ML20197B394

Memo Disclosing Previous Involvement W/Applicant Re Facility During Sept 1974.Author Head of Physical Sciences Dept of Rand Corp.Draft Research Proposal Re Facility Prepared by Physical Sciences Dept in Sept 1974 Encl

ML20197B394
Person / Time
Site:	Atlantic Nuclear Power Plant
Issue date:	10/27/1978
From:	Gilinsky V NRC COMMISSION (OCM)
To:
References
NUDOCS 7811060260
Download: ML20197B394 (10)

Text

n. - _ _ _

i. ,,J ~ d  %

37 . ~. m QT?'T3iTG DD(AGGtR GG41 ll?$ ,

~

UNITED STATES OF AMERICA Eli4 0CT. T1978> #

. NUCLEAR' REGULATORY COMMISSION l "\p . ,.,. .<. % ,-

.,a 3 h '

In the Matter of )

)

0FFSH0RE POWER SYSTEMS ) DocketNo.STN50-437h (FloatingNuclearPowerPlants)

)

MEMORANDUM TO COUNSEL FOR THE PARTIES The purpose of this memorandum is to disclose to the parties to this proceeding my previous involvement with the Offshore Power Systems i application during the period I was head of the Physical Sciences Depart-ment of the Rand Corporation. During September 1974 I discussed with AEC's Division of Reactor Safety Research Rand's ideas for possible research on the consequences of a core meltdown of an offshore nuclear power plant.

I subsequently submitted to the AEC a six-page draft research proposal on that subject which was prepared by four members of th staff of the Phy-

• / 1 sical Sciences Department under my supervision.~ The AEC did not award Rand a research contract, and I have had no subsequent exposure to this subject either at Rand or since I joined the Commission. After reviewing the matter, I have concluded that I have not had a personal and substan-tial involvement in evaluation of the question of core-melt accidents at floating nuclear plants and that I will be able to consider the issue pre- ]

sented by review of ALAB-500 in an impartial manner unaffected by the limited

-*/

A copy of that proposal and the letter transmitting it to the AEC is attached.

THIS DOCUMENT CONTAINS 7 8 H C16 Osty 6 P0OR QUAUTY PAGES -?

2 involvement discussed above. In these circumstances, I propose to participate in the pending review proceeding. I will, b ver, consider any objections to my participation received by Friday, November 3, 1978, and I will give careful consideration to any objection before deter-1 l

mining whether I should abstain from participating.

,f Commissioner 73 he 7 ) ,f Date 1

y/% t,,f.;l i ---

,,,,j 1 .

.. - I f%. o1 **

o f 4.3,0 *

/ ( .

r f.*.G - .u., .. ,,.m

. .n c,,]

M NI\.w avt.4,CA wi w .

=

~rn h l'

' +

VICTO.'t Cit 1NS)'Y Head 17 ScPtember 1974 bywat Scierm cerwtment .

Dr. Jerry !! arbour ~*" ' ' '

D!. vision of Reactor Safety Research p,,

AIC Ucadquarters

• Germantoun, Maryland 207o7 ,

.,. U ,/*'

Dear Dr. Harbour:

Last week, Dr. Saul Levine telephoned me and requested our '

ideas 'for research on the consequences of a core =eledevri .

of an offshore nucicar power picnc. He requested that vc g.;  ;

cend our preliminary draf t material to you for co==Lants, e.

You vill note char. we nave organized ::! c ---- # a' into three

.' .1 i

. parts: n ... s.: a.

Oo}

L', a' 4

C ./ I .". ,

- . 1. The reaction of the molten core with the ocean . . .

j ..

environment.

- y, ..

/ '1 1 The dispersion of the products of this inter- ,

~

p,h ,j 2.

action in the surrounding ocean. .

yoff , c,.. - . . . . /-

. ?g,, y The dispersion of these products into the . f r ~

. 3. '

atmosphere. *,,jAT3 ['%' w ' " ,'"..

• W M.)

