ML20134G624
| ML20134G624 | |
| Person / Time | |
|---|---|
| Site: | 05200003 |
| Issue date: | 11/07/1996 |
| From: | Joseph Sebrosky NRC (Affiliation Not Assigned) |
| To: | Liparulo N WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP. |
| References | |
| NUDOCS 9611130306 | |
| Download: ML20134G624 (19) | |
Text
~
. _ - - -... -.. - -...-. ~ -. - -...
72 -0D 3
/
\\
UNITED STATES I
is NUCLEAR REGULATORY COMMISSION i
WASHINGTON, D.C. 200SH001
%,..... p#
Novmeber 7,1996 l
Mr. Nicholas J. Liparulo, Manager Nuclear Safety and Regulatory Activities Nuclear and Advanced Technology Division Westinghouse Electric Corporation P.O. Box 355 Pittsburgh, Pennsylvania 15230 "cCT:
FOLLOWON QUESTIONS REGARDING REACTOR VESSEL INTEGRITY DURING SEVERE ACCIDENT CORE MELTS
Dear Mr. Liparulo:
In a May 22, 1996, letter the staff transmitted two reports to Westinghouse I
concerning reactor vesse's integrity during severe accident core melts.
, to this letter provided Idaho National Engineering Laboratory 1
(INEL) comments and questions on the application of the Department of Energy (DOE) report DOE /ID-10460, "In-Vessel Coolability and Retention of a Core Melt," dated July 1995.
This enclosure was provided to Westinghouse for information and consideration, and was to serve as the basis for future discussions on the topic.
A teleconference z s held with Westinghouse, INEL, Dr. Ti.cofanous, DOE, and members of the soif on October 25,1996,: to discuss Enclosure 1 of the May 22, 1996, letter.
Unfortunately, this teleconference did not prove productive in addressing the concerns identified in the letter.
Therefore, j
the staff has decided to reissue Enclosure 1 of the May 22, 1996, letter as requests for additional information (RAIs).
Please note that the only i
difference between Enclosure 1 of the May 22, 1996, letter and the Enclosure to this letter is that RAI numbers have been assigned to the INEL comments and questions.
Although this letter is being reissued unchanged except as noted above there are several typographical errors that remain in the enclosure.. They are listed here to ensure that you are aware of the problems with the letter. The errors include the following:
Figure 2 and 3 on page 3 should be labeled figure 1 and ? respectively.
The last sentence of the first paragraph on page 12 should have the parameters and the temperatures changed and should read as follows:
"Thus the radiative sink temperature, T,j,ture of approximately 960 K (Figure Q.10) and the melt surface tempera
,T
, of approximately t
f 1710 K would be inappropriate for such a configura ion."
The third paragraph under Finite Element Analysis on page 14 should have
.the first occurrence of 2,5_cm changed to 5.0 cm and should read as follows:
... Figure 4.6 also indicates that the wall thickness is 5.0 g\\
ca;...".
NRC HLE CBRER COPY
((TM h 3
A
Mr. Nicholas J. Liparulo November 7, 1996 If you have any questions regarding this matter, you can contact me at (301) 415-1132.
Sincerely, original signed by:
Joseph M. Sebrosky, Project Manager Standardization Project Directorate Division of Reactor Program Management Office of Nuclear Reactor Regulation Docket No.52-003
Enclosure:
As stated cc w/ enclosure:
See next page DISTRIBUTION:
D t ocket;J11ej PDST R/F JSebrosky PUBLIC BPalla, 0-8 H7 JMonninger, 0-8 H7 FEltawila. T-10 G6 F0dar, T-10 G6 CBerlinger, 0-8 H7 JKudrick, 0-8 H7 GHolahan, 0-8 E2 EJordan, T-4 D18 DISTRIBUTION w/o enclosure:
TMartin DMatthews TQuay TKenyon DJackson WHuffman l
WDean, 0-17 G21 JMoore, 0-15 B18 ABehbahani, T-10 K8 i
ArRS (11)
ARubin, T-10 K8 YChen, T-10 K8 S8asu, T-10 K8 l
4 DOCUMENT NAME: A:IVR SCSB.RAI f.9J AP600 DISK) i 's T3 seenive a espy of thle doewesent,inacete in the bec t : e Copy without ettschmentienclosure
- E' s Copy with ettschment/ enclosure "N' = No copy 0FFICE PM:PDST:DRPM l
.;.:$$E$19ESA D:PDST:DRPM
[
NAME JSebrosky:sg,' CJkudrkiNM TRQuay'ON DATE 11/'7/96 V'
11/Fs/96L?\\t' 11/1/96 ICIAL RECORD COPY g
\\
4
Mr. Nicholas J. Liparulo Docket No.52-003 Westinghouse Electric Corporation AP600 cc: Mr. B. A. McIntyre Mr. Ronald Simard, Director Advanced Plant Safety & Licensing Advanced Reactor Programs Westinghouse Electric Corporation Nuclear Energy Institute l
Energy Systems Business Unit 1776 Eye Street, N.W.
P.O. Box 355 Suite 300 l
Pittsburgh, PA 15230 Washington, DC 20006-3706 Mr. John C. Butler Ms. Lynn Connor Advanced Plant Safety & Licensing Doc-Search Associates Westinghouse Electric Corporation Post Office Box 34 Energy Systems Business Unit Cabin John, MD 20818 Box 355 Pittsburgh, PA 15230 Mr. James E. Quinn, Projects Manager LMR and SBWR Programs Mr. M. D. Beaumont GE Nuclear Energy Nuclear and Advanced Technology Division 175 Curtner Avenue, M/C 165 Westinghouse Electric Corp..ation San Jose, CA 95125 One Montrose Metro 11921 Rockville Pike Mr. Robert H. Buchholz Suite 350 GE Nuclear Energy Rockville, MD 20852 175 Curtner Avenue, MC-781 San Jose, CA 95125 Mr. Sterling Franks U.S. Department of Energy Barton Z. Cowan, Esq.
