ML20136C730

Rept of Subcommittee of ACRS on Detroit Edison Reactor

ML20136C730
Person / Time
Site:	Fermi
Issue date:	07/21/1955
From:	US ATOMIC ENERGY COMMISSION (AEC)
To:
Shared Package
ML20136C593	List: ML20136C588 ML20136C730
References
FOIA-85-646 ACRS-GENERAL, NUDOCS 8511210138
Download: ML20136C730 (8)

Text

W-e em e - - *e w e mese . --Mm. . g-,, . . . , , ,

~

y 4

e

+

REPORT OF SUB00MMITTEE OF ADVISORY COMMI'" TEE ON REACTOR SAFEGUARDS l

ON DETROIT EDISON REACTOR n!/7s1 8511210138 851028 PDR FOIA BALDQHBS-646 PDR

l. Introduction

(

Two subcomittee meetings have been held with the representatives of the Atonio Power Development Associates. The first meeting was held in Cambridge March 3. Present representing ACRS were Messrs. Brooks, Rogers, and Russell. APDA was represented by Messrs. Amorosi, Jena (NDA), and McCarthy.

The second meeting was held at the Detroit Edison offices April 20 At this meeting ACRS was represented by Messrs. Brooks, Rogers, McCullough, Wensel and Russell. The APDA group was represented by Messrs. Amorosi, McMath, Jens, Petersdorf, Yevick, n ecker, Everett, and Whipple. Professor Hans A.

Bethe also attended in his capacity as consultant to APDA. Formal presenta-tions by APDA were made at this meeting ar, follovas Coolant system, Petersdorf Reactor vessel and piping system, Yevick Plant site and building, McMath Fuel elements, Everett Nuclear characteristics, McCarthy

! Control characteristics and accident studies, B ecker f Schedule and financing, Amorosi No formal written report has yet been submitted by APDA to ACRS. The tentative reactor descriptien and conclusions contained in the present pre 14=4=v subcommittee repcrt are based only on informal oral presentations s by APDA and informal discussions with members of the subcommittee, 2 General Description of the Project The AFDA is a fast converter type reactor, sodium cooled, and egloying partially enriched uranium as fuel with a ratio of about three parts U-238 t

to one part U-235. The critical mass is about kl5 KG of U-235. The tet>al heat power of 300 megawatts is removed by a primary loop of sodium with an inlet tagerature of 5500 F. and a 2500 F. temperature rise through the reactor. The primary loop transfers its heat to an intermediate MaK secondary loop, operating about 500 F. lower, and the secondary loop trans-fers its heat to the steam generator, the steam condition being 600 psi.

! at 7000 F.

The reactor includes a breeding blanket both on the sides and on the ends of the fuel elements. Approximately 25% of the conversion occurs in the ccre to 75% in the blanket. The plutonium inventory in core and blanket builds up to about 120 KG after 16 ye , and is 75 XD at the end of h

, years. The neutron flux is roughly 1 neutrons /cm2 /sec.atthecenter of the core, with an average energy of about 0 3 mar.

Two types of fuel element design have been considered, and although a final decision has not been reached, both types - pin and flat plate -

involve the same external dimensions of the reactor and similar flow

( ,

.. . ~

. /-

conditions. For both elements the reactor cycle is conservatively designed

( for 15 burn-up of total uranium or 34 45 burn-up of the contained U-235.

Corresponding to this the mechanical design of the fuel allotas for 5%

expansion due to radiation damage to the fuel, he burn-up figure gives a==vinn=Puconcentrationof2000 grams / ton,butbecauseofthefast .

spectrumtheisotopicqualitycorrespondsto50gran/tonHanfordmateriale The pin fuel elanents are 0.125" diam. U - 25 2r wire diffusion bonded to 0.007" wall Zr cans. Each pin has a spiral rib which insures proper spacing of the 169 sub-assembly in a 1/16" pins constituting wall stainlessacan.

2 757he in flat across nata plate hexagonal design, which F. drop from fuel to coolant as has a much compared with lower temperature 105-2500 F. for pins drop,),(35-65 consists of loose plates of .0140" thick U - 2% Zr alloy sodium bonded in a brased radiator-like structure of stain-l less steel with the coolant Na formed through reactangular channels e

0.07h" by 0.2L" inside dimensions between the fuel plates.

In the pin design the core contains a total of 107 hexagonal sub-assemblies, of which 99 are fuel and 8 control; in the flat plate design there are 101 square sub-assemblies with 92 fue; and 9 control elements. In both designs

-l the =mv4=nm reactivity wtrth of one sub-assembly is under $1.00, i.e. less than sufficient to take the reactor from critical to prompt critical.

