ML19305A496

Draft Rept on Mission to States by B Roche & a Cayol from 790401-06 on Accident of Tmi

ML19305A496
Person / Time
Site:	Crane
Issue date:	05/23/1979
From:	STATE, DEPT. OF
To:
Shared Package
ML19305A495	List: ML19305A493 ML19305A496
References
NUDOCS 7908220171
Download: ML19305A496 (42)

Text

.

ar+

We i

<3 Y,'

REPORT OH ' A MISSION TO THE STATES BY MESSRS. B. ROCIE AND A.

FRQ4 APRIL 1st TO APRIL 6,1979 ON THE ACCIDENT OF THE THREE MILE ISLAND NUCLEAR PLANT This report was drafted following a mission to the States fro:n April 1st to April 6,1979, to collect the maximum data available on the accident which occurred on March 28, 1979 at about 4 a.m.

on Unit 2 of the Three-Mile Island nuclear plant near Harrisburg in Pennsylvania. This mission was in addition to the permanent team in Washington which is headed by Mr. P.

Zaleski, n'tclear attach 4 at the French Embassy.

It is a synthesis of data which were given by the Nuclear Regulatory Cc= mission (NRC) within the existing agreement between this organication and the Industry Ministry.

It seems already possible to determine and comment the significant events which happened during the first hours of the accident.

However, all details concerning the causes and the development of the accident will only be available when the inquiry in process in the United Statbs' vill be over within 3 to 6 months.

It all started with an incident which generally has no serious consequences and which happened in the non nuclear part of the installation:

loss in flow of feedwater to the steam generators.

However, following the violation of a technical specification (closing of valves which should have been open), the auxiliary 79082 2 0 (H 849 353

._ a

~

a feedwater system of the steam generdtors did not play its role.

This resulted in a drying of the steam generators which prevented a good cooling of the core. Parallel to or consecutively, a

relief valve of the primary system was not totally closed a6ain, which created conditions for a leak of primary coolont water in the containment.

At this point of the development of the accident which had no important consequences for the plant and the environment, the operator, taking into account an indication of the pressurizer level which was not reflecting the real situation, stopped for several minutes the flow of the emergency cooling system which hcd automatically started.

This action resulted in a lack of cooling for the core which consequences cannot yet be evaluated.

In a second phase which lasted approximately from 5 a.m. to 8 p.m.

on March 28, the operator tried to ensure the cooling of the core.

First, he took an initiative which rapidly proved to be inadapted to the situation by turning off the main coolant pumps.

He certainly feared their damaging due to the effect of cavitation.This stopping of the pringry pacps seriously weakened the cooling of the core and resulted b 1:rportant damage to the fuel bundles (it h estimated that 10 to 25% of them were broken) as well as in a zirealcy-water chemical reaction,- causing the release of hydrogen responsible for the formation of a bubble in the upper part of the vessel. However, contrary to

'some information, there was no noticeable fusion of the uranium oxyde.

849-354

v.

- 3'-

~

In the evening of March 28, primary pumps were again in operation thus allowing for a satisfactory cooling of the core of the reactor..

As of April 6,1979, the cooling down of the plant was still a delicate operation as for this purpose it is necessary to use a circuit which is located outside the containement building and, in case of leaks, there may be radi'oactive releases in the environment.

Radioactive releases in the environment were from two main origins:

- The venting of the primary coolant which run out accidentally in the auxiliary building when the tanks for liquid effluents overflowed.

This primary coolant vs, pumped from the reactor containment building after the beginning of the accident. When the security injection circuit started, it did not trigger the isolation of the containement (a mechanism special to Three Mile Island).

- Voluntary discharges of radioactive gases outside the circuit for processing gaseous effluents which storage capacity run the risk of being saturated following an unusual production of Eas caused by the accident,.

Although the consequences of the accident on the installation are particularly serious.-it is not possible at this time to make a decision on an eventual re-activation of the reactor - radioactive be releases in the environment proved to/rather weak.

The maximum dose received by the persen assumed to be most exposed (who i= supposed to remain permanently at the limit of the perimeter of the utility) was belev 85 mrem over a period of about five days. This dose is 849 355

~

T 4-equivalent to the dose received annually by an inhabitant of the region through natural radioactivity.

In spite of certain. difficulties due to coordination problems, the -

assistance plan was carried out well by local authorities.

Finally,-NRC is making a very important effort to analyse the accident.

The conclusions will be published in the United States and the technical aspects of this analysis will b e held at the disposal of foreign safety organisms within a few weeks.

849 356

.j A.

4 9

4 er

~

s w -

y-r n

w m-g-

awye ee-

--<w----

me eanew

. - _ ~.. -.

Q

.\\

MINISTRY OF INDUSTRY

=

Service Central for the Safety Paris, April 20, 1979 ME..

of Nuclear Installations E

5s[:;f Ut/L 1 8 1979

=.

5:fE"

.....z

Subject:

Report of the French experts mission on the Three Mile Island 2

="

incident 5EE

Dear Sir:

==

First I would like to thank you for the welcome by NRC to the French

$l experts, Messrs. Cayol and Roche, during their recent mission to the

E.'

United States. The readiness, they could get - thanks to your inter-si.~".

vention - to gather very useful information to accomplish their task

==

in the best way.

.z,,_.

I am sending you a copy of the two experts' report with attachments, to let you know how they interpreted the information received by American

_g experts on the Three Mile Island incident.

To inform with utmost clearness the French opinion particularly on the consequences of the int:ident toward the population and the environment, the authorities will, in short term, publish the report.

Also I will be very grateful, if you would be good enough to let me know NRC staff review, your significant connents in a short time.

In that case, I would be willing to call you by telephone, during next week, before the report is made public.

Thanking you again for your cooperation, best regards, The Chief of the Service Central

~=

as....

. i+-

= =.

849 357 y

p=

ciih m=

..iZ. _.Z... 7:.....::........Z."

" ~ ~

-~

~Z:.

H. c e y oc r

s 1

L.

CE!; TRAL OFFICE OF NUCLEAR FACILITY SAFETY 21UCLEAR PROTECTION AND SAFETY INSTITUTE - IPSN (NCCLEAR SAFETY DEPARIMENT - DSN)

JUL la g REPORT on the MISSION TO THE UNITED STATES by Messrs.

(

B. ROCHE (Min"muJ Engineer, Central Office of Nuclear Facility Safety) and A..CAYOL (Engineer, Atomic Energy Conunission, IPSN/DSN)

I - REPORT ACCIDE!C AT THREE MILE ISLA!O U!!!T 2 ON 28 MARCH 1979, AT APPROX. 4 A.M.

(LOCAL TIME) 849 358 i

I b

r w

INTRODUCTION This report pertains to a mission organized following the announcement of an accident which befell the Three Mile Island 2. reactor in the morning of 28 March.

Its basic objective is to present a preliminary smmnary of informa-tion gathered on the site of the accident, particularly from the Nuclear Regu-latory Cnmmisolon (NRC), the federal agency responsible for nuclear safety in the United States.

