ML20041A775

Reactor Safety Study:Assessment of Accident Risks in Us Commercial Nuclear Power Plants, Executive Summary

ML20041A775
Person / Time
Site:	Pilgrim
Issue date:	10/31/1975
From:	US ATOMIC ENERGY COMMISSION (AEC)
To:
Shared Package
ML20041A773	List: ML20041A772 ML20041A775
References
CON-#487-5067, FRN-46FR61132, RTR-NUREG-75-014, RTR-NUREG-75-14, RTR-WASH-1400, RULE-PR-50 2.206, WASH-1400, NUDOCS 8202220493
Download: ML20041A775 (12)

Text

w.

W:

WASH.1400 t,!0 if (NUREG 75/014)

~

e 2;

04~

Lh-

.:e.

>at

.a.

ty-g.;.

~..

g#

fn.

?!.;

ji;;;

=

,n,,.

c In-M REACTOR SAFEIT STUDY.-

e

^

w:

1:2

AN ASSESSMENT OF ~

p

.jhi.'

~

ACODENT RISES IN-U.S. COMMERCIAI!

tn.

4

• L n ;

e NUG. EAR POWER ' PLANTS 2 l.} gy i

7u

_q:

)..:

y:.

t 3

.g; };,

u -

g-

..._e

• 1 *.

%c

.i

.4 i :

M,y.. '

EXECUTIVE

SUMMARY

.ta s

g

, a s,.

y 2

b.

t.

1 l

V

.r V

V v

l h

U.S. NUG. EAR REGULATORY COMMISSION hlE.:

OCTOBER 1975 L

y p

-j

1 K x

4 I'e p

7 F.+

7 f5 9

t'.~;

3J,

8202220493 820110 PDR PR 50 46FRb1132 PDR m

s:

Secti:n 1 Introduction and Results

..s 1

Thel Reactor Safety. Study-wassponsorej b.

The likelihood of reactor' accidents by the U. S.' A**"mi e. Energy Commission

'is much smaller..than that of many

~

to ; estimate.ths.publie.. risks that couldin?

non-nuclear accidents having'r'imilar s

consequences. All non-nuclea acci-be involvedcin potential accidents commercialv nuclear power plants of the -

dents examined in this study, in--

type now in use.

It was performed unders cluding fires, explosions. toxic, ther independent-direction of Professor" chsmical release.B,- dam--failures,-

Norman C. Rasmussen of the Massachusetts!

airplane crashes, earthquakes, hur -

Institute of Technology. The risks had{

. ricanes and tornadoes, are much more-to be estinted,. rather. than measured, likely to occur and can have conse-:

because although there are about 50 such quences ccuparable to, or larger?

plants now operating, there have been no than, those of nuclear accidents.

nuclear accidents to date resulting in significant releases of radioactivity in Figures 'l-If 1-2 Rand' 'l-T ' compare" the U.S.

connercial nucleare power plants, nuclear reactor accident ~ risks predicted Many of the methods' used, to develop for the 100 plants expected to be oper-these.. estimates are based on those that ating. by. about 1980 with risks from were developed by -the Department 'o f

. other man-caused and natural events - to Defense and the National Aeronautics and which society is generally.already Space Administration in the last 10 exposed. The following..information is yean'and arewng-into-increasing -use contained-in-the-figuren

-n in recent' years.

~

w a.

Figures 1-1 and 2 show the likeli-m-

hood and. number of., fatalities from

$ o'biective'of Nhe study was to makem.a both nuclear and a variety of non-realistic. estimte of.thesa-risks.and' nuclear accidentsT** These figures to provide perspective, to compare them indicate that non-nuclear events are l wit O non,npelear risks'"to,which our society and.its individua1s are already "- - -

- about - 10,00CL times more likely'es

~

to produce large numbers of fataliti exposed.

This information - may be of

, than nuclear.' plants l',

help. in determining the future reliance by. society on nuclear. power as. a source of electricity.

b.

Figur'e 1-3 shows Idhe likelihood and The -results from this study suggest that dollar value of property damage as-the risks.to the public-from potential,.

sociated with nuclear and non-nucle-ar accidents.

Nuclear plants are accidents.in nuclear power plants are comparatively small. This is. based. on about'1000 times less likely to, cause comparable large dollar value the following considerations:

accidents than other sources. Prop-a.

The possible consequences of poten-

~~

tial reactor accidents are predicted

. ~'

(r z.-- to be no. larger,. and in. many cases

, much -smaller, than. those of non-

-.. 1... -.

n. nuclear accidents. The consequences The fatalities shown in Figs. 1-1 and are predicted to be smaller than 1-2 for the 100 nuclear plants are

. people have been led to believe by-those that would be predicted to occur previous studies whi~ch deliberately within a short period of time after the maximized estimates of,these conse-potential reactor accident. This was 9.

n quences.-

s.., 2

.n.

done to provide a. consistent comparison al-

.,.; u to the non-nuclear events which also

-x1 a-cause fatalities in the same time

..; m ;

nn.-

frame.. As in potential nuclear acci-dents, there also. exist possibilities

{*;L-

;.- a 2

.m

.e.

1. "-

rne work, originally. sponsored

;by, the for injuries and.. longer. term health c.U.S.- Atomic Energy Connission,. was. com-effects from-non-nuclear accidents.

pleted under.the-sponsorship of.the Data.nor predictions of_this type are U.S.

Nuclear. Regulatory..Corsaission, not available for noq-nuclear, events 1 'which came into being_ oncJanuary <19,

.and so comparisons.cannot; easily be o

made 7 '~ - ' 35_ i T.1.3 a

..;c._

c

~

s 1915.-

. 3.n 3...c U

Xn4.

.p C

a-

!S h*$.

$- f f h f f* $.. 'f f.hfi QQflffh

- - - =

--a w e

liCkD 10-10 a

i g

1 1

1 I

l l

l l

1 I

M ikU 8r's blif. flOkl'df h 8p-i_

_cg# _ _ l

- _l_ -_ _._

, i i

7 3

• g

(

l

\\

l l

-l'

_1 _ _ d_ _._ [ __ _

\\

_\\ b a '"

,no.

_ i_

in0_

I f - ~

. \\a ; <-

.g g

i

--.m c

3 rnm____s r-- -+ - --

E ' ' '"m----P I-- --

r.