The regulating control and safety rods are all identical in design. They are made of boron carbide having a total reactivity worth about 3 times that of an equivalent volume of fuel. The safety scram time design objective is 0.8 sec. for co lete insertion. The rods can be inserted in the fuel channels with " clearance, are actuated through the conter of the rod, and are dumped by sodium dash-pot action. The total excess reactivity re-quirement is esti.nated at $0.65 including:

Temperature coefficient $ 0.30 Burn-up 0.16 Growthoffuel(15 burn-up) 0.19 6 0.65 The rod withdrawal rate is limited to about 111 per sec.

i I

The reactor will be reloaded once a week with a 6-8 hr. down time, at j which tins 8-10 fuel sub-assemblies will be changed (replace or change i position)and20blanketelementsremovedforreprocessing.

l l At the time of the two meetings with the er.bcommaittee the site chosen for j the reactor was on the shore of Lake Erie about 20 miles from the nearest Canadian shore line and about an equal distance from the nearest part of the lI City of Detroit. Subsequently it has been learned that an alternative site l inland has been selected tentatively but not yet acquired. No further i

information on this alta nate site is available at present.

l', i i

.-. . .- --.- \

\

( 3. Potential Hasards of APDA Reactor The fast converter representn a radically new design so far as remotors of this power level are concerne,d. The special features include the use.of Na as a coolant, the fast spectrum leading to a neutron lifetime of about j

10"7 sec., the use of a breeding blantet with a significant inventory of 1 4 plutonium, and the .se of a new fuel element design. Nevertheless, if the assugtions underlying the safety calculations for this reactor can be sub-stantiated, it appears that the chance of a hasardous incident is negligible  !

owing to the inherent nuclear stability of the design. )

i We list below h principal hasardous features of the reactor, together with a discussion of how they are met in h design.

a. The danger of a chemical reaction between sodium or sodium-potassium and water.

The steam generator, having the greatest hasard of this sort, is isolatedfromthereactorbymeansoftheIHI(intermediateheatexchanger) loop eqploying non-radioactive NaK, The steam generator lies outside the l

reactor containment, and the layout is understood to be such that a major NaK-steam explosion would not damage the reactor or cause spread of con-tamination.

b. The danger of a radioactive sodium fire within the reactor containment.

The reactor primary loops, and IHI are contained in a sphere de-signed to contain ho, pai, overpressure up to 6500 F., corresponding to a design stress of 15,000 psi. The primary piping is all blanketed under an 1

inert atmosphere,

, o. The danger of a prompt critical accident causing moltina of the core.

This hasard naturally suggests itself in view of the shortness of the prog t neutron lifetime. It is possible to operate the reactor with not more than 30.65 excess reactivity. This is possible because of the absence i

of %e poisoning in the fast reactor. Furthermore, the reactor primary coolant pool is never allowed to fall below 3000 F. even in shutdown con-dition, and the frequent, staggered unloading schedule requires only a $0.16 allowanoe for burn-up and pisoning. Moreover, accident calculations described below show that 41.h0 excess reactivity added at the rate of $50 per sec. does not produce a fuel temperature in excess of 7700 C. (1h200 F.)

in the initial power transient. This is just within the beta range. This ,

conclusion depends strongly on the assu q tion that the fuel maintains its I

mechanical integrity and yield stress, even after radiation damage.

d. The danger from extreme sudden increases in reactivity resultira from the addition of moderator to the pri - y coolant.

The accident described under c. would result from the addition of  !

g approximately one liter of hydrogenous material to the coolant, advancing into m

bum >* g P-we-w- e,w- - ,-- , -, ,w-_, -- ,,,w m- , -w,-n--ww- M

. p l

the core in a sharp front. The danger of such an accident occurring is  !

I minimised by the use of t'a intermediate loop and by the absence of hydrogenous material from the reactor bdiding. This is a matter which will (,

require extensive discussion in any safety report.

[

o. Molting of fuel elements due to loss of pr4== v coolant.

[

The design .f the primary system is characterized by double walls I throughout, so that rupture of the prinary system would not involve un-covering the fuel elements. The fuel elements will not melt under shutdown ,

power as long as they are immersed in sodium.

f. Radiation embrittlement of the primary system.

