ML19322C385

Safety of Nuclear Reactors, Presented at United Nations 550808-20 Conference on Peaceful Uses of Atomic Energy in Geneva,Switzerland

ML19322C385
Person / Time
Site:	Crane
Issue date:	08/20/1955
From:	Mccullough C, Mills M, Teller E CALIFORNIA, UNIV. OF, BERKELEY, CA, LAWRENCE LIVERMORE NATIONAL LABORATORY, Advisory Committee on Reactor Safeguards
To:
References
TASK-TF, TASK-TMR Y-14463, NUDOCS 8001160920
Download: ML19322C385 (10)

Text

.

a

~

/uc(kc4 The safety of Nuclear Reactors By C. Rogers McCullough,* Mark M. Mills,t and Edward Teller,t USA REACTOR TECHNCLOGY ybesteSo far, there have been essentially no reactor accidents leading to serious consequences. For this In any new tic!d of technology, it is :spasmucqs reas n, statimeestkinspagnationalmostenetocaccidets, insenghJMapajtaelv.if possible;;as; many.fet tur'e s' kidheispiAas3scen's,tpSrtincr4for7 humatmve4 aghougitakapearablerdoesmot=saEicetta givesuselul farc.Juclear reactor technolog'v is such a field, and spyagategiappng thgnypenceded>by insur, no one looks to it with hope for many material ance.c mpaniesMomexample, m e4mnaue benedts for mankind. Among these bene 5ts are the t W-p ther words, to determme what is an possibility of electric power generation, propulsion acceptable risk, a certam amount of ju%=anddh by nucJear enercy, and the utilization of reactors as tan g - - = l% g w m mogag m b&c.

' branches of science and C'F988M""'Ohfa98tP cysdg research tools in many

. With all the mherent sateguards that can be put into a reactcr, there is s'ill no tool-proof system. Any

.W.n with a long !ist of possible attractive fea-tures of reactors, there are, unfortunately, certain

'J. stem can be deteated by a great enough fool. They :

dangerous characteristics. The Advisory Committee resMungenwasurswhen a#semdmnasim on Reactor Safeguards (see Appendix) has the re-Problems or reliabih. "*50"D N C*

ty, acequate control, adequate l

sponsibility of looking at the hazards connected with I

nuclear reactors. The members of this committee are supervision, must all be meluded. It is convenient to I k upon the concepts of reactor safety m the tol-exceedingl :.nxious to see rapid and fruitful devel-l cpment of' reactor technology, but because of the I *I"E ""F8 nature of the hazards involved, and because they One unponan* concept is the division of safety have been specifically requested to look at hazard pr blems into on-site and ott-site problems. The on-problems, tiiev feel it important that no undue risks site problems have to do with the protection of t

be taken in the development of nuclear reactors.

reactor operating personnel and other people who may be at the reactor site in order to make use of it, REACTOR SAFETY and the protection of the economic investment in the Innnedhtely, when one attempts to evaluate re-reactor facility Off-site problems have to do with i actor hazards, there is encountered the necessity for the protection of the general public, or persons who attempting to dedne the notion of reactor safety, and are not more or less directly connected with the op-w hat this notion shall include. Of course, absolu(e eration of the reactor. One way to minimize ad-site saictois not.possible audMnat,.is~.reallw.ineant in hazards is simply to locate the reactor at a remote conncction with,reactorhagnis igthe.Jninimizatigs 3nd unpopulated place. In terms of reactor utiliza-

~ ~

s of hazards.until one.has an acceptable calculated.rish tion and economics, this solution is otten unsatzstac-The operation of nuclear reactors appears safe tory. The economic utihzation or electric power rud it i3 in fact, deceotivelv safe. A nuclear reactor generated by reactors, for example, nearly always wid n. t run awar uniess n' number of serious mis-requires that the reactor he located reasonably cbse t potential users of this power. ihis means that ict take..i planning and operation should be committed.

economic reasens the reactor -hauld be located near It is however, impossible to conduct extensive op-crations over a km; time without occasional occur-Populous, industrial areas.

rences of such mistakes. We have been exceedingiv Substantial moral and ethical problems are in-luckv m far that nobody has as vet been killed by a v lved in connection with reactor hazards. On-site runa'way reactor. It is not possible to count on 'in-personnel, like persons working in other industries, ded iite continuation of such ud luck, knowingly and willingly submit iemselves to what-M9ddCND CJ"I"NE.3Leact m hazas m apciated,vith working near a OW#ktiFJ1'8KC2 haz,gigadqenggrlgrihowgmenchii'tNeactor,ac&g L

reactor because et satary reqmrements. special work-mg condiuons, or personal interest.

,,: Ch.vrman. Mvisory Com:ri: tee crc Reactor Safeguards.

For o Y-site peopM. on the ntner hand, who have n knowiedge or interest in tl e operation of the re-r,in ni California Radiation Laboratory, Liver-m c2% inia.

actor, it seems that prevention of danger to inear r

Nrar: ment of Physics. University of California.

persons or dama;c their property is a mandatory 79

~0

\\

'hjj(@fmMAM, 8001160[74

'-l

f 80 VOL. Xtli P/853 USA C. R. McCULLOUGH o el.

moral obligation in the operation of a reactor. This of these dssion products from the machine. The po-problem is more severe than in the case of dangerous tential ability of a reacter to run away makes it chemical or explosivessplants, because the radioactiv-possible for this radioactive material to escape to the itv contained in a reactor can constitute a hazard to a surrounding areas. The hazard is crudely analogous w'ide area if it escapes from a machine and becomes to conducting both explaire and eirulent poison di3persed.: This public hazard has been one of the production tmder the same roof.'