>A 'r s

The first of these is a needssary prelude to the second and

' '* * '-  : -c w ,

third; however, ec bo11cvc the methodolo~y of analysis and modci develope.cnc of the dispersion in ocean and atr.osphere

• can be* pursued simultaneously. Thus as the results of the first part of the study are obtained, they could be inserted '

into the ocean and atmospheric dispersion =odels. ,

I a n also enclosing, bior.raphic'al catorial on staff membersDr. , .,

and consultanta who could be involved in such rescarch.

Critton, Dr. Ca: Icy, and ponnibly Dr. Lecndertse plan to be in Washington in about two vecks and cou3d arrange to visic d' 6

• .e W n,M# - A L h. . !- J L 24. _C J

' )

ifl- N- .'<.fr wb .u c -.e .d> It a A c'A .

G l~ d .

~, J - . ui -. hlA.5lL e A L.!,,., L  ? n- - -- /

U I[ NA.N O (.Oi WalION, Irtil .stalN S l'R1 t I. Saf 414 pto.NicA. C u i *Of;NIA 'UM'. l'i nt N . f.11 Wl "'

- - - - .1.

,)p ' (. ,

j s .,

• ~'.**- ' s, ' ' , 9 .. l'

.. ) . .s

\~.. .

! l Dr. Jerry }! arbour ~2- 17 Septeraber 1974

- you on .the siternoon of October 1 or during the day ofThey Octo-bar 2 for further discucsions of ~ these study areas.

vill telephone you next week.

.Sinecrcly,

. . +

• ~

VG:pd *

Enclosure:

I Draf t Troposal, "Some Safety ' ,;;

Aspects of Offshore Nuclear

, Powerplants" . .

i

- a. *.  ;

Dr. Saul Levinc . .

ces . .

M. C. Davie l

bec: '

C. Gazley .

C, Gritton . .

J. S. King, Jr. , ,

l L. R. Koeni.g .

J. J. Leendertse .

. J, McCully ,- ..

A. S..Mengel '; .

' * \

G. Shubert .

G. K. Tanham (W[0).

I

' *  ! r .

• r ,

" . s. ,,. ..

,.,,.., ,i

+ 4  ;

. . % *. ? .

, e .

' 4 4 . < < . , , }

, , l l

, l l

9 ,

1

. . l 1

. . j l

e

) *

)

1

(

SOME SAFETY ASPECTS OF OFFSIIORE ttUCLEAR PO'!ERdLAftTS

\

I E. C. Gritton, C. Gazley, Jr. , L. R. Koenig, and J. J. Leendertsef i The Rand Corporation O

.. GN4.9 # '

Santa Monica , California '

September 17, 1974 l

It is the objective of the research pr'oposed here to study the effect sn the environment of an accident which involves a core meltdown of an 1

The probability of a meltdown pffshorebarge-mountednuclearpowerplant.

bn 'oe made extremely low through the use of adequate engineered safe-huards. Nevertheless, it is essential in licensing offshore plants to -

essess the consequence of a core meltdown and subsequent release to the environment. Key questions are: how do consequences of core meltdown accidents in offshore plants compare with those located on shore? How can ltheseplantsbemadesafer? For example, should core catchers be required lon all offshore plants? Should the breakwater be ecmpletely enclosed and impecvicus to the tre.r.cfer of wwater and dissolved or suspended materials? j Clearly, major design and cost issues are involved. A broad accident study could provide a sound technical basis for providing quantitative answers to these questions.

The analysis of a core meltdcwn in an offshore nuclear plant has .

three principal parts:

1. An analysis of the escape of the molten core and its interaction with the ocean environment.

l 2. An analysis of the transport ana dispersion of the products of

'the ocean / molten core interaction in the surrounding ocean.

3. An analysis of the transport and di:persion of the products of the ocean / molten core interaction in the atmosphere. l These are examined in detail in the following sections. Briefly ,

the first part involves heat transfer analyses of molten or high-temperature The objective of those i core materials and their interaction with the ocean.

analyses would bo to try to estimate the kinds of products that would be found if molten or high-temperature core material come in contact with the ocean. The key question is: would the hot core when it interacts with the ocean break up into finely dispersed particles which could be suspended i

I

O- } ,

2 .

o. . .,

and transported tcward shore by the ocean currents, or would it remain in fairly large pieces? .

