ML20141M612

Declaration of M Kaku.* Supports Contentions Submitted by Cooperative Citizen Monitoring Network Re Plant.Certificate of Svc Encl

ML20141M612
Person / Time
Site:	Millstone
Issue date:	08/23/1992
From:	Kaku M CITY COLLEGE OF NEW YORK, NEW YORK, NY, CO-OPERATIVE CITIZEN'S MONITORING NETWORK, INC. (CCMN
To:
Shared Package
ML20141M605	List: ML20141M603 ML20141M612
References
92-665-02-OLA, 92-665-2-OLA, OLA, NUDOCS 9209030071
Download: ML20141M612 (19)

Text

89-24-92 10:33 3 SnTeis q OL8tER ST WC 18DB .Tn PB2 1

1 UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSING BOARD la the matter of Nortbeut UtiEties (Millstone Nuclear Station Unit 2)

Docket Number 50 336 OLA

] Pbider Nunnber DPIt 65 4

, ASL'8 Panel 92 665-02-OLA (Spent Fuel Pool Design)

Declaration of Dr. Michio Kaku I, Dr. Michio Kalu, this 23rd day of August,1992, dechre a.nd state as follows:

1. I am a full profenor of theoretical nuclear physics at the Graduate Center of the City University of New York and also the City College of New York. I received rny B.A. degree in physics from Harvard University in 1968 (Phi Beta Kappo esmma cum isode). I received my Ph.D. In theoretical physics ficm the Lawrence Radiation Laboratory at the Univ of CaliL at Berkeley in 1972. In 1972 3, I was a lecturer at Princeton University. Since 1973, I have been a pmfessor at the City Univ. of New York. I have been a visiting professor at the following institutions:-Institute for Advanced Study at Princeton (1990), New York Univeralty (1988), and the Calif.

Inst, of Tech. (1976). I have written 7 books in physics, and pubilahed about 60 profeestonal papers in standard physics jounals. I have also contributed to 10 other books edited by felk= physicists. I am the author of Nweisse Power: Reih Sides, which has became a standard textbook concerning the nuclear controversy on many college campuses. My most recent book is Quatum Field bry: a Modern Inheduction (to be published by Oxford Univ. Press.) and Revoad Einefeis the 9209030071 920824 PDR ADOCK 05000336 G PDR

-E=00N 212 400 0405 06-24-02 tottaAM Poet #38

-l _ - . . - _ , .- ._ . .

m a4-se tem 1. sysTeis a ct.MR ST wt imm - 971 M -

l l

l 2 '

\

Cosmic Quest for the bry of the Uninrse (Bantam Books). I am also a Fellow of the American Physical Society, the largest organization of physicists in the U.S.,

which is an honor only held by the top 10% of physicist.:in this country.

2 I have read some, but not all,of the documents concerning Ncrtheast's Utility's rearrangement of the spent fuel pool for Millstone Unit 2. I understand that Region I will be partitioned into Region A and B, multing in a not replacement of fresh fuel with me,re depleted fuel. I understand that the rearrangernent of the spent fuel was originally proposed as one way in which to compensate for the unexpected rate of degradation la the BoroSex boxes, and also because an error of 5% was found in Combustion Engineering's crisiaal computer calculation of the neutron reactivity,-

, which resulted in k.a exceeding the NRC 95/95 lirnit of.95 for the pool.

3. _I am also aware of the utilities' main argument: that the rearrangement can only reduce the pool's storage capacity, and hence can only help make the poolless dangerous. Therefore it appears irrefutable that this rearrangement is in the interest of public safety. At first glance, this is an entirely reasonable assumption.

4. Unfortunately, a more careful reading of the documents does not bear this out. l I believe that the optimisen of the utility is premature. In fact, after having read some i

of the analysis of the spent fuel pool,I am rather disturbed at the sloppy roethodology and hasty conclusions of the utility. I shall address three main areas (a) reanalysis of j

the criticality study, abowing that the calculation of neutron reactivity rnay not be as rigorous as previously thought (b) reanalysis of the accident scenarios, showing that more realistic scenarios exist for a maxinuarn credible accident which are much more L

' serious than those analyzed la the FSAR (c) conclusions and recomrnandations for t

future action to correct sonne of the inadequacies of the utility's analysie.