NE-50 Eckert Seamans Cherin & Mellott 19901 Germantown Road 600 Grant Street 42nd floor Germantown, MD 20874 Pittsburgh, PA 15219 Mr. S. M. Modro Mr. Ed Rodwell, Manager Nuclear Systems Analysis Technologies PWR Design Certification Lockheed Idaho Technologies Company Electric Power Research Institute Post Office Box 1625 3412 Hillview Avenue Idaho Falls ID 83415 Palo Alto, CA 94303 l
Mr. Fran! A. Ross Mr. Charles Thompson, Nuclear Engineer U.S. Department of Energy, NE-42 AP600 Certification Office of LWR Safety and Technology NE-50 19901 Germantown Road 19901 Germantown Road Germantown, MD 20874 Germantown, MD 20874 4
t 4
i
I INEL Comments and Questions about DOE /ID-10460 INEL's review indicates that Reference 1 applied the approach shown below to demonstrate ves-sei integrity for cases with complete RCS depressurization and ex-vessel cavity flooding. As indi-cated in this figure, the authors must prove two assertions in order to !nsure vessel integrity:
For all heat fluxes at or below the CHF, the corresponding minimum vessel wall thick-nesses are sufficient that the vessel remains intact.
Heat fluxes to the lower head always remain below CHF.
As indicated in the figure below, various types of information (analyses, experiments, and assumptions) were used to suppor' 'ach assertion. Starting at the far right of this figure, the authors used CHF data from the UU tests tc estimate minimum vessel thickness for the struc-tural analyses and provide bounding heat fluxes for the thermal analysas. For the minimum ves-sol wall thicknesses, the authors performed structural calculations to demonstrate that the vessel remains intact if heat fluxes are at or be!ow the CHF (usertion 1). Thermal analyses are per-formed for two bounding configurations to determine maximum possible heat fluxes. These heat fluxes (and associated uncertainties) are shown to be below the CHF values (Assertion 2).
Hence, the authors concluded that the vessel remained intact for cases with complete RCS depressurization and ex-vessel cooling. The authors relied on peer reviewer concurrence to vali-date the adequacy of the information supporting each assertion, n.
e kh8888
^^"'""
=,,,,*a"* :--
- =t,,
- m.ao, 7.ell".""M
- g"g;;,;la
" " * - = -
- ETw""* ""l"
"= - --
CHF Haar Fhans uuuse..ee
,.e
".L,."s s;ls""llll e.se.
M m
o
. n,. e
, ise.
l
=g
."a"""ll%,ses e,annese 1"l""'"*"'"" '*"
ww=
. o.mann, se a,.ie n
. Deens so.my heen,r eiseason ame
. He e..nnuten m e.
. unanionv emensamen.enews,
. Remaean er* to,,gerenne
. Heat tenster I,em mouse pool ir
.amesyso. sonapnemennewwe veens==
.[n7
. e.
e as, basesmen mosanne
. Neoglais FCWnt.auf preenwe noes DOEllD-10460 strategy for demonstrating AP600 in Vessel Retention in helping NRC evaluate the conclusions from Reference 1, INEL is reviewing all of the assump-tions, experimental data, and analyses listed above. Thi, paper contains preliminary questions about the rssumptions, e.:perimental cata, and analyse used to support each assertion.
1
i j
Thermal Analyses Uncertaintles l
480.440 Uncertainty estimates provided in Reference 1 are incomplete and as such, tend to under-i state overall uncertainties in the thermalloads. Please revise the main Section 7.2 calcula-tions comprehensively considering appropriate values for the following uncertainties:
l Material properbes. As discussed below, Table 7.1 implies that Section 7 calculations l
assumed uncertainty distributions that were inconsistent with those presented in Appendix L.
i Furthermore, uncertainties in many material properties (melting temperatures, density, spe.
cific heat, and emissivity) appear to have been neglected.
)
Natural convection heat transfer correlation (unward downward. and local correlationst Nat-
)
ural convection correlations contain uncertainties associated with experimental conditions j
and uncertainties introduced in fitting a natural convection correlation to the data. As dis-I cussed below, there are also uncertainties associated with heat transfer because of phenom-l ena, such as transient phenomena and vapor transport effects.
l Decav nower curve uncertainties. There are non-negligible uncertainties associated with decay power curves. Furthermore, as discussed below, it appears that uncertainty distribu-
{
tions for pool power density are too small.
l Metal laver heat transfer. There is uncertainty associated with the assumed minimum thick-4 ness of the layer and the assumed heat transfer correlations. Furthermore, there is the poten-l tial for oxidation to produce significant heat sources in the metal and to reduce upward heat losses from upper surface of the metallayer.
480.441 As discussed below, there are other debris configurations that may produce more severe thermal loads to the vessel wall. Please provide appropriate probabilities for various postu-j lated endstates.
l Molten Pool Natural Convection Heat Transfer Assumptions l
480.442 Figures 1 and 2 compare several natural convection heat transfer correlations developed i
from various experimental and analytical studies, in these figures, correlations are plotted over the range of Rayleigh numbers for which they were developed. As shown in these fig-ures, there is considerable scatter between various correlations. Comparing the experimen-tally developed, Kymalainen and Theofanous correlations, one can see that the ratio of upward to downward heat transfer is lower for the Theofanous data in the higher Rayleigh number range. Thus, an analyses assuming the Theofanous correlations (based on mini-ACOPO data) will predict less heat being transferred to the metal layer above the ceramic debris and reduced heat fluxes from the metal layer to the wall. This trend is demonstrated in Figure 7.15, which investigates a sensitivity in which the Mayinger correlation was substituted for the mini ACOPO Theofanous correlation (the Maginger correlation is lower than the Theo-fanous correlation for pool Rayleigh numbers of 10 ). Larger differences exist between the Theofanous and Kymalainen correlations than the Mayinger and Theofanous correlation for 15 pool Raylei0h numbers at 10. In their response to Dr. Schmidt, the authors state that these differences are due to the shape of the lower boundary. However, this emphasis on geometry yields the author's data as the only data applicable to these tests. Furthermore, mini-ACOPO data do not agree with other hemispherical correlations at low Rayleigh numbers. Please per-form a sensitivity study based on the Kymalainen data.
2
~.
.=
- ~
.., ~
10*
West nyou. DEA 108 oce s so.se
,.a-
,e
("a -
WOAD 13386 10*
.. - *,, '**~
=
3 z
.e =.
10' W.yeg.r..sp Kymal.in.n..errments. t.ce.
spenm.nt.t lc M.ymg.r. nwmer.c.t h.m.pr..r.
M,n & K.ar.ck.
l Ku.i.es, &.Em.o no nm.nt.' cy moere..mpen M.y,ng *. es nm nt.i t.ct.ngol 10, c'
St mD in.or m ' &.A. men...apenenent. rect.ngul.'
nx... w n nw n. mien..
