ML042300607

Blue Ridge Environmental Defense Leagues Proposed Findings of Fact and Conclusions of Law Regarding Bredl Contention I

ML042300607
Person / Time
Site:	Catawba
Issue date:	08/06/2004
From:	Curran D Blue Ridge Environmental Defense League, Harmon, Curran, Harmon, Curran, Spielberg & Eisenberg, LLP
To:	Atomic Safety and Licensing Board Panel
Byrdsong A T
References
50-413-OLA, 50-414-OLA, ASLBP 03-815-03-OLA, RAS 8338
Download: ML042300607 (27)

Text

VR5-8338 August 6, 2004 UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION USNRC BEFORE THE ATOMIC SAFETY AND LICENSING BOARD August 12, 2004 (3:47PM)

OFFICE OF SECRETARY In the Matter of RULEMAKINGS AND Docket No's. 50-413-OLA, ADJUDICATIONS STAFF DUKE ENERGY CORPORATION 50-414-OLA (Catawba Nuclear Station, Units I and 2)

BLUE RIDGE ENVIRONMENTAL DEFENSE LEAGUE'S PROPOSED FINDINGS OF FACT AND CONCLUSIONS OF LAW REGARDING BREDL CONTENTION I Blue Ridge Environmental Defense League ("BREDL") hereby submits its proposed findings of fact and conclusions of law regarding BREDL Contention I.

Introduction

1. BREDL Contention I asserts that Duke Energy Corporation's ("Duke's") license amendment request (CLAR") to test plutonium mixed oxide ("MOX") fuel at the Catawba nuclear power plant is inadequate because Duke has failed to account for the differences between MOX and low enriched uranium ("LEU") fuel behavior; nor has Duke accounted for the impact of these differences on Duke's analysis of loss of coolant accidents ("LOCAs"). In particular, BREDL raises the concern that Duke has failed to consider the effect of MOX fuel relocation on Duke's ability to demonstrate that the LAR is in compliance with the emergency core cooling system (ECCS) acceptance criteria in 10 C.F.R. § 50.46.

2. We admitted, without objection, the Prefiled Written Testimony of Dr. Edwin S.

Lyman Regarding BREDL Contention I (July 1, 2004) (hereinafter "Lyman Testimony); and

-femp e =SCcy-6 CC 0L OZa

Rebuttal Testimony of Dr. Edwin S. Lyman (July 8, 2004) hereinafter "Lyman Rebuttal Testimony").

3. Dr. Lyman, a Senior Scientist with the Global Security Program at the Union of Concerned Scientists in Washington, D.C., is a qualified expert on nuclear safety and safeguards issues, including the issues raised in Contention I. He holds a Ph.D., a master of science degree and a bachelor's degree in physics. Lyman Testimony at 1. For over eleven years, he has conducted research on security and environmental issues associated with the management of nuclear materials and the operation of nuclear power plants. Id.

4. Dr. Lyman's research has included the safety and security implications of using mixed oxide fuel as a substitute for uranium fuel in nuclear power plants. He has also published articles on this topic. Lyman Testimony at 1. Dr. Lyman was co-author of one of the first papers to explore the possibility of using MOX fuel to dispose of separated plutonium from dismantled nuclear warheads. Tr. at 2515. Since then, he has conducted his own analyses of various safety issues associated with the U.S. program to use MOX fuel, including an analysis of the consequences of severe accidents. Id. He has also attended numerous meetings over the last 12 years regarding safety issues associated with MOX fuel. Tr. at 2516. We find that Dr. Lyman is a qualified expert on the issues raised by Contention I, whose testimony should be given full weight.

5. The ASLB also admitted the Testimony of Steven P. Nesbit, Robert C. Harvey, Bert M. Dunn, and J. Kevin McCoy on Behalf of Duke Energy Corporation on Contention (July 1, 2004) (hereinafter "Duke Testimony"); and the Rebuttal Testimony of Steven P. Nesbit, Robert C. Harvey, Bert M. Dunn, and J. Kevin McCoy on Behalf of Duke Energy Corporation on Contention (July 8, 2004) (hereinafter "Duke Rebuttal Testimony"). In addition, the ASLB 2

admitted the NRC Staff Testimony of Undine Shoop, Dr. Ralph Landry and Dr. Ralph 0. Meyer Concerning BREDL Contention I (July 1, 2004) (hereinafter "NRC Staff Testimony"); and the NRC Staff Rebuttal Testimony of Dr. Ralph Landry and Dr. Ralph 0. Meyer Concerning BREDL Contention I (July 8, 2004) (hereinafter "NRC Staff Rebuttal Testimony").'

Regulatory Requirements 10 C.F.R. § 50.46 Acceptance Criteria

6. NRC regulations at 10 C.F.R. § 50.46 establish acceptance criteria for emergency core cooling systems for light-water nuclear reactors. Essentially, the regulation sets limits on the extent of fuel damage that can occur during a design basis LOCA. Three of the acceptance criteria warrant special consideration for MOX fuel: peak cladding temperature ("PCT"), maximum cladding oxidation, and coolable geometry. Section 50.46(b)(1) requires that that PCT "shall not exceed 22000 F." Section 50.46(b)(2) provides that the "calculated total oxidation of the cladding shall nowhere exceed 0.17 times the total cladding thickness before oxidation." Section 50.46(b)(4) also requires that "[clalculated changes in core geometry shall be such that the core remains amenable to cooling."

Appendix K Evaluation Models

7. Appendix K to Part 50, whose requirements are referenced in 10 C.F.R. § 50.46(a)(1), sets forth required and acceptable features of ECCS "evaluation models" that are to be used in determining compliance with the criteria in 10 C.F.R. § 50.46. Appendix K does not include consideration of fuel relocation. The NRC did contemplate requiring fuel relocation to be included in Appendix K models, but decided such action was not necessary in the course of its resolution of Generic Issue 92. Exhibit 26, Memorandum from Ralph Meyer, NRC Office of l The Staff submitted a corrected version of its testimony at the commencement of the hearing 3

Nuclear Regulatory Research, to John Flack, NRC Regulatory Effectiveness and Human Factors Branch, re: Update on Generic Issue 92, Fuel Crumbling During LOCA (February 8, 2001). See also Exhibit 27, Memorandum from Ashok C. Thadani, Office of Nuclear Regulatory Research, to Samuel J. Collins, Office of Nuclear Reactor Regulation, re: Information Letter 0202, Revision of 10 CFR 50.46 and Appendix K (June 20, 2002).