l i

Errors in Criticality Analysis l

5. The rearrangement of the spent fuel pool may have a negative impact' on safety  ;

for several reasons. First, the marrangement allows for much rnore highly irradiated J

i I

a.eam

- 212 406 04W5 Os-24-92 10:16AW ' P003 s36

,4,-m --y-,--,-,-,m #v,w-w..-y.,-#v - , ,,... ,-%,3- ,,-w ,, .- ..mw. , ,,, ,_,,,y r,,-,. ,,m,-w-w,- ,,mo ,,.,,.,rr-w.mewww-

es-e4- m te:34 3. rsres o a.Ie GT wrc tams eTbTas4 3

spent fuel to be placed into the pool in Region B, which nay increase the total radiation inventory. The rearrangement trades new, fresh fuel (which has very little accumulated waste products) for depleted fuel (which may contain millions of curies per fuel assembly). Thus, the severity of an accident at the spent fuel pool becomes significantly larger, raising the possibility that fission products may escape into the environment.-

1 The main justiacation for making the rearrangement is that it reduces the storage capacity. However,it my turn out it may not reduce the neutron reactivity to below the level required by the NRC, which is .95. It may turn out that 6esh a e.mpletely loaded Region I md li u well as the " reduced " R*61 on A,B,C will have L greater than the NRC 95/95 limit of Q = .95, because of unexpected degradation of the Boro 8ex boxes and errors made in the criticality study. It is conceivable that the reduction' in L made by the rearrangement may aot be suscient to reluc 4 down to .95. For example, the NITWAL KENO Sa recalculation of the neutron reactivity for the old Region 1 conAguration estimated that k.a = .9812, whleh exceeded the .95 limit. It is therefore entirely conceivable that a correct calculation of Q for both Region 1 as well as for Region A and B will show that k.a still exceeds the NRC limit of ,95.

_7 k.a may be greater than .95 because no one knows precisely how much degra.

dation has occurred within the BoraSex boxes. Only 16% of the Boro 6ex boxes have actually been enmined, which is too small to give an accurate picture of the true nature of the degradation. The utility has made an estimate of the corrected neutron reactivity levels by making " conservative' g=== about the the average presence of gaps within the entire pond, i e. that these hypothetical gaps were randomly dis-tributed. However, it may turn out that more Baro 8ex degradation has occur <ed -

than expected, or that Baro 8ex gaps haw been concentrated in certain areas, mak-lag local distribution of neutrons much higher than the computer calculation for the A

a J... a u . . c.a .. - z . - > z m a n

9 W SS: 3 2. SYSTB15l CLWER ST WC 18Et3B . - Jr1 PB5 1

1 4

entire pool.' As a result, it seems prudent for the utility to throw away its earlier-

- calculation and actually examine all Boroflex boxes to determine the true extent of

~

degrs6dation, re.ther than making unwarranted amumptions.

8. What is disturbing is that actual examination of the Borollex boxes show a l large amount ci erosion, beyond the gaps found in earlier observations at other mactor

[ sites. Since nothing 1 known about the full extent of erosion among the various Boeofler boxes, the computer calculations may be totally obsolete. The neutron i

reactivity studies may not be modeling the actual state of the Baro 8ex boxes, where

an unknown about of degradation is causing gaps as well as unexpected erosion. Until

an inspection is made of all the Boro 6ex boxes, all computer programs are suspect,

, 9. The main problem facing any calculation of neutron reactivity is whether one-can properly model the distribution function of neutrons ((s,y, s, f) in the presence of -

[ the high absorbing Boro 8ex boxes The earlier calculation by CE did not, and hence caused the problem in the Aret place. Unfortunately, the diffusion method often used in these kinds of study is not idenity suited in calculating neutron reactivity with j thin,--highly abscrbing boxes. There are too many hidden assumptions which may l break down in the presence of highly absorbing boxes.

1J

9. One cd the problems is that the utility has not yet given me or the citisens of

- Connecticut a complete copy of their computer codes and calculations.' Therefore,it

! is impossible for anyone to impartially evaluate the cdfectiveness of their computer calculations and, more importantly, their assurnptions. Until this is dame, it is t' heir l word against the ward of their critia. Although the utility cites the public record i concerning the benchmarking of certain experiments performed to check the computer.

analysis, usually these benchmarked studies are highly idealised experiments that may -

- have little to do with the actual problem in question. For example, the reports admit .

that very little actual experimental data exists on performing benchmark calculations -

with highly absorbing Boro 8ex boxes. Therefore until a detailed description is given l -R" e sk' 212 406-0405: 0 8-2 4-s 2 10 : 16 Aas : r005 s3e

-w .r-._ ,r--. ..m. v.,--m-. > m4 -4,+, -u+. m+ . . ~ -.-.- .%,,_.--,,,u.m ,-u~~m. m.. --w,m~e.,~mm,. m,---., m-._ev -m , . . - .

m-24-92 191 % 3. SYSTEMS 9 CLIVER ST W: INDO 9'Y1 P06 5

of the computer calculation and the anumptions behind it, the summary of the results provided by the utility is of little use.

