ML20206H081
| ML20206H081 | |
| Person / Time | |
|---|---|
| Site: | Sequoyah, Surry, 05000000 |
| Issue date: | 05/09/1986 |
| From: | David Williams SANDIA NATIONAL LABORATORIES |
| To: | NRC |
| Shared Package | |
| ML20204G644 | List:
|
| References | |
| RTR-NUREG-1150 NUDOCS 8704150252 | |
| Download: ML20206H081 (8) | |
Text
..
~f
..-m
.a.
- ^ -
l May 9, 1986 L, #
l Some Sensitivities for Direct Containment Heating Loads D. C. Williams In a draft NUREG-1150 issue paper on DCH, Tim ime (NRC/RES) ar.d Farouk Eltawila (NRC/NRR) note a number of subissues which can affect the severity of the. containment loads that result from DCH. Their discussion includes some estimates of the magnitude of the effects that may result. The basis for these estimates is acknowledged to represent only " educated guesses" in some cases, and it is one purpose of this j
h-memorandum to provide some calculated results in support of the work of Lee and:Eltavila, examining the potential effects of the various issues 1
they raised.
Calculations cited here were performed with the DHEAT2 code, which is a revised ' version of the DHEAT code used in earlier work performed for the CLWG. In addition to more conitenient input options, the new code incorporates the enthalpy functions for corium constituents developed by M.
Pilch (6425) and currently implemented in the CONTAIN interim DCH model. Effects of chemical reactions upon the composition of the containment atmosphere and the corium are now included, as are options to permit hydrogen burns, reactions of metals with steam in the absence of sufficient oxygen, and dispersal into the atmosphere of liquid water along with the corium debris. As in the original DHEAT code, the
.mh.
pontainment is treated as a single cell and the dispersed debris is assumed to come into thermal equilibrium with the atmosphere, while heat
[ transfer from the gas and debris to containment structures is neglected.
The treatment is thus a purely thermodynamic calculation, with no rate-dependent calculations. Within these limits, it is believed to be reasonably accurate: in code ~ comparisons involving DCH in a dry atmosphere, DHEAT2 and CONTAIN have shown agreement to within 14 for final temperatures and pressures, given equivalent problems (e.g., heat transfer to structures switched off in the CONTAIN calculation).
Somewhat larger discrepancies may arise in analyses of atmospheres rich in steam, due to differences in the treatment of the water equation of state near the saturation curve.
The questions addressed by Lee and Eltavila were raised in the
(,7 context of two plants, Surry_ and__Sequoyah. Since DHEAT2 can not meaningfully. address effects associated with the ice condenser, only Surry is considered here. The effects of the parameters considered will, in general, depend upon the values of other parameters relevant to the problem. The present study has been largely limited to one-at-a-time variation of these parameters about a base case. The base case is as follows:
- 1. Initial containment pressure, Po, and temperature, To, equal to 2.67 bars and 403 K, respectively. These conditions are based upon CONTAIN calculations for the Surry TMLB' sequence, and represent b
8704150252 870408 PDR NOREQ 1150 C PDR WP
- . c.;
- ......
~~
^
~
conditions immediately prior to vessel breach. Initial gas composition is 674 water vapor, 2.774 H2, and the remainder air.
j p(t
~
1
- 2. RCS at full system pressure, 15.7 Mpa; blowdown steam is assumed to enter the containment atmosphere sufficiently rapidly so that it contributes with full effectiveness to the pressurization due to DCH.
- 3. 50% of the total corium inventory is ejected from the vessel-and participates with full efficiency in the DCH; i.e.,
its metal content completely reacts with oxygen and the combined chemical energy and thermal energy equilibrates between the dispersed corium and the atmosphere. Initial corium temperature is 2530'K and the quantity and composition are as specified in C0WG Standard J
i Problem 2.
- 4. No corium is quenched in the cavity (hence there is no steam spike), and no liquid water is dispersed with the corium. The i
absence of a steam spike means that pressures will be lower, but l'
temperatures higher, than in the "50% DCH" case considered in the earlier CLWG work since, in the latter work, it was assumed that 100% of the corium would be ejected from the vessel, and that any corium not participating in DCH would boil water to produce a steam spike.
- 5. Hydrogen initially present does not burn.
The parameters varied and the results obtained are discussed (i
briefly below. The numerical results themselves are summarized in Table 1.
