ML19220B585
| ML19220B585 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 04/13/1979 |
| From: | Meyer R Office of Nuclear Reactor Regulation |
| To: | Mattson R Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML19220B586 | List: |
| References | |
| NUDOCS 7904270057 | |
| Download: ML19220B585 (27) | |
Text
{{#Wiki_filter:l e d 8 S, UNITED STATES 3% 4 NUCLEAR REGULATORY COMMisslON ~ l 3I k.h WASHINGTON, D. C. 20555 APR 131979 MEMORANDUM FOR: Roger J. Mattson, Director, Division of Systems Safety, NRR FROM: R. O. Meyer, Reactor Fuel Section Leader, Core Performance Branch, Division of Systems Safety, NRR
SUBJECT:
CORE DAMAGE ASSESSMENT FOR TMI-2 Attached is our assessment of the core damage at TMI-2 for use in the SER for natural circulation. It represents our independent evaluation of the facts available and of the industry / vendor / licensee analysis, which we have heard in several briefings. An earlier estimate of fuel damage was made by Rubenstein et al, and a recent meeting was held at NRC with industry experts. Memoranda describing those evaluations are attached to this document. 'y Q 4 ' ~ Ralph 0. Meyer,/Section Leader Reactor Fuels Core Performance Branch Division of Systems Safety
Attachment:
As stated I cc: R. Tedesco K. Kniel P. Check C. Berlinger D. Crutchfield (FOIA File) D. Houston M. Tokar V. Stello B. Grimes G. Knighton J. Voglewede D. Powers PDR 7 9 0 4 2 7 0 0 5~T gg g 80-207 io6
/ ? CORE DAMAGE ] A. Introduction j For the usual analysis of hypotnetical accidents, initial core t i conditions are assumed and consequences are calculated. This would I involve complex thermal-hydraulic calculations and fuel behavior analyses. i At Three Mile Island, however, scme of the consequences are known (i.e., i' some information on fission product release, hydrogen generation, and instrument readings is available), so we will use " reverse engineering" as our principal method of backing out an assessment of core damage. i We start with the assumption that the core was uncovered and allowed to heat up for significant periods of time. Figure 1 shows the system pressure history for March 28, which includes three periods of significant uncovery. The periods of uncovery correspond approximately 3 with the major periods when system pressure was belcw the saturation pressure. We will assume that the first core uncovery began shortly after 92 minutes into the accident at which time excore ion chambers sFow a response 4 I j spike corresponding to the loss of water shielding. Althouch the two later periods of uncovery may have produced additional core damage, we will focus en the first period because decay heat was larger then and because that period produced the large radiation instrument reading (at 150 minutes) in the containment indicating major fuel damage. Because the fuel damage to be discussed below is so extensive, we will 80 208
i {[ e \\f !! ! Il 5 1 )g e 4 L 1 t o H, 3 \\ F \\ 1 E R0 } U2 S6 S 2 E = 1 RT P ( NO I 1 N\\ \\ T 1 A I R l UA] T E R 0 U S 1 S S D E E R M P US 9 A5 S S CR k 8 7 6 l 5 l l 4 _\\ \\g f l 3 \\ f 2 1 I 1 ER G U k G _{\\ f l 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 5 3 l 9 7 5 1 9 7 5 2 2 2 2 1 1 1 1 2c Nc ) ( ga ig#
i 4 j j conclude without demonstration that virtually all of the fuel rods in the core failed in the sense of experiencing defects large enough to release j gas. Furthermore, the rods probably failed by a LOCA-like ballooning-and-rupture mechanism. Because of the massive oxidation that followed, the mode i of failure is probably irmaterial. t ~ As a poini. of reference, Table I lists melting temperatures of the various materials used in the fuel system. B. Fuel Rods Fission product and hydrogen measurements at TMI-2 give important clues about the condition of the fuel rods. We will deal with fission product releases first. t Air and water samples containing fission products have been analyzed. While we have analyzed both for indications of fuel conditions, we have concentrated on the Xe-133 concentration in the air sample. This isotope was selected for analysis for several reasons: (a) it is a noble gas and will not react, plate out or condense, (b) it has a relatively long half life (5.29 days) and a high production rate (6.8 atoms per 100 fissions) and therefore will be abundant thus reducing measurement errors, and (c) fission product release correlations are much better established for noble gases than for other fission products. I Bettis (BAPL) has evaluated the Xe-133 activity and concluded it is equivalent to 31% of the total core inventory. We have independently checked this calculation (but, of course, not the sample activity) and agree (31.5%). 80 210
j. fi i. I 1 TABLE 1. ELTING TEFFEPATURES FATERIAL TEFFEPATURE, C TB PERATURE, F i lD2 2805 5080 ZIRC-4 1850 3362 Za0 2715 4919 2 INCONEL718 .1260-1286 2300-23 % 4 304SS 1399-1421 2550-2590 AL 0 -B C 2030 3686 25q As-In-CD 800 1472 UO -Go2 50* 2750 4982 2
- Two fuel assembites contained gadoliniItest rods.
