Global Nuclear Fuel

ENCLOSURE 2

MFN 15-076

NEDO-33798-A, Revision 1

Non-Proprietary Information– Class I (Public)

IMPORTANT NOTICE

This is a non-proprietary version of NEDE-33798P-A, Revision 1, from which the proprietary information has been removed. Portions of the enclosure that have been removed are indicated by an open and closed bracket as shown here [[ ]].

Global Nuclear Fuel

NEDO-33798-A Revision 1 September 2015

Non-Proprietary Information - Class I (Public)

Licensing Topical Report

APPLICATION OF NSF TO GNF FUEL CHANNEL DESIGNS

NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

INFORMATION NOTICE

This is a non-proprietary version of the document NEDE-33798P-A Revision 1, which has the proprietary information removed. Portions of the document that have been removed are indicated by an open and closed bracket as shown here [[ ]].

IMPORTANT NOTICE REGARDING CONTENTS OF THIS REPORT

PLEASE READ CAREFULLY

The information contained in this document is furnished for the purpose of obtaining NRC approval of the Licensing Topical Report, Application of NSF to GNF Fuel Channel Designs. The only undertakings of Global Nuclear Fuel – Americas, LLC (GNF) with respect to information in this document are contained in contracts between GNF and participating utilities, and nothing contained in this document shall be construed as changing those contracts. The use of this information by anyone other than that for which it is intended is not authorized; and with respect to any unauthorized use, GNF makes no representation or warranty, and assumes no liability as to the completeness, accuracy, or usefulness of the information contained in this document.

NEDO-33798-A Revision 1

Non-Proprietary Information - Class I (Public)

September 1, 2015

Mr. Jerald G. Head Senior Vice President, Regulatory Affairs General Electric-Hitachi Nuclear Energy Americas, LLC P.O. Box 780, M/C A-18 Wilmington, NC 28401-0780

SUBJECT: FINAL SAFETY EVALUATION FOR GLOBAL NUCLEAR FUEL – AMERICAS, LLC LICENSING TOPICAL REPORT NEDE-33798P, “APPLICATION OF NSF TO GNF FUEL CHANNEL DESIGNS” (TAC NO. MF0742)

Dear Mr. Head:

By letter dated February 13, 2013 (Agencywide Documents Access and Management System Accession No. ML13045A460), Global Nuclear Fuel – Americas (GNF) submitted NEDE-33798P, “Application of NSF [niobium (Nb), tin (Sn), iron (Fe)] to GNF Fuel Channel Designs,” to the U.S. Nuclear Regulatory Commission (NRC) staff for review.

By letter dated August 13, 2015, an NRC draft safety evaluation (SE) regarding our approval of Topical Report (TR) NEDE-33798P was provided for your review and comment. The NRC staff's disposition of the GEH comments on the draft SE are discussed in the attachment to the final SE enclosed with this letter. Please note that even though TR NEDE-33798P is proprietary, the enclosed SE is non-proprietary and will be publicly available.

The NRC staff has found that TR NEDE-33798P is acceptable for referencing in licensing applications for nuclear power plants to the extent specified and under the limitations delineated in the TR and in the enclosed final SE. The final SE defines the basis for our acceptance of the TR.

Our acceptance applies only to material provided in the subject TR. We do not intend to repeat our review of the acceptable material described in the TR. When the TR appears as a reference in license applications, our review will ensure that the material presented applies to the specific plant involved. License amendment requests that deviate from this TR will be subject to a plant-specific review in accordance with applicable review standards.

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

J. Head - 2 -

In accordance with the guidance provided on the NRC website, we request that GEH publish approved proprietary and non-proprietary versions of TR NEDC-33798P within three months of receipt of this letter. The approved versions shall incorporate this letter and the enclosed final SE after the title page. Also, they must contain historical review information, including NRC requests for additional information and your responses. The approved versions shall include an "-A" (designating approved) following the TR identification symbol.

If future changes to the NRC's regulatory requirements affect the acceptability of this TR, GEH and/or licensees referencing it will be expected to revise the TR appropriately, or justify its continued applicability for subsequent referencing.