The first column describes the parameter (s) varied from.the base case. The next two columns give the final pressure Pf and temperature Tf
(
that result. The last two columns give measures of the effect of the l
particular parameter variations considered. These measures are:
1 Delta-Pf - Pf - Pf(Base Case) and RPf - [Pf - Po]/[Pf(Base Case)-Po].
In some cases, the effect of a parameter variation upon the DCH pressure i
rise is expected to be additive as a first approximation, in which case i fs
, (,5 the first measure is more meaningful; in other cases, the effect is expected to be more nearly multiplicative, in which case the second measure is more generally useful. In either of these cases, the less general measure is enclosed in parenthases in Table 1.
In some instances, it is not clear that the effect of the parameter variation is either simply additive or simply multiplicative, in which case no
)
preference between the two measures-is expressed.
The base case yields a final pressure of 11.76 bars and temperature of 1317 K, which may be compared with values of about 13 bars and 1180
,l K,
respectively, obtained for the CUWG "50% DCH" case. The differences
]
(T) i
.-+
l,.,
-n,,.
_7 primarily reflect the absence of a steam spike in the present base case, l
(o..
but code differences also have some effect.
f The effects of the parameter variations will be discussed next.
Numbering corresponds to the subissues raised by Lee and Eltawila except for subissue 9, which was not considered in the Issue Paper.
(1). RCS Pressure at vessel breach. Cases la and Ib gives results for a RCS blowdown from 4.1 MPa (600 psi) and no blowdown, respectively. Reductions in Pf are quite significant, of the l
order of 1.5 bars, which is considerably larger than was l
suggested in the draft Issue Paper.
j
' ()
(2). Corium temperature. Calculations run for 1800 K and 2800 K (Cases 2a and 2b, respectively) show a moderate effect, very close to what was suggested in the Issue Paper.
4 (3). Fraction of Core Ejected. Cases 3a and 3b were run with 25% and 1004 core ejection, respectively. Although this fraction is indeed a very important parameter, pressure rise varies somewhat less than directly in proportion to it, for several reasons: some of the pressure rise is due to the blowdown-steam, heat capacities increase with increasing temperature, and increasing fractions.of the total energy still reside in the corium after thermal equilibration, as the total corium dispersed is increased.
(-.,
(4). In-vessel Oxidation of Metals. Cases 4a and 4b give results assuming a) 80% of the Zr is oxidized in-vessel (versus 304 in the base case), and b) assuming 80% of both the Zr and the Fe are oxidized in-vessel. Pressure reductions are substantial, as expected from the importance of chemical energy release to DCH.
l Note, however, that a large amount of in-vessel oxidation implies the generation of large amounts of hydrogen, and the reduction in pressure calculated here is dependent upon the
-questionable assumption that this hydrogen does not burn in the direct heating environment.
f, t,;
(5). Effect of Water in the Cavity. Water in the cavity might have a number of effects. At one extreme, it might quench debris so that oxidation does not occur and the thermal energy only goes into producing steam, not direct atmospheric heating. On the otherhand, it has been suggested that the water might be blown out of the cavity but remain suspended in the atmosphere during direct-heating, in which case it might not. suppress. oxidation j
i but could still be available to absorb energy. The Issue Paper suggested that no more than 15% of the corium energy would be l
subj ected to water quench; however, other investigators have suggested water quench could be much larger and 50% cases are therefore be considered here also.
I l
_ _ _ = _. -
/,6 Cases 5a and 5b give,results obtained when it is assumed that 15%
and 504, respectively, of.the ejected corium is quenched with steam i
f generation only from theithermal energy, and with no oxidation of the quenched fraction. Pressure reductions are significant, and are greater than was suggested in the Issue Paper, based upon the 15% case. In these cases, the amounts of water vaporized were about 4000 kg.and 13500 kg, respectively.
