ML19224A918
| ML19224A918 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 05/01/1979 |
| From: | Persiani P ARGONNE NATIONAL LABORATORY |
| To: | Lawroski S, Till C ARGONNE NATIONAL LABORATORY, Advisory Committee on Reactor Safeguards |
| Shared Package | |
| ML19224A917 | List: |
| References | |
| NUDOCS 7906130205 | |
| Download: ML19224A918 (4) | |
Text
{G ' '
[
... /MiGO N N E
~
N ATION AL INTR A-L ABOR ATORY MEMO L-A8O R ATORY May 1, 1979 TO:
C. E. Till, Director Applied Physics Divinion S. I.awroski ACES, Chemical Engineering Division FROM:
P. J. Persian / /gw Applied Physics Division SJEJECT: Three Mile Is.and Incident The purpose of this ce=orandum is to su==ari=c some of my thinking on the requence of events relating to the fuel element behavior and the consequent core conditions one nay expect. Although the following proposed core damage codel was based primarily on the ther=al-hydraul'ic conditions reported in the early public infor=ation releases, the consistency of the codel had to be explored with the occuple readings obtained shortly af ter the rc= oval of the hydrogen gas bubble.
The thercoccuple data of April 10,19 79 (furnished to
=e by Steve Lawreski), with the core decay heat at the level of about 5 MW were used in these esti=ates.
1.
Initial Loss of Coolant, The suggested codel assu=es that at the initial voiding of the core proper, the circalloy clad fractured under fission gas pressure and under locally induced ther=al stresses (radially and circu=ferently), thereby releasing the volatile fission products. The fractures occurred with the clad still in the rasilient (circalloy) state.
Consequently, the degree of dis-place =ent and cove =ent of the split clad is limited by the =utual interaction between adj acent similarly aff ected fuel ele =ents.
This cutual interference introduces constraints to the split clad notion and in effect =aintains the fuel pellets in a somewhat " quasi-columnar" distribution.
The local coolant flow areas although dramatically reduced in sore regions, were always adequate to allow ef fective cooling by the superheated steam.
That is, in cost of the regions the surface to volu=e ratio for heat transf er did not approach that of " spherical-type geocetries" with low surf ace to volute ratios which would have resulted if rest of the fuel pellets had tumbled into a heap.
2.
Oxidation of Zirealloy The oxidation of the =ircalloy procccded in-situ on a ti=c scale of hours consistent with the time period of core exposure.
The oxidation reaction, although rapid, was not " violent." The esticate on the rate and extent of the from the reported hf regon gas bubble size, d
oxidation is based in part pressure, and temperature condition, 1200 cu ft, 1000 psi, and 280*F, respectivelv (230 lbs of hydrogen).
It was also reported that hydrogen was released into the containcent and auxiliary buildings (2,100,000 cu ft) at concentration levels of 2 to 32 at STP conditions (230 lbs). The totsi urount of circa11oy oxidized 11 ) to be approxi=::tely 12,700 lbs (5.3 x 10 g) is then cocputed (91/4 x 2
which represents between 25 and 30% of the total circalloy (44,000 Abs) originally in the structure.
174 0 %
7906130205
-.w.,
(
4 Assuming the oxidation process occurred over twice the above 307. of the niad surf ace area (49 x 106 2), the weight gain of the circalloy (32/91 x cm g) per untc of area is estimated to be 0.063 g/c=2 5.8 x 10 g = 2.03 x 106 6
oc 6300 cg/dn'.
Referring to the attached curves on the oxidation of circonium, the 1 to 2 hour2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> period of cxidation would have been accomplished at circalloy te=peratures in the range of 1400 to 1500*C, with possible higher clad and fuel pellet temperctures in the hotter regions of the core.
With the melting te=perature of the :1rconia (2700*C) being close to that of the uranius oxide pellet (2800*C), the highest tceperatures would still be much below the celting point.
3.
Coolant Flow Area Reduction The extent of the core damage consistent with the proposed codel can be scoped by est?:ating (from ther=occuple data) the reduction of the coolant flow areas in the fuel channels.
Frem the core pressure drop and heat flow equations:
V2 (1)
AP = Tc A and Q = pC AVAT, (2) p the thermal hydraulic relation between fuel channel regions 1 and k is given bT
- 2/3 (3)
A /Ak" S /9k
- OT f0?1 g
i k
In the above equations; A = coolant flow area, V = coolant velocity, p = coolanc density, Q = heat generation, and AT = temperature change in the channel.
The ther=occuple daca applicable for this estinate are those satisfying the follcwing core conditions:
(1)
The resistance to coolant flew should result only from the reductica in flow area; and (2)
The ecolant should be in a liquid state.
It is assumed that these conditions were established af ter the hydrogen gas bubble was recoved and the system repressurized above 2000 psi. The representative thernoccuple data presented in Table I, were selected for regions in which the decay heat generation could be raticed consistent with the operating radial power distribution.
!74 097
.(-
_a_
TABLE I.
TMI April 10, 1979 "'hermocouple Data Enlet Coolant Temperature 280*F Representative Temperature Heat Cencration Area Reduction Core Region
'F Q
A /A f y 1.
liinimu:n or no 282 1
1 n:n,ge 1.
Ibdest Da= age 286 1
0.48 3.
Moderate 305 0.30 Danage 4.
Most Da= age 400 3
0.14 The range of crea reduction for the clave set of conditions are found to be about a facter of 7 for those regiona with the highest (400*F) ther=occupla readings, about a f actor of 3 in those regions with coderate (303*F) te=perature readings, and about a factor of 2 for those regions with =odest (265'F) tc=pcra=
ture readings. These estimates sec= to fndicate that one should expect to find the disposition of coct of the fuel ele =ents in a " quasi-columnar" configuration with cociant flow f eatures screwhat similar to the initial flow characteristics.
That is, the clad will be split, distorted, and exidized, but the general characteristics of fuel-clad-coolant area had been caintained.
Several daannels =ay be = ore severely codified than others with respect to coolant flow areas but not completely blocked.
Soce clad particulate spalling =ay have occurred when the c=ergency core coolant rcentered the fuel channels and interacted with the highly c=brittled zirconia. However, it =ay develop that the spalling was not extensive encugh to severely alter the general " quasi-colunnar" distribution of the fuel pellets.
It would be instructional in understanding under these abnor=al transients, the relative dif fusion behavior of the volatile fission products 133 134Cs and iodine, xenon and cesium, if the relative concentrations of Cs, 137Cs were =casured in the coolant and in the depositions f ound in the con-tain=ent structure.
As a final thought, the above expectatica of the core configuration =ay not be unique to the ther=al hydrat lic data.
e 174 098 O
('.
I 9600 ESCO 8000 7200 5400
'm 5600
=
STEAM 24 HOURS b
4800 G
E=
4000 3200 STEAM 1 HOUR 2400 1600 800 l
f 700 800 900 1000 1100 1200 1300 1400 1500 1600 TEMPERATURE,
- C OIIDATICN OF ZIRCALLOY 1/ 4 Cii 9 e *
( s.
t.
-S.
~~
PJP:mlb Distribution R. V. Lancy J. A. Kyger R. Avery J. Honeka=p M. Groltes R. A. Sevy L. LeSage Y. Chang Kalicullah V. Maroni A. /corosi E. Monson R. McConnell L. Burris D. Webster M. Steindler
- 3. R. T. Frost R. S. Zeno e.
174 100
__