10, Given the sensitive nature of the problem and the luge fission product inwn-tory of the spent fuel pond,it is essential to examine the assumptions inherent in such a calculation. This is imports.nt because the neutron density function p(x,y,s,t)is extremely eensitive to the presence of high dennity neutron absorption boxes. The difusion equation,in fact, may not be able to properly take into account such factors.

From the limited amount of information provided to me, I can draw certain con-clusks concerning the accuracy of the calculation:

(a) the calculation amumes that neutrons in a spent fuel pool behave very much like a gu, in which the neutrons do not travel very fast or very far and are mainly I

governed by elastic collisions. In particulu, one assumes Fick's principle that the f!ux is proportional to the gradient of the neutron density:

J = D94 (0.1)

However, this usumption is only true if the neutrons obey the kinetics ett uonly found in ideal gases, with thermalization and perfectly elastic collisions among the neutrons. In reallife, this assumption is violated by many factors, such as the presence

, of very fast neutrons which do not act like an iden.1 gas and can quickly travel across the entire spent fuel pond without many collisions. In particulu, Fick's principle may be violated in the presence of highly absorbing Boroflex boxes, In the presence of thin absorbers, there are large uncertainties in the gradient of the neutron density function. This would render much of the utility's computer calculations rather useless.

l (b) If one assume. Pick's principle, than the acxt aneurnption is that the neutron ,

density function obeys a diffusion equation, which is a second order partial differential

equation based on the net cnoservation of the neutron population within a small i ,

spherical volume. UnfortunatAy, the difusion equation cannot ha solmi aractly.

I W 4

a.eek 212 406 0406 0s-24-92 10:16Au Poes s3s

- - - - . .. - _ -- . . . = - - . - . - - - . . . -. . .. -. . - . .

3-24-E 10:Yi 3L SWTEMS 'l CL8@ ST RfC 1EEDB . S'TI PE n 6

Lacking an analytical solution, one must make even more assumptions concerning the peutron distribution function. Usually, a computer calculation divides space-time .

up into finite intervals. However, the lattice or cell approximation, again, may break down in the presence of thin, highly absorbing Borodex boxes, espeelally if the width-of the boms are smaller than lattice sise. If the lattice size is too large, then absorbing i

boxes cannot be modeled correctly. But if the lattice sise is small, then this requires inach rnore computatlanal power and time.

I (c) The next assumption is that one can break up the energy spectrum lnto discrete I

chunks, or " groups,' and thers write the net conservation of neutron number between j F

all the various discrete groups. Ultimately, the neutron calculations are performed pn the multi group equations, such u:

f5 y 51 6, - D, - ENE 6, + E0-1E 4r = -a, D, f W2)

A g + %; 4 g,,/ *e a=,+t , a h.: 4-.,

for the sth group, where eventually N -+ oo, where E, a are the. group transfer -!

eroes sections, and where D, is the inverse of the transport cross section. Ideally, N, the number of partitions of the energy, should be u high as possible, preferably in the thousands. The problem is that the calculation ., oased on the assumption that N = 27, which is not a very large number in which to partition the energy spectrum of the neutrons. Given the nature highly abeorbing borated walls, it would esem more appropriate to approximate infinity with the nu:nber N =.200 or 500 at the .

very minimum, rather than 27, which is too small to take into consideration the small edge efects that may occur around the boxes. More:wer,in the pusence of very high i

levels of neutron absorption, the concept of buck 11ag becomes less relevant and the  !

l neutron difusion equation itself begins to break down, so even if N -. oo, the results for the neutron dansity wlD be incorrect.

(d) There la also the assuraption that Monte Carlo sinmixtions can,in fact, provide 1.

