Text

Letter from David A. Repka to Administrative Judges Enclosing the Testimony of Duke Energy Corporation on Contention I
	ML041950059
Person / Time
Site:	Catawba
Issue date:	07/01/2004
From:	Repka D; Duke Energy Corp, Winston & Strawn, LLP
To:	Anthony Baratta, Elleman T, Austin Young; Atomic Safety and Licensing Board Panel
	Byrdsong A T
References
	-RFPFR, 50-413-OLA, 50-414-OLA, ASLBP 03-815-03-OLA, RAS 8086
	Download: ML041950059 (212)
	v • d • e

WINSTON & STRAWN LLP 1400 L STREET, N.W., WASHINGTON DC 20005-3502 202-371-5700 35 W W .r *00tooP_. 38314.3333 50w 101 WS . 43 R. -R-l , II AO4VOO 83H - Cm P.Kl. I 6 - l-0 0 IL 60601-0703 W.far00. MY I01 853 L- 01. CA 90071-1543 GSmP83 0 CA 94111-6584 I204 0 . 5 75110 aIos P 8312 L-W. .0 IDCZ NTf 312-556-5603 51 -204700 *113-15-1700 415-91-1000 41-52-317-75-75 33-6-53-44422 44207-1531025 DOCKETED USNRC July 6, 2004 (10:29AM)

July 1, 2004 OFFICE OF SECRETARY RULEMAKINGS AND ADJUDICATIONS STAFF Ann Marshall Young, Chairman Anthony J. Baratta Administrative Judge Administrative Judge Atomic Safety and Licensing Board Atomic Safety and Licensing Board U.S. Nuclear Regulatory Commission U.S. Nuclear Regulatory Commission Washington, D.C. 20555-0001 Washington, D.C. 20555-0001 Dr. Thomas S. Elleman Administrative Judge 5207 Creedmoor Road # 101 Raleigh, N.C. 27612 Re: Duke Energy Corporation, Catawba Nuclear Station, Units 1 and 2 (Docket Nos. 50-413-OLA, 50-414-OLA)

Dear Administrative Judges:

Enclosed for filing in the above-referenced docket is the testimony of Duke Energy Corporation on Contention 1.

Very truly yours, David A. Repka Counsel for Duke Energy Corporation Enclosure cc: See enclosed Certificate of Service i&npl41te S CLe gaE

UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSING BOARD In the Matter of: )

) Docket Nos. 50-413-OLA DUKE ENERGY CORPORATION ) 50-414-OLA

)

(Catawba Nuclear Station, )

Units 1 and 2) )

CERTIFICATE OF SERVICE I hereby certify that copies of "TESTIMONY OF STEVEN P. NESBIT, ROBERT C. HARVEY, BERT M. DUNN, AND J. KEVIN McCOY ON BEHALF OF DUKE ENERGY CORPORATION ON CONTENTION I" in the captioned proceeding have been served on the following by Federal Express overnight courier this 1" day of July, 2004.

Additional e-mail service has been made this same day, as shown below.

Ann Marshall Young, Chairman Anthony J. Baratta Administrative Judge Administrative Judge Atomic Safety and Licensing Board Atomic Safety and Licensing Board U.S. Nuclear Regulatory Commission U.S. Nuclear Regulatory Commission Washington, DC 20555-0001 Washington, DC 20555-0001 (e-mail: AMY@nrc.gov) (e-mail: AJB5@nrc.gov)

Dr. Thomas S. Elleman Office of the Secretary Administrative Judge U.S. Nuclear Regulatory Commission 5207 Creedmoor Road, #101 Washington, DC 20555 Raleigh, NC 27612 Attn: Rulemakings and Adjudications Staff (e-mail: elleman(eos.ncsu.edu) (original + one copy)

(e-mail: HEARINGDOCKET@nrc.gov)

Susan L. Uttal, Esq. Diane Curran Antonio Fernandez, Esq. Harmon, Curran, Spielberg & Eisenberg, LLP Margaret J. Bupp, Esq. 1726 M Street, N.W.

Office of the General Counsel, 0-15D21 Suite 600 U.S. Nuclear Regulatory Commission Washington, DC 20036 Washington, DC 20555 (e-mail: dcurrangharmoncurran.com)

(e-mail: slu~nrc.gov)

(e-mail: axf2(nrc.gov)

(e-mail: mjb5@nrc.gov)

Office of Commission Appellate Adjudicatory File Adjudication Atomic Safety and Licensing Board Panel Mail Stop 0-16C1 U.S. Nuclear Regulatory Commission U.S. Nuclear Regulatory Commission Washington, DC 20555 Washington, DC 20555 David A. Repka Counsel for Duke Energy Corporation 2

DC364735.1

July 1, 2004 UNITED STATES OF AMERICA DOCKETED NUCLEAR REGULATORY COMMISSION USNRC BEFORE THE ATOMIC SAFETY AND LICENSING BOARD July 6,2004 (10:29AM)

OFFICE OF SECRETARY RULEMAKINGS AND In the Matter of: ) ADJUDICATIONS STAFF

)

DUKE ENERGY CORPORATION )

) Docket Nos. 50-413-OLA (Catawba Nuclear Station, ) 50-414-OLA Units 1 and 2) )

)

TESTIMONY OF STEVEN P. NESBIT, ROBERT C. HARVEY, BERT M. DUNN, AND J. KEVIN McCOY ON BEHALF OF DUKE ENERGY CORPORATION ON CONTENTION I (MOX FUEL LEAD ASSEMBLY PROGRAM, MOX FUEL CHARACTERISTICS AND BEHAVIOR, AND DESIGN BASIS ACCIDENT (LOCA) ANALYSIS)

TABLE OF CONTENTS I. INTRODUCTION ...................................................... 1I II. OVERVIEW.......................................................................................................................5 A. The MOXFuel Project....................................................... 5 B. MOXFuel Experience....................................................... 7 C. LAR Overview ...................................................... 10 III. MOX FUEL APPENDIX K LOCA ANALYSIS .. 11..........................

1 A. Summary of Requirements ...................................................... 11 B. Overview of a PressurizedWater Reactor (PWR) Design Basis LOCA ................... 13 C. Overview of MOXFuel LOCA Analysis ............................. ......................... 18 D. Conservatisms in the LOCA Analysis ...................................................... 24 IV. DIFFERENCES IN LEU AND MOX FUEL BEHAVIORS ........................................... 26 A. Fuel-Related "Differences ...................................................... 27 B. Cladding-Related "Differences"...................................................... 37 V. FUEL RELOCATION AND RELATED ISSUES ..................................... 40 A. Fuel Relocation - Descriptionand Regulatory History .............................. ............. 41 B. VERCORS Tests ...................................................... 47 C. M5 T Ballooning...................................................... 51 D. Filling R ofatio R elocated Fuel...................................................... 54 E. Fuel-CladdingInteraction...................................................... 59 F. MOX Fuel Relative Power at High Burnup...................................................... 61 G. Conclusionson Fuel Relocation...................................................... 66 VI. UNCERTAINTIES ...................................................... 70 VII. CONCLUSIONS............................................................................................................... 74 i

July 1, 2004 UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSING BOARD In the Matter of: )

)

DUKE ENERGY CORPORATION )

) Docket Nos. 50-413-OLA (Catawba Nuclear Station, ) 50-414-OLA Units 1 and 2) )

)

TESTIMONY OF STEVEN P. NESBIT, ROBERT C. HARVEY, BERT M. DUNN, AND J. KEVIN McCOY ON BEHALF OF DUKE ENERGY CORPORATION ON CONTENTION I (MOX FUEL LEAD ASSEMBLY PROGRAM, MOX FUEL CHARACTERISTICS AND BEHAVIOR, AND DESIGN BASIS ACCIDENT (LOCA) ANALYSIS)

I. INTRODUCTION

1. (Nesbit) I, Steven P. Nesbit, am an Engineering Supervisor II employed by Duke Energy Corporation (Duke). I currently serve as the Duke Mixed Oxide (MOX) Fuel Project Manager. I have 22 years experience with Duke and 24 years experience overall in nuclear engineering and management, including both the commercial sector and U.S. Department of Energy (DOE) nuclear projects. I have particular experience in the area of design engineering, including nuclear safety analysis and nuclear reactor safety reviews. In my current position I direct the technical, Nuclear Regulatory Commission (NRC) licensing, and business activities associated with the MOX Fuel Project. A full statement of my Professional Qualifications is included as Attachment 1 to this testimony.

I

2. (Harvey) I, Robert C. Harvey, am a Senior Engineer employed by Duke, responsible for the Loss of Coolant Accident (LOCA) analyses supporting the Oconee Nuclear Station (Oconee), McGuire Nuclear Station (McGuire), and Catawba Nuclear Station (Catawba).

Prior to my employment by Duke, I was employed by Yankee Atomic Electric Company to perform LOCA and severe accident analyses to support various nuclear reactor licensing initiatives and core reloads. Overall, I have over 25 years of experience in nuclear thermal hydraulic and safety analyses. With respect to the Duke MOX Fuel Project, I provided oversight to AREVA (formerly Framatome ANP) in the performance of the supporting LOCA analyses discussed further below. A full statement of my Professional Qualifications is included as to this testimony.

3. (Dunn) I, Bert Dunn, am an Advisory Engineer employed by AREVA Framatome ANP Inc., located in Lynchburg, Virginia. I have 34 years experience in the nuclear engineering field - primarily in the area of LOCA analyses and safety analyses to support nuclear fuel design and licensing activities. For the Duke MOX Fuel Project, I am responsible for the LOCA analyses supporting Duke's proposal to utilize four MOX fuel lead assemblies at Catawba. A full statement of my Professional Qualifications is included as Attachment 3 to this testimony.

4. (McCoy) I, J. Kevin McCoy, am an engineer in the fields of metallurgy and materials engineering, employed by Framatome ANP, Inc. I have a Doctorate in Materials Engineering and a Masters Degree in Metallurgical Engineering. I have more than 20 years of experience in the nuclear industry, with my most recent work in the area of nuclear fuel including recent work on MOX fuel performance. A full statement of my Professional Qualifications is Attachment 4 to this testimony.

2

5. (All) Based on our specific job responsibilities, we are all very familiar with Duke's License Amendment Request (LAR), dated February 27, 2003. The LAR seeks NRC approval for Duke's proposal to use four MOX fuel lead assemblies at Catawba. The MOX fuel lead assemblies will be included in cores that will be predominantly comprised of Low Enriched Uranium (LEU) fuel assemblies.

6. (All) The purpose of this testimony is to specifically address Contention I of the Blue Ridge Environmental Defense League (BREDL), an intervenor in this NRC license amendment proceeding. That contention, as admitted by the Atomic Safety and Licensing Board (Licensing Board) in its Memorandum and Order of March 5, 2004, asserts that:

The LAR is inadequate because Duke has failed to account for differences in MOX and LEU fuel behavior (both known differences and recent information on possible differences) and for the impact of such differences on LOCAs and on the [design basis accident (DBA)] analysis for Catawba.'

7. (All) As discussed in this testimony, LOCA is the only DBA analysis at issue in this proceeding. In our testimony we show that, contrary to Contention I, Duke has filly accounted for "known" differences in MOX and LEU fuel behavior in its LAR; that Duke has performed conservative LOCA analyses using well-known and approved methodologies; and that the LAR meets applicable NRC requirements. Duke has also considered issues raised in a recent research proposal concerning "possible" differences in behavior between MOX fuel and

- LEU fuel, and has concluded that these issues do not affect the safety or compliance of the MOX fuiel lead assembly proposal.

l "Memorandum and Order (Ruling on Standing and Contentions)," Duke Energy Corp.

(Catawba Nuclear Station, Units I and 2), LBP-04-04 _ NRC _ (March 5, 2004, slip op. at 63).

3

8. (All) This testimony focuses in particular on Duke's LOCA analyses, which were completed in accordance with the methodology and acceptance criteria in NRC regulations in 10 C.F.R. § 50.46 and 10 C.F.R. Part 50, Appendix K. Consistent with the direction of the Atomic Safety and Licensing Board,2 this testimony on Contention I does not include issues and analyses related to LOCA dose consequences. Duke has in fact prepared assessments of doses following a design basis LOCA in accordance with applicable regulatory requirements and guidance.

Those analyses are addressed in the LAR, Attachment 3, Section 3.7.3, and in responses to NRC Staff Requests for Additional Information (RAIs). However, these matters are beyond the scope of Contention I and the scope of the testimony of this panel.

9. (All)Section II of this testimony provides an overview of the MOX Fuel Project and the requested regulatory approval.Section III provides an overview of LOCA analyses in general, and reviews the LOCA analyses that were performed for the MOX fuel lead assemblies.

Section IV addresses the potential differences in LEU and MOX fuel behavior that were postulated by BREDL.Section V reviews the fuel relocation issue raised by BREDL in more detail, and addresses potential fuel relocation and related phenomena in the context of the MOX fuel application. Section VI addresses the issue of uncertainties as they relate to lead test assembly programs and the LAR.Section VII provides the conclusion of this testimony, which is that the MOX fuel lead assembly program does not pose an undue risk to the health and safety of the public due to the LOCA issues raised in Contention I. In particular, MOX fuel does not pose a risk in the area of fuel relocation that is significantly different than for LEU fuel. That 2 Order (Confirming Matters Addressed at April 6 Telephone Conference), April 8, 2004 (at p. 2) ("With respect to Contention I, this contention encompasses those calculations involved in the determination of events up to and including LOCAs and DBAs, but does not include analyses related to any releases either in containment or offsite."); see also Tr. 1726-36.

4

risk is well-bounded by conservatisms in the regulatory acceptance criteria and in the LOCA analysis.

10. (All) Our testimony will show that the LOCA calculations performed by AREVA and submitted by Duke (i) meet the applicable regulatory requirements and (ii) adequately and conservatively characterize MOX/LEU differences. We will demonstrate that the MOX fuel LOCA analyses are inherently conservative, providing a large margin of safety for the public.

We will address purported MOX/LEU differences in detail, particularly those related to fuel relocation and related issues. We will address the BREDL concern over uncertainty in the context of the proposed Catawba lead assembly program. We will conclude that Contention I is not correct, and that the Catawba MOX fuel lead assembly program can proceed without adverse impacts on the health and safety of the public.

II. OVERVIEW A. The MOX Fuel Project

11. (Nesbit) On February 27, 2003, Duke submitted the LAR to the NRC to allow for the receipt and use of four MOX fuel lead assemblies at one of the McGuire or Catawba units.

On September 23, 2003, Duke amended the LAR to apply to Catawba only.

12. (Nesbit) Duke's proposal to use four MOX fuel lead assemblies is part of an important and ongoing nuclear non-proliferation program of the United States and the Russian Federation. The goal of this program is to dispose of surplus plutonium from nuclear weapons by converting the material into MOX fuel and using that fuel in nuclear power reactors. The current proposal for four MOX fuel lead assemblies supports the potential future use of larger quantities of MOX fuel at either Catawba or McGuire. Should such future use occur, however, it will be the subject of a separate NRC licensing action.

5

13. (Nesbit) The MOX fuel lead assemblies will be manufactured in France under the direction of AREVA. The MOX fuel assemblies will be based on the AREVA Advanced Mark-BW fuel assembly, a fully qualified LEU fuel assembly design that will be adapted for MOX application through changes to the fuel rod design. The Advanced Mark-BW fuel assembly is a standard lattice 17 x 17 fuel assembly specifically designed for use in Westinghouse reactors (such as Catawba). The Advanced Mark-BW adaptation for MOX applications, the Mark-BW/MOXI, is dimensionally and structurally identical to the Advanced Mark-BW with the only change appearing in the fuel rod internal design. The Advanced Mark-BW and the Mark-BW/MOXI share as a design feature M5Tm alloy fuel rod cladding, guide thimbles, and spacer grids. (MS is a trademark of AREVA.)

14. (Nesbit) Duke is currently targeting Catawba Unit 1, Cycle 16 (ClC16), with a Spring 2005 startup, for initial insertion of MOX fuel lead assemblies. Plans call for the lead assemblies to be irradiated for a minimum of two cycles to confirm acceptability of the fuel assembly design, verify the ability of Duke's and AREVA's models to predict fuel assembly performance, and confirm the applicability of the European database on MOX fuel performance to Duke's use of MOX fuel.

15. (Nesbit) The CIC16 core will be predominantly comprised of Westinghouse Robust Fuel Assembly (RFA) fuel assemblies. In addition to the RFA fuel and the four MOX fuel assemblies, there will be eight Westinghouse Next Generation Fuel (NGF) lead test assemblies in the core. The eight NGF test assemblies will be loaded into core locations that are not adjacent to the MOX fuel.

6

16. (Nesbit) On April 5, 2004, the NRC Staff issued a Safety Evaluation Report (SER), concluding that the health and safety of the public will not be endangered by operation with four MOX fuel lead assemblies at Catawba.

17. (Nesbit) The Advisory Committee on Reactor Safeguards (ACRS) has also considered the LAR. The ACRS Subcommittee on Reactor Fuels reviewed the matter during a public meeting (in which BREDL's consultant participated) on April 21, 2004. The full ACRS reviewed the MOX fuel lead assembly LAR in a public meeting (in which BREDL's consultant again participated) on May 6, 2004. In a letter dated May 7, 2004, the ACRS concluded that "under the restricted circumstances considered in both the Duke Power application and the NRC Staff's safety evaluation, the four mixed oxide lead test assemblies in non-limiting core locations that do not contain control rods can be irradiated in either of the Catawba reactor cores with no undue risk to the public health and safety."

B. MOXFuel Experience

18. (Nesbit) Standard light water reactor LEU fuel consists of sintered uranium oxide pellets enclosed in zirconium alloy cladding to form fuel rods. The fuel rods are bundled into a fuel assembly, and the fuel assemblies reside in the reactor core where they produce power through nuclear fissions in the fuel pellets.

19. (Nesbit) The uranium in LEU fuel is slightly enriched in 235U (typically in the range of 4-5% 235U). Initially, energy comes from fissions in the uranium, mainly from thermal fissions in 235U. As LEU fuel is used, it also produces plutonium from neutron absorption in 238U. 239 Pu and 241Pu, like 235U, are fissionable at thermal neutron energies. Later in the life of the LEU fuel, the initial 23 5U has been depleted and appreciable quantities of plutonium have built up, so LEU fuel energy comes from fissions in both uranium and plutonium.

7

20. (Nesbit) The plutonium that is produced in LEU fuel can be recovered by reprocessing the spent fuel. This involves discharging the fuel from the reactor, dissolving it in acid, and chemically separating the uranium, plutonium, and fission products from the resulting solution.

21. (Nesbit) MOX fuel is another type of nuclear fuel. MOX fuel is essentially identical to LEU fuel except that the MOX fuel pellets are comprised of a small amount of plutonium oxide mixed with the remainder uranium oxide (typically, depleted uranium in the range of 0.2-0.3% 235U). MOX fuel is used on a large scale in some European reactors, utilizing plutonium that has been recovered from reprocessing uranium fuel. This type of plutonium is classified as reactor grade (RG) because it contains at least 20% 240 Pu. Catawba will utilize weapons grade (WG) plutonium. WG MOX fuel contains plutonium with less than 7% 240 Pu.

22. (Nesbit) In the United States, MOX fuel development programs were conducted in the 1960s and 1970s in the anticipation that wide-scale reprocessing of spent nuclear fuel in the United States would result in substantial quantities of separated plutonium. MOX fuel lead test assembly programs were conducted at four United States commercial nuclear power plants:

Big Rock Point, Dresden, San Onofre, and Quad Cities. In the late 1970s, in an attempt to discourage the commercial separation of plutonium, the United States government instituted a policy that the United States would not reprocess commercial reactor fuel. As a result, United States plans for reprocessing and MOX fuel use were discontinued.

23. (Nesbit) In 1980, the NRC approved the loading and use of four MOX fuel assemblies at Ginna Unit 1 (out of a total of 121 assemblies in a Ginna core). The Ginna MOX fuel assemblies were the last MOX fuel assemblies loaded into a domestic commercial nuclear power reactor in the United States.

8

24. (Nesbit) Despite the policy change in the United States, in the last three decades there have been substantial advances in MOX fuel technology worldwide. Other countries reprocess spent nuclear fuel on a large scale, producing plutonium than can be fabricated into MOX fuel. Nuclear power plant MOX fuel use began in German reactors in the 1970s.

Currently, more than thirty reactors in France, Germany, Switzerland, and Belgium are using MOX fuel. European reactors operate with cores comprised of a mixture of LEU fuel assemblies and MOX fuel assemblies, with MOX fuel core fractions as high as 36%. MOX fuel fabrication and use has reached a state of industrial maturity in Europe.

25. (Nesbit) European experience indicates that MOX fuel performs well in nuclear power reactors. Fuel failure rates are commensurate with those of LEU fuel. Significantly, there has never been a MOX fuel failure attributable to the fuel pellet material.

26. (Nesbit) MOX fuel is fundamentally similar to LEU fuel. MOX fuel is predominantly uranium oxide, and physical characteristics such as thermal conductivity and heat capacity are very close between the fuel types. Differences such as fission gas release are accommodated by fuel design and operation. Plutonium is the primary fissionable element in MOX fuel, unlike LEU fuel which has predominantly uranium fissions at beginning of life. This difference affects nuclear characteristics such as thermal flux level and effective delayed neutron fraction. Again, these differences are well understood and accommodated by fuel assembly and core design.

27. (Nesbit) Irradiation of plutonium in MOX fuel is an effective means of rendering WG plutonium unattractive for theft or diversion into nuclear weapons. Irradiation of WG 3 BAW-10238P, MOXFuel Design Report (2002). BAW-10238P, Revision 1, was issued in May 2003, as referenced below. There is also a non-proprietary version of this report that includes the pertinent information.

9

plutonium in MOX fuel destroys much of the original plutonium and degrades the isotopics of the remainder.

C. LAR Overview

28. (Nesbit) The LAR consists of a cover letter and Attachments 1-6. Attachments 1 and 2 are marked-up copies of the McGuire and Catawba Technical Specifications, respectively.

The Technical Specification changes are to the following areas: spent fuel storage, reactor core design features, criticality, and Core Operating Limits Report approved methodologies. LAR provides the safety justification for the proposed changes. Attachment 4 is Duke's No Significant Hazards Consideration analysis. Attachment 5 is Duke's assessment of environmental impacts associated with the MOX fuel use. Attachment 6 contains a request for exemptions from selected NRC regulations. These exemption requests clarify the applicability of current NRC regulations to MOX fuel, and provide for the use of M5TM cladding.

29. (Nesbit) Attachment 3 to the LAR, the safety analysis, covers four major technical areas: description of the MOX fuel lead assemblies, effects of four MOX fuel lead assemblies on plant operation, safety analyses of operation with four MOX fuel lead assemblies, and the impact of operation with four MOX fuel assemblies on risk. In addition to the information in , the LAR and associated responses to NRC RAIs include by reference a number of topical reports that were developed, wholly or in part, to support the MOX fuel program. Those topical reports, reviewed separately by the NRC, are listed below:

BAW-10238P, Revision 1, MOXFuel Design Report (May 2003).

Background information on MOX fuel and evaluations of the impact of MOX fuel pellets on the fuel assembly and fuel rod design.

BAW-10239P, Advanced Mk-BW Fuel Assembly Mechanical Design (March 2002). Description of the fuel assembly design.

10

BAW-10231P, COPERNIC Fuel Rod Design Computer Code (September 1999). Computer code used for mechanical analyses of MOX and LEU fuel.

DPC-NE-1005P, Nuclear Design Methodology Using CASMO-4/SIMULATE-3 MOX (August 2001). Description of nuclear analysis methods for application to LEU and MOX fuel at McGuire and Catawba.

DPC-NE-2005P-A, Revision 3 (September 2002), Thermal-Hydraulic Statistical Core Design Methodology, Appendix E.

Thermal-hydraulic design methodology for the MOX fuel assembly design.

The LAR also references the following topical report that is independent of fuel type:

BAW-1 0227P-A, Evaluation ofAdvanced Claddingand Structural Material (M5) in PWR Reactor Fuel (February 2000). Evaluation of M5Th1 as an advanced cladding and structural material.

30. (Nesbit) Contention I raises issues that have been addressed in the LAR and associated correspondence. Specifically, the impact of four MOX fuel lead assemblies on a LOCA analysis is addressed in LAR Attachment 3, Section 3.7.1 and in a Duke RAI response submitted November 3, 2003. The relevant portions of these documents are included as Exhibits I and 2, respectively, to this testimony.

III. MOX FUEL APPENDIX K LOCA ANALYSIS A. Summary of Requirements

31. (Harvey, Nesbit) Contention I refers to the impact of MOX fuel on "LOCAs and on the DBA accident analysis" for Catawba. We read this to refer to the impact on a design basis LOCA scenario, and more specifically on the design basis LOCA analysis. As acknowledged by BREDL in response to Duke Interrogatory 4, "BREDL makes no assertions [in Contention I]

regarding design basis accident (DBA) scenarios or accident analyses other than for the design 11

basis LOCA."4 The relevant LOCA analyses (Section 3.7.1 of the LAR and the RAI responses) that Duke has provided in connection with its application address the requirements of 10 C.F.R.

§ 50.46 and 10 C.F.R. Part 50, Appendix K.

32. (Harvey) 10 C.F.R. § 50.46 states that each light-water reactor must be provided with an emergency core cooling system (ECCS) that must be designed so that its calculated cooling performance following postulated design basis LOCAs conforms to specific acceptance criteria. ECCS cooling performance must be calculated in accordance with an acceptable Evaluation Model (EM) and must be calculated for a number of postulated LOCAs of different sizes, locations, and other properties sufficient to provide assurance that the most severe postulated LOCAs are evaluated. The EM must include sufficient supporting justification to show that the analytical technique realistically describes the behavior of the reactor system during the postulated LOCA.

33. (Harvey) The 10 C.F.R. § 50.46 acceptance criteria are:

The calculated maximum fuel element cladding temperature

("peak cladding temperature" or "PCT') shall not exceed 22000 F.

The calculated total oxidation of the cladding shall nowhere exceed 0.17 times the total cladding thickness before oxidation. If cladding rupture is calculated to occur, the inside surfaces of the cladding shall be included in the oxidation, beginning at the calculated time of rupture.

The calculated total amount of hydrogen generated from the chemical reaction of the cladding with water or steam shall not exceed 0.01 times the hypothetical amount that would be generated if all of the metal in the cladding cylinders surrounding the fuel, excluding the cladding surrounding the plenum volume, were to react.

4 "Blue Ridge Environmental Defense League's Response to Duke Energy Corporation's First Set of Interrogatories and Requests for Production of Documents," April 14, 2004, at3.

12

Calculated changes in core geometry shall be such that the core remains amenable to cooling.

After any calculated successful initial operation of the ECCS, the calculated core temperature shall be maintained at an acceptably low value and decay heat shall be removed for the extended period of time required by the long-lived radioactivity remaining in the core.

34. (Harvey) Appendix K defines the required and acceptable features of the EM to be utilized in a LOCA analysis. These include:

Sources of heat during the LOCA

Swelling and rupture of the cladding and fuel rod thermal parameters

Blowdown phenomena

Post-blowdown phenomena; heat removal by the ECCS

35. (Harvey) The results of the LOCA analyses define the "LOCA limits" (allowable core power peaking limits) which will be employed to satisfy the criteria specified in 10 C.F.R.

§ 50.46.

B. Overview of a PressurizedWater Reactor (PWR) Design Basis LOCA

36. (Dunn) A large break LOCA evolves through three phases. A realistic (best estimate) analysis of the event would consider these phases as recognizable, but overlapping or mixed. However, for conservative deterministic approaches, such as that used by AREVA in its LAR Appendix K LOCA analyses for MOX fuel lead assemblies, the phases are treated as distinct. Figure 1 shows the cladding temperature response of the core hot pin during a design basis LOCA as calculated using the AREVA Appendix K methodology. This particular example is the MOX fuel case with the highest PCT (see Table Q14-1 of Exhibit 2).

13

Figure 1 Catawba MOX LOCA Temperatures 2200_ _ _ ____

T t 2000 f -pFhd UT& T_

XS3 180DI OZJ^S j1400 /

1000 __

800 600 0 10 300 400 500 600 Thme, s 14

37. (Dunn) The first phase - blowdown - initiates with the opening of a break in the reactor coolant system (RCS) and lasts until the RCS has depressurized to an approximate equilibrium in pressure with the reactor building. For the Catawba LOCA shown on Figure 1, the blowdown extends from accident initiation (0 seconds) to 25 seconds. During this phase, the core is cooled primarily by the flow of water through the reactor coolant system and out of the break. With the loss of the pressure boundary, the reactor coolant flashes or boils to steam within the RCS. These areas of steam (referred to as voids) reduce the moderating ability of the coolant, effectively shutting down the nuclear reaction and reducing heat production (power) to decay heat levels. Because the power within the nuclear material has dropped significantly, the fuel pellet temperature decreases. However, the increasing voids lead to departure from nucleate boiling on the exterior of the fuel pins, dramatically reducing heat transfer from the pins. As a result, the cladding heats up quickly. As shown in Figure 1, the initial cladding temperature peak occurs as the temperatures of the cladding and the pellet approach each other. At this point the capability to transfer heat from the pellet through the cladding to the coolant is approximately equal to the decay heat rate. This typically occurs at 4 to 10 seconds after accident initiation with temperatures in the 1400 to 1800TF range. Following this initial peak in cladding temperature, the continued reduction in power allows the cladding and fuel temperatures to decrease until just before the end of the blowdown phase.

38. (Dunn) During blowdown, the injection of emergency coolant has not provided substantial benefit because the RCS continues to expel coolant through the break. By the end of blowdown, the RCS, as modeled in deterministic Appendix K evaluations, is nearly empty, heat transfer has further degraded due to the lack of liquid, and the cladding and fuel begin to increase in temperature. The next phase - the refill phase - is defined as the time required to inject 15

sufficient emergency coolant to fill the lower head and lower plenum of the reactor vessel and re-initiate flow of coolant into the reactor core. In Figure 1, the refill phase lasts from 25 seconds to 38 seconds. During refill the coolant is pooling below the core in the lower head, and there is no mechanism for flow up through the core. Fuel pellet and cladding heatup is nearly adiabatic during refill, with the temperatures rising perhaps 200-300TF before the end of the phase.

39. (Dunn) When the coolant in the reactor vessel rises to the bottom of the core, the third phase initiates - reflood. In Figure 1, the reflood phase initiates at 38 seconds and extends for the remainder of the accident simulation. During the reflood phase water in the bottom of the core is heated to boiling, creating an upward two-phase flow that restores cooling to the upper portion of the core. Early in reflood, the coolant level in the core is low with a low boiling rate, so there is insufficient two-phase flow in the upper regions of the core to remove decay heat. As a result, the upper regions of the cladding continue to increase in temperature during the initial stage of reflood.

40. (Dunn) It is during this early reflood phase that cladding temperature and stress develop to the point where cladding swelling and rupture are possible. The RCS is at a very low pressure, maximizing the stress pushing the cladding outward. As the temperature increases, the conditions for cladding ballooning and rupture could be met. At the rupture, the cladding would pull away from the fuel pellet to a diameter greater than the original cladding diameter in that location. Figure 2 shows two views of a sample of cladding from a LOCA test at Argonne National Laboratory. At locations A and D, which are removed from the ballooning position, the cladding has not experienced significant strain. Closer to the rupture, location B, the cladding experiences higher strain. At the rupture, location C, the cladding undergoes the largest strain and is expanded outward asymmetrically around the rupture.

16

Figure 2

+ - 1.

II A BC D A B C

41. (Dunn) Several important heat transfer processes ensue at the ruptured location on the pin. The pellet-to-cladding gap is increased significantly, reducing the pellet-to-cladding heat conductance and lowering, for the moment, the energy transferred from the fuel to the cladding. The cladding surface area increases, substantially increasing the rate of energy transport from the cladding to the coolant. Furthermore, the intrusion of the cladding into the coolant flow channel creates flow turbulence and mechanical interactions between entrained water droplets and the cladding, both of which act to improve the rate of heat transfer from the cladding to the coolant. The net effect, shown in Figure I in the plot of clad temperature at the ruptured location (or "node"), is a short-term cooling of the cladding at the rupture location.

Eventually, the cladding at the ruptured location resumes heatup at an equivalent rate to the nearby unswelled portions of the hot pin.

42. (Dunn) As the water level rises in the core, boiling increases, upper region two-phase flow increases, and the heat transfer approaches the ability to remove decay heat. At this 17

point, the last cladding temperature peak occurs. For deterministic Appendix K evaluations, this reflood cladding temperature peak is generally the highest of the entire event. Thereafter, cooling in the upper core continues to improve, and the cladding and pellet temperatures decrease gradually until the core is quenched (covered by a two-phase mixture) and long term cooling thereby established.

C Overview ofMOX FuelLOCA Analysis

43. (Dunn) Except for the MOX fuel-related changes discussed below, the LOCA analysis cases discussed in Duke's MOX fuel lead assembly LAR and in the associated responses to NRC RAIs were performed using the NRC-approved AREVA deterministic LOCA EM. The EM was approved by the NRC for application to Westinghouse-designed four loop PWRs (such as Catawba) with LEU fuel. To apply the EM to MOX fuel, a review of potential differences between LEU and MOX fuel that could affect the calculation of LOCA results was conducted and, where necessary, changes were made to the evaluation techniques. These changes and the justification for the changes are also described in Section 3.7 of the LAR.

44. (Dunn) The LEU-based EM is easily adapted to the prediction of LOCA behavior for MOX fuel. This is because, with the exception of decay heat and initial fuel enthalpy (heat content of the fuel), the LOCA response is primarily controlled by system phenomena.

NUREG/CR-5249 5 captures the importance of the various phenomena impacting the results of a LOCA. The identified controlling phenomena are listed in 13 categories. Twelve of the phenomena categories are related to the reactor system and are independent of the fuel type.

Only one category includes phenomena that relate to the nuclear material (e.g., MOX vs. LEU).

5 NUREG/CR-5249, EGG-2552, "Quantifying Reactor Safety Margins - Application of Code Scaling Applicability, and Uncertainty Evaluation Methodology to a Large-Break Loss of Coolant Accident" (December 1989).

18

Within this one category, fuel pellet enthalpy at operating conditions, fuel decay heat, gap conductance, and cladding oxidation are listed as significant contributors. Only the first three of these relate to the nuclear material. Thus, most of the approved EM is clearly already appropriate to the modeling of MOX fuel, with no adjustments.

45. (Dunn) That being said, some characteristics of the nuclear material that relate to the generation and conduction of energy from the pellet through the cladding to the coolant have the potential to affect the results of the Appendix K LOCA calculation. To account for these characteristics, AREVA reviewed MOX fuel characteristics for potentially LOCA-significant phenomena. Areas of concentration during the review were reactor kinetics, decay heat, and thermal and mechanical properties, including MOX pellet enthalpy. The LAR, Section 3.7.1.1, documents: (i) the phenomena that were given specific consideration for the MOX fuel lead assembly LOCA evaluation; and (ii) the disposition that was given to each phenomenon.

46. (Dunn) For the MOX lead assembly core at Catawba, the reactor kinetics will be dominated by the LEU fuel because the MOX lead assemblies comprise only 2 percent of the core. However, those differences that do exist, lower beta effective and a more negative moderator coefficient in MOX fuel, will act to reduce the neutron power generation in the MOX fuel relative to the surrounding LEU fuel. As is discussed in Section III.D of this testimony, a MOXILEU decay heat comparison indicates a beneficial result for MOX fuel. For the first several thousand seconds, well beyond the time of PCT following a LOCA, the decay heat for a MOX fuel assembly, if operated at the same power as an LEU assembly, would be lower than the corresponding LEU fuel assembly. For both of these phenomena, however, the MOX assemblies were conservatively evaluated using neutron power-related characteristics and decay heat characteristics appropriate for LEU assemblies.

19

47. (Dunn) Important properties associated with the conduction of energy out of the pellet are the pellet thermal properties and the fuel to cladding gap coefficient (gap conductance).

Within the pellet, only the thermal conductivity differs in any substantive way between MOX and LEU pellets of the lead assembly design. The difference, though slight, acts to increase the initial enthalpy of the MOX pellet. Therefore, the MOX thermal conductivity from COPERNIC, approved for MOX applications, was incorporated into the LOCA evaluations. There exists no significant difference in heat capacity between MOX and LEU. Composition of the gas in the pellet-cladding gap can affect the gap coefficient (thermal conductance between the pellet and the cladding). As the fuel is used, higher fission gas release rates for MOX fuel can impact the gap coefficient. To provide for this, the AREVA LOCA analysis was based on MOX-specific gap gas compositions obtained from COPERNIC.

48. (Dunn) As discussed above, the MOX fuel lead assemblies will use MSTM cladding. For many years Zircaloy-4 was the standard cladding material used worldwide in PWRs. Over the past decade the nuclear industry has been moving toward the use of advanced cladding materials instead of Zircaloy4. M5Tm is an advanced cladding material that provides superior corrosion resistance, lower irradiation growth, and better ductility retention than Zircaloy-4. M5rm has been reviewed by the NRC and approved for use as a cladding material.

It is currently being used in fuel operated in nine United States reactors. M5Tm is also being used in foreign nuclear power reactors, including MOX fuel applications.

49. (Dunn, Harvey) ZIRLOTm is another recently-developed advanced alloy cladding material. Westinghouse developed ZIRLO1m in the same time frame that AREVA developed M5Tm. ZIRLOTm is used in the predominant co-resident fuel at Catawba, the Westinghouse RFA 20

design. Zircaloy-4, M5T, and ZIRLOTM are all predominantly zirconium, and the unirradiated properties of the materials in cladding applications are very similar.

50. (Dunn, Nesbit) The use of M5T for MOX fuel assemblies is not a difference "in MOX and LEU fuel behavior" as was specifically called out in Contention I. LEU fuel assemblies commonly use advanced cladding materials such as M5T and ZIRLO'. However, BREDL has alleged that the characteristics of M5TM may exacerbate MOX/LEU differences, so M5Th cladding performance is also addressed in this testimony.