(
10 ^10 ' 10' 1b' 1b'*
10" 10'
- 10
15
1 d
10
10" Ra r
l l
l Figure 2. Comparison of correlations for predicting upward heat transfer.
10*
u.,y.ng...
u.,,ng.,.
nm.nw suc K,
.m
.., w m.<
c.co.. om.nc.n,n. m.pn.m.nw n.m..en.,.
.n 102
- = w nt a*$on*
- * F " '8^
5 a :2 s'i.',no.5 D.a"&.w.
.a...s'.n= w =ct.nw.,
.m.m...et.nu,
m, n
p M.y To...
Ai.n.
n,,..~nm.nw Tn or.nw... wan. n.m.on.
n.m on..
s '"
p.e 10*
,u
~
DOFID.10 60 10' 1
j WCAP 13348 108 i
t 4
10 10' 1b' 1b' 1b
1b" 1b
1b'*
10
1b
10
10" Ra Figure 3. Comparison of correlations for predicting downward heat transfer.
3
480.443 Figure 5.8 compares ratios of local to average Nusselt numbers from several numerical and experimental studies. As noted in the caption, these curves have been area averaged to cor-rect for differences between tests run in slice and hemispherical geometries. As discussed during the March 21,1996, NRC/ DOE / Westinghouse meeting, this correction was performed by dividing tt,e ratio, Nu,_g/Nu
,fj,,, with the following ratio g
Nu* ~ #
, Nu"' #
dA "'"'
' N u _,g;,,
Nu _,g;c, (2nR) dR Nug_ s,,,,
g g
2 Nud-slice dA s,,,,
2nR Please clarify how this equation was applied to obtain the Figure 5.8 Mayinger curve. Assum-ing the Mayinger (1985) curve is based on information from Mayinger (1975)* and Jahn and Reineke (1974), INEL obtained the curve shown in Figure 3. Note that the peak local value (at 90 degrees) is lower than either the mini-ACOPO curve or the Mayinger (1985) curve shown in Figure 5.8. Please discuss the equations used to obtain the Figure 5.8 Mayinger (1985) curve and include the Kymalainen COPO local values in Figure 5.8.
2.0 l
/
i.s -
j -
m
/
y.
%tana,.
\\
, 1 o-
)
/,,-
/
o.s
/
,e
.... /
"o to ao ao lo s'o so 7o so so 0.v=.
Figure 3. Comparison of area corrected localized Mayinger data.
480.444 The potential for higher transient thermal loads (than those at steady state) at lower positions on the lower head is acknowledged on p. 5-10. As the authors discussed in their responses to Drs. Seghal and Schmidt, an addendum was included in Appendix D that discusses mini-ACOPO transient data to provide insights about the development, with time, of heat flux dis-tributions. Data presented in this addendum were taken from tests in which the time required for steady state conditions to occur was minimized.
- The DOEllD-10460 reference list did not contain a Mayinger (1985) entry. Hence, a was assumed that the Mayinger (1985) curve represents the Mayinger 0975) data.
4
I~
l
.~
i The authors and Dr. Schmidt exchanged several comments regarding whether transient nat-ural convection relates to the boundary layer or the bulk pool time constant. The authors note
}
that the boundary layer time constant is on the order of 2 minutes. However, e bulk pool time constant corresponding to a molten pool in the AP600 lower head would be considerably i
longer. In Min and Kulacki,2 the following relationship was developed based on experimental data for estimating the time required for steady state behavior.
c t,, =
23.34 ( ara 4 223)
(2)
Assuming the material properties in Appendix L and Table 7.2, this relationship suggests that l
transient behavior may last as long as 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br />. in Appendix V, Dr. Schmidt identifies this i
point as an open item and the authors state that they anticipate that ACOPO data will resolve j
any questions about the characteristic time for steady state conditions.
I Data in the addendum to Appendix D suggest how transient heat fluxes may vary as a func-tion of time. Data from several tests (A5, A10, and A6) suggest that at later times, the ratios l
of the local to average Nusselt numbers become larger at locations near the 90 degree loca-tion and that ratios of local to average Nusselt numbers become smaller near the 10 degree location (no cata were plotted for the O degree location). Hence, these data support the assertion that heat loads to the vessel at lower angles (where the CHF values are lower) may be higher during transient time periods.
To resolve questions related to transient behavior, please provide the following:
An explanation of the behavior observed at the 40 degree location in runs A3, A4, and A6.
The ACOPO data to resolve Dr. Schmidt's comments A sensitivity study using a distribution that bounds the flatter flux shape distributions that may occur during transient stages of natural convection.
480.445 In Dr. Tuomisto's comments, he noted that the presence of materials with lower vaporization temperatures in the ceramic pool could increase heat transfer rates from the ceramic pool to the metal layer. In Appendix 0, the authors responded that there is no mechanism for entrap-ping significant quantities of steel below superheated oxidic melt. However, there is a poten-tial for superheated steel and control materials to be mixed within the molten pool. The presence of such lower vaporization temperature materials is supported by voids measured in examinations of debris from the TMI 2 vessel and the LOFT fuel damage tests (porosities l
of at least 20%). Using correlations based on experiments performed by Greene, et al.,3 1
INEL walls.prformed calculations to estimate the impact of vapor on heat transfer to the side As shown in Figure 4, boiling significantly increases heat transfer for pool Rayleigh numbers of interest. Information in Reference 4 indicates that upward heat transfer experi-ences a similar increase because of vapor transport. Have calculations been performed to i
quantify the time required for these materials to vaporize from a ceramic pool? Please pro-J vide calculations assessing the impact of vapor transport effects on the margins to CHF.
Molten Metal Layer Heat Transfer 480.446 Please provide the following Chapter 5 Addendum information:
-identify the two-dimensional heat transfer model used for these calculations.
1
-Clarify how the vessel dimensions and heat fluxes shown in Figure 5.15 were derived.
j
-Clarify the impact of applying results from a vertical, flat plate analyses to a structure that is j
angled (the Figure 5.15 diagram should be at least 70 degrees as indicated in Figure 4.8).
-Explain why heat fluxes are higher within the unmelted vessel steel in Figure 5.16 (heat 2
fluxes of nearty 1600 kW/m are predicted within the vessel, but not on the surface).
l 5
0
10' Vset' 5'acer. s s.o 10*-
.......oi
.or
.._..os 10'
- 10' h'
10*-
M g.,
1 08 g.;-#
10' 1U' toa 10" 16* Id*
to" 1ba 1o*
10" to*
Rayleigh number Figure 4. Impact of boiling on heat transfer from a molten pool.
