ML19321B091
| ML19321B091 | |
| Person / Time | |
|---|---|
| Issue date: | 05/29/1980 |
| From: | Harold Denton Office of Nuclear Reactor Regulation |
| To: | Leyse R AFFILIATION NOT ASSIGNED |
| Shared Package | |
| ML19321B092 | List: |
| References | |
| NUDOCS 8007250611 | |
| Download: ML19321B091 (12) | |
Text
L 5$
s,p n
.,g UNITED STATES 8
NUCLEAR REGULATORY COMMISSION o
t,
- p WASHINGTON, D. C. 20555
\\
/
MAY 2 s 1380 r
{,IHIS DOCUMENT CONTAINS Mr. Robert H. Leyse T. POOR QUAUTY PAGE3 12525 Saratoga Creek Drive Saratoga, CA 95070
Dear Mr. Leyse:
In your letter to Comissioner Hendrie, dated April 2,1980, you stated that Figure 4 of SECY 90-107 was erroneous.
The basis for this opinion was your belief that the thermodyn1mics of the Zircaloy-steam reaction were ignored.
In a follow-up telephone discussion with a member of our staff, you stated that, specifically, the peak cladding temperature as displayed in Figure 4 was probably too high. Also, in your letter, you requested that the staff provide you with certain calculations of maximum flame temperature for the zirconium-oxygen and zirconium-water reactions.
The thermodynamics of the zirconium-steam reaction were, in fact, considered in the staff's calculations. The heat of reaction (AHr) was modeled as a constant value of 2808 Btu /lb of zirconium.
A more precise temperature de-pendent AHr could have been used.
However, the reaction heat is only a mild function of temperature, decreasing as the temperature increases (Ref-erence1).
Some important thermodynamic limitations were imposed on our cal-culations.
However, these limitations were clearly stated in SECY-80-107.
The most imports t limitation was the assumption of an adiabatic heat-up.
This assumption allows no heat to be transferred away from the surface of the rod.
This means that enthalphy changes in the gaseous reactants and pro-ducts (H O and H ) are not considered. Another limitation stated in SECY 2 2
107 was the assumption that sufficient steam was always available for the reaction.
That is, no material balance was made to account for the amount of water consumed or hydrogen produced.
Furthermore, it was assumed that the fuel rod's structural integrity would be maintained throughout the transient.
The staff has prepared the enclosed discussion paper addressing the issues identified in your letter.
In summary, we find that the calculations per-formed for SECY-80-107 did consider thermodynamics. The limitations were clearly stated therein.
Flame temperature analysis does not provide suffi-cient information about the relevant processes; specifically, flame tempera-ture analysis ignores the nature of massive zirconium reaction including kinetic effects. The calculations presented in the enc!osure show that heat transfer and the presence of other substances (e.g., U0 ) are very important 2
in determining maximum cladding temperatures.
,800725agg
'.. q,..
Mr. Robert H. Leyse 2-I trust the information contained in this letter is responsive to the concerns identified in your letter of April 2,1980. Mr. Leyse, I thank you for your interest in these matters and for taking the time to write us about them.
e^ -t :w:3 by
!s.fl.OCSICQ Harold R. Denton, Director Office of Nuclear Reactor Regulation
Enclosure:
Discussion i
,,,,3,
- - =,
y --
.o DISCUSSIO!! PAPER Some Thermodynamic Considerations Regarding Zirconium-Water Reactions The thermodynamics of the zirconium-steam reaction were considered in the staff's calculations. The heat of reaction (AH ) was modeled as a constant value of 2808 r
Btu /lb of zirconium.
A more precise temperature dependent AH could have been r
used.
However, the reaction heat is only a mild function of temperature, decreas-Ing as the temperature increases (Ref. 1).
Some imoortant thermodynamic limita-tions were imposed on our calculations.