8. Nevertheless, more recently, the NRC has acknowledged that omission of fuel relocation effects is a non-conservatism in Appendix K with a very large potential impact on PCT, and that an early "resolution" of this issue (i.e., Generic Issue 92) may have been in error or is no longer applicable because of new information. See Exhibit 26 and Exhibit 27, Attachment 4 at 4-5. See also tr. at 2520, 2532-33 and Exhibit 53, e-mail message from Ralph Meyer to various PIRT participants (October 13, 2000).

9. Moreover, fuel relocation is considered in the Westinghouse LOCA analysis that Duke relies on to establish compliance of its co-resident LEU fuel with 10 C.F.R. § 50.46. Harvey, tr.

at 2373-74. The model for this analysis is known as "WCOBRA/TRAC." Tr. at 2374.

Applicability of 10 C.F.R. § 50.46 and Appendix K

10. 10 C.F.R. § 50.46 and Appendix K apply only to uranium-based fuel, but Duke has requested an exemption from this limitation so that these requirements will apply to MOX fuel.

Contention I presents the legal and factual question of whether and how 10 C.F.R. § 50.46 and Appendix K should be applied to evaluate the safety of using MOX fuel at the Catawba reactor.

As discussed below, for a number of reasons, we agree with the testimony of Dr. Lyman that it is generally appropriate to apply the requirements of 10 C.F.R. § 50.46 to MOX fuel, as long as on July 14, 2004.

4

Appendix K is not strictly applied to exclude consideration of relocation of the fuel during LOCAs. See Lyman Testimony at 3.

Findings of Fact Summary of BREDL's Testimony

11. Dr. Lyman testified that Duke's design-basis loss of coolant ("DB-LOCA") analysis is inadequate because it does not address the uncertainties associated with relocation effects that M5-clad MOX fuel may experience under LOCA conditions. Lyman Testimony at 2. These uncertainties relate to Duke's assertion that the action proposed in the MOX LTA LAR will not result in a violation of the emergency core cooling system (ECCS) acceptance criteria in 10 C.F.R. § 50.46: peak cladding temperature ("PCr), maximum cladding oxidation, and the preservation of a coolable core geometry.

12. Dr. Lyman noted that the phenomenon of fuel relocation has been observed in experiments with irradiated LEU fuel under LOCA conditions. Lyman Testimony at 2. While no similar experiments have been done on MOX fuel (see tr. at 2411), there are technical reasons to believe that the impact of fuel relocation effects during a LOCA may be more severe for MOX fuel rods than for LEU fuel rods of the same burnup, due to differences in characteristics such as fuel fragment sizes and fuel-clad interactions. Moreover, calculations in Duke's LAR indicate that MOX fuel is generally more limiting than LEU fuel with respect to DB-LOCAs. Therefore, the consequences of fuel relocation, and the non-conservatism associated with neglecting them, may be of greater concern for MOX fuel rods than for LEU fuel rods with respect to compliance with LOCA regulatory criteria. Lyman Testimony at 2.

13. In his testimony, Dr. Lyman asserted that Duke has failed to address these uncertainties in MOX fuel behavior, and therefore its LTA application is unacceptable to satisfy 5

the requirements of 10 C.F.R. § 50.46 with respect to PCT, maximum cladding oxidation, and coolable geometry of fuel. Lyman Testimony at 2. In addition, by failing to address the uncertainties in MOX fuel behavior, Duke has not demonstrated compliance with the general reasonable assurance standard in 10 C.F.R. § 50.40(a). Ids

14. As Dr. Lyman testified, experts have concluded that MOX fuel may experience more severe relocation effects than U02 fuel at the same burnup. Lyman Testimony at 5. The Staff's testimony also confirmed the possibility that the amount of fuel that is relocated in MOX rods may be greater than the amount in LEU rods. NRC Staff Testimony at 14. In addition, the Staff confirmed that fuel relocation could cause the cladding temperature in the balloon to increase by several hundreds of degrees F. Staff Testimony at 17.

15. A 2001 study by the Institut de Protection et de Siretd Nuclkaire (IPSN, now IRSN) did not explicitly consider MOX fuel, but stated that "it must be pointed out that that results of corresponding calculations with... high burnup MOX fuels would have been more severe with regard to acceptance limits." Lyman Testimony at 5, citing Exhibit 29, Grandjean, Hache and Rongier, "High Burnup U02 Fuel LOCA Calculations to Evaluate the Possible Impact of Fuel Relocation After Burst," OECD/NEA Proceedings of the Topical Meeting on LOCA Fuel Safety Criteria, Aix-en-Provence at 7 (March 22-23, 2001). As Dr. Lyman noted, IRSN, the successor to IPSN, has reiterated these concerns, stating in a recent presentation that for MOX fuel, a "higher initial energy" and an "enhance [sic] of fuel relocation impact" results in greater increases in PCT and ECR associated with relocation. Lyman Testimony at 5, citing Exhibit 30, V. Guillard, C. Grandjean, S. Bourdon and P. Chatelard, "Use of CATHARE2 Reactor Calculations to Anticipate Research Needs," SEGFSM Topical Meeting on LOCA Issues, Argonne National Laboratory, slides at 8-9 (May 25-26, 2004). In the abstract for this presentation, the authors 6

state that "a lack of knowledge on theses [sic] parameters [important for relocation] for irradiated U02 and particularly MOXfiel [emphasis added] may lead to reduce [sic] safety margins."

Effect of Fuel Relocation on Peak Clad Temperature During a LOCA.

16. Fuel relocation refers to the movement of fuel pellet fragments into regions of the fuel rod where the cladding has ballooned during a LOCA transient. Exhibit 27, Attachment 4 at 4-5.