main concerns of the Advisory Committee on Re-IJntil really safe nuclear machines of the future actor Safeguards.

become available, we have to construct our reactors Froin another point of view, the safety of a nuclear with extreme circumspection and we must continue

, react %eam booedtM= " WtWof thinpMht to operate them with the same caution after ten years

!inympici,imilsnumstahdearandsreliabilitv*ofrthe uh of safe running as on the very drst day when they

• oittherspac1hane were started up.*

! clut-ickMor example, the reactivity may de-In order to emphasize the characteristic of the

!cr' esse rapidly with increasing temperature. In this special hazard due to radioactive materials in the case, it may be practicalt-impossible to exceed some reactor, a list of tolerances is presented in Table I.5 safe limit in temperature. This intrinsic stability is Although there has been a substantial effort in the very desirable. In fact, one may sav that a machine assessment of the effects of radiation on biological witit large intrinsic stabilitv c'an be so stable, be-systems, particularly systems resembling people, cause of fundamental phviical characteristics, that there is still a great deal to be learned.8 However, only a Maxwell demon can make it misbehave. An even allowing for considerable error in the qt;antita-ord'inary machine, which depends on the operation of tive assessment of this problem, it is still evident from the control system to set its power level, can be upset Table I that radioactive poisons are more hazardous by a mere gremlin! One would like to minimize the than chemical poisons by a factor of something like 108 to 10'. This is such an enormous factor that radio-dependence upon administrative control for safe

' operation of a reactor. However, asanppsesqqatprag.

active poisons essentially must be considered a quali-ticagypese reactors;will.nearly,ahvays,requiry tative new kind of problem. Furthermore. this imphes s

/

A=wence upos:adminastrative control fog I

a cer dliabIe. operation. This means that prob-Toble 1.5 Comporison of Toxic Substances in Air

• saidd

(* ""'"" " '" * #* I tem 4 arise connected with the loading and unloading of fuct, the startup and shutdown of the reactor, u,

g, proper manipulation of controls, and adequate ac-g,,,,,,

3.,,,,,,,

w,,,...

countme for all materials made radioactive by the re-C**'"'**'"**"'

actor, including both intentionally irradiated' material and any radioactive effluent associated with the oper-

[y""'

y6 8001 5

ation. Thu the normal, as well as the abnormal oper-Beryllium 1.5 x 10"

?

?

ation and behavior of the reacter must be carefully

,t,g ;,,,,,,,,,;,,,,,,

considered. It is clear that a reactor which m normal U" (m. sol) 1600 x 10*

1690 x 10d 10.000 operation is well run and under complete and precise Pu" 32 x 104 32 x 10*

10.000 control is much less likely to behave m an abnormal Sr" 1.3 x 10*

1.3 x 10" 10,000 fashion leading to a serious accident.

• 1t should be remembered that industrial poisons are THE CONTAINED RADIOACTIVITY usuapy in many ton quantities, whereas radioactive poisons are m ICO-kilogram quantities.

The most serious continuing hazcrd associated with.

t" Tolerance" for chemical poisons is denned as the max.

nuclear reactors is due to the large amount of radio.

imum tolerable level for S hours per day exposure. In the case f radi active p is ns : 1erance is the maximum level activity which they contain. LarEe reactors may wmch can be tolerated every day tor S hours equivalent to contain hundreds or, pounds of radioactive tission 0.043 rem per day.

products which correspond to many tons of radium

" Fatal Dose" in the case of chemical poisons is desned in conventional radioactive measure. Et all of these 25 @'

• rapidly fatal" dcse when the given concentration in air is mhaled for 30 mmutes to one hour. In the case of ra-tission products are as hazardous as radium, but dioactive material this means about 50'~e survival if the nearly all of them contribute substantially to the dose is acquired quite rapidly, for example, over a minute or hazar'd.8 There are two ways in which th'e hazard

(*[,s gunne an 84mur day This is equivalent to about of contained fission products may be minimized : One

" Adopted at meeting of the American Conierence of is to remove tission products during the operation of Governmental and Industrial Hygienists in Atlantic City, the reactor in such a war as to maintain a minimum N* J i" ^Pril 1931-concentration of such m'terial in the machine. This I. Industrial Hygiene and Toxicology. Frank H. Patty, a

Editor. Interscience Publishers. Inc 1949.

continuous removal of fiss. ion products regtnres some

.* 1taximum Permissible Amounts of Radioisotoxs in tvpc of :iuid fuel, either liquid or gaseous. in order the Human Body and 11aximum Permissible Concentrations in Air and Water. Handbook 52. National Bureau of Stand-to continue cleanup operations on the fuel during' the ards. Afarch 20. 1950.

operation oi the machm.e. The other way to mimmize it The Effects of Atc.mic Weapons. US Government the hazard is to mmimize the possibility of the escape printing orsce, Revisee senember 1950.

h h

$ff d h M T M( V 6

A m

THE SAFETY OF NUCLEAR REACTORS 81 that the problem of keep;ng radioactive materials Table ll. Deloveci Heat Power and Radioactivity

• within the reactor anri preventing the spread of radio-cirer Normal Shutdown active materials over populous areas is verv serious.

In Table II there is a summary of delayed heat

,,, ffMjIf?"'/,, g:

prmiuction and the corresponding radioactivity from n,,,,,f,,,

m.,

mm.,

645 ion products. For a machine of 250.000-kw heat

>^= ado r'a ac'm'i ss:

aerm's a3:

power ( 60.0CO-kw electric power), something like m

w

< = +,

6

<=^<>

39) million curies of activity remains at the end of 10 10 see 12.9 2.1 x 10' 11.000 1.8 x 10' one dar niter shutdown. This corresponds to 300 tons 108 1.7 min 8.0 1.3 x 10' 6800 1.1 x 10' of radium in terms of radioactivity. The sheer quan-10' 16.7 min 5.2 S.4 x 10' 4300 7.0 x 10' 10' 2.8hr 3.3 5.3 x 10' 2700 4.4 x 10' tity ot radioactivity is enormous.

