Fersonnel from E. G. and G. have been at KFK for about five months for two-phase flow testing. E. G. and G. ard KFK have performed joint tests I
of densitor.eters. Thereaf ter, tests were conducted on the instrumentation l used in the American 10FT and Semiscale tests. At present the latest model DTT (Drag-disk Turbine Thermocouple) is being calibrated. After completion of the D7T tests, further support will be given to KFK densitometer tests.
I The LOFT DTT worked well except at very high steam velocities. (These very high steam velocities are not of interest for the LOFT tests). There has been sore disagreement in the data between the test loop reference data, the LOFT instrunentation, and radiotracer data. This disagreement does not necessarily indicate a fault in the instWmentatinn, but is rather a result -
of imperfect understanding of some aspects of two-phase flow.
l i
7 De data have not yet been reduced, and because of their preliminary nature ,
will not be reported at this time. Seme of the data will probably be in- l cluded in a future report. Se information will also be reported in detail at the next FPSIM.
- 5. Upper Plenum Experiments.
No report in this period.
- 6. Upconing Events of, Interest.
The agenda and call for papers for the forthcoming ENS /ANS meeting in Brussels (October 16-19,1978) is attached.
Readers are reminded of the conference "Beaktortagung 1978" to be held at Hannover, West Germany, on 4-7 April,1978. Attendees should be reasonably fluent in Geman.
e e
,~ * .
. 8 ,,
Terperature profile ,
e a s ,Aetual boundary layer s
l ._ L s e
[ \__ _ p, Equivalent coundary laye'r Gas Film "elt/ Film Interface
/
Solid and Liquid Deeet pfsitson Produets Ca(002Decorposition T1GUFE 1.
Liquid Return Flow Liquid Carried up by Babbles 1
/i O y( ts
/ A.1 I [ Ak -
1 3 A , d
tv.oy.tR1ry-ylp - -) .k- . yI f .
l
\ Gas Filn [
Eeat tYan fe frez Ibol to film is cor.centrated at these points FIGURE 2 J
0
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ . ______ ____.___.._._m.. _ _ _ _ . . . _ _ _ _ . _ _ _ _ . _ _ _ . . . . _ . . _ . _ _ _ _ _ _ _ . . - _ _ _ . _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ - _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ __
, ,4;
'Y j')fbyY.D-YIk$f[dNEN( b E DU D NDE N YINI
- h
' C A LL for PA P Eris g hf i.'
j C< -
1 n the basis of the summaries, the Technical Program Comm'ttee will }<
!(! '
o INTETIN AllON AL MEE TINC oit NUCI EAR POWER RE ACTOrt S AB ETY c ccept papers for oral presentation in EnClish (20 minutes) and/or for public- . k p tion la transactions. The Technical Program Committee has the rlglet i TOPIC AL ME E TIN G - y <
li eject the papers which do not meet the criteria but, in such cases, w!Il im- $l October 16 -19 1978 - Deussels neignmen fement en appreal procedure. Notification should be espected for May 1.1978. ' Q.
e case of oral presentation. It is recommended to use slides (slaes 35 mm
, hmtEM&MiRcr#MNfYM**tW*MftMM1MBW r 2 s 2 in.) or full siac transparencies. Wording on slides 'and transparen- e iles must be kept to a minimum . using very large eine bold characters to l nsure legibility.
1 phe papero retained by the Technical Program Committee for presentation l nd/or publication must be presente.d in their final state on./or before October r
. 1978. These final pape rs - directly derived f rom the survimartes - cannot
, steed a total nt 12 pages including a maximum of 3 pages for figures and/or . .- The Eucopean Nuclear Society and the American Nuclear Society are jointly .I shlee. They must be in camera-reedy format (typed lines not exceeding 17 cm eponsoring an international meeting on the safety of nuclear power reactor's a length, double spaced, well contrasted). .
haviry attained commercial status (LWR. CANDU-PilW. ACR. IITGR. FDit).
1 I
l
- t le inter.ded to publish transactions within 6 months after the meeting. The 4 day meeting will be held October 16 through 19. 1974 In Brussele, and followed immediately by technical tours. It will be organized by the ANS Belgian Section. ,
i Contributed papers are called for and papers are invited in the subject-areae given in the following non-enhauettve llett
%v .
- ~
e , , * * *g* %
,g.
. 7 e
/
A /
m . __ .
. ,. 3
- Containment concepts declan, perinemince and special me,iti. asion# .
I t/ ttlek mescoement and enunicr-mearnece - Sitine c riteria and evaluation ,
- Itadwaste procenging
- Asses = ment of riske waste storage and disposal
~
. Safety related dreign bases and criteria *
- racility decommlooloning -
- Regulatory guidance and requiremente, licenelng
- llellability engineering and analyele safety emperience
- Nuclear plant safeguards and accountability, succese in enfcguarding -
A and special nuclear materiale A A
- Off-elte safeguards and accountability
- Current key safety leaves
- Value In pact and cnnt-benefit analysee Summaries of the papers, in Enstleh. not exceeding 1000 worde, are to
- Education and public acceptance of cuclear power p! ante be submitted not later than Marcli 85.1978. to:
2/ Accident analyele and substantiation M e, J. p, ,an th e,oet l ~~
Techtitcal Program Chairman c/o D e l g o ti u c l e a l r e .
1 Accident analysis and confirmatory activities .
rue du Champ-de-Mars. 25 D - 6050 Brussels - Delgium {
- Operational safety emperience
- Structurat dynamice
. ; The summaries must present facts that are new and significant in their conten0 I ' 3/ Pheno.rienology l...
. ii- .' - '
- or their Interpretation. . s.inth the work performed and the reeutte act leved i
must Se presented in a tiarrative of $00 worita minimum. The narrative .
- liest transfer ' '[ , -
typed doubly spaced must be placed between an introductory stateinent Indic-
- Chemical processes ,,
1- = ating the purpose of the work and a closing statement summariting the algnl.
A Metallurgical processes - 'p ficant new results or interpretatten. The latter le followed by a manimum of
~ '
- Mechanical proccesee 5 'i , , ,
3 pages of fisurce and/or tables, and by a list of proper references to all 8
closely related information that has been pubtlehed. Thee list should be kept
-f 7
. flydraullce ,
. f .* Nebtrontce ' C?-
.to a minimum. .
, ' h. . .
j '
" "I ' umm8 tIH . t 8mt Pflm F for the paper redew.
4/ Radio activlty generation and contro '
and on a separete page. typed singly spaced, a 10 line abstract must be pre -
- "* * "" '"*
- E*I' "" "**
- Radlological source term determination
- Radio-activity transport mechantamm anel control addrene of the principal author enuet he inificated me ' . le to this author that further correspondence will be directed.
Gk V - W , I = e
'.'}}