As Dr. Lyman testified, fuel relocation increases the local linear heat generation rate within the ballooned area. Lyman Testimony at 4. Thus it could increase the severity of a LOCA by resulting in a greater fuel rod peak cladding temperature (PCT) than in a situation in which fuel relocation did not occur. Id. This increase could be as much as several hundred degrees. Tr. at 2522-23.

17. Because transient oxidation during a LOCA increases with an increase in PCT, fuel relocation could also result in a greater maximum cladding oxidation. Lyman Testimony at 4.

Finally, the greater local linear heat generation rate requires a greater coolant flow around the ballooned area to ensure long-term core coolability. Id; see also Exhibit 28, slides presented by A. Mailliat and J.C. M6lis, IRSN, at "PHEBUS STLOC Meeting" with NRC Staff (October 23, 2003).

18. Robert C. Harvey and Bert M. Dunn, witnesses for Duke, testified that the impact of fuel relocation on the increase of the heat source in the ballooned region is mitigated by the lower density of the relocated fuel compared to the original pellet density, so that only very high filling ratios are a cause for concern. Duke Testimony at 46-47. Their assertion is undermined by testimony by the NRC Staff that the diameter increase in the balloon can be as great as 100%, so the cross-sectional area can increase by a factor of 4. NRC Staff Testimony at 12. Therefore, if the entire area fills with fuel, the linear heat source would double for a filling ratio of only 50%.

For the FR2 observed filling ratio of 0.615, the linear heat source would increase by a factor of 7

2.5. The significant impact of relocation on PCT and maximum clad oxidation for these filling ratios is demonstrated in Exhibit 29.

19. Dr. Ralph 0. Meyer, a witness for the NRC Staff, testified that "if fuel relocation has any effect, it would increase the temperature only in the ballooned region of the fuel rod... the ballooned region is seldom the location of the calculated peak cladding temperature when relocation is ignored." NRC Staff Testimony at 15. However, Dr. Lyman pointed out that when relocation does occur, FR2 Test E4 shows that the peak temperature of the ballooned region is within about 20 degrees F of the PCT at the unruptured node. Tr. at 2493.

20. This conclusion is also consistent with Exhibit 27, in which NRC states that "the fuel relocation effect on PCT may be significantly larger than that assumed in GS-92 [+460F]" that fuel relocation may have a +3130F impact on PCT. Id., Attachment 5 at 4.

21. Duke witness Mr. Burt Dunn asserted that if "reasonable" fuel relocation effects were to be incorporated into its deterministic evaluation model, the peak temperature at the ruptured location would still be very close to the peak cladding temperature at the non-ruptured location, within a reasonable bound of uncertainty, because the enhanced cooling at the rupture node would largely offset the greater linear heat generation due to relocation. Tr. at 2400. However, Mr. Dunn would not concede that the non-ruptured location would in all certainty remain the limiting case, but merely pointed out an example of one test in which the non-ruptured location did remain the limiting case, albeit by a small margin. Tr. at 2401.

22. In fact, Mr. Dunn's statement leaves open the possibility that if relocation is accounted for, it is possible that the ruptured node could in fact be limiting with respect to PCT.

Dr. Lyman presented evidence that such an outcome was possible in certain large-break LOCA transients. Tr. at 2502-2504. See also Exhibit 53 at 1. In such situations, relocation will have a 8

direct impact on PCT and could reduce the margin to the regulatory PCT limit. Duke has not attempted to estimate the magnitude of this reduction in margin.

23. In the 2001 IPSN study referred to in paragraph 15 above, the authors used the CATHARE2 computer code to calculate the impact of fuel relocation on the large-break LOCA PCT for a high-burnup U02 fuel rod as a function of the '"illing ratio," or the ratio of the volume of the relocated fuel material to the volume of the ballooned region.2 Lyman Testimony at 4. In that study, for the scenario evaluated, the authors found that the PCT in the absence of relocation effects was 9701C. For a filling ratio of 70%, the maximum considered, the PCT was 11441C. For a filling ratio of 40%, the PCT was about 20'C greater than for the no-relocation case. Thus the maximum impact on PCT of relocation in this study was a APCT of+ 174°C (3131F) for high-burnup UO2 fuel. It is not clear from the study whether higher filling ratios, and hence larger impacts on PCT, are possible. Lyman Testimony at 4.

24. While the phenomena involved in these processes are too complicated for simple back-of-the-envelope assessments, some observations can be made. According to the MOX LTA LAR at 3-43, the peak temperature at the hot pin rupture location is 1841'F. If the 313'F increase in clad temperature associated with fuel relocation with a filling ratio of 0.7 is added to this value, the resulting clad temperature at the rupture location is 21540F. From Figure 11 in Dukee's testimony, the PCT in a rod where relocation occurs appears to be about 20'F greater than the maximum temperature at the rupture location. Therefore, the peak clad temperature associated with an LEU rod with 0.7 filling ratio due to relocation could be as high as 21740F ---

a value with substantially less margin to the 10 CFR § 50.46 limit. Consideration of additional 2 These calculations are reported in Exhibit 29, C. Grandjean, G. Hache and C. Rongier, "High Burnup U0 2 Fuel LOCA Calculations to Evaluate the Possible Impact of Fuel Relocation After 9

MOX effects, such as a greater filling ratio, could shrink this margin even further. Lyman Rebuttal Testimony at 2.

25. NRC Staff witness Dr. Ralph Meyer testified that this approach was not "a bad way of getting some estimate on the outside effect" of fuel relocation on rupture node temperature during a LOCA. Tr. at 2669. But Dr. Meyer also claimed that the value of 313'F that the NRC Staff had previously cited as an upper limit value for the effect of relocation on clad temperature was no longer valid, and that more recent calculations by IRSN indicated that the upper limit value was 2700F. Tr. at 2634-2635. However, during questioning, Dr. Meyer admitted that the source of this new information (i.e., Exhibit 30) apparently contained inconsistencies, in that the numerical value of 2700F written on the Powerpoint slide was smaller by at least 90'F than the value obtained from the accompanying data plot. Tr. at 2660-2662. Dr. Meyer also admitted that he could not figure out the difference between the curves and the numbers. Tr. at 2662.