l l 80 211 ._ ~.. ~. -
f .E s i e Fission products including gases are normally retained by the U0 2 pellets. -A normal pellet release to the fuel rod internal voidage is only-1 or 2% (even for a successfully terminated LOCA) so that a 30% release indicates additional release from fuel pellets not just a release of the gap activity. t Fuel pellet releases are strongly dependent on temperature, and Figure 2 shows a correlation of release versus temperature for Xe-133 (from a recent ANS-5.4 draft standard). The correlation, however, is for steady-state releases and we are dealing with a transient. Further errors are possible because of kinetics changes due to oxidation to U 0 orU0. Nevertheless, it is a reasonable 4g 38 approximation and is consistent with recent short-time annealing ex-periments (private ccanunication 4-10-79, R. A. Lorenz, ORNL) and earlier annealing work (G. W. Parker et al., ORNL-3981 - See attachment A). Parker heated irradiated samples in a furnace for 5.5 hours. The samples had burnups ranging from trace to 4000 mwd /t (about the same as TMI-2). Parker measured releases of about 5% at 1600 C,15% at 1800 C and 40% at 2000 C with an uncertainty of about a factor of 2-in release. These experimental releases for conditions roughly similar are h to TMI-2, but for different isotopes,^ fair agreement with Figure 2. 80-212 Using Figure 2 we could conclude that (a) the fuel was uniformly heated (uniform in axial and radial directions) to about 1750 C, or (b) 30% of the fuel melted while 70% remained below 1200 C, or (c) U any intermediate condition existed. Because of the core uncovery sequence,
oao n i i i. _8wn M e i o (.D [ W J 8-LJ ga x es ~ (f) u t C _8 CD 8g e 3 z ~. E-. c CD e w -! E (D U g (n ea r C W ta H H g It-7r N., y >y eW M @(n \\. -8 { M w$ ~' y_J ~ l U \\. O e -? X if C i, C i O W Z Z C O o gC 1 -8 L m g i CN g-L a w 5 e e 8 tr 1 0*0 9*0 PO 00 0 3593138 19N011098J 80-213
G the fuel rods probably did not heat up uniformly in the axial direction. It is reasonable, however, to treat the fuel rods as isothermal in the radial direction because of the low heat flux. 4 Figure 3 illustrates this point with a comparison of a full-power radial temperature profile and a decay-heat-power temperature profile. There are physical limits on how hot the fuel can get during the periods of core uncovery because the fuel rods have a large heat capacity and a low heat generation rate. If one assumes zero heat removal (this would produce the most rapid heatup rate possible) during the first period of uncovery, the heatup rate is still fairly slow. Figure 4 shows the adiabatic temperature increase with time for the peak-power axial location, for the low-power ends of the rods, and for the average location. Since there must have been some heat removal thus further slowing down the temperature rise, pellet temperatures probably did not reach the melting point. Figure 5 shows the temperature changes with time for a surface heat transfer i 2 coefficient of 0.5 BTU /hr-ft - F, which is a very small value. The results on temperature distribution are, therefore, not conclusive. It is unlikely that fuel temperatures were uniform and no lower than 1750 C, and it is also unlikely that any fuel (UO ) melting took place. 2 The fuel, however, did get very hot compared with its normal operating temperatures. Oxidation of Zircaloy by steam and the attendant decomposition of wa'ter provided the major source of hydrogen in the TMI-2 vessel and containment. 80 214
RADIRL FUEL TEMPERATURES a. S-G d o Ej8~ U E' E 8-u+ E l9 o = = = = = = = = = a 8-LEGEND o - PRE-RCCIDENT 1.5 HOURS POST-RCC10ENT (CORE COVERED) o-cr o.0 o'. t o'. 2 o'. 3 0.4 0.5 RADIUS (CM) cn FIGURE 3.