10' 23hr 2.0 3.3 x 10' 1700 2.8 x 10' Opoan,on or this reactor for one year produces about 100 kiicgrams of 6ssion products. On the basis

• The radionctivity figures are for fission products only of 10-7 mg < cm3 this can contaminat-10* cubic kilo-(do not include radioactive fuels or components). It is as-meters of air to tolerance. Said another way, a layer sumed that the mean decay event corresponds to 1.0 Mev of air one km deep covering an area 1000 km on a side could 1.e brought to tolerance level.

of human reaction times and conventional external Another teature at radioactive poisons is that a emergency human actions, nevertheless, a nuclear lethal level is not detectable by human senses. Fur-thernue very serr.ais injury may n t be detected for reactor is a very sluggish device and does not pro-duce a nuclear explosion even remotely approximat-wme years aster exposure.-

ing that of an atomic bomb. Irbai.hddarse ESCAPE OF RADIOACTIVITY th _ _ T"ost nothagMiktsangsspkaiourealg, The war in which reactors can malfunction and empF r very fast react rs with a n n thermal lead to the' escape of fission products may be classi-neutron spectrum and heavily loaded with enriched 6ed as follows ; (1) a super-critical nuclear excursion uranium, it does appear possible to have an accident or nuclear rtmaway: (2) melt-down of reactor com-which is fast enough so that portions of the machine ponents, even with the chain reaction shut down, may be propelled with velocities of a few meters per j second. This again does not resemble an atomic bomb exP osion, or even the explosion of ordinary chemical l because of the delayed heat produced by the radio-l active 6ssion products; and (3) possible exothermic l

i

• • P osives; rather it is similar to the events that :

che nical reactions among the components of the re-might occur m an automobile accident. Therefore a actor itsdi. The latter, although it is clearly not nuclear runaway, m, itself, does not represent seri-present n the machine is operating normally, may be nunated bv a runaway nuclear cham reaction or by ous hazard to ort. site people.

delaved he'at melting. -

However, as pomted out above, a nuclear runaway can serve t do two thmgs It may disrupt the struc-These prdulems will be discussed in more detail

,aelo v. The drst two are unique to nuclear rentors ture ot the reactor suf6ciently so that radioactive Poisons may escape, or it may lead to exothermic as compared to other power sources, and have no true analogue in other areas of technology. They are dis-chemical reactions between different components of the reactor core, and a chemical explosion of con-cussed m. some detail, for research reactors,, and siderable violence. Indeed, for certain types of reactor nuclear power plants' elsewhere.

structures, it would appear that the chemical reaction THE PROBLEM OF NUCLEAR RUNAWAY that might follow a nuclear runaway would produce substantially greater energy and violence than the

. An outstanding characterisde of nuclear reactors runaway which preceded it.

is their potential abih.tv to ashieve extremelv high In order to make some of these notions more quan-power levels m. a short time if adequate control of titative, it is conveni;nt to talk about the rising period

,the machme is lost. A typeekandeep runaway.acgi-of a nuclear reactor. A nuclear reactor which is super-Ident,mansvart>and be over.in.t,irqcpappreciably 1%

critical increases in power level bv a factor of e at each than.agecond in this. respect they,a Edifferent.ftcEn interval of time corresponding to' the so-called e-fold-ayygMf,.

specaleraughing and it is this ex-ing time. In turn, the c-folding time is related to the tremely short time that makes it quite important that intrinsic neutron generation time of the reactor, and automatic control and safety systems be available, be the degree of supercriticality, by the so-called in-hour reliable, and be relauvely.;apid m their operation.,

equation. In Fig.1. a number of curves are shown a

Anomsmenaeure%=c - - : =:

-< a connecting the rising period of the reactor with its 1 tiniLdesaNt.seemm4e.1aary.violess A comparison' excess reactivity, i.e., the fraction of excess neu-between a nuclear rea: tor and an atomic bomb is very trons produced in one generation. The c-folding mi-leading and certainly not to the point. From a times shown in the 6gure are relatively lon~. This number of studies of possible reactor accidents of is due to the delayed neutrons. As you all'know, this type, it must be concluded that even though re-the 6ssion event produces certain 6ssina products actor accidents could happen quite rapidly in terms which. in turn. after periwis rangin t.p to 80 seconds

'I 82 VOL Xill P/853 USA C. R. McCULLOUGH st of.

the c-folding times are correspondingly reduced. Here 3

1o one may say that nuclur reactors represent a genume u

II departure from conventie.nal power sources in that q

l l

I enormous power level increases are possible in the t

?

to

-i h

i i l I

' I event of mal-operation in remarka!.ly short times.