Although the fission rate in the fuel is such that its radiation damage can be extropolated from experience in thermal reactors, the fast flux is of the order of 100 times that available in any current reactor. The critical reactor vessels and piping are protected fro.m this flux by the blanket and a thermal shield, so that the properties of these elements do not e

involve extropolation beyond current experience. Some question may be raised about fast neutron effects on the supporting structure of the core.

This does not appear serious, but should be examined critically.

g. Production of power excursions by entry of sodium " cold slug" into the core.

Calculations have been made on the effect of introducing a cold slug at various rates. The temperature reduction assumed was h300 F., i.e. from 6800 F. to 2500 F. sodium inlet temperature. Even for instantaneous insertion the power excursion does not exceed h times operating power, and the maximum fuel and coolant outlet tegeraturo remain well under danprouc levels. These conclusions appear to be true whether or not the safety system j operates properly.

h. The failure of flow in the primary cooling system.

This type of incident has been extensively studied by means of simulator calculations. The flow is assumed to decay exponentially with

' various time constants to h% of its full-power value, this last being main-tainable by natural convection. The worst case calculated corresponds to instantaneous decay of coolant flow and 80 sec. scram time and still gives a

==vi=== fuel temerature well below the molting point of the U-Er alloy.

The situation is not appreciably worse when no scram occurs.

h. Discussion of Factors and Assumptions Underlying Hasards Calculations,

s. Doppler effect. This arises because the non-uniform spacing of fission and capture resonances in the fuel leads to self-shielding which

, varies with tagerature owing to the change in the shape of the resonances k-

• t r~

- . . . ~ . - ~ .

g by Doppler broadening. In this reactor we are dealing with the situation in which the broadening is large compared with level spacing. This situ-ation has been analyzed theoreticany by Goertzel and Foshbadt of NDA with the oonaboration of Bethe. The effect of U-235 is to produos a small

.. positive temperature coefficient of reactivity while that of U.238 is to i produce a negative coefficient. The theory is much more reliable in giving the relative importance of the effects in U-238 and U-235 than in giving ab-solute magnitudes. The theoretical values for the actual composition and

~

structure of the reactor ares, ,,

for2540;$6x10-f/*C for 28 -1.36x10- /9 t a,shumptions. These with'a figuresnotare effect of zero being to be compared with apossible for the wor'g/0 C, computed for value of -5.36x10"0 longitudinal expansion of fuel. Even if the Doppler effect turned out posi-tive it would have to be substantiany larger than estimated to influence the computed stability of the reactor. It is to be emphasised that an these numbers are theoretical, however. Some check on the doppler estimates may be obtained from experiments now in progress on EBR. Here the estimated

,)

positive effect is much larger. ,

b. Integ*ity of the fuel elements. An important safety feature of the design lies in the use of fuel which is mechanically continuous for the full length of the core. This insures instantaneous reaction of the part of the negative temperature coefficient which arises from longitudinal expansion.

which begin to be signifi-This statement is true neglecting internal effects,$20 cant at reactivity insertion rates of the order of per sec. Inertial stresses in the fuel do not exceed the yield stress for insertion rates belou G50 per sec. Indications are that these assumptions are valid even after irradiation of the fuel to 1 percent burnup. ,

! c. Validity o_f the ten.perature coefficient due_ _to thermal expansion.

This win have no be checked experimentally before the reactor can be operated.

, Not only are the calculations themselves subject to some uncertainty, but also there is the possibility of unknown mechanical features of the reactor which give added temperature or power coefficients and whose contribution could a

conceivably be positive in some design circumstances. Consideration should alt be given to means for testing the longitudinal fuel coefficient indepenWntly.

d. The assumbtion that the excess reactivity required for operation will never be exceeded. The reloading scheme for the reactor involves loading fresh fuel and moving' fuel elements with an the safety and control rods in-sorted in the reactor. Thus it is theoreticany possible to overload the reactor to the point where it would go well over prompt critical if an the i

rods were withdrawn. On the other hand, calculations show that if the rods are withdrawn at their mechanically limited rates, the fast-acting temperature coefficients of the reactor win prevent any serious power excursions provided

- withdrawal is stopped by operation of the safety circuits. The normal excess reactivity of 00.35 over hot critical can be handled by the longitudinal fuel