51. (Dunn) The modeling of M5Tm cladding within the LOCA EM is specific to that material. Many characteristics (e.g., thermal conductivity, heat capacity) are very similar to Zircaloy-4. The most significant difference (relevant to a LOCA event) lies within the cladding swelling (ballooning) and rupture model.

52. (Dunn) Cladding ballooning occurs when the cladding is heated to elevated temperatures with an internal pressure that exceeds the external pressure (coolant pressure).

Depending on the extent of the ballooning, the cladding may rupture, relieving the driving pressure difference and terminating the ballooning effect for that rod. The cladding ballooning and rupture effects are modeled in the EM.

53. (Dunn) The modeling of rupture involves the prediction of the rupture temperature of the cladding as a function of cladding stress and the prediction of cladding strain as a function of the rupture temperature. Although a specific M5T rupture temperature versus stress correlation has been developed, the predictions do not differ significantly from the Zircaloy-4 correlations. The cladding strain, however, differs in magnitude, because M5TM is slightly less ductile prior to irradiation than Zircaloy-4. In other words, unirradiated M5T will 21

strain (deform) less before rupture than Zircaloy-4, for the same applied stress. The temperature distribution of strain also differs.

54. (Dunn) The established rule for deterministic LOCA calculational approaches, Appendix K to 10 C.F.R. § 50.46, requires that the degree of cladding swelling and incidence of rupture not be underestimated. Because all claddings tend to (i) embrittle with irradiation, and (ii) potentially accrue added strength due to pellet-cladding bonding, deterministic LOCA evaluation models use unirradiatedcladding properties to maximize the predicted strain. This same approach was incorporated within the LOCA evaluations for MOX fuel presented in the LAR.

55. (Dunn, Harvey) LOCA calculations are performed at a variety of plant conditions to establish LOCA limits (allowable peaking) which ensure compliance with the criteria of 10 C.F.R. § 50.46. For the MOX lead assembly calculations, the most severe results occur for axial power peaks at the 10.3 ft elevation (i.e., for power distributions that are peaked close to the top of the 12 foot active core). The most limiting results and the corresponding acceptance criteria are shown below, with the acceptance criteria in parentheses:

Peak Cladding Temperature (PCT) 2019.50 F (22000 F)

PCT at Ruptured Location 1750 0 F (22000 F)

Local Oxidation 5.2% (17%)

Total Core Oxidation 0.4% (1%)

This most limiting case was evaluated at a burnup of 30 GWD/t (see Table Q14-1 of Exhibit 2).

The fuel and cladding temperatures from this case are shown in Figure 1.

56. (Dunn, Harvey) The maximum calculated cladding strain for this case is 51 percent and the flow blockage due to this ballooning is 52 percent of the coolant channel 22

surrounding the hot pin. This amount is well within the coolable geometry limit (specified by the AREVA LOCA evaluation model) of 90 percent.

57. (Dunn, Harvey) Long term cooling is not directly assessed by the LOCA

_ calculation. Long term cooling is a system determination requiring (i) that a cooling path is available from pumped injection and (ii) that the flow pattern within the reactor vessel is such as to prevent the precipitation of boric acid. Both of these factors are dependent on core wide phenomena and neither factor changes with the inclusion of four MOX fuel lead assemblies within the core. The conclusion that the Catawba operating procedures meet these conditions is established in the Catawba Final Safety Analysis Report (FSAR) and is not altered with four MOX fuel lead assemblies.

58. (Dunn, Harvey) The analyses in LAR Section 3.7.1 (Exhibit I) show a difference of less than 40F between the MOX fuel and LEU fuel comparison cases, with MOX fuel being the higher value. This is not a significant difference in a LOCA PCT response. The MOX and LEU results are thus essentially the same. As discussed further in Section III.D of this testimony, the MOX results are actually conservative because AREVA adopted the simplified approach of using LEU characteristics when it was obviously conservative to do so for MOX fuel. In other words, credit was not taken for beneficialaspects of MOX fuel.

59. (Nesbit) The NRC Staff reviewed the LOCA analyses of the MOX fuel lead assemblies. Section 2.4.1 of the April 5, 2004 NRC Safety Evaluation Report6 states: ". . . the NRC staff concludes that the effect of four MOX LTAs has been conservatively evaluated and has been demonstrated to be in compliance with the requirements of 10 C.F.R. § 50.46."

6 R.E. Martin (NRC) to H.B. Barron (Duke Energy), "Safety Evaluation for Proposed Amendments to the Facility Operating License and Technical Specifications to Allow Insertion of Mixed Oxide Fuel Lead Assemblies" (April 5, 2004) ("SER").

23

D. Conservatisms in the LOCA Analysis

60. (Dunn) The conservatisms incorporated in the MOX LOCA evaluation performed for the LAR include conservatisms (or "margin") inherent in the deterministic Appendix K approach and specific conservatisms that we employed in the MOX fuel LOCA analysis.

61. (Dunn) The major conservatisms included by regulation within the approved Appendix K method are:

_* A decay heat approximately 10 percent higher than the 95 percent confidence level and 30 to 35 percent higher than the nominal expectation.

The use of the Baker-Just oxidation (metal-water reaction) correlation, which incorporates a reaction rate 30 to 50 percent above the available data.

The use of the highest allowable local power level. Plants actually

_ operate with peaking significantly lower than the limiting values defined to be acceptable by the LOCA analyses. This conservatism is quantified for representative Catawba operation cycles in Section IV of this testimony.

The retained use of full double-ended break areas. The NRC is currently reviewing the use of substantially reduced break areas based on the risk of the large break occurring. Leak before break analyses indicated that the probability of an instantaneous double-

_ ended break is very low.

Limiting ECCS bypass assumptions. Nearly all of the ECCS water delivered to the reactor coolant system prior to the end of blowdown is assumed to be bypassed, and therefore unavailable for core cooling.

62. (Dunn) The accrued margin provided by all of the conservatisms identified above, in terms of a change in the PCT, can be estimated from applications of "best estimate" models that include a realistic treatment for most of the conservatisms. In a recent analysis, AREVA performed both Appendix K calculations and best estimate calculations on a three-loop Westinghouse-designed plant. The Appendix K PCT obtained was between 2100 and 2200TF, 24

while the nominal or best estimate PCT lay below 1500'F. Because even the best estimate approach retains some conservatisms or biases (e.g., use of full double-ended break areas), a judgment can be made that the conservatisms in the Appendix K methods provide more than 600'F margin to best estimate LOCA results.

63. (Dunn, Nesbit) The best example of additional conservatisms incorporated in the specific MOX fuel LOCA calculations for the LAR is the use of the LEU decay heat model, which was referred to above. A comparison of LEU and MOX decay heat on a one-to-one basis (Figure 3 below) would indicate that MOX falls 3 to 5 percent below the LEU curve during the time period of importance. Because the PCT is strongly influenced by the decay heat, this effect alone is estimated to be a conservatism of up to 750 F on PCT.

Figure 3 MOX/LEU Decay Heat Ratio 1.000 r , ,I r I 0.980 0-960 I N I

_ I 3 u.94U 1 0.920 0.900 0.880 o 0s I 5 10 15 20 25 30 Time (minutes) 25

64. (Dunn) There are also conservatisms in the MOX fuel LOCA evaluation in the use of LEU fuel neutronics coefficients. For example, the void reactivity coefficient for the MOX assemblies will be more negative than that for the LEU assemblies. Therefore, a small additional depression of the neutron power in the MOX assembly will occur as voids shut down the reactor during early blowdown. Because no multidimensional analysis of the core during LOCA has been conducted, the magnitude of the neutronic effect (beneficial for MOX fuel) has not been determined.

65. (Dunn) As is discussed further below, BREDL points to one factor not included in the LOCA model and cites that as a non-conservatism: fuel relocation. Fuel relocation is not modeled for either LEU fuel or MOX fuel. However, the possible impact of fuel relocation on compliance with the acceptance criteria must be considered in the context of all of the conservatisms already built into the models and the criteria. Fuel relocation is discussed in further detail in Section V of this testimony.

IV. DIFFERENCES IN LEU AND MOX FUEL BEHAVIORS

66. (Nesbit) Contention I focuses on "differences in MOX and LEU fuel behavior" and "the impact of such differences on LOCAs." In an attempt to obtain some understanding of what those asserted differences and impacts might be, beyond those already addressed in the LAR, Duke posed interrogatories to BREDL. In BREDL's response 7 to Duke Interrogatory 4, BREDL listed 19 MOX fuel behaviors that "... will affect a LOCA scenario or analysis in a manner different than LEU fuel." In response to Duke Interrogatory 5, BREDL listed 8 MOX fuel cladding (MS')behaviors that ". . . will affect a LOCA scenario or analysis in a manner different than LEU fuel cladding behavior." BREDL added the statement that "BREDL does not 7 "Blue Ridge Environmental Defense League's Response to Duke Energy Corporation's First Set of Interrogatories and Requests for Production of Documents," April 14, 2004.

26

assert that all of these behaviors necessarily will have a significant impact." In fact, BREDL did not highlight any behavior as having a significant impact. In this section of the testimony, Duke examines each behavior listed by BREDL in some detail. Duke has determined that each behavior is either inapplicable to the design basis LOCA calculation, already addressed in the LOCA evaluation, clearly insignificant in terms of impact on the LOCA calculation, and/or specifically addressed later in the testimony.

A. Fuel-Related "Differences" Item 1: Rod centerline temperatureas afunction of rodpower-

67. (Harvey, Dunn) Fuel rod centerline temperature is not a direct input in the LOCA analysis. The related input parameter is the fuel rod stored energy (enthalpy) or fuel pellet volumetric average temperature. This parameter is obtained using an NRC-approved fuel performance computer code/methodology. The LOCA analysis calculations model the fuel pellet properties (fuel dimensions, conductivity, and heat capacity); model inputs are adjusted to match the fuel rod stored energy determined by the approved fuel performance methodology.

For the MOX fuel LOCA analysis, the COPERNIC fuel performance code is used. COPERNIC is also used in the LEU fuel LOCA comparison analysis presented in the LAR. The COPERNIC computer code considers the difference in fuel properties between the LEU pellets and MOX pellets, and therefore adequately represents the differences in fuel centerline temperature between the MOX and LEU fuel rods.

68. (Harvey, Dunn) Figure 4 is taken from Figure Q21-2 in Duke's November 3, 2003 RAI response (Exhibit 2). Figure 4 shows that the MOX and LEU fuel pellet temperature profiles at the beginning of life are very close at initiation of a LOCA. This is because of the fundamental similarity between MOX and LEU fuel in the areas of thermal conductivity and pellet radial power, as shown by Figures 5 and 6, which also are reproduced from Figures Q21-1 27

and Q7-1 in the November 3, 2003 RAI response (Exhibit 2). There is no significant difference between fuel types due to the initial temperature profile.

Figure 4 MOX and LEU Fuel Pin Temperature Profile Comparison at Loss of Coolant Accident Initiation (Figure Q21-2) 35M

-- -- -- - -- -- -- -- -- - ---- - -- - -- - -- -

3000

..... ----- * ----- ------ ----- - ---- I- ----

2500 --------

----- ----------- ------ ----

1 2000 - ---- ------ ----- -----

------ -----

- -

I150D LE------------- ---- ----------- ----

1000

.4 ueltMOXI 500 0 1 2 3 4 5 45 7 S 9 10 Radial Mesh POint (M) 28

Figure 5 Thermal Conductivity Comparison for MOX and LEU Fuel (Fuel porosity of 0.0479)

(Figure Q21-1) 9.OE-4 44

- ,LWEU § GWdtmtU

,LEU A @ 40 GWd/mtU

- 4wtMOX@OGWd/mthm 04 -- - - ------- ; . ri------------

\ . ' ' - K- 4wet%MOX~t400

\ :: : GWd/mthm: : :

0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 Fue Temperatu:

(F) 29

Figure 6 Pellet Radial Power Profiles 0.5 GWD/t Item 2: Magnitude, timing, radionuclide composition and power history dependence offission gas release (during normaloperation)-

69. (Harvey, Dunn) The fission gas release from the fuel pellet has two effects. First, the fission gas that is released to the pellet-cladding gap degrades the gas conductivity, which can result in a higher temperature drop across the gap. Second, the released fission gas also increases the fuel rod internal pressure. The internal pressure has a mild effect on gap conductance, and also impacts the fuel rod rupture, swelling, and flow blockage during the LOCA event. Fission gas release is modeled in the approved fuel performance code specifically 30

for both MOX and LEU fuel. For the MOX fuel LOCA analyses, the fission gas release results from COPERNIC are included as input into the LOCA calculations. COPERNIC explicitly models MOX fuel effects on fission gas release. As mentioned above, COPERNIC fission gas release results for LEU are used as input to the LEU comparison LOCA analysis presented in the LAR. Therefore, the fission gas released during normal operation is explicitly modeled on a fuel type-specific basis (MOX or LEU) in the LAR LOCA analysis. In BREDL's response to Duke Interrogatory 32.f,8 BREDL acknowledged that this behavior is not relevant to Contention I.

Item 3: Fuel-cladinteraction-

70. (Harvey) This phenomenon is addressed in detail in Section V.E of this testimony.

Item 4: Peak cladding temperature (PC) -

71. (Harvey) The PCT is the principal result or output of the LOCA analysis. The PCT acceptance criterion (<2200'F) is based on the performance of the cladding material, not the fuel pellet material. The MOX fuel-specific analyses summarized in Exhibit 2, Table Q14-1, demonstrate that MOX fuel meets the PCT acceptance criterion.

Item 5: Oxidation potential-

72. (Harvey) The oxidation of the fuel pellet is not modeled in the LOCA analysis method. Therefore, this effect does not impact the LOCA analysis for either MOX or LEU fuel.

Item 6: Linear heat generation rate -

73. (Harvey) The linear heat generation rate (LHGR) is an input to the LOCA analysis which directly affects the sources of heat in the LOCA analysis and the resulting PCT.

The fuel initial stored energy and the decay heat in the limiting fuel rod (hot rod) are both influenced by the assumed LHGR (or peaking) value. LHGR is explicitly modeled as an input to 8 "Blue Ridge Environmental Defense League's Response to Duke Energy Corporation's Second Set of Interrogatories and Requests for Production of Documents," June 8, 2004.

31

the LOCA analysis, irrespective of fuel rod type (MOX or LEU). Thus, there is no difference in the representation of the LHGR value for a MOX fuel rod compared to a LEU fuel rod. As described in Section III of this testimony, for a given LHGR, the fuel initial stored energy was explicitly modeled for MOX fuel, and decay heat was conservatively assumed to be the same as LEU fuel. Thus, the impact of LHGR on fuel initial stored energy and decay heat are addressed appropriately. Furthermore, MOX fuel lead assembly core designs will ensure that the MOX assemblies are not the peak power assemblies (highest LHGR) in the core during the nominal depletion. The key point is that the LOCA limits (i.e., the LHGR inputs to the LOCA analyses that produce results within the PCT acceptance criterion) are established as operating limits to ensure that the Catawba unit will operate at acceptable LHGRs for both fuel types (MOX and LEU).

74. (Harvey, Nesbit) There is substantial margin between the actual plant operating conditions and the assumed LOCA analysis peaking (FQ) values that are used to establish operating limits. FQ is monitored monthly through core flux maps (three-dimensional measurements of the core power distribution) and compared to the LOCA analysis FQ values that define acceptable operation. Figure 7 presents the Catawba Unit 1 Cycle 14 measured FQ values and the FQ LOCA analysis values for the resident LEU fuel. (Note that these measured FQ values do not include an allowance for measurement uncertainty.) It is apparent that actual core operating conditions are far removed from the limiting LHGR values assumed in the LOCA analyses. While Figure 7 is for an all-LEU core, similar margin is expected to be present for MOX fuel.

32

Figure 7 2.60 2.40

-

- 2.20 0

,U (U 2.00 U-co 1.80 0C

= * , , , * . . *

._I- 1.60 0E

CIC14 Core Max Fq_

[L 1.40 1.20

- LOCA Fq Limit j 1.00 I I 0 100 200 300 400 500 Cycle Exposure, EFPD 33

Item 7: Magnitude offission product release duringgap releasephase; Item 8: Magnitude of volatile fission product release during early in-vessel core degradation; Item 9: Rate of volatile fission product release during early in-vessel core degradation; Item 10: Magnitude of semi-volatilefission product release during early in-vessel core degradation; Item 11: Rate of semi-volatile fission product release during early in-vessel core degradation; Item 12: Magnitude of low-volatile fission product release during early in-vessel core degradation; Item 13: Rate of low-volatile fission product release during early in-vessel core degradation; Item 14: Radionuclide inventory offuel -

75. (Harvey) The purpose of the design basis LOCA ECCS analysis is to provide reasonable assurance that the plant ECCS will prevent significant core damage in the event that a LOCA occurs. Significantly, items 7-13 are phenomena associated with hypothetical severe accidents and are not relevant to the design basis LOCA ECCS analysis, because the design basis LOCA by definition does not involve the type of core damage associated with severe accidents.

Item 14, the radionuclide inventory of the fuel rod, is not directly modeled in the design basis LOCA analysis other than as it relates to the fission gas released from the fuel pellet discussed earlier. As noted above, dose analyses (either LOCA doses or theoretical severe accident doses) are not included within the scope of Contention I.

Item 15: Radialpowerdistribution; Item 16: Axialpower distribution-

76. (Harvey) 10 C.F.R. Part 50, Appendix K, Section I.A, requires that the maximum peaking factors allowed by the Technical Specifications (at Catawba, these factors are currently 34

contained in the Core Operating Limits Report) be bounded, and that a range of power distribution shapes and peaking factors representing power distributions that may occur over the core lifetime be studied. The selected combination of power distribution shape and peaking factor should be the one that results in the most severe calculated consequences for the spectrum of postulated breaks and single failures that are analyzed. This requirement is met for the LOCA analysis for both the LEU and the MOX fuel as the results of the LOCA analysis define the allowable peaking factors ("LOCA limits") that satisfy the acceptance criteria specified in 10 C.F.R. § 50.46.

77. (Nesbit) Catawba continually monitors the axial flux difference (AFD), defined as the power in the top half of the core minus the power in the bottom half of the core, divided by total core power.

AFD = Powerop - Powerbotom Powero,.a The core design process determines the AFD that would be required for a given power level to produce the peaking factor FQ that was assumed in the LOCA analysis. Figure 8 illustrates the margin between (i) the actual measured AFD for Catawba, Unit 1, Cycles 14 and Cycle 15 and (ii) the minimum AFD that would be required to produce core peaking at the peaking limits established by the LOCA analysis. It is evident that actual core operating conditions are far removed from the limiting conditions assumed in the LOCA analysis. Note that actual plant operating limits are more restrictive than the zero LOCA margin AFD values shown on the figure. This is due to (i) incorporation of allowances for uncertainty in the plant operating limits and (ii) the fact that other parameters (e.g., steady-state departure from nucleate boiling ratio) are more restrictive than LOCA in some cases, and therefore set some portions of the plant operating 35

limits. While the data shown below reflect cores comprised of all LEU fuel, similar margin is anticipated for cores containing MOX fuel lead assemblies.

Figure 8 110 - - -- -

100-210 so _ _ _

70 -4.,7

=60 _ _ _ _ _ _

E 50 =

CIC1 Measured AFD i 40 1

CIC14 Measured AFD 30 -_-AFD at Zero LOCA Margin 20 -U-AFD at Zero LOCA Margin. -

10 _ I 0-

-60 -40 -30 -20 -10 0 10 20 30 40 60 Percent Axial Flux Difference 36

Item 17: Potentialforffuel crumbling and relocationfollowing clad ballooning, Item 18: Particlesize distributionoffuel pelletfragments as afunction ofburnup; Item 19: Characteristicsoffuel relocation (filling ratio, increase in local linear power density) -

78. (Harvey) BREDL has presented no evidence or basis for a postulated difference in the characteristics of fuel relocation between MOX and LEU fuel. This issue is discussed further in Section V of this testimony.

B. Cladding-Related "Differences"

79. (Nesbit, Dunn) It should be noted that the cladding material to be used for the MOX fuel, M5TM, is being used extensively with LEU fuel worldwide. M5TM is also being used with MOX fuel in Europe. Zircaloy-4, the traditional light water reactor (LWR) cladding material, has been used with both MOX fuel and LEU fuel worldwide.

Item 1: Extent of clad ballooning and impact on fuel relocation-

80. (Harvey) Cladding ballooning occurs when the cladding is heated to elevated temperatures with an internal pressure that exceeds the external pressure (coolant pressure). The deformation affects the cladding temperature, coolant flow (blockage), rod internal pressure, cladding surface area, and the pellet-to-cladding gap. As discussed above, all of these effects are modeled in the evaluation models used for both Zircaloy-4 cladding and M5TM cladding. The cladding ballooning models used for each cladding type are developed using cladding-specific experimental data. Both MOX and LEU LOCA analyses use unirradiated cladding properties (be it M5' m or Zircaloy-4) that maximize the extent of the cladding ballooning. Cladding ballooning effects are discussed further in Section V.C of this testimony.

37

Item 2: Fuel-cladinteraction-

81. (Harvey) This effect is not modeled in Duke's LOCA analysis. It is discussed further in Section V.E of this testimony.

Item 3: Peak clad oxidation (outer surface);

Item 4: Peak clad oxidation (innersurface) -

82. (Harvey) An exothermic reaction occurs when zirconium and water come into contact. This process occurs at normal operating temperatures, but the reaction rate is very small. The reaction rate increases as the temperature of the cladding material increases. 10 C.F.R. Part 50, Appendix K requires that this reaction rate be calculated during a design basis LOCA using the "Baker-Just" rate equation. The Baker-Just equation was chosen since it resulted in a conservative calculation, over-predicting the amount of oxidation over the temperature range of interest in LWR LOCA analyses. The data used to develop the Baker-Just rate equation are based on Zircaloy-4 cladding. Oxidation experiments have been performed for M5TM cladding, and those data show that the Baker-Just model also bounds the M5Tm data. The MOX LOCA analyses have been performed using the required Baker-Just model for oxidation of both the outside and inside cladding surface. These analyses demonstrate that the MOX fuel with M5T1 cladding meet the 10 C.F.R. § 50.46 criteria with respect to peak and total core cladding oxidation.

Item 5: Hydrogen uptake-

83. (Harvey, Dunn) Hydrogen uptake into cladding occurs during normal plant operation. The uptake of hydrogen will embrittle the fuel rod cladding - decreasing its ductility. Ductility is important in a LOCA event to ensure that the cladding remains intact after core quench. Hydrogen uptake affects both LEU fuel and MOX fuel cladding, and is observable in both M5TM and Zircaloy 4. Data for M5TM cladding have shown that the uptake of hydrogen is 38

more than a factor of five lower than that observed in Zircaloy-4. Given the lower hydrogen uptake, the LOCA performance of M5Th should be better than Zircaloy4.

Item 6: Loss of ductility (as measured by ring compression tests) as afunction of clad oxidation and surface condition,for all burnups -

84. (Harvey, Dunn) Ductility is important in the LOCA event to ensure the cladding remains intact following core quench. M5"m retains its ductility better with irradiation than Zircaloy-4. With respect to cladding integrity, the important 10 C.F.R. § 50.46 acceptance criteria are the PCT limit of 2200TF and the 17% limit on local oxidation. To ensure that this limit remains applicable to the M5Th' cladding, a series of high temperature, highly oxidized cladding tests were performed. The tests showed that the 2200TF PCT and 17% local oxidation limits specified as the acceptance criteria apply to M5'cladding as well as to Zircaloy-4.

Item 7: Reaction with fissionproduct releases (especially tellurium)-

85. (Harvey, Nesbit) One difference between M5 Tm cladding material and Zircaloy-4 is the tin content (M5Tm has none). The lower tin content in M5TM cladding has been postulated to contribute to higher tellurium releases during hypothetical severe accidents. In BREDL's response to Duke Interrogatory 33.f,9 BREDL acknowledged that this question is relevant only to source term issues that were to have been considered in Contention II, which has been dropped.

This phenomenon is not within the scope of Contention I.

Item 8: Maximum flow blockage consistent with core coolability-

86. (Harvey, Dunn) Assembly flow blockage models are developed from single-rod burst (rupture) and pre-rupture strain tests. Bundle effects are included in the flow blockage determination through the use of multi-rod tests or through computer code simulation of the 9 "Blue Ridge Environmental Defense League's Response to Duke Energy Corporation's Second Set of Interrogatories and Requests for Production of Documents," June 8, 2004.

39

bundle. For Zircaloy-4 the single rod strain data is obtained from NUREG-0630 data.' 0 M5TM cladding has different high temperature creep characteristics and different acH-iJ transformation temperatures than Zircaloy-4 cladding. These differences affect the cladding burst strain and resulting flow blockage, as discussed further in Section V.C of this testimony. For this reason, the NUREG-0630 data is not used to develop MS'flow blockage models.

87. (Harvey, Dunn) Tests were conducted at the CEA EDGAR facility in Saclay, France to determine the M5TM characteristics relating to maximum assembly flow blockage.

Using these data, AREVA developed new ballooning and flow blockage models for M5T" cladding similar to the methodology developed in NUREG-0630 for Zircaloy-4 cladding. This is documented in an AREVA topical report." The NRC Staff has reviewed the AREVA approach and, in its safety evaluation for M5'cladding, found that approach to be either as conservative or more conservative than the flow blockage model in NUREG-0630.12 The LOCA analysis for the MOX fuel uses the AREVA M5TM flow blockage model, which demonstrates that core geometry remains amenable to cooling following a design basis LOCA.

V. FUEL RELOCATION AND RELATED ISSUES

88. (Nesbit) As reframed by the Licensing Board, Contention I encompasses original BREDL Contention 10, which cited a purported failure by Duke to ". . . account for uncertainties in MOX fuel assembly behavior during Loss of Coolant Accidents." The basis for BREDL

'° D.A. Powers and R.O. Meyer, "Cladding Swelling and Rupture Models for LOCA Analysis," NUREG-0630 (April 1980).

11 Framatome ANP Topical Report, BAW-10227-A, "Evaluation of Advanced Cladding and Structural Material (M5) in PWR Reactor Fuel" (February 2000).

12 S.A. Richards (NRC) to T.A. Coleman (Framatome Cogema Fuels), "Revised Safety Evaluation for Topical Report BAW-10227P: Evaluation of Advanced Cladding and Structural Material (M5) in PWR Reactor Fuel" (February 4, 2000) (see Section 5.3.4).

40

Contention 10 rested largely on a presentation by the Institute for Radiological Protection and Nuclear Safety (IRSN) to the NRC Office of Research on October 23, 2003. IRSN is an agency in the French government that conducts research and analyses related to radioactivity. It performs functions that are somewhat similar to those of the NRC Office of Research or to those of the national laboratories in this country. It is not the agency with actual safety oversight authority for French nuclear plants.

89. (Nesbit) This section of testimony addresses issues cited by BREDL and derived from the IRSN presentation and other sources. It should be noted, however, that IRSN was not addressing the United States MOX fuel project or the Duke LAR, and that IRSN has made no representations related to the safety of the MOX fuel lead assembly program at Catawba.

Moreover, the BREDL issues are derived primarily from BREDL interpretations of statements made by IRSN concerning possible research needs related to LOCAs and severe accidents.

A. Fuel Relocation - Descriptionand Regulatory History

90. (Harvey) Fuel relocation during a LOCA is a phenomenon in which fuel pellets lose their integrity and fall (relocate) to a lower portion of the fuel rod. In order to have fuel relocation, two factors much be present. First, the pellet must be irradiated to significant burnup, such that irradiation-induced cracks have developed and the pellet can lose its integrity (break into small fragments) when the constraining influence of the cladding is removed. Second, the cladding must swell (balloon) sufficiently prior to rupture to provide room for the pellet fragments to crumble and fall.

91. (Harvey, Dunn) The potential for fuel relocation during LOCA occurs early in reflood as cladding temperatures rise to around 15000 F with fuel temperatures on the order of 1600TF. At this time the cladding will expand and rupture providing gravity and vibration the 41

opportunity to pull pellet fragments down into the swelled region. Fuel relocation has been experimentally demonstrated in the PBF-LOC tests, FR2 tests, and the FLASH5 test. Most recently relocation was observed in the swelling and rupture tests at Argonne National Laboratory. Figure 9 shows a photograph of a cross-section of PBF LOC-6 rod 12 at the burst elevation which illustrates the fuel fragmentation in the presence of cladding rupture.

Figure 9 no Cladding Epoxy 2r_

Fuel - _!

270- I g0.

Previously Burst tips irradiated Rod 12, 74% total circumferential elongation, 1066 K estimated burst temperature 42

92. (Harvey, Dunn) The concern over relocation stems from the possibility that the relocated fuel may generate too much power in a localized area, and thereby overheat the cladding. It should be noted, however, that this potential for increased heat generation is also mitigated by heat transfer enhancements that are associated with cladding ballooning. The swelling of the cladding does two things in this regard: first, it alters pellet-to-cladding gap in a manner that provides less efficient energy transport from the fuel to the cladding; and second, it provides for greater cladding surface and therefore enhanced heat transfer from the cladding. In addition, narrowing the flow channel outside the ballooned cladding increases turbulence, which further enhances heat transfer from the cladding. The FLECHT-SEASET flow blockage tests demonstrate this effect.' 3

93. (Harvey) Cladding ballooning is necessary for fuel relocation to occur. Cladding ballooning can occur only when the cladding is heated to an elevated temperature in the presence of a pressure difference across the cladding (internal rod pressure exceeding the external pressure, namely the coolant pressure). Given the LOCA response of the Catawba plant, cladding ballooning is only possible for a large break LOCA. The conservative licensing calculations for the small break LOCA events do not achieve cladding temperature in the range where cladding ballooning is possible. The maximum PCT for the small break LOCA event is 1225TF which is significantly below the temperature where cladding ballooning is calculated to occur. The importance of this result is that fuel relocation requires a very unlikely large break LOCA event.

94. (Harvey) The concern over fuel material relocation following a LOCA with fuel cladding ballooning was identified in the mid-1980's. It was recognized that the phenomenon 13 NUREG/CR-3314, "PWR FLECHT-SEASET 163-Rod Bundle Flow Blockage Task Data Report" (October 1, 1983).

43

was not included in the LOCA evaluation models and, under certain assumptions, could result in a non-conservative assumption. This was a limitation in the ECCS performance analysis methods (LOCA analysis computer codes). However, at least one study by an NRC contractor organization concluded that known conservative features in the Appendix K models more than offset the effect. The NRC classified this item as Generic Issue (GI) 92 ("A Prioritization of Generic Safety Issues", NUREG-0933) and initially assigned this item a low priority based on the value/impact assessment. However, subsequent prioritization placed this issue in the "drop" category.

95. (Harvey) A combined meeting of ACRS subcommittees discussed the postulated relocation effect in a meeting on November 15, 2001. The purpose of this meeting was to discuss relaxing the Appendix K decay heat assumption. In the meeting the NRC identified the fuel relocation issue as a non-conservative assumption in the approved Appendix K methods that may need to be addressed if the Appendix K decay heat assumption is relaxed. There was no indication that this issue was something that needed to be addressed for currently approved Appendix K methods. In the meeting it was recognized that European research/analysis is being conducted in this area.

96. (Harvey) The NRC Staff is evaluating at least the possibility of changes to 10 C.F.R. § 50.46 and Appendix K based on research that has demonstrated the existence of the large safety margin (described in Section III.D of this testimony) between the regulatory acceptance limits and the expected plant behavior during a LOCA.14 Staff members have noted that as known conservatism is removed from Appendix K, sources of non-conservatisms also need to be considered. One of a number of potential non-conservatisms identified is the fuel 14 Thadani, A.C. (NRC) to Collins, S. (NRC), "Research Information Letter 0202, Revision of 10 C.F.R. § 50.46 and Appendix K" (June 20, 2002) ("Thadani memorandum").

44

relocation phenomenon. However, our MOX evaluations continue to use the current deterministic acceptance criteria, with the built-in margins discussed above, so the Thadani memorandum is not directly relevant.

97. (Harvey, Dunn) ECCS performance analyses that do not model fuel relocation show a reduction in the rate of cladding heat up at the ruptured location and possibly a reduction in the cladding temperature following fuel swelling (ballooning) and rupture. The response was also seen in FR2 test B3.115 and is illustrated in Figure 10. The cladding temperature at the location of the rupture shows a significant drop following cladding rupture. This test was performed using an unirradiated fuel rod and therefore no pellet relocation occurred.

15 LWR Fuel Rod Behavior in the FR2 In-pile Tests Simulating the Heat-up Phase of a LOCA, KfK 3346 (March 1983).

45

Figure 10 FR2 Test B3.1 for Low Exposure Fuel (No Relocation) 2000.00 - -

1800.00 - _

1600.00 - -

1400.00 -

U.

1200.00 -

1000.00 -

E K

E- 800.00 -

600.00 - _

400.00 - -

200.00 -

0.00 -

0.00 50.00 100.00 150.00 200.00 250.00 Time, s

98. (Harvey, Dunn) Fuel pellet relocation can increase the heat source in the location of the ballooned cladding. However, this effect is mitigated by the fact that the fuel cannot crumble, fall, and pack itself into a mass with anything close to the original pellet density (typically around 95% theoretical density). An assumption of a very high packing factor (or filling ratio) within the ballooned section is required to develop any concern about the amount of heat generation following relocation. FR2 test E-4 was conducted on fuel of sufficient burnup that relocation occurred at the time of cladding rupture. As shown in Figure 11 the ruptured 46

node temperature decreases slightly as swelling and rupture develop and then follows the excursion of the non-ballooned areas of the fuel pin.

Figure 11 FR2 Test E4 for High Exposure Fuel (Relocation) 2000.00 1800.00 - A( on Fuel Pin 1600.00 1400.00 U_1200.00 tured, aAone a lLocation on Fuel Pin E 1000.00 / _

C) 0.

E 800.00 600.00 400.00 _ _ _ _ _ _ _ _ _

200.00 -

0.00 0.00 50.00 100.00 150.00 200.00 250.00 Time, s B. VERCORS Tests

99. (Nesbit) BREDL used an October 23, 2003 presentation by IRSN as the basis for its concerns related to MOX fuel relocation during a LOCA.16 In particular, BREDL referred to data from the VERCORS tests as supporting the BREDL contention that fuel relocation would occur at a lower temperature in MOX fuel than in LEU fuel under LOCA conditions. This 16 A. Mailliat and J.C. Melis (IRSN), "IRSN Source Term LOCA Program in the PHEBUS Facility," Presentation to the NRC (October 23, 2003).

47

section of the testimony summarizes the VERCORS tests and explains why the data from those tests are not applicable to this contention.

100. (Nesbit) VERCORS is a series of tests conducted by IRSN and its predecessor organization on fuel undergoing thermodynamic conditions that are consistent with severe accidents. The tests involve a single fuel rod section (three irradiated pellets) in a test rig. The test apparatus controls the temperature of the fuel and provides for variable steam and hydrogen flow past the fuel rod. The tests were carried out at Cadarache in France, beginning in 1989 and concluding recently.

101. (Nesbit) The tests investigated a number of severe accident phenomena, including fuel relocation and radionuclide release. In the context of severe accidents, fuel relocation refers to the extremely high temperature phenomenon of fuel pellets losing their integrity and collapsing downward. Severe accident fuel relocation is thermally-induced; it involves liquefaction and slumping of the fuel material following an extended loss of core cooling. In the case of the VERCORS tests with MOX fuel, the relocation temperatures were in the neighborhood of 4000'F or more.

102. (Nesbit) The VERCORS RT2 and RT7 tests involved irradiated MOX fuel. The thermal conditions of the VERCORS RT2 and RT7 tests were much higher than fuel temperatures experienced during a design basis LOCA. A design basis LOCA would include restoration of core cooling within a few minutes of accident initiation, based on activation of the Emergency Core Cooling System (ECCS). In a design basis LOCA, PCT is maintained below 2200TF in accordance with 10 C.F.R.§ 50.46(b). Figure 12 is a stylized representation of LOCA temperatures along with VERCORS relocation temperatures.

48

Figure 12 MOX Fuel Fuel and Cladding Temperatures During LOCA Compared to Severe Accident Fuel Melt Temperatures I

aI I

I-0 100 200 300 400 500 600 700 800 90o 1000 Time (smads) 103. (Nesbit) Fuel relocation during a design basis LOCA is a distinct and different phenomenon from fuel relocation during a severe accident. Fuel relocation during LOCA is a mechanically-induced phenomenon which follows cladding ballooning and fragmentation of irradiated fuel pellets. Fuel relocation during LOCA occurs at much lower temperatures, in the neighborhood of 1500-2200'F. Temperatures are far removed from liquefaction conditions experienced in the VERCORS tests.

104. (Nesbit) To our knowledge, and based upon the publicly available materials, IRSN has made no observations about the implications of the VERCORS tests on LOCA analyses or the ability to cool MOX fuel during a LOCA.

49

105. (Nesbit) To our knowledge, and based upon the publicly available materials, IRSN has made no comments on the safety of the proposed United States MOX fuel lead assembly program or the associated LAR.

106. (Nesbit) To our knowledge, and based upon the publicly available materials, IRSN has made no effort to restrict the use of MOX fuel in French reactors due to VERCORS test results. Twenty French reactors are currently using substantial quantities of MOX fuel (generally about 30% MOX fuel in a 157 fuel assembly core).

107. (Nesbit) Based on the information from the VERCORS tests of which we are aware, the results of those tests have no implications with respect to core cooling of MOX fuel during a design basis LOCA at Catawba. LOCA performance is driven primarily by the thermal-hydraulic response of the reactor coolant system, not by the fuel pellet material.

108. (Nesbit) On April 26, 2004, Duke posed Interrogatory 45 to BREDL. Duke asked the following questions.