3 i

l l

TheS -

_-- haerarteaWtiedh

'O l

l mcutf w%;Lupg&uss wa@>merm m -

\\

t s\\

i i t i

t ge:g y

_ muom ene nka==*mta These mee3 are expecteu to have two characteristics.

l l

First of all. they should be self-contained and wholly automatic so that they are not subject to error of 10 q

t,

{

adjustment or maintenance and are not subject to 2

e l

g l

l intentional tampering. Second, these fuses are to be i M\\\\\\i cr.a e,raero'cooa -

activated by changes in the power level, essentially i d,'

m*ea G*a' secocat' changes in the flux level of the nuclear reactor. and 5

[\\ z have rapid enough response so that they will intro-2

-2

- \\\\'

'M duce a substantial negative reactivity in the reactor l M$c DI '

I in a time of the order of one second'or less. One of 0

I the continuing dif6culties in the development of these

,4 fuses is this latter requirement for short-time opera-6 3

,#c':cci 000i 0 0t 01 10

'O tion. It appears that successful development of wch Excess Reactmty,k-1 a ittse will soon be achieved, but because of the short-Figure 1. Variotica of rising Period with excess reactivity from time requirement this deve!opment is neither easy the lawoor egooriaa:

nor simple.

6 i = f. -- 6 [ d'

As a matter of practical fact, one must consider how a large excess reactivity might be achieved in a

~ "'

s nuclear reactor.' First of all, it is clear that it can emit delayed neutrons. The fraction of these delayed not really be achieved instantaneottsly although some-neutrons produced in UN fission amounts to some-thing analogous to instantaneous excess reactivity thing like Mo of all neutrons produced. If one makes can be obtained on startup of a reactor if only a the excess reactivity of the reactor so great that the weak source of neutrons is used during the startup chain can proceed without the delayed neutrons, then procedure. It is then conceivable that through some the reactor is said to be in a prompt critical condition, error, rapid removal of control rods would allow the that is it is critical or even supercritical on prompt reactor to be highlv supercritical before the power neutrons alone. In this condinon the c-folding time level had risen to s$mething approaching the normal becomes short and one may estimate it by means of power range. For this reason, startup accidents are the equation:

particularly to be avoided.

1 k.,

In arv event, one must consider not only the possi-

,, =

re to ble degree of excess reactivity but also the rate at In this equation, r, is the c-folding time, to is the in-which reactivity may be added to the machine. For_

trinsic neutron generation time which depends on the this reason one would like safety rods and shim rods type of reactor, and k,, is the excess reactivity above that move out rather slowly but which could be re-inserted rapidly at any point during withdrawal.f p'rompt criticalitv.

Trpically, neutron generation times are about one One would also like the degree of control residing millisecond for large thermal reactors, something like in the control and safety system to have a g aded l'o millisecond for water boilers and small thermal weight so that as the reactor becomes nearly critical.

reactors. and may be as short as a microsecond or less only sma!!er amounts of reactivity are introduced by for inst-spectrum epithermal machines. The value the withdrawal of control rods. Safety rods which that one may assign to k,, depends on the type of must be completely withdrawn and cocked before machine and its requiremenp3 for excess reactivity in they may be re-inserted are particularly undesirable.

order to conduct experiments, overcome temperature There is another point which is quite pertinent in eficcts, allow for burnup ci the fissionable material, or the serious consideration of how rapidly excess re-override 6ssion-product poisons. However, it seems activity might really be added to a given nuclear reasonabic to assume an excess k-value of about 0.01 reactor. Extremely fast reactors for example. look fraction for terms of discussion. If this is done, then

. I In uda t av iJ circumk>cution. all remarks concern-an c-foldinE. time for the large thermal machin.es of mg control and safety systems will be made as though these.

sccond 15 obtamed. Only seven c-fold.mg tunes.

were conventional ab<orber systems. of coune it is entirelv U

that is U second, are required in order to increase possible to increase the reactivity at a reactcr by..mertinh go the Imwer level of the machine by a factor of 1000.

by changmg the charseteristics or,ithdrawing an ab<prbe go "55' nable mattr:ai estead ef w a redector. Our &-cum..n F.>r mach.mes with shorter neutron generation times, will assume that MI controls are of an absorbice type.

THE SAFETY OF NUCLEAR REACTORS 83 part ularly danuerous because of the very short Since the negative reactivity coefdeient+ can lead c-iolding time that one can achieve with modest ex-to shutting od the nuclear accident without destruc-ces react;rity. Ilowever, part of this danger is spuri-tive edects, a few words about these coci6cients may ou3 hecause the reactor will become supercritical be desirable. First ni all lar;:e negative reacavity enou::h to rim through a complete rimaway accident coci6cients are clearly wanted. However, these co-before very much excess reactivity can be added by ef6cients must he quick geting, able to take etiect.nd ordinary methods of operation of controls. Only very shut down the reactor during the transient conditions sudden motinn of the control rods, motion so rapid of a runaway. Primary changes in temperature are that it u ould have to be induced by special pneumatic caused by the generation of Rssion heat in the fuel systems could Icad to a rapM, explosive type of acci.

elements.1;Ieating of the fuel elements may change dent with the>e fast reactors. For this reason. a care.

the reactivity or the machme negatively it the tuel fut study of the possible rate of reactivity increase, elements contain large quantities of US. This nega-rather than the total potential excess reactivity avail, tive change is due to the increased absorption of able. shouhl be carried out when the nuclear runaway res nance energy neutrons by Doppler broadenmg nr the Uns absorption resonances. For large lumped problem is considered.

thermal reactors. this edect amotmts to about 10-5 It seems that the prevention of nuclear runaway fracuan et reactivity per degree C temperature rise.

accidents is very closely associated with the problem

. A see ndary reason tor the temperature coef6cient of excess reacticitv and the rate at which excess

'5 ** heatmg or the moderator. In many reactors.'

reactivity might he'added to a given machine. This in turn depends on the technical details of any given

@'.S ".211 benencially reduce reactivity. However, the time tor heat to tiow from the hot fuel elements to machite. both in its neutronic behavior and in the the moderated portion may be sumciently long, sev-operation of control devices and possible other ways eral seconds to a mmute m somerises. so that a1-of changing the reactivity, perhaps because of the presence of experimental irradiation facilities. This though the moderated temperature coemaent is tav rable, it does not have time to come into play is not a problem that can be generally solved for all machines, but each nachine must be studied on its durmg a nuclear runaway. For example, the thennal diffusion time across a four-mch thickness or graph-own merits.