Cases Sc and 5d give the results obtained when it is assumed that the same amounts of water are available for quench, but that the quench occurs only after the oxidation energy is released. Here the effects are seen to be much less.'It is apparent that the dominant effect in the first scenario was due to the suppression of the release of chemical energy, not the reduced effectiveness of steam generation (relative to atmospheric heating) as a pressure source.
i Even given the release of the chemical energy, water quench does ra
- (s have the potential to be effective, but much greater quantities of water must participate. Case 5e shows results when it is assumed that oxidation does occur, but 90000 kg of water is available in the atmosphere for quench. Complete quench occurs, with vaporization of about 72000 kg of the water; pressure rises are much greater than would be calculated for quench of 50% of the core thermal energy in a standard steam spike, but are still well below estimated failure pressures for Surry. Finally, Case 5f gives results assuming oxidation occurs, and 36000 kg of water are available for quench, so that about half the corium energy (chemical plus thermal) goes into steam generation. Though there is mitigation with respect to the base case, the final pressure is still high.
(!
6). Effect of Structure. This subissue represents the effects structure may have on reducing the amount of debris that is transported to the main mass of atmosphere and the main supply of oxygen.
Its effect in DHEAT2 can be only simulated by reducing the amount of corium assumed to participate, yielding results equivalent to the variations considered under subissue 3.
Case 6 shows results obtained assuming structure reduces participation by a factor of 0.75, relative to the base case.
7). Incomplete Thermal and Chemical Interactions. If the same degree of incompleteness is assumed to be the same for both the thermal j
l and the chemical interactions, the representation of this subissue using DHEAT2 is again equivalent to varying the core
( -l, ejection fraction. Cases 7a and 7b give results obtained i
reducing this fraction by 0.95 and 0.5, respectively (the range suggested in the Issue Paper).
Chemical reaction may be " incomplete" because adequate oxygen is not available in the reactor cavity and lower containment regions in which the debris is ej ected, and metals may react with steam before j
transport to abundant oxygen supply occurs. (Indeed, re~ cent CONTAIN i
calculations suggest this scenario is quite plausible.) 1 Case 7e illustrate the effect of assuming Zr and Fe react with steam only.
Substantial reduction in peak pressures results from the reduced energy release. This effect is, of course, eliminated if the hydrogen produced
,' (;
i
y -
subsequently burns on a time scale sufficiently short so that its energy release is additive with that due to the DCH.
If it is assumed that Fe is inert in the. steam environment, so that j
only the Zr reacts with steam (Case 7d), the reduction in pressure is slight because the iron-water reactions contribute relativelf little energy. However, if it is assumed that the Zr also does not react (Case 7e),
substantial additional reduction in pressure occurs, because even the Zr-water reaction is quite energetic.
8). Hydrogen combustion. DCH involves heating of the atmosphere to high temperatures, and also disperses large quantities of oxide particles through the atmosphere; the surfaces of these hot
(,
particles may have a catalytic effect upon the hydrogen-oxygen L.
reaction. Both effects could strongly increase the flammability of the atmosphere, if hydrogen is present. Hence, even if the atmospheric composition is not within the usual flammability limits (hydrogen or oxygen concentrations too low, steam concentration too high), combustion of any hydrogen present at the time of DCH may result.
Calculations were performed for the base case assuming that 24, 44, and 64 by volume of the atmosphere consisted of hydrogen prior to vessel breach (additional hydrogen in the vessel blowdown was not considered).
Results are shown in Cases 8a, 8b, and 8c, respectively. Substantial increases in the final pressure are obtained, considerably larger than was suggested in the Issue Paper.
-(
For a given amount of hydrogen, the pressure rise due to hydrogen combustion is roughly additive to other pressure rises, but this additivity is not very precise. Case 8d gives results for a 44 hydrogen burn added to. Case 3a (254 corium ejection). Relative to Case 3a, the additional pressure rise due to the hydrogen is 1.75 bars, versus 1.39 bars in Case 8b, in which the same amount of hydrogen was burned. Thus, the effect of the hydrogen burn is somewhat greater when the effect of
)
DCH, by itself, would be milder. The reason is the increase in gas heat capacity with increasing temperature.
It should also be remembered that confusion can result when effects of hydrogen burns are expressed as a function of " percent hydrogen" j
unless the starting pressure is specified. For a given " percent hydrogen", the actual amount of hydrogen involved is proportional to the Cs,i pressure, other things being equal; hence, for a given percent hydrogen, 4
the pressure increment due to hydrogen combustion will be roughly proportional to the initial pressure.