Although NPC, in the first few days following the accident, inade infor-mation available to the public as soon as its reliability could be established, it nevertheless made it clear that its inquiry could not produce conclusive re-sults until weeks if not months later. Consequently, this report cannot pro-vide comprehensive information on all' aspects of the accident.

All documents which were obtained in the United states and used in the preparation of this report are included as appendices hereto.

849 359 e -

a g

vs~'

e r

PART I GENERAL DESCRIPTION OF THE MISSION I-l PURPOSE OF THE MISSION The mission was organized with the purpose of obtaining the greatest possible quantity of information on the accident at Three Mile Island 2 and deriving the first lessons from it.

It we.s made easier-by the existence of a bilateral information exchange agreement ' signed in 1974 by and between the NRC and the French Ministry of Industry.

In addition, the mission filled out the permanent team stationed in Washington under the leadership of Mr. P.

Zaleski, nuclear attache at the French Embassy.

I-2 DURATION OF THE MISSION AND PERSONS CONTACTED

• Mr. Roche arrived in Washington at'approximately 12 noon (local time) on 1 April; Mr. Cayol arrived at about 6 p.m. (local time) on the same day.

+ Mr. Cayol left the United States in the late afternoon of Thursday, 5 April; Mr. Roche departed the country in the late afternoon of the following day, Friday, 6 April.

We contacted Dr. Jo LaFleur, Assistant Director of the' Office of Inter-national Programs (OIP) of the NRC, who furnished to us all information bulle-tins issued by the NRC.

The OIP also set up an information meeting for visit-ors and foreign nuclear attachds on 3 April, as well as the tour of the site on 5 April, in which we were able to have a brief meeting with the plant opera-tors.

We were also present at a public meeting of the NRC Executive Committee held on 4 April. The meeting was intended to determine which specific actions (including possible shutdown) should be taken in regard to ope _ ting nuclear m

plants designed by Babcock and Wilcox (Three Mile Island 1, Arkansas 1, Crfstal River 3, Davis Besse 1, Oconee 1, 2 and 3, Rancho Seco). We should recall that 1

849 360

v t

2 the Governor of California called for the closing of the state's Rancho Seco

" j s

facility shortly after the accident at Three Mile Island. The meeting was called partly in response to the Governor's request.

On 4 April we attended a meeting of the Advisory Committee on Reactor Safegards (ACRS)-the equivalent of our " standing groups"

, which was called to study the accident.

Mr. Zaleski arranged an interview for us with Messrs. Larowski, Frailey and D. Okrant (the chairman, the secretary and a member of the ACRS, respec-tively) on 4 April.

C.AAo Rinally, we had a telephone interview with an engineer (Mr. gaero) from Westinghouse (Pittsburgh) in the morning of 3 April.

I-3 MAIN DOCUMENTS The principal documents we were able to obtain were as follows:

+ The bulletins issued by the NRC for distribution to U.S. leaders and to officials of the NRC itself and of other, foreign organizations.

These bulletins, numbered PNO 79-67 through 79-67K, are annexed as Appendix A.

They were distributed to all the foreign experts.

• The accident scenario prepared by Babcock and Wilcox engineers.

This document, which was kindly provided by Mr. Frailey, is annexed hereto as Appendix B.

+ The measures taken by the NRC (inspections and analysis complements) in regard to Babcock and Wilcox reactors presently in operation (Ap-pendix C).

• The report of the above-mentioned meeting of the NRC Executive Com-mittee on 4 April (Appendix D).

I-4 ACTIONS TAKEN By means of telephone conversations and telex messages, we kept the Central Office of Nuclear Facility Safety and the Nuclear Protection and Safety Institute advised of the latest daily information as it became available.

(Tne 850 001

-y aee

.m r

~

3

,r complete set of telex messages sent is annexed as Appendix E.)

I-5 FOREIGN EXPERTS PRESENT We noted the presence of engineers from the following countries: The Netherlands, Spain, Belgium, Italy, Federal Republic of Germany, Japan and Mexico.

In addition, representatives of the IAEA and the Committee of Euro-

~

Pean Communities were also present.

e s

850 002 i

s 9

t.

PART II BRIEF DESCRIPTION OF THE FACILITf Appendix F contains extracts from the plant safety report, which pro-vide general information on the facility as a whole and include diagrams of the primary loop and its relation to the containment building.

It might also be mentioned' that Three Mile Island Unit.2 had been put in service at the end of 1978 and that its fuel elements had been in operation for a total of only 60 day equivalents at full power at the time of the acci-dent.

The Three Mile Island plant is operated by Metropolitan Edison, which together with two other local companies has formed a group known as the General Public Utilities Corporation (GPU).

The operation of the two units on the site employs about 320 people.

The principal characteristics of Unit 2 are as follows:

Thermal Power 2,770 MR j

Electrical Power (Net) 880 MW l

+ Fuel Elements e UO at 2.5-3% enrichment 2

e 177 assemblies, 15 x 15 each.

+ Reactor vessel e Outside Diameter 4.7.m o Total Height 12.20 m Thickness in current portion 0.21 m o

+ Industrial Architect: Burns and Roe Inc.

• Turbine Manufacturer: Westinghouse.

850 003 4

l

PART III DESCRIPTION OF THE ACCIDE!C III-l DESCRIPTION OF THE F!RST SIXTEEN HOURS OF THE ACCIDENT ON 28 MARCH At a public meeting on 4 \\pril, minutes of which are included in Ap-pendix D, NRC engineers presented to the Commission's directors the accident scenario they had reconstructed from data recorded at the power plant.

Only data for the first 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br /> were examine'..

Taking as the time origin the moment when the accident started on 28 March (4 a.m. local time), the scenario is as follows:

Time 0

e Loss of normal supply to the steam generators, for a still unknown reason. This entails automatic shutoff of the turbine.

3-6 sec e Opening of the pressurizer discharge valve in response to signal of excess primary pressure (2255 psi - 153 bar).

From 9 e Emergency shutdown of reactor in response to excess primary pressure signal (2355 psi - 161 bar).

From 12 o Primary pressure falls to 2205 psi (152 bar) and then to to 15 sec 2147 psi (148 bar).

The temperature of the hot branch at the core outlet reaches 611*F ("*20*C).

The pressurizer exhaust vdive should have closed again at 2205 psi, but it fails to close.

30 sec e The three steam generator relief supply paraps operate at full pressure.

There is no flow through them:

the -

valves downstream of them are closed, in flaarant contra-diction of the technical soecifications.

60 see o The pressurizer level rises rapidly -

" Low" level of steam generators A and B.

2 min o Automatic cut-in of the high pressure safety injection in response to low primary pressure signal (1600 psi - 110 bar).

850 004 5

m r

r

=

6 s

i c.

From 4 o The pressurizer level exceeds the dial range. The operator

~ l to 11 min stops the two safety injection pumps at 4.5 min and 10.5 min respectively.

e At 6 min, steam flashing in the core. Primary pressure:

1350 psi (93 bar); core outlet temperature: 584*F (308'C).

e At 7.5 min, the sump pumps come on autcmatically.

e At 8 min', the operator opens the steam generaittors' emergency feed valves.

e At 8 min 18 sec, the pressure reaches a minimum in steam generator B.

e At 8 min 21 sec, the pressure (secondary loop side) within steam generator A climbs.