The second part involves the development of analytical techniques to predict the transport and dispersion of solutes and granular material which can be moved as a sediment. It appears that a sedimentation model could be developed in a general way such that the sour'ce term does not have to be kncwn initially. Work on this part would therefore proceed in parallcl to part one.

Then as the results of the first part of the study are ob-tained (physical characteristics of reaction products), they could be insert-ed into the transport and sedimentation model to estimate their dispersion.

The third part requires the modification of current plume or cloud transport models to make them suitable for this problem. As th'e hot core reacts with the ocean water, it would probably form a steam cloud which could entrain quantities of gaseous fission products. If this cloud were '

to move toward shore, the effective fallout from the cloud could be quite serious.

In parts two and three we would expect to utilize water and atmos-pheric transport models which have been developed at Rand.

The three study areas proposed above combine to form a coordinated program of research.

We believe that it is important initially to place primary emphasis on the first area--that of an analysis of the initial interaction of a molten core with the ocean -since this would provide quantitative input data for the other two problem areas. In the following, we describe each of the study areas in greater detail.

ANALYSIS OF INTERACTION OF A MOLTEN CORE WITH THE OCEAN ENVIRONMENT Physical and chemical phenomena occurring when a molten reactor core falls into sea water are not yet well-defined or understood. While it can be easily visualized as a violent " vapor explosion" in which large amounts of sea water are evaporated rapidly and the molten core is solidi-fled, it is not clear how rapidly this would occur, how the core materials i

would be distributed between vapor and liquid, and what the size distribu-tion of solid particlos would be. The distribution of nuclear material, in phase, fom, and size, is a necessary ingredient for subsequent evalu-ation of the transport and dispersion of that material in the ocean and i

O

v . s.

- J s ,

-3 ,

in the atmosphere. A knowledge of the physics of the. interaction is also l

necessary to answer other practical questions: Is a " core-catcher" neces-sary?. If not, what is the safest way to release the molten material--one large lump quickly, one slowly-flowing stream, many small streams, etc.?

In recent years, the " vapor explosion" has begun to attract some research attention due to its application not only to reactor safety '

problems but also to safety problems in the handling of molten metals and in the shipping.of large quantitics of liquified natural gas. A.

limited number of experiments have been made in which small quantities of molten materials (bismuth, lead, tin, uranium oxide) are dropped into cooler liquids (water, liquid nitrogen, and sodium). A number of mechan-isms (models) have been proposed for the interactions but none is fully satisfactory. The primary difficulty appears to be in the description of how the contact area increases so rapidly--i.e. , how the mo'1 ten mate-l rial breaks up. Initially, a vapor film 'is formed at the molten surface and the heat-transfer rate is characteristic of boiling heat transfer.

A very rap 1c increase in heat-transfer surface (by a fact.or of au'uut l 1000) occurs suddenly and the explosion takes place. There are indica-l tions that a vapor film instabili,ty grews rapidly due to some external disturbance, the film collapses, a jet of the cool liquid penetrates the molten mass, this jet vapori2es explosively and fractures the molten material. This process is then repeated for each fragement until the daughter fragments become small enough to freeze.

We believe that an analytical / numerical study of the physics of this process would enablo a clarification of the mechanisms, would provide a framework'in which to evaluate the available laboratory data, would suggest new experiments, and would provide a better basis to scale such data to the problem of interest. The latter is more complex than most of the experiments because of the higher temperature, the complex ccmposition of sca water, and the contaminants in the melt. The higher temperatures lead to appreciable heat transfer by radiation across the vapor film and to dissociation of water vapor in the film--both enhancing the rate of heat transfer. It appocrs to be feasible to model the stable vapor film (with its heat transfor by convection, diffusion, and radiation),

the instability characteristics of the film, its collapso, the subsequent penotration of the sea water into the molt, the vaporization of this water jet, end the consequent fracture of the molten material.

t ,i h _a_ .

's

e. , .