Does BREDL agree that the VERCORS tests and results are irrelevant to the issue of MOX fuel compliance with 10 C.F.R. § 50.46 emergency core cooling system requirements? Does BREDL agree that to the extent the VERCORS tests are relevant, it is in relation to dose analyses only? If not, describe the basis for BREDL's position.

BREDL responded "BREDL agrees that the VERCORS tests are relevant to dose analyses only when design-basis LOCAs are considered."17 BREDL therefore does not dispute Duke's conclusion that the VERCORS tests are not relevant to the demonstration of compliance with 10 C.F.R. § 50.46 emergency core cooling system requirements, which is the issue of Contention I.

17 Blue Ridge Environmental Defense League's Response to Duke Energy Corporation's Second Set of Interrogatories and Requests for Production of Documents, June 8, 2004.

50

C. M5 Tm Ballooning 109. (Dunn) The impact of fuel relocation could depend in part on the amount of strain experienced at the location of cladding swell and rupture. As explained earlier, cladding swelling just prior to rupture forms a local cavity within the cladding that can fill with fuel fragments from cracked pellets. The amount of the swelling (cladding strain) is directly proportional to the consequence of the relocation because more swelling provides more volume to be filled by fuel pellet fragments.

110. (Dunn) The amount of swelling experienced is dependent on the phase of the cladding crystalline structure at the time of rupture. As the cladding heats the structure changes from a close packed hexagonal form, termed the alpha phase, to a body centered cubic form, called the beta phase. When the cladding is in either form, cladding strain at rupture is maximized. However, during the transition from alpha phase to beta phase, the differing crystalline structures tend to retard the ability of the metal to deform, so strain at rupture is minimized. LOCA evaluations treat this phenomena with direct correlations of experimental results. Figures 13 and 14 illustrate the required information. Figure 13 correlates the rupture temperature as a function of cladding stress. A determination of rupture is made by computing the cladding stress and comparing the current cladding temperature to the temperature of rupture given by Figure 13. If the cladding temperature equals the Figure 13 temperature the cladding will rupture. The strain is then obtained from the correlation with cladding rupture temperature (Figure 14).

51

Figure 13 Figure M5 Rupture Temperature Correlation

- - - - - - - - - - - - - - - - - - - - --- - - - - - - - - - - - - - - - - - - - - - - - - - - - -- - - - - - - -- - - - - - - -- - - - - - - - - - - -- - - - - - - -- - - - - - - - - -- - - - - - - - -- - - - - - -- -

-- - - - - - - -- - - - - - - - -- - - - - - -- -- - - - - - - -- - - - - - - -- - - - - - - - - -- - - - - - -- - - - - - - - -- - - - - - - - - -- - - - - - - - - -- - - - - - -- - - -

U

-- - - - - - - - - - - - - - - - - - - - - - - - - - -- - - - - - - - -- - - - - - -- - - - - - - -- -- - - - - - -- -- - - - - - - - - -- - - - - - - - - -- - - - - - - - - -- - - - - - - - - - - - - - - - - -- - - - - -- - - -

E L

E-A a.

-- - - - - ---- - - - - - - - - - - - --- - - - - - - - - - - - - ----- - - - - - - - - - - - - - - - - - - - - --- - - - - - - - - - - - - - - - - - - - - - - - - - - - -- - - -

------------------------------------------ -------- ---------

-- -- - - - - - - - - --- - -- - -- -- - - - - -- ---- --- ------- - - - - - - - -- - - - - - - - -- -- - - - - - -- - - - - - - - - -- - - - - - - - - -- - - - - - - - - -- - - - - - - - -- - - - - - - - -- - - - - - - -

Engineering Hoop Stress, KpsI 52

Figure 14 Figure MS Rupture Strain versus Rupture Temperature Slow Ramp Rate

.;

/ \ / \,

..fv 4/ S\

. _

Rupture Temperature, C 111. (Dunn) As can be observed in Figure 14, the amount of strain varies substantially with the rupture temperature. Typical rupture temperatures for the Catawba plant lie between 1500-1600'F, placing the rupture during the transition of the cladding from the alpha phase to the beta phase. Typical strains lie between 50-75%.

112. (Dunn) The strains used in the LOCA calculations are based on tests on unirradiated cladding, because this maximizes the strain and therefore maximizes the flow blockage. The maximized strain would also maximize the consequences of fuel relocation.

113. (Dunn) The impact of irradiation on cladding ductility at high temperatures is primarily due to the pickup of hydrogen during normal operation. Hydrogen tends to embrittle the cladding, thereby reducing its ability to strain prior to rupture. Because MWTM picks up 53

relatively little hydrogen during operation, a M5T strain versus rupture temperature curve would show little change with irradiation. In contrast, Zircaloy-4 cladding picks up substantial hydrogen during operation. Accordingly, a Zircaloy-4 cladding strain versus rupture temperature curve would change more with operation (burnup) than would a M5TL curve. In this comparison, M5TM would experience less strain at rupture than Zircaloy4 in the unirradiated state, but the two materials have approximately equal strain potential near the end of irradiation.

114. (Dunn) Overall, therefore, there is little expected difference in the consequences of fuel relocation due to cladding differences.

D. FillingRatio of Relocated Fuel 115. (McCoy) Another factor in assessing the impact of postulated fuel relocation is the appropriate density of the material that might fall (i.e., "relocate") into the ballooned region of the cladding. This material would be made up of the fragments of pellets from the immediate vicinity of the ballooning. During LOCA, when the cladding balloons away from the fuel pellets, the pellets could break into fragments along the existing crack lines and fall into the balloon. The relevant question is whether this would be significantly different for MOX fuel relative to LEU fuel.

116. (McCoy) The potential for fragmentation can best be determined by reviewing available micrographs of irradiated LEU and MOX pellets. Figure 15 shows a LEU fuel rod cross-section following irradiation and examination. The pattern of cracks in the pellet is typical of irradiated fuel.

54

Figure 15 Micrograph of Irradiated LEU Pellet with Typical Cracking 117. (McCoy, Dunn) Fuel pellets crack during the first ascent to power. The central portion of the pellet is hotter than the periphery and therefore expands more as the pellet comes up to its operational temperature distribution. The result is a set of roughly radial cracks in the outer portions of the pellet. The orientation of the cracks is more variable near the center of the pellet and central "islands" may form, completely surrounded by cracks. The potential for fuel relocation during LOCA develops with burnup as gaseous fission products accumulate within the pellet on the grain boundary creating stress in the pellet leading to fragmentation along crack lines when external constraints are removed (i.e., when there is cladding ballooning). Figure 16 provides another pellet cross-section illustrating a differing crack pattern. In this case a crack has developed radially that passes through the central region of the pellet. Beyond the general pattern, the positioning and path of the cracks is stochastic.

55

Figure 16 Micrograph of Irradiated LEU Fuel Pellet with Transverse Crack 118. (McCoy, Dunn) The processes that cause cracking and the cracking pattern -

thermal stress being resolved along grain boundaries - are the same for a MOX pellet as for a LEU pellet. The crack locations are not influenced by the plutonium rich agglomerates unique to the MOX fuel. Thus, within the general cracking pattern of LEU pellets described above, no differentiation can be made between LEU and MOX pellets, and fragment formation during LOCA would be expected to be the same for both fuel types. Figure 17 is a typical micrograph of a MOX pellet. The dark black areas or spots are plutonium rich agglomerates. There is no evidence in the micrograph that the crack locations are influenced by the agglomerates. The general cracking pattern is the same as for LEU pellets as shown in Figures 15 and 16. Thus, no differentiation in pellet fragmentation should be made between LEU and MOX fuel.

56

Figure 17 Micrograph of Irradiated MOX Pellet (RG) 119. (Nesbit) In this regard, this issue was raised at an ACRS meeting on the MOX fuel LAR. Dr. Dana Powers of the ACRS reacted to the "radial cut of a high burnup MOX pellet" shown by Patrick Blanpain of Framatome ANP, as follows:

CHAIRMAN POWERS: This hits to a point that Dr. Lyman raised, a question that he raised in his presentation. I think it is fair to characterize it as a question. Was there any inherent difference in the fragmentation of this MOX fuel relative to what we have experienced in uranium dioxide 57

fuel a(s) we go up to these burnups? I'll have to admit, had you not told me this was MOX fuel, I probably would not have known otherwise.' 8 120. (McCoy, Dunn) The density of the fragmented fuel material after falling into the cladding balloon will depend on the shapes of the fragments, the size of the fragments, and the efficiency with which the fragments settle into the ballooned region of the cladding. IRSN presented a distribution of densities based on the limited experimental evidence available at a May 25-26, 2004 meeting' 9 in Argonne, Illinois. The sources were the PBF and FR2 experiments, which used LEU fuel. The results are reported as filling ratio, which is the mix density of relocated material divided by the original pellet density. For the cases reported the ratio falls between 0.55 and 0.8. However, the upper portion of this range may be discounted.

The most reliable relocation data is from the FR-2 tests, and those values lie in the 0.55 to 0.65 range.

121. (McCoy) In the end, in my opinion, the breakup of either LEU or MOX fuel during LOCA would occur by the same process and should produce nearly identical fragmentation. The available measures of relative fuel material density within the relocation area indicate that a filling ratio less than 0.7 would be expected for LOCA conditions. There would be, in this aspect, no significant difference between MOX fuel and LEU fuel.

18 Advisory Committee on Reactor Safeguards Subcommittee on Reactor Fuels, April 21, 2004, Tr. at 69.

19 Grandjean, Claude and Hache, Georges (IRSN), "LOCA Issues Related to Ballooning, Fuel Relocation, Flow Blockage and Coolability," Presentation to the Meeting of the Special Experts Group on Fuel Safety Margins, Organization for Economic Co-operation and Development's Committee on the Safety of Nuclear Installations, Argonne, Illinois, May 24-26, 2004.

58

E. Fuel-CladdingInteraction 122. (McCoy) Bonding of the fuel to the cladding is a complex process that occurs during operation of the fuel. The first requirement for bonding is intimate contact between the fuel pellets and the cladding. Fresh fuel at room temperature has a substantial diametral gap between the pellets and cladding, nominally about 150 Aum. The gap size changes during irradiation as a result of irradiation-induced and pressure-induced creep of the cladding, and fracturing, densification, and swelling of the pellets. The general trend is for the gap to shrink until the pellets and cladding are in contact. The chemistry inside the fuel rod also changes during irradiation, with fission products being produced and the oxygen potential increasing.

When the gap is sufficiently small, the interior surface of the cladding can oxidize by athermal diffusion. In this process, oxygen is dislodged from fission spikes near the surface of the pellet and implanted in the cladding. Finally, cesium, which is produced as a fission product, can react with the fuel pellet and the zirconium dioxide layer of the cladding to form a variety of cesium uranates, zirconates, and plutonates. It is these phases that bond the pellet to the cladding.

123. (McCoy) Despite the complexity of bond formation, the processes that lead to bonding are the same for LEU and MOX fuels, and the bond strengths are expected to be similar because of the similarities in fuel chemistry and operating conditions. Both fuels are predominantly uranium dioxide, and they have similar yields of cesium as a fission product.

Therefore, it is entirely reasonable to expect that pellet-cladding bonding will have similar effects on LOCA performance in MOX and LEU fuels (if indeed there is any pellet-cladding bonding effect on LOCA at all).

124. (McCoy, Dunn) BREDL has speculated that the extent of fuel-cladding bonding may be important to the fuel relocation phenomenon. It has also been speculated, however, that bonding is beneficial in controlling relocation. Although there is insufficient information to 59

quantify the benefits of bonding, we list potential benefits below. Nonetheless, AREVA has taken the conservative approach by taking no credit for the possible benefits in the MOX fuel lead assembly LOCA analysis.

125. (McCoy, Dunn) One possible benefit of bonding is that it may provide some additional strength to the cladding and therefore may reduce the extent of ballooning. Reducing the extent of ballooning might not prevent relocation, but it would reduce the volume into which fuel fragments could relocate and thus lower the local decay heat at that location. At present, there is insufficient data to confirm this effect.

126. (McCoy, Dunn) A second possible benefit is that, under some circumstances, the bond and the pellet structure near the pellet surface may be strong enough to allow a pellet fragment to adhere to the cladding interior surface during swelling and rupture. If this occurs, the bonded fragment may obstruct the motion of the other pellet fragments, slowing the filling of the balloon during fuel relocation, and possibly even reducing the final amount of relocated fuel.

This effect also has not been observed in experiments to date.

127. (McCoy, Dunn) Fuel bonding might also provide a coating that would remain with the cladding after thermal expansion causes the cladding to separate from the pellet during the first seconds of a LOCA. This is not specifically a fuel relocation effect, but if present any coating effect would be seen following cladding rupture. Such a coating effect might partially protect the cladding interior from high temperature oxidation. While there is no direct experimental confirmation of the phenomena, if present it would provide additional mitigation against local oxidation approaching the 17 percent acceptance criterion.

128. (McCoy, Dunn) At bottom, there is no evidence that pellet-cladding bonding differs significantly between MOX and LEU fuel. Even if one were to assume that it did, the 60

discussion above makes it clear that pellet-cladding bonding might act to improve LOCA performance by reducing the extent of cladding ballooning. Furthermore, the most recent studies of fuel relocation made no provisions to include postulated effects of pellet-cladding bonding.

These reports are included with this testimony as Exhibits 3, 4, and 5.20 Therefore, hypothetical MOX/LEU differences in this area do not affect the current regulatory position that relocation effects, if present at all, are adequately bounded by the inherently conservative nature of Appendix K LOCA analyses.

F. MOX Fuel Relative Power at High Burnup 129. (Nesbit) BREDL hypothesizes that relocation in the MOX fuel lead assemblies may be a more significant impact than relocation in LEU fuel because the power in MOX fuel is higher at end-of-life than the power in LEU fuel. This theory appears to be based on no more than a statement by IRSN that "... this question is particularly important for end-of-life MOX fuel where power generation is not reduced, unlike for U02 fuel." 2 '

130. (Nesbit) Higher power at end-of-life is adverse for fuel relocation because fuel that is operating at higher power prior to the postulated accident gives off more power (decay heat) during the accident. The relocated fuel fragments would have a higher power, contributing to higher PCTs.

20 M. Lambert, et al. "Synthesis of an EDF and Framatome ANP Analysis on Fuel Relocation Impact in Large Break LOCA", proceedings of the Topical Meeting on LOCA Fuel Safety Criteria, Aix-en-Provence (March 2001) (Exhibit 3); C. Grandjean, et al. "High Burnup U0 2 Fuel LOCA Calculations to Evaluate the Possible Impact of Fuel Relocation after Burst", Proceedings of the Topical Meeting on LOCA Fuel Safety Criteria, Aix-en-Provence March (2001) (Exhibit 4); and V. Guillard, et al., "Use of CATHARE2 Reactor Calculations to Anticipate Research Needs", presentation at the SEGFSM Topical Meeting on LOCA Issues, Argonne National Laboratory (May 2004)

(Exhibit 5).

21 "IRSN Source Term LOCA PROGRAM IN THE PHEBUS FACILITY," IRSN presentation to the NRC (October 23, 2003), Slide 21.

61

131. (Nesbit) However, the IRSN statement about higher power at end-of-life is apparently based on European MOX fuel experience. For the reasons discussed below, the Catawba MOX fuel lead assemblies will operate at lower power at end-of-life than LEU fuel.

This will provide a MOX fuel benefit (not a penalty) with respect to fuel relocation (relative to the LEU fuel at Catawba).

132. (Nesbit) First, we consider the type of plutonium in the fuel. European reactors have been operating for decades with "reactor grade" (RG) MOX fuel. The fuel is referred to as "reactor grade" because it is initially comprised of RG plutonium blended with depleted (or natural) uranium. As noted in Section II.B of this testimony, RG MOX fuel contains plutonium with more than 20% 240 Pu. In contrast, the Catawba MOX fuel lead assemblies will initially contain "weapons grade" (WG) plutonium, which is less than 7% 240 Pu. The initial isotopic content of the plutonium (RG vs. WG) affects the nuclear characteristics of the MOX fuel, including the profile of power vs. burnup, as discussed in the paper "Basis for the Design of Reactor Cores Containing Weapons Grade MOX Fuel." 22 133. (Nesbit) WG MOX fuel is more similar neutronically to LEU fuel than is RG MOX fuel, as illustrated below. Figure 18 shows k-infinity (kit) for three mechanically identical fuel assemblies as a function of burnup, as calculated by the CASMO-4 computer code for an infinite lattice configuration. K. correlates to the ability of the fuel assembly to generate power.

As the fissile material depletes with burmup, k. decreases.

22 Steve Nesbit and Jim Eller, "Basis for the Design of Reactor Cores Containing Weapons Grade MOX Fuel," Advances in Nuclear Fuel Management III, American Nuclear Society, October 5-8, 2003.

62

Figure 18 A LEU 120 ___

-X-RG MOX 1.15 -- AdzMO 1.10 1.05 __ _ __ _

1.00 _ _ _ _ _ _ _

0.95 0.90 = = =__'

0.85 0.80 0 10 20 30 40 50 60 Fuel Exposure, GWD/t 134. (Nesbit) At about 22 GWD/t burnup, the curves intersect; at that point the three fuel assembly types would have about the same relative power if loaded in the same core location.

135. (Nesbit) The slope of the RG MOX k. vs. burnup curve is flatter than the curve for LEU fuel. This means that, all other things being equal, at burnups higher than 22 GWD/t, the RG MOX fuel would have a higher relative power than LEU fuel.

136. (Nesbit) The WG MOX fuel k. vs. burnup curve is much closer to the LEU fuel ki vs. burnup curve. Therefore, at higher burnups (e.g., 40-50 GWD/t), the WG MOX fuel relative power would be much closer to LEU fuel. In other words, the effect noted by IRSN (a 63

tendency for higher end-of-life power in MOX fuel) is much less important for Catawba, which will use WG MOX fuel lead assemblies, than it is for French reactors that use RG MOX fuel.

137. (Nesbit) Second, we consider the nature of the lead assembly program itself. The purpose of lead test assemblies is to gather additional data on new fuel assembly designs that may not have extensive prototypical operating experience. Accordingly, lead test assemblies are placed in "nonlimiting core regions." 2 3 The Catawba cores containing MOX fuel lead assemblies will be designed to ensure that the MOX fuel assembly power during the nominal depletion is not the highest assembly power in the core.

138. (Nesbit) Duke has developed a fuel cycle design for CIC16 containing four fresh MOX fuel assemblies and 72 fresh LEU assemblies. Those MOX and LEU assemblies have been carried forward into preliminary scoping analyses for Cl C17 and C I C 18. The results from those analyses were used to calculate anticipated fuel assembly peaking for each of the 76 assemblies loaded into C1C16, as shown in Figure 19.

23 Catawba Nuclear Station Technical Specifications, Section 4.2. 1.

64

Figure 19 1.6

`1A IA AAAA AL 1 1.2 ID 0 0 0 ~ AA IL

1.0

MOX LTA It 0.8 A Maximum LEU

.D 0.6 O Average LEU I I A *A A A A A A O

_ CA 0.4 0.2 mu * * * - * * *

W

_ 0.0 0 200 400 600 800 1000 1200 1400 1600 Accumulated Cycle Irradiation In Effective Full Power Days 139. (Nesbit) As shown in Figure 19, the MOX fuel lead assembly relative power (squares) is lower than the corresponding average LEU assembly relative power (circles) at all times except at the very beginning of each of the first two cycles.

140. (Nesbit) The MOX fuel lead assembly relative power (squares) is lower than the maximum LEU assembly relative power (triangles) at all times.

141. (Nesbit) Third, we consider the fact that the decay heat from a MOX fuel assembly will be lower than the decay heat from a LEU fuel assembly with the same burnup and same operating power level. Figure 3 appearing earlier in this testimony shows that the MOX fuel decay heat should be 3-5% lower than the corresponding LEU fuel decay heat during the time period of interest for a LOCA. It appears that the IRSN did not acknowledge the decay heat effect, which is a benefit for MOX fuel.

142. (Nesbit) In summary, the IRSN statement about higher power in MOX fuel at end-of-life is not applicable to the Catawba MOX fuel lead assemblies. LEU fuel assembly operating power will be higher than MOX fuel lead assembly operating power at end-of-life at 65

Catawba, due both to the nature of WG MOX fuel and to core design. Power in relocated fuel fragments is also affected by decay heat. LEU fuel decay heat will be higher than MOX fuel lead assemblies decay heat at the same power level and burnup. To the extent that assembly relative power and decay heat might affect LOCA performance through fuel relocation, that effect will be beneficial for MOX fuel relative to LEU fuel. In other words, these MOX/LEU differences would act to mitigate, not exacerbate, the impact of MOX fuel relocation during a postulated LOCA.

G. Conclusions on Fuel Relocation 143. (Dunn) As discussed above, the issue of fuel relocation during LOCA was identified and evaluated by the NRC in the mid-1980s. The potential for relocation affects all nuclear plants using fuel within a sealed cylindrical cladding - that is, the majority of nuclear reactors world-wide. The issue was reviewed by the NRC, assigned a low priority, and subsequently given a "drop" designation. The NRC's review cited the conservatisms inherent in a 10 C.F.R. § 50.46 evaluation as more than bounding the potential impact of relocation.

144. (Dunn) As also discussed above, in the Thadani memorandum the NRC Staff has recognized that any removal of the substantial existing embedded conservatisms (or margin) in the regulations that might be associated with risk-informing the regulations should also be accompanied by either (i) a quantitative assessment that the retained conservatisms remain sufficient to bound fuel relocation or (ii) an explicit relocation model included in the modeling.

In this respect fuel relocation is just one of several potential conservatisms and non-conservatisms. Aside from the concern over removal of current conservatisms, the NRC's position on fuel relocation remains that the issue is well-bounded and does not require explicit modeling under the current, deterministic regulatory approach (such as was utilized for the MOX fuel lead assemblies).

66

145. (Dunn, McCoy) The industry has generally held that the impact of fuel relocation on calculated cladding temperatures or cladding oxidation predictions would be small or inconsequential if incorporated within comprehensive LOCA modeling. We do not expect fuel relocation effects to be significantly different for MOX versus LEU fuel.

146. (Dunn) At the meeting of the Organization for Economic Co-operation and Development's Committee on the Safety of Nuclear Installations in Aix, France in 2001, two evaluations were presented. One, by Electricite de France (EDF) and Framatome ANP, produced only a limited increase in cladding temperature (30'C or 540 F), resulting in the presenter's conclusion that relocation should not be taken into account in safety regulation 24 .

The other, from IRSN and using more conservative assumptions, indicated a 130'C (2340 F) potential impact on cladding temperatures.

147. (Dunn) Neither calculation took credit for increased heat transfer due to the rupture-induced geometric perturbation of the flow field. Even a partial accounting for these effects indicates a potential reduction of cladding temperatures of 100-1500 C (180-270'F).

Thus, the presentations at Aix indicate that a comprehensive evaluation of the effect of fuel relocation would either reduce the cladding temperature beyond that calculated by current models or make essentially no difference in the predicted temperature.

148. (Dunn) Experimentally, the most comprehensive treatment of the phenomena are the KfK tests performed at the FR2 reactor in Germany. Here testing was done for fuel that 24 M. Lambert, et al., "Synthesis of an EDF and Framatome ANP Analysis of Fuel Relocation Impact in Large Break LOCA," Proceedings of the Topical Meeting on LOCA Fuel Safety Criteria, Special Experts Group on Fuel Safety Margins, Organization for Economic Co-operation and Development's Committee on the Safety of Nuclear Installations, Aix, France, March 22-23, 2001. As noted above, this paper is Exhibit 3 to this testimony.

67

cracked and relocated and for fuel that did not relocate, providing a unique opportunity for comparison with limited cross test complications. Figures 10 and 11 appearing earlier in this testimony provide the results for two such tests.

149. (Dunn) When the fuel does not relocate, Figure 10, the cladding temperature near the ballooned and ruptured location cools substantially followed by a delayed heatup and a peak approximately 120'C (216'F) lower than that of cladding above and below the rupture location.

This cooling is caused by the combined effects of the increase in cladding surface area, the enhancement of heat transfer due to geometry effects, and the decrease in cladding-to-pellet gap conductance.

150. (Dunn) For the test with fuel relocation, Figure 11, there is only a slight reduction in temperature at the time of rupture followed by a heat up that is very similar to the other locations on the fuel pin. This indicates that the benefits of increased surface area and increased heat transfer from geometric effects are a near match for the detrimental effect of the increased decay heat and lack of expanded gap associated with the fuel repositioning.

151. (Dunn) Thus, the most comprehensive experimental evidence available for the evaluation of fuel relocation indicates that the phenomenon is not of substantial consequence with respect to PCT.

152. (Dunn) Finally, the most pessimistic of the relocation impacts calculated by IRSN

- can be added to the results of the AREVA LOCA analyses that were used to establish LOCA limits for the MOX fuel lead assemblies (Exhibit 2, Table Q14-1), and the results still do not exceed the LOCA acceptance criteria of 10 C.F.R. § 50.46.

153. (Dunn) At the conference in Aix in 2001, the IRSN presented a calculation of fuel relocation showing an increase in cladding temperature at the swelling location (fuel relocation 68

site) of slightly less than 320TF. The increase in local oxidation was 7% at the ruptured location.

The IRSN calculation was for a filling fraction of 0.7.

154. (Dunn) The highest PCT at the ruptured location in the LOCA calculations for Catawba described in the MOX fuel lead assembly license amendment request was approximately 1750TF and the local oxidation on that fuel pin is 3%. Adding the IRSN predictions to the Catawba MOX fuel results gives an estimated PCT of 2070TF and a local oxidation of 10%. Thus, even if the pessimistic IRSN predictions are simply added to the current Catawba MOX fuel LOCA evaluations, the results remain well below the acceptance criteria of 10 C.F.R. § 50.46.

155. (Dunn) At a conference in Argonne in May 2004, IRSN presented calculations of a fuel relocation PCT impact as high as 150'C (270TF) for LEU fuel, and a further fuel relocation PCT impact of 10C (18TF) for MOX fuel. 25 A 180 F difference between MOX and LEU fuel is negligible with respect to calculated LOCA PCTs, and supports the position that fuel relocation is not primarily a MOX fuel issue.

156. (Dunn) The IRSN presentation at Argonne notes a total PCT increase of 160'C (288TF) for MOX fuel during relocation, and traces the 100 C MOX fuel increment to higher initial stored energy in MOX fuel. As shown earlier, the MOX fuel lead assemblies at Catawba will have lower initial stored energy than LEU fuel due to operation at lower peaking and to lower decay heat. Accordingly, the MOX fuel lead assemblies at Catawba would actually see a benefit, not a penalty, relative to LEU fuel and potential fuel relocation impacts.

25 Guillard, et al., "Use of CATHARE2 Reactor Calculations to Anticipate Research Needs," Presentation to the Meeting of the Special Experts Group on Fuel Safety Margins, Organization for Economic Co-operation and Development's Committee on the Safety of Nuclear Installations, Argonne, Illinois, May 24-26, 2004. As noted above, this paper is included with this testimony as Exhibit 5.

69

-

157. (Dunn) Summing up, the impact of fuel relocation is generally held to be inconsequential. This view is supported by the most comprehensive experimental results available, the KfK tests in the FR2 reactor. Moreover, MOX/LEU differences at end-of-life would act to mitigate any adverse fuel relocation impact in the MOX fuel lead assemblies. Even so, superimposing IRSN's conservative evaluation of the impact of fuel relocation on top of the conservative Catawba MOX fuel LOCA predictions would result in consequence predictions within those allowable under 10 C.F.R. § 50.46. Any uncertainty associated with fuel relocation in MOX fuel is sufficiently bounded by the existing analyses.

VI. UNCERTAINTIES 158. (Nesbit) Contention I reduces to a question of what constitutes acceptable "uncertainty." Uncertainty is inherent in LOCA analyses. The computer calculations used to simulate a LOCA involve rapidly changing thermal-hydraulic conditions coupled with neutronic and mechanical phenomena. Recognizing this inherent uncertainty, the NRC established the conservative LOCA acceptance criteria in 10 C.F.R. § 50.46 and prescribed conservative LOCA modeling approaches in Appendix K to 10 C.F.R. Part 50.

159. (Nesbit) Duke used these conservative modeling approaches to demonstrate that the conservative acceptance criteria are met for the four MOX fuel lead assemblies following a postulated large break LOCA at Catawba.

160. (Nesbit) 10 C.F.R. § 50.46 and Appendix K were promulgated in 1974. In the ensuing years, NRC and others have developed more and more sophisticated analytical tools for analyzing design basis LOCAs. In addition, NRC and others have invested substantial resources in performing separate effects tests and integral tests for the purpose of improving and validating LOCA computer models. As discussed earlier in this testimony, one of the results of this work has been to confirm the fundamental overall conservatism inherent in both the 10 C.F.R. § 50.46 70

acceptance criteria and the Appendix K models such as those used to analyze the four MOX fuel lead assemblies.

161. (Nesbit) The October 23, 2003 IRSN presentation provided no evidence that the MOX fuel lead assembly LOCA analyses do not provide reasonable assurance that the LOCA acceptance criteria are met. In fact, the IRSN presentation addressed neither the Catawba MOX fuel lead assembly LAR nor the broader program to use MOX fuel to dispose of weapons grade plutonium.

162. (Nesbit) Based on remarks by IRSN, BREDL is speculating that fuel relocation may be different in MOX fuel than LEU fuel, and that such a difference, if present, may cause

_ fuel relocation in MOX fuel to have an adverse impact on a design basis LOCA. Again based on remarks by IRSN, BREDL claims the M5 cladding used on the MOX fuel lead assemblies is likely to have higher blockage ratios during a LOCA, resulting in coolant blockage that could lead to an unacceptable loss of core geometry and an uncontrolled core melt. BREDL states that these effects cannot be fully assessed in the absence of integral LOCA MOX fuel-bundle tests, such as those proposed by IRSN. BREDL claims that this alleged uncertainty is sufficient cause for denying the Duke application to use four MOX fuel lead assemblies until the testing proposed by IRSN is carried out.

163. (Nesbit) As discussed in Sections IV and V of this testimony, BREDL's concern

- that fuel relocation and M5Tm cladding ballooning during design basis LOCA are especially significant uncertainties for MOX fuel is fundamentally wrong. However, even if one were to assume for the sake of argument that the concern has merit, BREDL's position presents a classical "chicken vs. egg" paradox. In order to conduct the tests suggested by IRSN, irradiated MOX fuel rods would be required. In order to have irradiated fuel rods, a reactor would have to 71

operate with MOX fuel. However, BREDL would not allow a reactor to operate with MOX fuel until the tests are performed.

164. (Nesbit) In response to Duke Interrogatory 15, BREDL reaffirmed its position that the use of MOX lead assemblies at Catawba requires prior execution of a LOCA test program such as the PHEBUS tests proposed by IRSN.2 6 BREDL addressed the "chicken vs. egg" paradox by allowing for the possibility that PHEBUS tests using M5-clad reactor grade MOX fuel (presumably obtained from European reactors) may be adequate for understanding the relevant phenomena, if accompanied by additional analyses or separate effects tests to understand the additional impact of plutonium isotopic composition on design basis LOCAs.

(Note: BREDL has postulated no mechanism for the isotopic composition (WG vs. RG) of the MOX fuel to adversely impact the response to a LOCA.) PHEBUS tests are complex and require significant preparation, and the reduction and analysis of the data would require additional and substantial time. Even if the necessary arrangements could be made, the BREDL prescription would obviously delay the MOX fuel lead assembly program (and therefore the large-scale disposition of excess weapons plutonium) for a number of years in the name of "reducing uncertainty." Finally, the BREDL response to Duke Interrogatory 15 ignores the question - what should be done to resolve the "chicken vs. egg" paradox if European RG MOX fuel turns out to be unavailable, or if the isotopic issues cannot be resolved to BREDL's satisfaction?

165. (Nesbit) In making provisions for lead test assembly programs at commercial reactors, NRC envisioned the need to conduct limited irradiations of fuel assemblies such as are proposed for the MOX fuel lead assemblies. NRC implicitly recognizes that a small number of 26 Blue Ridge Environmental Defense League's Response to Duke Energy Corporation's First Set of Interrogatories and Requests for Production of Documents, April 14, 2004.

72

lead test assemblies, located in non-limiting core locations, can be used without posing an undue risk to the health and safety of the public. Over the past decades, the nuclear industry has made substantial improvements in fuel design that have contributed to improved fuel performance and utilization. Lead test assembly programs have contributed to improvements in cladding material, development of debris-resistant features, and implementation of integral neutron absorbers in fuel pellets, among other advancements. These types of improvements are beneficial from the standpoint of public health and safety, and they provide for more economical utilization of the uranium in reactor fuel.

166. (Nesbit) The logical extension of the BREDL position, if adopted wholesale by the NRC, would be that lead test assembly programs cannot be carried out unless there is sufficient data such that the lead test assembly program is not needed in the first place. This is obviously not a desirable result because it would inhibit future fuel design improvements that might provide additional safety and economic benefits.

167. (Nesbit) In alleging that differences between LEU fuel and MOX fuel could potentially exacerbate MOX fuel relocation during a LOCA relative to LEU fuel, and that this potential impact introduces an unacceptable amount of uncertainty, BREDL does not acknowledge that there are actually already differences among LEU fuel pellet designs (approved and operating in reactors) that equally could be postulated to have similar effects on fuel relocation. This inherent "uncertainty" remains well within the acceptable margins. The MOX fuel lead assemblies fall within a similar range and are well-bounded by conservatisms in the analysis and acceptance criteria.

168. (Nesbit) For example, some LEU fuel pellets contain integral burnable absorbers (e.g., gadolinium, erbium) within the fuel matrix. Other LEU fuel pellets are coated with 73

absorber material (e.g., zinc diboride). These materials can impact pellet properties such as fuel conductivity and fuel pellet internal gas generation.

169. (Nesbit) Even without different materials, LEU fuel pellets are designed with different sizes and shapes (e.g., length, diameter, dish, chamfer), and such differences could be postulated to affect the characteristics of fuel fragments resulting from a design basis LOCA with fuel relocation.

170. (Nesbit) LEU fuel pellets are also manufactured in different plants using different processes. LEU pellets have different physical characteristics (e.g., theoretical density). These physical differences too could be postulated to affect fuel relocation as well.

171. (Nesbit) At bottom, it is not necessary in a LOCA model to address and to quantify the impact of every difference in order to reduce uncertainty to an acceptable value.

The fundamentally conservative nature of Appendix K LOCA analyses provides reasonable assurance that postulated LOCA relocation effects for either LEU or MOX fuel do not threaten the health and safety of the public. BREDL's desire for perfect certainty with respect to relocation effects of MOX fuel should be considered in the context of the inherent differences among the various LEU fuel designs operating in the United States and abroad.

VII. CONCLUSIONS 172. (All) As described above, the LAR and the supplements in the RAI responses demonstrate that Duke has satisfied 10 C.F.R. § 50.46 by performing LOCA analyses in accordance with 10 C.F.R. Part 50, Appendix K. The AREVA LOCA analysis uses an approved methodology that has been appropriately modified to address differences between MOX and LEU fuel behaviors. The results demonstrate conservative compliance with conservative acceptance criteria.

74

173. (All) At the April 21, 2004 meeting of the Advisory Committee on Reactor Safeguards Subcommittee on Reactor Fuels (Tr. 246), Dr. Ralph Meyer of the NRC Staff stated:

The loss of coolant accident, the effect of MOX - Well, first off, let me say that there clearly are neutron physics effects of MOX, and these can be and are being handled. But when you talk about the fuel part of that, for the loss of coolant accident any connection, any difference between MOX and LEU at this time is purely speculative, and I don't think there is any evidence that there is a difference, although we are, of course, interested in looking.

174. (All) We agree with Dr. Meyer's statement. Contention 1, as explained by BREDL, is pure speculation based on statements made by research organizations addressing issues outside the context of the Catawba MOX fuel lead assembly LAR. There is no evidence for an assertion that MOXILEU differences might significantly and adversely impact design basis LOCA analyses or LOCA fuel performance.

175. (All) Indeed, the BREDL contention was initially based on certain VERCORS tests at severe accident conditions that clearly do not establish a relocation issue at design basis LOCA conditions. BREDL offers no other data or testing and its arguments now pertain primarily to "uncertainty" related to the possibility of fuel relocation at design basis LOCA temperatures.

176. (All) In our testimony we have described the analyses performed to demonstrate that the MOX fuel lead assemblies will meet the relevant regulatory acceptance criteria for LOCA. We have shown that these analyses address the pertinent MOX/LEU differences through explicit modeling or conservative assumptions.

177. (All) In our testimony we have explained the substantial conservatisms (and associated margin) embedded in the NRC's criteria that more than compensate for any uncertainty related to fuel relocation. In addition, we have explained the conservatisms (and associated margin) in the MOX fuel LOCA analyses. We have also explained why, in our 75

opinion, MOX fuel relocation effects should not be significantly different from postulated effects for LEU fuel. In fact, we have shown that the MOX/LEU difference in end-of-life power would be beneficial to MOX fuel with respect to any fuel relocation effect. Therefore, taken in context, any "uncertainty" related to fuel relocation does not pose a significant issue.

178. (All) We have also shown that even superimposing IRSN's most pessimistic estimates of the impact of fuel relocation on top of the conservative Catawba MOX fuel LOCA predictions, the results would be within the acceptance criteria of 10 C.F.R. § 50.46.

179. (All) Finally, the inherent assertion in Contention I that all uncertainty must be eliminated prior to approving a lead assembly program is entirely inconsistent with the fundamental purpose of lead assembly programs and would have the undesirable effect of stifling future fuel advancements and delaying an important nuclear non-proliferation initiative.