However, these limitations were clearly stated in SECY-80-107. The most important limitation was the assumption of an adiabatic heat-up. This assumption allows no heat to be transferred away from the surface of the rod. This means that enthalphy changes in the caseous reac-tants and products (H O and H ) are not considered.
Another limitation stated I
2 2
(
in SECY-80-107 was the assumotion that sufficient steam was always available for the reaction. That is, no material balance was made to account for the amount of water consumed or hydrogen produced.
Furthermore, it was assumed that the fuel rod's structural integrity would be maintained throughout the transient.
Your letter implies that there may be a significant reverse reaction of Zr02 and H forming steam and zirconium, and that the equilibrium would be pressure 2
dependent.
Since one mol of 11 forms for each mol of H O in the forward reac-2 2
tion, the thermodynamic expression for equilibrium is:
P i
AF0 = - RT Lin i
_H2j2 lPH20 u
or P
exp
-AFD K*
H2
=
=
2RT P
H2O b ee*p w--
a m-
-- T D
b-
_,m
(
Where: AFD = standard free energy of reaction R
= gas contant T
= absolute temperature PH O = Partial pressure of water vapor 2
Ke = equilibrium constant P
= partial pressure of hydrogen H2 When solid Zr0 is formed, it is well known that the free energy of this 2
reaction is very large and negative.
The resulting value of Ke is 1010 for all temperatures.
Therefore, the reaction tends strongly to go in the forward direction to completion. Although liquid phase thermodynamic properties are not well known, experience with molten 7r,02 (Zircon arc lamp) shows expected stability.
This property of ceramics is of course quite typical.
The rate of reaction of zirconium with oxygen would be expected to be pressure dependent since one of the reactants is a gas and the product is a solid.
However, the rate of reaction of zirconium with steam would not necessarily be expected to be pressure dependent since there is no change in volume in the system by the reaction.
Experimental data (Ref. 2) have shown that there is 0
no effect of pressure at 1500 psi at a temperature of 2000 F on the kinetics of the reaction between Zircaloy and steam. However, at lesser temperatures a pressure enhancement of up to a factor of 3 has been reported (Ref. 3 & 4).
Thermodynamically, a mixture of hydrogen and oxygen will react spontaneously to form water at room temperature.
Kinetically, it never happens in the absence of an ignition solrce.
. ~ _.
-3.
Since the oxide formed on Zircaloy is protective and remains solid until dissolved by molten Zircaloy, the concept of a flame temperature for the reaction of massive Zircaloy (not a fine powder) with steam is not particularly meaningful.
Experiments (Ref. 5-9) have demonstrated that massive zirconium and Zircaloy do not ignite in steam and do so in high pressure pure oxygen only if there is no oxide film present on the surface of the metal, regardless of the temperature of the metal. Since the oxide is soluble in the molten metal, the molten metal can ignite in oxygen if the oxide can be kept from covering the surface.
But even molten zirconium will not ignite in steam, as the heat of reaction is not high enough to overcome the heat losses to the surroundings.
The paper of A. V. Grosse (enclosed with your 'etter), concerning the combustion of metals, describes the use of fine zirconium powder, not massive zirconium, in an oxygen jet to produce a high temperature flame.
For very fine powder particles, the ratio of surface area to volume is sufficiently high that only a little reaction is required to melt the particle. Any finely powdered metal will produce a " flame". in an oxygen jet if it can form an oxide at such a temperature. Only those that do not produce a protective reaction product film can " ignite" in an oxidizing atmosphqre in the massive form where conduction of heat away from the reacting surface can prevent its temperature rise.
There are two other very important factors that need to be considered in the determiniation of maximum temperature associated with the claddina-steam 1
e n,,y n
,,c.-
i 4-reaction. These are the thennal proximity of the U0 and Zircaloy and the 2
contribution of the decay heat source. These factors are examined below with the aid of the attached Figure 1.
The upper claddina temperature curve is identical to the calculation shown in SECY-80-107. The lower curve is the same calculation but without any zirconium-steam reaction.