1 Repa* 212 400 0405. 08-24-92 IO:36AM P001 s)$-

. --__ ..,e -- .y , , , , . . . . , ,,.,,.,_.__.__-.,m.,. _ ,,,-# ,n ,.._r.m, , ,.g., , -y7pe ,y -,, w.

m-24-9218 71 3 EYST9151 CLDS ST Rrt 1E30 - FY1 P90

-I L 7 i

reasonable estimates of neutron reactivity. The point of introducing Mente Culo simulations is that they can reduce the number of computations necessary to perform I

> a discult calculations by several orders of magnitudes. However, there h a price that one pays. The Monte Carlo simulation is crucially dependent on the number of eyclan or iterations that are performed to approximate the neutron rectivity. If the I number is small, then the Monte Culo calculation will not conwrge very well to the cmcrect result. Because of the unusual geornetry introduced la Region A and B, one l suspects that an unusually large nurnber of iterations will be necessary to provide any l tensonable approximation. Furthermore, there is the temptation to use shortcuts to reduce the number of iterations. Fbr crample, apparently the original CE calculation i arrived at an erroneous value for kg because it tried to use the buckling as a way in which to reduce the number of iterations between the spectral and apatial portions of the amiti-group calculation.

11. The point I am raising is that there we a large number of assumptions that are hidden behind may neatron rectivity calculation, and all these assumptions we, in turn, sensitive to the presence of highly absorbing borated thin walls. A strong case can be made that too many approximations are made in the computer algorithm to give reliable figures. - And the benchmarked experiments, in particular, may be unale== because they are too idealized to describe the system at hand. To revolve these uncertainties, NU must be willing to make public its computer codes and the assurnptions that we behind tham. Otherwise, their claims are just a matter of ~

i speculation, rather than science. .

12. The fact that the presence of highly absorbing boxes can render a neutron reactivity calculation useless is amply demonstrated by the original error mode in the CE calculation. It was precisely the presence of these highly absorbing materials that made certain approximations incertect, such as errors introduced by replacing the notal neutron crees ocetion with the transport neutron cross sectica, and Incorrectly 6

a. set 212 406 0405 Os-24-s2 to:IsAu:-tcos 1Je

m-84-er 18:3s . 2. psraMs 9 CL8@ ST M 1agr y n per B

i handling the buckhng term, tioth errors are highly sensitive to the presence of the dain, highly abeorb!ng BoroDex tmen aud high levels vi neutivu <sboorbing materials.

The errors may seem ernall (59%) but they are not small-when one considers that they may lead to a violation of the .95 limit and create a spent inel pool which is dangerously close to ackleving criticality, in which came uncontrolled arnonnts of

, radiation and heat may eventually be released.

Maximum Credible Accidents

13. ' Die rearrangement advocated by NU willincrease the fission product inven-tory of the spent fuel pool, so it is vital that one analyse the maximum credible accident. There are 5 x 108 curies per fuel assembly after 21 days decay, according to NU, and there will be on the order of 10" fuel assemblies la the pool, So the amount of radioactivity in the pool, roughly speaking, will be on the order of 10' curies, or one billion curies, which is on the order of magnitude of a nuclear reactor core inventory. So one should treat this problem with the same critical analysis given to power reactor ar.cidents. (One should keep in mind that the amount of radiation released by the Chernobyl accident was measured in millions of curies, not billions.

There is more radiation stored in the spent fuel pond that the radiattua released by the Chernobyl mactor.)

14. The utility states that the FSAR's accident analysis provides an upper limit to what might happen at an accident at the spent fuel site. The maximum accident, they claim,is the dropping of a fuel casket, weighing 200,000 pounds causing the breaking of 587 fuel assemblies. Howewr, the eject of this accident is mitigated because most of the impa.t takaa, ples.= ju water. The preena.c of coollag we.ter acts to shield the outside from radiation and dissolve water-soluble fission products from the broken assemblies, so the radiation damage is rather limited. % FS A'R estimated a 241 anecm donc, which is insignificant and well within 10 CPR Part 100 limits. I do not i i bellow that this is the maximum credible accident.

L r----rve -<e er -w,w-+-w-w,-sv.--.-,--,v-w.m we w w-- w r v-,- e s v e vr-', +-eve +-ww-se Peu-*'t=g' *s- 4 e **v'-ta=-- k- tv 41----rv e - r-

6 eventually lead to a ruaximurn Credible accident, or

15. A preview cf what mi ht beyond design basis accident, occurred just a few weeks ago, when the unexpected happened. On July 6,1902, at the Millstone 2 spent fuel pool, there was some loss of power to the cirenlating pumps Without these purnee tv siculaw water, temperatutes begse to rise, and water levcis bc5an to drop la the pool about 3 feet.

Water apparently backed up into the reu. tor centainment, causir.g the sump pumps to kick on. Water eventually had to pumped frem the reactor mntainment buk into the spent fuel pool. Although no fuel assemblies were uncovered, this accident revea that an accident involving a dangerous loss of :ooling water at a spent fuel pool is possible.