480.447 Several references discuss experiments that have been performed with an internally heated fluid layer below a second unheated, immiscible fluid layer.
G. Fieg," Experimental Investigations of Heat Transfer Characteristics in Uquid Layers with Intemal Heat Sources," Proceedings of the ANS-ENS Intemational Meeting on Fast Reactor Safety and Related Physics," Chicago, USERDA Report CONF 761001-P4,1976, pp. 2047 2055.
R. Schramm and H. H. Reineke," Natural Convection in a Horizontal Layer of two Differ-ent Fluids with Intemal Heat Sources," Proceedings of the Sixth International Heat Transfer Conference, Toronto, Canada, Vof. 2, Paper NC 20,1976, pp. 299 304.
F. A. Kulacki, A. A. Emara, J. H. Min, and A-T. Nguyen, " Experimental Studies of Steady arad Transient Natural Convection with intemal Heat Sources in Enclosed Cavi-ties," Proc % dings of the Fourth Water Reactor Safety Research Meeting, Gaithers-burg, MD,1976.
i F. A. Kulacki and A-T Nguyen, Hydrodynamic Instability and Thermal Convection in a
=
Horizontal Layer of Two Immiscible Fluids with Intemal Heat Generation, NUREGICR-2619R3,1980.
Correlations developed from Fieg and Kulacki, et al. for upward heat transfer are compared with the Globe Dropkin correlation in Figure 5. As shown in this figure, the Globe Dropkin cor-relation applied in Reference 1 predicts higher upward heat losses from the metal layer.
1 Assess he impact on results if the Fieg and Kulackl data were used instead of the Globe-Dropkin correlation.
Meterial Properties 480.448-A liquidus temperature of 1573 K is assigned to the metallic layer on p. 5 2 with reference to Figure 6.1. It appears that the specified temperature corresponds to a Zr mole fraction of i
-0.1. It is not clear why other temperatures (corresponding to other possible Zr mole frac-tions) are not considered. Appendix P indicates that 1600 K is the lowest possible tempera-j ture of any Fe-Zr eutectic in which iron is the dominant component. However, it is not clear 6
1 4
I i
71M)
~
}
-'~ KuencewEma f
SW)-
/
,. ~~
/
/
,/
M" e**
,/
p a ***
e**'*"
,,... = "
- p_._
10' 10*
10" i o
e, eena ese l
Figure 5. Comparison of correlations for predicting upward heat transfer from the metal layer.
that iron is necessarily the dominant component. Best-estimate SCDAP/RELAP5 calculations have been completed for 1 scenario similar to the one considered in Reference 1. Results from those calculations, rate that the mole fraction of unoxidized Zr (before lower core plate molting) could appuc,* 50%. Higher Zr mole fractions may be possible in some loca-tions if the metallic melt is not well mixed. In their response to Olander, the authors noted that "the composition cannot be specified with any degree of accuracy." However, no sensitivities were performed to consider the impact of this temperature.
For example, it appears that Figure 4.1 was obtained by using the best estimate values for critical heat flux (Figure 3.3), vessel wall thermal conductivity (32.0 W/K m per Table 7.1),
and irN zirconium liquidus temperature (Figure 6.1). If one considers the uncertainties in the vessel thermal conductivity (i 10% according to Appendix L) and the iron zirconium liquidus temperature (variations in the mole fraction of zirconium may lead to liquidus temperatures
)
between 1201 and 1948 K), the CHF-limited wall thickness could be as small as 1.5 cm (com-pared to the 2.5 cm shown at 90 degrees in Figure 4.1). If one were to consider the uncertain-ties estimated by others for similar experiments (i 20%),5 the CHF limited wall thickness would be reduced even further as shown in Figure 6. As discussed in the structural analyses comments, this thickness (and other uncertainties related to the structural analyses pre-sented in Section 4) reduces the robustness of Section 4 conclusions, and the finite element calculations should be revised using an appropriate vessel wall thickness 480.449 Page 7-1 indicates that values for thermophysical properties and their 2a values are listed in Table 7-1 based on information in Appendix L However, Table 7-1 lists single values for many of the properties in Table 7.1 (e.g., density, specific heat, emissivity, and molting point).
With the exception of viscosity, most of the ceramic pool properties listed in Table 7.1 are not shown to vary as a function of temperature. Furthermore, there is no indication that these properties were varied with temperature in the calculations for estimating the pool Rayleigh numbers. Please discuss how ceramic pool temperature-dependent properties and debris and vessel material property uncertainties were considered in these calculations.
7
.._..-._..~.-......_..._.__._...m.-
_.-..m._.
1 I
i; 0.1$
f l,
--
- Uppet hourW asewnpaone
+
-DOEAD 10eso sesumpeone j
i,
..- Loser bound soeumpeone a
o.1o-E i
g o.os e
- o io ao ao eo so so 7o ao sa Deer =
a,,-.
Figure 6. Possbie CHF-limited vessel wall thicknesses (Upper bound assumptions correspond to maximum vessel melting temperature and reduced CHF values; Lower bound assumptions corre-spond to minimum vessel melting temperatures and maximum CHF values).
480.450 Table 7.1 suggests that many of the Appendix L material property uncertainty estimates were neglected in estimating the overall uncertainties in the wall heat fluxes. Uncertainties in melt-ing temperatures, density, specific heat, and emissivity were not included in the values given in Table 7-1 even though there are uncertainties in these parameters associated with compo-sition and temperature as noted in Appendix L. In addition, the uncertainty estimates pro-vided for the effective thermal conductivity of the ceramic crust are not consistent with those presented in Appendix L Table L 3 shows that the thermal conductivity of the crust can vary between 5.6 i 1.1 to 2.010.3 W/mK, depending upon the composition in the crust, yet Table 7.1 used 2.8 i 0.4 W/mK. Appendix L material property uncertainties do not include uncertainties associated with prototypic crusts such as the influence of other materials in the crust or porosity of the crust. For example, the thermal conductivity for the ceramic debris i
samples taken from the TMI 2 lower pienum varied from 1.0 to.3.8 W/mK over a temperature range from 1000 to 1800 K. The sample to-sample variation at 1800 K was approximately 1.7 i
to 3.8 W/mK.' Please correct discrepancies between Appendix L and Table 7.1 material property values and revise calculations to consider swepr.' ate values and uncertainties.