Therefore, it is unclear from the information in Exhibit 30 whether the revised IRSN predictions for the maximum impact of relocation on clad temperature has decreased or increased relative to the values used in Dr. Lyman's testimony.

26. Bert M. Dunn and Steven P. Nesbit, witnesses for Duke, testified that the effect on Duke's LOCA analysis of using the LEU decay heat curve instead of the MOX decay heat curve is a conservatism of "up to 750F on PCT." Duke Testimony at 25. Dr. Lyman agreed that this is a conservatism, but he pointed out that the effect of using the LEU decay heat curve on PCT is considerably smaller than the effect of relocation on PCT, which could be on the order of several hundred Fahrenheit degrees. Lyman Rebuttal Testimony at 4. Both effects should be considered in any LOCA analysis involving MOX fuel, to ensure that any interactions between the two Burst," OECD/NEA Proceedings of the Topical Meeting on LOCA Fuel Safety Criteria, Aix-en-10

effects are properly accounted for. See also tr. at 2510-11 and Exhibit 36, Grandjean and Hache, "LOCA Issues Related to Ballooning, Fuel Relocation, Flow Blockage and Coolability,"

SEGFSM Topical Meeting on LOCA Issues, Argonne National Laboratory (May 25-27, 2004).

27. Duke witness Bert Dunn testified that although the 1841 IF peak rupture node temperature provided in the MOX LTA LAR (Exhibit 1) was the correct result of a valid LOCA analysis, the Catawba plant will not allow the operation of a MOX fuel pin at the power level used in that analysis. Tr. at 2347. Mr. Dunn also testified that he believes that the maximum peak clad temperature for the rupture location is never going to exceed 17501F for any design basis LOCA.

Tr. at 2353. However, Mr. Dunn also testified that in his personal opinion, differences in calculated LOCA peak clad temperatures of less than 100 degrees Fahrenheit are not insignificant, and "within plus or minus 100 degrees it's probably the same answer." Tr. at 2392.

Therefore, in Mr. Dunn's opinion, given the uncertainties associated with Appendix K analyses, a result for peak rupture node temperature of 1841°F is "probably the same answer" as a result of 17501F. Dr. Lyman's estimate of the peak rupture node temperature resulting from fuel relocation effects based on the higher temperature is appropriately conservative, given this uncertainty.

Effects of Fuel Relocation on Maximum Clad Oxidation

28. In addition to limits on PCT, the maximum clad oxidation must also be limited to less than 17%. Since the maximum clad oxidation typically occurs at the ruptured location, as a result of double-sided oxidation, a significant increase in clad temperature at the ruptured location due to relocation would result in a significant increase in maximum clad oxidation, with the potential to exceed the 10 CFR § 50.46 limit. The oxidation rate for M5 is substantially Provence (March 22-23, 2001) (hereinafter "Grandjean, Hache, and Rongiere).

11

greater at 21740F than at 18410F. Lyman Rebuttal Testimony at 2. See also tr. At 2548-2549; Exhibits 56, Yan, et al, Post Quench Ductility Results for Zry-4 and M5 Oxidized at 10000 C, and 11000 C (January 15, 2004); and Exhibit 57, Yan, et al, Post Quench Ductility Results for Zry-4 and M5 Oxidized at 12000 C, Slow Cooled to 8000 C and Quenched (March 23, 2004).

29. The IPSN study (See Exhibit 29) evaluated the impact on the maximum cladding oxidation for the ruptured region (two-sided oxidation). The equivalent cladding reacted (ECR) calculated by the Cathcart-Pawel rate law (a surrogate for "maximum cladding oxidation") was 12.6% for the no-relocation case, and 19.7% for the 70% filling ratio case. Thus the maximum impact on ECR resulting from relocation was calculated as AECR = 7.1%.

30. Duke witness Bert Dunn testified that if relocation is taken into account for the MOX LTAs, the peak clad oxidation could be estimated by adding the above value for AECR to the value originally calculated by Duke. Duke Testimony at 69. However, Dr. Lyman pointed out that in contrast to the estimate for PCT, it is not appropriate to estimate the effect of relocation on ECR merely by adding the AECR from the IPSN study to the maximum clad oxidation that Duke has calculated for the MOX LTAs using the Framatome Appendix K model, because ECR is a nonlinear function of PCT. Tr. at 2516-2517. In addition, the IPSN calculation was terminated at less than 200 seconds, whereas the Catawba LOCA analysis was run for 400 seconds. Tr. at 2664.

MOX Fuel Characteristics Which Exacerbate Relocation Effects

31. Dr. Lyman concluded that MOX fuel may experience more severe relocation effects than UO2 fuel at the same burnup because several characteristics that are important for relocation may be less favorable for MOX fuel. These include pellet fragment size and fuel-clad interaction. Lyman Testimony at 5.

12

MOX Fuel Pellet Fragment Size

32. Dr. Lyman testified that the IPSN calculations cited in Exhibit 29 demonstrate the high sensitivity of fuel relocation-induced increases in PCT and ECR to the filling ratio. The filling ratio, in turn, is a function of the average particle size of the relocated fuel fragments, in that smaller particles will in general result in greater packing of the relocated area and hence higher filling ratios.

Lyman Testimony at 5.

33. The vulnerability to fuel relocation is associated with the development of the high-burnup "rim" region known to emerge in LEU fuel for burnups exceeding about 40-45 GWD/t.

IPSN states that "fuel fragmentation is clearly associated to [sic] burnup, with finer fragments at higher BU." See Lyman Testimony at 6, Exhibit 29. In addition, Dr. Lyman cited CABRI test data acquired by IPSN which demonstrate a relationship between fragmentation and high burn-up rim structure. Tr. at 2467-72. See also Exhibit 51, Papin, et al, Synthesis of CABRI-RIA Tests Interpretation (undated).