From 11 e The pressurizer level can be read'again.

to 12 min e High pressure emercency core cooling system camps are turned on manually.

15 min e Rupture of the explosion membranes of the pressurizer dis-charge balloon at approximately 200 psi (13.5 bar).

From 20 e Pressure and temperature of primary loop remain stable; pri-to 60 min mary pressure: 1050 psi (72 bar); core temperature: 550?F (287*C) (boiling conditions).

I hr 15 min e The operator steps the'two primary pumps of loop B.

"Some damage had certainly already occurred to the fuel rods at this stage, but the worst was yet to come" (NRC citation).

I hr 40 min e The operator shuts off the two primary pumps of loop A.

I hr 45 min e The core begins to heat up rapidly.

to 2 hr e The core exit temperature exceeds 620*F (327'C) and goes beyond the readout range for 14 minutes.

o The core inlet temperature falls to 150*F (65'C).

Frcm 2 hr o The primary pressure rises from 700 to 2150 psi (48 to 148 15 min.to 3 hr bar).

B50 005 y.--

vm

7 w

t e At 2 hr 20 min, isolation of steam generator B.

Discharge of secondary steam into atmosphere through controlled dis-charge valves.

e At 3 hr, opening of the pressurizer discharge valve (manually).

o At 3 hr, detection of radioactivity in the sump drain system.

3 hr 15 min e Peak pressure: 5 psi (0.34 bar) in the pressurizer discharge balloon.

3 hr 50 min e New pressure peak: 11 psi (0.76 bar) in the pressurizer dis-charge balloon.

e The primary pressure is estimated as 1750 psi (120 bar).

e The pressure in the containment building reaches 4.5 tsf' (0.31 bar).

o Automatic isolation of the ' containment building as pressure drops to 4 psi (0.3 bar).

From 5 e Primary pressure rises from 1250 to 2100 psi (85 to 145 bar).

to 6 hr 7 hr 30 min e Pressurizer discharge valve opened in order to reduce primary pressure.

Note: The discharge valve was operated nunerous times.

It seems to have remained useable to some extent through-out the accident.

From 8 e Primary pressure decreases to 500 psi (35 bar).

to 9 hr e The temperatures measured by the cere thermocouples (.52 measure-ments) vary widely.

Some thermocouples go off the reading scale (620 to 700*F - 325 to 370*C).

Boiling is certain to be taking place in part of the core.

e The accumulators are partially emptied into the dome (accumu-lator load pressure: 600 psi - 41 bar).

10 hr o 28-psi (1.9-bar) pressure peak in the containment building (probable hydrogen explosion).

850 006 y

n

A 8

s e Start of the containment building sprinkling system, which is then cut off (manually?) after discharging 5000 gallons (20 m ) of borated water in the building.

I 13 hr 30 min e Pressurizer discharge valve closed to increase the primary e

pressure in order to:

e reduce the size of a bubble of incondensable materials and/or steam, and e make it possible to start the primary pumps.

e Primary pumps of loop A are turned on.

o At this time, the large temperature difference between the core inlet and exhaust (cold and hot branches) indicates that I

the flow in the core is practically nil.

13 hr 30 nin e Primary pressure' increases from 650 psi to 2300 psi (44.8 bar to' 158 bar).

16 hr o Core outlet temperature decreases to 560'F (293*C).

e Core inlet tmeperature increases to 400'F (204*C).

e Steam generator A functions and discharge its steam into the condenser, where a' vacuum has been established.

Sixteen hours after the accident, the general condition of the facility is as follows:

s

• Severe damace to fuel rods. NRC engineers estimate that approximately 10% to 25% of the fuel rods have broken. Their deformation hinders the cooling of the rods in the central area of the core. Nevertheless, later measurements of the-primary water radioactivity level show a low concentration of solid fis-sion products, indicating that no sienificant fusion of the fuel has taken piece _.

• Existence of a bubble of incondensable vaoor at the too of the core.

The presence and size of this bubble could be estimated by correlating the pres-sure variations in the primary loop, the hot and cold branch temperatures and the pressurizer level.

i 850 007 r--

- -~i r -*

--n

9 s

The bubble consists basically of hydrogen generated by a =ircalloy-

~

t water reaction which occurred during the core drying stages, when the tempera-ture of the sheaths made such a reaction possible (a value of 3500*F or 1920*C was mentioned).

• Local boiling in the core, as indicated by the core thermocouples.

~

+ Highly contaminated water in the core, with a radioactivity level of 800,000 Ci/m (see Part IV).

• Contamination of the nuclear auxiliary building after approximately 3

40 m of contaminated water is dumped in it, having been drawn out of the con-tainment building by the pumps. The spillage is due to the overflowing of the liquid effluent treatment tanks.

+ Cooling of the core is ensured by the flow of primary water (use of a primary pump) in loop A and removal of the heat through steam generator A.

This procedure requires the operation of the make-up water system and, more specifically, a discharge of highly contaminated primary water at the rate of approximately 80 liters / minute toward the chemical and volume control system loops (RCV).

+ Let us recall that the primary pressure is on the order of 2000 psi (138 bar) and the hot branch temperature is 560'F (293

• C) ; the core thermo-couples indicate the presence of areas of boiling within the core.

III-2 DEVEICPMENTS AT THE FACILITY FROM 28 MARCH TO THE EVENING OF 6 APRIL 1979 We were unable to obtain precise information on the exact evolution of the plant's parameters after the first 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br />. Nevartheless, the main ac-tions of the plant operator, undertaken under the control of NRC engineers with the concurrence of Babcock and Wilcox engineers, were intended:

(a) With respect to cooline the core, to reduce the occurrence of 850 008

10 boiling in the core while gradually lowering the primary pressure.

The cold shutdown of the primary loop and its cooling by the shut-down cocling loop were made very delicate for several reasons:

+ The presence of a bubble of incondensable gases above the core:

the reduction of the pressure which would have required passage over the shut-down cooling loop might have resulted in an increased bubble size and a conse-quant partial pumping out of the core, with a risk of cavitation in the pumps (or of a potential stoppage of the primary loop flow).

• The pumps and exchangers of the shutdown cooling loop are located in the nuclear auxiliary building rather than in the confinement building, so that there was a risk of radioactive discharges into the environment (poesible losses in the loop).

+ The defor=ation of the fuel ele =ents could lead, at low pressure, to the poor cooling of the assemblies.

(b) With respect to the containment building, to reduce the hydrogen content by activating recombiners. This use of recombiners required the use of some 400 tons of lead briquettes to limit the dosage received by the per-

{

sonnel.

The lead was not available at the site and was obtained from manu-facturers over the course of two days.

Some of it was even brought in by air.

j The reco=biners were put in operation on 3 April, in the following 3

one in service, treating about 1.7 m / min, and the other in reserve.

manners Note: (1) The hydrogen level in the containment building had reached 2.2% just before the recombiner was put into operation, compared to a maximum concentration of 1.7%

in the morning of 3 April.