16. This accident is also important because it demonstrates how vulnerable spent fuel pools are to a loss of water. By NU's own estimate, it only talass on the order of 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> or so for the spent fuel pool to reach the bohg point of water if the purnps were to fail. Thl , iu turn, can cause a disastrous overheating of the pool and awntual uncowring of the fuel enemblies.

i

17. In reality, a more realistic model than cask (ah.1. provided by the Brown's Ferry accident, where raultiple failures and human failures wer. neported becau a worker carelessly used candle light to seuch for a leak, and wound up setting off a major conflagration in the insulation. The uncontrolled fire caused major loss of con, trol of the reactor and an omioous drop in cooling water. The accident overwhelmed the local emert/.acy teams on the site, until it was Anally put out by the local fire de-

, partment. This accident scenario, wh!ch caused major damage to the safely systems, millions in repair costs, and almost initiated a LOCA, was never anticipated by the industry.

, 18. The faulty assumption in the FS Aftis that accident scenarios arc oclyinitiated by " single event failure,' such as the dropping of a single fuel assembly or cask.

However, this is highly idealised. In actual reality, all major acddeets of the past,

. *g ,

4 d

l t

~

212 TWLTOS C6-24-92 10:16aN P00s s38

aa-a4-EE 13:40 3. 5M5 TBS 'l O.hER 5* WC 1008 Em P8a 10 such u TMI, Chernobyl, Farrd.!, Sr .1, Brown's Ferry, EBRd, Dresden, etc., were caused by multiple mode failures coupled with human failure and design flaws. In fact, no where in the entire accident record do we have an actual major accident proceeding according to the idealbed predictions of WASH 1400. The celebrated

'dooble ended guillotine break in the cold-leg pipe of the primary system, intensimly studied in reactor acddent courses, hu never occurred in blatory, while the bulk of actual uddents is bued on multiple mode and human failures.

19. A more realistic muimum credible seddent might be a loss of cooling water, causing overheating of the pond. This common snode failure might be initia,td by a singte event (f re, chemical explosion, sabotage, earthquake, airplanc cruh, lightning bolts) which causes multiple failures. Given the arnple precedent of previous accidents, one can usume that a fire or chemical explosion can cause major damage to the reactor building, causing an evacuation of the site. Elo-trical power is lost and the spent fuel pool is unrupervised, and within hours the temperature rises sufficiently fut to cause boil off and rapid evaporation, ew.ntually uncovering the fuel rods.

20. Without circulating cooling water, the temperature rises rapidly. Within 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br />, boiling may occur u ternparatures rise to 2128 F desmes. At 360' F, radio-lodine in the rods begins to boil and leak out. At 1250* and 1800, radio cesium l and tellurium begin to boil. At 12000, ballooning and distortion of the zirconhun cladding occurs, releuing Ruion products into the water. Radioactive xenon and krypton gases are then releaml directly into the environment. At 1400', the cladding swells and finally ruptu ee. At 1800', the zirmnium starts to oxidize rapidly, creating large quantities of hydrogen gas via the metal-air and meta!-water reaction. Then any flame or spark can create an hydrogen su explosion that will pulverize most of the fuel suemblies, causing highly radioactive debris to escape into the environment.

21. The hydrogen gu explosion scenuiohu ample procedent. On the afternoon of the first day of the TMI accident in March,1970, enough hydrogen su from airoonium l

i

)l l

I a-e rw 217 ane n a ta, 08-24-02 10:16AM P010 #28 j 1

1

B-24--92 18: 4& 3. SYST96 'l CLI@ $7 Nyt 12238 5"rl P8&

d l .

~

11 i.

oxidation was released into the containment from the oxidation of aircordum to cause an explosion. Fortunately, the containmut wu able to withstand the impact of this hydrogen gas explosion. A hydrogen gas explosion has also been implicated in the Chernobyl accident.

I

22. In addition, sabotage cannot be ruled out. In fact, we han several highly publicised cases where sabotage was either carried out or was shown to be possible.

At one reacter site, disgruntled workers walked up, unirnpeded, to the spent inel pool and poured sodlum hydreadde directly Into it.- Fortunately, these workers did not

know enough about reactor physics to cause rnajor damage. Although the sodlurn hydroxide could be cleaned up, without any darnage to the fuel rods, a more skilled

, , worker might haw, for example, released certain valves and lowered the water level, or siinply dynamited the spent fuel pool. In real life, and not the alatively sterile i

world of " single-event tree analysis," individual workers get angry during strikes and will sometimes deliberately damage sensitive equipment in unforeseen ways.