A 3 I B -S i s ll O_ -g 3 tbb H sss g 222 C i, Ex1 m cn u, c met i 8 8 O aa4 e _g 5 ~ HT co C H m Ox -R -S W $C s amt m ecce exst ccm exa o (3*EJ N M 80~21G
e m n i i bDH i C Id -S r CUC o E EEE g M 98? l ~g s,., m _n - C wa M C i: b __3 .g tr C Ed Z -9 u; 13 c M i 0 (O M3 N M 80-217
f The Containment Systems Branch has estimated the amount of hydrogen present in the plant (Attachment B) after the periods of core uncovery that caused fuel damage. They included amounts (a) consumed by the hydrogen explosion (226 lb mole), (b) remaining in containment after the explosion (80 lb mole), and (c) in the primary system bubble (76 lb mole), which was corrected for radiolysis. Comparing the above amounts with the total amount of hydrogen that.could have been produced if all of the Zircaloy in the fueled region reacted with water, we get 41%. As with the temperatures, an ambiguity exists. This could mean that (a) about 40% of the cladding wall thicknese is unifomly oxidized throughout the core, or (b) 40% of the fueled region of the core has fully oxidized cladding, or (c) any intemediate condition exists. Figure 6 shows the time required for total wall thickness oxidation as a function of temperature (Cathcart-Pawel correlation). It is clear from Figure 6 that complete oxidation is possible in cladding segments that reached temperatures of around 2000 C during the period of core uncovery. It is also clear from Figure 6 that all of the cladding did not experience sustained temperatures of around 1750 C else it would all have oxidized. This is further evidence that fuel temperatures were not unifam throughout the core, and that temperatures locally were very high. Based on early estimates by the Analysis Branch of core uncovery, wa will assume simplified uncovery histories shown in Figures 7 and 8 for the 80-218
TOTRL OXIORTION TIME El EP ~., TIME FOR TOTRL CLADDING REACTION AS A FUNCTION OF TEMPERATURE g.
- 9 u
E-. D ' Z ~ { M N. b8- \\ e-. ~ e S 6 .g=.,. g O. C w 1000 1100 1200 1300 11OG 1500 1600 1700 1800 1900 2000 TEMPERATURE (DEG. Cl w Ficune6.
CORE UNC0VERY CASE 1 ~ ?- ~, l .\\ l o -n- ~. (n - La l I 0 Zo l .-. o-m _1 l id gin 8-l i ~ t-4 Z (r o. __I m l O o 's a tJ o - (r
- N.
o l o ~. ~. : CZ) o. 9 1.5 2 2.5 3 N POST-SCRAM TIME IHOURS) No FIGURE 7.