In the first calculation the reaction was complete at 406 seconds and the maximum cladding temperature was 6490 F.
At 406 seconds in the second case the claddina temperature was 3000 F.
Since both cases started at the same temperature (223 F) and had the same decay heat, the temperature contribution from decay heat in both cases was 2767 F.
The reaction heat contribution to the maximum temperature was 3490 F (see Figure 1).
Within 10 seconds after completion of the reaction in Case 1, the cladding temperature drops to virtual thermal equilibrium with the fuel.
From this time onward the cladding and fuel temperature were within 10 F of each other.
Several " flame" temperature calculations were performed to compare with the computer code calculations in Figure 1.
The results are summarized in the at-tached Table 1.
The standard state for the reaction was considered to be at one atmosphere 0
and 3000 F.
The " products" listed in the table refer to reaction products (Zr0 and H ) and inerts (UO ) whose enthalphy change was considered in the 2
2 2
flame temperature calculation. The higher value of reaction heat (256400 Btu /lb-mol) is equivalent to tne single value used in our computer code T9pDEE2. The value of 210160 Btu /lb-mol value is from Brassfield's (Ref. 1) equation at 3000 F.
The flame temperature formula used was:
?-
m- -
. T i
^H(3000)"fmax (NjCpi)dT (Eq.1) r
- 3000 i
AH (3000) = heat of reactior. at 3000 F (Btu /lb-mol)
Where:
r T
= maximum (flame) temperature (OF) max N
= number of mois of product or inert substance g
C
= heat capacity of product "i" (Btu /lb-mol OF) pj Heats of dissociation were ignored since the reaction goes to virtual completion.
0 The maximum temperature difference (ATmax) is Tmax - 3000 F and is given to show the reaction heat contribution.
The error in AT may be on the order of 20%
max because of the uncertainty in high temperature thermodynamic properties.
Cases 1 and 2 are tabulations of the T00DEE2 results shown in Figure 1.
Case 1, of course, is the actual cladding temperature, whereas Case 2 is the average of the fuel and cladding temperatures.
This would be the condition if fuel and cladding were in thermal equilibrium.
The degree to which Case 1 departs from equilibrium will be explained shortly.
Case 3 is a flame temperature calculation using Equation 1.
The U02 was considered to be in thermal equilibrium with the cladding and was included in the i'ight-hand side of Equation 1.
In principle and in fact, Cases 2 and 3 are nearly equal. Case 3 then provides a check on the intended thennodynamics of the computer calculation.
. Case 4 and flame temperature calculations throunh Case 7 used the lower value of reaction heat from Reference 1.
Comoarison of Cases 3 and 4 shows that by using a tenperature dependent reaction heat the maximum temperature is reduced more than 300*F. This is because the heat of reaction becomes less at higher temperatures.
Case 5 can best be described as a " pure" flame temperature.
Only the reaction products were considered. The temperature reached was the Zr0 boiling ooint 2
(~7800 F).
Case 6 considers the UO and the hydrogen. As expected, this produces the low-2 est temperature. This temperature difference (1950 F) is the maximum metal-water reaction effect that could be expected if, within a few seconds after complete reaction, good thermal contact exists between fuel and claddina.
Case 7 considers only Zr0 in the enthalphy balance. This case may be compared 2
to Cases 2 and 3 to illustrate the effect of considering UO in the balance.
2 The U0 and Zr0 boiling points were taken to be 7800 F which was the tempera-2 2
ture reached.
Even Case 1 in which the U0 and claddino are coupled only by 2
heat transfer is much closer to Cases 2 and 3 than Case 7.
Case 8 is presented for completeness per your request.
This is the result for the zirconium-oxygen reaction. The oxide boiling point is also reached for this case.
In sumary, the calculation of SECY-80-107 did consider thermodynamics. The limitations were clearly stated therein.
Flame temperature analysis does not provide sufficient information about the relevant processes; specifically.