We will now turn to the characteristics of a nuclear ite-moderator m a large lumped thermal machme is runaway, assuming that it is actually underway. As nearly a minute. This ume is so long that the coem-pointed out above a nuclear runaway is not particu-cient ass ci ted with moderator heatmg plays no part larlv violent but it does take place in a remarkablv during the runaway.

3hort time. The runaway will proceed according ta

.One ot the isnportant techm, cal areas associated the following steps. Firs't of all, excess reactivity is w th understandm, g nuclear rtmaway behavior at a inserted, the reactor thcn rises exponentially in power

.ct r is that of heat transfer under transient con-level with a neariv constant e-folding time until ditions. Relatively little knowledge m this area has enough energy is acetuuulated in the structure to been available beca,use most heat transfer studies are a:Tect the behavior cf neutrons. These early eEects c nducted under steady-state conditions.

are characterized by the term temperature coef6cients It appears likely that a nuclear runaway will cause

,.y of the reactivity. These may be either positive, that is

'". ugh disruption of the reactor structure so that making the reactor more reactive, or negative, mak-6ssion products wd,1 start to leak out of the reactor ing the reactor ! css reactive and tending to shut it nto the surrounding area. How fast th,s escape of i

down. If an increa>c in power level tends to make the 6ssion products wdl be depends upon the type of reactor more reactive and increase the power level reactor and type of reactor accident. It may be possi-not ontv further but make the further increase more ble to show that there will be very little mechanical rapid. one sometimes savs that this is an autocata-

" k'. ice utside the reactor shield so that if the Ittic reactor. Such a rea'ctor appears to be particu. buildmg which houses tite reactor can be made gas-tight. then escape or fiss,on products to areas outside U

l'arte dangerou<, and can possibly achieve really short i

c-folding times. A few strongly autocatalytic reactors the reactor building can be greatly reduced. Wemislt j

to enmuhnsisenthastiramass easectIFUEildm"ME I

are known. For most reactors, negative reactivity co-e[..

f 4-

=-

" N##

1. y8'N'N ef6cients will take eficct and lead to a lengthening

1. q.

in the c-fohling time.

DELAYED ENERGY PRODUCTION

~

Finally, a third phase of the runaway will occur Nuclear reactous>hamiumothetaiM,.,4 uulo.org 1

when enough of the reactor structure is actually bleclima*Astic. Because of the accumulated fission melted, vaporized or other vise adected (in most products, and the accompanying exothermic radio-j ca3e> reactivity coefdeients will not be adequate to active transformation of these 6ssion products. a

,9 l nt the reactor down without destructive eEects, nucsearsellesse.wilirconemogItzr orad =-@g,q '

i jalthough for some reactors this wih mdeed be the whesthv-*-WNrenmowiAahur dom The ca3e t, and these destructive edects will shut down energy produced by the 6ssion products has I;cen the nuclear chain reaction and stop the runaway.

studied, and the result for power prorhterion,ighg m.

p.

3 1

i l

84 VOL. Xill P/853 USA C. R. McCULLOUGH et al.

reactor which has been operating for a long time may

~

':~+*-'s, he summarized in the following equation:

v:" g*T y'w" P%a = 0.07 P,,,,,,o [t( sec y ]-", t > 1 sec io.cc:

/

IIere P.-,,o is the normal operating power level of

'/

the reactor, I6 is the delayed heat power level of LO::

the reactor in the same units as the normal power level, t is the time in seconds. and the 0.07 is an ex-ioc f

/

perimentally determined coefticient. Although the eamam a.cc u rission products individually decay exponentially, the 5

result of their statistical production is to make this delayed heat decay with the relatively weak power t

law indicated. For about one second after the reactor e'

3,,,, c,,,,,

i mee.nnes icco n.

is shut down the delayed power levelis approximately i

, 7"c of the normal gglgel.

C' l _iThrdM"E_eedise'l$iisifaTsETo*uftM ~

liGe m ;ure m1he com.;rsystenfshdW*h,'A c ei breakdown qi pumpt loss of pumping power, me-chanical failure of cooling piping, then even if the i

c oa, nuclear chain reaction is immediately shut down by

'c' io' ic*

ic' io'

"' ^"" s no t s

, we inserting control or safety rods there will still be left a substantial heat load which must somehow be Nv'. 2. coic ior.d t. p. rotor. ris. ofrer shutdown or th. choi.

r. action. { Heat capacity ossumptions are indicated.) (Court.sy of disposed ot.

w. H. zinn, Argonn. National taborotory)

For example, if the fuel :lements from the Mate-1 rials Testing Reactor wer7 suddenly removed from gegarrangemeesta$cBthreHer%Iirclk'nWghtbe the reactor and lett standmg m the open air, they rathw - H atabfin-6giwitt equipment. The j

would melt down by themselves by delayed heat pro-agie. righters' would then approach the reactor and.

J duction. If they were suddenly immersed in water, make suitable connections and force through emer-probably this melting would not take place.

gency cooling. A third possibility a to have stand (v One may consider the dehved heat ~W~n tr "

force @yecuerkcoolinggsimilar' to the main cooling

~

the comt of view or suddenly stopping forced conhn?

system but connected to a special power suoplv and m the reactor ano am suumo.wonmctb +""

with special separate piping.