Sincerely,

/RA/

Mirela Gavrilas, Deputy Director Division of Policy and Rulemaking Office of Nuclear Reactor Regulation

Project No. 710

Enclosure: Safety Evaluation

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

J. Head - 2 -

Sincerely,

/RA/

Mirela Gavrilas, Deputy Director Division of Policy and Rulemaking Office of Nuclear Reactor Regulation

Project No. 710

Enclosure: Safety Evaluation

DISTRIBUTION: RidsNrrDpr RidsOgcMailCenter JGolla RidsResOd RidsACRS_MailCTR MPanicker RidsNrrDprPlpb RidsNrrDss RidsNroOd JDean RidsNrrDssSnpb PClifford RidsNrrLADHarrison AMendiola

ADAMS Accession Nos.: ML15233A110 *via e-mail NRR-106 OFFICE NRR/DPR/ NRR/DPR/ NRR/DSS/ NRR/DPR/ NRR/DPR/ PLPB/PM PLPB/LA* SNPB/BC PLPB/BC DPR/DD NAME JGolla DHarrison JDean AMendiola MGavrilas DATE 8/27/2015 8/25/2015 8/27/2015 9/01/2015 9/01/2015 OFFICIAL RECORD COPY

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

GE-Hitachi Nuclear Energy Americas Project No. 710 cc:

Mr. Jerald G. Head Senior Vice President, Regulatory Affairs GE-Hitachi Nuclear Energy P.O. Box 780 M/C A-18 Wilmington, NC 28401 [email protected]

Mr. James F. Harrison GE-Hitachi Nuclear Energy Americas LLC Vice President - Fuel Licensing P.O. Box 780, M/C A-55 Wilmington, NC 28401-0780 [email protected]

Ms. Patricia L. Campbell Vice President, Washington Regulatory Affairs GE-Hitachi Nuclear Energy Americas LLC 1299 Pennsylvania Avenue, NW 9th Floor Washington, DC 20004 [email protected]

Mr. Brian R. Moore Vice President, Fuel Engineering, Acting Global Nuclear Fuel–Americas, LLC P.O. Box 780, M/C A-55 Wilmington, NC 28401-0780 [email protected]

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

SAFETY EVALUATION BY THE

OFFICE OF NUCLEAR REACTOR REGULATION

LICENSING TOPICAL REPORT

NEDE- 33798P, “APPLICATION OF NSF TO GNF FUEL CHANNEL DESIGNS”

GLOBAL NUCLEAR FUEL - AMERICAS

(TAC NO. MF0742)

1.0 INTRODUCTION

By letter dated February 13, 2013 (Reference 1), as supplemented by a letter dated June 10, 2015 (Reference 2), Global Nuclear Fuel - Americas (GNF) requested review and approval of an advanced zirconium alloy, NSF, for application to existing boiling water reactor (BWR) fuel channel designs. NSF derives its name from its primary alloying elements: niobium (Nb), tin (Sn), and iron (Fe). Recent operating experience has shown that channel distortion and associated control blade interference continues to be a major problem in the U.S. BWR commercial fleet. The goal of introducing NSF channel material is to resolve this issue. In addition, GNF requested a minor modification of the core-average, cell-average bow input to the channel-bow dependent critical power ratio (CPR) calculations.

In 2013 the U.S. Nuclear Regulatory Commission (NRC) staff approved GNF’s expanded lead use channel (LUC) program for the NSF zirconium alloy (Reference 3). The purpose of the expanded LUC program was to allow greater numbers of NSF channels to be exposed to varying in-reactor operating strategies, nuclear conditions, and water chemistry, in order to gain experience and gather data for batch application.

The NRC staff’s review was assisted by Pacific Northwest National Laboratory (PNNL). The staff’s conclusions on the acceptability of NSF for batch application is supported by PNNL’s Technical Evaluation Report (TER) (Reference 4).

2.0 REGULATORY EVALUATION

Regulatory guidance for the review of fuel system materials and designs and adherence to General Design Criteria (GDC)-10, GDC-27, and GDC-35 is provided in NUREG-0800, “Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants” (SRP), Section 4.2, “Fuel System Design” (Reference 5). In accordance with SRP Section 4.2, the objectives of the fuel system safety review are to provide assurance that:

• The fuel system is not damaged as a result of normal operation and anticipated operational occurrences (AOOs), • Fuel system damage is never so severe as to prevent control rod insertion when it is required,

Enclosure NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

- 2 -

• The number of fuel rod failures is not underestimated for postulated accidents, and • Coolability is always maintained.

