ML20042A024

Forwards Initial Review of Util steady-state Fuel Performance Code (Frosstey).Recommends Frosstey Be Subj to Further Evaluation

ML20042A024
Person / Time
Site:	Vermont Yankee File:NorthStar Vermont Yankee icon.png
Issue date:	12/29/1981
From:	Beyer C Battelle Memorial Institute, PACIFIC NORTHWEST NATION
To:	Voglewede J Office of Nuclear Reactor Regulation
Shared Package
ML20042A017	List: ML20042A019 ML20042A024
References
NUDOCS 8203220569
Download: ML20042A024 (5)

Text

4 December 29, 1981 ggg Pacific Northwest Laboratories Mr. John C. Voglewede  !

[; gi,'nd %ashington U.s.A. 993s2 Core Performance Branch Telephone <so9)

Division of Systems Integration Telex iss Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washington, DC 20355

Dear Mr. Voglewede:

T.he initial review of the Yankee _ Atomic Electric _. Comp.any steacy-state fuel performance code (FROSSTEY) has beehlo~mpldted by Pacific Northwest Labora-tory (PNL) as a part of the Steady-State Code Applications-II (B2343) pro-gram. Two sets of documentation for the FROSSTEY code review were transmitted to PNL by letters, dated June 5,1981, and August 24, 1981, from the Core Performance Branch. The PNL staff who participated in this

. initial review were M. E. Cunningham, D. D. Lanning, F. E. Panisko, W. N. Rausch, and myself. Each individual reads the FROSSTEY documenta-tion and composed comments and questions, which were discussed with you by telephone on December 2,1981. The following are general comments con-cerning the FROSSTEY code.

FROSSTEY has its origin in the GAPCON-THERMAL-2 (GT-2) code, which wL, developed by PNL, and is used by the Core Performance Branch in its audit calculations. The FROSSTEY code includes modifications which in several areas, e.g., rod internal pressure calculations at low burnups, appear to improve the predictive capabilities compared with the original GT-2 code.

It is not clear whether alterations in other areas are actually improve-ments because of the lack of data needed to support these changes, particu-larly at ..igh burnups. Those model changes in FROSSTEY that are judged to have a significant effect on fuel performance predictions are:

e Fission gas release - includes a burnup enhancement effect at low temperatures and burnups above 25,000 mwd /MTM e Fuel thermal conductivity - includes the change in thermal conduc-tivity cue to pellet cracking by assuming that the Maxwell-Eucken porosity correction applies e Clading creep - implicitly calculated in the code with the BUCKLE equations e Internal rcJ pressure - includes fuel crack volume. ,

~

G 8203220569 820222 PDR ADOCK 05000271 P PDR

Mr. John C. Voglewede .

December 29, 1981 Page 2 .

The FROSSTEY code appears to do a reasonable job of.predictir., beginning-of-life and low burnup fuel performante data, i.e., fuel temperatures, internal pressures, etc. However, the code is verified against a very small amount of gas release and rod internal pressure data, i.e., 4 and 6 fuel rods, respectively, and these are at low burnups (3,000 to 14,000 mwd /MTM). Consequently, the burnup dependence of the enhanced gas release model used in conjunction with the Beyer-Hann nodel cannot be verified in the burnup range where its effect becomes significant, i.e.,

>25,000 mwd /MTM. . -

The most significant limitat#on of the original GT-2 code was its lack of verification against fuel performance data above about 10,000 mwd /MTM burnup. High burnup fuel performance data have been used since to verify and, in some cases, modify the original published version of the GT-2

' code. From the documents submitted (YAEC-1249P and YAEC-1265P), it is clear that FROSSTEY has not been verified against high burnup data. For example, the majority of the fuel rods used in the FROSSTEY verification were at beginning-of-life with only 2 rods exceeding 10,000 mwd /MTM burnup,.

i.e., both these rods had average burnups of ~14,000 mwd /MTM. Fuel .is cur-rently being irradiated to -30,000 mwd /MTM and future cores are expected to reach 40,000 to 50,000 mwd /MTM. Consequently, the review and verifi-cation of this code beyond 14,000 mwd /MTM burnup will be limited to com-parisons against other fuel behavior models and those benchmark fuel performance data selected by PNL and NRC.

Specific questions concerning the F' OSSTEY R code are:

1. Figures 2.6, 2.7, 2.8, 2.11, and 2.12 (within YAEC-1265P) show a sharp upward bend in the centerline temperature versus power level plots. This is not seen in GT-2 calculations or other p.iblished fuel performance codes we are aware of, nor is it apparent in Vuel perfor-mance data.

Q. What is the reason for this sharp change in slope for. these plots?

2. Fuel densification values are used as input for verification of FROSSTEY.

Q. Where or how were these densification values obtained? Many of

the or'iginal references from which these data were taken did not "

quantify the amount of fuel densification.

3. The FROSSTEY code calculates stored energy by two methods.

Q. Which method if any, is used for fuel licensing applications?

e O

Mr. John C. Voglewede December 29, 1981 Page 3

4. Helium gas production and release does not appear in the FROSSTEY documentation. Postirradiation puncture data from an experimental (nonpressurized) Halden fuel rod (Rod 8 from IFA-432 a,sembly) irra-diated to 22,000 mwd /MTU burnup shows more helium to be present

(-25%) than would be expected from the initial amount of fill gas introduced during rod fabrication (Reactor Safe'ty Research Programs Quarterly Report, October-December 1980, S. K. Edler, ed.;

NUREG/CR-1454/PNL-3380-4).

Q. Would the inclusion of a helium production and release model in FROSSTEY have a significant effect on end-of-life rod pressures in YANKEE ATOMIC plants?