~

2 :-* -

...l.

h

c. L I -

kli

.k.t-j'.

I s;- --., - g.

. inm0-__

2:

_ j

- -j _ ---t-- _ -

,.l

• \\,

Ib i

[. anum-i+-- ]

1 l.

y

-..i I

l l

1

- _ 1-

. i

---

l

- - -]----

bn0.000-- - -- - -- y _ - - d - - - h -- -

\\

} in0.000-

't g-l

- \\ l g

j l

g l*

\\l i

1 l.

~

'l

_,-g _ --}- --

..,n ;,_

_ y[ ___q I

..t. _ _ _-

som -- - _ _,

100 Nect LSJ. *

.l Power Pie,itsj g

gg,

+

l... '"**"j

. l 1 - :-

l

.i

't

' -- A- - -.

in.0my___ -1 I _ _ _;i c.u _+-_ -

in.02.0m_

8..

.i I

-=-1

.I 4

t 4,

I i

t

-1 I

l

-i 4

. i

1 I

I I

1 p-l

.{

h l

1/10.000.03) l j..

g 14 0.000.000 10 100 1000 10.000 100.000 1.000.000 10 100 1000 10.000, e. 300.000 1.000.000

. Fatal"**

~ -

F au%.

y' FIGURE l-1 Frequencyo,f}atalitiesdue~to' Man-FIGURE S-2 Frequency of Fatalities due to Natural Caused Events t-Events.

L.

>:.e.n n - -: :

r.t; c'-

Not.s.. 1.' Freilt.i 6.e t..we..c' e.ats ar. not.ho=*. b,cou e.ts er.

1..f a n tw.s.no.i men.e.anarr.ru th. imsse

. ;.. - an i a

_ ca mooiunac.ws

" e iss

, e w,,n e a.a.ii.,,,a, du..i erob o

.,nm.i.o =

m=====8

*=

• 1rn id i 5~i= asaire==

a... i 2. as.a

. - i, s m.

.i.,i.e.a e n.

e.sr.o.at.6 ery fortart ef f /4 sie a en _.

.ayweve.e pie e

'A

, i ur.wa w n7 e, *cieri e m.ae s oa ir.a.eca p., u, ;, a asw n== wi m => i.e m

-=

.mme i. m,

.--.* t

'w= > e 1 '*.ad 8.a a===== ~eaw 3

S.,

...~a m.-

e nc

-ar, e,.

- ^ - "_

• ie tv haa

• 15.a8 6.a r.*.th s

m.i., e i.,

< e __ - :

,mos... nmo i. i.

4 m

m

.e av.cr e irm.ae s. sme,

,m.i e

-t

u i:

.c?. <.

the latest year for7 which these data

. yf e..

c.

were.' tabulated when.this

-s tudy was

's

[? erty'damagl isIassociAted with three perfomed e mThe u t predicted 4: nuclear accident risks.are very small compared

'. - -effects:

i t.

s-uic..:r i to other possible icauses.sof. : fatal

.r-2 cr.r.:

t. :

Lnjuries.

1.

the cost of relocating people 0

N -

M' F'--

In addition'to> fatalities -and property away from contaminated areas,- - -

damage, a number of other thealth ef fects 2.. the decontamination of land to could l>e caused by nuclear accidents.

to These. include injuries and long-term

~

avoid overexposing - people -

health' effects'such as cancers, ganetic W

radioactivity..

effects,-and thyroid gland illness. The

~

L i"N 3; ~;the cost of ensuring that people early illness' expected in potential ac-

'~

-- are' not exposed -to

-potential cidents would be about 10 times as large sources of radioactivity in food as the fatalitius shown in Figs.'l-1 and and water supplies.'1 '

l-2; for comparison there are 8 million P

injuries caused annually by other acci-

~

+-

• 8 Iri addition to-the overall risk informa-dents.

The number of cases of genetic tion in Figs. 1-1 'through 21-3, it is effects and long-term cancer fatalities tiseful to consider the risk 'to individu-is predicted - r to be. smaller than the als of-being fatally-injured by various normal ~ incidence rate of these diseases, types of accidents. The bulk of the in-Even for a large accident, the small in-formation shown in Table 1-1.is taken creases:in these diseases vould be dif-from the 1973 ' Statistical %bstracts of ficult'-to detect fron the normal inci-the U.S. and applies to the year 1969, dence rate. - _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _.

1 RGURE 13 Frequency 02 Property Damage due to Natural and Man-caused Events l

I g

g I

i l

,,,,# 1 '1/10

.. 1

> =. 1 *..-.,.ea_

.ene e w n,ks,,~*

__1---

,,,a.,m

., b yey m aves em c.,,,

e I

esaw 615 benea anaer seen vos.

l I

~

2

-.w....,.

pennewee w toevers o 14 ene 2 en 2 perimse. eas l

Cauand l

l r

w escion et t ab FWI I er FUheOOM46 I

$"*"t' I

/100 3 Se weww ee a sr com.e occure.ae.e ww weer.c4 'e= y.e -

l l

l ee,e, a

8r.

.e a==

e

..g

}g L

w i.ce-..e 1m

s. s j

g

\\

weerta-vv.

iiiOOO

_ _ __ __ _ l _ _ h._ _.

I I

1 1/10,000 100 rewooer I

g I

Po,

d l

l l

P6ents

.___f.__

if100.000 i

i i

l I

i 1/1,000,000 l~-

I l

l l

I I

I i

I l

l f

I 1/10,000,000 10 10 10 108 10'O 1011 6

7 8

Property Damage (Dolirs)

TABLE l-1 AVERAGE RISK OF. FATALITY BY VARIOUS,CAUSES

" Individual Chance Accident Type Total Number per Year Motor Vehicle 55,791

- l in 4,000 Falls 17,827 1 in 10,000 Fires and Hot Substances 7,451

.1 in 25,000 Drowning 6,181

, 1 in 30,000 Firearms 2,309 1 in 100,000 Air Travel 1,778 1 in 100,000 Falling Objects 1,271 1 in 160,000 t

Electrocution 1,148 1 in 160,000

^

Lightning 160 1 in 2,000,000 Tornadoes 91 1 in 2,500,000 Hurricanes 93 1 in 2,500,000 All Accidents 111,992 1 in 1,600 Nuclear Reactor Accidents (100 plants) 1 in 5,000.000,000 e

l i

"'hyroid illnesses that might result from

. observed during a period of 10 to 40 a large accident are mainly the forma-years following the accident.

tion of nodules on the thyroid glands these can be treated by medical proce-While the study has presented the esti-dures and rarely lead to serious conse-mated risks from nuclear pcwer plant quences. For most accidents, the number accidents and compared them with other of nodules caused would be small com-risks that exist in our society, it has pared to their normal incidence rate.

made no judgment on the acceptability of The number that might be produced in nuclear risks. The judgment as to what

.i very unlikely accidents would be about level of risk is acceptable should be equal to their normal occurrence in the made by a broader segment of society j

exposed population.