Spike.)Although not addressed in the Issue 9). Concurrent Steam Paper, the effect of p concurrent steam spike may be worth considering. Here it is assumed that additional"corium, not 3
part of the 50% participating in DCH, is ejected from the vessel at the same time, and quenches in water to generate steam. The scenario differs from those considered in Cases 5a -
5f in that the quenched corium is in addition to the DCH
(:)
_.,_,___m.-_
_~. _,
- ~
~
~[
.{ : 4.~....
i.
corium, rather than at the expense of the DCH corium. Cases 9a and 9b give results for 25% and 50s steam spikes, respectively,
- (..
being added to the base case. Significant enhancements in the
)
final pressure result. It should be noted that Case 9b is the scenario which actually corresponds to the "50% DCH" case reported in the earlier work performed for the CLWG.
It is essential to remember that the preceding results could be very misleading if applied to actual accident scenarios without any allowance for important mitigation factors associated with time-dependent processes not treated in DHEAT2. For pressures to even
('
approach those cited above, it would be necessary for debris to be efficiently transported to the main containment volume, thermal and chemical interactions to be completed, hydrogen to burn to completion, steam from vessel blowdown and/or steam spikes to be added, etc., all in a matter of a few seconds. Since these conditions are not likely to be
- met, the results cited here can only be taken as indications of the limiting-case potential importance of the effects treated; they certainly do not constitute actual predictions. Making useful predictions of DCH for actual accident analysis must await the development and application of detailed mechanistic models for the governing phenomena, such as the models currently being implemented in the CONTAIN code. Fortunately, rapid progress is being made by the CONTAIN proj ect in this area, and it is safe to assume that all significant uncertainties concerning DCH will be resolved by 4:30 this
(.~
afternoon, or Monday morning at the very latest.
i
1 I
Table 1
)
DCH in Surry Conditions just prior to vessel breach are P, - 2.67 bars, T, - 403K Case Definition Pg (bars)
Tg (K)
g 11.76 1317 0.0 1.0 Base Case
- 1. Primary System Pressure a) RCS Pressure 600 psi (4.13 MPa) 10.44 1522
-1.32 (0.85) b) No Blowdown 10.13 1583
-1.63 (0.82)
- 2. Corium Temperature a) 1800 K Corium 10.35 1159
-1.41 0.85 b) 2800 K Corium 12.16 1361 0.40 1.04
- 3. Core Ejection Fraction a) 25% of Total 8.37 925
(-3.39) 0.63 b) 100% of Total 16.33 1875 (4.57) 1.50
- 4. In-Vessel Oxidation of Metals a) 80% Zr oxidized in-vessel 10.01 1112
-1.75 0.81 b) 80% Zr and Fe oxidized 8.99 988
-2.77 0.70 (b
- 5. Corium Quench in Water a) 15% quench, prevents oxidation 11.05 1184
(-0.71) 0.92 b) 50% quench, prevents oxidation 8.98 874
(-2.78) 0.69 c) Same water as a), oxidation occurs 11.60 1247
.(-0.16) 0.98 d) Same water as b), oxidation occurs 11.16 1099
,(-0.60) 0.93 e) Oxidation reactions occur, 9.E4 Kg water for quench (7.2E4 vaporizes) 6.62 429 (5.14) 0.435 f) Oxidation occurs, 3.6E4 Kg quench 9.83 806
(-1.93) 0.79
(
. ~..,.
Table 1... continued
. ('
- 6. Reduce corium participation by 0.75 10.18 1132
(-1.58) 0.83
- 7. Incomplete Thermal, Chem. Reaction a) 954 complete 11.46 1281
(-0.30) 0.97 8.37 925
(-3.39) 0.63 b) 50% complete c) Zr, Fe react with steam only 9.41 1028
(-2.35) 0.74 k,.,"
d) Zr + H O the only reaction 9.05 989
(-2.71) 0.70 2
e) No Chem. reaction 7.22 788
(-4.54) 0.50
- 8. Hydrogen Combustion a) 24 H 12.46 1405 0.70 (1.08) 2 b) 44 H 13.15 1495 1.39 (1.15) 2 c) 64 H 13.85 1586 2.09 (1.23) 2 d) 4% H Burns in 25% Corium case 10.12 11.36 1.75*
(1.31)*
2
(_,-
- 9. Concurrent Steam Spike a) 25% of total corium inventory 12.51 1231 0.75 1.08 j
b) 50% of total corium inventory 13.23 1163 1.47 1.16
- Relative to Case 3a O
O
-