ML20042F339

Suppl 1 to Responses to Questions on C-E Rept CEN-386-P, 'Verification of Acceptability of 1-Pin Burnup Limit of 60 Mwd/Kg for C-E 16X16 PWR Fuel.'

ML20042F339
Person / Time
Site:	Arkansas Nuclear
Issue date:	04/30/1990
From:	ABB COMBUSTION ENGINEERING NUCLEAR FUEL (FORMERLY
To:
Shared Package
ML19302E106	List: ML20042F335 ML20042F339
References
CEN-386-NP-S01, CEN-386-NP-S1, NUDOCS 9005080197
Download: ML20042F339 (17)

Text

. . . . . _ . . . . . . . . _

CEN-346-NP $UPo 1-NP m

RESPONSES TO QUESTIONS ON COMBUSTION ENGINEERING REPORT CEN-386-P

" VERIFICATION 0F THE ACCEPTABILITY OF A 1-PIN BURNUP LIMIT OF 60 20/KG FOR COMBUSTION ENGINEERING 16X16 PWR FUEL" i

m April 1990 ABB Combustion Engineering Nuclear Power 1000 Prospect Hill Road

, Windsor, Connecticut 06095

"@['0$$k$ bb '.,88

h* "

1 LEGAL NOTICE .

THIS REPORT WAS PREPARED AS AN ACCOUNT OF WORK SPONSORED .

BY COMSUSTION ENGINEERING, INC. NEITHER COMSU$T80N' ENQiNEERING.

NOR ANY PERSON ACTING ON ITS BEHALP:  !':

A. MAKES ANY WARRANTY OR. REPRESENTATION, . EXPRESS OR IMPLIED INCLUDING THE WARRANTIES OF PITNESS POR A PARTICULAR . , .

PURPOSE OR MERCHANTABILITY, WITH RsSPECT TO THE ACCURACY..

COMPLETENESS, OR USEPULNESS OF THE INPORMATION CONTAINED IN THIS '

' REPORT, OR THAT THE USE OF ANY INPORMATION, APPARATUS, METHOD.'-

OR PROCESS DISCLOSED IN THIS REPORT MAY NOT INFRINGE PRIVATELY' OWNED RIGHTS; OR S. ASSUMES ANY LIABILITIES WITH RESPECT TO THE USE OF, OR FOR DAMAGES RESULTING PROM THE USE OP, ANY INPORMATION, APPARATUS,-

METHOD OR PROCESS DISCLOSED IN THIS REPORT.

s cn

a> , , ,

l, L

l QUESTIONS and RESPONSES Ouestion fl! '

l The burst- s1Iain data from fort Calh'oun cladding with local burnup levels .

[ between $$ to 63 mwd /kgM, presented in Section 4.1.5.a of the topical report, show very low claddittg strains between 0.03 to .0.1%. Section  :

'4.2.II.A.2(g) of the Standard Review Plan ' (SRP) -(Reference 1), which ,

i-addresses ' acceptance ' criteria' to preclude pellet / cladding interactica  ;

' (PCI) failures, states that uniform strain (elastic plus' plastic) of the cladding should not exceed 1% for normal operation and' anticipated -

operational occurrences (A00s). This strain limit has also traditionally been applied as'a limit for cladding strain in Section 4.2.II.A.l(a) of the

~

SRP. The cladding burst data- from Fort Calhoun' suggests that Combustion i Engineering (C-E) fuel cladding may fail at uniform strains significantly [

below the 1% strain limit recommended in the SRP, Therefore, several

! questions arise from the data:

j l E

l. (a) How applicable are the burst tests and measured strains from the Fort >

l Calhoun cladding to failure mechanisms from normal operation and A00s j in C-E commercial reactorst This response should address those j fa11Lre mechanisms identified in Section 4.2 of the SRP. If this data L is applicable to these failure mechanisms or if their applicability is

{ unclear, please address the following additional questions.

(b) Should the uniform strain limit of C-E cladding for normal operation and A00s be decreased to a level below 1% when local bcrnups exceed 55 l mwd /kgM?

i d,

i (c) Will the fuel cladding become even more embrittled at local burnups I.}7 above 63 mwd /kgM?

[

(d) Will fuel failures become more frequent' as the number of fuel- vods t that exceed local burnups' of 55. mwd /kgM increa ses from commercial L

operation?