These would be than that involved in this s'

'y.

! l M

hgg [".

. ;. *.w.,.-

- - * : a, _

T ': : +

',-j % w. - %

e. y n.

• • ::* :. _ : M - t '. ! g e,=,

y ss c:

.z w e

.,.g_ w. g. a r.

%-- e

c

_ Quest,ons and Answen About the Study.= ti.whr 1.~.

i.,1 -

= :..a w,w. n,.e

,n m e

. g....

..u.

a-

.-_7x e --.

u-

.:r..

c.-w _

• w, g

,7 m

.r - D.

.t.,,..

-t

a This section of -the summary presents inhich is 'needed both to cool the fuel more information about the details - of and C to raaintain

thei 21ssion schain' the study than was covered in the intro-reaction'.

"W.

r u.sr duction.

It is presented in -question and answer format for ease of reference.

The heat released in the' uranium by the fission process '.' heats 6the water and 2.1 WHO DID THIS STUDY AND HOW MUCH-forms steam; the steam turns a turbine EFFORT WAS. INVOLVED?

to :. generate electricity. - Similarly,-

coal and-oil plants gener&te electricity -

The study was done principally at the using fossil fuel'to boil water.

Atomic Energy Commission headquarters by t-

. g-n, a group of scientists and engineers. who Today's nuclear power plants are very had the skills needed to carry out the large.

A typical plant.has an electri-study's tasks. They came from a variety cal capacity of 1,000,000 kilowatts, or-of organizations, including the AEC, the 1,000 megawatts.

This is enough elec-national laboratories, private laborato-tricity for -a' city of about five hundred-ries, and universities..About 10 people thousand people.=

i-were AEC employees.- The Director of the c.

ir-

~;-" :r a

study was Professor Norman C. Rasmussen R 3 CAN A EUCLEAR POWER PLANT EXPLODE of the Departsnent of Nuclear Engineering LIKE AN ATCM BOMB 7 of the Massachusetts Institute of Tech-ic:

.a nology, who served as an AEC. consultant No~. pit is impossible.for nuclear power during the course of the study.: The pTents to explode'like a nuclear weapon.

Staff Director who had the - day-to-day The';1aws of physics do'not permit this responsibility for the project was because the fuel contfains only a small Mr. Saul Levine of the AEC.-

<The study fraction 5(3-51)acf 'the special. type of was started - in - the summer of 1972 and uraritum N{ called uran'iaan-235) that must took three years to complete. A. total gW :sr ded in weaponse:.4hMr.

t.-

u

':.a '.-

- I :.z s -. e--e,_ n of 60 people, r variouswconsultantse -70

. :~E%i- - N man-years of :.'f fort, and about four-mil r lion dollars were. involved.--

t : - --,...

'2. 4 HOW IS RISK DEFINED?

- W m.: m sg yn m,.

~.~.,

..y f-2.2 ^WHAT KIND OF NUCLEAR POWER PIANTS The idea of risk involves both the 1ike-l

'ARE COVERED BY TTHE STUDY?

7. - -

"11 hood -end.. consequences of an event.

"hus, -to estimate--ithe -risk. involved :..in l

4 riving.an automobile, one.would need to The study considered large power reac-tors. of the pressurized water and boil-know-the likelihood of van accident:.in ing water type being use'd in 'the vU.'s.

which, ; for naxample,- en individual could today.' " Reactors of the present" genera-be cl).cinjured or+2) killedc Thus: there C

tion are 'all water cooled, and therefore are etwo different consequences, -injury 2

the study " limited ~itself to this : type.'

for:f atality /- each with yits..own like11-Although-high temperature gas cooled and bood.*

For

-injury,. an individual's liquid metal-fast -breeder reactor -de-chance persyear;is about-one.in 130.and signs'are now under development,2 "reac-formf atality;. ittis about one -in 4000.

tors-of this -type are not expected to This type of data concerns the risk to have any significant role in U.S.

'elec-individuals and can affect attitudes and tric power production "in this decades habits that-individuals.have

-toward

]'-

driving. -

thus they were -not-conside, red.

,r, Nuclear power plants produce electricity However, from an overall societal. view-by the fissioning (or splitting) of Point,- different. types,-of data are of uranium atoms. "The nuclear-reactor. fuel.

interest. -Here, =1.5.million injuries in 'which the uranium -atoms fission is in per ryear ~and 55,000 f atalities per year a ?large steel-vessel. The reactor: fuel due to automobile -accidents represent consists of about 100 tons.- of uranium, the kind;of-infornation that might be of The uranium is*inside metal rods.about use-in making decisions on highway and 1/2 inch in diameter and about 12 _ feet autcecbile: safety.

long.

These rods are formed into' fuel bundles of about'50-200 rods each.

Each

-The same type of logic applies to reac-reactor contains several hundred bun-tors.

Froma the viewpoint of a person I

dies.

The vessel is filled with water, living in _ the. general vicinity - of a l

l

, l l

reacto r, tha likalihood of being killcdo,y,'.... tial radiccctiva releccGo might occur in any one year in a reactor accident is'"

" 'and has identified those that determine one chance in 5 billion, and the likeli-the risks.

This involved defining the hood of being injured in.any cnenyeac xid. m n. ways' in whichpthe fuel in the core could n

a reactor accident is one chance in mele and the ways in which systems to 75,000,000.

control the release of radioactivity could fail.

1

.1 m

r.