n C\\1 L_1 m -" <n g-- O aor 1 w r K ~s L1 r c-O x O C.) en z i --) s in a n_) - n c. x O O co LLJ l m -u_ I I l 8 I I I Ot1 021 001 CB 09 Ok 02 0 (S]HONI) 13A31 INW1000 3800 80 221
/ [' following calculation. Fuel that is covered will be considered to be 1 cold (i.e., no cladding oxidatien). Fuel that is uncovered will be allowed to heat up; fuel that heats up will be given a heat transfer l coefficient that is adjusted such that the total integrated oxidation is 40't. These calculations give the oxidation distributions shown in ~ Figures 9 and 10, and these distributions are insensitive to many of the assumptions that were made. Figures 9 and 10 thus are more probable distributions than 100% oxidation over 40% of the core or 40% oxidation over 100% of the core. Figure 11 is a recent best-estimate embrittlement correlation (Kassner et al., ANL) that shows high-temperature fragmentation of quenched tubes at about 30% oxidation. Using this correlation, Figures 9 and 10 indicate that a fragmented regiun of about 5 ft. in height exists near the top cf the core. It may well be right at the top of the core as a result of simplifications in our analysis. In any event, at least 4 to 6 ft. of intact (but partially oxidi;ed) fuel rods remain standing i at the bottom of the core. Figure 12 shows fragmented Zircaloy cladding after oxidation in a simulated-LOCA test. Kassner (ANL-78-25 and ANL-78-49 reports that at high temperatures (> 1250 C) many fragments are produced whereas at lower temperature the rod may simply break into two pieces. Inasmuch U as TMI-2 temperatures were higher than 1250 C and oxidation was severe, small fragments of the size shown in Figure 12 should be expected along with larger tube-like pieces. 80-222
CORE OXIDATION CASE 1 ~ I l l t 1 I I am l t. (J d - l 1 N O I 1 ~x l \\ a CD R_ l t zo Q l t. O e G~ J e O l e b a-l r b l N / e o l t. et n l fr d - l tu .i ~. 3 m. - ' ' ~ ~ ~,,. o o-0 1 2 3 4 5 6 7 8 9 10 ti 12 i N CORE ELEVRTION (FEET) N G3 Ficuae9. O
O 4 S CORE OXIDATION CASE 2 8 0 s O I ta d - N O. 8 O C.D R ' l 'I Zo I. O 1 Ox J O l l Ia-Oo 1 Z O I. ~ b .e 3 o TN e' i d~ 'e tu y** 't*w,
- o.....
I a i I a s 1 I a i 5 T 0 1 2 3 4 5 6 7 8 9 10 11 12 9 CORE ELEV8 TION (FEET) y FIGWE 19. 4
8 E l I I I I le i l I w O O z He e e as e 8 - 8 e E ed o e a e 8 -z z e e o O r< 0 0 ~ 2 8 d= e %o 3 m e m z a m s = =z gg tl6C# 0 0 z g a H E g 5 88 ke_ @ E e se g 8 o 2a
- o0 O
E g e aN e 5 Es 'WilP 3 G " z o a ez3 g d 5 85 %l_ g -O m s $z y m o e g 6 9 QG [OE g 4 z< m O e m d 8 H e a 2 g J,-- e 2 d 8 gP O qgpo 8 G g e s2 8 o a 0 p O m p a0 0 T h oc O <z a O C: e t QOgh o $C x e C N 6 O gg r O 000 Of O O e lg e 2 O e O w 8 E s %p8cco o ~ Jb a O Oo O r;o ~ g OG3 O G 1 O 8 w i N e g l l l l l I l ll l _O
- 8 8
8 o 8 8 8 8 8 e o (%) G310V38 DNIGGVlO IN37VAinO3 80-225
6* ( '. . l-s ~ 3.r.. m m^--- ~ %,t 4- ..=...._y e - r;.- 32 ,,iI.:t .e. 2
- 3.,
rs >-g
- E.. - Cm
,. r. 4,c3,.. ~ r-t - ~., -..+;. - ri -
- .i l !-3 _. _.,e
,. 4.. - - e.: i, s w .; e t:: .j_ a _, u _.. _3 g. u e. rr!! ;a. - t: 1 r_+ :^ '.m-". . " ' /- r ~ e - m. r-c. . m h n 11 r e ,o 3 t i r; - g , _.m.. r, C* *g *3 .,.. ~. _, pg. n WW_TC 55 ' ".1 li Q, I Q. e-t -- ?- TC#6. ~ l1%.. ~ y -g ,-._ p! w... _ a E2. T ' g f. ~ ;53, ' ;, "[- 3 a
- -3 ;-
r cc ? L3 sd m L g4 'I i A%g L' V c,a *. ' "T"I -- g g.,=.3 - ' . e '31 s' -313 1. s, 4" ' m F- - Eg. sg
- f. _R'.: '.~~ ' t
' Y. d.,, W ' l IN-. ~~ i . rc. ,y 4' 3 - 4 "- - r4, p +u w L3.2 u y[,.. a c L a_ . =f.~ .'. - f , ~~. 6... t ^ ~ - - + - .t. 7._1
- a-r., _.;;.
