ML20215A516

Safety Evaluation Accepting B&W Methodology in Special Rept JHT/86-011A, Creep Collapse Analysis for B&W Fuel, for Cladding Collapse If Calculated Cladding Temps Do Not Exceed Upper Limit Defined in Rev 2 to BAW-10084P-A

ML20215A516
Person / Time
Issue date:	12/05/1986
From:	Office of Nuclear Reactor Regulation
To:
Shared Package
ML20215A466	List: ML20215A461 ML20215A516
References
NUDOCS 8612110309
Download: ML20215A516 (11)

Text

.

ENCLOSU;E SAFETY EVALUATION OF SPECIAL REPORT Jilt /86-011A,

" CREEP COLLAPSE ANALYSIS FOR B&W FUEL"

1.0 INTRODUCTION

Creep collapse of fuel rod cladding depends on the existence of two events occurring in a fuel rod: 1) The formation of an axial gap of sufficient size, i.e., greater than 0.5 inches, and 2) sufficient cladding creepdown that will permit the cladding to collapse into the axial gap.

Cladding collapse was initially observed in the early 1970s in commercial PWR fuel rods that were unpressurized during fabrication, i.e., only one atmosphere of internal fill gas. Fuel pellet hang-up and densification were found to be the causative factor for axial gap formation. Once the axial gaps were formed, the substantial pressure differential between primary system pressure and internal rod pressure subsequently led to

. cladding creep collapse into this gap.

The technical fixes used by industry to remedy cladding collapse have consisted of reducing the amount of fuel densification by changes in fuel fabrication and pre pressurizing PWR fuel rods. Since the implementation of these fixes over ten years ago, and other refinements, there have been no further evidence of cladding collapse in U.S. commercial fuel rods.

8612110309 861205 PDR TOPRP EMVBW C PDR

During the early 1970s, the regulatory staff introduced an analysis approach (Ref.1) to be used by industry to insure that future fuel designs would not experience cladding collapse. This analysis approach assumed that axial gaps existed in all fuel rods in a core such that the fuel rod was calcu-lated as a free standing tube of infinite length with no fuel pellets to prevent its collapse. In addition, the regulatory staff required each fuel vendor to utilize conservative creep models and analysis methods in the evaluation of creep collapse. Obviously, this is a very conservative approach, but was initiated by the AEC regulatory staff at a time when fuel densification, pellet hang-up and creep collapse were not very well under-stood.

Babcock & Wilcox (B&W) has proposed a new method and criterion for pre-venting creep collapse. This approach relies on the fact that a:dal gaps of sufficient size to allow collapse, i.e., gaps greater than 0.5 inches, have not been observed in commercial fuel rods since the implementation of pre pressurized rods and fuel with low densification characteristics over.

ten years ago. A discussion and evaluation of this new method and criterion for evaluating creep collapse is presented in the following sections.

2.0 EVALUATION

a. B&W's Current and Proposed Approach for Evaluatino Creep Collapse Currently, B&W performs a cycle-by-cycle analysis for creep collapse l on each fuel design. This analysis is based on the conservative i

assumption that the fuel rod is a free standing tube of infinite length without fuel pellets to prevent its collapse. The analysis must show that creen collapse is not possible in a free standing tube up to the last cycle of operation.

B&W has proposed (Ref. 2) that in place of the above approach, they use a fuel manufacturing and operating envelope in which cladding creep collapse is not expected to occur based on past experience.

The B&W envelope proposes nominal values for cladding wall thickness, initial ovality, pellet density, pellet densification, and initial fill gas pressure with lower limits for these nominal design values.

These are the important input parameters used in the creep collapse analysis. The lower limits of this envelope correspond to the 2a fabrication tolerances of their current Mark B 15x15 design, with.the exception of the lower limit for initial fill gas pressure, which falls below the lower tolerance limits of the current design. In addition, there is a limit on burnup such that the maximum rod average

burnup must remain below 55 mwd /kgM. Consequently, B&W proposes that,-

as long as their design remains within this fabrication and operating envelope, creep collapse will not occur and thus, a cycle-by-cycle analysis of creep collapse is not required. If the design falls outside of this envelope, B&W will perform a creep collapse analysis for this particular instance.