_ ;., a c

2. 52 WHAT-CAUSES THE RISKS ASSOCIATED ; s WITH NUCLEAR PONER PLANT ACCIDENTS?~

2.7. HOW MIGHT A CORE-MELT ACCIDENTra tu l

OCCUR?:.

n. :m

n:.

=

i The-risks.from nuclear power plaats:are:-

..w o

...u ;

...; 3

.3 l

due to the. radioactivity formed; by. thet It is significant that in some 200 fission: process.4 In, normal operationt reactor-years of e conunercial'. operation of; nuclear.. power. plants-4 release..-minuter reactors of the type considered: in.T the h

amounts of. this radioactlyity.undertcon 2 report there have been no f:.el melting trolled conditions,.-

In..thet.eventeco f_.

accidents. To melt-the fuel requires a--

3, n

highly unlikely accidents, larger failure in -the cooling:. system or.thes.

t amounts of radioactivity, couldc.be rg ;-

occurrence of a. heat. -imbalance. that:

?

leased

.and could -cause significant:

would allow-the fuel. to. heat.up to its.:

risks.. -..

^4

.i.

melting point, about 5,000*Fa. n -

c.; 1 a.

t.tw

..u :

a..

a::.: -..- The - fragments. ofc the urppium, atom. that.

u-

.-, n, :...,7a.

remain after it fissions. pre.a radioac -

Tol;those. unfamiliar with the-character--

tive.

These radioactive atoms are istics of. reactors,Jit, might. seemi thate j

called. fission;productswLThey.disinte..

alle,thatt is. required:i to prevent fuela 3

grate further withi thm release of from: overheating _ iJs' at system to promptly-

[1 nuclear radiations. Many of them decay stop,: or shut-down, I the. fission - proce ss;.

A away quickly,q in-a matter of, minutes 9E at-the:first sign of trouble.

Although:

hours,. to:non. radioactive, forms.u Othersg reactors..have.csuch systems, they alonau decay away: more slow,1y; : and; reLuizer ars not :enough since the. : radioactive.

},

months,.and;in.a:.f- 'casse, many yearsi decay of v. fission fragments in the fueIt h

tai decay. : Ther, f A.

tontproduc.ts accumuN cohtinues: to generate-heato.(called decay 4 lating-in the.. fuel. rode (include-; both, heat)? that-must : be rentoved even af ter.

y(

gases and solids.

Included..are iodined the-fission process. stops.n Thus, redun-gases like krypton and xenon, and solids dant decay: heat removal: systems are also:

t like cesium and strontium.

provided 'in : reactors.. : In -- addition,?

8, -

"*. MC E~c 7

' 09

4.. S emergency core. cooling.. systems (ECCS).

{

2.6 HOW CAN RADIOACTIVITY BE RELEASED?

are provided to cope with a series of

}

- sr sus-

.T potential but unlikely; accidents, caused; t

j The only way that potentially large by ruptures..in,. and loss

of. coolant amounts of radioactivity': could) be red from, the normal cooling system.

i4 leased is by melting tha -fuel in the

.3 e i

reactor core. The fue1 that is removed from a reactor after use and stored at The Reactor Safety Study has. defined tvo the plant site"als'o contains tcensiderw broad types of situations that might po.

• j able amounts of radioactivity > Bodever,-

tantially. lead to a melting of.the reac'.

4 accidental releases from such used:. fuer tor corea the loss-of,-coolant accidenf.

iIf were found - td d be quite unlikely and (IDCA), and transients.. In the event. of i

small compared to potential releases '-of a potential loss of coolant,.the normal radioactivity from the fuel in the reac4 cooling water would be. lost from the tor core.

c wr

n. T cooling systems and core melting would 1l3 l

l 1<

. s.m t e.

be prevented by the use of the emergency The safety design of reactors irtcludes a core cooling system (ECCS). However, i t lI series of systems to prevent thet over-melting could. occur in.a loss of coolant e

heating of fuel and to control potential if the ECCS were to fail to operate. ~ ~

releases of radioactivity from;theefuel;l Thus, for a pote%1al accidental release of radioactivit r to 'the environment <-to The term " transient" refers to. any one lJ occur, there mc st be a series.of sequen-of a number of conditions which could tial failures that would cause the:- fuel occur in a plant and would require, the 4

to overheat and ~ release-its radioactivi4 reacto r. to be shut. down.

Following i

ty.

There would aIso have to-be cfail-shutdown, the. decay. heat. removal systems r

fe ures in the systems designed"to-remove would operate. to= keep-.the core.fgom and contain the radioactivity.

overheating. -.Certain-failures in either i

-l rn

... :r.;.,. :<.a

acer MT the: shutdown.orrthe' decay heat removal t

hs' study has "eximined a~very 1&rge ntan' systems also have the potential to;cause

'u

[

ber'of potential paths-by 'whicht poten" melting of.the core.<

na

..e

. ;r

--.- - - _ m...,

,gm g

,.44.g

-g;,,,,jgg"'

2.8 WHAT FEAWRES ARE.PROVIDED IN

, accident. "Stich 5ccidents e.rs* 1ese like-REACTORS TO COPE WITH _A CORE ELT ACCIDENT?

ly.but.could release a lirger amount of airborne radioactivity athe- ~have; more seri s consequences. t?Thd Zeonsequences NuclearI power',piaMs have n'umerEus;sys g tems designed-to. prevent' core meltingP' N

• • • 1**s likely' accidents 9have"been processesandadditional',_ features.th'at-]

f Furthermore, there are inherent _ physical 1-1 th ug come into play to remove and contain the '

radioactivity released from ~the molten 2.9 HOW MIGHT THE LOSS-OF-COOLANT ACCI-fuel should core melting ' occur.' Al-DENT LEAD TO A CORE EI,T?

though there are features provided to keep the containment building from being damaged for some time after the core Loss of coolant accidents are postulated melts, the containment would ultimately to result from failures in the normal fail, causing a_ release of radioac, tivi-reactor cooling water system, and plants ty.

are designed to cope with such failures.

The water in the reactor cooling system's is at a very high pressure (be tween, J 0 An essentially leaktight containment to 100 times the pressure in a car tire) building is provided to.. prevent the and if a rupture were to, occur in the initial dispersion pf the airborne ra-pipes, pumps, valves, or vessels that dioactiv.ity into the environment..Al-contain it, then a " blowout" would. hap-thoug h --the containment would f ail, in pen.