( .y

_ . . _ . _ _ _ _ _ _ n

. .. . L

,t l temperature coefficient alone withmtt molting of the fuel, but if the rods continued to be withdrawn without any other action the neutralising negative coefficient would eventually bring the reactor to equilibrium at tempers -

tures in excess of the iaolting point of the fuel. Under these conditions, however, boiling of the sodium would occur first, resulting in additional negative reactivity effects. However, a real hasard may still be presented by overloading and this should be studied, ,

s. The primary container and piping' retain their integrity. Type 3l47 stainless steel has proved vulnerable to stress oracking in steam A ant ex-perience. As a result of a careful r6 view of the situation, the A*DA group has decided to use type 3014 stainless. The themal cycling of the primary system is restricted by the temperature drops and operating procedures .

adopted, and the operating design appears to be conservative in this agard. l

5. Ocuments on Desirability of Prototype or Safety Tests. l

~

The APDA group has presented to the Subcommittee tentative cost estimates )

for the construction of various corts of prototype tests. For a prototype .

d of 20 percent power (one loop instead of three) a total cost of fourteen  !

1 million is estimated. This is to be compared with fifty-four udllion esti-mated for the complete project. For a borax-type experiment not operated at power the cost would be about ten million. It is debatable whether enough could be learned from a prototype experiment to justify its cost. The borax-type experiment - to be designed, constructed, and operated with AT, funds -

should be studied further. However, such an experiment would probably be justified only if the results could be generalised for a variety of fast re-actor designs, and this matter needs independent examination.

6 Conclusions Recornendations and Questions.

i* Since no formal project has been submitted, the Subecessittee can only present very tentative conclusions at the present time. These are listed below.

a. Should an explosive incident occur this reactor presents an unusual hasard because of the high plutonium content. On the other hand, because  !

I of the stabilising effect of the prompt negative temperattre coefficient, an incident involving release of fission products appears virtually impossible provided the assugtions of the safety transient calculations can be sub-stantiated experimentally. A program for this is suggested in b, o, d, e, and f below.

b. The doppler theory of NDA should be applied to the E R geometry,

o. Oscillator experiments on EBR at high power with restricted coolant flow should be cogleted and their interpretation fully understood in the light of b.

.6-(

._ l l

>~> , - . - .- -  : t w ,_ _-- - - - _ - . .- _ -. . . . _ -- _, - JL

> c

d. Differential experiments on U-235 and U-238 to look for holes in W neutron transmission at intemediate energies should be undertaken to

.I.. check h asswqptions regarding level spacing which go into the prediction of doppler coefficients.

e. A direct measurement of the doppler coefficient of mi.tures of 25 f and 28 (Frost-type experiment) should be undertaken in the Argonne fast critical.

f. Having checked W ory to the greatest extent possible by b, o, d, and es if W same theory applied to the APDA design ves a positivo doppler coefficient less than one-third of the calculated or measured ocefficient due to the longitudinal fusi a::pansion, then the assumptions of the safety calculations appear warranted as far as nuclear design is concerned. '

g. Isothemal . temperature coefficient measurements must be planned on -

W fully loaded APDA reactor before start-up. An effort should also be made

• to verify the calculated longitudinal fuel coefficient in the Argonne critical.

Should the coefficients be substantially less than predicted, operation of ,

the reactor might have to be postponed until a remedy could be found,

h. The loading procedures should be re arm =ined with a view to providing g greater assuran:e against accidental overloading of the reactor with the safd.ies in position. A positive mechanical means of preventing overload, as for example by special keying of core and blanket rods, is to be preferred.

i. The effect of radiation damago and of thermal cycling on the ooefficient of expcusion of the fuel end fuel subassemblies and on their yield stress must l be verified by seperate tests in order to substantiate the asetaptions of W safety calculations. .

j. Ostalled evidence concerning the effect of the high fast neutron flux -

on the mechanical integrity of the reactor core and primary containment and i piping should be obtained and presented. l

k. The possibilities for introduction of moderator to the coolant, and a

' realistic evaluation of the rate of reactivity addition and total reactivity i should be made.

1. Full scale mock-up tests of the control and safety system should be made and reported. i

m. The subcommittee sees no intrinsic factor in the design at present  !

which makes it unrcasonably basardous at h proposed site. We do not believe  !

the small generation time alone constitutes such an intrinsic hasard, and provided the reactor can be shown by positive experimental and theoretical evidence to be sc3f-stabilising to the extent assured, we believe it may eventually be approved for operation at this site vi h ut a prototype test at a remote sito.

n. We believe it would be vise to give serious consideration to a safety experiment for fast reactors in general at Aroo.

(

o. A satisfactory plan for inc,uring the presence of an adequate source in the reactor during shutdown should be presented. 1

-o, -  ;

. . . , , rt