23. On another occasion, safety of!!cials,in a test, placed a gun in a suitcase (sealed

in plastic so no one could get hurt). They then, with relative eue, walked with the I

briefcase past the security guards and entered the control room of the reactor. Had they been real terrorists, they could easily have seized control of the nuclear power I

plant and peam.ed unlimited mayhem,i.e. unscramming the reactor, shutting odf the RPI and LPI within the ECCS, etc.

24. One should not diamiss lightly the statement made recently by a. representstive h of the Yugoslavian government that they will deliberately sabotage nuclear power

plants in the Went if force is used against Yugoslavia. National governments, with all their resources, can cause damage far in excess of the damage caused by Individual

!~

workers. If Fw crisis worsens, then every reactor is fair game.

[ 25. One should not dismiss'the possibility of an earthquake,_which may set off -

snultiple mode failure within the reno 6er med the spent fuel pond. NU% own analysie i

i i

l l

a-osx  :: 4es oses. co-24-on sortaiu voti-ese F . .._._.___.__.._1 . ._ . _ . . . _ . - _ _ . _ . _ . _ . _ . _ , _ . - - . _ . _ _ _ _ . . _ _ _ _ _ . _..

~

I!E-24-92 18i42 ~1.~ SYsTes 1 CL5ViiR 5T M2 8MU) LTiH (M 12 considers earthquake damage resulting from strenes which cause .09 g uceleration in the horisontal direction, and .06 g acceleration in the vertical direction. However, i

these stresses ue far below the actual stresses found near large earthquakes, which can cause accelerations approaching 1 g. Although euthquakes are unlikely, no one knows how to predlet their frequency in the Northeast. Unlike the San Andreas fault, which is a clean, isolated fault line where paleo seismology can utimate the rough cycle time for estthquakes, earthquakes la the Northeast do not lie along such simple earthquake faults. The area is much more irregular, meaning that paleo-seismology dom net give an indication of the frequency of earthquakes in the Northeut. The point is th.t an earthquake can set of a common mode failure, resulting in a scenulo which can damage valves, pumps, set off fires, etc. which may cause the spent fuel pool to leak or lose water.

96, in 1975, there ww. two la.ndmark studies done on nuclear accidente, WASH-1400 and the American Physical Society Light Water Reactor Safety Report study (published in the Rsviews ofModern Physics,47,1, p. St.S-123.). They took a major step forward in calculating what might happen if allsafety systems failed at a nuclear

pown plant, regudlem of how unlikely that might be. They calculated what might happen if up to 75% of the core invootory of a reactor breached the containmet and wu released into the environment. This was important because, before then, the nuclear industry insisted that " defense in depth" wu su5cient to render such catutrophic accidents impossible, so therefore there was no need to analyse such accidents. Since then, because of the FOIA, we know that during the height of the TMI acddent, the NRC Commisioners secretly discumed whether WASH-1400 eosaarloe could actually happen if the reactor went out of control. The " impossible' accxlent scenarios of WASE1400 suddenly became the main topic of conversation at one of their important meetings during the crisis..We also had graphic proof of the usefulness of such studies when the Chernobyl accident released over 5% of its core-I a - s e ,s 212 406 040s f.8-24-92 10 16 Ang P012 838 y ~ -

m-24-92 $8:42- 1. SWT96 9 CL86 ST HYC imJe 9h) P1'i 13 inventory over the Ukraine.

27. A. a con cquccee,I believe that the NU ebould do the counterpart of WASH-1400 and the APS study, i.e., it should calculate what rnight happen if 75% of the invetory of fisalon products from the spent fuel pool were released into the envi-ranment. Specifically,i; ekould calculate the density function of the fluton products released in an accident by solving the standard difusion equation:

K + + x(2,y,1,f) = d*j#I .(0,3) shone solution is the standard Gaussian distribution for ead Assion products

~

dz,y, r,Q = (4,t)3f,(g g,g )3f, up g +

M where K 4are related to the standard Pasquill coemcients. ,

28. The proposed study should calculate the deposition of fission products over Connecticut, given the fact that the radioactive plume will rotate, much like a light-house, beesume of changing wind patterns over time. Then the study should ' calculate the fission product density deposited on the ground and the rate at which ingestion of these radioactive products takes place into the human body. We can then calculate the total ingestion of Aselon products (in person rems) by the population by Integratlag the denalty function over the depooltkm area:

D= #[x(r,1)p6 Frdr = [1 - exp(-Aff- 4/s) R) : (0.5) where p is the average population density of Connecticut,6 is the breathing rate, and F is the dose conversion factor in rem per curie inhaled.