[

J l l l.

l '(

~~

QUESTIONS and RESPONSES

(-  !

l Egiponse to Question #1  :

i .

I Use of the high strain rate burst test data'to evaluate the cladding strain  !

capability against pellet / cladding interaction (PCI) failure caused by i normal operation or A00's is conservative for the following reason: ,

c ,

i Strain rates anticipated during normal operation and A00's are expected to I e be lower than those employed in the burst' testing. At lower strain rates, .

the material ductility is expected to be higher. i In addition, it is to be noted that the data presented in Section 4.1.5.a [

(Page 21. first paragraph) of the topical report refer to uniform plastic l circumferential burst strain of fuel cladding. These values do ne'. contain the elastic component. At local burnup levels .of 54.7 to 62.5 mwd /kgU, the '

! measured uniform plastic strains ranged from 0.03 to 0.11%. The uniform  ;

plastic strains were calculated by subtracting the elastic component from  ;

I-the total uniform strain corresponding to the maximum pressure point on the ,

j pressure volume expansion curve for the burst specimen (corrected for the p Av response of the system induding the specimen). Elastic strain capability of the cladding needs to be added to-the uniform plast!c strain to obtain the total uniform strain (elastic plus plastic strain) capability of the material. Since the elastic strain of the sample cannot be accurately determined from the test data, an estimate of the sample elastic-strain was obtained as follows: The measured yield strength of cladding in the same burnup range as quoted above ranged from 115 to 125 ksi. The

• reported Young's modulus of non-irradiated Zircaloy at 600'F is -10x106 psi 3-(Reference 2). Assuming no changes in the Young's modulus as a result of j

irradiation, the elastic strain capability of the cladding is estimated to  :

be at least [ ). Therefore, the above data show that >

l-the total (elastic plus plastic) strain capability of the cladding ranges from ( ). Thus, the data on irradiated cladding show that-the measured strains exceed the 1% minimum limit recommended in the SRP.

5 l

l QUESTIONS and RESPONSES Resoonte to Question #1 (continued)

Consideration of the following factors further reduces concern for .

PCI-related failures at extended burnups.

1. At extended burnups, the heat rates of the fuel rods are significantly lower than the heat rates eariter in life. The heat rates at extended s burnups (>52 mwd /kgU) are expected to be significantly lower than the power threshold for PCI failures.

2. As a result of the establishment of tight fuel pellet cladding contact at high burnups, an interface interaction layer (main constituents Ir, U, and 0) forms between the fuel pellet and Zircaloy cladding. This interface layer helps to distribute the mechanical load exerted by the fuel pellet on the cladding. Once the interface layer forms, the application of a concentrated localized stress (necessary for the PCI-failures) is less likely. This layer also improves heat transfer ,

between the fuel pellet and cladding. As a result, fuel-swelling induced stresses in the cladding are expected to be reduced in. the case of power transients.

3. C-E has successfully irradiated fuel rods to local burnups of about 60 mwd /kgU in three commercial PWRs: ANO-2, Calvert Cliffs-1, and Fort Calhoun. The C E fuel exhibited satisfactory performance both during the normal operation to 60 mwd /kgU burnup and subsequent post-irradiation handling. Failures associated with reduced ductility of the irradiated material were not observed. This C E experience indicates that the erobability of cladding failure due to reduced ductility is low and setting a strain limit below 1%.is not-required.

In summary, the burst test conditions are more severe for ductility considerations compared to the conditions that are analyzed for normal operation and A00's. However, even under the more severe conditions

i l

l l QUESTIONS and RESPONSES l

ResoonsetoQuestion#1(continued) l _

i

imposed during the burst tests, the C-E cladding mechanical property data l exhibited total uniform (elastic plus plastic) strains higher than the minimum 1% limit. Therefore, it is concluded that the failure strain of

- C-E cladding for every failure mechanism identified in Section 4.2 of the '

SRP will be greater than 1%.

J References

1. U.S. Nuclear Regulatory Commission, July 1981. 'Section 4.2, Fuel System Design." Standard Review Plan for the Review of Safety Analysis Reoorts for Nuclear Power Plants . LWR Edition. NUREG 0800, Revision 2. U.S. Nuclear Regulatory Commission, Washington, D.C.