7 480.451 SCDAP/RELAP5 data suggest that thermal properties for stainless steel (specific heat and thermal conductivity) differ from the iron properties assumed in Appendix L and Table 7.1 for the stainless steel in the metal layer. For example, the conductivity of stainless steel at 1700 K appears to be closer to 36 W/m-K than the 25 W/m-K value cited in Table 7.1. Why weren't stainless steel properties used in these calculations and what impact would the appropriate properties have on the results?
8
Decay Heat Assumptions 480.452 The total decay power shown in Figure 7.1 appears reasonable compared to SCDAP/
RELAP5 calculations, which agree with the ANS 5.1 Standard. However, the ANS Standard includes estimates of uncertainty, which should be considered. For example knowledge of the core power within -2% and knowledge of each of the 23 U, 238U,and 239PU power fractions within -1%, the Standard indicates a 1 o uncertainty in decay power of ~5%.
In order to capture -95% of the uncertainty in decay power for those assumptions, one would need to place error bands of e 10% (t 2 c) on the decay power curve in Figure 7.1. In order to conclude that thermally-induced fai!ure of an extemally flooded, AP600-like reactor vessel is " physically unreasonable," more than a 2 e uncertainty should be considered. Please assess the impact of decay heat uncertainty.
480.453 Discussion on p. 7-9 indicates that the most likely time for core relocation lies between 4 and 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> with a supporting reference to a MAAP calculation. However, SCDAP/RELAP5 cal-culations indicate relocation begins at 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br />. A recently completed, revised MAAP calcula-tions also indicates a much earlier challenge to the lower head, with the beginning of relocation to the lower plenum shortly after 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />, and depletion of water in the lower head at approximately 2.7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br />. Please quantify the impact of earlier relocation times on your results by shifting the distribution of Figure 7.7 by 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />.
480.454 Results from SCDAP/RELAPS calculations indicate an initial molten pool power density of 3
-1.5 MW/m. This value is conservative relative to the treatment given in the subject report because uniform mixing of molten pool components is assumed in SCDAP/RELAP5. Specifi-cally,if one assumed metallic segregation (as is done in DOE /ID-10460), the power density would increase to an initial value of -1.7 MW/m. On that basis, it would seem that the maxi-3 3
mum value of 1.4 MW/m given in Figure 7.8 is too low. Please consider a distribution that accommodates the possibility of these higher power levels.
Assumed " Bounding" End State Condition 480.455 As noted by Levy and a number of the reviewers, the assumed end state cer, ation shown in Figure 2.2 is not necessarily the bounding or even most plausible configurL m for the stratifi-cation of trozen and molten material. Figure 7 illustrates several possible auymmetric, ideal-ized configurations that may be more challenging than the one assumed in Reference 1.
Configuration A corresponds to results from SCDAP/RELAPS calculations for a similar sce-nario. (This calculation is not yet complete in that a further assessment of lower head heat transfer and the potential for additional relocation of steel is still to be completed). The config-urat6on is characterized by a large oxidic pool (~85% of the core inventory) with a small metallic coinWaent (-5600 kg of unoxidized Zr and ~3400 kg of stainless steel). This config-uration developed primarily as a result of localized relocations through the reflector sidewall.
The stainless steel component accumulated from the molting of gray rods and the addition of lower plenum structures that melted as a result of being submerged by the debris. Calcula-tions suggest that this configuration could persist for a period of time necessary to completely bolt water in the lower head and then melt portions of the lower core plate. During this period, the metallic layer of -10 cm would remain well below the threshold of 15 cm, which was given as the thickness necessary to insure vessel integrity under ' focused' heat loads. Further-more, asymmetries couldbo expected in the core region as a result of the break location and the conditions that could develop in the vicinity of other inajor vessel penetrations (e.g., accu-mulator flow lines, core makeup line). These asymmetries that could significantly reduce steel molting and, ultimately, the thickness of the ' focusing' layer, in the main results, please assess the impact of a thin metallic layer on lower head integrity, as a result of potential relo-cation scenarios and as a result of acknowledging asymmetries in core melt progression.
9
4 3
P
~~
Con #gurason A Conrguraton B Figure 7. Possble endstates that may be more challenging than the Reference 1 endstate.
In Configuration B, a second pour would relocate onto the core plate forming a crust above the core plate. Any subsequent pours result in a second molten pool with a thin metal layer above it. As discussed above, multiple pours are possible because there are non axisymmet-ric heat transfer boundaries. The core plate would be heated from below by the molten pool and above by the overlying crust. Metal in the core plate would become molten and eventu-ally superheated, possibly resulting in a more severe heat load to the vessel than that postu-lated in the report. Hence, the idealized configuration shown in Figure 2.2 is only one of many plausible configurations. As discussed in Appendix 0, factors such as the " random nature of boiling crisis (as found in ULPU), natural asymmetries in the thermal loading," etc. will make it highly unlikely that the actual endstate in the lower head is axisymmetric. Hence, there are a number of additional scenarios that could have been considered that may be more plausible than the configuration presented in Figure 2.2. As Dr. Levy noted,if the molten stainless steel becomes trapped below a crust of ceramic material, the molten stainless steel must either continue to heatup above the melting point of the ceramic crust or mechanically break through the crust to migrate to the top of the ceramic layers. In the event that molten steel becomes trapped below a crust of ceramic material, please assess the thermal loads on the vessel and the margin to vessel failure in the steel region.
480.456 The calculated upper surface temperatures in the metallic layer, Figure Q.8, and inner tem-peratures for the upper core barrel temperatures, Figure 0.10, are well below the melting point of stainless steel. What mechanism caused the upper stainless steel structures to melt and produce the quantity of molten metal assumed in Figures 2.2 and 7.57 480.457 The metallic layer shown above the oxidic pool in Figure 5.1 is a basic assumption of the analysis outlined in this report. Is there evidence to show that this segregation will develop and persist in a turbulent natural circulation environment? This point is noted as an open issue by Dr. Turland (item 7). Are there conditions where lower head integrity could be adversely impacted if this segregation does not develop? If segregation occurs, could other configurations (i.e., a thin layer of absorber materials) adversely impact vessel integrity?