76

Attachment I STEVEN P. NESBIT Duke Power 526 South Church Street Charlotte, NC 28202 QUALIFICATIONS:

Mr. Nesbit has 24 years nuclear engineeringand management experience in the commercialsectorand on Departmentof Energy (DOE)projects. He is the Mixed Oxide (MOX") Fuel ProjectManagerfor Duke Power, which is playing a key role in the DOEprogram to dispose of surplus weapons plutonium He has 22 years experience with Duke Power. In addition, Mr. Nesbit has managed activitiesfor the Managingand OperatingContractorto DOE's Office of Civilian Radioactive Waste Management. He also has expertise in nuclearsafety analysis technology. Mr. Nesbit has extensive experience interacting with the Nuclear

_ Regulatory Commission and he has authored numerous topical reports and technicalpapers.

EDUCATION/TRAINING:

ME, Nuclear Engineering, University of Virginia, 1982 BS, Nuclear Engineering, University of Virginia, 1980 Graduate course work, Environmental Science Supervisory Development Program, Duke Power PROFESSIONAL AFFILIATIONS/CERTIFICATIONS:

Registered Professional Engineer, North Carolina Registered Professional Engineer, South Carolina American Nuclear Society EXPERIENCE:

3/99-Present Engineering Supervisor I - Duke Power Manages Duke Power's activities as part of the project to dispose of surplus United States weapons plutonium using mixed oxide (MOX) fuel. Directs technical, licensing, and business activities. Serves as a public spokesperson on the MOX fuel project.

09/96-3/99 Consulting Engineer - Duke Power 1

Steven P. Nesbit Led Duke Power's feasibility investigations regarding using MOX fuel at the company's three nuclear plants to support DOE's surplus weapons plutonium disposition program. Served as a representative on the Nuclear Energy Institute's Working Group on Surplus Weapons Plutonium Disposition. Interacted with external groups (Congress, DOE, and the public) in support of the MOX fuel project 11/95-09/96 Engineering Supervisor II - Duke Engineering & Services (DE&S)

Supervised the Design Basis and Project Integration Section of the DOE Office of Civilian Radioactive Waste Management (OCRWM) Management and Operating Contractor. Developed environmental design criteria and performed design basis accident evaluations for an interim storage facility for spent nuclear fuel.

05/94-11/95 Manager, Regulatory Interactions Section - DE&S Manager of the Las Vegas Regulatory Interactions Section of the Regulatory and Licensing Department of the Management and Operating contractor for the DOE OCRWM. Responsibilities of the seven-person section included interactions with the Nuclear Regulatory Commission (NRC) staff and on-site representatives, the Advisory Committee on Nuclear Waste, and the Nuclear Waste Technical Review Board; development of regulatory positions; regulatory reviews; Site Characterization Analysis comment responses; regulatory commitments; and NRC issue resolution activities.

12/92-04/94 Engineering Consultant - DE&S Licensing Engineer in the Las Vegas Regulatory and Licensing Department of the Management and Operating contractor for the DOE OCRWM.

Provided nuclear power plant licensing experience and general support to the DOE Yucca Mountain Site Characterization Office. Assisted with interactions between the DOE, the National Academy of Sciences, the Environmental Protection Agency, and the NRC, related to the development of an environmental standard for the potential repository at Yucca Mountain.

1991-1992 Utility Engineering Group (UEG) Site Engineer - DE&S Site Engineer in Washington, D.C., for the DE&S Utility Engineering Group. Provided utility perspective and experience to the DOE for the New Production Reactor Project. Served on the staff of the Chief Engineer of the project. Provided day-to-day liaison with the various project areas. Served as Project Engineer for the UEG. Managed the DE&S Washington, D.C.,

office.

2

Steven P. Nesbit 1990-1991 Senior Engineer - Duke Engineering & Services Worked in the safety review area of the UEG. Provided utility perspective and experience to the New Product Reactor Project in the area of nuclear reactor safety.

1988-1990 Design Engineer - Duke Power Lead engineer in the area of nuclear safety analysis technology, a work group comprised of five engineers. Worked on developing mass and energy release analysis capability for high energy line breaks at Oconee, McGuire, and Catawba Nuclear Stations. Used the RELAP5/MOD002 transient analysis computer code and wrote in-house analytical codes. Worked to develop reactor building analysis capability for large dry and ice condenser containments, including applications of the FATHOMS (COBRA-NC) and CONTEMPT computer codes. Tested the upgraded Oconee training simulator and evaluated vendor performance. Represented the Babcock and Wilcox Owners Group (B&WOG) on the Project Management Group of the Multi-Loop Integral System Test Facility, a thermal-hydraulic research project sponsored by the B&WOG, the Electric Power Research Institute (EPRI) and the NRC. Served on the Duke Power Crisis Management Team.

1982-1988 Design Engineer/Assistant Engineer/Junior Engineer - Duke Power Lead safety analysis engineer for the Oconee Nuclear Station, a work group of up to five engineers. Served as Duke Power representative on the B&WOG Analysis Committee. Participated in the Technical Advisory Group, a committee comprised of B&WOG, EPRI and NRC representatives, which evaluated the need for thermal hydraulic testing related to once-through steam generators. Helped develop symptom-oriented emergency procedures for Oconee. Performed extensive RETRAN benchmarks of plant transients and helped prepare a safety analysis methods topical report for submission to the NRC. Served as one of 12 auditors for the inaugural Duke Power Self-initiated Technical Audit, patterned after the NRC Safety System Functional Inspections. Participated in fuel loading and start-up physics testing at McGuire Nuclear Station. Participated in zero power physics testing at Oconee. Performed system and containment analyses of the Oconee plant. Prepared technical justifications for emergency Technical Specification changes which prevented unnecessary unit shutdowns.

3

Steven P. Nesbit 1979-1982 Reactor Operator/Reactor Operator Trainee - University of Virginia Reactor Facility Reactor Operator Trainee and licensed Reactor Operator for the 2-MW research reactor in Charlottesville, Va. Duties included shift operation work, training and fuel handling.

AWARDS/HONORS:

"Doer of Deeds," Yucca Mountain Site Characterization Office, February 2, 1994.

NewcombfTomton Fellowship, University of Virginia, 1980-1981.

Bachelor of Science with Highest Distinction, University of Virginia, 1980.

PUBLICATIONS:

Nesbit, S. P., Scott, M. W., Eller, J. L., Verbos, F. J., and Costello, M. V., "Non-LOCA Safety Analysis for Operation with Weapons Grade MOX Fuel Lead Assemblies," American Nuclear Society Winter Meeting 2003, New Orleans, LA, November 2003.

Nesbit, S. P. and Eller, J. L., "Basis for the Design of Reactor Cores Containing Weapons Grade MOX Fuel," Advances in Nuclear Fuel Management III, Hilton Head, SC, October 2003.

Anderson, S. L., Gilreath, J. D., Nesbit, S. P., and Laubam, T. J, "Mixed Oxide Fuel Effects on the Integrity of the McGuire and Catawba Reactor Vessels," Fifth Topical Meeting on Spent Nuclear Fuel and Fissile Materials Management, Charleston, SC, September 18, 2002.

Buckner, M. R., Bengelsdorf, H. D., and Nesbit, S. P., "American Nuclear Society Nonproliferation Position Statement," Fifth Topical Meeting on Spent Nuclear Fuel and Fissile Materials Management, Charleston, SC, September 18,2002.

Clark, R H., Dziadosz, D., and Nesbit, S. P., "MOX Fuel Irradiation Program for Disposition of Surplus United States Plutonium," Fourth Topical Meeting on Department of Energy Spent Nuclear Fuel and Fissile Materials Management, San Diego, SC, June 7, 2000.

Nesbit, S. P. and Bengelsdorf, H. D., "A Comparison of Surplus Weapons Plutonium Disposition Technologies," Third Topical Meeting on Department of Energy Spent Nuclear Fuel and Fissile Materials Management, Charleston, SC, September 9, 1998.

S. P. Nesbit, "A Utility Perspective on Surplus Weapons Plutonium Disposition in Existing United States Light Water Reactors," Advances in Nuclear Fuel Management II, Myrtle Beach, S.C.,

March 1997.

S. P. Nesbit, S. J. Brocoum, M. A. Lugo, J. A. Duguid, P. M. Krishna, "Regulatory Perspective on NAS Recommendations for Yucca Mountain Standards," 7th Annual International High-Level Radioactive Waste Management Conference, Las Vegas, NV, May 1, 1996.

4

Steven P. Nesbit J. Carl Stepp, Silvio Pezzopane, Quazi Hossain, Michael Hardy, Steven P. Nesbit, "Criteria for Design of the Yucca Mountain Structures, Systems, and Components for Fault Displacement,"

FOCUS '95 - Methods of Seismic Hazards Evaluation, Las Vegas, NV, September 20, 1995.

J. Carl Stepp, Michael P. Hardy, Quazi A. Hossain, Steven P. Nesbit, J. Timothy Sullivan, "Seismic Design Methodology for a Geologic Repository at Yucca Mountain," 6th Annual International High-Level Radioactive Waste Management Conference, Las Vegas, NV, May 4, 1995.

D. Stahl, S. P. Nesbit, L. Berkowitz, "Approach to Compliance with the NRC Substantially Complete Containment Requirement at the Potential Repository at Yucca Mountain," 6th Annual International High-Level Radioactive Waste Management Conference, Las Vegas, NV, May 3, 1995.

S. P. Nesbit, S. J. Brocoum, "New Public Health and Safety Standards for Yucca Mountain and Their Impact on the Carbon-14 Issue," Waste Management '95 Conference, Tucson, AZ, February 26, 1995.

S. P. Nesbit, R. J. Gerling, and G. B. Swindlehurst, "Qualification of the Oconee RETRAN Model by Comparison with Plant Transient Data," Nuclear TechnoloU, Volume 83, December 1988.

TOPICAL REPORTS:

DPC-NE-1005P, "Duke Power Nuclear Design Methodology Using CASMO4/SIMULATE-3 MOX," Duke Energy, August 2001.

YvP/TR-003-NP, "Seismic Design Methodology for a Geologic Repository at Yucca Mountain,"

U. S. Department of Energy, October 1995.

DPC-NE-3003-P, "Mass and Energy Release and Containment Response Methodology," Duke Power Company, August 1993.

BAW-2079, "Technical Advisory Group Investigation of Once-Through Steam Generator Thermal-Hydraulic Data Requirements," Babcock and Wilcox, March 1989.

DPC-NE-3000, "Thermal-Hydraulic Transient Analysis Methodology," Duke Power Company, July 1987.

SECURITY CLEARANCE:

DOE "L" Clearance (active)

REFERENCES:

DOE and commercial references available upon request 5

Attachment 2 ROBERT C. HARVEY Duke Power 526 South Church St Charlotte, NC 28202 QUALIFICATIONS:

25 years of Thermal Hydraulic and Safety Analysis experience supporting the reload and licensing of pressurized water reactors. Mr. Harvey has performed numerous safety analysis calculations using the RELAP4, RELAP5, RETRAN, TOODEE-2, CONTEMPT-LT, and MAAP computer codes.

EDUCATION:

Nuclear Engineering Graduate Studies, University of Lowell (1980-1982)

BS, Nuclear Engineering, University of Lowell, 1979 Supervisory Training, Yankee Atomic Electric Company, 1994 MAAP Code Utilization and Phenomena seminar, Fauske & Associates Two-Phase Gas-Liquid Flow Seminar, University of Houston Nuclear Power Reactor Safety Seminar, Massachusetts Institute of Technology (MIT)

Simulator Training, Combustion Engineering (CE)

Two-Phase Flow and Heat Transfer, Rensselaer Polytechnic Institute EXPERIENCE:

Senior Engineer - Duke Power Company 2/99 - present Lead Engineer responsible for the LOCA analysis supporting the Oconee, McGuire, and Catawba nuclear plants. Responsibilities include providing interface and oversight of the vendor analyses. In addition, performs LOCA mass and energy release calculation used as input to the containment analysis and performs UFSAR Chapter 15 non-LOCA safety analysis. Specific accomplishments include supporting the Oconee reanalysis to support steam generator replacement and the transition to best-estimate LOCA analysis methods for the McGuire and Catawba units. Serves as a member of the Emergency Operations Facility (EOF) in the position of Accident Assessment Manager.

Provided an independent assessment of the Texas Utilities LOCA analysis supporting the transition to Westinghouse fuel.

1

Robert C. Harvey Engineer - Duke Engineering & Services - (12/97 - 1/99); Senior Nuclear Engineer -

Yankee Atomic Electric Company (5/91 - 11/97)

Lead Engineer for pressurized water reactor (PWR) LOCA analyses supporting licensing for the Yankee Rowe, Maine Yankee and Seabrook Nuclear Power Stations. Areas of involvement included LOCA and containment analyses and severe accident analyses related to Individual Plant Evaluations (IPEs). Specific accomplishments included supporting the Maine Yankee small break loss of coolant accident (SBLOCA) analysis to justify a return to 2440 MWth operation, and providing oversight of Siemens Power Corporation SBLOCA re-analysis. Served as a response team member to Maine Yankee 1996 Independent Safety Assessment.

In addition, supported General Electric (GE) in severe accident analysis for simplified boiling water reactor (SBWR) certification and provided consulting to the Siemens fuel user group in the area of LOCA analysis.

Provided support to Northeast Utilities on severe accident management guidelines (SAMGs) for Millstone Units 2 & 3 and the Seabrook Nuclear Power Station and performed a technical review of ABB/CE reload analysis of St. Lucie Unit 2 for FP&L.

Senior Engineer - Yankee Atomic Electric Company 5/88 - 5/91 Lead Engineer for Yankee Rowe and Seabrook LOCA analysis related activities and for all severe accident analysis activities. Duties involved reload licensing analysis for the Yankee Rowe plant and vendor oversight of Seabrook LOCA analysis activities.

Supported the plant specific model development and certification of the Yankee Rowe plant simulator. Participated in the Yankee Rowe plant life extension (PLEX) effort providing support in severe accident evaluations and pressurized thermal shock (PTS) analysis. Also, provided training to Texas Utilities personnel in LOCA analysis method applications.

Nuclear Engineer - Yankee Atomic Electric Company 5/85 - 5/88 Lead Engineer for Yankee Rowe LOCA analysis related activities. Activities included large break loss of coolant accident (LBLOCA) model development and applications related to reload licensing, steam generator tube rupture (SGTR) analysis and plant request responses. Participated in the development of Yankee Rowe, plant specific emergency operating procedures (EOPs) based on the generic Westinghouse Owners Group (WOG) emergency response guidelines (ERGs). Performed a plant specific analysis to support deviations from the generic WOG guidelines. Also, provided training to Korean Power (KEPCO) engineers in LOCA analysis methods.

2

Robert C. Harvey Engineer - Yankee Atomic Electric Company 6/79 - 5/85 Performed LBLOCA analyses in support of reloads for the Yankee Rowe and Maine Yankee plants. Contributed to model enhancements of the LBLOCA methods used for Yankee Rowe. Participated in developing and assessing the RELAP5YA computer code used for PWR SBLOCA analysis.

_ PROFESSIONAL AFFILIATIONS/CERTIFICATIONS:

American Nuclear Society (ANS), Member The Research Society of Sigma Xi, Associate Member Registered Professional Engineer North Carolina (Registration # 027387)

South Carolina (Registration # 22237)

SELECTED PUBLICATIONS:

1. Maine Yankee Steam Generator Tube Sleeving Thermal-Hydraulics and Safety Analysis Impacts, co-authors K. R. Rousseau, S. Palmer, P. A. Bergeron, presented at

_. the American Power Conference, Chicago, Ill., 1995.

2. Maine Yankee Cycle 15 Core Performance Analysis, YAEC-1907, co-authors, January 1995.

3. Yankee Rowe Pressurized Thermal Shock, Thermal-Hydraulic Analysis, International Heat Transfer Conference, co-authors P. A. Bergeron, N. Fujita, August 1993.

4. Maine Yankee Level II PRA Results, ASME/JSME International Conference on

- Nuclear Engineering, co-author K. E. St. John, March 1993.

5. Thermal Hydraulics Analysis of the Yankee Plant Due to a Stuck Open PORV Using RELAP5/MOD3 Computer Code, RELAP5/RAC-B International Users Seminar, co-authors W. S. Yeung, R. K. Sundaram, November 1991.

6. Yankee Plant Small Break LOCA Analysis, YAEC-1732, co-authors S. Mihaiu-Westerlind, R. K. Sundaram, July 1990.

7. Yankee Nuclear Power Station Core 21 Performance Analysis, YAEC-1731, co-authors, July 1990.

8. Yankee Nuclear Power Station Severe Accident Closure Submittal, YAEC-1711, co-authors, December 1989.

3

Robert C. Harvey

9. Plant-Specific Analysis to Support the Yankee Emergency Operating Procedures, YAEC-1663, co-authors, April 1989.

10. Seabrook Station Risk and Plant Response for Low Power Operating Conditions, YAEC-1623, co-authors, March 1988.

11. RELAP5YA Simulation of LOFT Small Break Experiments L3-6 and L5-1, Transactions American Nuclear Society, Volume 55, co-authors, L. Schor, November 1987.

12. Estimate of Peak Clad Temperature and Its Uncertainty in a Large Break LOCA at Yankee Nuclear Power Station, YAEC-1431P, co-authors, R. K. Sundaram, K. E. St.

John, May 1984.

13. RELAP5YA - A Computer Program for Light Water Reactor System Thermal-Hydraulic Analysis, YAEC-1300P, co-authors, R. T. Fernandez, R. K. Sundaram, J.

Ghaus, A. Husain, J. N. Loomis, L. Schor, R. Habert, October 1982.

14. RELAP4 and RELAP5 Calculation of LOFT L3-5 and L3-6 Experiments: Comparison to Data, ANS Specialists Meeting on Small Break Loss-of-Coolant Accident Analyses in LWRs, co-authors, L. Schor, J. N. Loomis, A. Husain, August 1981.

15. RELAP4 Analysis of CREARE Flashing Transients with Reverse Core Steam Flow, Transactions American Nuclear Society, Volume 38, co-authors G. J. Brown, A.

Husain, August 1981.

16. Applications of a Lower Plenum Phase Separation Model to Yankee Rowe Large Break LOCA Analysis, YAEC-1231, Revision 1, co-authors, March 1981.

17. Maine Yankee Cycle 5 Core Performance Analysis, YAEC-1202, co-authors, December 1979.

4

Attachment 3 BERT M. DUNN AREVA 3315 Old Forest Road Lynchburg, VA 24501 EDUCATION: BS in Physics, Washington State University, 1968 MS in Physics, Lynchburg College, 1973 EXPERIENCE: AREVA Framatome ANP Inc. (formerly Framatome Nuclear Technologies Inc., Nuclear Power Division of the Babcock & Wilcox Company), Lynchburg, Virginia 3/84-Present Advisory Engineer AREVA representative to high burnup industry forums responsible for providing recommendations on licensing criteria to NRC. Lead on AREVA consultation with NRC on irradiated fuel LOCA testing at Argonne National Laboratory.

Member of U.S. NRC High Burnup Fuel Design Basis Accident Technical Evaluation PIRT (Phenomena Identification and Ranking Table) Panel.

Technical lead for the development of LOCA and Safety Analysis techniques for the licensing of new fuel cladding materials.

Technical lead for the development of evaluation techniques to determine the outcome of inherent boron dilution events.

Technical co-lead for the evaluation of best estimate LOCA licensing techniques.

Project lead for development of Loss-of-Coolant-Accident analysis capability for Westinghouse and Combustion Engineering plants.

2/83-3/84 Senior Product Manager Project Manager for Pressurized Thermal Shock (PTS) B&W Owners Group Task Force. Lead development of a probabilistic risk assessment for PTS in B&W plants.

1

Bert M. Dunn 1975-1980 Unit Manager. Emergency Core Cooling System Analysis Responsible for all ECCS evaluations of the performance of B&W-designed nuclear power plants.

1970-1975 Engineer/Supervisory Engineer Licensing of B&W's ECCS evaluation models and development of techniques for evaluating reactor building subcompartment pressure forces.

1968-1970 Engineer - Douglas United Nuclear Corporation. Richland. Washington Reactor physics, fuel engineering, operations.

PUBLICATIONS: Major contributor or the principal author of:

a. BAW-10034, "Multinode Analysis of B&W's 2568 MWt Nuclear Plants During a Loss-of-Coolant Accident," October 1971.

b. BAW-10045, "Multinode Analysis of B&W's 205-Fuel Assembly Nuclear Plants During Loss-of-Coolant Accident,"

May 1972.

c. BAW-10052, "Multinode Analysis of Small Breaks for B&W's 2568 MWt Nuclear Plants," September 1972.

d. BAW-10064, "Multinode Analysis of Core Flooding Line Break for B&W's 2568 MWt Internals Vent Valve Plants," April 1973.

e. BAW-10091, "B&W's ECCS Evaluation Model Report With Specific Application to 177-FA Class Plants With Lowered-Loop Arrangement," August 1974.

f. BAW-10091, Supplement 1, "Supplementary and Supporting Documentation for B&W's ECCS Evaluation Model Report With Specific Application to 177-FA Class Plants With Lowered-Loop Arrangement," December 1974.

g. BAW-10102, "ECCS Evaluation of B&W's 205-FA NSS," June 1975.

h. BAW-10104, "B&W's ECCS Evaluation Model," May 1975.

2

Bert M. Dunn

i. BAW-10106, "- QUENCH - Digital Program for Analysis of Core Thermal Transients During Loss-of-Coolant Accident,"

May 1975.

3

Attachment 4 Dr. J. Kevin McCoy Advisory Engineer- Materials FramatomeANP, Ina General Background Dr. McCoy, a veteran scientist/engineer, has more than 20 years' experience, largely in the nuclear industry. He has proposed, planned, and executed research projects for the Nuclear Regulatory Commission (NRC), the Air Force Office of Scientific Research (AFOSR), the National Science Foundation (NSF), and the Electric Power Research Institute (EPRI). Dr.

McCoy's research included developing models for waste glass degradation, identifying a new mechanism for densification of ceramics, vectorizing computer codes to run 500 times faster, and studying boundaries between quasi-periodicity and chaos as it applies to hydrogen embrittlement. In his research, Dr. McCoy developed methods for calculating the positions of atoms at a crack tip. He even found a method for calculating the energy barrier that must be surmounted to break one atomic bond.

Dr. McCoy worked on the Yucca Mountain Project (YMP) for nearly eight years, providing leadership and technical expertise on materials selection and nuclear waste behavior. He coordinated the efforts of Framatome ANP, Inc. and Lawrence Livermore National Laboratory to identify material properties and make material selection for the waste packages slated for use in the proposed geologic repository. Also, Dr. McCoy led the revision of a comprehensive report on nuclear waste behavior. This report summarizes $25 million worth of research and is one of nine major documents used to support the site recommendation for the nation's first high-level radioactive waste repository.

His recent work has focused on the behavior of commercial nuclear fuel. In the past three years he has written or revised topical reports on dry fuel storage and the performance of a mixed oxide fuel assembly. He has also written several reports on poolside postirradiation examinations of irradiated fuel.

Education

Ph.D., Materials Engineering, Purdue University

M.S., Metallurgical Engineering, Purdue University

B.S. (with Highest Distinction), Metallurgical Engineering, Purdue University Oualifications/Certifications Good conversational and written French Listed in Who's Who in Science and Engineering

Member, American Nuclear Society Publications Civilian Radioactive Waste Management System Management & Operating Contractor, Waste Form Degradation Process Model Report, TDR-WIS-MD-000001 REV 00 ICN 01, July 2000.

1

Dr. J. Kevin McCoy Representative of more than 30 peer-reviewed journal articles authored or co-authored are:

D. Stahl, J. K. McCoy, and R. D. McCright, "Impact of Thermal Loading on Waste Package Material Performance", in Scientific Basis for Nuclear Waste Management XVIII, 671-678, ed. T. Murakami and R. C. Ewing, Materials Research Society, Pittsburgh (1995).

J. K. McCoy, D. Stahl, and T. A. Buscheck, "A Corrosion Model for Waste Package Corrosion-Allowance Materials", in Proceedings of the Sixth Annual International Conference on High Level Radioactive Waste Management, 565-567, American Nuclear Society, La Grange Park, Illinois, and American Society of Civil Engineers, New York (1995).

J. K. McCoy, "Fuel and Cladding Oxidation Under Expected Repository Conditions", in Proceedings of the Seventh Annual InternationalConference on High Level Radioactive Waste Management, 396-397, American Nuclear Society, La Grange Park, Illinois, and American Society of Civil Engineers, New York (1996).

Publications - continued

J. K. McCoy, "Mechanical Failure of Commercial Spent Nuclear Fuel Cladding",

in Proceedings of the Sixth International Conference on Nuclear Engineering, 632-633, American Society of Mechanical Engineers, New York (1998).

_* J. A. Blink, T. W. Doering, J. K. McCoy, R. W. Andrews, J. H. Lee, D.

Sevougian, V. Vallikat; D. G. McKenzie, and J. N. Bailey, "Factors Affecting Performance of Engineered Barriers", in Proceedings of the Eighth Annual InternationalConference on High-Level Radioactive Waste Management, 290-292, American Nuclear Society, La Grange Park, Illinois (1998).

Applicable Work Experience Advisory Engineer II, Framatome ANP, 1993 - Present

Revised topical report on performance of a mixed oxide fuel assembly. Report compares materials and performance of mixed oxide and low-enriched uranium fuels

Wrote topical report on behavior of spent nuclear fuel in dry storage. Provided justification for increasing the allowable burnup limit

Prepared six reports on poolside postirradiation examinations of nuclear fuel assemblies

Predicted behavior of high-level radioactive wastes and support selection of waste container material 2

Dr. J. Kevin McCoy

Led revision of a comprehensive report on nuclear waste behavior. Report summarizes $25 million worth of research and is one of nine major documents that supported site recommendation for the nation's first high-level nuclear waste repository

Critically reviewed models for creep rupture of spent fuel cladding. Determined that previous models did not adequately account for cladding texture. Showed that inclusion of the effects of texture increases predicted creep life by a factor of six

Developed novel mathematical approach for describing degradation of spent nuclear fuel. The new approach is based on a deep insight into the similarities in the degradation behavior of fuel rods that start to degrade at different times.

Lengthy performance calculations are now performed more than ten times faster

Developed computer model to simulate bending and breaking of fuel rods during earthquakes. Proved that earthquakes strong enough to break fuel rods occur only once in a million years and would have very little effect on repository performance.

Principal Research Scientist, Battelle, Metals and Ceramics Department, Columbus, Ohio (1981-1992)

Proposed, planned, and executed research projects. Analyzed physical processes in materials, mostly with C code. Provided own research support. Served as system administrator for departmental UNIX computer system with twenty users

Developed methods for calculating the positions of atoms at a crack tip. Found method to calculate the energy barrier that must be surmounted to break one atomic bond. Knowing the height of the barrier is critical for calculating crack growth rates. This breakthrough came after several years of unsuccessful efforts by other scientists

Developed technique for instrumenting hot isostatic pressing. Instrumentation provides a continuous record of material behavior and increases the amount of data obtained from costly experiments by three to five times

Managed personal computer resources for department with fifty users. Planned hardware acquisitions and allocated resources.

3

UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSING BOARD In the Matter of: )

)

DUKE ENERGY CORPORATION )

) Docket Nos. 50413-OLA (Catawba Nuclear Station, ) 50-414-OLA Units 1 and 2) )

)

TESTIMONY OF STEVEN P. NESBIT, ROBERT C. HARVEY, BERT M. DUNN, AND J. KEVIN McCOY ON BEHALF OF DUKE ENERGY CORPORATION ON CONTENTION I EXHIBITS

I Exhibit 1 Relevant Portions of Duke Energy's License Amendment Request Submitted to the NRC, February 27, 2003 (Attachment 3, Section 3.7.1)

I 9

I a.

Attachment 3 Description and Technical Justification enhanced security requirements during receipt, handling, and storage of unirradiated MOX fuel assemblies. The specific changes to this plan will be submitted separately with the intent of having additional security measures and associated plan changes approved in the same time frame as the license amendments.

3.7 SAFETY ANALYSIS OF MOX FUEL LEAD ASSEMBLIES The MOX fuel lead assemblies have slightly different nuclear and thermal-hydraulic characteristics from the resident Westinghouse LEU fuel assemblies. The effect of these differences on the design basis transients and accidents described in the UFSAR were evaluated to verify that acceptance criteria continued to be met for the MOX fuel lead assemblies.

3.7.1 Impact of MOX Fuel Lead Assemblies on Loss of Coolant Accident Analyses The effects of MOX fuel lead assemblies on core operating and safety limits with respect to loss of coolant analyses (LOCA) were evaluated. With the conservative calculation approach described herein, there were no significant differences in the predicted performance of MOX fuel relative to LEU fuel for LOCA. This conclusion is based on an evaluation of MOX fuel with respect to isotopic content, decay heat, fuel material properties, and on representative LOCA calculations.

MOX fuel phenomena that have the potential to affect LOCA results are addressed in Section 3.7.1.1. Some adjustments to the Framatome ANP large break LOCA evaluation model are required to model MOX fuel. These adjustments are discussed Section 3.7.1.2. A limited set of large break LOCA calculations comparing MOX fuel lead assemblies to LEU fuel assemblies are summarized in Section 3.7.1.3. Section 3.7.1.4 contains a description of the set of MOX fuel lead assembly large break LOCA calculations that will be performed prior to operation with the lead assemblies. Section 3.7.1.5 addresses potential MOX fuel impacts on small break LOCA evaluations. Section 3.7.1.6 discusses potential mixed core loading effects for the MOX fuel lead assemblies.

3.7.1.1 MOX Fuel Phenomena and Lead Assembly Design Features that Potentially Affect LOCA This section addresses the effects of the MOX fuel isotopics on LOCA performance. It is concluded that the changes in delayed neutron fraction and void reactivity feedback are not significant for the lead assemblies and the use of the LEU decay heat standard is shown to be conservative for application to MOX fuel.

3-20

Attachment 3 Description and Technical Justification 3.7.1.1.1 Fissionable Isotope The key difference between MOX fuel and LEU fuel is that Pu-239 is the predominant fissionable isotope in the MOX fuel. The substitution of a MOX fuel assembly for a LEU fuel assembly affects the assembly neutronic behavior, its neutronic interaction with the rest of the core, and the fission product concentrations. Neutronic interaction between MOX and LEU fuel assemblies occurs through the energy spectrum of the neutron flux. It is primarily embodied in a change of the delayed neutron fraction (Ioff), the void reactivity effect, and the prompt neutron lifetime. The Doppler reactivity effect between MOX and LEU fuel is similar and not of consequence in predicting the peak cladding temperature during a LOCA. The differing concentrations of fission products and nuclei activation alter the decay heat rate between MOX and LEU fuel pins. However, as discussed in Section 3.7.1.1.2, LEU fuel decay heat modeling required by current NRC regulations remains conservative for application to MOX fuel.

3 Delayed Neutron Fraction ( ff)

The fraction of delayed neutrons (Peff) is lower in MOX fuel than in LEU fuel. As an example, the delayed neutron fraction for a 40 percent MOX fuel batch application will be reduced from around 0.0063 to about 0.0050 at beginning-of-life (BOL) conditions. This difference has two effects: (1) reactivity changes imposed on the core will produce a larger change in fission power, and (2) the neutron source for shutdown fission power will decrease.

Both effects act to lower the power of the MOX fuel assembly relative to the LEU assembly during the transient.

Change in Void Reactivity Feedback During LOCA, the void effect is responsible for achieving reactor shutdown and maintaining low fission powers in the unquenched regions of the core. Figure 3-2 provides a comparison of a void reactivity curve (effect on assembly k.) for a reference Framatome ANP designed LEU fuel assembly with a void reactivity curve calculated for a weapons grade MOX fuel assembly at the same conditions. A larger negative reactivity insertion occurs for the MOX fuel assembly than for the LEU assembly for all void fractions. This effectively suppresses the MOX fuel assembly power relative to the LEU assembly throughout a LOCA.

Prompt Neutron Lifetime The prompt neutron lifetime decreases for MOX fuel cores. For a 40 percent MOX fuel batch application the lifetime can decrease by approximately 25 percent. This change will not affect LOCA calculations because the prompt neutron lifetime only becomes important for positive reactivity insertions greater than Neff.

Use of Pre-LOCA Peaking throughout LOCA Simulation The LEU fuel LOCA evaluation model assumes constant local peaking factors throughout the accident simulation. If k. of any assembly does not monotonically decrease with increasing voiding, then local assembly peaking (assembly power relative to core average power) can increase during portions of the accident. This could increase the hot pin peaking factor for the fission component of the pin power and bring the assumption of constant peaking into 3-21

Attachment 3 Description and Technical Justification question. However, an examination of the void reactivity function for the plutonium concentrations anticipated for the lead assemblies, Figure 3-2, shows that the local k., for both the MOX fuel and the LEU fuel assemblies is monotonically decreasing with increasing void fraction. Thus, the hot assembly (highest void fraction) power levels are continuously suppressed during the evolution of the accident and the application of the initial peaking factors is justified and conservative for MOX fuel as well as for LEU fuel.

Combined Effects on LOCA Each of the neutronic effects identified as significantly differing between MOX fuel and LEU fuel results in a potential benefit in the MOX fuel parameter value over the corresponding LEU fuel value. Taken together these changes assure that the heat load within the MOX fuel lead assembly during LOCA will be lower than that in the resident LEU assembly. Thus, with all other processes being equal, core cooling mechanisms will more effectively control the cladding temperatures in the MOX fuel assembly than in the LEU fuel assemblies. The actual changes for the lead assemblies will not be significant because the effect of four assemblies on the core neutronic behavior will be limited and the MOX fuel assemblies will be substantially driven by the surrounding LEU fuel assemblies. Because the trend of the neutronic parameters is to the benefit of the MOX fuel assembly, it is conservative, as is done herein, to use LEU fuel neutronic parameter values in MOX fuel LOCA calculations.

3.7.1.1.2 Decay Heat The fission product decay heat rate for MOX fuel assemblies, representative of the lead assembly design, was determined using the 1994 ANSIANS 5.1, "Decay Heat in Light Water Reactors." The actinide heat rate was determined using ORIGEN-S with the SAS2H procedures in the SCALE code system (Reference 9). The result, including the appropriate uncertainties, is that the sum of the decay heat and actinide heat for the lead assemblies, for fully saturated decay chains, falls substantially below that used for LEU fuel cores. Figure 3-3 shows a comparison of decay heat plus actinide heat for MOX fuel, the curve fit applied in the Framatome ANP evaluation model for LEU fuel, and the 1971 proposed ANS 5.1 Standard required by 10 CFR 50.46 Appendix K The MOX fuel curve includes uncertainty factors sufficient to provide a 95 percent level of confidence that there is a 95 percent probability that the decay heat and the actinide heat are over-predicted. The Framatome ANP curve is a conservative fit to the 1971 proposed decay heat standard required by Appendix K Both the Framatome ANP curve and the 1971 standard curve include a 20 percent increase in the decay heat and best-estimate actinide heat prediction.

The MOX fuel decay heat curve is consistently below the Framatome ANP LOCA evaluation model curve for the first 36,000 seconds (10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br />) and, except for times less than 0.1 seconds, consistently below the 1971 proposed ANS standard to 1,000 seconds. Beyond 1,000 seconds, there is no significant difference between the MOX fuel curve and the 1971 proposed standard. Integrating the decay and actinide powers, the total energy represented by the Framatome ANP curve up to the approximate time of peak cladding temperature, 150 to 400 seconds, averages more than 12 percent higher than the MOX fuel curve. Therefore, it is conservative to use the same decay and actinide heat rate for MOX fuel of the lead assembly 3-22

Attachment 3 Description and Technical Justification design as was approved for LEU fuel. No change to the evaluation model is required for MOX fuel decay heat effects.

3.7.1.1.3 Thermal and Mechanical Properties The MOX fuel thermal-mechanical properties are very similar to those for LEU fuel. Six primary fuel properties are used in LOCA evaluations: thermal expansion, thermal conductivity, specific heat, emissivity, elastic modulus, and Poisson's ratio. The COPERNIC fuel rod performance code (Reference 3) differentiates a MOX fuel correlation only for thermal conductivity. 4 For each of these physical properties, the MOX LOCA evaluations will be conducted with close approximations over the LOCA temperature range to the appropriate COPERNIC correlation (MOX or LEU).

3.7.1.1.4 Steady State Fuel Temperature Prediction The Framatome ANP LOCA evaluation model requires that the initial fuel temperature for a LOCA simulation be determined by a NRC-approved fuel performance code. For LEU fuel Framatome ANP has typically used the TAC03 code as discussed in References 10 and 1 1.

However, COPERNIC, a more recent Framatome ANP code has models capable of predicting MOX and LEU fuel performance. Accordingly, Framatome ANP has applied COPERNIC for the determination of the steady state performance of the MOX fuel lead assemblies and for the initialization of comparison LEU fuel calculations. The following subsections discuss the changes to the LOCA evaluation model necessitated by the adoption of COPERNIC for LOCA initialization.

Transient Initialization The main effect on LOCA evaluations due to the change from TAC03 to COPERNIC is that the improved fuel conductivity model alters the RELAP5 fuel-to-clad gap initialization. With TAC03, the RELAP5 gap model was initialized at steady state. Agreement with TAC03 initial volume-averaged fuel temperature predictions was achieved by adjusting the multipliers on the gaseous conductance term coefficient. Multiplier values varied from 0.8 to 2.0. Although the multipliers were retained throughout the transient, they did not impose a significant change in the gap coefficient. With COPERNIC, an adjustment to only the gaseous conductance would require larger multipliers than are deemed appropriate for application throughout the LOCA transient. An alternative approach was chosen for the MOX fuel lead assembly analyses, specifically to initialize RELAP5 with the COPERNIC fuel temperatures and gaseous conductance multipliers of 1.0. The core model will not be in steady state at transient initiation but the gap coefficient will be appropriate for use during the transient. The lack of a time zero steady state is not consequential because the cladding response to a LOCA is a rapid heatup during the first one or two seconds of the transient.