In this' case' some of the water

'ntil would flash to, steam and' blow out.of the

~

time if the. core were to melt, u

that time,..the radioactivity ~~ released hole.

This could be serious since the from the fuel.would be deposited by fuel.could melt if additional ' cooling natural processes on the surfaces inside were not supplied in,,a are provided with systems to contain and a,, ' rather short the containment.

In addition, plants time.

trap the -radioactivity released within u.

...,3-.

,y the containmant building.. These systems

~

include such things as water sprays and Th'e loss of' hormal'c^o'oliid..in the event'.

pools to wash radioactivity out of the of a LOCA would stop ihe ch'ain reaction, building. atmosphere.and filters -to -trap so that the amount of heat produced radioactive particles prior -tes,their. -re,

would drop very rapidly-to a-few. percent lease.. -since the containment buildings of its operating level. However, after are made essentially.leaktight,-...the this sudden. drop 'the amount. of -heat radioactivity :is contained.as long as being+ produced would-decrease.anch more the building remains intact.

Even gif slowly and would be controlled by the the building were;to have sizable leaks, decay ;of the radioactivity in the fuel..

large amounts of ethe radioactivity would Although this decrease ir -heat '. genera-1ikely, ebe... removed..by the. systems pro-tion.cis helpful, it: would not be enough.

vided s-for that purpose or-would,be to prevent the fuel-from melting *unless deposited,on.; interior. surfaces of the additional cooling were supplied. To a

deal with this situation, reactors have building by, natural processes.

y w

emergency core cooling systems (ECCS) t

-.a -

..g 7

whose function is to provide cooling for Even though the containment building just such events., These systems have pumps, pipes valves, and water supplies, l

would be expected to remain intact for which are capable of dealing with breaks j

some time following'*a core melt, even-of various sizes.

They are also tually the molten mass would tie expected designed to be redundant so that if some.

to eat -its way through the concrete floor into the ground below.

-Following components f ail to, operate, the core.can, still be cooled.

this, much of the radioactive material would be trapped'in the soil; however, a

,u.

small amount woald escape -to the surf ace r.

e..,

and be released. Almost all of the non-g,; study' ha's' examine'd' a." large numbe^r'"cif ase us dioactivity would be trapped-potential seguences of;

~

LOCAs of various sizes., events following i

~

O

~ 'all

'In-almost of,the cases., the L' CA must be fol1' owed l

3

.r.

It is possitili Eo 'postu'late ' core melt ing system fo'r the core to melt.j)The accidents.in.which

..the containment principal exception to,this,is the: mas-building would. fail'by'overpressuriz.a-sive " failure,of the large pressure,ves-tion or. by missiles. created..by..the.

sel that contains the - core. - itowe ve r,

[

. I

2.13 WHAT ES THE NAZURE OF THE HEALTE the accumulated experience with pressure E

^

M LT ^

i vessels indicates that the chance of-H E

such. a failure is small. In fact the study found that the likelihood of pres-It is possible for a potential core melt sure vessel failure was so small that it accident to release enough radioactivity did not contribute to the overall risk,

. f s d#.xm from reactor accidents.

within a short time (about' one.' year) j after the accident. Other people may be

!j exposed to radiation levels which _ would produce observable ' effects which would require medical attention but from, which 2.10 HOW MIGa.

3EACTOR TRANSIENT LEAD they would. recover.

In addition, some TO A CORE MELT 7 people may receive even lower exposures, which would produce no noticeable ef-facts but might i;. crease the incidence The tena " reactor transient" refers to'a -

of certain dise ar.e s over a period of k

number of events that require the reac ;

many years.

The observable effects tor to be shut down. These range from-which occur shortly af ter the accident normal shutdown for such things as re ;

are called early, or acute, effects.

fueling to such unplanned but expected'

- a y

events as loss of power to the plant.

The delayed; or latent, ef fects of radi-

. i from the utility transmission lines; ation exposure could cause-some increase

[l The reactor is designed to cope with:

in the incidence of diseases such as b

unpir.nned transients by automatically" cancer,- genetic' ef fects,- and thyroid T

shutting down.

Following shutdown,l gland' illnesses in the exposed popula-cooling systems would be; operated to tion'.-

• Irr general these-effects would remove the heat produced by the radioac" appear as-an-increase in these diseases tivity in the fuel. There are several, over a 10 to 50- year period following different cooling systems capable of re -

the exposure.

Such effects may be dif-moving this heat, but if they all should ficult to notice because the increase. is fail, the heat being produced would be exlected to be small compared to the' sufficient to eventually boil away all_

normal-incidence rate of these diseases.

the cooling water and melt the core.

~

The - study has estimated the increased In addition to the above pathway to core.

incidence of potentially fatal-cancers

melt, it is also possible to postulate' over the 50 years following an accident, cora melt resulting from the failure - of-The number of -latent _ cancer fatalities.

the reactor shutdown systems-following a' are predicted -to be relatively small-transient event.

In this case it would compared to their normal incidence.

be persible for the amounts of heat-Thyroid illness refers mainly to small.

generated to be such that the-available' lumps, or nodules, on the thyroid gland.

I cooling systems might not cope with-it -

The nodules are treated by medical pro-and core melt could result.

cedures that scuetimes involve-simple surgery, and these are unlikely to lead.

l to serious consequences. Medication may -

also be needed to supplement the gland function.

2.11 HOW LIKELY IS A CORE MELT ACCIDENT?

~

Radiation is recognized as one of the factors that can produce genetic ef fects which appear as defects in a subsequent The Reactor Safety Study carefully exam.

ined the various paths leading to core generation. From the total population ~

melt.

Using methods developed in recent-exposure caused by the accident, the, years for estimating the likelihood of expected increase in genetic effects in such accidents, a probability of occur-subsequent generations can be estimated %

l rence was determined for each core melt.,

These effects are predicted to be small; i

accident identified.

These probabili-.

compared to their normal incidence rate. _

ties were combined to obtain the total!

probability of melting the core.

Thei-i value obtained was about one in 20',000.

2.13 WHAT ARE THE MOST LIKELY CONSE-per reactor per year. With 100 reactorsi 00ENCES OF A CORE MELT ACCIDENT 7 -'

l operating, as' is anticipated for. the.

l U. S m by about 1980, this means that 'the; As stated,'the probability of a core chance for one such accident.is."phe ih melt accident is on the average one in,,

,j 200 per year'.