From this, one can truly estimate the real impact of_a spent fuel pool accident.

Coneluelon and Recommendations 4

-a c6* 212 406 0405 Os-24-92 10:1sAW P013 eso

es-24-T 18:43 3. SWTEPS 10.hER ST rM taEIS SM P14 ~:

e 3-14 j 29. la conclusion, I am not so optimistle that the rearranged spent fuel pool, when fully loaded in the future, will meet the criteria that koe < .05. Although the utility states that reducing fresh fuelin the spent fuel site can only reduce the neutron levels, I am not conviamd. The assurnptions behind the computer calculations are j not sufEciently re!!able, specially in the presence of the highly absorbing Boro 6ex l boxes. In fact, many of the assumptions behind neutron transport theory begin to l break down precisely because of the presence of highly aboarbing thin walls. One's i- conclusions are only as valid as one's assumptions.- Or, as they say in the industry, j " garbage in, garbage out." This discussian is not purely academic, because the As.

l elon product Inventory of the pool will eventually reach one billion curies, which le

gable to what is found in a nuclear power plant.

30. The prvvio-e reactivity study by CE done on the spent fuel pool was in error by 59E, mm.alf bemme of the diSculty in modeling the Borofler harm by the neutmo diffusion' equation. I am not convinced that the newer neutron reactivity strdy is sensitive enough to truly calculate the sfiect of neutron abnorption' by the Borciex boxes, especially because of the degradation and unexpected erosion of the boxes (whose full extent has never been determined by the utility). -The neutron l reactivity calculations using Monte techniques studies have inherent uncutainties in them (given the assumptions inherent within the model) that may be too large to make reliable estimates of k.s for the fully loaded pool.

31. Given the fact that more spent fuel will be storal at the site, near populated areas, with about one billion curies d fission products, I think that NU should model a more realistic accident scenario. It should abandon the simplistic single mode failure model (which has never happened in a majee nuclear accident) and adopt a more _

4 Saxible and realistic multimode failure / human failure model, which agrene more with the history of past nuclear melting incidents and fission product release accidents.

32. Speci$cally, a credible scenario exists in which the water level dmpa danger.

4 a295N 212 406 c406 Os-24-92 10itsIM-' rote was I

,____.2_.~., . _,_....-.;,,___......._-_....._, - , _ , _ ,

_ EB-20-4B2 18 44 3. SYSTM 1 CLMR ST tht 10030 9"M PAS -

l.

15 ously in the pool. For exunple, a fire or chemical explosion may cause an evacuation ,

of the alte, teuttag to a power failure. Without anyone monitoring the pool, one can imagine the water level dropping due to Inks, boll of, and evaporatica as the temperature rises. It only takes about 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> to cause bolling withic the spent fuel pool. When the fuel usemblies uc uncovered, the temperature may be suScient to i j

i cause hydrogen gu generation and then an explosion, dispersion large amounts of '

fission products into the envirunment.

33. In light of these difficultles, I would lilee to make several recommendations:

First, that the utility cury out i, full ecale evaluation of the Boro 6ex boxes to deck for new sspe as well as measure the rate of erosion. Until this is done, all computer programs are largely useless. The utility should also perform rigorous benchmark studies using Boroflex boxes with the the a,etual geometry found in the spent fuel pool, not just idealiv.ations of the. geometry.

34. Second, the utility should carry out the rusonable demands of citisens groups, such as releasing a copy of its neutron reactivity calculation, and placing neutron ,

detectors around and inside the pool. This is reasonable, since deter. tors have a proven worth, 7br exampla, the paveence of such a detector (which could measure the level of water at TM1) could have prevented an accident whleh has already east GPU 81.5 billion. Neutron counters could give a rough ladication of whether the peal had higher than expected neutmn reactivity before an accident goes out of control.

35. Third, the NU should be required to do a realistic analysis of a maximum credible wddant, i.e. the release of 75% of the Assion product inventory late the -

environment. Like existing studies of nuclear reacters, one should assume that all.

safety systems me somehow volded, and that large amounts of fission products escape into the environment la the form of a plume. Slace the distribution of Basion products in diferent from a conventional nuclear reactor, one should obtain diferent results for l

a spent fuel accident. The fact that,50 years into the nuclear age, such as basic study L-i-sex att see esos- os-as-sa ioneu f rois ass

- =

r l 15-24-92 10145 3. Sh1liTP16 *1 CLM.P. ST NYC 10838 FT1 P15 16 for a spent fuel site does not exist is a testunent to the fact that nuclear waste has alwasys been given low priority. However, now that nucleu power plants we gradually filling up . pent fuel sites and we beginning to consolidate and repadage spent fuel, it la vital that such a study be. done.