?

la order to support the conclusion (and criterion) that creep collapse will not occur within their proposed fabrication and operating envelope, B&W has presented observations of past experience (Refs. 3, 4 and 5) that show no fuel failures have occurred due to creep collapse in any B&W fuel designs and that no significant axial gaps exist in their current Mark B 15x15 design. B&W has irradiated more than 800,000 fuel rods from their 15x15 designs without any evidence of creep collapse. In addition, forty fuel rods of the current 15x15 fuel design have been gamma scanned with only very small axial gaps detected, i.e., no gaps greater than 0.1 inches, and no axial gaps were observed in rods with burnups of approximately 50 mwd /kgM.

b. Evaluation of B&W's Proposed Creep Collapse Approach The principal areas of concern in evaluating creep collapse are:

1) do axial gaps of sufficient size for creep collapse exist, and 2) if a gap exists, can cladding collapse occur within the proposed i operating envelope? The latter question will be addressed first because if creep collapse is not possible in a free standing tube of infinite length, the concern about axial gap formation does not need to be addressed. B&W has performed a conservative collapse analysis l

of their current Mark B 15x15 design utilizing the lower fabrication

~

and prepressure limits from their manufacturing envelope and found that cladding collapse is possible at a burnup of 32 mwd /kgM and

greater. Consequently, it must be determined if axial gaps large enough to allow collapse are possib?e at or above the burnup, i.e.,

greater than 32 mwd /kgM, at which cladding collapse is possible in the Mark B 15x15 design using the lower limits of their envelope.

For the B&W-designed fuel rod, there are two possible mechanisms for axial gap formation. The first is fuel densification, and the second is collapse of the bottom plenum spring.

As noted earlier in the INTRODUCTION, gap formation due to fuel den-sification was a problem in earlier fuel designs, but examination of several current vendor designs have shown (Ref. 3) no significant axial gaps since the introduction of high density, i.e., greater than 94%

theoretical' density (TD), low densification fuel. The current Mark B 15x15 fuel design utilizes a high density, i.e. , 95% TD, low densifying fuel, thus, axial gaps due to fuel densification are not anticipated for this design.

Axial gaps up to 0.3 inches were observed (Ref. 3) in an earlier Mark B 15x15 design from Oconee-2, in which the fabricated fuel density was 92.5% TD. B&W has stated that they believe the higher fuel density _of the current Mark B 15x15 design, i.e., 95% TD, is the reason for the smaller axial gaps observed in the current fuel versus those observed in the earlier lower density fuel in Oconee-2. The staff concurs with this evaluation.

w The B&W-designed fuel rod is unique to other vendo; designs in that it utilizes a bottom plenum and plenum spring. The bottom plenum spring function is to hold the fuel stack in place. The concern in this review is relaxation of the plenum spring as a result of irradiation and thermal creep of these springs. If pellet hang-up were to occur prior to relaxation and collapse of the bottom plenum spring, it is possible that an axial gap could form in the fuel column. The question of plenum spring relaxation and axial gap formation was presented to B&W (Ref. 6). B&W responded (Refs. 4 and 5) that a total of 40 fuel rods of the current Mark B 15x15 design have been examined by gamma scanning with no axial gaps larger than 0.1 inches. Of these 40 rods, eight rods were examined from ANO-1 after one cycle of irradiation with no measurable gaps. A group of twenty fuel rods from five assemblies in Oconee-1 were examined at the end of four cycles of irradiation (average burnup about 38 mwd /kgM) with no gaps larger than 0.1 inches observed. Four of these twenty rods were examined at the end of five cycles of irradiation along with twelve other rods from the same assembly not examined previously.

The average burnup of these rods was about 50 mwd /kgM with no observable gaps (the detectable limit was 0.05 increi). All of the four rods examined at the end of five cycles of irradiation had displayed small axial gaps at the end of four cycles, indicating that the gaps had closed ep during the fifth cycle. A likely mechanism for the closure of the axial gaps is fuel swelling. This confirms that axial gaps, due to either bottom i

plenum spring relaxation or fuel densification in the current Mark B 15x15 design, are not of sufficient size, i.e., greater than 0.5 inches, to allow cladding collapse.

1 As a check on the bottom spring relaxation at extended burnups, B&W l has also examined the bottom position of the fuel stack from the gamma scans of the thirty-two fuel rods irradiated for four and five cycles. The rod average burnups of these thirty-two rods varied between 36 and 51 mwd /kgM. B&W has stated (Ref. 5) that from their measurement of the bottom position of the fuel stack, there is no evidence that the bottom plenum spring has relaxed. In addition, B&W has indicated that the diameter of the bottom plenum spring of their t

l

\

current design was increased to prevent any spring relaxation due to irradiation.

A statistical probability of finding an axial gap of a particular size l

i in the current Mark B 15x15 design can be calculated from the gamma scan data if some conservative assumptions are made on the mean gap size and standard deviation of this data. The actual gap sizes ob-served were not recorded for the forty rods examined, but it was noted that they were less than or equal to the maximum gap of 0.1 inches.