2. D. B. Scott, " Physical and Mechanical Properties of Zircaloy-2 and Zircaloy-4", WCAP-3269 41, May 1965.

l l

I 1

1 1

QUESTIONS and RESPONSES

~

Ouestion M Please discuss how the standard deviation, a, is calculated for oxide thickness at rod average burnup levels of 60 mwd /kgM from the 14x14 and 16x16 fuel rod data in Section 4.1.2.1.a. Should this data be separated-because there are inherent differences 'betwen the corrosion behavior of l 14x14 and 16x16 fuel rods or should they . be combined because the differences in corrosion are- not uniquely design dependent? -Also, it appears that the estimates of x + 3a,- provided in this section, assume that a is independent of burnup, while the corrosion data in Figures 4.1.2.a-1 i and 4.1.2.a 2 suggests that a becomes larger at higher burnups. What i impact would a more variable and larger calculated a have on the 3

performance analyses of the 16x16 fuel rods (e.g., cladding stress) at- 60 mwd /kgM?

Please provide the effect as a percentage change 'from the condition of no cladding wastage due to corrosion. >

I Resoonse to Question M The Calvert Cliffs and Fort Calhoun data are shown in Figure 4.1.2.a-1 in the report for information. They were not included in the curve fit. The

{

curve fit and the standard deviation were determined from just the ANO-2 fuel rod data. This is shown in the attached Figure 1, which plots the data, the curve fit, and the 3a limit. Figure I shows that the 3a line is a conservative bound to the ANO 2 ' fuel rod data and that the standard i deviation for these data .is not burnup dependent. As mentioned above, Figure 4.1.2.a 1 includes data from reactors other than ANO 2.

j Consequently, the spread noted in that figure reflects the influence of a

{

variety of plant specific factors and not just burnup alone. These factors include differences in the fuel rod operating history (e.g., heat flux),

coolant temperature, and the temperature at the oxide / metal interface.-

I i

j 5-

.--r_ ._ _ _ .

e-3 -

< 'm 2 w 2 >

a m ,

i W

m m .

O.

@ g OO .

@ O ,

O O .

g .. t y ,_ - -

'- c A

w cn.

t O x>

o

.x N

t U

Ol uo l "

3:

2 Oea ~o~

s z

2o w --tw--a x

o

<we m

2

- C- -

m -

I k ., Cm i, ,

! an "

C-

. a .

. 1 i i ,

1 -

i L

j i

~

.,. I, 1

i i 4 i i > i i o

, , , i. m o o o o o o o o o N

- o e m N e e v o

m-e n

o o ,

t (sves otus) SS3N>iOlH1 3ClXO  !

6-,

. - ~ . . , .

QUESTIONS and RESPONSES Ouestion #3 The results of the cladding collapse calculation -in Section 4.1.4.a are based on an assumed finite " hot" axial gap length in the fuel column -of modern C-E designs. In this analysis it is implied from post-irradiation examination (PIE) data that " hot" axial gaps greater than this assumed size have a low probability of existing in C E's 16x16 design, but no probabilities are calculated ' based on this PIE data. What is the probability of the C-E 16x16 design having an axial gap-of this assumed size' or greater? The probability . may be calculated using PIE data . of

" cold" axial gap sizes measured from modern fuel designs other than. C-E's 16x16 - design, but these designs should be comparable in fuel length, densification characteristics, and density. A correction between

• hot" and measured " cold" gap size is permissible but assumptions made in this B

correction should be stated. The calculated probabilities may also take into account that today's fuel designs typically form smaller' axial gaps in the fuel column than previous " Sider"- fuel designk because of changes in fuel fabrication. For example, three separate populations of axial gap size can be identified according to the-following fuel characteristics 1) older densifying fuel, 2) older nondensifying fuel with low fuel densities (i.e., <94% theoretical density), and 3) newer nondensifying fuel with higher fuel densities (i.e., >94% theoretical density); with fuel with the latter charach istics displaying the smallest . axial gap sizes following g irradiation.

Response to Question #3 As discussed in Section 4.1.4.a. C-E performs cladding collapse calculations using very conservative input asst $ntions. The criterion for selecting the length of the hot axial gap in the fuel column is that the value must be at least as large as the maximum predicted hot axial. gap, at 95% probability and 95% confidence level. Predicted hot axial gaps are based on adjusting cold measured axial gaps to hot operating conditions.