The main focus of the limited SRP guidance with respect to BWR fuel bundle channels is control blade interference and insertability. SRP Section 4.2.II.1.A.v states:

Control blade/rod, channel, and guide tube bow as a result of (1) differential irradiation growth (from fluence gradients), (2) shadow corrosion (hydrogen uptake results in swelling), and (3) stress relaxation, which can impact control blade/rod insertability from interference problems between these components. For BWRs, the effects of shadow corrosion should be considered for new control blade or channel designs, dimensions (e.g., the distance between control blade and channel is important), or materials. The effects of channel bulge should also be considered for interference problems for BWRs. Design changes can alter the pressure drop across the channel wall, thus necessitating an evaluation of such changes. Channel material changes can also impact the differential growth, stress relaxation, and the amount of bulge and therefore must be evaluated. If interference is determined to be possible, tests are needed to demonstrate control blade/rod insertability consistent with assumptions in safety analyses. Additional in-reactor surveillance (e.g., insertion times) may also be necessary for new designs, dimensions, and materials to demonstrate satisfactory performance.

With respect to ensuring control blade insertability under externally applied loads (i.e., safe shutdown earthquake (SSE) and loss-of-coolant accident (LOCA)), SRP 4.2 Appendix A, Section IV states:

For a BWR, several conditions must be met to demonstrate control blade insertability – (1) combined loads on the channel box must remain below the allowable value defined above for components other than grids (otherwise, additional analysis is needed to show that the deformation is not severe enough to prevent control blade insertion) and (2) vertical liftoff forces must not unseat the lower tieplate from the fuel support piece such that the resulting loss of lateral fuel bundle positioning could interfere with control blade insertion.

The NRC staff’s review of NEDE-33798P is to ensure that the introduction of NSF does not adversely impact the ability of existing BWR channel designs to satisfy these requirements.

3.0 TECHNICAL EVALUATION

The staff’s review of the NEDE-33798P is summarized below:

• Verify that the fuel channel design requirements are consistent with regulatory criteria identified in SRP 4.2 or otherwise acceptable and justified. • Verify NSF material properties based on existing material property databases and supporting mechanical testing database. • Verify that the NSF channel designs satisfies regulatory requirements. NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

- 3 -

• Verify that the GNF experience database (in-reactor residence, post-irradiation examinations, and out-of-pile testing) supports the operating limits being requested and provides reasonable assurance that no anomalous behavior will occur during batch implementation. • Verify that the impact of the GNF channel designs on the reload design methodology, safety analyses, and setpoints has been properly addressed. • Define the range of applicability and allowed manufacturing tolerances/variances (e.g., alloy composition, microstructure). • Define future surveillance and reporting requirements, as necessary.

The staff’s review builds upon the NSF enhanced LUC program (Reference 3) and the operating experience and data collected from past and ongoing surveillance programs.

3.1 BWR Channel Design Requirements

In Section 1 of NEDE-33798P (Reference 1), GNF defines the design requirements for BWR channels under normal operating conditions as follows: (1) provides a guide for control blade insertion, (2) directs reactor coolant flow over the fuel rods effectively defining the flow envelope, and (3) controls the coolant flow leakage at the channel/lower tie plate interface. In addition, for transients and accident conditions, the channel (1) provides the structural stiffness for the fuel bundle to withstand fuel drop and seismic/LOCA loads and (2) transmits fuel assembly seismic loads to the top guide and fuel support. To perform these functions the channel must (1) maintain structural dimensions (i.e., avoid excessive channel distortion) and (2) maintain structural integrity (i.e., avoid failure due to stress and fatigue, avoid excessive metal thinning from corrosion). These design requirements are consistent with the SRP and, therefore, are acceptable.

3.2 NSF Composition and Microstructure

In Section 2.2 of NEDE-33798P (Reference 1), GNF describes the composition and microstructure of NSF. Zirconium metal is comprised of a hexagonal close-packed (HCP) crystalline structure at room temperature up through operating conditions (288 °C). When alloying elements are added, they either go into a solid solution or precipitate out as a second phase depending on the concentration and temperature. The alloying elements in NSF, like Zry-2 and Zry-4, result in a two-phase microstructure. In NSF, the Sn, O and portions of Nb will be in solid solution with the α-Zr matrix at operating temperatures while the remainder of the Nb and essentially all of the Fe combine with Zr to create small second phase particles (SPPs) that are distributed uniformly within an α grain structure. The SPPs in NSF have been identified to be of the Zr(Nb,Fe)2 type at low and intermediate temperatures.