(2) The NRC experts estimated that some 11 deys would be re-quired to lower the hydrogen concentration to 1%.

(c) Mith respect to waste materials, to reduce the quantity cf e=is-850 009

11 s

t.

sions to the outside (rare gases) by venting the effluent tank degassing products into the interior of the containment building through a temporary line. We were unable to learn when this line was installed.

III-5 STATUS OF THE :7ACILITY ON 5 APRIL 1979 AND POTENTIAL PROBLEPE On 4 April, measurements within the core indicated that there were no further traces of boiling inside the core.

4 Maximum temperature inside cores 477'F (248'C) (only three thermo-couples indicate over 400*F, or 205'C).

(The water saturation tempe-rature at 69 bar is 285'C.)

+ Primary pressure:

1000 psi (69 bar).

• Average core inlet and out' set temperatures:

280*T (138'C).

The residual power output of the core was of the order of 5 MW.

Esti-mates suggested that the bubble of incondensable gases had decreased signifi-cantly. We were not given any quantitative indication o'f its volumer in com-3 parison, it had amounted to approximately 50 cubic feet (1.4 m ) on 3 April, and the most pessimistic earlier estimates were in the range of 900 cabic feet 3

at 875 psi (25 m at 60 bar).

+++

When we visited the site, a Metropolitan Edison representative told us that the company intended to return the plant to natural flow, completely re-filling the pressurizer. This operation would eliminate the following problems:

• Availability of the primary pumps--which have undoubtedly been damaged in the accident, so that their permanent operation may be randomly affected.

• Availability of external electricity-a loss of grid power would nean the steppage of unrelieved pumps.

• Availability of pressurizer level measurements-note that one of 850 010

12

=

g s

the three level sensors broke down on 3 April.

Nevertheless, going into natural flow presupposed that establishment of the thermosiphon could be assured. The efford of Babcock and Wilcox's en-gineers had this objective in mind, and they even considered conducting a thermo-siphon establishment test on another reactor of the same type.

. Switching to cooling by means of the shutdown cooling system was not considered in the short run, primarily because of the potential radioactive waste discharge and contaminatiion problems that might result from leakages from the loop outside the containment building, in which a particularly radio-active fluid would be carried.

III-4 HIGHLIG*dTS OF NRC'S CONCLUSIONS In the conclusion of its meeting of 4 April, the NRC indicated that the accident had resulted from the following six causes:

(1) The failure of the steam generator (ASG) relief feedwater system, related to the closed position of the isolation valves of the system, which violated the technical specifications. The valves had no doubt been closed two weeks earlier in order to permit maintenance or test operations in the loop.

(2) The failure of the pressurizer discharge valve to.close completely after the primary pressure drop.

(3) The unusable indications given by the pressurizer water level.

(4) It was not remembered that the containment building is automatically isolated by means of the emergency ~ core cooling system in the Three Mile Island reacter. Nevertheless, the activation of the isolation is supposed to have i

stepped the water removal pu=ps, thus preventing the overflow of the primary effluent tanks and consequently limiting the level of waste vented to the out-side.

850 D\\\\

9 13 s

9 t.

(5) The operator prematurely turned off the emergency core cooling system.

(6) The shutdown of the primary pumps contributed significantly to worsening the damage to the fuel rods.

III-5 DISCUSSION

~

The accident may be subdivided into successive stages. -

Stage 1 - from t = 0 tis t = 60 minutes, approximately.

In this stage, an accident occurs, resulting from the combination of two events (a) The loss of the normal steam generator feedwater system-a clas-sical transient which has been thoroughly considered in safety studies and which entails no damage to the reactor.

(b) The failure of the steam generator relief feedwater system, linked with a violation of procedures.

This ec=bination of events had not been considered in the facility's safety studies.

In the reactor, there was a significant i= balance between the heat pro-duced in the core (nuclear reaction during the first 10 to 15 seconds, followed by residual heat after the control rods dropped down after this period) and the power removed by the steam' generators, in which the water level fell rapid-ly because of the lack of feedwater.

The forced circulation Babcock and Wilccxpeam Note:

rator con-tains only a small amount of water; the rate f flow throuch it

_ g _"

is five to ten times faster than in a 0-tube recirculating steam generator (Westinghouse, Combustion Engineering and KWU).

The behavior of the plant at the beginning of an accident of this kind is characterized by a rapid heating of the primary fluid accompanied by a simi-850 012

14 u

s larly rapid increase in pressure. The pressuri=er discharge devices open normally.

(According to the NRC, the two safety valves of the pressurizer did work properly.) The failure of the discharge valve to close can be ex-plained, since it was forced to operate in water or in a water-steam mixture, conditions for which it was not designed.

The failure of the discharge valve to close undoubtedly aggravated the event, inasmuch as it caused a depressurization of the primary loop and a concurrent deterioration of the heat exchange at tha fuel rod sheaths.

The increase in the measured pressuri=er level is related to the ex-pansion resulting from the depressuri=ation.

It should be recalled that the pressuri=er level is evaluated on the basis of the pressure difference in a water column, which measure becomes less significant in the case of a water-stean emulsion.

The closing of the emergency cooling system by the operator between t = 4.5 minutes and t = 12 minutes led to a mass boiling in the core, which

=ay have resulted in damage to the fuel rods (deformation, and possible rup-tures). The reasons for the closing may be related to the herator's wish to keep the pressuriser level within the instrument scale.

Several incidental er induced events took place during this stage:

(a) The automatic starti-up of the containment sump pumps, re-sulting from the presence of water at the bottom of the s1

'Ibe water un-6 doubtedly c partly from the pressurizer discharge (total volume, I

3 c

24 m ), whose valves are set to approximatel' 7.5 bar The avb "e Lvalves e

were in operation until t = 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br />, at which time the containment vessel was isolated in response to the "High Containment Pressure" signal.

l i

(b) The rupture of the discharge disks (at about 13 bar) of the pres-s;. ricer discharge tank. This rupture was no dcubt due to the large cuantity 850 013

15 r

1

'O of water discharged by the discharge valve.

- 1 In any case, the opening of the steam generator feedwater isolation valves by the operator at t = 8 minutes, together with the renewed cut-in of the emergency core cooling system at t = 12 minutes, led to the reactor being in a nearly stable condition at t = 60 minutes, the end of the first stage.

But conditions'in the secondary loop (pressure of the order of 64 bar and tem-Perature of the order of 280*C) made it impossible to ensure a temperature clearly lower than the saturation temperature in the pri-y loop, where the pressure was 70 bar. This explains the presence of steam in the primary water.

Stage 2 - from t = 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> to approximately t = 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br />.

At t = 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> 15 minutes and t = 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> 40 minutes, respectively, the operator closes the pri=ary pu=ps (tiwo' per loop) of loop 3 and loop A.

No reasons were officially made known for this shutoff. According to the NRC's empson, th2y are probably related to feared or observed cavi-tation in the punps. The primary fluid contained significan* gap which could in fact cause cavitation events.