This causes the cladding to pull away from the pellet. Under this condition, the gaseous 4

COPERNIC has been approved by NRC for use with U0 2 fuel. NRC review of COPERNIC for application to MOX fuel is underway with approval expected by January 2003.

3-23

Attachment 3 Description and Technical Justification conductance is the only significant contributor to the gap coefficient. Thus, the approach improves the gap modeling for the LOCA transient relative to the current EM. A sensitivity study documented in Section 3.7.1.3 shows that the effect on peak cladding temperature of changing the gaseous conductance by a factor of 2.0 is small.

Initial Fuel Temperature Uncertainty The use of COPERNIC for LOCA initialization necessitates a determination of the initial fuel temperature uncertainties to be applied to the average core, the hot assembly, and the hot pin.

The measured-to-predicted distribution for COPERNIC, Reference 12, demonstrates that a constant temperature increment should be added to COPERNIC predictions to assure that 95 percent of the data are bounded with 95 percent confidence at high temperatures. Thus, the LOCA simulation for the hot pin should be initialized at the COPERNIC prediction plus the incremental adjustment. Assuming that the uncertainty distribution for COPERNIC is approximately normal, the relationships between the hot pin, the hot bundle, and the average core initial temperature predictions developed for TACO3 in Reference 13 (and approved by the NRC in Reference 14) remain reasonable for application to COPERNIC predictions.

TACO3 applications required that 11.5 percent be added to the hot pin initial temperature to assure a 95/95 prediction and that 3.0 percent be added to the hot assembly to assure a 95/95 confidence. The corresponding temperature adjustments for core initialization with COPERNIC are: 1) no adjustment of the COPERNIC prediction for the average core, 2) the hot assembly predicted temperature is increased by 26 percent of the COPERNIC incremental adjustment, and 3) the hot pin temperature is increased by the full COPERNIC incremental adjustment.

3.7.1.1.5 Plutonium Concentration in Fuel Pins A MOX fuel lead assembly contains three regions or zones of fuel pins, with each region having a different plutonium concentration. The differing plutonium concentrations will have an effect on the material properties of the pin, as described in Section 3.7.1.1.3. This effect is explicitly modeled in the analyses described in Section 3.7.1.3, and the results indicate that the effect is negligible.

3.7.1.2 Evaluation Model Adjustments Required for Lead Assembly LBLOCA Calculations This section describes the changes made to the approved Framatome ANP LBLOCA evaluation model (References 4 and 15) for use in MOX lead assembly calculations. The changes described are directly related to MOX fuel effects.

3.7.1.2.1 Adjustments for COPERNIC The technique for the lead assembly LBLOCA calculations is altered as a result of the use of COPERNIC to specify initial fuel conditions. The alteration involves the initialization of RELAP5 with COPERNIC initial fuel temperatures without adjusting the fuel-to-clad gap 3-24

Attachment 3 Description and Technical Justification coefficient to produce a thermal steady state condition. The fuel is in a transient condition at the start of the LOCA simulation. As discussed in Section 3.7.1.1.4, this approach offers the benefit of preserving the gaseous conductance term of the fuel-to-clad gap coefficient throughout the transient. Additionally, the initial fuel temperature uncertainty adjustments were altered as described in Section 3.7.1.1.4 to reflect the measured-to-predicted distribution from the COPERNIC benchmarks.

3.7.1.2.2 Adjustments for MOX Fuel Physical Properties The approved evaluation model uses fuel materials properties characteristic of LEU fuel. The evaluation for the MOX fuel lead assemblies uses fuel materials properties based upon the COPERNIC code, which is under review for application to MOX fuel. Although these properties do not differ substantially between MOX and LEU fuel, the thermal conductivity correlation within COPERNIC (for LEU fuel or MOX fuel) is improved over the conductivity modeling previously incorporated in Framatome ANP evaluation models.

3.7.1.2.3 Rupture Modeling for Mid-Span Mixing Grids This section describes how the approved fuel pin rupture model will be applied to fuel assemblies incorporating mid-span mixing grids (MSMGs - non-structural grids centered between structural grids). For the purpose of determining bundle blockage characteristics following cladding rupture, the Framatome ANP LOCA evaluation model assumes that the incidence of rupture is distributed throughout the upper two-thirds of the structural grid span within which rupture is calculated. For cores containing fuel assemblies with MSMGs, the modeling assumption is that the rupture density at the location of maximum blockage is not altered from that of a core containing no fuel assemblies with MSMGs. Rupture cooling is modeled in the hot assembly at only one elevation for cores with either type of grid configuration.

3.7.1.3 Representative LBLOCA Calculations To provide validation of the expected LOCA results for the MOX fuel lead assemblies, a set of large break LOCA comparison cases for LEU and MOX fuel assemblies, both of the lead assembly design, were run. All cases simulated a full double-ended guillotine break at the cold leg pump discharge with a CD of 1.0 and an initial power distribution peaked toward the core outlet (10.3-ft elevation). All cases incorporated the evaluation model adjustments described in Section 3.7.1.2, except as noted below for Case 2. The three cases are described below.

Case 1: MOX fuel base case with nominal gap conductance (See Section 3.7.1.2.1).

Case 2: MOX fuel case with 2.0 multiplier on nominal gap conductance.

Case 3: LEU fuel case otherwise identical to Case 1.

3-25

Attachment 3 Description and Technical Justification These calculations demonstrated that no significant difference exists between the two fuel types.

Table 3-2 lists the plant parameters and their values used in the calculations. As indicated in this table, the MOX fuel lead assemblies were held to a total peaking limit (FQ) of 2.4, four percent lower than the limit for the resident LEU fuel. A sequence of events for Case 1, the base MOX fuel lead assembly calculation, is provided in Table 3-3. Table 3-4 shows the results for fuel pins of three differing plutonium concentrations representative of the MOX fuel lead assemblies.

Table 3-5 compares the base MOX fuel evaluation case (Case 1) with the same MOX fuel assembly initialized with a fuel-to-clad gaseous conductance coefficient multiplier of 2.0 (Case 2). Increasing the clad-pellet gaseous conductance coefficient to twice its value approximates the type of core initialization that is used when the initial fuel temperature is obtained from TACO3. The peak cladding temperature changes by about 13 degrees F. The comparison of the MOX fuel (Case 1) and the LEU fuel (Case 3) results show a difference of 37 degrees F. This is expected, given the relatively minor differences in the modeling of the two fuel types.

Figures 3-4 and 3-5 provide information about the evaluation model and the input models for these calculations. Figures 3-6 through 3-11 provide the time dependence for important LOCA parameters based on Case 1. Note that there are no essential differences in calculation results between LEU fuel and MOX fuel with the modeling assumptions and conservatisms used.

The conclusion from these comparison calculations is that:

1) The calculated LOCA performance of MOX fuel and LEU fuel is substantially unaffected by the difference in the fissionable isotope even when no credit is taken for the expected reduction in decay heat in MOX fuel,

2) The impact of the EM core initialization technique, removal of the forced thermal steady state requirement, is small, and

3) The effect of different plutonium concentrations on peak cladding temperature (PCT) is insignificant and need not be specifically modeled.

3.7.1.4 LBLOCA Analytical Basis for Operation The LOCA analytical basis for operation of the lead assemblies will be developed during 2002 and early 2003. It is expected that the results will validate the allowed peaking employed in the sample calculations as shown in Table 3-2. The following calculations will be performed to validate lead assembly operability.

3-26

Attachment 3 Description and Technical Justification

1) Time-in-Life (Burnup) Sensitivity Study to 60 GWd/MThm (assembly burnup)

2) Steam Generator Design Effects Study (Three of the four McGuire/Catawba units have replacement steam generators of slightly altered design and lower tube plugging.)

3) Power Distribution (LOCA Limits) Study to Validate Kz These calculations will employ the model adjustments as described in Section 3.7.1.2.

3.7.1.5 Small Break LOCA (SBLOCA) Evaluation The primary SBLOCA issue is determining the core mixture level as a function of time. After such a determination is made, steam production below the mixture level is used with convection-to-steam and radiation heat transfer models to determine cladding temperatures above the mixture level. For the MOX fuel lead assembly core, the resident fuel assemblies dominate the core mixture level prediction, and the existing licensing calculations are applicable to the lead assemblies. Steam is rapidly diverted from the hot assembly to the average core to achieve a relatively uniform steam velocity across the core. Hence, the steam flow in the hot assembly at the location of the hot spot is characteristic of the average core flow and is essentially independent of the hot bundle power or configuration. Therefore, so long as the surface area for heat transfer or other local film coefficient effects are not altered, there will be no effect on the predicted cladding temperature between the lead assemblies and the resident LEU fuel assemblies. The lead assemblies have the same heat transfer surface area as the resident assemblies. The allowed local power of each MOX fuel lead assembly will not exceed that allowed for the resident fuel assemblies. Therefore, the calculated peak cladding temperatures for the lead assemblies will be less than those calculated for the resident fuel assemblies and it is appropriate for the lead assemblies to use the existing SBLOCA evaluation as their licensing basis.

3.7.1.6 Mixed Core Loading Effects The MOX fuel lead assemblies will reside within a core of Westinghouse LEU fuel assemblies. The lead assemblies will be surrounded by resident LEU fuel assemblies having the same physical dimensions and very similar hydraulic characteristics. The MOX fuel lead assembly design employs MSMGs and the resident fuel design uses intermediate flow mixing grids (IFMs). The design of these mixing grids is such that the MOX fuel lead assembly pressure drop is less than four percent lower than the pressure drop for a resident Westinghouse fuel assembly at design flow rates. Hence, flow diversion favoring one fuel assembly at the expense of the other design is expected to be inconsequential. Therefore, there will be no mixed core impact on the LOCA performance of the resident Westinghouse assemblies. The complete set of lead assembly LOCA calculations will be done with the average core modeled to simulate the hydraulic performance of the resident assemblies, providing a direct evaluation of the resident fuel effects on the MOX fuel lead assemblies.

3-27

Attachment 3 Description and Technical Justification 3.7.1.7 Conclusions There are no significant differences in calculated LOCA performance between LEU and MOX fuel with the modeling assumptions and conservatisms selected. No adverse consequences due to the presence of four MOX fuel lead assemblies in the resident core of LEU fuel assemblies are expected. Therefore, during a postulated LOCA, the MOX fuel lead assemblies behave essentially the same as the resident LEU fuel assemblies and the calculations for the resident assemblies can be applied to the lead assemblies. However, the resident LEU fuel assemblies rely on a best estimate LOCA model as the licensing basis, and the calculations described herein were performed with a deterministic model. To reconcile this difference, the 95/95 bounding LOCA results for the resident assemblies are compared to the lead assembly representative results in Table 3-6. This table will be reconstructed when the final licensing basis calculations are performed. The differences between the calculation approaches and the assembly designs are identified within the table. These differences can, if necessary, be applied to future resident assembly calculations to establish the expected impact on the lead assemblies. This eliminates the need to perform calculations on both resident LEU fuel assemblies and the MOX fuel lead assemblies in the event that revised LOCA calculations are needed. If the need for recalculation specifically concerns the performance of the lead assemblies, specific lead assembly calculations will be made with the models described herein and the relationship between the resident fuel and MOX fuel lead assembly LOCA results reestablished.

3.7.2 Impacts of MOX Fuel Lead Assemblies on Non-LOCA Analyses All of the non-LOCA transients and accident analyses described in Chapter 15 of the McGuire and Catawba UFSARs were reviewed to determine the impact of MOX fuel lead assemblies on the results and to verify that acceptance criteria continue to be met. In addition, the mass and energy release analyses in Chapter 6 of the UFSAR were also reviewed for any effect due to MOX fuel. Potential effects due to fuel assembly design differences are addressed in Section 3.7.2.2. The evaluation of MOX fuel effects resulting from changes in core average physics parameters is provided in Section 3.7.2.3. Some design bases transients and accidents are potentially sensitive to local physics parameters, and those are evaluated in Section 3.7.2.4. Potential decay heat effects are addressed in Section 3.7.2.5.

3.7.2.1 Transients and Accidents Evaluated The transients and accidents evaluated and the associated UFSAR sections are listed below.

I) Mass and Energy Release Analysis for Postulated Loss-of-Coolant Accidents (6.2.1.3)

2) Mass and Energy Release Analysis for Postulated Secondary System Pipe Ruptures inside Containment (6.2.1.4) 3-28

Attachment 3 Description and Technical Justification

3.9 REFERENCES

1. MOX Fuel DesignReport, BAW-10238(NP), Revision 0, Framatome ANP, March 2002.

2. Advanced Mark-BWFuelAssembly MechanicalDesign, BAW-10239(P), Revision 0, Framatome ANP, March 2002.

3. COPERNICFuel Rod Design Computer Code, BAW-1023 1P, Revision 0, September 1999.

4. David B. Mitchell and Bert M. Dunn, EvaluationofAdvanced Cladding and StructuralMaterial (M5) in PWR ReactorFuel, BAW-I 0227P-A, February 2000.

5. Duke Power Company NuclearDesign Methodology Using CASMO-4/ SIMULA TE-3 MOX, DPC-NE-J00SP, August 2001.

6. Duke Power Company McGuire/CatawbaCore Thermal-HydraulicMethodology Using VIPRE-01, DPC-NE-2004P-A, Revision 1, February 1997.

7. Duke Power Company Westinghouse Fuel Transition Report, DPC-NE-2009P-A, September 1999.

8. Duke Power Company Thermal-HydraulicStatisticalCore Design Methodology, DPC-NE 2005P-A, Revision 3, 2002.

9. NUREG/CR-0200, CCC-545, ModularCode System for PerformingStandard Computer Analyses for LicensingEvaluation, Oak Ridge National Laboratory, November 1993.

10. D.A. Wesley and K. J. Firth, TAC03 - Fuel Pin ThermalAnalysis Code, BAW-10162PA, October 1989.

11. ExtendedBurnup Evaluation, BAW- 101 86PA, Revisions land 2, and "Supplement 1 to BAW-10186P Revision I Mark-BW Extended Burnup."

12. Letter Framatome ANP to U. S. Nuclear Regulatory Commission, Response to Request for Additional Information on BAW-10231P "COPERNIC Fuel Rod Design Code," Response to Question 3, February 5, 2001.

13. Letter Framatome Technologies to U.S. Nuclear Regulatory Commission, "Modeling Refinements to Framatome Technologies RELAP5-Based, Large Break LOCA Evaluation Models - BAW-1 0168 for Non-B&W-Designed, Recirculating Steam Generator Plants and BAW-10192 for B&W-Designed, Once-Through Steam Generator Plants," FTI-00-55 1, February 29, 2000.

3-38

Attachment 3 Description and Technical Justification

14. Letter U. S. Nuclear Regulatory Commission to Framatome ANP, "Safety Evaluation of Framatome Technologies Topical Report BAW- 101 64P, Revision 4, "RELAP5/MOD2-B&W, An Advanced Computer Transient Analysis (TAC Nos.

MA8465 and MA8568)," April 9, 2002.

15. RSG LOCA - B WNT Loss-of-Coolant Accident Evaluation Model for Recirculating Steam GeneratorPlants, BAW-10168PA, Revision 3, December 1996.

16. Duke Power Company Thermal-HydraulicTransient Analysis Methodology, DPC-NE-3000-PA, Revision 2, December 2000.

17. Duke Power Company MultidimensionalReactor Transients andSafety Analysis Physics ParametersMethodology, DPC-NE-3001-PA, December 2000.

18. Federal Guidance Report No. 11, Limiting Values ofRadionuclideIntake andAir ConcentrationandDose Conversion Factorsfor Inhalation, Submersion, and Ingestion, U.S. Environmental Protection Agency, Office of Radiation Programs, Washington, DC, 1988.

19. Surplus PlutoniumDisposition FinalEnvironmentalImpact Statement, DOE/EIS-0283, U.S. Department of Energy, Office of Fissile Materials Disposition, November 1999.

3-39

Attachment 3 Description and Technical Justification Table 3-2 Plant Parameters and Operating Conditions Used In LOCA Evaluation Parameter Value Reactor Power (MWt) 3411 Pressurizer Operating Pressure (psia) 2310 System Flow (gpm) 382,000 Hot Leg Temperature (degrees F) 616 Cold Leg Temperature (degrees F) 555 Core Average Linear Power Generation Rate (kW/ft) 5.69 I.-

Highest Allowable Total Peaking for MOX Fuel Assembly (F1) 2.4 Hot Pin and Hot Assembly Radial Peaking Factors 1.60 Core Axial Peaking Factor 1.50

Increased to include 102 percent of rated power 3-41

Attachment 3 Description and Technical Justification Table 3-3 Case I - Sequence of Events for MOX Fuel Lead Assembly Calculation Event Time (seconds)

Leak Initiation 0 Accumulator Injection Begins 12.8 End of Blowdown 25.3 Bottom of Core Recovery 39.7 Rupture in Hot Assembly 73 Peak Cladding Temperature (unruptured node) 130 Table 3-4 Plutonium Loading LOCA Results Comparison Calculation Results 2.3 % Pu 3.6 % Pu 4.4 % Pu Calculation Pin Pin Pin Peak Cladding Temperature (degrees F) 2018 2017 2017 PCT Location (f) 8.8 8.8 8.8 Peak Cladding Temperature at Rupture Location (degrees F) 1841 1841 1841 Hot Pin Rupture Location (ft) 9.7 9.7 9.7 Hot Pin Rupture Time (sec) 73 73 73 3-42

Attachment 3 Description and Technical Justification Table 3-5 LBLOCA Sample Calculations Comparison Case 1 Case 2 Case 3 Results MOX Fuel MOX Fuel LEU Fuel 2x Gap Factor Peak Cladding Temperature Data (Peak Pin Data)

Peak Cladding Temp. (degrees F) 2018 2005 1981 PCT Location (ft) 8.8 8.8 8.8 Rupture Node Data Peak Temperature at Rupture 1841 1783 1753 Location (degrees F)

Hot Pin Rupture Location (fi9).7 9.7 9.7

-

Hot Pin Rupture Time (sec) 73 73 71 Oxidation Data Max. Local Oxidation (percent) 4.5 4.6 4.0 Location of Max. Oxidation (It) 8.8 8.8 8.8

Local Oxidation at the end of 400 second simulation.

3-43

Attachment 3 Description and Technical Justification Table 3-6 Comparison of Resident Fuel LOCA Calculation to MOX Fuel Calculation MOX Resident Dfeec Fuel Fuel Difference (95 percentile)

Highest Allowable Total Peaking (FQ) 2.4 2.5 -0.1 Peak Cladding Temperature (degrees F) 2018 2056 -38 Maximum Local Oxidation (percent) 4.5 10 -5.5

After 400 seconds 3-44

Attachment 3 Description and Technical Justification Figure 3-2 Reactivity Insertion versus Void Fraction Infinite Lattice U0 2 at MTC = 0.0 pcmldegree F 2.5000

-5.000 __ _ _ _ _ _ _ _ _ _

C

_2 I2 -12.500 L __

.16.000 ___

-17.500

-20.000____

-LEU foci

.22.500 - -- MOX fuel

-27.501 - I .- 4l 0.00 0.10 020 0.30 0.40 0.50 0.60 0.70 0.80 0.0D 1.00 Vold Fraction 3-49

Attachment 3 Description and Technical Justification Figure 3-3 Decay Heat Rate Comparisons MOX and LEU Fuel Fission Products plus Actinides 0.10

- 0.09 0.08 0.07

- A, 0.06 A..

C 0.0 4 0.04 0.03 -

0.02 -

0.01 -

0.00 _

0.01 0.1 1 10 100 1000 10000 Decay Time, seconds 3-50

Attachment 3 Description and Technical Justification Figure 3-4 Framatome ANP Recirculating Steam Generator LOCA Evaluation Model Codes Ir EOB Conditions I

RELAP5/MOD2-B&W REFLOD3B Blowdown RefillVReflood (Systems Calculation) (Systems Calculation)

EOB Conditio nIs BEACH*

Refill/Reflood (Hot Channel Calculation) Transient BCs

[Strip & Restart RELAP5 Case]

BEACH is a set of reflood heat transfer subroutines in RELAP5.

3-51

Attachmnent 3 Description and Technical Justification Figure 3-5 RELAP5/BEACH Core Noding with Mid Span Mixing Grids Heat Elevation In Active Structure Core Region, Ft Number avs ~ ~12.0 345 29 2445 11.34 344 . 28 444 GMi 10.69

= -S~~acer 343 -

3226

~~i 7_ 169443 10.41 342 25 10.14

_ 341 - 25_-9i 441 >mog 340 - M 24 8440 339 23 439 338 g = cerGrl 22 92438 8.98 0 .7 336 - 26 8.70 19 843 8 16 334 18 7.8642 E e3325 333 15 6 7E19XF 17 . 473 S2116 7 7.27 4 0331 15 43.99 0 330 14 0.72 328 2 6.15 325 93 42 322 -6 42 0 12_ 4.45

-

323 7142-. 9 -

S ce 5i.85 322 6042 2.91 320 4=U8 2 33215

-324 .-= 8e-U. 0.7 @ 1 317 712 0.05 32Core Segment Lent28F 321 ~4""fi3-522

Attachment 3 Description and Technical Justification Figure 3-6 RCS Pressure for MOX LOCA Calculations during Blowdown (10.3 ft Axial Peak) 25001 2000

.5

1; K 0~

1 000 500 0 .

0 4 8 12 16 20 24 28 32 Time, seconds Figure 3-7 Reflooding Rate for MOX LOCA Calculations (10.3 ft Axial Peak) 10 8

6

.I e 4 C01

-2l

-2 - ^ ---

e- -- --- A--- -C A--

U Du SW IOU zVU ZU JUL0 j35 400 Time, seconds 3-53

Attachment 3 Description and Technical Justification 200 250 300 350 400 Time, seconds Figure 3-9 Fuel to Clad Gap Multiplier Study Ruptured Node Results (10.3 ft Axial Peak)

U.

X 1500 K

r-0 50 100 150 200 250 300 350 400 Time,. seconds 3-54

Attachment 3 Description and Technical Justification Figure 3-10 Hot Pin PCT MOX Lead Assembly vs. LEU Assembly (10.3 ft Axial Peak)

. LEGEND MOX LA 25001 - - LEU FA 2000 i 1500 1000 500 0 50 100 150 200 250 300 350 4 0 Figure 3-11 Hot Pin Ruptured Location PCT MOX Lead Assembly vs. LEU Assembly (10.3 ft Axial Peak) 3000 LEGEND IM OX LA 2500 - - LEU FA 2000 i 1500 150 200 Time, seconds 3-55

W Exhibit 2 Relevant Portions of Duke Energy's Response to NRC Request for Additional Information Submitted to the NRC, November 3, 2003 3

II I

I 4

ad

12. Provide the appropriate regulatory criteria to be satisfied by the information in section 3.7, i.e., how this section meets the general design criteria specified in the Standard Review Plan.

Response (Previously submitted October 3, 2003)

Section 3.7 contains the safety analysis of three distinct subject areas; loss of coolant accidents (LOCA), non-LOCA accidents, and radiological consequences. The appropriate regulatory criteria for each of these topics are summarized in Tables Qi12-1 through Q1 2-3.

LOCA Criteria The LOCA acceptance criteria of 10CFR 50.46 (b) were established for light water reactors fueled with U0 2 pellets within cylindrical Zircaloy cladding. The MOX fuel lead assemblies have M5Tm cladding and mixed oxide fuel pellets. The applicability of the 10CFR 50.46 criteria to the MOX fuel lead assemblies is established in Table Q1 2-1.

Non-LOCA Criteria The criteria used to evaluate the non-LOCA transients/accidents in the Updated Final Safety Analysis Report are summarized in Table Q12-2 and except for rod ejection accident criteria are the same criteria used for analysis of non-LOCA transients/accidents in LEU fuel cores.

Provisional Rod Ejection Accident Criteria The current acceptance criteria for a rod ejection accident (REA) at Catawba are described in Section 4.1.2 of Reference Q12-1. These criteria are based on Section 15.4.8 of the StandardReview Plan (Reference Q 12-2), and are summarized below.

1. The radially averaged fuel pellet enthalpy shall not exceed 280 cal/gm at any location.

2. Doses must be "well within" the 10 CFR 100 dose limits of 25 rem whole-body and 300 rem to the thyroid, where "well within" is interpreted as less than 25% of those values.

3. The peak Reactor Coolant System pressure must be within Service Limit C as defined by the ASME Code, which is 3000 psia (120% of the 2500 psia design pressure).

With the exception of the enthalpy limit of 280 calgm, those criteria are equally valid for mixed oxide (MOX) fuel as for low enriched uranium (LEU) fuel during a REA. The dose acceptance criteria relate to the radiological consequences to the public, not the fuel type. The primary system pressure acceptance criterion relates to the integrity of the pressure boundary, not the fuel type.

The enthalpy limit was established to ensure coolability of the core after a REA and to preclude the energetic dispersal of fuel particles into the coolant (Reference Q12-3). The current pressurized water reactor regulatory acceptance criterion of 280 cal/grn is based primarily on experiments such as SPERT that were conducted by the Atomic Energy Commission. More recent REA experiments conducted at the Cabri facility in France, among others, suggest that a lower enthalpy limit may be appropriate, particularly for high 13

burnup irradiated fuel. The Electric Power Research Institute (EPRI) has used the more recent experimental data, coupled with cladding failure predictions using the Critical Strain Energy Density (CSED) approach, to develop proposed REA enthalpy limits as a function of burnup. The work is documented in EPRI's Topical Report on Reactivity Initiated Accident: Bases for RIA Fuel and Core Coolability Criteria" (Reference Q12-4),

which as been submitted to the Nuclear Regulatory Commission (NRC) and is currently under review.

Four MOX fuel rods have been tested under simulated REA conditions as part of the Cabri test program. Of those tests, three experienced no cladding failure with peak enthalpies of 138, 203, and 90 cal/gm. However, the Rep Na-7 test saw a cladding failure with fuel dispersal at an enthalpy of 120 cal/gm. The Rep Na-7 rod had a burnup of 55 GWd/MThm and a cladding oxidation layer of 50 microns (Reference Q12-4, Table 2-1).

Based on the results of that test, it has been postulated that differences in fuel pellet microstructure between MOX and LEU fuel may make MOX fuel more susceptible to disruptive cladding failure at lower fuel pellet enthalpy values.

Accordingly, for the MOX fuel lead assemblies, Duke proposes to use a radial average peak fuel enthalpy limit that is substantially more conservative than the current NUREG-0800 acceptance criterion for LEU fuel. Duke proposes to use a value of 100 cal/gm at all burnups as the acceptance criterion for MOX fuel rods experiencing a power excursion from hot zero power (HZP). This criterion is considered to be appropriate and conservative, for the reasons provided below.

1. The value is significantly lower than enthalpies at which disruptive failure has been experienced in any MOX fuel REA tests.

2. The value is significantly lower than the Fuel Rod Failure Threshold curve for LEU fuel as proposed by EPRI (Reference Q12-4, Figure S-1).

3. MOX fuel rods will be clad in M5Th. Fuel rod corrosion is considered to be a contributing factor to cladding failure under REA conditions. MTS has demonstrated extremely low corrosion relative to Zircaloy-4, the cladding material that was used in all MOX fuel REA tests (see Figure 6.1 of Reference Q12-5).

4. MOX fuel lead assembly rod burnup will be limited to less than 60 GWd/MThm.

5. Applying the criterion only to accidents from HZP excludes accidents initiating from hot full power with a high initial enthalpy (reflective of full power) but no rapid energy deposition in the fuel pellet.

Duke will use the SIMULATE-3K MOX computer code to perform three-dimensional reactor kinetics calculations of licensing basis REAs for all cores containing MOX fuel lead assemblies. Duke will verify that the peak enthalpy in all MOX fuel lead assembly rods remains below the 100 cal/gmn acceptance criterion during postulated REAs.

SIMULATE-3K MOX, described in Section 2.4 of Reference Q12-6, is an extension of SIMULATE-3K. Application of SIMULATE-3K for REAs at Catawba has been reviewed and approved by the NRC (Reference Q12-7) for cores containing LEU fuel.

Analyses of representative cores containing MOX fuel lead assemblies are summarized in 14

Section 3.7.2.4 of Reference Q12-8 and further detail will be provided in the response to Reactor Systems RAI Question 33.

The above criteria are conservative provisional criteria for the MOX fuel lead assembly program. To support the batch use of MOX fuel, Duke intends to propose alternative REA acceptance criteria. Duke plans to document the batch use MOX fuel REA acceptance criteria and REA analytical methodology in a MOX fuel safety analysis topical report and submit the report to the NRC for review in 2004.

References Q12-1. DPC-NE-3001-PA, MultidimensionalReactor Transients and Safety Analysis Physics ParametersMethodology, Duke Power Company, December 2000.

Q12-2. NUREG-0800, U. S. Nuclear Regulatory Commission StandardReview Plan, Revision 2, July 1981.

Q12-3. Meyer, R. O., McCardell, R. K., Chung, H. M. Diamond, D. J. and Scott, H. H.,

A Regulatory Assessment of Test Datafor Reactivity-InsertionAccidents, Nuclear Safety, Volume 37, No. 4, October-December 1996.

Q124. EPRI Technical Report 1002865, Topical Report on Reactivity Initiated Accident: Basesfor RIA Fuel and Core CoolabilityCriteria,June 2002 (currently under NRC review).

Q12-5. BAW-10238(P), Revision 1, MOXFuel Design Report, Framatome ANP, May 2003 (currently under NRC review).

Q12-6. DPC-NE-1005P, Duke PowerNuclearDesign Methodology Using CASMO-4/SIMULATE-3 MOX, August 2001 (currently under NRC review).

Q 12-7. DPC-NE-2009-P-A, Revision 2, Duke Power Company Westinghouse Fuel TransitionReport, December 2002.

Q12-8. Tuckman, M. S., February 27, 2003 Letter to U.S. Nuclear Regulatory Commission, Proposed Amendments to the Facility Operating License and Technical Specifications to Allow Insertion of Mixed Oxide Fuel Lead Assemblies and Request for Exemption from Certain Regulations in 10 CFR Part 50.

Radiological Dose Criteria General radiological criteria are provided in 10CFR 20, 10CFR 50 Appendix A, 10CFR 50.67 and IOCFR 100. These are not published as uranium specific criteria, but have been consistently applied to reactor applications by the nuclear industry. Some of these regulations also apply to other applications, such as nuclear medicine. The applicable acceptance criteria in 10CFR are determined by the purpose or scenario for which the consequences must be calculated, rather than by the source term or specific isotopes involved.

The purpose of modeling the event and projecting consequences is to protect the health and safety of the public. To that end, there must be a standard for comparison to draw a definitive conclusion as to the impact upon the public. In order to compare the biological effects from the various isotopes which are produced in nuclear applications and 15

industries, the concept of dose equivalent (or committed dose equivalent) was adopted.

Usually expressed in Reins or Sieverts, these units provide a comparison of biological effects by accounting for the energy deposition and the relative biological effectiveness from radiation emitted by isotopes.

Since dose is a measure of the cumulative biological effect of the emitted particles and rays regardless of the isotope of their origin, there is no need to specify specific dose acceptance criteria for a reactor using MOX fuel. Furthermore, the criteria which are currently in regulations for the protection of the health and safety of the public and control room operators can be applied for the same purpose and application that they currently are being applied within a plant's licensing basis. The dose acceptance criteria in 10 CFR can be applied in the same manner as applied for LEU fuel. Standard Review Plan guidance can continue to be applied in accordance with a plant's licensing basis as it has been for LEU fuel. The specific regulatory dose criteria used to analyze MOX fuel events are summarized in Table Q 12-3.

16

Table Q12-1 Applicability of 10CFR 50.46 Criteria to MOX Fuel Lead Assemblies 506 (b) :Applicability to MOXFuel LeadAssemblies This criterion concerns the performance of the fuel pin cladding material during LOCA and is, therefore, primarily related to cladding properties. The MOX lead assembly fuel rods will be constructed using Framatome ANP's MT cladding.

The 2200 0F criterion has been approved by the NRC as applicable to M5

cladding in granting the licensing of replacement fuel for several light water Peak Clad reactors over the last few years. The basis for approval is experimental evidence Temperature that M5Th behavior during LOCA conditions is equivalent to or superior to Zircaloy

< 2200 OF and is documented in BAW-1 0227P-A, 'Evaluation of Advanced Cladding and Structural Material (M5T) in PWR Reactor Fuel," February 2000."

This temperature criterion has no dependence on the fuel pellet design or makeup and is equally applicable for use with either U02 or MOX fuel pellets.

This criterion is fully applicable to the MOX fuel lead assemblies.

This criterion concerns the performance of the fuel pin cladding material during LOCA and is, therefore, primarily related to cladding properties. The MOX lead assembly fuel rods will be constructed using Framatome ANP's MST dadding.

The 17 percent criterion has been approved by the NRC as applicable to M5 cladding in granting the licensing of replacement fuel for several light water reactors over the last few years. The basis for approval is experimental evidence 17% Local that M5 behavior during LOCA conditions is equivalent to or superior to Zircaloy 17iOiocal and is documented in BAW-10227P-A, "Evaluation of Advanced Cladding and Oxidation Structural Material (MST) in PWR Reactor Fuel," February 2000."

The oxidation limit criterion controls the amount of hydrogen available to develop zirconium hydrides which increase the brittleness of the cladding in the post-accident environment. The criterion is not affected by the type of fuel pellet.

This criterion is fully applicable to the MOX fuel lead assemblies.

This criterion assures acceptable conditions within the reactor building and is 1%Core- unrelated to the core fuel and cladding so long as the hydrogen produced per wide percent cladding reacted is unchanged. Because the reaction for both M5 and Oxidation Zircaloy is between zirconium and oxygen, the hydrogen produced per reaction percent is the same for both materials. The criterion is unaffected by the use of M5'cladding and is fully applicable to the MOX fuel lead assemblies.

Core This criterion controls the geometry of the core following a LOCA. As a criterion, it Amenable to achieves its purpose regardless of the cladding material or the fuel pellet makeup.

Cooling It is fully applicable to the MOX fuel lead assemblies.

17

Long-term This criterion controls the availability of long-term cooling systems and core Core conditions. As a criterion, it achieves its purpose regardless of the cladding Cooln material or the fuel pellet makeup. It is fully applicable to the MOX fuel lead Coo assemblies.

18

Table Q12-2 Acceptance Criteria for Non-LOCA Transients/Accidents with MOX Fuel Lead Assemblies TransientlAccident A Criteria

. ^Dsriton.;

'Description :Acc' ince Criteria.

6.2.1.3 LOCA Mass and Energy

Containment design margin is maintained.

Release and Containment

Environmental qualification of the safety related equipment inside 3ressure/Temperature Response containment is not compromised.

3.2.1.4 Secondary System Pipe

Containment design margin is maintained.

Ruptures and Containment

Environmental qualification of the safety related equipment inside 3ressurewTemperature Response containment is not compromised.

15.1.1 Feedwater System Malfunctions that Result in a

Bounded by excessive increase in secondary steam flow analysis Reduction in Feedwater in Section 15.1.2 and same criteria apply.

Temperature

Peak RCS pressure remains below 110% of the design limit 5.1.2 Feedwater System (<2750 psia)

Malfunction Causing an Increase

Fuel cladding integrity shall be maintained by ensuring that the n Feedwater Flow calculated DNB ratio remains above the 95/95 DNBR limit based on an acceptable DNBR correlation.

Peak RCS pressure remains below 110% of the design limit 15.1.3 Excessive Increase in (-<2750 psia) 1econdary Steam Flow

Fuel cladding integrity shall be maintained by ensuring that the calculated DNB ratio remains above the 95195 DNBR limit based on an acceptable DNBR correlation.

Peak RCS pressure remains below 110% of the design limit 15.1.4 Inadvertent Opening of a (<2750 psia)

Steam Generator Relief or Safety

Fuel cladding integrity shall be maintained by ensuring that the Valve calculated DNB ratio remains above the 95/95 DNBR limit based on an acceptable DNBR correlation.

Peak RCS pressure remains below 110% of the design limit

(<2750 psia)

The potential for core damage is evaluated on the basis that it is acceptable if the minimum DNBR remains above the 95/95 DNBR 5.1.5 St Sst P limit based on an acceptable DNBR correlation. If the DNBR falls 1a1ur. eam ystem Ppig below these values, fuel failure must be assumed for all rods that do not meet these criteria. Any fuel damage calculated to occur must be of sufficiently limited extent that the core will remain in place and intact with no loss of core cooling capability.

Offsite doses calculated shall not exceed the guidelines of 10CFR100.

19

Table Q12-2 Acceptance Criteria for Non-LOCA Transients/Accidents with MOX Fuel Lead Assemblies tTransient/Accident .- - Ac aneCira :

Description Acceptnce Citeri 15.2.1 Steam Pressure Regulator

Not applicable, there are no pressure regulators In the McGuire or Malfunction or Failure That Results Catawba plants whose failure or malfunction could cause a steam n Decreasing Steam Flow flow transient.

15.2.2 Loss of External Load

Bounded by turbine trip analysis In Section 15.2.3 and same criteria apply.

Peak RCS pressure remains below 110% of the design limit

(<2750 psia) 15.2.3 Turbine Trip

Fuel cladding Integrity shall be maintained by ensuring that the calculated DNB ratio remains above the 95195 DNBR limit based on an acceptable DNBR correlation.

15.2.4 Inadvertent Closure of

Bounded by turbine trip analysis in Section 15.2.3 and same Main Steam Isolation Valves criteria apply.

5.2.5 Loss of Condenser

Bounded by turbine trip analysis in Section 15.2.3 and same Causing a Turbine Trip critera apply.

Peak RCS pressure remains below 110% of the design limit 15.2.6 Loss of Non-Emergency (<2750 psia)

AC Power to the Station Auxiliaries

Fuel cladding integrity shall be maintained by ensuring that the calculated DNB ratio remains above the 95/95 DNBR limit based on an acceptable DNBR correlation.