Heat Addition Due to ChemicalInteractions 480.458 Section 7 calculations omit heat addition due to the metallic layer oxidation. As discussed by Pro-fessor Olander, ignoring the heat generation associated with the oxidation of the metallic materi-10
~
i j
)
j als in the core is not only non-conservative, but clearly is at odds with the behavior that has been i
demonstrated in numerous experiments with prototypic materials. Professor Olander also noted i
that the metallic layer will oxidizo quickly in the presence of steam and even more quickly in air, j
Although Professor Olander's comments were specific to the oxidation of zircaloy present in the metal layer, stainless steel will also oxidize rapidly in steam and air.
While the heat of reaction for stainless steel is less, 270-650 kJ/ mole (depending on the final j
products formed), than that of zircaloy,1200 kJ/ mole, oxidation of stainless steel is important for j
several reasons. First, if it cannot be shown that the total oxidation rate is limited by the availabil-Ity of steam and air, the total heat addition considering both stainless steel and zircaloy is signifi-cantly more than for zircaloy alone. Second, the reaction rates for stainless steel and zircaloy are j
comparable at or near the molting point of stainless steel. In fact, the reaction rate data from sev-i eral investigators, including Bittel, Wilson, Leistikow, and Baker,e omgh u how that the oxida-s l
tion rate for stainless steel exceeds that of zircaloy as the melting point of stainless steel is j
reached. Third, the oxidation of stainless steel results in swelling and foaming of the material as l
molting temperatures are reached. This results in highly porous debris in regions where the stain-t less steel mixes with the fuel and zircaloy. For example, in many of the integral bundle heating F
and melting experiments conducted at low pressure in CORA and ACRR, the frozen U-Zr-O in j
regions where stainless steel structures were present exhibited a froth-like appearance, resulting l
in a density of the U Zr-O substantially lower than UO or ZrO. Conduction through this porous 2
2 region provides some insulating material on top of the metal layer, enhancing heat transfer to the i
side walls. Once oxidation is complete, the oxidized metallic layer provides additional insulation j
above to ceramic debris and increase downward heat loads to the vessel. Fourth, the presence
{
of smc4 mounts of iron oxides in debris composed primarily of (U,Zr)O can significantly impact 2
the properties of (U,Zr)O. For example, thermal diffusivity measuremerits of TMI-2 ceramic 2
debris varied as much as 150%. As noted by Uetsuka,e the most influential factor in that variation
)
was the presence of a second phase composed of ferrous oxide.
{
in their response to Olander, the authors noted that oxidation of the metal layer could be neglected because (a) there would not be sufficient steam present to allow the reaction to pro-coed at the rate noted by Olander ("any containing supply of steam, as assumed, has to come from the containment atmosphere, together with lots of air") and (b) the increase in emissivity of 0.8 would more than compensate for additional heat generation. However, as Olander pointed out, the addition of air dramatically increases oxidation of steel and zircaloy due to the large increase in the heat of reaction.
480.458 a) Was the water initially present in the lower head during relocation considered as a source for oxidation?
b)
Please document the mass flow rates of steam and air into the vessel and demonstrate that oxidation of metals in the core would be prohibited during later stages of the accident.
c)
Please quantify the effects of the energy possible from stainless steel and zirconium oxida-tion, the reduced upward losses from a metallic layer with an oxidized upper surface, and the increased focussing associated with a thinner, unoxidized metallic layer.
Radiative Boundary Condition on the Upper Surface of the Molt 480.459 The racRative heat transfer model for the debns upper surface assumes an idealized set of condi-tions that maximize heat transfer from that surface. In one instance, it appears that this model was applied to a configuration that violate modeling assumptions. It is noted in the report that the
. radiation exchange between the uppsr surface of the melt is dominated by the temperature at the melt surface since (a) the radiation from the remaining structures is smal! because of the large f
surface area relative to th6 melt surface and (b) the radiation from the steam / air / aerosol atmo-sphere can be neglected because the atmosphere was transparent and any aerosols had been i
11 l
1
~
- - - - - -. -. - _.... -. - _ _ _ - -. ~
~
j previously deposited on colder surfaces. In addition, the detailed model assumes a very ideal-ized geometry with a plate of a single thickness between the melt surface and the vessel wall.
This may or may not be appropriate depending on the actual geometry of the structures remain-ing above the molten pool. For example, in the extreme parametric case it appears that this model was used under conditions where the assumptions would not be valid. In this case, where
)
the molten metallic layer was in contact with the lower support plate, a relatively small fraction of the molten metallic layer would be able to see the colder vessel structures. Thus the radiative sink temperature, T.,,, of approximately 920 K (Figure O.10) and the melt surface temperature, T.,i, of approximately 960 K would be inappropriate for such a configuration.
480.459 a') What calculations have been performed to insure that the absorption and emittance by the atmosphere in the vessel can be neglected, particularly if aerosols are being released from the surface of the metallic layer?
b) What calculations have been performed to verify that the aerosol generation and deposition rates for the conditions represented by Figure 2.2 would result in the absence of aerosols in the vessel atmosphere? If available, please provide supporting MAAP results conceming the atmosphere above the pool.
c) A very idealized geometry was assumed for the radiation heat transfer. What thicknesses, Ss and So, and surface area, S., were used in the calculations presented in Chapter 7 for the Figure 2.2 scenario and the " extreme parametric study"?
Structural Analyses 480.460 The objective of Section 4 is to demonstrate that structural failure won't occur in cases where the vessel wall is reduced to critical heat flux limited thicknesses. The following logic was applied by Theofanous et al. to prove that vessel failure will not be governed by structural failure.
Section 4.1 establishes the required thickness to support dead loads (no thermal stresses) when the vessel is at full strength, in this section, it is concluded that the required thickness is 0.15 mm.
Section 4.2 should demonstrate that after thermal stresses are considered, the vessel still has the capacity to support the dead loads. It states that the load carrying capacity is govemed by the wall thickness that is under tension and that this thickness is approx-imately 5 mm.
Though never directly stated, the reader is supposed to compare the 0.15 mm of required wall thickness to the 5 mm of wall under tension and conclude that the vessel can easily carry the dead load.
The structural analysis considers the vessel as a thin-walled hemispherical shell. The vessel is assumed to be completely depressurtred, so only dead loads and buoyancy forces are consid-ered. The vessel is also assumed to have a linear temperature distribution through its thickness, with an outside wall temperature of 400 K and an inside temperature of 1573 K. The upper tem-perature is based on the liquidus temperature of a specific iron /zircaloy eutectic composition.