As with existing Zry-2 and Zry-4 channels, the standard microstructure of NSF (except in weld-zones) in channels is a fully recrystallized grain structure with uniform distribution of SPPs. The normal manufacturing process to produce channel strip begins with an ingot that is triple melted for homogeneity. The ingot is then forged, hot rolled, and beta

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

- 4 - quenched. The beta quench dissolves all of the SPP’s into solid solution at temperature to precipitate them into finer, more dispersed particles upon cooling. After beta quenching, the final thickness of the strip is then produced by a sequence of hot and cold rolling to final size with intermediate and final anneals. The cumulative annealing after beta quenching determines the SPP size while the final anneal fully recrystallizes the grain structure.

The development of NSF builds upon decades of nuclear experience with Zr-Nb-Sn-Fe alloys. NSF’s nominal composition by weight percent (wt%) is: Zr (base) with 1.0% Nb, 1.0% Sn, and 0.35% Fe.

Table 2-1 of NEDE-33798P provides a comparison of nominal composition between NSF, ZIRLOTM, and E635. Examination of this table reveals similar alloying composition. The specific concentrations of the major alloying elements are controlled during the ingot melting process. The allowable range of major alloying elements is provided in Table 2-2 of NEDE-33798P. Control of impurities is described in Section 2.2.3 of NEDE-33798P. In response to a request for information (RAI) regarding the range in allowable alloy composition (RAI-1, Reference 2), GNF stated that the allowable range is analogous to the composition ranges defined for Zry-2 and Zry-4 in ASTM B352/352M-11 and that quality controls similar to American Society for Testing and Materials (ASTM) and in accordance with 10 CFR Part 50 Appendix B would be maintained.

PNNL’s technical assessment of NSF’s composition and microstructure is provided in Section 2.0 of the TER (Reference 4). PNNL had concerns regarding the impact of variation in Sn and O content on channel performance. In response to RAI-1c (Reference 2), GNF described the impact on ultimate tensile strength (UTS), creep rate, and channel distortion. As described in further detail below, PNNL found the response acceptable. PNNL concluded that the allowable range in alloying was acceptable. Based upon this assessment, the staff finds the NSF composition, allowable range in alloying content, and microstructure acceptable.

A limitation on allowable alloying content and microstructure will be included in the staff’s approval to ensure future NSF channels exhibit the same thermal, mechanical, and nuclear properties and performance as described in NEDE-33798P and in response to RAIs.

3.3 NSF Material Properties

In Sections 2.3 through 2.15 of NEDE-33798P (Reference 1), GNF describes the thermal, mechanical, and nuclear properties of NSF. PNNL’s technical assessment of NSF’s material properties is provided in Section 3 of the TER (Reference 4). For a majority of the material properties, GNF demonstrated that the NSF and Zry-2 properties are equivalent and the use of Zry-2 properties in design calculations is reasonable. For a few key properties, NSF showed improvement relative to Zry-2. However, this improvement was not always credited in mechanical design calculations. Table 3-1 of Reference 4 summarizes the material properties used in downstream design calculations.

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

- 5 -

In response to an RAI regarding the impact of changes in Sn and O content on NSF’s UTS and yield strength (YS) (RAI-1c, Reference 2), GNF stated that design strength of NSF or any zirconium alloy is actually controlled in the material specification and requires objective evidence that the material lots meet the strength requirement. This accounts for any variability in the material strength from variation in chemistry. In Section 3.4 and 3.5 of Reference 4, PNNL reviewed the modified stress-strain models for NSF along with the design values. PNNL found these models acceptable.

In response to an RAI regarding the impact of changes in Sn and O content on NSF’s creep rate (RAI-1c, Reference 2), GNF stated no significant difference in creep rate would be expected over the allowable composition range. This conclusion was supported by experimental data. PNNL found this response acceptable.

In response to an RAI regarding differences in channel growth between NSF and Zry-2 shown on Figure 2-13 of NEDE-33798P (RAI-5, Reference 2), GNF described the bases of the growth database and further evidence that NSF does not experience the breakaway growth exhibited by Zry-2. PNNL found this response acceptable.