34. During manufacture of MOX fuel using the MIMAS process (which will be used for the Duke LTAs), plutonium agglomerates --- macroscopic clumps of plutonium-rich particles --

occur in the fuel. Lyman Testimony at 6. Because the fissile material is concentrated in these clumps, very high local burnups result, due to the fact that the fission is occurring in a heterogeneous fashion. The ratio of local burnup within the agglomerates is on the order of 4-6 times the rod-average burnup, depending on the irradiation time. For instance, the agglomerate bum-up reaches about 60 GWD/t when the rod average is only around 18 GWD/t, and reaches 100 GWD/t when the rod average is only 28.4 GWD/t. As a result, high-burnup rim-like regions emerge in the outer layers of the plutonium agglomerates for much lower rod-average burnups 13

than 40-45 GWD/t, because the local burnups within the plutonium agglomerates increase much more rapidly than the rod-average burnups. Id.

35. In Dr. Lyman's opinion, the CABRI test results described in Exhibit 51 demonstrate that the high bum-up rim region in LEU fuel and the high bum-up structure in the plutonium agglomerates in the MOX fuel undergo micro-cracking as a result of the fission gas accumulation. Tr. at 2472. In the event of a rapid gas expansion, this micro-cracking can lead to fragmentation. Id. at 2473.

36. Thus, as Dr. Lyman testified, it is reasonable to expect that the onset of fuel relocation in MOX fuel may occur at lower rod-average burnups than in LEU fuel. Lyman Testimony at 6. This would imply that MOX fuel will be vulnerable earlier in its irradiation history (and consequently for a longer time) than LEU fuel. Also, the particle size distribution in MOX fuel will be smaller than in LEU fuel at the same rod-average burnup, to the extent that fine fragments are generated in the ultra-high burnup plutonium agglomerate regions. Id.

37. Fuel fragmentation can also be caused by the stress induced by the stored-energy redistribution during the blowdown phase of a LOCA. Lyman Testimony at 6. See also Exhibit 31, A. Mailliat and M. Schwarz, "Need for Experimental Programmes on LOCA Issues Using High Burn-Up and MOX Fuels," NUREG/CP-0176, Proceedings of the Nuclear Safety Research Conference at 436 (May 2002) (hereinafter "Mailliat and Schwarz"). Because MOX fuel has a lower thermal conductivity and a higher radial temperature gradient than LEU fuel, it could experience greater fuel fragmentation during the blowdown and more severe relocation effects as a result.

38. As Dr. Lyman pointed, out, according to two out of four NRC experts who participated in the 2001 PIRT panel on LOCAs and high-burnup fuel, the composition of fuel 14

(i.e. a specified MOX composition) is of "high importance" for consideration of fuel relocation effects because it "may affect the amount of fine grain material after relocation. Fuel structure and mechanical properties are influenced by fuel type." Lyman Testimony at 6, citing Exhibit 32, NUREG/CR-6744, "Phenomenon Identification and Ranking Tables for Loss-of-Coolant Accidents in Pressurized and Boiling Water Reactors Containing High-Burnup Fuel," Appendix D, Table D-1 at D-67 (December 2001). One expert concluded that fuel composition was of moderate importance to relocation, stating that "the consequence of fuel fragments relocation (higher local decay heat and higher cladding temperature) could be more effective with MOX fuel than with U02 fuel" but that "the viscoelastic properties of the MOX should impair the fuel fragments relocation at high burnup." Id. at D-67. A fourth expert concluded that fuel composition would be of only low importance to relocation. Id. at D-67. As Dr. Lyman testified, this difference of expert opinion highlights the inadequacies of the experimental database with regard to integral tests of MOX fuel under design-basis LOCA conditions, and underscores the significant uncertainties in Duke's design-basis LOCA analysis. Lyman Testimony at 6.

39. Dr. Meyer testified that, according to the results of recent high-burnup integral tests at Argonne National Laboratory ("ANL") "it appears that the small particles or fines are blown out of the burst opening when the rod depressurizes," implying that "there would be few or no small particles in the ballooned region" of the type that could make a difference in the relocated fuel mass in MOX fuel and LEU fuel. NRC Staff Testimony at 14-15. As Dr. Lyman pointed out, however, Dr. Meyer fails to take into account the fact that the tests in question were performed on BWR fuel rods and not PWR fuel rods. Lyman Rebuttal Testimony at 3. The two types of fuel are sufficiently different that it is difficult to come to any conclusions about the behavior of PWR fuel rods during a LOCA from experiments on BWR rods. This is why ANL 15

plans to repeat these tests using rods from the H.B. Robinson PWR. See Exhibit 40, excerpt from Yan et al. slide presentation at 32 (undated).

40. Moreover, as Dr. Lyman also observed, the excess fine particles generated in MOX fuel originates in the rim regions surrounding the plutonium agglomerates, which are distributed throughout the fuel pellet. Dr. Lyman's observation was confirmed by the Staff's response to Question 39. See Staff Testimony at 14.

41. As demonstrated in the Staffs testimony at 14, the rim material in LEU fuel is generated only around the circumference of the fuel pellets, and thus is adjacent to the clad and the burst opening. The apparent absence of fine particles in the vicinity of the burst opening does not provide evidence that fine particles escape from the fuel rod at circumferential locations away from the burst opening (in the case of LEU fuel) or from locations throughout the entire fuel rod cross-section (in the case of MOX fuel).

42. In fact, as Dr. Lyman testified, to the extent that the phenomenon observed in this test indicates that fine-grain rim material near the burst opening is blown out of the rod during depressurization of LEU fuel rods, the result only strengthens BREDL's contention that MOX fuel rods contain greater quantities of fine-grain material than LEU fuel and hence may be subject to more severe relocation effects.

43. Dr. Lyman also testified that in at least one of the tests in the the same Argonne LOCA series that was referred to by Dr. Meyer, there was evidence that fine fuel particles below 0.3 millimeters in size remained within the fuel rod during the LOCA transient, and fell out only when the rod was handled after the test was over. See tr. at 2529-2531. See also Exhibit 52, Yan, et al, LOCA results for advanced-Alloy and High-Burnup Zircaloy Cladding (undated).