An approximation to a closed form solution for thermal stresses in a cylinder (Equation 4.1),
away from end conditions, was use to determine the distribution of stress in the ablated section of the vessel wall at the melt poci's metallic layer interface. This development neglected the radial stress component which has a lower maximum magnitudes throughout the wall thickness than the azimuthal or axial components of stress. INEL calculations of the radial stress, calcu-lated elastically, indicate radial stresses of about 10 MPa compared to the elastic tangential stresses of around 1500 MPa. These tangential stresses will be limited to the yield stress at tem-perature, which is about 250 MPa (See Fig. 4.6) in the segment of the vessel wall expected to carry the dead loads. Therefore, neglecting the radial stress is appropriate. With that assumption, 12
j i
Section 4 indicates that there is an inner region of the vessel wall that remains elastic and is l
available, with the outer segment in the vessel wall that is in tension and slightly yielded due to i
thermal stress, for some additional load-carrying capacity to withstand the net gravity force of the core melt after the buoyancy force of the water in the flooded reactor cavity is considered, if one assumes elastic-perfectly-plastic material, there can be no additional tensile load-carrying capac-ity in the material previously tension-yielded. So the load capacity margin of the vessel discussed j
in Chapter 4 is not nearly as high as is inferred.
A more rigorous indication of structural capacity of the ablated section was demonstrated by the elastic-perfectly plastic thermal structural analysis performed using a finite element approach in the Addendum to Chapter 4. This additional analysis was meant to answer reviewer concems for i
ductile tearing and answered several concems left open in the original development. This analy-sis apparently did not include creep effects but did give a good idea of how thermal stress would
)
redistribute and confirmed the existence of an inner segment of the vessel wall that cou!d be called upon for longer term dead weight loading capacity of the vessel wall. There was no detail l
on the boundary conditiens, mechanical loading, or mechanical properties used and one is left to assume that these have been appropriately modelled for the problem at hand.
Approximate Solution i
480.460 a)
The approximate Equation 4.1 is meant to determine stresses in a cylinder's wall well away i
from its and conditions. What is the effect of end conditions (the vessel head below and the j
thicker cylinder wall above) on the stresses calculated in the ablated area?
b)
The stress goes up more than 10 fold for a system pressure differential of 1.5 MPa. Table 7.4 indicates that IVR may be use as a mitigation scheme for cases 1 APC and 3DC which we i
partially depressurized conditions. Given that IVR is to be considered for such conditions, i
please discuss the margin in the vessel for cases with low pressure.
]
c)
On page 4-3, Theofanous et al. state:
'Thus the load carrying capacity of the vessel will be governed by the outer thickness of the wall that is under tension. Below, we will show that for the conditions of this problem, this outer region has a thickness of about one fifth the wall thickness; that is -5 mm thick for the j
limiting case considered in Figure 4.2."
i The load carrying capacity of the vessel is not govemed by the outer wall under tension. In a
}
hanging vessel, such as this one, dead loads create tensile stresses. Given that an elastic-i perfectly plastic assumption was assumed in the analysis and that the thermal stresses have plastically yielded the outer tensile region, the load carrying capacity for dead loads is gov-l orned by the portion of wall which has not yielded in tension (the inner four fifths). Referring j
to Figure 4.6, most of the outer one-fifth wall has yielded in tension and is not available to carry more tension.
d)
SA508 and SA533B1, two typically used pressure vessel steels, have structural properties that are quite similar. However, SA106B has considerably different properties at higher tem-peratures. Neither SA1068 nor SA533B1 maintain full strength (355 MPA) up to 900 K. At 900 K, the yloid strength for SA106B drops to 133 MPais and SA533B1 drops to about 200 MPs.1e Why was the full yield strength used in this strength determination? The lower actual yield stresses at temperature reduce the vessel's capacity accordingly. Please demonstrate that me AP600 steel properties are closer to the properties of SA106B than SA533B1 as stated on Page 4-4.
e)
The reader is referred to an unspecified place in Section 7 to see the basis for the liquidus temperature, although the phase diagram is actually found in Figure 6.1. Note that Figure 6.1 shows that the liquidus temperature for this eutectic composition may vary by several hun-13
~ _ _ _ _
dreds of degrees. However, Section 4 only indicates that one temperature was considered.
As discussed above,if the range of possible temperatures were considered along with other uncertainties in predicting the wall thickness, the range of wall thicknesses shown in Figure 6 are possible. Please revise this analysis assuming an appropriate minimum wall thickness.
480.460 f)
This analysis claims all of the unyielded inner-wall region and the outer plastically stressed region due to the accident thermal loads can be applied to the vessel's ultimate capacity to hold the dead weight loads. While there is some indication that the vessel might withstand these losds, this calculation alone does not adequately prove that, and margins of safety are not clear from the analysis in the original part of Chapter 4.
- 9) The finite element analysis was not well defined. What were the mechanical boundary condi-tions of the vessel finite element model? What was the mechanical load applied? What mechanical properties were used? What were the inner surface wall temperatures above the 72 degree location shown in Fig. 4.7?
h)
The original content of Chapter 4 (Page. 4 6) refers to Appendix G of the report for a support.
ing argument that creep will not substantially change the result. Appendix G discusses a creep analysis of a vessel with internal pressure of 400 psi and a 5 cm. thickness and was run for an accident time of 15 hours1.736111e-4 days <br />0.00417 hours <br />2.480159e-5 weeks <br />5.7075e-6 months <br />. While this analysis is interesting for general trends of stress redistribution in the vessel wall, it offers little guidance in determining margins to failure for the specific, ablated vessel wall case considered in Chapter 4.
i) in Figure 4.6, the authors present results of a finite element analysis which includes only ther-mal stresses. Were dead loads included in this finite element analysis? Figure 4.6 also indi-cates that the wall thickness is 2.5 cm; whereas Figures 4.1 and 4.2 shows that the wall thickness in the critical region (between 70 and 90 degrees) should be approximately 2.5 cm.
As discussed above, INEL analyses indicate that the minimum wall thickness may be consid-erably lower. Please repeat this analysis considering dead loads and an appropriate mini-mum wall thickness.
j) The analysis of the Addendum to Chapter 4 is meant to address ductile tearing and calcu-lates strains in the 7 to 18% range. Please quantify the strain levels that would be of concern.