Therefore, the actual gaps observed in these rods were between the minimum detectable limit of 0.05 inches and the maximum gap size observed of 0.1 inches. If one conservatively assumes that the mean

gap size, x, observed was 0.1 inches with a standard deviation, o, of 0.1 inches, the axial gap size is calculated from a one-sided student t statistic, at a 0.05% level, and a sample size of 40 to be x + t(o) = 0.1 + 3.55(0.1) = 0.455 inches.

Consequently, there is a 99.95% probability that a gap of 0.455 inches i or smaller will be observed in the current Mark B 15x15 design.

If only those gamma scan data from the thirty-two rods at burnups greater than 32 mwd /kgM are utilized, where axial gaps due to spring relaxation and cladding collapse is of concern, the sample size is i

reduced to 32. If one once again conservatively assumes the mean and standard deviation of the gaps in these rods are 0.1 inches, the gap size at the 0.05% level becomes 0.1 + 3.64(0.1) = 0.464 inches.

This result is not significantly different than that obtained for the 40 rods.

An additional concern specific to the B&W approach is whether the proposed B&W envelope for preventing creep collapse contains and bounds the important parameters for creep collapse. There is one i

important parameter to creep collapse that is not included in the B&W operating envelope, and this parameter is cladding temperature. A question as to why cladding temperature was not included in their operating envelope was presented to B&W (Ref. 6). B&W responded (Ref. 4) that the current approach for creep collapse analysis utilizes- l J

w-C-we,=- i w-vvve-'w -w e- mer kewe +r -wm +- n t -v -v9 **7-WWe-

a conservative cladding temperature (Ref. 7) calculated from the maximum linear heat generation rate (LHGR) allowed by the technical specifications; and that this bounds all possible cladding temperatures for current B&W designs. The staff agrees with this conclusion, but if the maximum calculated cladding temperatures should increase utilizing the calculational methods in Reference 7, the issue of creep collapse would need additional review. The staff concludes that the remaining parameters and bounding conditions for the B&W envelope are acceptable because they are based on bounding conditions of their et* rent manufacturing tolerances and operating data of the current Mark B 15x15 design.

3.0 CONCLUSION

S B&W has presented gamma scan data from current Mark B 15x15 fuel rods irradiated up to rod average burnups of 51 mwd /kgM that show no axial i

gaps greater than 0.1 inches have formed. A conservative statistical l analys'is o,f this data has shown that there is a very small probability l

that a gap of sufficient size to allow cladding collapse will form due l

l to e'tner fuel densification or bottom plenum spring relaxation in the l

! current Mark B 15x15 design. The new B&W manufacturing envelope for evaluating creep collapse is based on the lower two-sigma fabrication l tolerances of the current design from which the most recent axial gap i data were obtained.

l

. In summary, the staff finds that the above B&W methodology for cladding collapse is acceptable with the following condition: that the calculated cladding temperatures do not exceed the upper limit for cladding temperatures as defined in BAW-10084P-A, Rev. 2 (Ref. 7).

4.0 REFERENCES

1) " Technical Report on Densification of Light Water Reactor Fuels,"

WASH-1236, Regulatory Staff USAEC, November 14, 1972.

2) Letter, J. H. Taylor (B&W) to C. O. Thomas (NRC),

Subject:

Creep Collapse Analysis for B&W Fuel, JHT/86-011A, dated January 31, 1986.

3) W. M. Adams, et al., "CEPAN Method of Analyzing Creep Collapse of Oval J

Cladding, Vol. 5: Evaluation of Interpellet Gap Formation and Clad Collapse in Modern PWR Fuel Rods," EPRI NP-3699-CCM Volume 5, Project 2061-6, April 1985.

t

4) Letter, J. H. Taylor (B&W) to D. M. Crutchfield (NRC),

Subject:

Creep 1

Collapse Analysis for B&W Fuel (JHT/86-011A), JHT/86-180, dated l

August 4, 1986.

l l 5) Letter, J. H. Taylor (B&W) to D. M. Crutchfield (NRC),

Subject:

Creep 1 Collapse Analysis for B&W Fuel (JHT/86-011A), JHT/86-230, dated l September 17, 1986.

i

_c. . .. ___ - __ .

1, .

, 3

.

• g e

% t A '

~

6) Letter, D. M. Crutchfield (H;RC) to J.\ H. Taylor (B&W),'

Subject:

Cre'ep g 5 Collapse Analyses for B&W Fuel'(JHT/86-011A), dated June 26, 1986.

?

i

7) A. F. J. Eckert, et al., " Program to Determine In-Reactor Performance of B&W Fuels - Cladding Creep Collapse," BAW 10084P-A, Rev 2, October 1978.

6 l .

f l ,

-. . . . _ _ _ ___ _. .