^ '

_re

-1 ne feel elements, and otner materials in It is clear that a:delayedsascadentMit,.could;ng the reactor in close thermal contact with the fuel be.k@Mensmismighhverwwelldead to elements will then start to increase in temperature.

sufficient 7 disrupoond the-reactancore.to-allowr>6e-The rate of temperature rise will be proportional t sion,prphra a=mpe S t is a:so clear that this again I

the precedmg steady power level or the machme, and will not be of itself. very violent event, and agas 2 the rate of temperature rise will be reduced it there as in the case of the nuclear runaway it is probable '

is a large heat capacity in intimate thermal contact that an accident of this kind can be minimized a good with the fuel elements. In tact, smce the rule of du deal by providing a gas-tight building around the Long and Petit mdicates that the heat capacity or reactor.

solid materials is proportional to the number of atoms they contain, a crude rule of thumb would state that CHEMICAL REACTIONS l

the rate of temperature rise is proportional to the Either a nuclear runaway or a delaved heat accident l

power of the reactor per atom of material in good may cause considerable meltin'g an'd mixing of re-thermal contact with the fuel elements. The simple actor components and lead to exothermic chemical expre<sion we have given for the delayed power indi-reactions between these components. A simple exam-cates that the time rate of temperature rise should be ple of this type is that of an air-cooled graphite re-proportional to the 0.8 power of the time. In Fig. 2 actor. A sudden temperature rise in the uranium are curves showing the rate of temperature rise fol-fuel may be suf6cient to ca'ise it to melt and heat the lowing uncooled shutdown from nonnal operation for adjacent graphite so that both the uranium and a few reactors? One concludes that this temperature graphite can burn in the cooling air. If the air sup-rise, although not so rapid as to constitute a sudden ply is not turned off, it is likely that a substantial event in terms of human reaction times, is neverthe-portion of the reactor could be consumed in this wav.

le33 rapid enough to be quite troublesome. In Table II This would then disperse radioactive fission produc'ts are summarized some delayed heat power levels.

into the surrounding area through the exhaust portion l

desiggs are suggested.

of the cooling system.

Th

! One is to have a standbEemmglinggrsteg Another example is that of a heavy-watenmoder-which works either by gravity tiow of coolant or by ated-and-cooled natural uranium reactor. In a ma-I natural convection. Another is to have staudhgaemeg.

chine of this sort a runaway accident cou:d melt the

[\\f \\

M @M N b l

,n

THE SAFETY OF NUCLEAR REACTORS 85 uranium and allow it to mix intimately with the be taken that a failure cannot put the controls out of water. In this ca-c the thermodynamic potential indi-operation.

cates that an exothermic chemical reaction can take In order to prevent a delayed heat accident. it is place. Whether or not such a reaction would be rapid important that enough natural convection heat trans-and vinient i, m,t clearly known. In this case one has ier can take place in the overheated core to dispose to deal with the chemical kinetics of a heterogene-of the delayed beat. perhaps just into the ground.

ously reacting chemical sy3 tem (among other things Even if the structure is damaged, one must try to the probable degree of di3p-rsion of the uranium into keep the temperature lower than that temperature the water i not known ). Presumably, the rate will which would start a substantial pressure rise in the depend upon the intrin3ic molecular kinetic process reactor structure. In that case. the dssion products in the conventional chemical sense, but it will also may be kept inside the reactor shield. This means depend upon the degree of dispersion of the uranium that some coolant contained in the core should havet into the water, the rate at which reacting molecules a large surface to which it can ransfer heat by natu-can ditiuse throu:;h the uranium oxide layer that ral convection or byAoiling convection, and that this would be formed between the uranium and the water, degree of cooling should be suf6cient to keep the bulk and the degree of turbulent mixing and scrubbing of of all volatile materials below their boiling points. It the two reactants against each other. This latter ei-may be remarked that boiling heat transfer is known iect might be generated by the reaction itself. This is to be especially ef5cient so that in any event there clearly a complex problem and a great deal more will tend to be a ceiling put on the temperature rise needs to be learned. However, one can say this: If at about the boiling temperature of the original cool-an exothermic reaction of this type goes to comple-ant employed. This in turn implies that by appro-tion, the resulting energy release will nearly always priate construction one may limit the pressure inside be substantially greater than the energy generated in the shield to a few atmospheres. Thus it may be a preceding nuclear runaway. Thus it is important to rather easy to make sure that the fission products are determine the possible chemical reactions. A substan-kept inside the shield.

tial increase in reactor safety can be achieved by the rinally, the problem of chemical reaction among elimination of possible reacting components in the reactor components can often be minimized. For ex-reactor structure.

ample, already-reacted components might be used in some cases. Cranium oxide rather than uranium SAFE DESIGNS metal in an air-or water-cooled reactor may serve it seems worth while to summarize the preceding as an example.

discussion with a few remarks concerning the ap-The other general conclusion that tha Safecuard_ f proach to >aie reactor designs. First ot all, it is desir-Committee has come to is that explosive hazard in i able to provide a large tyegative reactiv ty coemcient.

reactor accidents is minor, at least for people not at This can,ually be aducted by thermarcouphng of the reactor site. Indeed, for many reactors, it appears the n elements to those portions of the reactor unlikely that ther,e will be much' mechanical violence uhv give a substantial reduction to the neutron external to the reactor shield. For this reason, a gas-multiphcanon when heated. For example, m the case tight building, or a moderately gas-tight buildin~

ot ennched, water-moderated reactors, close thermal which mav con 6ne the nision products during a cool-contact between the fuel elements and the moderated ing perio[1 and from which the 6ssion products are water can iced to enough heating and vaporization exhausted into scrubbers and out a high stack, may or the water to reduce the water density m the event serve to prevent the spread of dssion products fol'-

or a nuclear runaway, and > hut down the reactor lowing a reactor accident. For some reactors the before serious damage is done. y,uccessful tests of con 6ning building will have to be a gas-tight pres-

1. j tin,- > ort have been made.1.\\ design or this sort must sure vessel. Safe-desi:m procedures represent an im-S " ),

i.

be thonght through carefully m order to make sure portant field of nudear reactor development.