QUESTIONS and RESPONSES -

Resoonte to Question #3 (continued)

L Axial gap data were obtained as a result of a post-irradiation examination of fuel from the San Onofre Unit 2 Cycle 3 core. This fuel is typical of the current generation 16x16 (and 14x14) Combustion Engineering high density nondensifying fuel. Thirty axial gap measurements were obtained from 17 fuel rods. These cold axial gap measurements were analyzed and

" hot" axia? 3aps were determined by accounting for axial thermal-expansion.

The largest cold gap measured was~ 0.9 inches. It was calculated that thermal expansion of the fuel column during reactor startup reduces the largest cold gap to 0.3 inches at normal operating conditions.

The calculation of axial thermal expansion was based on an evaluation which considered the effects of axial varietion in linear heat rate, local pellet clad gap conditions (which affect fuel temperatures), changes in these parameters with time and/or burnup, and the existence of more than one gap (where applicable) in a single. fuel rod. 'It was assumed .that the -

fuel column segment below an observed gap would thermally expand axially upward to reduce the gap size. Once the axial gap is closed, the fuel

{

column above the gap would be pushed -upward, closing successive gaps, if they existed.

i Based on the assumption that the distribution of the data is normal, statistical analysis of the hot' gap data has been performed. The finite hot axial gap length assumed in the cladding collapse calculation ~ is well in excess of that expected at a 95% probability and 95% confidence-level.

Reference 3-1 3-1 CEN-386-P, " Verification of the Acceptability of a 1 Pin Burnup t.imit ,

of 60/ mwd /kg for Combustion Engineering 16x16 PWR Fuel",- Combustion Engineering, Inc., June 1989.

8-

• l

QUESTIONS and RESPONSES Question #4 i

From the small number of FATES 3B predictions of low temperature fission gas release data provided in Table 4.1.6.a 3, it appears that this code may be underpredicting this data by a small amount (i.e., I to 2% release when rod average burnups exceed 54 Wd/kgM). What is the effect, if any, on fuel performance ca%1ations at low temperatures if the FATES 3B code predicts

1% release when 2.5% release'is the actual amount releas6d?

Rg nonte to Question #4 i

An empirical model to predict fission gas release was developed by C E for use in FATES 3 and was described in CEN 161(B)-P, (Reference 4-1). The-model accounted for the effects of temperature, burnup, and grain size, and i was calibrated against the data from 00 fuel available at that time. The 2

(

FATES 3 . verification data were limited to'.burnups up to 48 mwd /kgu.  !

Although the burnup of the verification data base for FATES 3 extended up to i 48 mwd /kgu, the fuel centerline temperatures associated with the highest burnup data were characteristicelly below 1250'F. The NRC completed the '

safety evaluation and approved the FATES 3 model for C-E safety analysis but i also im9md a restriction on the grain size used for fission gas release calcu 61ons (FATES 3A Reference 4-2). Once additional high-burnup, high-temperature experimental data became available, C-E analyzed these  !

data and found that modifications to the FATES 3 fission gas release model were required to predict these high burnup data on a best-estimate basis.

Changes were incorporated into FATES 3 (forming the FATES 3B version, Reference 4-3) to increase the burnup dependence, temperature dependence.

and modify the kinetics of grain growth. The model was specifically' tuned to predict high burnup, high-temperature fission gas release at power .;

levels and temperatures characteristic of- fuel performance licensing calculations at high burnup. Consequently, the FATES 3B predictions of gas release from high burnup (>54' mwd /kgU), low-power test rods reported in Table 4.1.6.a-3 of Reference 4-4, are slightly underpredicted, but by less 9

, m

QUESTMNS and RESPONSES l' t Response to Dunstion #4 (continued) l than 0.5% on average. Based on the formulation of the model of Reference  !

4-3, the high temperature release is not affected by this low temperature' underprediction.