~

~

~

^ ' ' ',

20,000 per reactor per year.

The sost

I

- I y

m

.., ~

w,,(

w%v 4 g,,3, % gg,

-um likely consequences of' such an $ccident in' that a wide range of ' consequences is are given below. r--

c.;y.,

,., c n.g_ g,,,, g 3 possible... depending..qthe exact condi,

A-under_which the. accident._ occurs.

~.E :

Ingthe., case of a core melt, the conse-l o? :-r,; 2

. nr p

MOST;LIKELY CONSEQUENCES.0F.A v;,..

7,quences;would;. depend, mainly. on three

,, factors 2 the amount of radioactivityi CORE. MELT. ACCIDENT - - u..

released,.the way.it is dispersed by-the' t-Consequences prevailing weather conditi.ons,.and the.

a.

number of people exposed to the 'radia-Fatalities.

r*

Al tion.-

With -these three factors known, it is.possible.to make a ~ reasonable

.m r-t;

<1 estimate of the consequences.

Injuries.,,

Latent Fatalities per year (1

The study calculated the health' effects and the probability of occurrence for Thyroid Nodules per year

<1 140,000 possible combinations of radio-active release magnitude, weather type, Genetic Defects per year

-<1 end population exposed. The probability j

of a given release was determined from a

+-

Property Damage (a)

<S1,000,000-careful examination of the probability of various reactor system failures. The (a)This does not include damage that probability of various weather condi-tions was obtained from we athe r - data might occur to the plant or costs for collected at many reactor sites.

The replacing the power generation lost

.,by.such damage _

probability of various__ numbers of people being exposed ~ was obtained from_ T.S.

census data for current and planned U.S.

2.14 EOW DOES THE AVERAGE ANNUAL RISK

, FROM NUCLEAR ACCIDENTS COMPARE, reactor sites. These thousands of com-TO O.T.H. ER COMM.ON,-R. ISKS?, aE.

putations 'were carried out with the aid pg.a large digital computer.

m Considerikg'.. the l5hiillioE ^pec'p#1e who

^ ~.,.

.N.1,I

~

live wi'thli'25 miles' of current.or; plan-

'Thesi results showe'd thdti the, probabili-more fatalities: fs7,resul' ting ty.of an accident.

in.10.or ned U. S.' ' Treactor.' sites, 'and based on reali:ted to -be about rates'in the II.S.

.the current. accidenK' of fatali' ties "ana,,inju-1~in,3,000,000 per p[aEt,pe'r(yeir'.

.The annual ' numbers' ries expected from various sources are probability of 100 or more fatalities is shown in the~ table below.

predicted to be about 1 in 10,000,000, and for 1000 or more,31 in 100,000;000.

.- r.

The largest-value reported in the..-study v..r ANNUAL' FATALITIES AND INJURIES was-3300. fatalities-p with a probability EXPECTED AMONG THE~15 MILLIOR PEOPLE bf'about one in~a billion.

i

'm%'

W L-l

- LIVING WITHIN 25 MILES OF U.S.

Teh e '

l' i

REACTOR SITES The above estima't'isare derived Yromb consequence model which includes statis-Accident Type _ _ Fatalities __.Tnjuries

.T.ical. calculations..;.to describe. evacpa-Automobile 3

.c:.4,200-r tions of 1.peopleJ. cut of3 the path'.of

. 375,000 airborne radioactivity. This evacuation model was. developed from. data describing Falls 1,500 75,000 evacuations'~that..have.been performpd iluring non-nuclear events.

'.s g Fire

560 22,000

., g

^f 100 similar plants ate Eldctrocution 90 Yf a group o

l considered, then the chance of an acci-

' dent causing 10 or more fatalities is.1 l

Li_gl_itni n..g.--

_8..-.._. -

in 30,000 per year.

For' accidents in-Reactors (100 2,'

20 volving 1000 or, more fatalities the 3

plants) number is 1 in'1,000,000 per year.

"In-

?

.terestingly,.' 'this value coincides with

,the probability that a meteor would 2.15 TdAT IS THE NUMBER OF ' FATALITIES strike a U.S., population center and AND INJURIES EXPECTED AS A RESULT cau'se' 1000 f atalities.

OF ' A CORE MELT ACCIDENT?

~ " -

The table shown below can be used to A core melt accidenti iis similar to many compare the likelihood of a nuclear other types of major accidents such as accident to non-nuclear accidents that fires, explosionsi dam failures, etc.,

could cause -the

-same consequences.

l l

• :. 9

I s

,.-m.

,r u :..

.n 3m.

.1'AVERiGE PROBABII,ITT' Or MK OR MAN-CAUSED

• AND NATURAL EVENTS '

' -' ' 8 e -W l d

~'#

Probability of

. w rProbability of

'1000 or More ~"

a a : e '.

a.:.-100 of' Mare T

+-1

~'~M Fatalities-J"*

Fatali~ tie s Type of' Event

y. ~

,r >

Man-caused

- ; ", n2

~'"

"'123 1

.. q

.n.

ii in-2' years 1 in 2000 years 1*

Rial Airplane;CFash J

'+17in'7 years 1 in 200 years Fire' '

4 C

• M ml'in 1B^ years 1 in 120 years'd e ;

Explosion Toxic Gas:

1 in 100 years 1 in 1000. years

.. t

.:s "t'

.s be:...,

,,e

-'1 0'a 1

Natural * '. -

. 2,.

1-

- ~ - '.

^

s.

T A-

-: E -- -

1 in 5' years very sm Tornado 2-1 ini 5 years '

1 in-2 5,

. - ' - ' d Furricane:

1 in 50 years

- : - I in -20: years 1 in 1,000,000: years 2,

Earthquake-

~ - 1 in--100',0 00 ye ars.

"r-i Meteorite impact

.... i;. u l

-< m g

~

1 I l 319

a ;;

.. g 3;

.g#g j

Reacto rs ' 'S i ^ ",, ; '- b l-y.

m,

.:c':

n c.a

,. y

,y

_, ; gg

.: 2.;.c :.c w ' ; bin :

y

- - sc -

.w

,. :,.e 1,000,000cyears 100 plants 1,' 'in -140,,000 yearso, T

' m '*

i 1

a.