5. The FROSSTEY code includes fuel grain growth and restructuring models (Section 6.5, YAEC-1249P).

Q. Are these models used in licensing analyses performed by FROSSTEY? If these models are used, additional evaluations may be needed. It should be pointed out that most cocmer.-

cial fuel operates below the restructuring temperatures.

6. Equation 3.8b used for cladding corrosion (Section 3.4, YAEC-1249P) does not appear to agree with the corresponding equation in MATPRO 11.

Q. Are the two equations equivalent? If not, why not, since MATPRO II is referenced as the source of this model?

7. The FROSSTEY code uses the lower bound of an irradiation cladding growth model developed by EPRI (Section 4.5.2, YAEC-1249P).

Q. Is the EPRI model appropriate for cladding used in Yankee Atomic plants? Using the lower bound may be conservative for calculat-ing rod internal pressures, but is this also used for calculating rod clearance with the top grid plate of the assembly? Also, please provide a figure plotting AL/L versus time. for a variety of fluences typical of your commercial operation, since we have had difficulty in comparing the FROSSTEY model to other irra-diation growth models with which we are familiar.

8. It appears that a factor of w (calculation of internal void volume) , -s is missing from Equation 5.25 (Section 5.4, YAEC-1249P). .

9. It is stated that a "new computational structure has been devised to allow an unlimited number of time-steps in a power history analysis,"

using Sou1hier and Notley's approach (Section 6.9, YAEC-1249P). From the description provided, it is not clear how the Soulhier and Notley approach could have unlimited time-step capability.

Q. Please supply more information to assist in our understanding of how this is accomplished. '

. l

.. l Mr. John C. Voglewede -

December 29, 1981 Page 4

10. It appears that FROSSTEY will be used to calculate the behavior of urania-gadolinia fuels.

'Q . How will the material properties be modified to 6: count for the effects of gadolinia additions to urania.

11. Fission gas release and fuel swelling models are both dependent on fuel temperature and burnup.

Q. Are radial as well as axial burnup variations included in the local calculation at each fuel node? For example, is burnup calculated individually for each radial node and used in cal-culating the release for this node?

'12 . Three sets of constants are given for the solid gap conductance, H solid, calculation (Table 5.1, Section 5.2, YAEC-1249P).

Q. Have these constants been comparea against Garnier and Begej-data (Ex-Reactor Determination of Thermal Gap and Contact Conductance Between Uranium Dioxide: Zircaloy-4 Interfaces Stage I: Low Pressure, NUREG/CR-0330, PNL-2696; also see NUREG/CR-1224, PNL-3232 for Stage II: High Gas Pressure report)? Also, which of these sets of constants is used for fuel licensing applications?

13. It is not clear that decreasing the thernal conductivity of the fuel to allow for the effects of fuel cracking results in a conservative estimate of stored energy. This modeling approach may increase the centerline temperature; however, it will also decrease the fuel su'rface temperature which can result in a lower calculated stored energy.* Reference 47 (in YAEC-1249P), used to support the conser-vatism of using a thermal conductivity degradation model, does not relate to the conservatism of such a model in a fuel performance code i for stored energy predictions. This unpublished reference emphasizes j that the reduction in thermal conductivity associated,with pellet l cracking biases and increases the uncertainties in estimated stored energy derived from fuel temperature measurements.

Q. Please demonstrate, in a quantitative manner, that FROSSTEY _ _s predictions of stored energy are conservative. -

1 ~

l *The fuel volune at the surf ace is much larger and, because stored energy

! is related to volume-average temperature, the decrease in surface tempera-ture will have a larger effect on stored energy than the increase in centerline temperature.

t I

- O,

l Mr. John C. Voglewede December 29, 1981 Page 5 -

14. The Lloyd model or temperature jump distance model has been shown to give nonconservative (low) values relative to other accepted ~ gap conductance models and also relative to measurements. The multiplication of the Lloyd model values by 1.8 occurs in both the open and closed gap regimes in GT-2, and this modification is accepted as predicting realistic, if not conservative, values.

However, FR0SSTEY fails to modify the Lloyd value in the open-gap regime, except for the very-near-contact-condition.

. . . . . ~ - . . . - -

Q. Please quantify the effect of not using the 1.8 value in the open-gap region on gap conductance.

The following specific comment is offered concerning the buenup-enhanced fission gas release model used in FROSSTEY. The Bellamy and Rich model

'that is used to model a burnup enhancement, is based on the mathematics for athermal knockout release. This model is contrary to current theories for burnup-enhanced release which indicate that it is a thermally-activated process, due to the collectior, of bubbles at grain boundaries which result in bubble interlinkage to a free surface. Thermally-activated release is .

considered to be operational to temperatures as low as 9000C, particu-larly at high burnups. Athermal knockout release is not considered to be dominant until fuel temperatures are below 9000C. This does not mean that the FROSSTEY model is unacceptable for licensing calculations, but rather it will be considered to be an empirical relationship that, when used with the other models within FROSSTEY, must adequately predict high burnup gas release data.

. Based on this initial review, we recommend that FROSSTEY be subjected to further detailed evaluations, particularly the fission gas release, fuel thermal conductivity, axial rod'greath, and cladding creep models. The extent of these evaluations will depend on how well the FROSSTEY code predicts the benchmark data chosen by NRC and PNL.

If you have any questions concerning the above, or have further suggestions for this code review, please contact me.

Sincerely,

. -s C. E. Beyer Nuclear Fuels Section CB:bjb cc: M. E. Cunningham M. D. Freshley D. D. Lanning .

F. E. Panisko W. N. Rausch

.