s w g3 r j
- -3
-e..: b.,- 3~ L .,,".,.2 ', ..~.h,;3rk:.~ 'hj h _"L. - '. [' Y[i- -..z.. m,v. - e. 4.#2 g ie g 31 V 1 .. 42 ; -.- ~ L: n: ~~:,n a 24 n 4 -':P~- %n.nm, ..c. . cl. h. - -- u .~-# i. L: 3 kh.~m$._ - d-r E 'l' 'ChN.y<.v~:<.x.N-hN. b) fi'..x.,,r ,... - ]. = l.,, e w -; o z: - H.-m. ~ -g.m ;y=.,. y-g ya.. _ ..... _, 9 ,sn ~..- . _ m _. -c e Fig. m.12 Fig. m.13 Zircaloy-4 Cladding after Thermal-rhock Fail-Zirealoy-4 Cladding after Thermal-shock Fall-ure Showing Location of Thermocouples That ute Showing Location of Thermocouples That Produced the Temperature-vs-Time Curves in Produced the Temperature-vs-Time Curves in Fig. m.10. ANL Neg. No. 306-78-223. Fig. m.11. ANL Neg. No. 306-78-224. FIGURE 12, ~ 80-226
- ar
.e-w=mee.--%o.m.---
- -. = =... +
- -. -. - - - = = = -.--.m.-- -. = =. = =. -
W Fuel pellets normally crack during operation and crack healing can occur at power. Figure 13 is a typical example of a cracked pellet. Quenching during core flooding may also promote fragmenting of the pellets. Severely fragmented regions are comonly seen in fuel pellets as a result of extreme temperature conditions in test reactors. Powdered regions in fuel pellets have also been seen in some PBF tests,-but these tests are characterized by very high pcwehs (> 20 kw/ft) and very steep temperature gradients unlike the low-power unifonn (radial) temperature TMI-2 fuel. Therefore we would expect the TMI-2 fuel to be in millimeter-size granules and larger pieces including whole pellets. C. Unfueled Comoonents (Control rods, guide tubes, etc.) Figures 14 through 17 show the control rt's, the burnable poison rods, the power shaping rods, and the central instrument tube. All of these rods and the instrument > tube are inserted into Zircaloy guide tubes in the fuel assembly. The materials of which these components are made are indicated on the figures. An important clue about the condition of unfueled components is provided from instrument readings. The fact that all 52 thermocouples worked med throughout the accident and continue to give credible information 3 suggests that a central tubular structural member survivied. It is tempting to conclude that all Zircaloy guide tubes also survived, but this may not be the case since the thermocouple is well protected by multiple barriers. 80 227
e i 1 7 *Laram m -asur m2x_ ,r. w 4 1 ~ .jkMT, y ~9 c g g! Q '3 ,unw&1 sw[#-:,(yy3. 3,,st ifw::.s..v . s 5: 6 M.t ? p) .> Q p .c ng;r v a, , e,.iv,D c N s k - d E 'y = ~ *7 Y) .:e. '. ~ ..Q. 4 a .h 7 $h N f % N 1k b(?$.@i?) +- s:ws.ege. w 1._.