16

This contradicts Dr. Meyer's assertion that all small particles or fines would be blown out of the burst opening when a fuel rod depressurizes during a LOCA.

Fuel-Clad Interaction

44. Dr. Lyman expressed a concern about differences between MOX and LEU fuel with respect to fuel-clad bonding and the impact of such differences on fuel relocation behavior during a design-basis LOCA. Lyman Testimony at 7. According to IPSN (now IRSN), tight fuel-clad bonding may delay the onset of fuel relocation. Exhibit 31 at 433. During NRC's recent expert elicitation (PIRT) process on LOCA issues for high-burnup fuel, all four participating experts agreed that "chemical and mechanical bonding between the fuel pellet and the cladding..." was of high importance to the fuel relocation phenomenon, because "bonding could significantly affect the relocation characteristics by impeding pellet fragment movement." Lyman Testimony at 7; Exhibit 32, Table D-1 at D-69. It has been confirmed that MOX fuel is more resistant to clad failures due to pellet-clad mechanical interaction (PCMI) than LEU fuel, even at high burnups. Lyman Testimony at 7; Exhibit 34. Nuclear Energy Agency, NEA/NSC/DOC(2004)8, International Seminar on Pellet-Clad Interactions with Water Reactor Fuels, at 20 (May 6, 2004). This phenomenon is not well-understood but may imply that the pellet-clad bond is weaker for MOX fuel, in which case MOX fuel may have a greater propensity to earlier and more extensive fuel relocation than LEU. Lyman Testimony at 7.

45. In Duke's April 14, 2004, Response to BREDL's first set of discovery requests, Duke stated that the Framatome design-basis LOCA analysis for the MOX LTAs did not assume any fuel-clad bonding and was therefore "conservative" with 17

respect to the requirement that the degree of cladding swelling not be underestimated. Id at 14. However, in the absence of an assessment of whether and to what extent the pellet-clad interaction is weaker in MOX fuel than in LEU fuel, there is no way of knowing the degree to which this assumption is conservative for MOX fuel. Therefore, Duke's failure to properly account for this phenomenon contributes another uncertainty to the safety margin associated with Duke's design basis LOCA calculation. Lyman Testimony at 7.

46. Moreover, as Dr. Lyman testified, there is evidence to contradict Duke's assertion that "deterministic LOCA evaluations typically based on data taken from unirradiated cladding" are conservative with respect to clad swelling. According to IPSN (now IRSN), results from the PBF-LOC experiments found that irradiated rods experienced greater clad deformation than unirradiated rods during design-basis LOCA conditions. See Exhibit 31 at 432. Both Duke and the NRC Staff have testified that this result from the PBF-LOC tests have been explained, but the explanations they provide are different. See Tr. at 2533-2534. This indicates that the underlying phenomena are still not well understood, and the possibility that the results were due to irradiation effects cannot be ruled out. There is simply no way to determine whether Duke's design-basis LOCA analysis underestimates or overestimates the degree of clad swelling (and hence the degree of fuel relocation) for MOX LTAs without additional experimental data from integral LOCA tests of high-burnup MOX fuel rods. Lyman Testimony at 7. We agree with Dr. Lyman that given the lack of data, the assertion by Duke and the NRC that "a major effect is not expected" with regard to differences in pellet-clad bonding between MOX and LEU is speculative and unpersuasive. Id. at 15.

18

Impact of Clad Balloon Size on Severity of Fuel Relocation During a LOCA

47. Duke plans to use M5 cladding for the MOX LTAs, as compared to the Zircaloy-4 or ZIRLO cladding that is extensively used in US PWRs. According to IRSN, M5 will form larger balloons than Zircaloy-4 in a design-basis LOCA because it remains more ductile during irradiation. October 2003 IRSN presentation to NRC at 24. The greater retained ductility of M5 as a function of burnup compared to Zircaloy-4 can result in a greater M5 balloon size during a design-basis LOCA for fuel rods of the same burnup. Larger balloons increase the space available for fuel fragments to fall and hence result in a greater propensity for fuel relocation during a LOCA, with an associated increase in PCT and local clad oxidation.

48. In a recent presentation at Argonne National Laboratory Recently, a group of experts from Electricitd de France (EDF), Framatome ANP and the French CEA challenged IRSN's assertion that M5 cladding would form bigger balloons during a LOCA than zircaloy-4.

We agree with Dr. Lyman that the EDF presentation does not respond adequately to the issue that IRSN has raised. Their claim is that the Edgar creep tests -- which indicated a greater ductility and a larger balloon size for M5 than for zircaloy4 -- are not the appropriate tests to actually evaluate balloon size during LOCAs. Ramp tests utilizing pre-hydrided cladding samples, which EDF asserts are more representative of LOCA conditions, indicate that the balloon size for M5 is not actually greater than for zircaloy4. Lyman Testimony at 8.

49. Obviously, a ramp test would be more similar to the conditions experienced during a LOCA than a steady-state creep test. However, neither creep tests nor ramp tests utilizing pre-hydrided but unirradiated cladding materials adequately simulate all the relevant phenomena that could affect balloon formation during a LOCA involving high-burnup fuel. For example, a well-known property of M5 cladding is that it generates a thinner oxidation layer during normal 19

irradiation as a function of burnup than zircaloy4. Zircaloy4 at high burnups tends to generate a thick oxidation layer that's prone to spalling. Spalling will cause spatial inhomogeneities in the clad temperature that negatively affect ductility, leading to earlier cladding ruptures during a LOCA and hence smaller balloon sizes. Lyman Rebuttal Testimony at 8.

50. The ramp tests described by EDF do not appear to take that effect into account.

Therefore, we don't believe that the EDF presentation fully addresses the differences that would be observed in actual irradiated fuel with regard to the ductility and the balloon size of M5 Compared to that of zircaloy-4. See Lyman Testimony at 8. As Dr. Lyman noted in his testimony, this question remains unresolved because there is an absence of experimental data on the performance of high-burnup, M5-clad fuel, under design-basis LOCA conditions. Id. at 9.