In response to an RAI regarding differences in corrosion and hydrogen uptake between NSF and Zry-2 (RAI-3, Reference 2), GNF provided oxide thickness and hydrogen measurements from samples at nearly identical operating conditions. Examination of this database reveals that NSF has a higher corrosion rate, but significantly lower hydrogen pickup fraction relative to Zry-2. Based upon the information provided in NEDE-33798P and in response to RAI-3, PNNL found the corrosion performance of NSF channels acceptable.

In response to an RAI regarding differences steam oxidation under accident conditions between NSF and Zry-2 (RAI-3f, Reference 2), GNF provided data on oxidation measurements of NSF with and without a pre-oxidation film at 1000 ºC and compared the results with a Zr-2 specimen, which was not pre-oxidized. The NSF material with and without the pre-oxidation film behaved better than Zry-2. All of the specimens tested exhibited oxidation rates less than the Cathcart-Pawel (CP) relationship. PNNL found this response acceptable.

Based upon PNNL’s assessment of NEDE-33798P and responses to applicable RAIs, the staff finds the material properties used in the mechanical design calculations for NSF channels acceptable.

3.4 NSF Operating Experience

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

- 6 - program (RAI-9, Reference 2), GNF provided details of past, ongoing, and future NSF LUCs. Upon its completion, the NSF LUC program will have amassed a significant channel performance empirical database.

The enhanced LUC program includes a monitoring program which is designed to provide a reasonable level of assurance against unanticipated channel distortion. In addition, the enhanced LUC program includes a post-irradiation inspection plan which is designed to gather data during subsequent reload cycles to identify negative trends and confirm expected performance. The inspection plan dictates specific requirements on the number and type of inspections to be performed on NSF LUCs. It is expected that the data being collected in accordance with this inspection plan on the vast quantity of NSF LUCs will provide reasonable assurance of in-reactor performance and model validation. This data will continue to be collected and will lead batch application of NSF channels.

The NSF channel operating experience database includes C-, D-, and S-Lattice plants up to a maximum of 52 GWd/MTU assembly average burnup, 2,796 days (residence), and 43,600 inch-days (effective control blade exposure (ECBE). In addition, the largest number of NSF LUCs monitored and inspected will have operated in S-Lattice plants, which are more susceptible to channel interference (due to smallest gap between blade and channel).

PNNL’s technical assessment of NSF’s channel surveillance program is provided in Section 5 of the TER (Reference 4). Based upon the information in NEDE-33798P and in response to RAI-9, PNNL states that the operational experience in this LUC phase supports the conclusions that NSF is resistant to fluence gradient-induced bow, resistant to shadow corrosion-induced bow, and similar to Zircaloy in creep corrosion performance.

In response to an RAI regarding the need for a future NSF surveillance and reporting requirement (RAI #10, Reference 2), GNF cited the monitoring, surveillance, and reporting requirements of the ongoing NSF LUC program. GNF identified the number of past, present, and expected future NSF channels included in this program and the large quantity of data collection. Based upon the information presented in response to RAI-9 and RAI-10, the staff finds the NSF LUC monitoring, surveillance, and reporting requirements acceptable with respect to providing sufficient confirmatory information ahead of batch applications. In addition to collecting and reporting the channel performance data, the NRC agrees with GNF that certain empirically-based models should be re-calibrated as more data becomes available. Section 5 lists conditions and limitations of the staff’s approval of NSF channels. Included are requirements that the NSF LUC irradiation and data collection program continue, annual reports are provided to NRC staff, and GNF notify and provide the bases of changes in channel performance models.

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

- 7 -

3.5 NSF Performance Evaluation

GNF has requested use of NSF material on existing approved GNF channel designs. Except as described below, existing approved mechanical design requirements and calculational methods will be used to confirm the performance of GNF channels manufactured with NSF material. The staff finds the continued use of these design requirements and methods, along with the material properties described in Section 3.3, acceptable for NSF channels.

Recent operating experience has shown that channel distortion and associated control blade interference continues to be a major problem in the U.S. BWR commercial fleet. The goal of introducing NSF channel material is to resolve this issue. Contributing factors for channel distortion include (1) creep bulge, (2) fluence gradient-induced bow, and (3) shadow corrosion-induced bow. Each will be addressed below.