- .; %..., x n. :w g p,,,5
, cac ~ i 5 a y. $y,, 1 k b [A, '"E Ohb$h $+ .&a,,~w &b., .._ yMn. a r s W,: a =m -1 e j'" _ l - l :y l r i- ~ y:...;.'- _>..;. 4 ; -C. y .v O =~e, f g pe i *
- s. -{
- l N~s
&Y ^' a f.: Q _ . v f ~, W 0 80~228 FIGURE.13.
I ~ l R 1 N!\\ I \\ COUPLING x - { 3 O$ -.6 $ m J r n I SPfDER ( S S Grade CE3M) ,e c f 3 - ~ +. -+ +. ~~+%, l ~ .m '~'%- -- C. 0 2 l l*,,, ,/ 3 0 4-SS x I \\ (0 44O A. 0:D.[
- )
l
- J
\\' O $~ 'O C
- s
? TCP VIEW I N.~ +. l ~' ~ ~. ~ - ,X. (, u ide. '7~u L *r ~- a - e. NEUTRCN ABSCRStNG MATERIAL 'j g +. O. Y 9 7 J.0. x o. f70 o D -{ TLA 7 IT7., - C Cd) SC 729 CONTRCL RCO - - - ~ 4: kte t-Control Rod Assemb t . FIGURE 14. \\y' y y y yy
- ~
j' x h COUPLING s I (3o4 51) 'N I t s .1. SPIDER _(.s3.s,e4cneyt l t t l [ c 10 1__ ~+-+_ \\.,7, g -+-+ / J (0 / i *s.) (,f g Oxs 0 0 4 TCP VIEW b_,.-. v w a.1 da-<-l ~> If. Nw3 NEUTRON ABSCRBING MATERIAL (3AA 7e/KZr.,, S' C)) ( 3 r. /.. / u y f 4 ) 80-230 ~~ ~ AXfAL POWER SHAP!NG RCO 5 l$ to d;;UI Axial Power Shapin ,I i Rod Assembly M .T UD F wuas 15. (jf1 = / V
I COUPLING -(704__SS) [ i se r. ,/* / C SPIDER F -bb S f
- dt.
C. F ] \\ g 2. BURNABLE PCISCN E*V) I 'l d R00 Z r- + b4 T /,pi / r n / b o gu,.) l l kf (0.410 00) C /e )) ~
- g. -
.g \\M / e' ~( tj ,/ ki, a e 1 TOP VIEW ~ ~ ~n,_ v / I BURN ABLE POISON- - W ATERIAL p w +C'~Alg0) S 3 i.: V 80~231 I \\ m 7f Freuas16. O i Burnable Poison ] ] - \\/ \\/ Rod Assembly 4
Fixed SPND assembly cross section 7-INCONEL OVERSHEATH / CAllBRATION TUBE hl BACKGROUN.D ihi DETECTOR ' D_ \\ ~ hI SELF-POWERED NEUTRON DETECTOR (7) h! o \\ THERMOCOUPLE (1) Fleuae17. 80-232
Jf INEL has made calculations of guide tube temperatures by parametrically varying heat transfer conditions (see Attachment C). Their results show that guide tube temperatures lag the fuel rod temperatures by only about ?0 F. Babcock & Wilcox has perfont.ed similar calculations and concluded that there is a much larger spread in temperatures. We believe the INEL calculations are more nearly correct and that temperatures of unfueled compcnents were close to fuel rod temperatures. Since fuel rod temperatunes are believed to have exceeded 17500C in the hot region of the core, then in that region (a) Ag-In-Cd and its stainless steel cladding would have melted, (b) Inconel spacer grids would have melted, (c) Zircaloy guide tubes would have oxidized, and (d) Zircaloy cladding of the burnable poison rods would have oxidized. In the cooler parts of the core below about the 4 to 6 ft. elevaticn, we would expect all unfueled components to be intact, althouah perhaps damaged, just as the fuel rods are expected to be intact. Control rod segments could have only fallen about 3 inches if reiccad by melting in the hot region, and the Ag-In-Cd absorber should be in place tecause it is an insoluble metal. Although the burnable poison rods muld also be expected to be in place, their poison is probably lect; boron is known to leach out of B C-A1 02 3 pellets when exposed 4 to sater in a radiation environment. SG'233
dY O.