Thus, it may be supposed that the operator hoped to cool the core by natural flow.

Nevertheless, according to NRC experts, the presence of steam

" locks" in the primary water (at the U bends of the' hot pipes at the top of the steam generators), in conNnction with the condition of the secondary loop, undeniably prevented the thermosiphon from becoming established.

The exact reasons for the stopping of the pumps, which now seems to have been a serious error in $udgement, will not be known until the completion of the detailed inquiry now being pursued by the NRC.

In any case, the core was quite empty and had no cooling at all after the loop A pu=ps were turned off at t = 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> 40 minutes.

The core (or at least its central portion) dried out several times, leading to overheating of 850 014 e- - - -

-w

e 16 s,

the sheaths and significant zircalloy-water reactions. As a result, hydrogen was released.

It should be noted in this regard that the design of the safe-guard systems is intended to ensure that al:nost no zircalloy-water reaction (a strongly exothermic reaction) occurs in any design accident, by limiting the temperature of the sheaths to values below 1200*C.

It is difficult to discern the exact reasons for the various actions taken by the operator during this stage of the accident. One may suppose, never-theless, that the operator found himself faced with an unknown situation ands (a) attempted to determine just what the situation was, and (b) tried various approaches leading to reestablishing a stable flow in the core.

Inparticular,onemayattempttoexplainhisvariousactionsasfollow$:

• Handling of the Pressurizer Discharge Valve Prell=inary Note:

It is not exactly known whether the operator con-trolled the valve itself or the (normally open) isolation valve located upstream of it.

The operator closed this valve in order to increase the primary pres-sure with a view to cendensing the steam in the loop; the condensation was meant to make the primary fluid homogeneous and thus allow the thermosiphen a.k.dM.

to function.

But the only result of the c.;pr was to increase the Iri=ary pressure without condensing the steam in the core.

The operator then tried the opposite action. By opening the pres-surizer relief valve, he sought to reduce the primary pressure, thinking he would thus be able to go into the shutdown cooling system (NRC information).

The lowered pressure resulted in an increased flow of the high-pressure emer-gency core cooling system and the automatic start-up of the accumulators (cpening pressure, approximately 40 bar), whose operation was not inhibited 850 015

17 s

as it should have been under the normal sequence of passage to the cold shut-

)

down mode.

In conclusion, it may be supposed that the operator again closed the pressurizer valves at approximately t = 13 hours1.50463e-4 days <br />0.00361 hours <br />2.149471e-5 weeks <br />4.9465e-6 months <br /> 30 minutes, in an attempt to increase the primary pressure in order to condense the steam in the entire loop and thus m*ake it possible to start the primary loop A with limited cavit:sticn.

His choice of loop A is explained by the fact that the pressurizar is in this loop.

The latter operation also made it possible to properly cool the core.

The reasons for the success of the operation are difficult to determine at this time. The arrival of cold water from the accumulators must have contributed to it but'is not sufficient to fully explain it.

The condition (pressure) cf the secondary loop also must have played a part. Ne cannot formulate a defi-nitive conclusion without knowing this condition.

• Isolation of the Steam Generators The reasons for the isolation of the steam generators are not fully known in detail.

perhaps the operator acted in this manner in order to pre-vent discharges of the contaminate'd primary fluid into the envi v-ment, which

=ight have resulted frem preexisting leakages in the steam generator pipes.

At any rate, the operator let out the steam from steam generator A (the only one that was back in operation) through the condenser at t = 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br />, in order to prevent direct venting of s. team into the atmosphere.

+ General Remarks (1) The NRC was advised of the accident about 7-8 a.n.

(that is, at t=

3-4 hours) ; NRC " Inspection and Enforcement" teams arrived at the site at approximately 10:30 a.m.

Babcock & Wilcox's engineers were also notified (at the sane time as the NRC).

Thus, it can be said that by 10-11 a.m.

(that is, 850 016

18 s

a*.

= 7-8 hours) the optrator was being advised by technicians who undoubted-

~

ly were able to suggest courses of action to him.

p Q Q,4 (2) In his presentation on 3 April, Thompson of the NP.C stated 14skt the coerator, act! f M _ m _ u,6n, had certainly attempted to

--, S'M n~-

g prevent damage,to the facility; this should be kept in mind when studying the shut-off of the primary pumps at t = 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> 40 minutes.

Stage 3 - From t = 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br /> (about 8 p.m. on 28 t'. arch) to S April.

• Core Cooling bv Means of Forced Circulation Throuch Loon A The information obtained from the NRC indicates that the operations undertaken during this stage were intended to reduce the primary pressure with a view to an eventual cold shutdown.

It may be said that the primary concern of the operator was to reduce the residual power and that the teams of experts under the leadership of Babcock & Wilcox tried to estimate the size of the gas bubble by looking at the variation in the pressurizer level as the te=perature and pressure changed.

The remarks that might be made in regard to this stage of the acci-dent have to do with how the size of the gas bubble was estimated by the ex-perts and with the actions taken to reduce it.

• Formation of the Bubble The hydrogen content'in the reactor vessel, which was higher than 2%,

clearly indicated that a significant circalloy-water reaction had taken place during stage 2.

This reaction was the cause of the formation of the gas bubble above the cere; the gases contained in the fuel rods damaged in the ac-ident centributed only slightly to the formation of the bubble, according to the N ?.

• Estimation of Bubble Size M:\\

The NRC (or the o,m* M ^ _2iny) requested a working group con-sisting cf technicians frem various organizatiens (NFC, builders, universities) 850 017

19 s

to develop a correlation that would enable the volume of the bubble to be estimated from available parameters (pressure, temperatures and pressurizer level). The group calculated a suitable correlation within two days.

• Reabsoretion of the Bubble According to the NRC, the bubble was reabsorbed because of:

o the transfer of the gases through the hot branch A toward the pres-surizer, from which they were removed inside the containment building by means of a remotely controlled opening event; o the partial recombination of the hydrogen and oxygen; o hydrogen leakages in the control rod mechanism housings (this ex-planation is open to doubt, in our opinion).

It is also undeniable that the improved calculation precision made it possible to eliminate the safety =argins initially adopted in esrimating the volu=e of the bubble.

850 018 e

e-e n--

s w

s PART IV RADIOLOGICAL ASPECTS OF THE ACCIDEUT IV-1 SOURCES OF WAS"E DISCHARGES Radioactive discharges into the atmosphere had three sources:

(1) The assumed leakage from steam generator a may have resulted in a

~

release of rare gases and iodine into the environment in the first few hours after the accident. We have been unable to obtain any estimate of the amounts released in this' manner.

(2) The overflow of the head tanks of the liquid effluent treatment system resulted in the spillage of approximately 40 cubic meters of highly contaminated water in the auxiliary building. After filtration (absolute filter and iodine filter), the ventilation system of the auxiliary building ejected the rare gases (krypton and xenon) contained in the water.

These uncontrolled emissions were responsible for the radicactivity observed near the facility and in the plume in the first few days following the accident. On a date which was not specified, the company installed vinyl covers over the water contained in the auxiliary building, in order to reduce the rate of degassing and, consequently, the amount of radioactive a hsions.