Peak RCS pressure remains below 110% of the design limit 5.2.7 Loss of Normal Feedwater (<2750 psia)

Flow

Fuel cladding integrity shall be maintained by ensuring that the calculated DNB ratio remains above the 95/95 DNBR limit based on an acceptable DNBR correlation.

Peak RCS pressure remains below 110 % of the design limit

(<2750 psia) for low probability events.

15.2.8 Feedwater System Pipe

Fuel cladding Integrity shall be maintained by ensuring that the Break calculated DNB ratio remains above the 95/95 DNBR limit based on an acceptable DNBR correlation.

No hot leg boiling occurs.

Peak RCS pressure remains below 110% of the design limit 15.3.1 Partial Loss of Forced (<2750 psia)

Peactor Ccolant Flow

Fuel cladding integrity shall be maintained by ensuring that the calculated DNB ratio remains above the 95/95 DNBR limit based on an acceptable DNBR correlation.

20

Table Q12-2 Acceptance Criteria for Non-LOCA Transients/Accidents with MOX Fuel Lead Assemblies Transient/Accident Acce^tance.Criter:

. .. .Acc'ep3tance Criteria Description

Peak RCS pressure remains below 110% of the design limit 15.3.2 Complete Loss of Forced (<2750 psia) eactor Coolant Flow

Fuel cladding integrity shall be maintained by ensuring that the calculated DNB ratio remains above the 95195 DNBR limit based on an acceptable DNBR correlation.

Peak RCS pressure remains below 110% of the design limit

(<2750 psia) 15.3.3 Reactor Coolant Pump

Any fuel damage calculated to occur must be of sufficiently limited Shaft Seizure (Locked Rotor) extent that the core will remain in place and intact with no loss of core cooling capability.

Any activity release must be such that the calculated doses at the site boundary are a small fraction of the 10CFR100 guidelines.

15.3.4 Reactor Coolant Pump

Bounded by reactor coolant pump shaft seizure analysis in Shaft Break Section 15.3.3 and same criteria apply.

Peak RCS pressure remains below 110% of the design limit 15.4.1 Uncontrolled Rod Cluster (<2750 psia)

Control Assembly Bank

Fuel cladding integrity shall be maintained by ensuring that the Withdrawal From a Subcritical or calculated DNB ratio remains above the 95/95 DNBR limit based Low Power Startup Condition on an acceptable DNBR correlation.

Fuel centerline temperatures do not exceed the melting point

Peak RCS pressure remains below 110% of the design limit 5.4.2 Uncontrolled Rod Cluster (<2750 psia) 1o4trol Assembly Bank

Fuel cladding integrity shall be maintained by ensuring that the Withdrawal at Power calculated DNB ratio remains above the 95/95 DNBR limit based on an acceptable DNBR correlation.

Fuel centerline temperatures do not exceed the melting point.

Peak RCS pressure remains below 110% of the design limit 15.4.3 Rod Cluster Control (< 2750 psia) assembly Misoperation (system

Fuel cladding integrity shall be maintained by ensuring that the Malafunction or Operator Error) - calculated DNB ratio remains above the 95/95 DNBR limit based aodDrop on an acceptable DNBR correlation.

Fuel centerline temperatures do not exceed the melting point

Peak RCS pressure remains below 110% of the design limit

(<2750 psia) 5.4.3 Rod Cluster Control

Fuel cladding integrity shall be maintained by ensuring that the ssembly Misoperation (System calculated DNB ratio remains above the 95/95 DNBR limit based alfunction or Operator Error) - on an acceptable DNBR correlation.

ingle Rod Withdrawal

Fuel centerline temperatures do not exceed the melting point

Any activity release must be such that the calculated doses at the site boundary are a small fraction of the IOCFRIO0 guidelines.

21

Table Q12-2 Acceptance Criteria for Non-LOCA Transients/Accidents with MOX Fuel Lead Assemblies Transient/Accident l AcceanceCriterial Description

Peak RCS pressure remains below 110% of the design limit 15.4.4 Startup of an Inactive (<2750 psia)

Reactor Coolant Pump at an

Fuel cladding integrity shall be maintained by ensuring that the Incorrect Temperature calculated DNB ratio remains above the 95/95 DNBR limit based on an acceptable DNBR correlation.

15.4.6 Chemical and Volume

Peak RCS pressure remains below 110% of the design limit Control System Malfunction that (<2750 psia)

Results in a Decrease in Boron

Fuel dadding integrity shall be maintained by ensuring that the Concentration in the Reactor calculated DNB ratio remains above the 95/95 DNBR limit based Coolant on an acceptable DNBR correlation.

-

15.4.7 Inadvertent Loading and

Any activity release must be such that the calculated doses at the peration of a Fuel Assembly in site boundary are a small fraction of the IOCFR100 guidelines.

an Improper Position

Peak RCS pressure remains below 120% of design for very low probability events (< 3000 psia).

Fuel cladding integrity shall be maintained by ensuring that the calculated DNB ratio remains above the 95/95 DNBR limit based 15.4.8 Spectrum of Rod Cluster on an acceptable DNBR correlation.

Control Assembly Ejection

Any fuel damage calculated to occur must be of sufficiently limited Accidents extent that the core will remain in place and intact with no loss of core cooling capability.

The fission product release to the environment is well within the established dose acceptance criteria of 10CFR100.

See provisional cal/gm acceptance criteria attached.

Peak RCS pressure remains below 110% of the design limit 5.5.1 Inadvertent Operation of (<2750 psia)

Emergency Core Cooling System

Fuel cladding integrity shall be maintained by ensuring that the During Power Operation calculated DNB ratio remains above the 95/95 DNBR limit based on an acceptable DNBR correlation.

15.5.2 Chemicaland Vonume

Bounded by inadvertent operation of emergency core cooling ncreaSesReactorcont system during power operation analysis in Section 15.5.1 and nvrentsReatrColn same criteria apply.

Peak RCS pressure remains below 110% of the design limit 5.6.1 Inadvertent Opening of a (<2750 psia) ressurizer Safety or Relief Valve

Fuel cladding integrity shall be maintained by ensuring that the calculated DNB ratio remains above the 95/95 DNBR limit based on an acceptable DNBR correlation.

15.6.2 Break In Instrument Line or Other Lines From Reactor Coolant

Any activity release must be such that the calculated doses at the Pressure Boundary That Penetrate site boundary are a small fraction of the 10CFR1 00 guidelines.

Containment 22

Table Q12-2 Acceptance Criteria for Non-LOCA Transients/Accidents with MOX Fuel Lead Assemblies TransientlAccident A Criteria

Description

Fuel cladding integrity shall be maintained by ensuring that the 15.6.3 Steam Generator Tube calculated DNB ratio remains above the 95/95 DNBR limit based Failureon an acceptable DNBR correlation.

Any activity release must be such that the calculated doses at the site boundary are a small fraction of the 10CFR100 guidelines.

23

I I I I I [ r r I I I [ I 1 r F Table Q12-3 Regulatory Dose Criteria For Accidents with MOX Fuel Lead Assemblies

'Accident ~ Classic Source TermReference ~ AlternateeeRefeence Source Term Offsite Doses (EAB and LPZ) _ _ _-_ _

RG' 1.195 LOCA 300 Rem Throd 10CFR100.11 25 Rem TEDE RG 1.183 25 Rem WB SR'1.. p.AIOCFR50.67 Steam Generator Tube Rupture 300 Rem Thyroid RG 1.195 25 Rem TEDE RG 1.183 with fuel failure or pre-incident Iodine spike 25 Rem WB SRP 15.6.3 SRG 15.195 Steam Generator Tube Rupture 30 Rem Thyroid 10CFR100.11/ 2.5 Rem TEDE RG 1.183 with concurrent Iodine spike 2.5 Rem WB SRP 15.6.3 1TOCFR50.67 SRG 15.195 Main Steam Line Break 300 Rem Thyroid RG 1.195 25 R TEDE RG 1.183 with fuel failure or pre-incident Iodine spike 25 Rem WB SRP 15.1. App. A 5 em IOCFR50.67 SRG 15.195 Ap Main Steam Line Break 30 Rem Thyroid RG 1.195 2.5 Rem TEDE RG 1.183 with concurrent Iodine spike 2.5 Rem WB SRP 15.1.5 App. A Locked Rotor Accident 2.5 Rem 30 Rem Thyroid WB RG 1.19583 SRP 15.3.32.ReTEEG118 Rod EjectionRdEetoAciet75 Accident 6.3 Rem Rem Thyroid WB2 RG SRP1.195 15.4.8 App A 6.3 Rem TEDE RG 1.183 FulHnln ciet75 Rem Thyroid RG 1.195 Fuel Handling Accident 6.3 Rem WB2 SRP 15.7.4 6.3 Rem TEDE RG 1.183 Control Roomn Doses Al50 Rem Thyrid 3 RG 1.195 Al5RmWB Appendix Al 5 Rem TIEDE 1CRG1.183 50 Rem skin 1GOCF5196 WB= Whole body, RG=Regulatory Guide, SRP= Standard Review Plan 2 Where a conflict exists between SRP and RG 1.195 on the whole body dose limit for a particular accident, the more current guidance is shown.

3 RG 1.195 specifically states that this criterion may be used in lieu of the one in the SRP.

24

13. To allow the NRC staff to perform confirmatory analysis, please provide both the McGuire and Catawba loss-of-coolant accident (LOCA) input decks for the low enriched uranium (LEU) as well as the MOX fuel rods. Provide the decks in an electronic format, including nodalization diagrams.

Response (Previously submitted October 3, 2003)

The accompanying compact disc includes two RELAP5/MOD2-B&W input decks in UNIX format as follows:

r5moxnrc.in - Input deck for MOX fuel pins, power peaked at 10.3 ft.

r5uo2nrc.in - Input deck for LEU fuel pins, power peaked at 10.3 ft.

These are blowdown input decks used in the deterministic evaluations of MOX and LEU fuel pins reported in the license amendment request. The deterministic MOX fuel calculations comprise the licensing basis for the MOX fuel lead assemblies.

Deterministic LEU fuel calculations were included to address the relative LOCA performance between MOX and LEU fuel.

Figures Q13-1 and Q13-2 are node diagrams for the decks. Figure Q13- 1shows the loop node arrangement while Figure Q13-2 shows the reactor vessel node arrangement.

Figure 3-5 of Attachment 3 to Reference Q13-1 provides some additional detail specific to the core region.

RELAP5/MQD2-B&W is a derivative of the INEL code RELAP5/MOD2. Many changes were made to the INEL code to create the approved Framatome ANP deterministic LOCA code. Because the input for these changes may not be recognizable by other versions of RELAP5, the following list of related input card images is provided to assist the NRC staff.

Card 190: EM Choking Model Specification Card (Activates Framatome ANP specific choked flow break modeling.)

Card 192: EM Critical Flow Transition Data (Activates Framatome ANP specific critical flow break modeling.)

Card 195: Interface Heat Transfer Weighting (Activates Framatome ANP specific interface heat transfer weighting.)

Cards 10000020-10000029: Heat Structure Cards (Activate Framatome ANP specific filtered flow model - I OCFR50.46 Appendix K requirement.)

Cards 10000S80-1000OS99: Reflood Grid and Wall Heat Transfer Factor Data (Activate Framatome ANP specific grid model for droplet breakup and convective heat transfer due to grids.)

Cards lCCCG80I-lCCCG899: Left Boundary Heat Structure Cards Cards ICCCG901-lCCCG999: Right Boundary Heat Structure Cards 25

(Activate the Framatome ANP specific EM heat transfer package.)

Cards 19997000-19999999: EM Pin Model Specification (Activate Framatome ANP specific EM core package providing for dynamic fuel-clad gap conductance and fuel rod swell and rupture. Also provide the M5S' cladding properties.)

Reference Q13-1. Tuckman, M. S., February 27, 2003 Letter to U.S. Nuclear Regulatory Commission, Proposed Amendments to the Facility Operating License and Technical Specifications to Allow Insertion of Mixed Oxide Fuel Lead Assemblies and Request for Exemption from Certain Regulations in 10 CFR Part 50.

26

I r I I I I I XX r I I I I r r I Figure Q13-1 Loop Noding Diagram

- ISV a MSIV ThpIe Loop Single Loop 27

Figure Q13-2

.- Core Noding Diagram

_ l 364 Upper Head 360 IIUpper Plenum 1 358 356 354 OultAnnulus CV 170 I 1o00 CV280 1 352 i} CV200 Lower Plenum 28

14. Provide the reference to the best estimate LOCA model noted in section 3.7.1.7.

Response (Previously submitted October 3, 2003)

Based on RAI Questions 14, 15, and 16 it appears that some clarification is needed with respect to the LOCA analysis performed for the MOX fuel lead assemblies and how this analysis is used to support the lead assembly cores. In summary, the licensing basis for the resident Westinghouse RFA fuel remains the best estimate large break LOCA analysis performed by Westinghouse. Framatome ANP Appendix K analyses demonstrate that changing the fuel pellet material to MOX has no significant impact on peak cladding temperature following a large break LOCA. Framatome ANP Appendix K analyses provide peaking limits that ensure the peak cladding temperature for MOX fuel rods following a large break LOCA remain within the regulatory limit. The following discussion provides a further description of the analysis performed for the resident fuel assemblies as well as the MOX fuel lead assemblies.

Resident Fuel The resident fuel in MOX fuel lead assembly cores will be the robust fuel assembly (RFA) design that is supplied by Westinghouse. The large break LOCA analysis that supports this fuel design is the Westinghouse best estimate method described in Reference Q14-1. The analysis is based on the WCOBRA/TRAC method and includes detailed treatment of the uncertainties associated with the computer models and the inputs related with plant operation. As part of the analysis, Westinghouse performed sensitivity studies to address transition or mixed core effects. This was necessary because the RFA fuel was initially introduced into cores containing Framatome ANP Mark-BW design fuel. The conclusion of the mixed core sensitivities was that the presence of the Mark-BW fuel assemblies had an insignificant impact on the calculated results. Westinghouse also performed small break LOCA calculations for McGuire and Catawba using the NOTRUMP methodology as described in Reference Q14-2. A mixed core penalty of 100 F was assessed and applied to the small break LOCA results to accommodate the presence of the Mark-BW fuel assemblies. Given that the MOX fuel lead assemblies are more similar hydraulically to the RFA fuel than the Mark-BW design fuel, the mixed core penalty developed for the Mark-BW fuel assemblies bounds the MOX fuel lead assemblies. Therefore, the Westinghouse LOCA analyses for the resident RFA fuel remain valid in the presence of four MOX fuel lead assemblies.

MOX Fuel Lead Assemblies To address the MOX fuel lead assemblies, Framatome ANP performed deterministic large break LOCA calculations consistent with the requirements of 10 CFR 50 Appendix K. In order to model accurately the effect of changing the fuel pellet material to MOX, Framatome ANP made modifications to their deterministic large break LOCA method as described in Reference Q14-3. These modifications are described in Section 3.7.1.2 of Attachment 3 to Reference Q14-4. Next, Framatome ANP performed large break LOCA calculations for a MOX fuel lead assembly as well as a Framatome ANP LEU fuel assembly, with both analyses assuming the hydraulic characteristics of the Advanced Mark-BW fuel assembly design. This sensitivity study was performed to assess the impact of the change in fuel rod parameters (MOX vs. LEU) on the calculated results. As discussed in Section 3.7.1.3 of Attachment 3 to Reference Q14-4, this sensitivity study showed that there is essentially no difference between the LOCA results for the MOX 29

fuel and the LEU fuel (APCT of 37 0 F). The Framatome ANP MOX fuel lead assembly results were also compared to the Westinghouse best estimate results to illustrate the similarity of the results. Given the differences in the two analytical methods, a direct comparison of the results is not completely valid. However, the comparison illustrates that the MOX fuel lead assembly with the lower peaking assumptions yields lower peak cladding temperature results (APCT of-380 F).

Following submittal of the MOX fuel license amendment request, Framatome ANP completed additional cases to investigate the impact of steam generator type, time in life, and axial power shape. Two different steam generator designs were examined:

Westinghouse Model D5 steam generators (Catawba Unit 2), with a 10% tube plugging assumption; and BWI steam generators (Catawba Unit 1), with 5% tube plugging. The study concluded that the Model D5 steam generators with the 10% tube plugging assumption are limiting with respect to the Framatome ANP deterministic large break LOCA analysis. The D5 case provided the base case input for the other sensitivities cases.

Framatome ANP performed time in life sensitivities to assess the large break LOCA results as the stored energy in the fuel rod varies with cycle burnup. At burnups greater than 30 GWd/MThm, a KBU factor is applied to limit the PCT for these cases. The KBU factor reduces the FQ (total peaking factor) as well as the FAh (enthalpy rise factor or radial peaking factor).

Furthermore, using the limiting bumup case which uses a KBU of 1.0, i.e., the 30 GWd/MThm case, Framatome ANP evaluated power peaks at different elevations. The purpose of these sensitivities was to establish LOCA limits as a function of core height.

At elevations above the 8 foot elevation a Kz factor was applied. The Kz factor reduces the FQ as well as the axial peaking factor (Fz).

A summary of the sensitivity cases is provided in Table Q 14- 1. The resulting LOCA peaking requirements for the MOX fuel lead assemblies are shown in Figure Q14-1.

These peaking requirements will assure that the MOX fuel will comply with the regulatory limits for LOCA as provided in the response to Reactor System RAI Question 12.

MOX Fuel Lead Assembly Licensing Basis The licensing of the MOX lead assemblies will be based on analysis to determine the relative accident performance between the MOX and resident LEU assemblies because of the different fission source materials. As presented in the license amendment request, large break LOCA calculations, using the Framatome ANP deterministic LOCA evaluation model, have been performed for both LEU and MOX assemblies. The LEU calculations applied the evaluation model as approved by NRC. The MOX calculations applied the evaluation model with specified alterations, described in the LAR, necessary to simulate MOX fuel. The comparison of these two calculations demonstrated the expected result: that there is essentially no difference in the large break LOCA performance between fuel, of comparable design, using MOX pellets and fuel using LEU pellets. An evaluation of the small break LOCA, provided in the LAR, also determined that there would be no differences in the calculated results between the MOX and LEU fuel assemblies. Therefore, the assessment of the Catawba LOCA performance for the 30

cores with four MOX lead assemblies is that LOCA performance is not altered. This result, in combination with a reduction in the allowed peaking factor for the MOX fuel pins, provides the licensing basis for the MOX fuel lead assemblies assuring that all of the criteria of IOCFR50.46 are met.

References Q14-1. WCAP- 12945P-A, Volume 1 Revision 2 and Volumes 2-5 Revision 1, Code Qualification Documentfor Best-EstimateLoss of CoolantAnalysis, March 1998.

Q14-2. WCAP-100564P-A, Westinghouse Small Break ECCS EvaluationModel using the NOTRUMP Code, August 1985.

Q14-3. BAW-10168P-A, Revision 3, RSGLOCA -BWNTLoss-of-CoolantAccident Evaluation Modelfor RecirculatingSteam GeneratorPlants, December 1996.

Q14-4. Tuckman, M. S., February 27, 2003 Letter to U.S. Nuclear Regulatory Commission, Proposed Amendments to the Facility Operating License and Technical Specifications to Allow Insertion of Mixed Oxide Fuel Lead Assemblies and Request for Exemption from Certain Regulations in 10 CFR Part 50.

Table Q14-1 Summary of MOX Fuel Lead Assembly Large Break LOCA Sensitivity Cases Model D5 SGs with 10% Tube Plugging TIL ' Elevation, (Gd/~u) ~Lkz Feh Fz F PCT (OF)

BOL 6.8556 1.0 1.0 1.6 1.500 2.4 1919.2 20 6.8556 1.0 1.0 1.6 1.500 2.4 1943.6 30 6.8556 1.0 1.0 1.6 1.500 2.4 1948.8 50 6.8556 0.867 1.0 1.387 1.500 2.08 1824.4 60 6.8556 0.8 1.0 1.280 1.500 1.92 1787.6 30 4.7001 1.0 1.0 1.6 1.500 2.4 1815.0 30 8.5656 1.0 0.993 1.6 1.490 2.383 1964.0 30 10.2756 1.0 0.972 1.6 1A58 2.332 2019.5 31

Figure Q14-1 MOX Fuel Lead Assembly Total Core Peaking Factor 2.5-2.4- - - - - - - - - - - - - - -

M 0L 0

--a-8.5656 ft

-*-12 ft 1.7 0 10 20 30 40 50 60 Bumup (GWdIMThm) 32

15. Provide the uncertainty analysis that was performed for the LEU and MOX LTA demonstrating that the 95/95 peak cladding temperature has been calculated for the core.

The response is expected to include a complete discussion of the statistical methodology used.

Response (Previously submitted October 3, 2003)

The MOX fuel and LEU fuel LOCA analyses that support the use of the MOX fuel lead assemblies are deterministic calculations and therefore no uncertainty analysis was performed. See the response to Reactor Systems RAI Question 14 for additional explanation.

16. Section 3.7.1 states that the LOCA model used for the LEU fuel is a best estimate model.

Provide the Phenomena Identification and Ranking Table for the LOCA analyses performed with the best estimate model and reference the best estimate model used for the analysis.

Response (Previously submitted October 3, 2003)

The Phenomena Identification and Ranking Table (PIRT) used in the Westinghouse best-estimate LBLOCA analysis is contained in Reference Qi6-1. Since this method was not used to directly support the MOX fuel lead assemblies, this PIRT is not applicable to the MOX fuel lead assembly analysis. See the response to Reactor Systems RAI Question 14 for additional explanation.

Reference Q1 6-1. WCAP- I 2945P-A, Volume I Revision 2 and Volumes 2-5 Revision 1, Code QualificationDocumentfor Best-EstimateLoss of CoolantAnalysis, March 1998.

17. Provide the experimental data base used to assess the biases and to determine the uncertainties in the fuel rod behavior for the MOX LTA.

Response (Previously submitted October 3, 2003)

The database is provided in Chapter 3 of the COPERNIC topical report (Reference Q17-1). Additionally, at the NRC's request, several MOX fuel rods from the Halden experiments were analyzed with COPERNIC to end-of-life burnups in the range of 50 to 64 GWd/MThm.

Reference Q17-1. BAW-10231P Revision 2, COPERNICFuelRodDesignComputer Code, July 2000.

18. In sub-section 3.7.1. 1.1, nothing is mentioned about the MOXALEU interface behavior.

Provide a qualitative and quantitative discussion regarding the neutron flux behavior at the interface of the MOX and LEU fuel assemblies.

Response (Previously submitted October 3, 2003)

Duke used the CASMO-4 computer code to model pin cell neutron flux and power at the intersection of four quarter-assembly lattices. These "colorsets" provide detailed two dimensional neutronic calculations that account for interface effects between dissimilar fuel assemblies. MOX fuel assemblies and LEU fuel assemblies of equivalent lifetime 33

21. How does the lower fuel conductivity of the MOX fuel impact the maximum pellet centerline temperature during a LOCA as compared to LEU fuel? Please provide a qualitative and quantitative discussion of the differences.

Response

There is only a slight difference in the fuel pellet conductivity between MOX fuel of the lead assembly design and plutonium concentration and comparable LEU fuel. Figure Q21-1 compares the thermal conductivity for MOX fuel pellets of the lead assembly design to comparable LEU fuel pellets for both un-irradiated fuel and fuel irradiated to 40 GWd/MThm. The thermal conductivity values shown in Figure Q21 -1 are from the fuel performance code COPERNIC (Reference Q21-1). COPERNIC has been approved by NRC for use with LEU fuel and is under review for MOX fuel applications. Although thermal conductivity values in Figure Q21-1 change with burnup for both MOX fuel and LEU fuel, the offset, approximately two percent, is constant.

The analyses presented in Section 3.7.1 of Attachment 3 to the license amendment request (Reference Q21 -2) directly compare the effect of the MOX to LEU offset in conductivity in conjunction with the other differences in the fuel pin designs. Figures Q21-2 and Q21-3 provide a fuel pin temperature profile comparison between MOX and LEU fuel pellets at the accident initial conditions and at the approximate time of peak cladding temperature. As expected, there is little difference in the temperature distributions between the two fuel types. Figure Q21-4 provides the evolution of the centerline fuel temperatures with time for the MOX and LEU fuel at the location of peak cladding temperature. The two fuel temperatures differ slightly during the course of the transient.

The variation is attributed to fuel pellet thermal conductivity and to other differences in the fuel pin design. As an example, the LEU fuel pin has a higher pre-fill pressure than the MOX pin. The higher pressure increases the hoop stress resulting in a slightly lower calculated rupture temperature and earlier calculated rupture time. Combined with all of the models interacting to determine the cladding and pellet temperatures the LEU fuel centerline temperature is 40'F cooler at the time of peak cladding temperature.

The difference in thermal conductivity between MOX fuel of the lead assembly design and comparable LEU fuel is small. The effect of this difference on LOCA calculational results is nil and not distinguishable from the effects of normal fuel design variations.

References Q21-1. BAW-10231P Revision 2, COPERNICFuel Rod Design Computer Code, July 2000.

Q21-2. Tuckman, M. S., February 27, 2003, Letter to U.S. Nuclear Regulatory Commission, Proposed Amendments to the Facility Operating License and Technical Specifications to Allow Insertion of Mixed Oxide Fuel Lead Assemblies and Request for Exemption from Certain Regulations in 10 CFR Part 50.

41

Figure Q21-1 Thermal Conductivity Comparison for MOX and LEU Fuel (Fuel porosity of 0.0479) 9.OE-04 O0 1a- LEU OGWdtmtU 8.OE-04 a a a a - LEU t240 GWd/mtU a '-e-4wt%MOX@OGWd/mthm 4Z 7.0E-04 i a ,

a -K- 4wt%MOX@40 ' a

\ ' ' 'GWdtmthm S6.0E-04

- -- t- }, -' - ------ - - ------ - - - - - - - - - - - - - - - t I 8 5.OE-04 U

i 4.0E-04 3.0E-04 a a a a a I a a 2.OE-04 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 Fuel Temperatre (F) 42

Figure Q21-2 MOX and LEU Fuel Pin Temperature Profile Comparison at Loss of Coolant Accident Initiation 3500 - CP add_

_ I I d i 3000 --------

l I i

-- ------

2500 - - --------- -- r -- I- I- - - -- - --- - --

2000 --- -- -- - -- - --- --- - -- - --- ---- -- --- r- ------- -- --

0 1 2 3 4 5 6 7 8 9 10 Radial Mesh Point (#)

43

Figure Q21-3 MOX and LEU Fuel Pin Temperature Profile Comparison at Time of Peak Cladding Temperature 3500 3000 I P I WI I Fd 2500 r-r ---- ---- - --------- I------ -I-

--

_b 2000 a-1000 4wt%/MOX PCrTfre - 131secI 4 I 500 0 1 2 3 4 5 6 7 8 9 10 Radial Mesh Point (#)

44

Figure Q21-4 MOX and LEU Fuel Centerline Temperature Comparison for Loss of Coolant Accident 3500 3000

@ 2500

- 2000 I..

U X1500 1000 500 0 50 100 150 200 250 300 350 400 Time (seconds) 45

22. The first paragraph of section 3.7.1.1.2 states that "The result, including appropriate uncertainties, is that .." Please state the uncertainties that are being referred to in this section along with what is considered to be appropriate.

Response

References Q22-1 and Q22-2 are industry standard tools for calculating decay heat for low-enriched uranium (LEU) cores. Analysis of highly burned LEU fuel shows that it produces the majority of its energy from the fission of plutonium isotopes. Therefore, these standard tools are appropriate for calculating decay heat in cores containing MOX fuel and for determining the uncertainties to be applied.

The uncertainties included in the MOX fuel decay heat analysis include:

(1) ANSI/ANS-5.1-1994 standard uncertainties for infinite irradiation by isotope, (2) ANSI/ANS-5.1-1994 "ISO standard" for energy released from fission (the "Q" value),

(3) ANSI/ANS-5.1-1994 standard for absorption burnup correction factors, Gmax (t), and (4) actinide decay uncertainties.

Many of these values are a function of time after shutdown. Table Q22-1 shows the effect of time after shutdown on each of these uncertainties.

To obtain a reasonable statistical (95/95) tolerance/confidence factor to apply to the one sigma uncertainty, the Appendix K requirement and the standards were examined. As explained in the ANSIANS-5.1-1994 standard, the 1.2 uncertainty factor was based on work reported in the Bettis Technical Review by K. Shure. Shure's work stated that a relative uncertainty of 20% would bound all positive deviations in decay periods less than 1 seconds. The measured data indicate that the one sigma uncertainty is about 10%. Thus, there is a factor of two in the Appendix K requirements between the sigma and the bounding value. This implies that a tolerance/confidence factor of two is acceptable to use as a 95/95 percent level of confidence in the determination of conservative decay heat calculations. The MOX fuel decay heat model uses a tolerance/confidence factor of two applied to the uncertainties.

The 95/95 actinide decay heat fraction and the 95/95 fission product decay heat fraction are calculated and summed to produce the MOX fuel decay heat model. Comparing the results of the 95/95 MOX model with the standard Appendix K decay heat model for LEU fuel shows that the LEU model produces higher values of decay heat than MOX fuel. This is shown in Figure 3-3 of Attachment 3 of Reference Q22-3 for LOCA-typical decay times.

References Q22-1. American National Standard for Decay Heat Power in Light Water Reactors, ANSI/ANS-5.1-1994, American Nuclear Society, 1994.

Q22-2. O.W. Hermann, R.M. Westfall, ORIGEN-S: Scale System Module to Calculate Fuel Depletion, Actinide Transmutation, FissionProductBuildup and Decay, andAssociated Radiation Source Terms, NUREG/CR-0200, September 1998.

46

Q22-3. Tuckman, M.S., February 27, 2003, Letter to U.S. Nuclear Regulatory Commission, Proposed Amendments to the Facility Operating License and Technical Specifications to Allow Insertion of Mixed Oxide Fuel Lead Assemblies and Request for Exemption from Certain Regulations in 10 CFR Part 50.

47

I I i I r r r I Ir r I Table Q22-1 Effect of Time after Shutdown on Decay Heat Uncertainty Factors Uncertainty Uncertainty Value Uncertainty Range (From Text Parameter (1.0 Second after (100 to 107 Seconds Reference of Q22) Shutdown) after Shutdown)

(1) One sigma uncertainty for U 2.8% 1.7 - 2.8% ANSI IANS-5.1-1994, Page 14 fission product decay heat (1) One sigma uncertainty for U 9.0% 3.8 - 9.0% ANSI /ANS-5.1-1994, Page 18 fission product decay heat 9

(1) One sigma uncertainty for 23 Pu 4.5% 3.6 - 5.3% ANSI /ANS-5.1-1994, Page 16 fission product decay heat (1) One sigma uncertainty for 24FPu 5.4% 4.4 -10.0% ANSI IANS-5.1-1994, Page 20 fission product decay heat (2) Q-sigma for 235U (MeV per Fission) +0.5 NA ANSI /ANS-5.1-1994, Page 38 (2) Q-sigma for 238U (MeV per Fission) +1.0 NA ANSI IANS-5.1-1994, Page 38 (2) Q-sigma for 239pU (MeV per +0.7 NA ANSI/ ANS-5.1-1 994, Page 38 Fission) ______________________________

() Q-sigma for 241Pu (MeV per+/-07NASIAN-.194Pae3

2) Fission) 07NANIAN51194Pae3 (3) Gmt (t) (Note 1) 2% 2.0 - 18.1% ANSI/ ANS-5.1-1994, Page 26 9 10% NA Note 2 (4) 23 U decay heat one sigma uncertainty (4) 239Np decay heat one sigma 15% NA Note 2 (4) o neigmuncertainty % NA Note 2 (4) Decay heat for all other actinides 20% NA Note 2 one sigm a uncertainty _ _ _ _ _ _ I__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

NA - Not applicable because there is no apparent time dependence of this parameter.

Note 1: Gm,,x(t) is the maximum correction relative to the nominal value of G(t).

Note 2: The actinide decay heat uncertainties are estimated based on the accuracy of ORIGEN-S and measured data.

48

24. Section 3.7.1.1.4 discusses the LOCA transient initialization and the changes made to accommodate using the COPERNIC code instead of the TAC03 code, including the adjustments made to some of the parameters. Provide additional information on the adjustments made, how the adjustments were developed and include any data used to develop the adjustment. Additionally, since these values are used in RELAP5 initialization, please show that throughout the fuel lifetime, the TAC03 and COPERNIC codes predict consistent values for the different fuel parameters used as input for the LOCA analysis.

Response

The discussion in Section 3.7.1.1.4 of Attachment 3 to Reference Q24-1 concerns alterations in the approach used to determine the fuel-to-clad gap conductance and in the values used for the initial fuel temperatures in the three core heat structures of the LOCA simulation. The approach to the fuel-to-clad gap conductance is described in detail in the response to Reactor Systems Question 25. The following discussion presents additional detail regarding the determination of the initial fuel temperatures for the core heat structures.

Because COPERNIC is NRC-approved for LOCA application to LEU fuel and includes modeling for MOX fuel properties, it was selected for the prediction of initial fuel temperatures for the MOX simulations and for the LEU comparison case. COPERNIC is an advanced fuel performance code relative to TAC03 and predictive consistency between COPERNIC and TAC03 should not be expected.

The Framatome ANP deterministic LOCA evaluation model, used to evaluate the MOX fuel lead assemblies, incorporates a two coolant channel, three heat structure core model to assure that the coolant and pin conditions for the hot spot are appropriate. The two coolant channels represent flow in the average core and flow in the hot fuel assembly respectively. The three heat structures represent the average core, the hot bundle, and the hot pin. Both the hot bundle and the hot pin couple thermal-hydraulically with the hot fuel assembly fluid channel. Figure 3-5 of Attachment 3 to Reference Q24-1 illustrates the arrangement. The NRC approved this core representation in, Reference Q24-2.

LOCA calculations include provision for appropriate uncertainties in both transient and initial conditions. One of those uncertainties is the initial fuel temperature or initial stored energy used in the core simulation. To determine the initial fuel temperatures, an NRC-approved fuel performance code, such as COPERNIC or TACO3, is run in accordance with the plant boundary conditions and core power distributions to be simulated. These codes produce best estimate predictions of the core temperature distributions that are transferred, after adding appropriate prediction uncertainties, to RELAP5/MOD2-B&W for the LOCA calculations. The uncertainties are determined from the benchmarks of the fuel performance codes and the make-up of the core region being modeled in RELAP5/MOD2-B&W.

For the hot pin, the LOCA calculation resolves a conservative representation of a single region of fuel pellets in a single rod. The appropriate level of uncertainty to add to the hot 50

pin initial temperature prediction is a temperature increment that gives a 95/95 confidence that the resultant temperature is not under predicted. For COPERNIC, the fuel performance code used for MOX simulations, this would comprise an addition of [ ]

to the prediction of the fuel temperatures along the entire hot pin. For a TAC03-based evaluation, 11.5 percent of the predicted fuel temperature would be added.

For the average core, the LOCA calculation resolves a representation of a large group of fuel pellets in many rods. The appropriate level of uncertainty to add to the initial temperature predictions includes the integration of individual pellet uncertainties over this entire group and a determination of the 95/95 confidence band for the entire group. With the size of the group involved, the aggregate uncertainty is near zero and it is appropriate to initialize this group, the average core, at the fuel performance code prediction without adjustment. With this selection, the COPERNIC [ ] the benchmark temperatures is conservatively ignored.

The more interesting initialization is that for the hot bundle representation. The purpose of the hot bundle is to provide the coolant conditions with which to cool the hot pin. As such, the hot bundle configuration is selected to represent the aggregate of the eight fuel pins immediately surrounding the hot pin. For TAC03, the appropriate 95/95 confidence level for the aggregate initial temperature or stored energy of a group of eight pins requires that the TAC03 prediction be increased by about 2.5 percent. The modeling approved by the NRC in Reference Q24-2 stipulated that the temperature prediction be increased by 3.0 percent to provide a small additional conservatism. The determination of the increase is dependent on the distribution of the uncertainty and bias for the fuel performance code. The TAC03 uncertainty distribution is a Gaussian or normal distribution and the difference in a temperature adjustment to achieve 95/95 confidence between a single member set and the set representing the eight fuel pins surrounding the hot pin is significant, 11.5 percent for the hot pin and about 2.5 percent for the surrounding pins. If the uncertainty distribution for COPERNIC is close to Gaussian, there will be little difference in the relative temperature adjustment. That is, the appropriate adjustment will be the same fraction of the 95/95 adjustment factor for both codes. In determining the uncertainty adjustment for COPERNIC applications, it was assumed that the COPERNIC uncertainty distribution was sufficiently close to Gaussian to employ this logic.

The justification of this argument only requires that the distribution of uncertainty for COPERNIC be reasonably normal and that the temperature adjustment providing a 95/95 confidence for a single member set be known. That the COPERNIC uncertainty is reasonably normal can be observed in a comparison of the TAC03 uncertainty distribution to a histogram of the COPERNIC benchmarks. This comparison is presented in Figure Q24-1 as normalized predicted minus measured data. By observation, the uncertainty distribution for COPERNIC, if correlated, would not differ markedly from that of TAC03 except for a slightly different bias. Thus, for the LOCA evaluation of the MOX lead assemblies, the COPERNIC prediction of the hot bundle initial temperature was increased by the ratio of hot bundle to hot pin adjustment for TAC03 times the hot pin adjustment for COPERNIC, 3.0/11.5 times [ ].

51

Reference Q24-1. Tuckman, M. S., February 27, 2003, Letter to U.S. Nuclear Regulatory Commission, Proposed Amendments to the Facility Operating License and Technical Specifications to Allow Insertion of Mixed Oxide Fuel Lead Assemblies and Request for Exemption from Certain Regulations in 10 CFR Part 50 Q24-2. Letter, U.S. Nuclear Regulatory Commission to Framatome ANP, Safety Evaluation of FramatomeTechnologies Topical Report BA W-10164P Revision 4, RELAP5/MOD2-B& W, An Advanced Computer Programfor Light Water Reactor LOCA andNon-LOCA TransientAnalysis, April 9,2002.