We propose the following specific research: -

1. Evaluation of existing analyses of core meltdown to determino melt characteristics: temperature, composition, etc.

2. Analysis of existing experimental data on vapor-explosion character'stics .and determination of parameters which dominate the phencmenon.

a. . wuioion eI enaijtit-mtmer4c medcb of tho-:ac1 ton cc-e 9ter . . . . - - . i action with sea water; this model will include: rad'iant heat transfer and disscciation in the vapor film, vapor-film insta.

• bility and collapse, and penetration of water jet into the melt with subsequent vaporization and melt fracture. '

4. Application of the model results to: preliminary estimate of . .

dispersion fractions of radioactive material into vapor and liquid phases and corresponding particle sizes, the design of necessary experiments to improve and verify the model, and the ,

design of core catcher / distributors which minimize the release of radioactive materials.

ANALYSIS OF THE TRANSPORT AND DISPERS_ ION OF THE PRODUCTS OF THE OCEAN / MOLTEN CORE INTERACTION IN THE SURROUNDING OCEAN The core meltdown of an offshore nuclear power plent may distribute ,

--through many mechanisms--radioactive particles of a wide range of sizes and densities in the waters around the nuclear plant. The fate and

. distribution of these particles in time is the principal objective of this part of the study.

It can be imagined that fine light particles will be moved more or less as a dissolved substance by the tides and currents in the vicinity -

of the installation and only drop out very gradually. Heavier material of small particle size will likely move over the bottem of the ocean more or less in the same manner as sand is moved in the offshore area.

Very heavy material of large size will likely have only very limited movement and it can be imagined that they will become buried in the local bottan sodinent.

t

{

l. - , ..

- ~

. . .s. .

In this proposed study, we would first classify anticipated material which would result from such an accident as to sizcs and density distribu-tion and radioactive properties. Subsequently, it would be necessary to develop the mathematical models of transport for each class. This devel-opment would be particularly complicated for materials which are similar to sand and fine gravel as their movements can be a movement over the sea bed (bed load), movements in the water column, or a combination of both. While moving, the radioactive material becomes mixed witir natural sediments.

Once this movement mechanism, in part influenced by wave motion, is described it would be p'ossible to suggest practical methods of solution, which will incorporate already-developed Rand computer models.

ANALYSIS OF THE TRANSPORT AND DISPERSI_0N OF THE PRODUCTS OF THE OCEAN / MOLTEN CORE INTERACTION IN THE ATMOSPHERE _

If a hot molten core enters the ocean, a cloud of water, water vapor, tea salt, and reactor debris will be generetad and M11 hubMa: narhaps

~

explosively, upward to the ocean's surface. It will be hot upon reaching the surface and a warm buoyant plume of this debris will rise in the atmosphere, mix with air, and spread downward. As the plume cools, condensation will occur and material may precipitate from the cloud and be distributed over a rather large area. Solid debris from the reactor my also be carried upward and fall back to the surface at varying dis-tances from the source; and noncondensed products will be widely dispersed, perhaps never returning to the grnund," perhaps returning after being absorbed on particulates or dissolved in droplets.

The usual way in which the downwind nature of the plume would .be p edicted is by means of equations in whch the plume dimensiens and component concentrations are specified; a priori, based on the anal nature of the source (point, area,line), the wind speed and atmospheric stability. This is the classical Sutton approach.

However, in assossing high-censequence low-probability events little experimental data are typically available to guide decisions. It is therefore essential to depend on sophisticated mathematical models and simulations to a grr.tcr degree than is usually the caso. It is important

y)

, in our problum to achieve a rigorotis treatment of the ecmpicx processes including chemical and physical changes that may occur within the plume as well as interactions of the plume and its contents with the environ-ment.

Rand has developed and refined a cloud-scale (10 'ei domain) model of atmospheric cumulus convection that has been used for a number of related -i studies. It contains a basic' core that computes atmospheric dynamics onto which may be attached modular units which specify phenomena of interest to a particular problem. Its building-block approach permite I adaptation to modeling the dispersion of a pollutant cloud from a nuclear

• accident.

l

- r

?

s t l

i 4

P 4

e e

b a

4 i

- - - - - - - - - _ _ - _ _ - _ _ - _ - _ _ _ _ .