Please discuss if these strain levels are temperature dependent.
k)
A number of different structural failure scenarios have been discussed at INEL, such as vibratory loading (from bubble formation in the cavity) of the lower nead under the degraded wall conditions, shear loads in the degraded wall due to unbalanced loading of debris bed and the buoyant force, and thermal shock effects on the vessel. Thermal shock was t,rought up by the reviewers, Professor Olander and Dr. Tuomisto, with a focus on the possibility of brittle behavior of the vessel wall during the fuel melt relocation onto the lower head. The outer vessel wall will actually undergo tensile stress conditions during two phases of the acci-dent: (1) during cavity flooding in which the vessel wall outer surface temperature will drop at a rapid rate from about 550 K to 373 K with the vessel depressurizing from operating pres-sure to some nominally low pressure (0.5 MPa); and (2) later in time during the melt reloca-tion to the lower head when only thermal stresses cause the tensile stress on the outer surface of the vessel. The author's response to the reviewers' questions is based on the future des $n specifications for the vessel material to provide material that can withstand an accidental flooding and an end of life RTNDT sufficient to ensure that the material will stay ductile. It is not clear whether the RTNDT discussed is low enough that the above phase (1) cannot shift the RTNDT to a higher temperature level under the thermal load rate of the con-dition to cause brittle behavior and possible crack initiation. If preliminary crack damage to 14
I the outer segment of the vessel wall were to occur during phase (1), then the available tensile load capacity required in phase (2) would be diminished. Has thermal stress load rate been considered in the pressurized thermal shock of phase (1)?
}
480.460 1) The vessel material design RTNDT is given in degrees Fahrenheit on page 7 22 and in degrees Centigrade on page T-77. Please clarify which scale is appropriate.
j Typos / Errors l
480.461 a) The Equation numbers in Figure 5.8 are incorrect.
b) The footnote for Figure 5.8 is missing (see responses to Dr. Seghal's item 8 and Dr.
j Schmidt's item 6).
j c) Equation 5.12 should be Equation 5.22 in paragraph (b) on page 7-13.
d) The word, '1 ease" should be '1 east"in paragraph 2 on page B-4.
g j
e) Figures E.20 and E.21 are switched.
l f) The word, '1ashion"is misspelled in item 3 on page 016.
f
- 9) Figures V.1 and V.2 from Appendix V do not contain heat flux values estimated to have caused the hot spot and the captions for these figures are incorrect. These figures contain the average heat fluxes and vessel temperatures estimated when nominal and lower bound values from TMI-2 data (debris mass, composition, decay heat, etc.) were assumed. As noted in the Stickler reference, the global vessel temperatures estimates from either of these case were considerably higher than vessel temperatures inferred from the TMI-2 vessel steel samples. In fact, for the pressures occurring during the TMI-2 accident, these average vessel temperatures were sufficient to result in vessel failure. Hence, additional cooling, not cur-rently considered in severe accident analysis models, must have occurred during the TMI-2 accident. Section 5 of the Stickler reference discusses calculations in which local heat fluxes were superimposed on the vessel in order to induce inner vessel temperatures correspond-ing to the hot spot temperatures inferred from the TMI 2 vessel steel samples (1400 K).
References
- 1. T. G. Theofanous, et. al., in Vessel Coolability and Retention of a Core Melt, DOE /lD-10460, July 1995.
- 2. J. H. Min and F. A. Kulacki, Steady and Transient Natural Convection with Volumetric Energy Sources in a Fluid Layer Bounded from Belo w by a Segment of a Sphere - Annual Report 6, 1
July 1976 September 1977, NUREG/CR-000s, February,1976.
- 3. G. A. Greene, O. C. Jones Jr., C. E. Schwarts., and N. Abuaf, Heat Removal Characteristics of Volume Heated Boiling Pools with Inclined Boundaries, NUREGICR 1357, BNL-NUREG-51157,1980.
- 4. C. M. Altison, J. L Rempe, and S. A. Chavez, Design Report on SCDAP/RELAPS Model Improvements - Debds Bed and Molten Pool Behavior, INEL 94-0174, November 1994.
- 5. Personal conversation between J. Rempo, INEL, and T. Y. Chu, SNL, April 1996.
- 6. H. Uetsuka, " Thermal Props ty Measurement of TMI 2 Debris," presented at SARJ-94 Tokyo, Japan, November 1,1994.
15
_-__...________.._.m i
- 7. C. Allison, et. al., SCDAP/RELAP5/ MOD 3.1 Code Manuals, Volume 4: MATPRO-A Ubrary of Materials Properties for Light Water Reactor Accident Analysis, Idaho National Engineer-ing Laboratory Report, NUREG/CR-6150, EGG-2720, June 1995.
j
- 8. Letter from Westinghouse to NRC, dated April 4,1996.
l
- 9. J. T. Bittel, L. H. Sjodahl, and J. F. White, " Oxidation of 304L Stainless Steel by Steam and j
Air," Corrosion 25, pp. 7-14,1969.
i l
- 10. R. E. Wilson and C. Bames, Chemical Engineering Division Semiannual Reports, January.
June 1964, ANL-6900, August 1964, p. 239.
4 l
- 11. R. E. Wilson et al., Chemical Engineering Division Semiannual Reports, July-December j
1965, ANL 7125, May 1966, p.150.
- 12. R. E. Wilson et al. In " Chemical Engineering Division Semiannual Reports," July-December 1966, ANL-7325, April 1967, p.136.
j
- 13. S. Leistikow," Comparison of High Temperature Steam Oxidation Kinetics Under LWR Acci-i dont Conditions: Zircaloy-4 Versus Austenitic Stainless Steel No 1.4970," Proceedings of the Sixth Intemational Symposium on Zirconium in the Nuclear Industry, June 1984 pp. 763 778.
i
- 14. L Baker, An Assessment of Existing Data on Zirconium Oxidation under Hypothetical Acci-l dont Conditions in Light Water Reactors, ANLJLWRISAF 83 3.
i 15.J. L Rempe, et al., Light Water Reactor Lower Head Failure Analysis, NUREGICR 5642,
{
EGG-2618, October 1993.
i
- 16. L A. Stickler, et al., Calculations to Estimate the Margin to Failure in the TMI2 Vessel, NUREG/CR-6196, TMi V(93)EG01, EGG-2733, March 1994.
j l
i
)
i t
I 16