(3) The exposure to air of the primary water, which was removed at the' rate of 80 liters / minute by the volume and chemical control system, led to periodic controlled emissions.

In fact, the radioactive rare gases initially dissolved in the primary water were collected in tanks designed for storage and treatment of gaseous effluents.

The intake of gas was higher than the-storage capacity and, in order to prevent untimely emissions through the valves of these tanks (venting pressure, 7-8 bar), the ecmpany deemed it more advisable (with the agreement of the NRC) to engage in centrolled releases.

l At a later, non-specified date, the cc=pany returned the gases to the 850 019

21' containment building, in order to reduce the level of emissions into the environ-ment.

Moreover, it appears that the radioactive discharges into the river resulted from a conscious decision (made with Tl authorization): 230 cubic meters of slightly contaminated waste water were released.

~

IV-2 EXPOSURE OF PERSONNEL In the morning of 29 March, during a primary water sampling cperation in the nuclear auxiliary building, two employees (one operator and one chemist) received total doses of 3.1 and 3.4 rems respectively. The maximum allowable dose for workers is 5 rems per year and 3 rems per three-month period.

Ten other persons in the duty team received; doses of the order of 2 to 3 rems in the first few hours after the accident.

Finally, by 5 April the exposure levels in the control room were 0.1 mrem /hr (compared with 0.4 mrem /hr on 2 April). Masks were not recuired, but the operators did wear them, out of precaution, during a few hours after the accident.

IV-3 PRIMARY WATER RADIOACTIVITY AND DCSIMITRY IN THE BUILDINGS IV-3.1 Primary Water Radioactivity Analysis of a sample of primary water indicated a contamination level of 800,000 Ci/m (confirming the significant rate of emission of fission prod-ucts in the core).

The radioactivity, which amounted to a dosage rate of 1,000 mrem /hr frem a 100 cm sample, was due pri=arily to rare gases (xenon and krypton),

iodines, cesium and, to a very small extent (we were unable to obtain the exact results of the isotope analyses), to non-volatile products (strontium).

This indicates that no accrec_iable_ meltine occurred in the fuel elements.

850 020 L_

22 IV-3.2 Radioactivity and Exposure Levels Inside the Containnent Building No exact values were provided of the radioactivity level inside the containment building. Values of 700 Ci/m on 31 March and 70 Ci/m on 2 April were announced by the NRC and then denied.

The radiation sensors installed within the contatnment building in-dicated conflicting values, namely:

• 10 to 20 rad /hr at the service floor,

+ 10,000 to 30,000 rad /hr at the' top of the building.

We cannot guarantee the accuracy of these readings, which were given to us orally. We were also told that the sensor at the top of the building was a shielded but unsealed sensor, intended to measure the direct radiation from the core, so that the infiltration of ra'dioactive gases under the shield-ing would be responsible, according to our infor=eri for the very high reading.

In our opinion, the sensors inside the containment building were not designed for such levels of radioactivity, so that the values they registered had no relation to +2ality.

On the other hand, radiation measurements in contact with the outside wall of the containment building (approximately 1.2 meters thick) indicated a Q

dosagie of less than one nrad/hr.

Considering that the containment building walls have an attenuation coefficient of 10, the radiation level inside the building certainly did not exceed 10,000 rad /hr.

IV-3.3 Radioactivity Levels Inside the Auxiliary Building The radioactivity levels inside the nuclear auxiliary building are likely to have caused access problems, including, among others:

• the difficulty of installing the hydrogen recombiner;

• the dosqpes in contact with the gaseous effluent storage tanks (60 rad /hr), which prevented certain interventions; 850 021

)

i i

23 r

+ the des ge rates (10 rad /hr) after the transfer of contaminated water to the auxiliary building.

IV-4 EXPOSURE OF THE PUBLIC IV-4.1 Direct Excosure Wind direction on 28 March was toward the north-northeast, so that

~

~

the town of Middletown (12,000 inhabitants), five miles distant. from the reactor, was in the path of wind from the plant.

Data from two of seventeen dosimeters which have been installed in fixed stations for three months show accident-caused 32-hour integrated dosages of:

• 65 = rad, 700 meters north of the reactor, and

+ 22 mrad, 1,000 meters north-northea'st df the reac*wr.

nese values do not include the radiation due to the rare gas enissions after noon on 29 March.

The occasional uncontrolled releases of rare gases on 29 and 30 March resulted in transient readings of up to 25 to 30 mrad /hr at the facility bound-ary. They lasted for only a short period (certainly less than~ene hour).

The NRC's esti=ates of the public's exposure balance over the five days following the accident are as follows:

• hypothetical maximum individual dose:

80 mrem;

• mean dose for the 2,000 inhabitants nearest the plant: 9 nrem, compared to the annual dose due to background radiation, equal to more than 100 crem.

By 3 April, the dosage rate outside the facility had fallen to less than 0.01 to 0.04 mrer/ar.

IV-4.2 Iodine Contamination A total of 152 atmospheric samples were taken. Only eight of them l

850 022

- i l

J

24 a

s t.

~

indicated iodine-131 activity levels from 0.3 to 2.5 pCi/m.

The highest observed activity was equivalent to one-fourth of the maximum allowable con-centration in the United States for a continuous emission (U.S. regulation:

10 CFR 50, Appendix 1).

Within a 20-kilometer radius of the power plant, a total of 56 milk

~

speciments were obtained (cow's milk and goat milk) in 20 farms. The measure-ments showed that 19 samples were not interpretable and the renaining 38 were

-lower than the measurement sensitivity threshold.

A detailed analysis of nine sar:ples revealed radioactivity levels ranging frcn 10 to 40 picoeuries of iodine-131 per liter of milk.

In comparison, th. Federal Department of Health sets a limit of 12,000 picoeuries of iodine-131 per. liter of milk before cattle are required to be fed on corn feed.

Note:

(1) The maximum allowable concentrations in France (Decree of 20 June 1966) are as follows for iodine-131:

3

• in air:

200 pCi/m in soluble form;

• in water or nilk:

1,000 pCi/m in soluble forn.

(2) The decrees of 10 August 1976 limit the mean weekly volume-specific activity for a normally operating 3,000.W reactor unit to a maximum of 0.2 pCi/=3 for aerosols (essentially iodine).

850 023

s e

s.

PART V PROBLElis REIATED TO THE EMERGENCY PIAN V-1 THE 'IWO EPS.RGENCY PLANS A distinction should be made between

• The power plant's internal emergency plan, which is the plant ope-rator's responsibility as long as the accident is unlikely to affect areas out-side the plant site.

In the United States, this plan is regulated by Appendix E of 10 CFR 50 (Code of Federal Regulations, equivalent to France's Journal of-ficiel).

• The external emergency plan, which is the responsibility of the state governor and local authorities. This plan is not covered by any specific regu-lation.