Figure Q24-1 TACO Uncertainty Distribution Compared to COPERNIC Benchmark Histogram 52

25. Section 3.7.1.1.4 discusses RELAP5 initialization, stating that the core model will not be in steady state at transient initialization. Since a false declared steady state can lead to errors from an imbalance, please provide justification for why the RELAP5 model will not be in steady state at transient initiation and how steady state conditions for initialization are assured.

Response

The RELAP5/MOD2-B&W code includes a fuel pin model that represents the fuel rod in accordance with the requirements of IOCFR50 Appendix K. This model explicitly considers the fuel pellet, fuel-to-clad gap and clad-to-coolant heat transfer. It allows for specification of material conductivities for the pellet, gap, and cladding. The gap conductance term accounts for gaseous conductance, fuel pellet-to-cladding contact and radiation.

The initial fuel thermal conditions for LOCA are determined by an NRC-approved steady-state fuel performance code. For the analysis of the MOX lead assemblies, COPERNIC is used. The following input from COPERNIC is transferred to RELAP5/MOD2-B&W:

- Fuel rod temperatures after adjustment for uncertainties (29 axial and 10 radial nodes),

- Fuel pellet and cladding radial geometry,

- Fuel-to-cladding contact pressure,

- Initial internal fuel pin pressure,

- Fuel thermal conductivity, and

- Gas composition.

COPERNIC provides a best-estimate calculation of the initial fuel temperature distributions. To provide suitable inputs for RELAP5/MOD2-B&W, appropriate uncertainties are added to the predicted temperatures when they are transferred. This increase in temperature combined with the fact that COPERNIC and RELAP5/MOD2-B&W have slightly different gap models means that the steady-state initial fuel temperature predictions for the two codes will differ. Previous LBLOCA analyses, based on the TACO3 fuel performance code, accounted for the differences by the application of a gap gaseous conductance multiplier. The multiplier, which was held constant throughout the transient, forces the initial fuel temperature prediction of RELAP5/MOD2-B&W to match the fuel performance code prediction plus uncertainty.

An evaluation of the RELAP5/MOD2-B&W and COPERNIC gap conductance models was performed to understand the differences between the models and to determine whether the application of a constant gap gaseous conductance multiplier (determined at steady-state) remained the appropriate method for accounting for the differences between the models and for the uncertainty adjustment of the initial fuel temperatures. Figure Q25-1 illustrates the differences between the RELAP5/MOD2-B&W and COPERNIC gap gaseous conductance models. The figure shows the multiplier on the RELAP5/MOD2-B&W term that would be necessary for it to match the COPERNIC prediction as a function of steady-state gap thickness. The gap thickness effectively translates to the inverse of time-in-life, where open gap conditions exist at BOL and the gap closes and contact pressures develop with increasing burnup.

53

The results of the evaluation determined that RELAP5IMOD2-B&W and COPERNIC provide similar gap conductance results when the gap thickness is relatively large.

However, there was a noticeable difference in the gap conductance when the gap is small.

The accounting of gaseous conductance for gas space between rough surfaces in contact differs between the two codes. Although a gaseous conductance multiplier would allow RELAP5/MOD2-B&W to generate an initialization that matched the uncertainty-adjusted COPERNIC fuel temperatures, the multiplier value would be large for small gaps and applicable only so long as the gap remains small.

Figure Q25-2 demonstrates the transient gap thickness for LBLOCAs initialized at BOL and 45,000 MWd/MtU. For BOL, the gap is initially open and the increased transient gap does not significantly alter the gaseous conductance. A multiplier of between one and two could be applied without significantly affecting the transient simulation. However, for exposed fuel, the initialization multiplier based on the gaseous conductance model may be as high as six and would only be reduced to between two and three by application of the COPERNIC fuel pellet temperature uncertainties. Such a multiplier would quickly become inappropriate as the gap opens during the transient. Because RELAP5/MOD2-B&W does not have the ability to modify the gap gaseous conductance multiplier during the transient, and it is apparent that the multiplier should be less than two after about five seconds, the gaseous conductance multiplier approach was deemed inappropriate for COPERNIC-based LOCA calculations.

RELAP5/MOD2-B&W does have the capability to directly specify the initial fuel rod temperatures independent of the gap conductance. It is, therefore, possible to force the initial heat structure temperatures to the correct values, albeit by giving up a strict steady-state configuration. To determine the effects of starting the core in a non-steady-state condition, a study of several fuel pins with differing gap coefficients was performed.

LOCA simulations with multiple hot fuel pins, each with the same initial fuel temperature distribution (input specified), but with gaseous conductance multipliers varying from 0.5 to 2.0 were run. The results, Figure Q25-3, demonstrated timing differences in cladding heating and cooling rates, particularly in the first few seconds of the transient. However, the overall cladding and fuel temperature trends were preserved and no significant peak cladding temperature differences were noted. The initial heatup of the cladding and cooldown of the fuel pellet occurred quicker with a high gaseous conductance multiplier.

For reduced gaseous conductance, the opposite was true. After the initial heatup, however, the offset of the cladding and fuel temperatures is aligned to compensate for the differences in the gap conductance and the cladding temperature response are thereafter consistent in both timing and magnitude. Because the fuel energy decrease is delayed for the lower gap conductance, fuel temperatures tend to remain higher during the refill and reflood portions of the LBLOCA, resulting in a tendency for a slightly higher cladding temperature during this phase. Furthermore, because the cladding temperature response is, for the most part, consistent, it can be inferred that the core energy transmitted to the reactor system, which is initialized at steady-state conditions for the plant power, is consistent and that there is not a significant effect on the evolution of the remainder of the primary system during the LOCA transient. Therefore, because Figure Q25-2 shows that the gap opens quickly during a LBLOCA and Figure Q25-1 shows that there is little difference between the gaseous conductance of RELAP5/MOD2-B&W and COPERNIC for open gaps, the best solution is to apply no gaseous conductance multiplier (i.e. a factor 54

of 1.0).

In conclusion, the system model in the MOX demonstration cases was initialized to steady state at the desired peaking conditions and the initial fuel temperatures were set to the COPERNIC-predicted temperatures with appropriate uncertainties added. The method ensures an appropriate specification of the initial fuel stored energy and a proper calculation of the gap conductance during a LBLOCA transient.

Figure Q25-1 Multipliers on RELAP5 to Match COPERNIC Gap Thermal Model A k 5.0 - _

- Ar____ LBLOCA Transient Gap Size Near Peak Power j 3.0 and PCT Node after -S aV seconds.

a - I -

a 1.0 -- _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ _ _ _

n n - _I I__ _ I _ _ _ _ I_ _ _ I__ _ _ I___ _ _ _

0 25 50 75 100 125 150 175 200 Hot Mechankal Gap (micron) 55

Figure Q25-2 LBLOCA Transient Hot Mechanical Gap Sizes 250

-4BOL -- 45GWd/mtU]

C - - - - - ----

2 200 ----- - -

E 0 150 I. I2 I I

VU V _v I I To aAI Iiti secl 100 5 101I053 Tim afe LL Iniito sc 56

Figure Q25-3 LBLOCA Transient Cladding Temperatures at PCT Location I6 TME (s)

(Mg = Gaseous Conductance Multiplier)

26. Provide the basis for assuming that the uncertainty distribution for COPERNIC is a normal distribution.

Response

The actual assumption was that the COPERNIC uncertainty distribution was approximately normal. This assumption and the basis for this are explained in the response to Reactor Systems Question 24.

27. Please provide the basis for the COPERNIC temperature adjustments for core initialization in section 3.7.1.1.4. Additionally, please provide the basis for why the TAC03 temperature predictions are reasonable for application to COPERNIC predictions.

Response

TAC03 temperature predictions have no application to COPERNIC predictions. What was involved in the fuel temperature initialization of the LOCA core simulation was that the relative uncertainty for a specific region of the core, originally developed based on the TACO uncertainty distribution, was applied to the COPERNIC fuel temperature prediction. The application of the same relative uncertainty and the basis for it are 57

explained in the response to Reactor Systems Question 24.

28. In sub-section 3.7.1.6, the subject of mixed cores is discussed. In the middle of the paragraph it is stated that the MOX LTA pressure drop is less than four percent lower than the pressure drop for a resident Westinghouse fuel assembly at design flow rates.

Please provide additional detail on the cause of this pressure drop difference, how it was calculated, and the impact including the consequences of this pressure drop. Also, please provide the design flow rate used for this analysis.

Response (Previously submitted October 3, 2003)

The pressure drop difference between the resident Westinghouse Robust Fuel Assembly (RFA) fuel and the MOX fuel lead assemblies is due to mechanical design differences in the grids and the top and bottom nozzles of the fuel assemblies. Even though the rod geometry, pitch, and axial grid locations are the same, unique design differences in the grids and nozzles themselves cause differences in hydraulic resistance. This overall difference was calculated by evaluating full core RFA and full core MOX models with the VIPRE-01 thermal-hydraulic code and comparing the overall calculated Ap. The code represents these hydraulic differences by means of vendor-provided form loss coefficients for each grid design, top, and bottom nozzles. The design flow rate for these evaluations was the current Technical Specification minimum flow rate of 390,000 gpm.

The impact of this difference in pressure drop is flow redistribution between fuel types in a mixed core environment. This redistribution varies with axial elevation in the core as a direct effect of the difference in local grid form loss coefficients. The consequences of this pressure drop difference result in the need to account for this flow redistribution in the analyses of fuel assembly lift, departure from nucleate boiling ratio (DNBR) in steady state and transient analyses, and fuel assembly performance issues such as maximum allowable crossflow. Flow redistribution is accounted for in these analyses by modeling the hydraulic differences directly in a conservative representation of the mixed core fuel assembly geometry.

29. The staff presumes that a mixed core analysis will be performed to account for the use of four MOX LTAs in the core. Therefore, provide the mixed core penalty that was calculated. If a mixed core calculation was not performed, provide a technical justification for not performing the analysis.

Response (Previously submitted October 3, 2003)

The mixed core MOX fuel lead assembly DNBR penalty is explicitly calculated for the entire range of conditions analyzed in a reload cycle. With the currently licensed Duke Power analysis methodology, maximum allowable radial peaking limits are calculated for a range of axial peak locations and magnitudes as described in DPC-NE-2004P-A. This family of peaking limits is repeated for the various sets of reactor statepoints (power level, pressure, temperature, and flow) analyzed to support cycle reload analyses. This entire set of limits is used to represent the limiting fuel assembly in the core.

To model the mixed core, a bounding model of a single high powered MOX fuel assembly at the center of the core surrounded by a remaining core of resident Westinghouse RFA fuel assemblies was used to calculate the explicit peaking limits. This model contained 58

Exhibit 3 I

A

Lambert, M., et. al., "Synthesis of an EDF and FRAMATOME ANP Analysis on Fuel Relocation Impact in Large Break LOCA," NEA/CSNI/R(2001)18, Aix-en-Provence, March 2001

NEA/CSNI(R(2001)[8

10. Synthesis of an EDF and FRAMATOME ANP Analysis on Fuel Relocation Impact in Large Break LOCA Michel Lambert, Yann Le HWnaff, EDFISEPTEN, Jean-Luc Gandrile, Framatome ANP, France Paper summary During a Large Break (L1) LOCA, there is a rapid loss of coolant, inducing depressurizationand loss of cooling of the fuel rods. Considering the honest fuel pin, simulated with conservative assumptions, the following phenomena happen:

increaseof the stress (depressurization),

- cladding strainand burst,

- violent gap depressurizationof the broakenfuelpins,

- and possible relocation offragments-of the pellets in the birst area.

Locally (a few. centimeters), geometry, and residualpower are therefore potentially modified. Qualitative analysisshows that dominant parametersare:

- thefill up ratio of the balloon by thefragments.

- the balloon diameterin the affected area of the pin,

- the thermalconductiviry of thefragmentedfuel material,

- and the heat transferbetweenfragments andcladding.

Calculationsareperformed using the CATHARE LB version 4Df the CATHARE code (this version is based on CATHARE 2 V).3LI version with additional models specific to LB LOCA) The code has been approved by FrenchSafety Authorityfor Ldrge Break safety analysis.

'The impact of relocationis estimatedusing conservativeassumptionsfor the transient,

- penalizing size and location of the break (stagnationpoint in the core),.

- SERMA + 2 aloawforresidualpower,

- Penalizing pmaximum localpower.

but realistic approachfor relocationtmodels (mainily balloon size, filling rate and hear transfer models) In assembly creep experiment, an azimuthal hot spot is observed so the Keusenhoffs azimuthal temperature difference model, qualifed on REBEKA experiment, is-used to calculate the balloonsize.

Calculationresults show a limited impact on the Peak Clad Temperature (PCI) (-30°C).

Therefore. it can be concluded that relocation phenomenon should not be taken into account in safety regulation,consideringthat-the effect of relocation on claddingtemperature is low, even for the more penalizing transient, simulated with conservativeassumptions (exceptfor relocationphenomenon which is simulated with more realism),

- overall conservatism in the LB safety analysis is higher than relocationpenalty.

- the relocation phenomenon, if it exists, is limited to the burst area (afew centimeters), and to the hot rods.

231

-

FiRA.ATQME ASP Synthesis of an EDF and FRAMATOME ANP analysis on fuel relocation impact in Large Break LOCA GUS Lanmbt(EDFSEPTEN)

Yam Le Hdnaff(fDfSEPTEV)

Jewa-Lc GanMle (FR4M4TOWEANP)

~13 Ut I~w3a~mwjc*Au OCA Wd Meft~

0= 232 mwc I AgAMATOME. A*P RA locatinn in I rgea Rrik LEQA Plan

introduction

description of te phenomenon

conductivity model

burst strain

effect on cladding temperature

conclusion C) WtWMUWLhKk=3rWXUA*Mu LWCAt" O= Wmd22 23 2

A{RAMATOME ANP Reloontion in IJrgp Rrpak I -CA Introduction During LB LOCA, with the conservative conditions of the Safety Reports:

.licOhddessudston

= stress

'=adcl Wain =g cddi bust

=,violent a depressurisat'on w possible eiocatlon of firagnents peet In burst area

Main question for the plant operator:

Ols It necessary to consider the Welocatlon phenomenon hi saety studies ?

W)lEvIz!"vUK9TmrWt11Jm=uOTWUa.AM1 LOCAWm mbookOCOE Uu 22 23 3 U

fRA~MATOME J?

,AInnation in I aVrA Brpak I O-QA Description of the phenomenon

main parameters:

Ofil up ratio Obaloon diameter I.

0 Conductivity of fe fragmented area 0 Exchange between fragmerts and cladding

~) RVICI4oSfU LOCAW U _ . Ct2223 OC~w 4

RBPInatinn in Largp, Ireak l OCA Conductivity of the fragmented area

Many conductivity correlations describing an heterogenous area exist.

Thelmura-Yagi model (ICARE2 model) is the best one with regard to the temperature level and fragment size.

awrOWSzM UIVKZ UU.E .J1A LICkMMO=U.I*222n .

?RCIFSA-6 lAT QME-ar A)t4P

. Rt-nat~ion i6 LArcip- qrPRk UIOCA -,%-_

Conductivity of the fragmented area it X AX -

... kWa: De Va, CM:.

ano: nDs W-* ha2

..

.e ) MK t~wm~worwKAm LOCAW*. hZI2223 r'mag~cE

4 i M ATO ME AM.P RA-nlrtion in I arpe Rrsk I OrCA Burst strain

Realistc approach:

Oin assebly creep experiment, existence at an azimuthal hot spot (hot-s tr t effect)

FRAMATOMECEA-Genoble qualified an hot spot model (keusenhoff) on REBEKA expermet.

O In Keusenhoff model, the azimutha hot spot temperature difference Is governed by the eccentricity parameter.

mvIa7aunmuo-arMaJAi3LOC~rudi p. O=EMud. 2222 7 1 Burst strain Sa CA MGO 410 WA. WA MO0.

LOCAWa m~ OCDCMUNd222 a

_ . ..

ARtAMATOME AN P Rnlonstion in Itrgp Rrra-k LOCA Calculation

To estimate the effect of relocation, CATHARE GB (LB) was used.

CATHARE LB Isapproved by French safety authorities.

O Based on CATHARE 2 VI1 .3L1 version developped by CEA-Grenoble.

O Specific models developed by FFRAMATOME or LB LOCA (refill phase.

cross-flow, WSt models .. ).

z u m LOCAvswds OEr VlCh2223 U

.RAMATQAE AMP RF llonatinn in1 1r.E Rrank I OCA Calculation

Relocation is not taken into account in the present safety regulation.

= a realistic approach can be use

With realistic LB LOCA transient: no relocation

Conservative transient to estimate relocation impact

Description of the reference case (3 loops PWR):

rconservative transient

penanzig size break

residual power SERMA+2a

penalizg maxknmu local power (FOmax)

' relocallon model: realistic approach SOvME*nugnnT FOO1S1IEMnES wU*A3U LOCArAM.egoOMbbLdz 2223 ID

AiRAM.ATMEA~P RCalcultinn in I Rrelk LOCA _rp-Calculation I.

laddlag tepsretbure

=) %i.lwEnsenw 71E"AQM1kT"UaL9AMU LOCAasn UdOCOE 2223 11 4

RAMATQAEL ,4NP Rolonation in I grg Orp-ak I OCA Conclusion EDF considers that relocation phenomenon should not be taken Into account in safety regulation:

3The ffect of relcation dcdding temperature Is weak even with csathoe assurnp ms.

O Al the fethei conseavatisms in te LB saty cauations are higher than relocatoperiatty.

O Relocation pheno non doesn' exist with realistic condons.

OThe nluture node is not necessarly te hot spot node.

OThe relocatin Is lImited to the burst area (a lew centimeters) and to the hotest rods.

ua %JoESZT m q(rgmuvc~tUm S vmIuR LOCA&W e MA~dh G =n~ 22 23 12

Exhibit 4 I

I

Grandjean, C. et. al., "High Burnup U02 Fuel LOCA Calculations to Evaluate the Possible Impact of Fuel Relocation After Burst," NEA/CSNI/R(2001)18, Aix-en-Provence, March 2001

HIGH BURNUP U0 2 FUEL LOCA CALCULATIONS TO EVALUATE THE POSSIBLE IMPACT OF FUEL RELOCATION AFTER BURST C. GRANDJEAN, G. HACHE, C. RONGIER Institut de Protection et de Sdret6 Nuclcaire CEN Cadarache, FRANCE.

Abstract A literature review, conducted at IPSN, of available results of past LOCA in-pile experiments with irradiated fuel has revealed that irradiatedrods behavior was signficantly different from that of unirradiatedrods under similar condirions.

Partzularly, as suggestedfrom the resultsftom PBF-LOC FR2 and FLASH5 experiments indicating a general occurrence of the relocation offragmented fuel within the ballooned cladding, a main concern was raised regarding the possible impact offuel relocation on peak clad temperature and local oxidation rate.

In view ofobtaining some insight into thefuel rodperformance followingfuel relocationin a PWR high BU U02 rod under LOCA, calculations were being performed. using the French CATHARE-2 code with, specific modifications in thefuel routines so as to describe a fuel accumulation after burst in the ruptured mesh of the rod cladding.

Main results indicate that the peak'clad'temperature maj increase significantly. bu still remains'below'the ECCS acceptarice limit on PCT.On the other han4 the maximum cladding oxidation rate may exceed the 17%

acceptance limit when the initial (in service) oxidation rate is cumulated with the transient oxidation rate.

However, alternative embrittlement criteria based on residual thickness of ductile metaL such as the Chung and JKassner criteria indicate a fair remaining margin to the thermal shock embrittlenent limit, whereas the handling embrittlement limit may be exceeded 1 INTRODUCTiON hI thetfollowing of-tbe studies that were jointly conducted by IPSN and EDF in order to investigate, the behavior of high 'burnup fuel cladding under LOCA conditions, IPSN' has been re-examining the problem or Loss-of-Coolant-Accidents with consideration of specific aspects related to fuel and cladding irradiation, so as to identify the remaining needs for further studies and experimental daia.

These concerns have led IPSN to initiate new studies in order to provide the answers to pending qiestions regarding the behavior of irradiated rods and'assemblies under LOCA conditions.

In a preliminary step, in view of obtaining some insight into the fuel rod performance following fuel relocation in a PWR high BU U0 2.rod under LOCA. calculations were being perforined, using the

.French CATHARE-2 code with specific modifications in. the fuel routines so as to describe a fuel accumulation after burst in'the ruptured mesh of the zod'cladding.

2 BACKGROUND 2.1.. Irradlatedfuel rodbehavlor 2.1.1 Uterature review There exists a few number of available results of such experiments with irradiated fuel rods under LOCA conditions: main issues were found in results from the. PBF-LOC tests in the USA[1,2], the FR2 tests in Germany[3]; and the FLASHS[4] test in France.

A process of fuel relocation was clearly evidenced from the experimental observations made in these tests series : in all irradiated rods of the PBF-LOC, FR2 and FLASH5 tests, fuel relocation has occurred as a result of slumping of pellets fragments from upper locations into the swollen region of the burst cladding. Fuel relocation phenomena is not restricted to high burnup fuel since fuel fragmentation occurs as soon as low burnup levels (it was thus noticed on LOC5-7B rod, fresh rod pre-conditioned up to 48 MWdft).

A main question concerns the instant of fuel relocation occurrence in these experiments. It is not easy to mnke it perfectly clear for most of tests but, in FR2 tests E3 and E4 that were specially instrumented for that purpose, it was demonstrated that the fuel movement initiation occurred at the time of cladding burst, possibly initiated by the pressure difference between rod plenum and coolant channel with assistance of gravity slumping.

The fuel movement was probably favored in PBF-LOC and FR2 experiments where the fuel-cladding gap was not totally closed, due respectively to low burnup or to the inverted rod internal-external -

pressure difference during initial irradiation at low temperature. A tight bonding between fuel and clad was supposed to counteract the fuel motion inception. However, in FLASH5 experiment with high burnup fuel (50 GWdlt), and in spite of a low clad ballooning (not higher than 16%) post-test examinations have shown that fuel fragments were no more stuck to the cladding : the transient temperature rise combined to clad deformation may be sufficient to suppress fuel-cladding bonding.

2.1.2 Main concern For irradiated fuel rods, as observed in the PBF-LOC results, the clad deformation is expected to be larger than for fresh rods, as a result of a more uniform temperature distribution associated to pellet-clad gap reduction following clad creepdown during rod irradiation. The increase in clad deformation will leave more space for fuel fragments to relocate. Since the fuel fragmentation is clearly associated to burnup, with finer fragments at higher BU. a pellet stack slumping is likely to occur after burst resulting in more or less compact filling of clad balloons. A major question is then what could be the impact on peak clad temperature and final oxidation ratio of the local increase in lineic and surfacic power and of the associated local decrease in fuel-clad gap ?

It should be emphasized that this question is particularly important for U0 2 fuel at beginning-of-life and for MOX fuel at end-of-life where power generation is not reduced unlike for U0 2 fuel.

2.1.3 Early evaluations The State-of-the-Art Review performed by P.D. PARSONS et al. for CSNTJPWG-2 and published in 1986[5], thus after PBF-LOC and FR2 tests completion, reports two calculation studies addressing the impact of fuel relocation on peak clad temperature.

2.1.3.1 Cakulatonsin Sweden The first one was conducted in 1978-79 in Sweden by Bergquist[6l, within the frame of the ECCS evaluation for the Ringhals 3 power plant. It consisted of a series of parametric transient calculations, performed with the TOODEE-2 code, so as to evaluate the response on clad temperature with/without fuel relocation in the balloon after rod burst.

The main assumption for fuel relocation was a uniform redistribution of fuel in the deformed meshes of the balloon with a density taken to 50% of the theoretical density in the base case, a fuel thermal conductivity of 0.6 Wml1K and heat exchange between fuel and cladding dealt with an exchange coefficient of 5000 W/m 2iK. In the reference case, the hoop strain of the most defonned mesh of the balloon was 42%. and the peak clad temperature (PCI) without fuel relocation did not exceed 20000 F

(=10930C).

Calculations with fuel relocation showed that the evolution of clad temperature in the ruptured mesh is essentially dependent of the power rating in that mesh, in relation with fuel average density:

- at 43 kWMm linear power (Fq = 2.09). the clad temperature evolution remains of classical shape, with a PCr around 2050"F (1 l2lC);

- at 47.87 kWlm linear power (Fq = 2.32) the clad temperature evolution exhibits a significant rate increase around 2000°F with a subsequent temperature escalation after 45 s in the transient;

- at 43 kW/m linear power, but with a 60% fuel theoretical density in the balloon (instead of 50%),

the clad temperature evolution again exhibits a significant rate increase after 45 s, reach of the 2200°F limit around 62 s, followed by subsequent temperature escalation.

Although these early calculations had to be considered with large reservations, it may look surprising that they were not much discussed nor compared to counter-calculations, in consideration of the possible importance of calculated trends with respect to safety analysis.

2.1.3.2 INEL Calculations The second evaluation was a steady state thernal analysis of a ballooned fuel rod following a fuel redistribution, the amount of which based on PBF-LOC tests results. This analysis was performed by T.R Yackle[71 as a response to a NRC request ; it is also mentioned by Broughton in the PBF-LOC3/LOC5 test report (I].

Fuel redistribution in the ballooned cladding is modeled by a series of up to 7 concentric rings of different width to take account of large particles of original fuel and small particles of additional fuel, neighbor rings being separated by gas gaps. Only radial heat transfer is considered, with a rod power corresponding to ANS decay heat 100 s after scram (- 3% -original power), and a flat radial power profile. A cladding surface heat transfer coefficient, of 60 Wh 2/K. was assumed, a fuel thermal conductivity constant at 2.6 W/m/K and no radiative transfer between fuel particles.

The amount of fuel redistribution has been determined from the results of the PBF LOC-3 and LOC-5 tests. A line fit through the available data of fuel relative increase as function of cladding relative volume increase indicates an average filling ratio closed to 0.65. Three calculations have then been considered, corresponding to clad strain of 0, 44 and 89%. The following table gives the temperatures at fuel centerline and clad outside surface obtained in these three calculations.

Clad strain (%) Tdd (K) Tuo20(K) 0 1095 1180 44 1120 1620 89 1320 2450 For the worst case, with 89% clad strain allowing the redistribution of 160% additional fuel, the.

outside clad temperature is 225 K larger than for the reference case (without deformation), while the corresponding increase on maximum fuel temperature is 1270 K.

The conclusion that was drawn at that time appears presently quite surprising, as it was stated that

'fuel relocation into a balloon (with conditions such those calculated) will not pose a significant problem during a LOCA since both fuel and clad temperatures remain well below the corresponding melting points"...It may be thought that the relatively close occurrence of the TMI-2 accident is likely to explain such shift of concerns from LOCA to Severe Accident issues.

3 IPSN CALCULATIONS 3.1 Reference code and calculationprocedure The French CATHARE-2 code has been chosen as a base tool due to its capability to provide a best-estimate evaluation of the thernal-hydraulic evolution in hot assembly as well as in mean core subchannels. The code organization allows to run stand-alone calculations of the fuel module (the

CATHACOMB module) in order to provide rapidly information about the behaviour of specific fuel rods subjected to a given set of hydraulic conditions ; these hydraulic conditions will thus not be influenced by the behaviour of such specific rods. The hydraulic conditions may be retrieved from a previous CATHARE whole calculation, or may be that of the current CATHARE computation with which the stand-alone fuel module is carried out in parallel.

For the purpose of this study, the calculations were performed as follows:

- in a first step a whole CATHARE-2 computation was run, for a large break LOCA transient occurring on a typical French PWR, with an input deck corresponding to a fresh fuel "mean core" rod and a high burnup U0 2 fuel hot assembly" rod ; this computation provided the hydraulic file used as input in the.following calculations;

- in a second step a series of stand-alone CATHACOMB calculations were run, using the previously created hydraulic file. for a specific hot rod of the hot assembly, and with the inclusion of specific modifications in some fuel routines in order to simulate fuel relocation after burst in the ruptured mesh; these modifications will be briefly described in the following.

3.2 Code version The CATHARE-2 V13L code version was used as starting version, according to the known improvements implemented in this version to calculate the reflooding phase of the LOCA transient.

Slight modeling improvements have been added in the fuel routines so as to compute at the end of each time step the oxidation weight gain (using both Cathcart-Pawel and Baker-Just rate laws) as well as the thickness of a-Zr[O] oxidation layer.; the .former:variable allows a direct calculation .Qf the equivalent oxidation rate ECR, while the latter variable allows to derive the remaining thickness of the central P-Zr layer.

3.3 Basic Input options for the initial CATHARE-2 whole calculation 3.3.1 Accident transient conditions The main assumptions are in agreement with those retained in the Standard Safety Report for 900 MWe French PWRs:

- double ended break on cold leg of the. loop bearing the pressurizer,

- core at 102% nominal power at accident initiation,

- residual power_= ANS71 + 20%.

3.3.2 Fuel rods description Basically, three fuel kqdsinay be described for a CATHARE calculation:

- the mean core rod, with a weight of (N. -I)x N,,.

where N. is the number of assemblies in the core and N,. the number of active rods per assembly

- the hot assembly mean rod, with a weight of Np,

- one (or several) hot rod(s) in the hot assemnbly Each of these rods is described in terms of geometry, cladding oxide thickness profile, power profile.

A typical axial meshing with 40 meshes was chosen.

In agreement with the options retained in sensitivity studies performed at EDF some years ago, the lineic power of the mean core rod was chosen as that of beginning of life (BOL) while only the hot assembly rods were chosen irradiated to 57 GWj/tU ; the ratio F.% of hot rod power to mean core rod power was chosen to 1.28 (as compared to the 1.55 value for BOL case), and the ratio F, of hot rod power to hot assembly mean rod power was kept to 1.05 identical to BOL case.

Prior to the LOCA initiation, the reactor core is then supposed to be subjected to a transient evolution that brings the three above mentioned -rods respectively to: 68.20 kW, 83.25 kW and 87.41 ,W, and with a truncated cosine axial power profile.

The irradiated rods bear an external oxide layer on the zircaloy clad, with a thickness profile typical of 4 cycles irradiation, the maximum thickness reaching 106 pim at 2.79 in elevation. The pellet-cladding gap is supposed to be closed in the irradiated rods at transient initiation. The internal pressure in hot conditions in irradiated rods is significantly higher (-15.6 Mpa) than in mean core fresh rods.

Two hydraulic channels are associated to the mean core and hot assembly rods. The thermo-mechanical behaviour of the hot rod(s) is influenced by the hot assembly channel hydraulics during the blowdown and refill phases and by the mean core channel hydraulics during the reflooding phase.

3.4 Reference case behaviour-(wlthout fuel relocation)

A reference calculation without fuel relocation was first performed for the hot rod of the hot

-assembly. It must be pointed out that a best-estimate treatment of the clad ballooning and burst for the irradiated rod was not searched here: the standard clad deformation and burst models for fresh fuel were kept unchanged in CATHARE However, in consideration of the results of the TACIR experiments (oxidation and quenching tests) on irradiated cladding [8], having clearly indicated that the initial oxide scale was no more protective for high temperature oxidation, it was chosen to suppress the protective effect of the initial oxide scale towards transient oxidation of the clad, that is normally active in the standard oxidation model of CATHARE.

The rod cladding appeared to rupture at.30.2 seconds on mesh 24 (elevation 2.15 m) with a hoop strain of 56.3%. All the following results, unless explicitly stated, will refer to the ruptured mesh

elevation.

Figure l displays the evoution of the fuel centerline and clad outside temperatures: the clad outside temperature rises to a maximum of 9700C while the fuel centerline temperature remains below I 100°C during the heatup phase.

Figure 2 displays the equivalent cladding reacted ECR evolution, as calculated with Cadtcart-Pawel rate law, for the ruptured mesh and the two neighbor meshes. Foi the non ruptured meshes, due to unprotected oxidation on the external face only, the oxidation rate ECR is increased by about 1.7% in

absolute value, while on the ruptured mesh, due to two-sided oxidation, the ECR rises from an initial value at 9.2% to 12.6% at the end of the transient.

Figure 3 compares-the equivalent cladding reacted ECR evolutions, as calculated with Cathcart-Pawel and Baker-Just rate laws, for the ruptured mesh. It can be noticed that both correlations give very close results in the corresponding range of clad temperature. Acceptance criterion on clad maximum oxidation rate (<17%) is clearly well satisfied.

Figure 4 displays the evolution of the remaining thickness of the clad O-Zr layer, showing the sharp drop in thickness (from -520 to -330 Aim) corresponding to clad ballooning up to rupture, followed by a slow decrease corresponding to high temperature oxidation. The final thickness remains just above 300 inm, indicating a fair remaining resistance to thermal shock embrittlement, with reference to embrittlement criterion proposed by Chung and Kassner[9J, while the handling limit proposed by these authors is just reached.

3.5 Fuel relocation case 3.5.1 Basic assumptions and modeling options

With reference to the FR2 experimental results discussed before in section 2. .1, we assumed that fuel pellets crumbling and relocation occurred immediately after the cladding burst, leading to a partial

filling of the inside volume of the ballooned cladding ruptured mesh. This volume was calculated as that of a cylindrical volume with clad inner radius at burst and mesh height. It was then assumed that this ruptured mesh volume was filled homogeneously with fuel fragments up to an user's input filling rate (= ratio of dense fuel volume to new mesh volume). A base calculation was performed with a filling ratio of 61.5% corresponding to a value measured in the FR2 experiment ES. Two other calculations were conducted with values of the filling ratio of 40% and 70% in order to evaluate the sensitivity of the results to this main parameter.

The fuel fragments were assinilated to spherical particles with user's input diameter. These fragments are in contact with the cladding, leading thus to a closed fuel-cladding gap.

The effective thermal conductivity of the fuel fragments was derived from the Irnura[IO] correlation and taking into account the radiative transfers between particles according to the Yagi theoretical model. The resulting model, so-called "Imuia-Yagir model, had been implemented in the SF1D code ICARE2 of IPSN after it had been validated against Sandia DCO experiment.

According to the results of the TAGCIR experiments mentioned in previous section, the protective effect of initial oxide scale towards LOCA transient oxidation was again suppressed in all the following calculations involving fuel relocation.

3.5.2 Results of the base case (61.5% filling ratio)

A basic calculation was .performed with a filling ratio of 61.5% corresponding to a local value measured on a sample taken from the ballooned region (with 67.5% total circumferential elongation) of the the FR2 experiment E5. The particle diameter was taken as the average size determined in FR2 experiments, i.e. 2.7 mrm .

Figure 5 displays the evolution of the fuel centerline and clad outside temperatures: the clad outside temperature reaches a maximum level around 11000C while the fuel centerline temperature remains below 1200'C during the heatup phase.

Figure 6 shows the evolution of the oxidation rate ECR, as calculated with Cathcart-Pawel rate law, for the ruptured mesh and the two neighbor meshes. It appears that the oxidation rate rises from 92%

to near 18% on the rupture/relocation mesh, while the increase in ECR does not exceed 2% on the neighbor meshes. Figure 7 displays the evolution of ECR values at ruptured mesh, as calculated with Cathcart-Pawel and Baker-Just rate laws ; compared to the corresponding curves for the calculation without fuel relocation (figure 3) a clear distinction can now be made between both evolutions, corresponding to the increase in clad temperature. The maximum value of ECR calculated with Baker-Just rate law is 19.4%, thus exceeding the current acceptance limit.

Figure 8 displays the evolution of the remaining thickness of the clad j3-Zr layer, showing the same sharp drop as in reference case corresponding to clad ballooning up to rupture, followed by the decrease corresponding to high temperature oxidation, with a final thickness just below 250 pm.

Since the maximum oxygen content at this temperature level remains below 0.9 wt %, it appears that the Chung/Kassner criterion would be satisfied for the thermal shock limit but not for the handling limit.

3.5.3 Sensitivity to the balloon filling ratio Finally, comparative calculations were performned with filling ratio values of 40% and 70%, the latter value corresponding to the fuel void fraction measured by gamma decay counts in some PBF.LOC experiments. The particle diameter was kept at the same value s in the previous calculation (2.7 nun).

Figure 9 displays the evolution of clad outside temperature with increasing value of the balloon filling ratio: for 40% filling the temperature level is similar to that of reference case without fuel relocation whereas peak clad temperature reaches 11440C with 70% filling.

Figures 10 and 11 display the evolutions of the oxidation rate ECR. as calculated with Cathcart-Pawel, and of the remaining thickness of the clad P-Zr layer respectively, with increasing filling ratio

values. For highest filling value, the total oxidation rate ECR reaches 19.7% (22% with the Baker-lust rate law) while the A-Zr layer remaining thickness remains near 230 pm at the end of LOCA transient.

4

SUMMARY

AND.CONCLUSIONS LOCA transient calculations have been performed with an adapted version of the French code CATHARE-2 in order to evaluate the possible impact of crumbling and relocation of irradiated fuel in the ballooned region of a cladding after burst.

Focus has been put on the sensitivity of-peak clad temperature and final oxidation rate on the filling ratio of the ballooned cladding with fuel crumble.

The calculations do not intend to give a best-estimate view of the detail behaviour of high burnup fuel rod under LOCA transient. In particular, the thermo-mechanical properties of irradiated zircaloy were not available for the calculation of cladding def6rnation and burst with irradiated material.

The results indicate that for fuel relocation- in the ballooned region with a filling ratio up to the values obtained in FR2 or PBF-LOC experiments, the peak clad temperature may increase significantly, but still remains below the ECCS acceptance limit (1200 0C) on PCI.

On the other hand, the maximum cladding oxidation rate exceeds the 17% acceptance limit when the initial (in service) oxidation rate is cumulated-with the transient oxidation rate and when the initial oxide layer is assumed -no more protective for transient oxide growth. However, alternative embrittlement criteria based -on residual thickness of ductile metal, such as the Chung and Kassner criteria, indicate a fair remaining margin to the thermal shock embrittlement limit, whereas the

handling embrittlement limit. appears exceeded.