~

that enough heat transier surtace is avail,ble and

/

ral.id enou"h heat flow will take place to shut down a gammuw&EnhEqt machm, e during an accident.

The control system must be carefully designed so Although good administrative control of the re-that in the event of too high a power level, too high a actor does not lead to the same degree of conddence rate of rise of power level, a serious reduction in in the good behavior of the tachine that intrinsic coolant dow, or any major failure of fuel elements l gremlin-f ree built-in stability does. nevertheless, the reactor wCl shut down in a time interval small good administrative control does enhance the safety enough to minimize damage. All potentially danger-and reliability of reactor operation. Indeed, goo &a6 ous failurc< should be monitored by instruments and mimayaabeibdentreLinnand--f_W thbeg6pcGi control channels leading to shut-down or " scram.

wkht =_w% From the h

These monitors and channels shouhl be at least in point of view of public hazard ca6-dupliente, independent of each other, and preferably tiogL -

="smahaptvesghauiis&W of dirTerent types. In addition particular care should tha

86 VOL. Xill P/853 USA C. R. McCULLOUGH st of.

llowever, thjd r-. _ ~r Ma&

radioactive and should not be allowed to accumulate unduly, or to be lost or to be handled in an irrespon-

--- a pp,

..t._ e %n, _ w Throughout the design and construction of the re-sible manner.

actor. thorou:,h supervision, careful design for relia-CONSEQUENCES OF AN ACCIDENT nihty and thorough testing of all reactor components

including coolant system, od-gas systems, chims, We believe,the following discussion outlines the

, e ?,

f, main features which can make a nuclear incident

. saictv and control mechanisms, and all control and to e.

'/'

! operating instrumentation should be carried through.

dangerous. In the event of a reactor accident, there "i

!.\\ll these comnonents should be given svstematic and will probably result a release of radiotetive material from the reactor. Operating personnel may be seri-

{ thorough shak'edown te3 ting before the' reactor is put ously injured or perhaps even killed. The reactor i into operation and before it becomes radioactive so itself may be damaged beyond repair or recovery.

i that moditication, correction, and maintenance can be The reactor building and its associated equipment done with less dinicultv. Indeed, it is extremelv disi.

are very likely to be heavily contaminated and indeed, cult to empl.asize how" important it is to have com, II may n t be possible to clean up the buikhng sum-plete, thorough. systema *ic shakedown of all portions ciently to put it into operation again. Design of the of reactor control and i tstrumentation.

buildmg so that possible cleanup operations are as In the design of the reactor, careful attention easy as possible,s desirable." Smooth, clean sur: aces

);

i should be given to the problem of maintenance after Perhaps clad m stamless steel, would make cleanup i

it is placed in operation. It should be possible to enter Perations easier. Fmally, radioactive materials can all instrument areas, most of the control areas. and escape from the reactor site altogether. Fission prod-obviously the central control room, after the machine ucts may be carried m the wmd and spread over has started up and been operating for some time.

adjacent" populated areas and constitute an acute

, Fuel-element failure, a contmumg problem. may hazard. Radioactive material may escape into the allow radioactivity to enter portions of the reactor ground and be carried by the percolating ground structure which normally would be expected to be water to adjacent rivers or other water supplies.

radiation-free. Th,s should be taken into account in Although a great deal needs to be known about the i

the origmal design.

character of radioactive material that might escape Once the reactor is placed into operation, contmu.

from a reactor, whether it is in large or small particles, mg close supervision is essential. Maintenance pr -

whether it is indeed gaseous, whether it would rise cedures should be carefully followed and mamtenance high into the air or seep slowly along and into the checks should be scheduled in an appropriate way.

ground, nevertheless, some notion of the possible The period of reactor startup is a particularly critic tl spread of the hazard could be obtained by study of one, and should be followed very closely. Reactor the meteorology, and hydrology at the reactor loading and unloading are delicate operations, par-site.m28 " It is desirable, for exitmple, to have the ticularly the unloading of now-radioactive fuel ele-prevailing wind to blow from the reactor to uninhab-ments. Startup of a reloaded reactor must be carefully ited are:rs. It is also desirable to have the reactor site considered since the reactivity may have been affected not be located on a main watershed. From the point by a new fuel loading. The normal, or routine, day-of view of the hazard alone, it is of course desirable to-day operation requires close super ision so that to have the reactor site far from populous or vital troubles may be detected at an early date and cor-industrial areas. It will not always be possible to rective measures taken. Clearly, careless operation of obtain this remote location and still obtain economic the controls may lead to a supercriticality accident utility from the reactor. For this reason, the Safe-a and the manipulation of the controls should be carried guard Committee is continuing to emphasize the out only by people who are thoroughly familiar with importance of safe reactor designs, the development the characteristics of the reactor and its associated of contained fuses to minimize the possibility of a runaway accident. and the use of gas-tight containing equipment.