It is covered only by an exchange of letters with the NRC, which grants its approval in principle to the provisions of the plan, and the local authori-ties, who strive to implement certain r.eans in case of an accident. Indeed, it appears that the plan of M.c State of Pennsylvania had not been fully developed; none of the American states which have nuclear facilities have an IEC-approved emergency plan at this time.

The emergency plans designed to cover accidents in civilian nuclear facilities seen to be valid as well in the case of accidents in :2ilitary instal-lations.

V-2 RESPCNSIBILITIES OF THE VARIOUS PARTIES It should be made clear that the decision as to whether or not to evacuate the populatinn is the responsibility of the State Governor.

'"he IEC plays only a consultative role in this regard.

In view of this situatien, the NRC could not provide to us exact indications on the implementation of the erergency plan.

850 024 25

- l

26

=

t.

s Evacuation operations are the responsibility of local authorities (State and Counties).

In regard to the various technical decisions that must.be made with respect to the operation of the plant after an accident, the plant operator submits its proposed solutions to the NRC, which should give its concurrence.

But the operator is responsible for the cent squences of its actions, neverthe-less.

V-3 CEVEIDPMENT OF THE EMERGENCY OPERATIONS Although the accident occurred at approximately 4 a.m. on Wednesday, 28 March, the NRC and the State of Pennsylvania were not informed of it until n'early 7 a.m.; this may mean that the plant operating company was not irnediate-ly aware of its seriousness..The alarm was given at about 10 a.m.

Starting on the same morning of the 28th, 40 people from the Brock-haven laboratory (New York) arrived at the site with their ecuipments this organization is particularly well-trained in obtaining soil and water samples.

Several NRC mobile units arrived at the site in the afternoon of 28 March, charged with the tasks of collecting c;:__ h '

kW

......s and, if necessary, under-taking decontamination measures.

U.S. Ar=y units fron the nuclear test center in Nevada arrived with airplanes and helicopters in the afternoon of Thursday, 29 March. These units were responsible only for collecting at:nospheric spe-cimens.

On Friday, 30 March, experts from the federal Food and Drug Adminis-tration reached the site.

According to information obtained from the NRC, tne first few days in-diffi mediately following the accident seemed to be characterized by some &c in ensuring proper organization and coordination ar.ong the various in_:r.

.._og parties (local authorities, fe deral agencies, private corporations). These 850 025

27 s

were most apparent in the following respects:

+ The ccmplete saturation of telephone lines in the area affected by the accident. The telephone company quickly arrived and very rapidly installed several direct lines to remove the bottleneck.

• The difficulty encountered in using the various specimens collected by the different sampling organizations, due to the difficulty of adequately referencing the collection sites and the lack of sample standardizations three fourths of the milk and water specimens collected in the first two days after the accident were not usable, uk'. kJk q,

On Friday, the op w ^ g ~ n my decided to permit significant emis-sions at the 3:mmunstack (1.2 rem /hr, measured at the samisestack). This led the governor, acting upon a suggestion from the NRO, to go on television to recommend the evacuation of pregnant women and pre-school children within a radius of five miles (8 kilometers) of the power plant. This recc==endation was interpreted by the population as an evacuation order. Within a radius of 8 tp 10 miles (13 to 16 kilometers) around the plant, residents attempted to leave their towns, giving rise to numerous bottlenecks and traffic jams, cars lined up to top off their tanks at gas stations, which were able to meet the demard; and depositors lined up at banks, withdrawing a total of 10 to 15 mil-lion dollars in one and a half' days.

According to information we were able to obtain, the man in charge of the emergency plan at the local level was not in favor of the evacuation.

It seems that a large part of the people who left their homes soon returned.

In fact, about 15,000 people (pregnant we=en and young children) remained away fr:m their homes for a prolonged period.

On Tuesday, 2 April, the authorities completed the study of an evacua-i t i:.- plan covering the 600,000 inhabitants of the area within a radius of 20 850 026

28 9

'h miles (32 kilometers) of the power plant. The plan would no doubt have been implemented if it had been necessary to take specific measures to suppress the bubble at the top of the reactor core, or if the hydrogen content inside the containment vessel had reached a dangerous level (above 4%).

Discussions on the effectiveness of ingesting bdine as a preventive

~

measure had been taking place for many months; this non-radioactive iodine is potentially capable of saturating the thyroid e-d thus preventing the absorption of radioactive iodine. The Three Mile Island accident put an end to these dis-cussions: on Saturday, 31 March, those responsible for the emergency plan con-cerned themselves with the distribution of potassium iodide doses to the in-habitants. As it happens, however, no potassium iodide was available on U.S.

l soilinthemorningof3][g] March. 'Thanks to the efforts of an American chemical company, 50,000 doses of the chemical had been brought to the Harris-burg airport by Satusday evening. None was distributed, not even to the power plant's employees, since the company management though it useless.

Furthermore, the authorities reco== ended that cattle be fed prepared feed, but the sale of milk was not prohibited.

V-4 CONCLUSIO!!

In spite of certain difficulties during the first few hours or even the first two days after the start of t.he accident, the local authorities were able to adequately manage the implementation of an emergency plan.

850 027 l

l l

- l l

PART VI SITE VISIT on Thursday, 5 April 1979, the NRC arranged a visit to the site by car.

The tour included members of the NRC, foreign visitors and foreign embassy nuclear attaches (see list in Appendix G).

It was not possible to enter the property of the power plant.. We were able, however, to look at the plant from a lookout on the left bank of the river, to the right of the island on which the plant is located. The only sign that there was anything at all wrong at the facility was the presence of journa. lists and technicians whose mobile units were parked near the lookout.

4 Some cows grazed on meadows a few kilometers from the power plant. We noticed that a small ornamental plant store halfway from the nuclear plant and the town of Middletown remained open.

Going through Middletown, which lies about six kilometers from the plant, at about 2 o' clock in the afternoon, we noticed no signs of anything unusual--no extra police, a few children and adults in the streets.

In short, the town did not appear deserted, although it also did not seem too lively.

We could not say whether the many houses.in the countryside around the town were occupied or not.

While in Middletown, we found a representative of the power company, who was able to briefly answer a number of cuestions.

850 028 29 l

w-m-

w e.

e-m w

.e

PART VII ADDITIONAL INFORMATION CBTAINED VII-l ORGANI::ATION ESTABLISHED BY THE AUTHORITIES

~

VII-1.1 NRC Personnel From the start of the accident, the NRC installed an appropriate organization with the objective of ensuring control of the situation, cen-tralizing and distributing information to the responsible parties, and in-forming the media.

In addition, the NRC made all appropriate arrangements to obtain the greatest possible amount of information in order to permit a detailed analysis of the accident to be made (hearing of all persons who vere in the control room at the time of the accide.nt). This analysis will be made public as soon as it is available, within two to six months.

PNO 79-67D (Appendix A) lists the personnel assigned to the site by the NRC, which included:

+ 17 inspect rs, from three Inspection and Enforcement regions U

b b LRtW,

44 r=M "'ri _ M et bn agents from three I&E and National 4

i Reactor Regulation (NRR) inspection regions 3 sr^'~ ~ to pro de information, from the NRR

[L h

o-v.