SUMMARY
Many or all fuel rods may have ballooned and ruptured, but this mode of initial defecting is probably irrelevant in 1-ight of later e more extensive damage. i F 7 In the hot upper central region of the core, fuel temperatures probably exceeded 1750 C releasing large quantities of fission products; about 30% of the total core inventory of noble gases was released. About 40% of the Zircaloy cladding reacted with water. Tais region of severe oxidation was localized above the 4 to 6ft elevation and may not have included peripheral bundles. The severely oxidized fuel probably fragmented into pieces ranging from millimeter size to whole sections of rods. The temperature of unfueled components lagged the temperature of fuel rods by only about 200F so that they also experienced temperatures above about 17000C. Consequently, in the hot region of the core Zircaloy components should have oxidized, and components with Inconel, stainless steel, and Ag-In-Cd should have melted. Because of many layers of protection, the thermocouple tubes have survived even in the damaged core region, although the outer sheath of the instrument tube may be badly damaged. Nearly all of the broken and oxidized fuel debris should remain trapped in the upper core region because the upper end fittings have a grillage that would act as a screen. Furthermore, the compaction of fuel debris 80-234
25' 's limited because it is fabricated with a packing fraction of about i 46% and the theoretical maximum packing fraction (for a bed of spherical particles) is only ab.out 63%. It is very likely that fuel r debris are also trapped in some mixing cups (See Figure 18) contributing to non-uniform thermocouple readings. An earlier estimate of fuel damage in TMI-2 was made at NRC by Rubenstein, Meyer, Tokar, and Johnston. That estimate is in general agreement with the present estimate although our current evaluation is more refined. A memorandum summarizing the earlier estimate is attached as Attachment D. E. Recomendations Reactor fuel is rugged, and it is unlikely that limits for natural circulation conditions will be related to fuel behavior. The general criterion with regard to the fuel should be that additional Zircaloy oxidation and fission gas release should be avoided. Significant oxidation rates do not occur until 900 or 1000 C (See Figure 6). U Significant fission gas releases do not occur until even higher temperatures (See Figure 2). These temperatures should be avoided in the (relatively) undamaged regions of the TMI-2 core, but these temperatures are so high that other limits will probably prevail. By now the adiabatic heatup rate is low (See Figure 19) and ample time will be provided to detect fission gas or hydrogen releases. Therefore, on-line methods of such detection, if feasible, should give adequate 80-235
sD}- h t< i i t \\ ) '.. S 'i
- 3 0
m5 y yk $k \\ r hc t a a Ly l nb Y um L b B oe ~. d n /// d n i A / S o 7 S l l A {_e W 0 l iC l TA L s l 1 \\ i ~ \\ f \\ D \\ D b L 0 t l l l t 8 o@ fA G l l IM, rlA I TT IF D JIE V REPP U f 81 ~ E %kN,NNN /4 R tG / s wfF I 4! m
~ o t d 9 7 9 1 d 5 1 i L I R TTT P PFF D /// d R N WWW ) E S G 964 R R E U O L O F ooa H s( P E M U I T T 4 RE H C I T R B 3 R I D R 9 't 1 eauc i F ~ o 5 8R 8 se g t G ua$5Ei d' d 0O nc% 0 2
M warning of fuel damaging conditions. l A discussion of instrument responses relevant to fuel behavior was l held with a group of fuel experts from across the industry. A t summary of those discussions was prepared by W. V. Johnston and is attached as Attachment E. One consensus of that group was that in-core thermocouple readings should be recorded continuously. A recommendation for such data recording was made and is attached as Attachment F. 80-238 -. _ - -.}}