The results -of the present.study give some insight into the possible impact of the crumbling and relocation of high burnup U0 2 fuel in a LOCA transient, a phenomena that was observed previously in in-pile experiments and which might significantly affect the late evolution of accident transient and associated safety issues. It must be pointed out that results of corresponding calculations with low burnup UO2 or high burnup MOX fuels would have been more severe with regard to acceptance limits.

The results of the present calculation study- give some support to the need for further experimental data, to be provided by irradiated fuel LOCA experiments involving fuel relocation. A.best representativity should be obtained with in-pile experiments, so as to maintain heat generation in fuel fragments whatever their displacemient may he.during the: relocation process: Such expenments are currently under planning by Halden Readtor Project and by IPSN.

REFERENCES

1. J.M. BROUGHTON et al., PBF LOCA Tests Series. Tests LOC3 and LOC5 Fuel Behavior Report.

NUREG/CR 2073, June 1981.

2. J.M. BROUGHTON et al., PBF LOCA Test LOC6 Fuel-Behavior Report.

NUREG/CR 3184. April 1983.

3.. Eli. KARB et al., LWR Fuel Rod Behavior in the FR2 in-pile Tests Sinmlating the Heat-up Phase of a LOCA.

KFK 3346. March 1983

4. M. BRUET et id., High Burnup Fuel Behavior during a LOCA Type Accident The FLASH5 Experiment.

IAEA Technical Committee Meeting Behavior of Core Material and F.P. Release in Accident Conditions in LWRs, Cadarache, France, March 16-20,1992.

.5. -P.D. PARSONS, E.D. HINDLE, CA MANN, The Deformation, Oxidation and Embrittlernent of PWR Fuel Cladding in a Loss-of-Coolant Accident. A State-of-the-Art Report.

NEA/OECD-CSNI Report 129, December 1986

6. P.. BERGQUIST, A Parameter Study Conceming the impact on the Calculate Peak Clad Temperature of a Redistribution of the Fuel After Cladding Swelling and Rupture.

FV-79-0017/2

7. TR. YACKLE, 'Steady State Fuel Rubble Thermal Analysis", HIZ-317-80 correspondence from HJ. Zeile, EG&G, Idaho, to RE. Tiller, DOEIID, Idaho Falls, September 29, 1980.

8. C. GRANDJEAN, R: CAUVINC. LEBUFFE, N. WAECKEL, French Investigations of High Burnup Effect on LOCA Thenrnomochanical Behavior. Part Two: Oxidation and Quenching Experiments under Simulated LOCA coinditions with High Burnup Clad Material.

24th Water Reactor Safety Information Meeting, Bethesda, Md, USA, October-21-23, 1996.

9. H.M. CHUNG, T.F. KASSNER, Embrittlement Criteria for Zircaloy Fuel Cladding Applicable to Accident Situations in Light-Water Reactors: Summary Report Argonne National Laboratory, NUREGICR-1344, ANL-7948, January 1980.

-10. S. IMURA, E TAKEGOSHI, Effect of Gas Pressure on the Effective Thennal Conductivity of Packed Beds.

Heat Transfer Japanese Research, Vol 3, NoA, p. 13,1 97 4.

L24(M BWU) 1R-24 14.................... ....... ................ ~JI t~at --..

... .:..- :-.. :.... .. ..

12 smaZ IMn Ii

, ...............

to.

I. ~~....... ..... :

I7. ... ... ,.........

@ l i B 9 ISM

. . . ....... ..... @_......... . .. .

I * - 6.,

S. d...:..

...... .: .. .. ......-... s....

. 4.

30.

n. 1.

S.

e.

1 . B1 ...........

I. gD. 3 2. IC M.) IS 1. . -e M. 0. W. ID. 33. 12. IC. 14. W M.

Lucw &d -Lf.O HA Rod. QlGU 57 jWA ] w Od LOCA. Hot Rod. eSU =S7 MAJ ICA lb idA gm atdwWkiait hWad& 1V1.3 pc Figure 1. Figure 2.

7.. gVs lcbw Imum)s ETA24

\:::::::: Is 3...t ,... ' .....

.... ...- ...'.......... .. .. '. _ . . . ... . AE o G. .. .. .. . ...

. ..

4.M r...

.. ., .. ;. .. ...... . ...

3M. 25. . .. .. .. .

. .. . .. . . . . .. . ..

~~~~... XC '..2....... 1A'. 14 18a3 300 RI 3g. -: -; , :-' ..............

-:--;-: - ;--

2. .... A.v~..... .. *........ s... .@ .. .. w 3 23 40 08 XC 1 0 13 4 SO. .... e......... ...... .... :...- ....

3.....

.. ............

.......

-. E .4 .4 .E . ID. 1 .M 'f 8. n. ......

2..0L W ......

3 I. .. *M.. M. 2M WW&akLOCA. Ht McLW 57 MVJ/WCAK UkL7 FkLOC t Rod. 6U 1JI

. 9 cArDNE2 NOW gdtR ca L NDp kCI e ee Of -,j cod& V,, IN WI Adi M No wfcm effi of lA M* VII Pc Figure 3. Figure 4.

4-:-

214 W IZ b-e:_S~a BMAZ 16.

14.

i 12.

i

, ,

6. ' ' *'

t.o=

2 ' ~

-_ . Z. C. so. . 1la. 13. 14I 1m. 1 soD.

lB LOCA. Ht Rod. p = 57 6dtW . CA1UM2 Wk. M Rod. Rw =9MAU 57.ANE./

Fud wirabmn nm i meSh .61.Z of ulIocn vokm! TLl pUI F ml~ Di mn L eh, MS ofhl1 WU= i z e,,,

Figure S. Figure 6..

24. 0 BE24

3 . . = b 14.

16 14 3D.-

TI lE I CUE E..34D06D 3 203 3 . 163. 13. X}. ot 3). 40 I 13 lo 4 . I 3 13.. 2E.

lB LOCA. Hot Rod. fm = Y Wdf CA- M2 LB UC. Ht Rod. &imLq = 9 GVdMUdCAW2 W ~mkmfin nm. sh , 61i ofhllo volteie Vi , Fri.mlaibn nW. na;h 61.5Z of W11. vobnme 1. p Figure 7.F Figure 8.

TatL24 CPiIM4

mt C

,fl&2 6 IULB.

LB LOCA. tht Rod. &niup = S7 G~d/tti CATH4JE LB LOCA. Hot Rod. Gumup = 57 G~dftUi CA1HAOE Fwl zaIocatbn in D4um mesh. Whso of lfi1IkW raIA Pb Vum!mlocatkn in navurmd mesh. klftmm of Wtvkstatlo jl n "i

Figure 9. Figure 10.

&h La= tR rms (wm)

MD-ESETA24 Hfasnms 4W.

~3M.

301W n.e....

... . .... ... .. .. ... . .. . ... . .. . .. I UD- 2D. 40. W. 00. 100. 120. 140. UD. IM. 280.

LB LOCA. Hot Rod. &+/-rriq, = 57 G~d/WJ CAW Fuel mlcatkmhin nauxm! wesh. Wraa of CUMiwbo VO P figure 11.

I l Ir I I I I r r I r I I I I

- -0 - -

Burntu If v.asy e; POS!

C. GRANDJEAN, G. HACHE, C. RONGIER IPSN, Cadarache,France OECD Topical Meeting on LOCA Fuel Safety Criteria, Aix-en-Provence, March 22-23, 2001

I II tI((r I r I t I[ I? I 111I I I Hfigh Burnup Fuel LOCA Calculations to Evaluate

___-_the Possible Impact of Fuel relocation after Burst BACKGROUND (1)

Mainfindings were provided by the results of: PBF-LOC, FR2, FLASH5 experiments:

>P fuel relocation was observed in all irradiated rods as a slumping of fuel fragments from upper locations into the swollen region

> fuel movement initiation occurred at burst in E3 and E4 FR2 tests

> fuel motion, (favored in FR2 due to non closure of gap) is supposed to be counteracted by a: tight fuel-clad bonding 4 bonding was not observed on FLASHS (50 GWdlt) despite low clad ballooning (16%)

Important issue:

Fuel relocation t* increases local power and reduces drastically pellet-clad gap l_ (b impact on Peak Clad Temperature and Oxidation Rate ?

importance: U02 at BOL, MOX at EOL C.GmndJenn *OECD 7opicalMeeting on LOCA Fuel Safety Criteria Akxen-Provence, 22-23 March, 2001

[ I I I I I tr r r r r r I rI I rPe-Transient Posf-Trinslent Neutron radiographs of rod Fl (burnup 20 000 MWd/tU).

Comparison between status pre-transient and post-transient

I IirI I f r I I r I II High Burnup Fuel LOCA Calculations to Evaluate the Possible Impact of Fuel relocation after Burst BACKGROUND (2)

BERGOUIST (Sweden, 1978-79): Parametric transient calculations with TOODEE-2 code

- impact of fuel relocation after clad ballooning and burst / reference case without relocation

- main sensitivity to peaking factor (Fq) and density of relocated fuel (Preloc)

Results: ref case (Fq =2.32) w/o relocation -4 PCT - 2000OF = 10930C relocation, Fq =2.09, Preioc = 50% Ptheor - PCT - 20500 F = 1121 0C relocation, Fq =2.09, preloc 60% Ptheor -4 Tclad 71 above 21500 F and subsequent escalation

YACKLE (INEL, 1980) : Steady state thermal analysis of a fuel rubble in clad balloon

- fuel relocation ratio. extrapolated from PBF-LOC experiments

- relocated fuel modeled as a series of 7 concentric nodes with stagnant steam gaps

- power: ANS decay heat at 100 s; flat radial power profile; Results: worst case: 89% cladding strain 4160% fuel redistribution W Triad = 1320 K (+ 225 K) and Tcentfuel = 2450 K (+1270 K)

Conclusion:

fuel relocation = not a problem since both Tare well below melting points!!

C. Granrdjean OECD Topical Meeting on LOCA Fuel Safery Criteria Aix-en-Provence, 22-23 March, 2001

I I I I I l It I rl I I I I I I I High Burnup Fuel LOCA Calculations to Evaluate the Possible Impact of Fuel relocation after Burst.

IPSN Calculations: Large Break LOCA calculations with irradiated fuel rods Code version CATHARE2 V1.3L with specific modifications to:

)> simulate fuel accumulation in the ruptured mesh after burst

> calculate oxidation rate ECR and P-Zr remaining thickness CalculationProcedure

1't step: whole CATHARE2 LOCA computation run without fuel relocation

% provides the hydraulic conditions for following calculations nd

+2 step: stand-alone fuel module (CATHACOMB) calculations

> under imposed hydraulic conditions retrieved from previous step

>' without fuel relocation (reference case)

> with simulation of fuel relocation after burst, according to user's input characteristics for the filling of the clad balloon C.Grandjean OECD Topical Meeting on LOCA Fuel Safety Criteria Aix-en-Provence, 22-23 March. 2001

I I I I I I r I r r I r r r I I I High Burnup Fud LOCA Calculations to Evaluate the Possible Impact of FEel relocation after Burst Whole CATHARE-2 standard -calculation:

> large break LOCA (double ended break on cold leg)

>~ mean core rod :fresh fuel

> hot assembly rods :irradiated fuel 57 GWd/t Photwmd /P ~cor 1d2'.28 (1.55 it BOL) ; Fq 1.94 (2.35 at BOL)

> at accident initiation:

core at 102% of nominal power, a cosine axial powcr profile, a pellet-clad gap closed in irradiated rods, N rod internal pressure P't15.6 Mpa C.Grandjean OECD Topical Meeting on LOCA Fuel Safeyr Criteria Aix-en-P rovence, 22.23 March. 2001

I 1 I I I. i I I I r r r r t r I r I High Burnup Puel LOCA Calculations to Evaluate the Possible Impact of Fuel relocation after Burst Stand-alone CATHACOMB calculations:

> suppression of the protective effect of initial oxide scale (according to the results of TAGCIR experiments on irradiated cladding)

> homogeneous filling of the balloon Pilling ratio (= 1- void ratio):

. base case: 61.5% ( value measured in FR2 experiment E5)

sensitivity.study: 40% and 70%

> fuel fragments assimilated to spherical particles in contact with the cladding wall (res. gap = 1gm) particle diameter: 2.7 mm (- average value in FR2 experiments)

> thermal conductivity derived from a debris bed model, including convective and radiative heat transfer between fuel particles:

IMTJRA/YAGI model, validated against the DC1 experiment (SNL)

C.Grandjcan OECD Topical Meefing on LOCA Fuel Safety Criteria ALixen-Provence, 22-23 March, 200)

f II I I I r[ I r r r r r I r V -

30*.

1} . .....

. I a'

. . '; t I ..

. ' **

.a

.

' a,

.

, I

.

..

TRQD.24 .

t a^

a ss a a

a

.sat a .

a .

_.;.__;_

-,

a a -

a a

a 4

a

..........

a a l;j_

ECIRtot I . a I a. a a. a a I,

D .......b.........r.....l... ..... .... l.....

. . .. . . .. ',

. l a I _ Ia _L In 1.- , .

a ,

a, a a AUCm 8 Aa

. . ........

. .. .. . .. . . III'.I.I...

. . .. . . . . . . . . C, 200 ................. ' - .' ....

20 .. ...J .......... 490. a. . 2M, 1.

a. a

.

a.

.a

.

. .

.. . .. . .... .

Va . ' * ,' '. 2.. ................

.0 .

a . a. . . ,,,l ..

M O. . . .. ... 4... . .. . T. . ... . . . a . . . *

a 6 .. .a....4....I.. ......

, ,

I. 4 4 I I

{ ......... .. t ........... .. -' I '-- t- ---

5.a . . . . . .

sor0.. 40...,..-0. 0. 100. 12. 14. 1.. 10. 2 . 3. , i . a . a.

U. 20O 40. 611 60. 1DO 12 140. 160. 1T00 203 Lm B kLOCA. ot Rod. 5 WiU CAqCME La Brek LOCA. Hot Rod. JU=579 G/W JCATAW NO Nel~relocation.. No pmbtcM~ affet of k-iNWa oxkde ViA pdv No fee vloceton. No pmteUe effect of ilbt oxid Vi.1 p&

I I A oct I ast~ ~ &t - Ai

besk `22, it .. is m a es S

1 I . 1 I I I I I

I 1 r I I I I -

5.IERot WX) eLkL Tdrftss (ff}m S

I. EcR24 * .

I ..

.

....

. I EBETA24

. .. . . ..

. . . . . . . . .

. . -.. .*.. . *. .a..

.13. . . . . I. . . .1. . ... . . . ......

. _

. I

'

.

. .. ... I

§ ........

.. ... .

A~u

12. ~~~~.... ......... !!'

Ii.. 9 EMh . aI * *a . - *

. I . . .. . * . I 10.

.. .. . . . . . . .'.

I *I .

a

.9.

B.

a a a a a I . I ' 3tv,' ... A.....b.. .J... -....

7. . . . . .. . . . SOD. .. ... ..... ... . .. ... . s..........

a

.

,a .,

a '. a

. .

.

..

.

.6.......@.... J....j..................J..............t......

I a

, . . . . - * . a

.5'1.'._!......L..§._....I....I..__..

T ( ff D. 20. . ,0. S.a .0. . 0.

. 2.* 14., ,60 I .0 20

a, a~e a I a* .. 'a a

5. 2 Wv a a. a a a a aB§

4. .

I.

I I I a

a a a

a a

a 3.

a' . a

. a a I a I a MS .

n1 n a..... a....................a ....

a .

AI D. 20. 40. 60. SO. lD. 1. 140. lCD. 10. 20. * . 20 4 0 W. W. l. 1 20

. 140. lC . 1 0D 2 W.

la Bred LOCA. Hot Rod. IL 57 GVI/W CAHA Large Bteak LOCA. Hot Rod. BU = 59 GVj/W CACAI Uatn. leffect of Wob .NVia No e mieocaon. No promtmve effect of NW ox VIA "

Ir I r r t r r t I I CcSSDS ZZ25)* a I .

-2.1 tcR~tot 22i . . .. .. a i . .... . .. .

20 .. . .. . .. . . . .. .. . .......

..............- Amai Cmh3b~

A A MoJ. . .......

12. .... .... ---

10. .

. 2M

.. ... . ......... .. 40 1 . (0 2 Di 20. 40. ( ... '00 "oI'm 1.

..... 1(0. 2(00 LB LOCA. Hot Rod. &itnup = 57 G~d/tU 'CANAE2 LB LOCA. Hot Rod. Bainp =59 GWd/tU CAWHM~

Fuel rlocation in nupt. mesh : 61.5% of baloon volum III pdv Fuel relocaon innop. mesh i 61.51%of baloon voltme VIA p&i

I I I I I I Ir r r r [rf r rI I ECRtot (X) I till L,

LAab Rl ss (mbcbrn)

. . ..... I I ECR24 , . . . . . . .

EBETA24 22 ' a . s5. .a ..

20 . . .. . .. . .

&Ca~ut ssr. , . ., a . , , ALbI^a t8 . ..... .. . . 8akJsl IZ I , *

a . ,

14. . . . . . . . m. *.. G1. ,

a\ , a a . ,

14. . . *. ; *. . a .........

. .'

J,,e....4... . ....... J....*......

. , I I Ia . . I I8 ...

.... . . .. .j. .. ...- . s. .... .... . .......

. a a va I I ., a 300. I a . . . a 1......................,

OU I . a ., a a, , a

... a.r. .I,,.. .. , ,... ....

H. ...... @,..........Z.. .. .... '... .. '.. 2m.

2 .. a . . . . .

150 . a a Ia Ia a 100.

02*~~.. * .. a. ....... *..4 s.

111 ...........................................

D. 20. 40. ') 1 1 0 . 140. 150.I1 . 20. vU. 20. 40. 60. 0. 100. 10. 140. 15. 150. 20.

LB LOCA. Hot Rod. Birmp = 57 GVd/tUl CARE LB LOCA. Hot Rod.. mp = 7 GWdItU jCATHE2 Fud rIocation innpt. mesh',1 of booi vohfe of.Sx jR Vpdv lFedoeton in* I.5Z.b olom meshI, cvol VlI

140.cACIJ (CELSRIS, Z2IS)

TCL-24 1300. , ;

A fill4OZ 1200. Nz~7t B fgl=6i.Sz C fmU=ZOZ 7

600. . ---. -M

,----,----T-- - - -

-r- - - --- - - - --- - - - v - - - --- - - - -l-e l M*. I -- - - -- -

100. - - - - -

W.- .- - - - - - - 7 -- - - - --- - - --.-. -

320. 0 , ,

D0., .0.

.10 . 4. . 8.

0 M.

0 .I.10 1.. 6-8.

I.0.

00.

LB LOCA. Hot Rod. Bumup = 57 GWd/tU CATHARE2 Fuel relocaUon in ruptured mesh. Influence of filling ratio V1.3L priv

I I r I I I I I I I I I I Y -

cRMt (XJ h.I

,. . . .

OJR24

, . . .

A1fa a MA.tS Cox 5.

, .... .a. .a.. .

, .1 . . . . .

10.

S.

AI I UD 2. 0. 0. 0. . 0. 14. . 1,. 2.

LB LOCA. Hot Rod. Biunup = 57 GWd/UW CARE Fuel relocton in npmd meh. Influe of fft mUo V1. priv

I I f I r f Ir t I I I I I I High Burnup Fuel LOCA Calculations to Evaluate the Possible Impact of Fuel relocation after Burst

LOCA transient calculations have been performed with an adapted version of the French code CATHARE2 in order to evaluate the possible impact of crumbling and relocation of irradiated U0 2 fuel in the ballooned region of a cladding after burst.

Focus has been put on the sensitivity of PCT and final oxidation rate on the filling ratio of the ballooned cladding with fuel crumble.

+ Results indicate that with a filling ratio up to values obtained in FR2 or PBF-LOC experiments:

> the PCT increasessignificantly but still remains below the 12000 C limit;

> the oxidation rate ECR may exceed the 17% acceptance limit when initialand transientoxidation are cumulated and when initialoxide is considered no more protectivefor transientoxidation;

> alternative embrittlement criteriabased on residual thickness of ductile metal, such as the Chung and Kassner criteria,indicate a fair remaining margin to the thermal shock embrittlement limit, whereas the handling embrittlement limit appearsexceeded; C.GrandJean OECD Topical Meeting on LOCA Fuel Safety Criteria AL&ren-Provence. 22-23 March. 2001

I iI II 1 I I I I I I I IHigh Burnup Fuel LOCA Calculations to Evaluate the Possible Impact of Fuel relocation after Burst

The results of the present study give some insight into the possible impact of the crumbling and relocation of high bumup UO2 fuel in a LOCA. transient, a phenomena that was observed previously in in-pile experiments and which might significantly affect the late evolution of accident transient and associated safety issues.

Results of corresponding calculations with low burnup U0 2 or high burnup MOX fuels would have been more severe with regard to acceptance limits.

These results bring some support to the need for further experimental data, provided by irradiated fuel

LOCA experiments involving fuel relocation. A best representativity should be obtained with in-pile experiments so as to maintain the heat generation in fuel'fragments whatever their displacement may be during the relocation process.

C. Grandjean OECO Topical Meeting on LOCA Fuel Safety Criteria Aix-en-Provence, 22.23 March, 2001

NEA/CSNI/R(2001)18 Discussion:

Comment byM.EI-Shanawany: The UK carriedout similaranalysis (1988) using the computer code BART which included a number of models such as the grid rewening effect that were not taken into account in your paper.

The UK analysis used informationfrom a number of experiments such as KfK, PBFUKAEA. HALDEN and URN The analysis indicated that the peak clad temperature may increase but the temperature did not exceed the 1204 C limit.

Hence, it is not clear what is new in IPSN's analysis, and how your analysis is adding value to our understandingoffuel pelletfragment axial relocation.

(REF: Calculationsof the effect of pellet fragment axial relocation on the peak- clad temperature duritg a loss of coolant accident in a pressurisedwater reactor. K. T. Routledge, M. E-Shanawany & D. Utton.

Second UK National Heat Transfer Conference, 14-16 September 1988, Universityof Strathclyde. Glasgow )

Question from the audience: Was a rupture induced improving of cooling taken into account?

Answer by C. Grandjean: The improving of cooling associatedto clad ballooningand ruptureis not taken into account in the standard version of CATHARE, nor in the version modifiedfor the calculationspresented here. Such influence may however have been taken into account in some calculationsperformed by vendors with their own evaluation models.

Question by R. Meyer: What assumptions made a 10-times difference in peak cladding temperaturebetween your calculationsand the previous papers?

Answer by C. Grandjean: IPSN calculationsdifferfrom EdFcalculationson some options'in modelling, particularlyheat transfer in post-DNB conditions, resulting in different burst time and burst strain, and on optionsforfuel relocation in balloonedarea, mainlyfilling ratio and residualgap (or not) betweenfuel fragments and clad.

G. Hache added: There was too much azimuthal temperature variationin the EdF model.

Question by H.M. Chung: Tight pellet-cladding bonding, commonly observed in high-burnup cladding, is likely to strongly influence the degree of azimuthal temperature variation at burst, and hence. burst size, wall thinning, susceptibility to thermal-shock fragmentation. and fuel relocation. Would

.

-NEA/CSNIIR(200 1)18 the degree of azimuthal temperature variationat high burnup be largeror smaller than at low burnup, and why?*

Answer by C. Grandjean: The azimuthal temperaturegradientis reducedfor irradiatedfuelas a result of fuel fragmentation and relocation associated to cladding creepdown -during reactor normal operation. Fuel relocation is supposed to be a bit more compact at high BU with possibly fine fragments interspersedamong larger ones. However, fuel rearrangementhas been observed to start with early fuel conditioning and the resulting effect on azimuthal temperature variationmight not be much largerat high burnup than at low burnup.

Comments by J. R. Jones: The assessments of heat transfer in the blockage region of the ballooned fuel assembly is sensitive to the dynamics of entraineddroplets and in the UK we have found it necessary to depart from the mean-diameter approach in favour of a multi-group representation of the droplet spectrum.

In the early 1980s work was reportedby Garlick et al. on the observation of relocation of fracturedfuel pellets in simulated LOCA conditions. The degree bf pellert relocation around the time of burst was modest, and certainlyless that the reported simulationsassumed.

I question whether the relocation observed post test in PIE had occurred laterthan the time of interest.

Let me pointed out that pellets can be conditioned by rampling to high reactorpower, priorto atest and relocation can be. examined without.the need to achieve high fuel burnups. This was confirmed by Dr. Wolfgang WIESENACKofHalden-who did thisfor the IFA 54x test sceries.

Answer by C. Grandjean: In the FR2 experiments, only two tests (E3 and E4) have been instrumented in such a way to allow to trace the inception of fuel relocation. These two tests have clearly demonstrated that the fuel stack collapse startsat the time of claddingburst.

Comments by C. Vitanza Halden LOCA tests with internaland external thermocouples showed that it was difficult to have a completely uniform azimutal temperature distribution. Even if one tries to get very uniform boundary conditions at the circumference, there are always small azimutal temperature differences which tend to reduce the effect of balloning.

Exhibit 5

!

I L

I i

I c

Guillard, V., et. al., "Use of CATHARE2 Reactor Calculations to Anticipate Research Needs," SEGFSM Topical Meeting on LOCA Issues, Argonne, May 2004

SEGFSM Topical Meeting on LOCA Issues Argonne National Laboratory, May 25-26, 2004.

USE OF CATHARE2 REACTOR CALCULATIONS TO ANTICIPATE RESEARCH NEEDS V. Guillard', C. Grandjean, S. Bourdon, P. Chatelard Institut de Radioprotection et de SOiret6 Nuclkaire IRSN/DPAM: BP3 - 13115 St Paul-Lez-Durance, France valia.guillard(a)irsn.tr ABSTRACT To analyze the consequences of the introduction, in Nuclear Power Plants, of advanced fuels at high burn-up, decided by most of the utilities in western countries in order to reduce the fuel cycle costs, IRSN has initiated a research program focused on the study of such PWR fuel behavior in LOCA conditions.

A first step of this program, comprising analytical and experimental parts, has been to identify the main physical phenomena, linked with thermomechanical behavior of irradiated rods in bundle geometry, to be taken into account in reactor safety analysis: cladding deformation and flow section restriction in bundle geometry, mechanical interaction between neighbor rods or structures, axial extension of balloons ; cladding oxidation and secondary hydriding ; fuel fragmentation and relocation, balloons filling rate and FP release ; fuel rods thermal behavior in bundle geometry during reflooding conditions, rewetting of the claddings around ballooned regions with fuel relocation mechanical resistance of irradiated claddings in post-quench conditions.

This paper summarizes an analysis of sensitivity calculations performed with CATHARE2 "Best-Estimate" code, used in France in the frame of realistic methodology to evaluate safety margins.

The objective of these calculations is to point out, among parameters affecting last-mentioned phenomena, those for which taking into account basic uncertainties lead to important uncertainty on global code response (Peak Cladding Temperature, oxidation rate ...). That is the case of fuel relocation phenomena, whose impact is highly dependent on parameters such as, in the example of LB LOCA transient, cladding radial and axial deformations in bundle geometry, burst criteria, balloon filling rate, thermalbydraulics around balloons. A lack of knowledge on theses parameters for irradiated U02 and particularly MOX fuel may lead to reduce safety margins.

This study may provide some elements to identify future research needs to complement present experimental data base, reduce uncertainties and develop more realistic calculation models, which may better fit the thermomechanical behavior of advanced irradiated fuels.

1Corresponding author

~

, .R

,:iN T S'1

,: . , ON O.., I USE OF CATHARE2 REACTOR CALCULATIONS TO ANTICIPATE RESEARCH NEEDS V. GUILLARD, C. GRANDJEAN, S. BOURDON, P. CHATELARD Institute for Radiological Protection and Nuclear Safety (IRSN)

Major Accident Prevention Division I Fuel in Accident Situations Department BP3 - 13115 St-Paul-Lez-Durance Cedex - France Content of the presentation

> Introduction

> Main physical phenomena to be modeled

> Main hypotheses of the calculations

> Main results of the LB LOCA calculations

> Conclusion and perspectives valia.auillardOirsn.fr ANL, May 25-27, 2004 I I I I I I i I I J I I J I I II

UGeneral Background 4 Xoi w rInn 4 o ~~ OCA-P oram: Al.',,<

Pfr tud ofPWM'r06d- .... :_:_..... . >, I>Of fue1ea.s O hrmomechancIbhvo c~add~g~y

-w 'Ap 00,

~ .in LOCA conditions E CATHARE2

> French thermalhydraulics system code with CATHACOMB fuel module I

. ..

QJ Objectives

> Identify future research needs for new generation of fuels

> Improve knowledge, models and calculation methodologies D Approach used to reach this objective

> Identify physical phenomena involved in thermo-mechanical behavior of advanced Irradiated rods in bundle geometry (State Of the Art by C. Grandjean & G. Hache)

> Take them into account in the modeling

> Quantify basic uncertainties (CIRCE tool and sensitivity calculations)

> Evaluate global uncertainty on CATHARE2 code response (SUNSET tool)

. .> . nt dfine boundary conditions for further more precise

R icua~ culations and analysis under LOCA conditions valia guillard@irsn.fr ANL, May 25-27, 2004 I I I J II J I I I I I I I I I j I J J

. 7.

Reference:

C. 0 "LOCA Issues Related to Ballooning, Relocation, Flow Blockage-and&Coo16 Ut.';

Main Findings from a Review of Past Expe=04 P Temperatures ('C)

-FFuel I in -Clad TYPICAL LB LOCA TRANSIENT EASY NOT SO EAS BY Fuel fragmentation (debris size, 1200 granulometry, porosity)

Clad coolability around ballooned regions 800 with fuel relocation Clad Fuel rods thermal behavior ballooning during refloding conditions Mechanical interactions between neighbo' 00 rods or structures Post-quench

,.. residual ductility Axial extension of balloon 50 100 150 Time (s)

Fuel U Clad a Bundle

- -t C"C r ,' o ,. I.

valia.giuillardD-irsn.f ANL, May 25-27, 2004 IjII I i I I i i I I I I I I II

,, . ; 9`72, ' NA 70I~L i i"

--

Ir

.t 'v-l 4..k .,

4

.tA ! -4

.4'

.. ahd I Q Calculation of several transients including Large & Intermediate breaks LOCA

-l Use of standard CATHARE2 versions (V1.3L, V2.5)

El Basic uncertainties on CATHARE2 models taken into account

> Consistent with PIRT implications C Zry-4 cladding without hydrogen uptake effect on mechanical properties O Hypothesis linked with the use of irradiated fuel at high burn-up

> Modification of thermal properties

conductivity and thermal capacity laws from SCANAIR code

> Deletion of protective effect of initial oxide layer on transient oxidation

outcome from experimental ANL and TAGCIR program analysis on irradiated Zry

> Introduction of Baker-Just correlation to calculate oxidation rate 0 Initial state of irradiated rods given by METEOR code

> Gap width and pressure, radial and axial fuel power profile, external cladding oxidation profile, cladding thickness ...

Example: HU o of tVie impact of irradiated fuel relocation Btk ,conditions on a PWR900Mwe valia.guillard(irsn.fr ANL, May 25-27, 2004 iI I i I j. I I i I I I I I I I I

CI Main hypothesis S

Balloon filling rate Vf by relocalized fragments fnn_

Relocation mesh

= burst location mesh Upper bound value filled by fuel debris 80 from upper meshes 70' Hea

Filling ratio of the balloon 0 0 x A

= function (balloon size) 60 A

Balloon size derived from S 50 CATHACOMB calculation of 0 mean elongation 40 o PBFILOC-gammascanning

PBFILOC-micrographles Basic calculation value = 61.5 %

Modification of the power factor 30 A FR2 and the fuel mass in the - Upper bound relocation mesh 20' ° EdF calculation x IRSN reference calculation

Considering relocated fuel as 10' + IRSN sensitivity calculation a porous medium (Imura-Yagi model for the conductivity calculation) I' U

.L 20 30 40 50 60 70 80

Flow blockages: impact on T/H Ballooning (%)

and fuel cooling not taken into account I

11 .i

1. I

-I I ,, I-.. -

-

,

I I

,

.4 t' "S"IfIll'", valia.auillardlirsn.fr ANL, May 25-27, 2004 I I I I I I I I I I i I i J I I I

O High sensitivity of the relocation impact on PCT

> Burst at 24 s on stress criterion deduced from EDGAR experiments (aat rupture = 57.5%)

LB LOCA - Irradiated U02 Hot iard cWding temnpertiure ut bum Inte d 1200.0 _ I 000 Fuelrelo cation impact depends on:

/ -rs

. .bwhwhich determines the energy

_re-distnh tibn in the burst mesh

baI;bensiba and filling ratio, 600.0 500.0 500.0 ..

400.0 CLAD BURST 300.0 FUEL RELOCATION 200.0 -

-

Without relacatbn With relocation (filing ratio: 61.3%)

1

+ 1 OC to 1500 C on PCT 1 oo~o Withre~ocC ng0rCton70P0%

- to(fi (depending on balloon filling ratio: 61.5 to 70%)

0.0 20.0 40.0 60.0 80.0 100.0 120LO 140.0 160.0 160.0 200.0 Time [sI RS'u 5 N 'a fn Cte valia.auillardlirsn.fr Iroa g ~nteb _researneeds ANL, May 25-27, 2004 I IIIiII IJ. i I I I I I I I I I

El Fuel relocation impact on ECR and residual ,B-layer

> Burst at 24 s on stress criterion deduced from EDGAR experiments (£at rupture = 57.5%)

LB LOCA - Irradiated U02 LB LOCA - Irradiated U02 Hot rod oxidation ratett burst Inesh Nor rod beta-layerthickness ut busrst nesh 20.0 600.0 18.0 55L.0 50.00 16.0 430.0 -

14.0 400.0 [

12.0 3S0.0

.t 10.0 300.0 U

w IE 6.0- 250.0 200.0 150.0 4.0 1 00.0 2.0 50.0 0.0 _ 0.0 0 .0 20.0 40.0 50.0 60.0 100.0 120.0 140.0 150.0 160.0 200.0 0.0 20.0 40.0 50.0 80.0 100.0 120.0 Time is] rim*e []

.- %to7% onECR 550 to 67km on residual p-layer thickness (9 6pn balloon filling ratio: 61.5 to 70%) ;(dep nding on balloon filling ratio :61.5 to 70%)

i ~ ~ _ on~toneeds - :SE,, F'_ T d_ pR valia auillardOirsn.fr ANL, May 25-27, 2004 I I I I I I 1 i I I I I I I I

El Fuel relocation impact on PCT

.3 > Burst at 18 s on stress criterion deduced from EDGAR experiments (£ at rupture = 58.7 %)

LB LOCA - Irradiated MOX Hot rod cuddin&remperaureat burst incnh9

. . - i ^ .nitiai erngy.

Enihance of fuel relocation impact

- , CLAD BURST FUEL RELOCATION

- Withoul r.eocaton

-With rebcationr I

+60 IC on PCT Time (5]

valia__uillardirsn. ANL, May 25-27, 2004 I I I I I I I I I II I i I I I II

1 Fuel relocation impact on ECR and residual P-layer

> Burst at 18 s on stress criterion deduced from EDGAR experiments (a at rupture = 58.7 %)

LB LOCA - Irradiated MOX LB LOCA - Irradiated MOX Ho! rodl oxidaifon rate at bursn mensh Hotrod beks-hayer resdual tidckesrburji mesh 0

E U

w T.

I Tirne is] Time (a]

4.arge uncertainty 4Higher Impact ,ryarogenuptake effect 'NOdto reduce oE i nd residual P .layer nd r ie h n al p o e i s c rt inties d.,-I reepending on clad alloy valia gui11ard&1rsn.fr ANL, May 25-27, 2004 I I I I I I J I I I J I I I I I I

Use of CATHARE2 code for high burn-up fuel analysis under LOCA conditions

> Example of Large Break transient

> Emphasis on fuel relocation phenomena impact

> Main results: PCT and ECR increased D High uncertainty on global code response due to identified lack of knowledge

> Instant of fuel movement HALDEN?

imposed as burst time in the simulation

depends on clad ballooning/deformation and burst criterion

> Balloon size - HALDEN ?

which is also linked with ballooning/deformation model

> Filling ratio 4- FR2, PBF, HALDEN?

> Relocated fuel properties

fragments size, granulometry, porosity, conductivity, . ANL, HALDEN?,

Bundle effects - Need of integral test Axial extension of balloon Flow blockages

Clad coolability around ballooned regions with fuel relocation iRSHi valia guiIlard§Dirsn.fr aliatuiladfihne ds ANEGMay ANL, 26-27, 2004 May27 20 I I I I I I I I I I I I I I I I I I

' - C. -

O Perspectives

> Modification of clad mechanical properties to take into account hydrogen uptake effect

> Study to be complemented by calculations using CATHARE / FRETA

rods 3D thermo-mechanics

rod-to-rod interactions models

cooling and reflooding models for overall bundle

> Use of NEPTUNE 3D local module for flow blockage cooling calculations

> IRSN plans to develop a new code for fuel LOCA calculations O StatQft &Art C. G. Hache) + Analytical studies m 4idtIficIf knowledge improvements

irr~iat~ed rodbbehavior under LOCA conditions

_o n eton of fuel and bundle models

. Reu Egain Ade~tairs and new margins valia.guillardlirn.fr ANL, May 2527, 2004 IIIIJI I I I I I I I I I I I I I

ML041950059

Contents

Text

Dear Administrative Judges:

REFERENCES:

3.9 REFERENCES

Response

Response

Response

Response

Response

Response

SUMMARY

Conclusion:

Reference:

Navigation menu

ML041950059

Text

Dear Administrative Judges:

REFERENCES:

3.9 REFERENCES

Response

Response

Response

Response

Response

Response

SUMMARY

Conclusion:

Reference:

Navigation menu

Search