The Electric Power Research Institute (EPRI) and Areva (parent company of Framatome ANP) apparently continue to deny NRC access to samples of irradiated high-burnup M5-clad LEU fuel for testing at Argonne National Laboratory. Exhibit 35, Letter from Ashok C. Thadani, NRC, to David Modeen, EPRI (April 21, 2004). See also tr. at 2536-37 and Exhibit 55, letter from James F. Mallay to Ralph Meyer (May 5, 2003); Exhibit 54, letter from James Mallay and R. Reynolds to Ralph Meyer (December 15, 2003). This lack of cooperation can only cause further delays in the ability of NRC to obtain the experimental data it needs to confirm the safety of high-burnup M5-clad fuel (whether LEU or MOX).

51. In his testimony, Dr. Lyman underscored the admission of M. Blanpain of AREVA during the ACRS Reactor Fuels Subcommittee Meeting on April 21, 2004 that MOX fuel irradiated in France is predominantly clad in Zircaloy-4, and only "some M5 fuel rods with MOX for experimental purposes" have been used in France. See Transcript at 61-62. For some reason, France is reluctant to use M5-clad MOX fuel domestically and is primarily producing it for 20

export to Germany (and now to the United States). However, even in Germany the use of M5-clad MOX has been extremely limited. Dr. Lyman is not aware of any integral LOCA tests performed with irradiated M5-clad MOX fuel.

Impact Of Fuel Relocation on the Ability of the MOX LTA Core to Satisfy the Regulatory Requirement for Coolable Core Geometry.

52. As Dr. Lyman testified, fuel relocation increases the local linear heat generation rate in the ballooned region. Lyman Testimony at 9. The maximum flow blockage that will preserve a coolable geometry depends on the assumed heat source and the heat transfer properties of the fuel bundle. As IRSN points out, acceptable bundle blockage ratios were derived based upon arrays of unirradiated fuel rods, and did not take into account fuel relocation and its associated impacts on the redistribution of the decay heat source within the fuel rods. Exhibit 28 at 29.

IRSN restated its concern in a recent presentation:

"The impact of fuel relocation in fuel rod balloons, as was observed in all in-reactor tests with irradiated fuel, leading to an increase in local power (lineic and surfacic)..., on the coolability of the blocked region, is still fully questionable and should be addressed by specific analytical tests with a simulation of fuel relocation."

Exhibit 36 (emphasis in original).

53. We agree with Dr. Lyman that any analysis that does not take this into account is incomplete and is likely to be non-conservative. Lyman Testimony at 9. Lack of consideration of this phenomenon will be of greater concern for the MOX LTA core to the extent that the MOX LTAs have a smaller margin to regulatory limits than LEU fuel.

Smaller Safety Margins for MOX Fuel With Respect to PCI in a LOCA

54. As Duke's calculations have demonstrated, the PCT in a design-basis LOCA is higher for a MOX rod than for an LEU rod in the same position in the core. Exhibit 1, Duke 21

MOX LTA LAR at 3-43. The margin to the 10 CFR §50.46 PCT limit of 22000F is therefore smaller for a MOX rod than for an LEU rod in the same position.

55. At high bumups, the linear heat generation rate for MOX fuel is generally higher than that for LEU fuel. This, in turn, results in increased centerline temperature and stored energy, therefore reducing the margin to design-basis LOCA regulatory limits. BREDL maintains that every reduction in margin associated with MOX fuel use, coupled with the non-conservatism of ignoring fuel relocation effects, reduces confidence in Duke's design-basis LOCA analysis of the MOX LTA core.

56. Mr. Harvey testified that the Westinghouse method is a "best estimate LOCA method" which is "more state of the art" and has "more margin" to the 2,200 degree limit in 10 C.F.R. § 50.46. Tr. at 2376. He stated that the MOX method is based on Appendix K, which is "a much more conservative analysis." Id. Mr. Nesbit testified that the degree of conservatism in Appendix K is 600 degrees or more. Tr. at 2383. As Dr. Lyman testified, however, this margin of error is relevant only to "best estimate" calculations of PCT. Tr. at 2518. An applicant that elects to use an Appendix K deterministic analysis, as has Duke, may not take credit for any additional safety margin that may have resulted from utilizing a "best estimate" method. Instead, it must simply show compliance with the performance limits in 10 C.F.R. § 50.46, using an Appendix K analysis. Tr. at 2519. In this respect, Dr. Lyman and Mr. Dunn agree that to compare a best estimate and an Appendix K analysis would be to mix "apples and oranges." Tr.

at 2377-78.

57. Moreover, as Mr. Harvey previously admitted (see par. 9), the best estimate LOCA method used for the co-resident LEU fuel in the Catawba core takes fuel relocation into account.

However, the deterministic Appendix K method used for the MOX LTAs does not take 22

relocation into account. Therefore, because Duke's application does not consider relocation for both MOX LTAs and for the co-resident LEU fuel, it is impossible to determine whether the MOX LTAs will be limiting with respect to PCT if fuel relocation occurs.

58. As Dr. Lyman testified, the balance of conservatisms and non-conservatisms that applies when Appendix K models are used for LOCA analyses of uranium fuel rods is upset when Appendix K models are applied to MOX fuel analyses without taking into account additional effects, such as relocation, which may be more severe for MOX fuel than for uranium fuel. Tr. at 2521. The slight difference in the decay heat curves for MOX and LEU fuels is another effect that should be accounted for, but this difference would likely be overwhelmed by the potential effect of relocation. Tr. at 2521.

59. Because there is little or no experimental data to conclusively validate the impact of relocation on either LEU or MOX fuel, a design-basis MOX LTA LOCA analysis that takes relocation into account would be highly uncertain --- with a resulting large uncertainty in the calculation of the relocation-associated increase in PCT and ECR of a MOX LTA fuel rod compared to the relocation-associated increase in PCT and ECR of an LEU fuel rod. For instance, if the MOX filling ratio is 70% and the LEU filling ratio is only 40%, because of a greater quantity of fine fragments in the MOX fuel, the increase in rupture node temperature could be nearly three hundred degrees Fahrenheit greater for MOX than for LEU (assuming that no other MOX-related effect, such as a greater initial linear heat generation rate, results in an even more severe increase in PCT associated with relocation), with a correspondingly large increase in the rate of clad oxidation.