..y se

.g... ;

Q m 3.) 3 3,-Q.;g 3 g

.g u 4 *:.

c.- a s

n The5e..inclNh '. man-cadsed: MwEll5I$d onif'c'identYP' tite case offiarge:,

and -Thi$L

'in' J

~U natural events.~ Many'of.tnese,piob'abila less likely ac The'e J

s nodules ities are obtained' from' hist 6rical are easily diagnosed' and treatable'by records, but others are so small that no medii af3 ' br" surg'ical-procedurerF" The incidence ' o'f other effects

• woul& be'~ low J'

~

such2;evest "has ever'beeh obseivedT 'In the latter> cases the ~ pr6b' ability 'has and should%ot' b'e' discesnible in* view-ef been calculated using techniques similaf the hith* normal incidence of these'two to th,ose.dsed' for the nucleari plant. *

' diseases". * '

J Y --

<.3. zen

,r-a.,:

1: 1 ~, ;.,

. n: -- :.

..x w ma.

h s:-?.

.... a..

- -.;a. 2 n. a. C t'

t m

-\\

Irr regard to; injuries' from potential l

nuclear power plant accidents) thes. num-INCIDENCE PER YEAR Cr LATENT PEALTH ErFECTS j

ber.of r injuries that would require medi::.

roLIm1N.G,A POTENTIAI, REACTOR ACCIDENT lh cal attention shortly af ter an-accident

'sorma1IDI k.2 :

h is about 10 times larger than-the number 4

1; of fatalities predicted..

. : chance per Incidence

. _ anactor-per__. _

. _. Rats __

V in -

'. Year 1ExposeO 2.16 WHAT IS THE MAGNITUDE OF THE LATENT, He'alth one i one in Population OR'LONG-TERM, HEALTSI EFFECTS? -.:

Effect (pec yean 20,000gAhr.

1,000,000(a)

(per year)

,3

[;

As with the'short-term; effects,..the'in-g,,

cidence -of latent cancers,. treatable Cancers'

<1 170 17,000 '

latent thyroid illness.: and: genetic 1

effects would vary with the exact Thyroid accident conditions.

The table below Illness

<1 1400 8000 M'I illustrates the potential size of such

$*,,d events.

The first column shows the ts

<1 25 8000 consequences that would be produced by i

core melt accidents, the most likely of which has one chance in '20,000 per (a)The rates due to reactor accidents are -

reactor per year';of occurri'ng.. 'The temporary and would decrease with time.

Jhe bulk of the.caneers and thyroid second column shows the consequences for siodules would occur over a few decades..... _

an accident that h4 a chance' of"'l in'.a and the genetic effects would be sig-million of occurring.; The third column

- nificantly reduced in five generations.

(b)This is the normal incidence that would shows the normal incidence rate.-- ' ' ". '

be expected for a populition of en -: ;a.

c.,..ad

..vc.r3 SMJ p.

10,000,000 people who might receive p,

.- i otQ i : s..

c-

..c- : n:ce some exposure in,a very large accident.

4-s.

Int-these. accidents konly the induction over the. time period.that the potential:yo

't j

of thyroid nodules. would. be 2 observable,

. reactor:pegident, effects might occur.

-?

- },

b

.2.17 WHAT TYPE OF PROPERTY DAMAGE MIGHT

.l.,000,000 -per Slant per year of causing A CORE MELT ACCIDENT PRODUCEJ..:

3 damage of..c.about :.fone r: billion. dollars.

. m

.:i u.-

n.

nn dhe ssnaximum va2 ue would be' predioted.-to A. nuclear accit.ent would cause.mo gphysi-be.,about 2,14 abillj oa dollars.,.,wisth : a cal damage to property beyond the nplant

.9Cobabi.lity. a wf.

.a bout..

one ain a

site but may contaminate it wiith radio-

% 004-000,000 peg plagt pec. year, q :r..a activity. At high levels of contamina-e:o:. c bed nr-n.r.3 arson n : --. e

tion, people would have to be relocated Rhis uproperty damqgt yish.;,from nuclear

.4ccidents can be. compared do other -risks from their homes.until decontamination

' procedures permitt'ed ' their 'returni..'At tin daeveral ways, -The Margest man-caused levels lower than this,' but invo1Ving :,a

, events that.have -occurred wre -fires. -In larger area,: decontamination procedures

-recent gears -there have been an average would also be" required, but. people :would

-of,three fires with damage in excess of be'able to continue to live'in' theiarea.

.-10 i million. idollars i.every year. About The area requiring decontamination ^would conce every ;two years : rthere is

.a fire involve a few hundred to. a few thousand

.vith.. damage, in the :50-to.100 million square miles. ' The principal. concern '.in dellar range. There have been four hur-this larger. area would be to monitor

.gicanes in _the,last 10 years

. hich w

farm produce to keep the' amount'.._of

~' $1111on dollars., 'Recent, earthquake es-caused damage:in the range of 0.5 to

'S radioactivity ingested through the. food chain small. Farms in this area :yould timates' suggest that a one billion have their produce monitored,and any

^* dollar earthquake can be expected in'the produce above a safe level could not be

_U'S..about once every~50 years,*

used.

j g.

i yg

g

~.Ascomparison %f Jthe~; preceding costs i:.: -

e-"

The core melt-accident having a.tikeli-shows 'that, Athough a' severe reactor hood of one in 20,000' per plant per year

. accident woul4 be'very. costly., the costs would most likely result in.Little or 'no Nodld"be;.within.the' ' range df 'other contamination. The probability :cf an

'.setious; accidents. experienced by society s

accident. that requires relocation of 20 and ~ the : probability of such' a nuclear

'y square ctiles is one in ::100,CD0 aper

accident is $stimated to ibe : smaller 'than

.e reactor - per yea r.-

Eighty per cent.cf

that 'of the other events.i - ~

'v

_all oore melt accidents would bemxpect-7f '

ed to be less severe than~ thism.The y..I,9 ; WHAT.WILL BE THE CHANCE OF A to largest ' accident might : require reloca

'I gl." REACTOR.MELTDQWN.7N 'THE YEAR

'"-}' '2000'IF NE HAVE '1000* REACTORS ~ ~., '

tion :from.290.r equare. nilesn:: In... an

ar r this,.- Tagriceltural i T.Q]%;OYERATING7, *

~

~ ME~

accident. : such products, particularly mil.k, would have

'und

'. E j '

,' _, f

,,, y.