• 19 NRR specialists 83.

g The NRC personnel at their disposal significant amounts of equip-ment, including in particular:

+ 3 DOE (Department of Energy) helicopters 2 DCE laboratory tru;hr b

+ 1 NRC instrument 'auck7 \\/M, III also assembled a crisis staff at its Bethesda facilities, con-sisting of specialists itself and other NRC divisions.

(Although SRC head-30 850 029

31 s

quarters are located in Washington, D.C., nest of its facilities are in Bethesda, Maryland, 20 kilometers from downtown Washington.)

This crisis staff centralized the information and published bulletins with the latest developments of the situation (NPO or Preliminary Notification of Event or Unusual Occasion, which are included in Appendix A).

The bulletins were addressed only to the NRC's leaders at the start, but after 30 March they were also distributed to other federal organs, including:

• EPA - Envorinmental Protection Agency

+ FDA - Food and Drug Administration

+ DCE - Department of Energy, and in particular to a " White House Situation Room", a crisis organization specially created by the President of the United States.

k pdc.A,

VII-1.2 Other We noted that the personnel of the Middlebourg air base, 5 kilometers frem Three Mile Island, had been placed on alert.

We were also told, moreover, that the total number of federal govern-ment personnel (including the NRC) mobilized to the site was of the order of 270 persons.

VII-1.3 Press.Information i

We did not collect exact infor=ation en this natter.

We were only

.able to learn, through NRC engineers, that the U.S. President had requested that all informatien be given to the press through the NRC, on or about 30 March.

The PNO's annexed in Appendix A were not distributed to the press.

However, the NRC did make all the inforr.ation they contain available to the press in its press conferences.

Generally speaking, it was evident that the NRC wished to make in-850 030

32 a

t.

formation available to the public within as short a period as possible.

VII-1.4 Information for Foreign Excerts The experts sent by foreign countries were welcomed by the NRC office

~

of International Programs, which organized the following activities for them, in particular:

• a meeting on 3 April, in which an NRC inspector replied very openly to written and oral questions put to him: and

• a site visit on 5 April.

Furthermore, CIP also distributed all PNO's to the foreign representa-tives as soon as they were issued. Finally, the NRC informed the foreign re-presentatives of the conclusions of its analyses of.the accident.

VII-2 REACTIONS AT TE TIME OF TE MISSION We have not analyzed the public, press or industrial reactions to the Three Mile Island accident.

We were able to observe some significant reactions, however:

+ Nuclear Controversy. The accident will give new strength to a centroversy that had been losing speed.

In particular, it will be an excellent springboard for Ralph Nader, whose public opinion ratings had been in a marked decline. The cartoon published in a local paper (York Daily Pecord) and en-closed in Appendix H is significant.

In conclusion, opponents of nuclear power distributed the tract a sample of which is included in Appendix I at the publie

=eeting on 4 April.

• Press Reactions.

By reading the daily papers (The Washington Post, basically) and listening to the radio and TV, we were able to observe that the nedia kel.t the public informed of the situation frem one hour to the next.

The newspaper articles,- as well as the radio and "Y co=mentators, were extremely 850 031

33 a

s to measured and could in no way lead to panic reactions. Further.7 ore, the press started to discuss the financial implications of the accident as early as Tuesday, 3 April, at which time the situation had been brought under control.

.c VII-3 THE CASE OF OTHER BABCOCK & WILCOX REAC" ORS The NRC took very quick action--on 1 and 5 April 1979-in regard to other Babcock & Wilcox reactors under construction or in operation (see Ap-pendix C).

These actions, which are detailed below, pertained essentially to

• a special inspection efforts

• additional analysis;

• expected changes in facilities and operating procedures.

In any case, the NRC did not deem it necessary to take more rigorous measures with respect to these reactors at its meeting of 4 April, since the results of its inquiries were not yet known.

It should also be noted that copies of the PNO bulletins were distrib-uted to all parties holding reactor construction or operation permits, regard-less of the reactor designer.

The questions raised by the NRC in its Bulletin 69-05 Revision A, of 5 April 1979 (see Appendix C), may be summarised as follows:

A - Study of the accident, in order to examine what errors were made, study the usual behavior of the reactor at the time of incidents such as feedwater loss, and make personnel more aware of the seriousness of the consequences of wrong actions.

B - Verify the safeguard systems maintenance and test procedures, and particularly the retention of the functions provided by these systems during tests and their later return to proper operating position.

850 032

CONCLUSION This conclusion is voluntarily limited to the analysis of the develop-ment of the accident, the only subject on which the NRC was able to furnish exact information.

First of all, we must point out that the IDE: did not hesitate in making available the greatest possible amount of information to the foreign visitors who came to learn about the accident.

We believe it is premature to attempt to drawn definitive conclusions from the accident, particularly with rsspect to its exact causes (design de-feet, defective operation and control procedures, etc.) : it will be necessary to wait for the results of the inquiry now being conducted by the NRC.

It may be noted at this time,'however, that the accident followed the non-availability of the steam generator emergency feedwater supply system after the violation of a technical specification which called for the valves in this circuit to be kept open.

The consequences of the resulting accident sequence to the facility were not considered in the plant safety studies and operation procedures.

The operator thus found himself in an unexpected situation and made inappro-priate decisions--shutting off the emergency core cooling system between t =

4.5 minutes and t = 12 minutes and, particularly, shutting off the primary loop pumps at t = 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> 40 uinutes.

The latter operation had the most serious i= pact on the core.

The failure to close the pressurizer valve seems to us to be more a consequence than a cause of the accident, even though it may have contributed to the core cooling difficulties.

Furthermore, the operator was undoubtedly strongly hindered in his 35 850 033 b-

36 s

e appraisal of the situation by the absence of indications of the condition of the primary loop (water level in the pressurizer).

It is very likely that, during the first stage of the accident, he was not aware of the exact condi-tion of the facility.

In addition, the very rapid drying time of steam gen 6-rators of the Babcock & Wilcox type allowed him too little time to take more well-considered decisions, at least in the initial stages of the accident.

The operator does not seem to us to have known how to take advantage of the possibilities offered him by the secondary loop.

It is important to note that the accident as a whole had severe con-MAO

k. 4 sequences for the fuel elements, although no c';ni--et_'""#^"

f the fuel took place. There occurred a strong zirealloy-water reaction, which undedbted-ly had a greater imeact than that of the reaction traditionally contemplated W6 O for certain '----- - ' Te_ accidents (such as complete rupture of a large pipe in the primary loop) when conducting the safety studies for power plants of this type. The reaction resulted in the formation of a bubble of incondensable gases which contributed greatly to preventing the operator from bringing the reactor back to a safe condition.

E

2b, Nevertheless, the safeguard systems normally installed in the #

- f Nbk.

to diminish the consequences of accidental sequences different from the one that actually happened were ab1'e to play their part in preventing unacce le enissions into the environment.

In this regard, the presence of iodine in the auxiliary building ventilation system made it possible to avoid unac-ceptable emissions that might have led to a population evacuation.

8,50 034

.