60. We agree with Dr. Lyman that in order to provide reasonable assurance of safe operation of the Catawba plant using MOX fuel, these significant uncertainties should be 23

reflected in Duke's analysis, and NRC approval should be contingent upon a demonstration that uncertainties of this magnitude do not undermine reasonable assurance of adequate protection of the public health and safety. Lyman Testimony at 10.

Inaequacy of Experimental Data Base

61. Dr. Lyman testified that currently, the data base is insufficient to permit a demonstration that the significant uncertainties associated with MOX fuel behavior in a LOCA do not undermine reasonable assurance of adequate protection of public health and safety. Dr. Lyman clarified that perfect certainty is not required. Tr. at 2517. Rather, experimental data is needed to reduce uncertainty to a level that a reasonable assurance of safe operation can be provided. Id It is noteworthy that while the NRC and nuclear industry have accumulated 50 years of experience with LEU fuel, there is no comparable data base for MOX fuel. Tr. at 2518.

62. We find persuasive Dr. Lyman's testimony that the only way to fully and reasonably address the uncertainties associated with the behavior of high-bumup, MS-clad MOX fuel during LOCAs is to conduct integral LOCA tests of such fuel, fabricated with the same specifications as the lead test assemblies that are under consideration here, and irradiated to a range of burnups, including the maximum of 60 GWD/t that Duke has requested in its LAR. The proposed Phebus test series would likely make a substantial contribution to reducing the level of uncertainty associated with MOX fuel behavior during LOCAs. Lyman Testimony at 11.

63. Dr. Lyman testified that these integral tests could be supplemented with separate-effects tests specifically designed to look at fuel relocation as a function of burnup for both MOX and LEU fuel, and to measure the relative susceptibility to relocation of the two types of fuels. The Halden IFA-650 test, which apparently is being designed to examine fuel relocation effects in LEU fuel, could help to resolve some of these questions. But similar tests on mixed oxide fuel will also 24

be needed. And separate effects tests cannot reproduce the complex, interrelated set of thermal-hydraulic and mechanical phenomena that would occur during a LOCA and would affect fuel relocation. Lyman Testimony at 11.

64. Dr. Lyman pointed out that neither the testimony of Duke nor the testimony of the NRC staff offers any experimental evidence to support their claims regarding the ability of the MOX LTAs to comply with 10 CFR § 50.46 criteria in the presence of relocation effects. Lyman Rebuttal Testimony at 1. Instead, Duke offered a single micrograph of an irradiated, reactor-grade MOX fuel pellet of unspecified burnup. Id. Thus the testimony of Duke and the NRC Staff tends to support, rather than refute, BREDL's contention that the LOCA analysis in Duke's LTA LAR lacks adequate experimental support.

Conclusions of Law

65. It appears that all parties are in agreement that it is generally appropriate to apply the requirements of 10 C.F.R. § 50.46 to MOX fuel. The question then arises as to whether the effects of fuel relocation should be considered in Duke's Appendix K analysis for determining compliance with 10 C.F.R. § 50.46. Given the potentially significant effects of fuel relocation on the results of a LOCA analysis, and given that Duke relies in part on a LOCA analysis in which fuel relocation effects were considered, we agree with Dr. Lyman that Appendix K should not be strictly applied to exclude consideration of relocation of the fuel during LOCAs. See Lyman Testimony at 3.

66. We agree with Dr. Lyman that these uncertainties cannot be addressed with mere calculations or analyses based on LEU performance. In his professional opinion, the only satisfactory way to address these uncertainties would be to conduct integral tests of MOX fuel 25

assemblies under LOCA conditions in such a manner that the impacts of the phenomena described in his testimony can be measured with reasonable accuracy and precision.

67. Certain characteristics of MOX fuel appear to exacerbate the effects of fuel relocation, thus leading to greater maximum cladding oxidation resulting from higher rupture node temperatures and potentially to higher PCTs. While there are several other known non-conservatisms in Appendix K, this one in particular appears to be relevant to the MOX LTA LAR because of its disproportionately large impact on the MOX LTAs compared to the LEU assemblies that comprise the remainder of the core. Given the potential impact on the ECR and PCT of relocation effects, it is not appropriate to omit consideration of this phenomenon from the Appendix K analysis for MOX fuel, especially in light of the fact that Duke considers fuel relocation in its best estimate models for the LOCA analysis for the co-resident LEU fuel.

68. In their testimony, Duke's witnesses claim that BREDL is presenting a "chicken and egg" paradox by arguing that LOCA testing of irradiated MOX fuel rods is necessary to support the safety case for the MOX LTA program at Catawba. Duke Testimony at 73-74. We don't believe that this is a fair characterization of BREDL's position. As Dr. Lyman testified, BREDL has pointed out that the necessary testing can be conducted with MOX fuel irradiated in Europe at European test facilities, of which there are several. The MOX LTAs that Duke proposes to load at Catawba are not merely incremental modifications of the LEU fuel that is in use at US reactors, but are radically different fuels with completely different microstructures. BREDL believes it would be irresponsible to use the Catawba station as a test reactor for these novel fuel assemblies. BREDL also notes that a loss-of-coolant accident at Catawba when MOX LTAs are present in the core would likely put an end to the MOX program in the United States. Thus, BREDL argues, it would be prudent to ensure that the likelihood of such an occurrence is low.

26

69. Indeed, under the circumstances, we believe that it is reasonable and necessary, in order to support a finding that use of MOX LTAs will pose no undue risk to public health and safety, to require the accumulation of additional test data regarding the behavior of plutonium MOX fuel under LOCA conditions.

Respectfully submitted, jane Curran Harmon, Curran, Spielberg, & Eisenberg, L.L.P.

1726 M Street N.W., Suite 600 Washington, D.C. 20036 202/328-3500 e-mail: Dcurrangharmoncurran.com August 6,2004 27