• highf',bt ' 'te,mpted.to t,.kenthe.per to be monitored for a month or two.: over g

an area about 50 times larger until the plant' probability gf 'a particular 'reac-iodine, decayed away.,,After that, the tot'..gccident'.and multiply it by 1000 to

,y b) area requirin.g monito' ring'dould be,very accident]'in

-- edtimate' the chance of an much. smaller.

'.,, 3 e the year 2000.

This is not a valid es 7 c ;, 7c c; m e.r.

calculation, however, because it assumes

^

2.18 l'WiiAT WOUW"BE 'THE ' COST OF'TifE *l that the reactors to be. built during 'the

'J 4

CONSEQUP'.vCES OF A COFf HELT*

next 25 years will be the same~as ' hose

1. ~ * ' ' ' *~~

WI#

,being built, today.

Experience ~with on

• ' ACCIDENT 7 r)

-. :- r --

t.-.i 1

.c..ie e pt a t: r-7other#technolo3ies, such as automo'bil'es As with the other consequences,C he' cost

and 'hircraft' for example, generally t

would depend upon the exact circum-shows that, as more units are built and stances of the accident. The cost cal-more experience is gained, the overall culated by the Reactor Safety Study safety record improves in terms of fewer included the cost of moving and housing accidents occurring per unit. There are the people that were relocated, the cost changes in plants now being constructed caused by denial of land use and the

+ hat appear to be improved as compared cost associated with the denial of use to the plants analyzed in the study.

of reproducible assets such as dwellings and factories, and costs associated with 2.20 HOW DO WE KNOW THAT THE STUDY HAS the cleanup of contaminated property.

INCLUDED ALL ACCIDENTS IN THE The core melt tecident having a likeli-ANALYSIS?

hood of one in 20,000 per reactor per year would most likely cause property The study devoted a large amount of its damage of less than

$1,000,000.

The effort to ensuring that it covered those chance of an accident causing.

potential accidents of importance to

$150,002,000 damage would be about one determining the public risk. It relied in 100,000 per reactor per year. The heavily on over 20 years of experience probability would be about one in that exists in the identification and

-1.1-

(

cnclyaio of potential rx.ctor cccidents.

tha N':tionni Atron:utic7 tnd Sp':ca Ad-I

{i It also went considerably beyond earlier ministration were u ed in-the ctudy. As L

l analyses that-have been performed"by used in this study, these techniques, considering a large number of potential,

called ' event trees-and' fault " treer, -

I failures that had never before ' been-helped to define pottntual accident analyzed. ~ Pot' example, the failurei of paths and -their* likelihood aof -occur'-

reactor systems that can lead to core rence. ~

~

i melt and the 7f ailure of systems that affect the consequences-of - core melt An event tree defines #an initial'failu'r's have been analyzed. :The consequences'of within the plant. It then; examines-the the failure'of theimassive steel

• reactor course of' events which follow as detei".-

f'ilure 'of vessel and of' the containment were con-mined by' the': operation or a

sidered for the first time. The likeli~-

various systems that are' provided to 1

hood that various external forces such prevent the core-from ' melting and'. tD as earthquakes, floods,-. and tornadoes prevent the release.of radioactivity to could cause accidents was also analyzed.

the environment. Event trees were~ used

-- 0 in this, study-to define thousands of In additio'n there '

further factor's potential accident paths which were ex-are amined to, determine their likelihood of high 4egree of confidencs cccurrence and'the amount.of radioactiv-that give. a that the important and significant accf-l dents affecting risk.have been included.

ity that they might' release.

~ ',

~

~

~

These are: 1) the identification of ' ail tth significant sources of radioactivity Io-p[ulttreeswereusedto determine.

3 cated at nuclear power plants,

.)

the likelihood of failure of the various fact that a 1,arge release of radioactiv-systems identified in the event tree ity can occur only'if the reactor ffue'l' - - accident, pathai :-A'.,fmult Ctree starts-'

were to melt, and 3) knowledge of th'e with thea dafinition' of an undesired l,

physical phenomena which'can cause fuel eventc.such as-the. f allure of a-systest to melt.

This type of

• approach le'd t to operate, and then determines,. us4ag the screening of thousands'of potential engineering and mathematical-logic,. the accideni paths to identify those that ways in whicbc the. system can failt.

.ll-would essentially. determine ~ the i publi*c Using. data covering. 1) the failure.of risk.

components such - as'. pumps,. pipes-and g

while th'ere' is 'n'o way "Y 'f;J:provihg tha"t valves, 2) 5 the-likelihood of operator V

errors, and 3). thet-likelihood of mainte-o 3;

i all possible accident": se'quences which nance. errors, it is possible to: estimate contribute to public risk have been the likelihood of_ system failure, even g:

considered in the study, the systematic where.no, data ort total system failure f:

approach,,used,in identifying possible exist,s: :o accident sequences makes it.'unlikely

)

I that an acciden; was overlooked, which wculd 'significantly change the over Il The likelihood 'anci ~the' ' size of 'radioac-risk.

' *7 tive releases from~ potential accident 1 ~.

paths were used in combination with the WHAT TECHNIQUES WERE USED IN likelihood of various weather conditions.

{

2.21

"j and population distributions in the vi-3 I

PERFORMING THE STUDY?. ~

cinity of the reactor to.alculate the:

Methodologies developed over the past 10 consequences of the various potential 7

U years by the Department of Defense and accidents.,

4

.4 d

'(

m o

i m

e

..l" n:

e ;,

s

.a

.it J.

AA

,)

LI

,j

',,4

,J

'3' d

$ ?*

e

-a -

e

.m.

t-.: a?

}%

. s

*e I

u.n N n:.

s a

0

- u nq 2.n

. :2 m

. c ~.

.**.e.

• cf t-sw

• C E N taVil. O -. '. A'41*Jn.i

,.. /. f

-a 4.

• ~

I hj n :J ' c *. ; e g r. i. ::c re

.N

o

^..-

A jt

.e m...

n.

p, if

~ a l'

n