. flame temperature analysis ignores the nature of massive zirconium reaction The calculations presented herein have shown including kinetic effects.
that heat transfer and the presence of other substances (e.g., U0 ) are very 2
important in detemining maximum cladding temperatures.
(_
g-
[
~
Table 1 Maximum Zirconium-Steam Reaction With 0
Reactants Temperature at 3000 F Standard State Case Calculat'on Products (Bru/fb-mol) max (OF)
ATmax (OF)
AH T
1 T00DEE2 256400 6490 3490 Zr0, UO2 2
2 T00DEE2 256400 5330 2330 Zr0, U02 2
3 Equation 1 256400 5404 2404 Zr0, U02 2
4 Equation 1 210160 5075 2075 Zr0, U02,
2 5
Equation 1 210160
~7800
~a800 Zr0, H2 2
6 Equation 1 Zr0, UO, H 210160 4950 1950 2
2 2
7 Equation 1 Zr0 from H O 210160
-7800
~4800 2
2 8
Equation 1 Zr0 from 0 473800
~7800
~a800 2
2 n.
.mme w w A1 -
&*-----a
I References 1.
Brassfield, et. al.," Recommended Property and Reaction Kinetics Data for Use in Evaluating a Light-Water Cooled Reactor loss-of-Coolant Incident Involving Zircaloy-4 or 304-SS-Clad UO, GEMP-482" April 1968.
2 2.
R. E. Pawel, J. V. Cathcart, J.J. Camobell, S. H. Jury, " Zirconium Metal-Water Oxidation Kinetics V. Oxidation of Zi;caloy in High Pressure Steam," Oak Ridge National Laboratory Report, ORNL/NUREG-31, December 1977.
3.
C. Eicheldinger (Westinghouse) letter to V. Stello (USNRC), March 31, 1975.
4.
D. A. Powers (USNRC) note to D. F. Ross (USNRC), March 31, 1978.
5.
" Zirconium Fire and Explosion Hazard Evaluation," USAEC Report, TID-5365, August 7, 1956.
6.
F. E. Littman, F. M. Church, and E. M. Kinderman, "A Study of Metal Ignitions II.
The Spontaneous Ignition of Zirconium," Journal of Less-Common Metals,
_3_, p. 379-397 (1961).
7.
L. F. Epstein, " Correlation and prediction of Explosive Metal-Water Reaction Temperatures," Nuclear Science and Engineering,10,, p. 247-253 (1961).
8.
L. Baker and L. C. Just, " Studies and Metal-Water Reactions at High Temperatures III.
Experimental and Theoretical Studies of the Zirconium-Water Reaction," Argonne National Laboratory Report, ANL-6548, May 1962.
9.
N. I. Sax, " Dangerous Properties of Industrial Materials," Van Nostrand Reinhold Company, 1975.
o O
m
/l o
.0 n
i i0 0
n u
i
.0 i
t n
c 9
o a
c te r
uR i
o Z
nm ta 0
hn i e 0
to Wt ii S
.0
/Wt c
m
.0 a
u 8
N i
no 1
E m
5 c
r
.0 a
s M
e
\\
i t
t C
0 c
S
.0
.0 7
f i
ii0 0
i.
.0 S l
//_
.0 E 06D
. N
. O
. C 0
. S
.S D
.0 N
0 O
5 C
E
. E
. M S
.0I
~
.0. T 8
-m 4
g' I
f
.0 0
4 M
O R
F 0
Yom v
f l
.0 PU
.0 T
0 USgy y
3 A
^
l l
E H
C A
G [; I y
.0 I
0 T
.0 A
.0 B
.E A
I D
[
40 0
A
.0 1
.0
.1 ER l.
U G
0 I
.0 E
- L
~
~
~
~
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0
'0 0
0
/
G S
4 3
2 1
8 l s
mJJ:U U (s U UC i & b'r
. li
.I
'ii' g
i.
'll
.i I I
j1i'.
t i i *