For licensing calculations the power history data are selected to give .

conservatively high fuel temperatures, high fission gas release, and high hot internal gas pressure. Therefore, because fuel performance licensing  :

, calculations are not. performed at low temperatdre, the slight'

! underprediction of the data in Table 4.1.6 a 3 (Reference 4 4) would have an insignificant effect on licenting calculations. '

(

l However, the impact of an additional 1.5% fission gas release at high  ;

burnup was evaluated. The degradation in gap conductance due to the '

additional gas release results in an insignificant increase in fuel temperatures, on the order of 3 5"F. at 7 8 Kw/ft. The increase in rod internal pressure at 55.0 mwd /kgU due to an additional 1.5% fission _ gas '

release is less than 100 psi (75 psi- for a typical low temperature Calvert Cliffs test rod). However, it should be noted that the average underprediction is only on the order of 0.5% as discussed above. 4 t

References:

4-1. CEN-161(B)-P, " Improvements to Fuel Evaluation 'Model", Combustion l

Engineering, Inc., July 1981, i l_ 42.CEN-161(B)-P-A, " improvements to Fuel Evalut. tion Model" Combustion Engineering, Inc., August, 1989.

• 4-3. CEN 161(B)-P Supplement 1-P, " Improvements to Fuel E' valuation Model",

Combustion Engineering, Inc., April 1986. k

• i

(

c 4

e + - . - ,, .a. , . . . , , - . . . , . ..n.... .-- -a,b,,.- ~ ~ . . . , , , .,.-.n.m

3f QUESTIONS and RESPONSES Resnonse to Question #4 (continued) -

4 4. CEN 386 P,'" Verification of the Acceptability of a 1 Pin Burnup Limit of 50 mwd /kg for Combustion Engineering 16>JL PWR Fuel". Combustion Engineering, Inc., June, 1989.

i l

i

-l

, i

-l i,

(

t

=

t ,

y --,

[l _g..a

I l

l QUESTIONS and RESPONSES Question #5 In Section 4.2.1.a. on guide tube wear, it is stated that References 4.2.1.a-2 and 4.2.1.a 3 have verified the conservatisms in C E analytical predictions of guide tube wear for these assemblies. Please provide a comparison of measured and predicted maximum wear for 16x16 unsleeved assemblies along with their burnup levels. What is the maximum wear predicted for the C E 16x16 assemblies at the maximum residence times expected for the burnup levels requested?

Response to Question #5 '

Figure 4.2.1.a 1 (attached) provides a comparison of measured and predicted guide tube wear for unsleeved System 80 fuel assemblies (System 80 is the only unsleeved 16x16 fuel design). . Assessments of the measured guide tube wear signals (voltage readings from Eddy Current Testing) employed conservative assumptions to maximize the calculated' volume loss associated with the ~ wear indications. The measured wear volumes shown in Figure 4.2.1.a-1 are the maximum volumes from any guide tube measured during the two inspection caapaigns (eighty guide tubes inspected at Palo Verde 1 and forty guide tubes inspected it Palo Verde 2). -

l l

1 Inspection of Figure 4,2.1.a-1 shows significant margin between the maximum measured wear volumes and the corresponding predicted wear volumes. The maximum predicted wear volume at the maximum residence time associsted with  !

the extended burnup levels is

[ ).

Analyses have been performed that demonstrate that the minimum amount of volume loss that a guide tube could sustain without violating any design i criteria is [

). The' unsleeved System 80 design is, therefore, concluded to be acceptable for operation to the extended burnup levels since the maxirum predicted wear is less than the minimum wear necessary to violate any design criteria and since there is significant margin between the marimum measured wear volumes and their associated predictions.

1 12-

FIGURE *1.2.1.a-1 COMI'ARISON OF MEASURED AND 1*REDICTED GUIDE TUBE WEAR VOLUMES .

m td T

O Z

. 52

m D

O -

ta '

2 D

-J O

rr

<t td 3

i Ol'El<A~llim TIME, llI

slightly higher than the levels for which irradiation performance  !

l characteristics (helium release, internal void volume, and pellet swelling) '

were determined, as reported in CENPD 269 P, Revision 1-P. The C E model i accounts for the effect of higher levels of boron carbide by increasing the total volume of helium gas released and by. increasing the pellet swelling.

No additional post irradiation examinations have been . performed or are planned.

l 1

I 15-l i

_ _ _ _ _ _ - _ _ . _ _ _ _ _ - . _ _ _ . . - _ _ _ . _ _ . - . _ _ _ . . . , , _ . . _ , _ . - , ,