36, UntG these recommedations are carried out. I cannot truthfully state that a fully loaded spent fuel pool in the new reurangement is esfo On the contruy, it may ewn prove to be a health hazard.

I declue, subject to the pain and penalty of perjury, the foregoing is true and correct, to the best of my knowledge.

I Sisnad I

bb W Michio Kaku, Ph.D i

i I*88% 217 408 0&C% Os-24-G2 10'14AW . Done w20

r

. ' t:L;

.a UNITED STALES OF AMERICA NUCLEAR REGULATORY COMMISSION 92 AUG 27 All:10 In the Matter of

. !; a : e i ,,,u

. , i . -

NORTHEAST NUCLEAR ENERGY COMPANY Docket No.(s) 50-S$0llid,j,i' W (Millstone Nuclear Power Station, Unit No. 2)

CERTIFICATE OF SERVICE I hereby r/ d e that copies of the foregoing C(Ny ac drd,,y #s.p,May de -~/f ,

have been served upon the following persons by U.S. mail, first class, except as otherwise noted * ' ' '

Office of Commission Administrative Judget DecArL S U.S. Nucle +are,-ve Beh Ivan W. Smith, Chairman Regulatory Commission- Atomic Safety and Licensing Board Washington, DC 20555 U.S. Nuclear Regulatory Commission Washington, DC 20555 Richard M. Kacich Michael J. Pray, AIA Director, Nuclear Licensing 87 Blinman Street '

Northeast Utilities New London, CT 06320 P. O. Box 270 Hartford, CT 06101 Patricia R. Nowicki Nicholas S. Reynolds, Esq.

Associate Director John A. MacEvoy, Esq.

EARTHVISION, Inc. Winston &-Strawn 42 High1and Drive 1400 L Street, N.W.

South Windsor, CT 06074 - Washington, DC 20005 Da4eA 8 WWM a-l.1.4l%-- e. a c..s.s lospW/ A _ iYize nfo $0as be/~/ kLbnv -

9 e .

e

UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION In the Matter of NORTHEAST NUCLEAR ENERGY COMPANY Docket No.(s) 50-336-OLA (Millstone Nuclear Power Station, Unit No. 2)

CERTIFICATE OF SERVICE I hereby certify that copies of the foregoing LTR MARUCCI TO LB DTD 8/24 have been served upon the following persons by. U.S. mail, first class, except as otherwise noted and in accordance with the requirements of 10 CFR Sec. 2.712.

Office of Commission Appellate Administrative Judge Adjudication Ivan W. Smith, Chairman U.S. Nuclear Regulatory Commission Atomic Safety and Licensing Board Washington, DC 20555 U.S. Nuclear Regulatory Commission Washington, DC 20555 Administrative Judge Administrative Judge Charles N. Kelber Jerry R. Kline Atomic Safety and Licensing Board Atomic Safety and Licensing Board U.S. Nuclear Regulatory Commission U.S. Nuclear-Regulatory Commission Washington, DC 20555 Washington, DC 20555 John T. Hull, Esq. Richard M. Kacich Office of the General Counsel Director, Nuclear-Licensing U.S. Nuclear Regulatory Commission Northeast Utilities Washington, DC 20555 P. O. Box'270 Hartford, CT 06101 Patricia R. Nowicki Mitzi S. Bowman Associate Director Coordinator EARTHVISION, INC. DON'T WASTE CONNECTICUT 42 Highland Drive 97 Longhill Terrace South Windsor, CT 06074 New Haven, CT 06515

Docket No.(s)50-336-0LA LTR MARUCCI TO LB DTD 8/24 Nicholas S. Reynolds, Esq.

John A. MacEvoy, Esq. Mary Ellen Marucci Winston & Strawn 104 Brownell Street 1400 L Street, N.W. New Haven, CT 06511 Washington, DC 20005 Michael J. Pray, AIA Frank X. Lo Sacco 87 Blinman Street 4 Glover Place, Box 1125 New London, CT 06320 Middletown, CT 06457 Joseph M. Sullivan Rosemary Griffiths 17 Laurel Street 39 South Street Waterford, CT 06385 Niantic, CT 06357 1

Dated at Rockville, Md. this 28 day of August 1992- / - /

h &

Office of the Secretary of the Commission l

l t

, , ,s ;r ---,- , , - , e r- -v -, rn,. -. --e... nr-