If a reactor is employed as an irradiation facility, vessels and builf gs.

it is po33ibic for experiments to give rise to sudden Perhaps it is important agaits to emphasi:e the change > of reactivity. Experiments should be pir med, degree of public hazard that might follow a reactor l the plan reviewed by the administrative staff, and accident. Assuming that good luck prevails and no suitable emergency procedures decided upon, before one is killed, it may nevertheless be necessary to f in. erring experiments into the reactor.

evacuate a large city, to abandon a major watershed, '

Finally, there is one phase of the administratively and very probably it would be necesary to make the controlled reactor which is usually taken for granted reactor site itself a forbidden area for some years to but may require a word er 3o: The careiut accounting come.

ior all materials which have been irradiated in the Despite all these possib!e dire consequences. it is machine. There is usually available a number of test the helici of the Advisory Committee on Reactor hvies in which experimental irradiations may be car-Safeguards that nuclear reactors will scon st:rt to ried out. The samples so irradiated can be highly produce substantially increasing n aterial benedts for 4

.1

3

1 THE SAFETY OF NUCLEAR REACTORS 87 humanity. We believe that useful electric power in REFERENCES

!arge quantities can be generated by nuclear reactors.

1. Teller E., testimeny. Atomic Power Dewbrment and

.It is our concern that rapid progress shall be made Private Entertrise, Hearings before the Joint Commit-but that enou:;h caution be observed so that no cata,.

tee on Atornie Energy,83rd Congress.1:t Session. June-July 1953 (Gou. Prindng OEce, Wanngton, D. C.).

trophic event will t'elay the fruition of reactor devel-

2. Weil, G. L., Ha:crds cf Nucitar Power Plants, Science epmen.

121, No. 3140, 315 (1953).

APPENDIX. THE ADVISORY COMMITTEE ON REACTOR 3.11 organ, K. Z. and Ford, IL R., Developments in In-urnd Don Drunninans, Nucle mes 12, Na 6, 26 SAFEGUARDS TO THE UNITED STATES ATOMIC (1954).

ENERGY COMMISSION

4. Teller, E., Reactor Hazards Predictable, Nucleonics

-The Advisory Committee on Reactor Safeguards 11, No. II, s0 (1953).

was formed by combining the Reactor Safeguard

5. McCullough, C. R., General Criteria for Saft Reactor [

Con mittee and the Industrial Committee on Reactor Dign and 0;cration, ASME Annual Meeting, Nov.

S-Dec. 3,1954.

Location Problems. At this time members are: M.

6. Mas, M. M, Hazards of l.ow Powu Renanl: Rcac.

Eenedict, Massachusetts Institute of Technolo27; fors, Proc. 3rd Annual Oak Ridge Summer Symposium, II. Brooks. Harvard L. iversity; %,. P. Conner, Jr.,

TID 5031 (Technical Information Services, Oak Ridge.

n l

Hercules Powder Company; R. L. Doan, Phillips Tenn.,1951).

Petroleum Company; H. Friedell, Western Reserve

7. Hurwitz, H., Jr., Safeguard Considerations for Nuclear [

University: I. B. Johns, Monsanto Chemical Com-Power Phnts, Nucleonies 11, No. II, SO (195?).

pany: C. R. McCullough, Chairman, ACRS; M. M.

8. Way, K. and Wigner, E. P., Radiation from Pission Mills University of California Radiatiod Laboratory; Products, Phys. Rev. TO 115 (1946),

K. R. U-1,orn. Allied Chemical and Dye Corporation;

9. Zinn, W. H., Private communication.

D..\\. Rogers Allied Chemical and Dye Corporation;

10. Dietrich, J., E.rterimental Determination of the Self C. R. Russell, Secretary, ACRS; R. C. Stratton, Regulation, and Safety of Water Reactors, P/481 Vol.

Travelers Insurance Companv: E. Teller, Depart-13, Internati aa! conference on the Peaceful Uses of At nue Enugy, August M,19:s.

ment of Physics, University o'f California; H. Wex-

11. Gilbest, F. W., Decontaminati:n of the Canadian Reac-ler, United States Weather Bureau; and A. Wolman, tor, Chem. Eng. Progr.,0, 267, (1954).

The Johns Hopkm.s L..mversity.

12. Fhzgerald. J. J, Haitz, H, Jr, and Tonics, I,

The Reactor Safe;uard Committee was formed in

.1fethod for Evaluatsng Radiatson Ha:ards from a Nu-1947, and the Inc strial Committee on Reactor Loca-cz,ar incid,nt, KAPL-IN5 (Knolls Atomic Power tion Problems wcs formed in 1949. The following Laboratory, Schenectady, N, Y '1954). Available from were also associatrd with these committees for pro.

the Office of Technical Services, U. S. Dept. of Com-merce, Washington, D. C.

longed periods: Cmdr. J. Dunford, US Atomic Energy Commission; Col. B. Holzman, US Air

13. Fitzgerald, J. J, Hurwitz, H., Jr. and Tonks. L Force; J. Kennedy, Washington University, St.

1. 'Q,n#[#'"#I#",, I,"7s(1 54j g9 y

I.ouis : F. Seitz, University of Illinois; G. Weil,

14. Mester, R. B., and widdoes, L. C., Evaluating Reactor tormerly Divuion or Reactor Development, US Hazards from Airborne Pission Products, Nucleonics AEC; and J. A. Wheeler, Princeton University.

12, No. 9, 39 (1954).

i,

^

.D

4 8

1. s 1

~ -,7 -, ~

,,. a..

~

a

~.

i, b

i.

Proceedings of the International Conference 4

on the Peaceful Uses of Atomic Energy Held in Geneva 8 August-20 August 1955 Volume 13 Legal, Administrative, Health and Safety Aspects of Large-Scale Use of Nuclear Energy l

-/N-UNITED NATIONS New York 1956 I

Y-12 TECHY'"iL LIBRARY Y-140.3 3 ganwgr y[ghfd h{[dN$id(n" w

m

..