Creep Bulge: Creep bulge in channels occurs because of the differential pressure between the inside and outside of the bundle. At a given axial position, the pressure drop is effectively a constant stress on the channel face that induces an elastic bulge that over time results in permanent strain. Channel deformation due to creep bulge has not been a major concern in the industry and, by itself, has not lead to control blade interference issues. The purpose of this review is to provide reasonable assurance that the use of NSF does not exacerbate creep bulge and/or introduce a new problem.

In Section 2.9 of NEDE-33798P (Reference 1), GNF describes the creep bulge mechanics and measured bulge database. PNNL’s technical assessment of NSF’s creep bulge model and supporting database is provided in Section 5.3 of the TER (Reference 4). In response to an RAI regarding comparable NSF and Zry-2 measured bulge (RAI-4b, Reference 2), GNF stated that the data depicted in Figure 2-11 of NEDE-33798P for both materials is from the same channel design and plant type. PNNL concluded that the creep rate of NSF channels is comparable to Zry-2 channels.

Based upon the measured creep data and PNNL’s assessment, the staff finds NSF channel performance with respect to creep bulge acceptable.

Fluence Gradient-Induced Bow: Irradiation growth is mainly attributed to the anisotropic redistribution of irradiation-induced vacancies and interstitials into dislocation loops on preferred crystallographic planes. Channel bowing occurs when a flux gradient across the channel box induces differential growth on opposite faces of the channel box. In bundles located toward the core periphery, a higher neutron flux would be experienced on the channel face toward the core interior, relative to the face toward the core periphery. Channel deformation due to fluence gradient-induced bow has been a major concern in the industry and, coupled with shadow corrosion-induced bow, has resulted in control blade interference issues. The purpose of this review is to provide reasonable assurance that NSF channels provide improved or equivalent performance or, at least, do not exacerbate fluence gradient-induced bow and/or introduce a new problem.

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

- 8 -

In Section 2.11 of NEDE-33798P (Reference 1), GNF describes the irradiation growth mechanics, model, and measured growth database. PNNL’s technical assessment of NSF’s irradiation growth model and supporting database is provided in Section 5.1 of the TER (Reference 4). In response to an RAI regarding the growth database (RAI-6c, RAI-6d, Reference 2), GNF provided additional data for NSF channels with EBCE > 4500 inch-days and clarified the ECBE weighting factor. Based upon the growth database, PNNL concluded that NSF channels exhibits lower and more controlled growth than Zry-2 channels. In addition, PNNL reviewed the fluence gradient- induced bow model and found it acceptable.

The database supporting the fluence gradient-induced bow performance and model is limited. As described in Section 3.4, ongoing and future NSF LUC programs are expected to provide a significant amount of new data to confirm performance and validate models. A condition on the staff’s approval has been developed to document and report the NSF LUC data as it is collected. See Section 5.

Based upon the irradiation growth database, ongoing and future LUC data collection, and PNNL’s assessment, the staff finds NSF’s fluence gradient-induced bow performance and models acceptable. Revisions to the bow model based upon future data collection are allowed under the provisions described in Section 5.

Shadow Corrosion-Induced Bow: Shadow corrosion is an enhanced irradiation corrosion mechanism that occurs on zirconium alloys when a dissimilar material (such as a stainless steel control blade) is near the zirconium surface (such as a BWR channel) and the water chemistry is oxygenated. When a fuel bundle is controlled early in life, the increased corrosion on the blade side relative to the non-blade side results in a difference in hydrogen absorbed in channel material. Hydrogen is absorbed into the metal as part of the corrosion process and causes a volume change resulting in channel bow.

Because direct measurement of shadow corrosion-induced bow is only possible when the fluence gradient is zero, shadow bow is generally observed by accounting for the fluence gradient induced bow. After accounting for fluence bow in the data, the end-of-life channel bow correlates well with the total effective control blade exposure (ECBE).

Prediction of shadow bow involves two parts. The first part is calculation of ECBE for each channel (this is a measure of the susceptibility to shadow-corrosion induced bow). The second part is using an empirical correlation to convert the ECBE to a corresponding shadow-induced bow.

In Sections A.2.3 and A.2.4 of NEDE-33798P (Reference 1), GNF describes the shadow corrosion-induced bow mechanics, model, and supporting database. PNNL’s technical assessment of NSF’s shadow corrosion bow model and supporting database is provided in Section 5.2 of the TER (Reference 4). In response to an RAI regarding the inferred shadow bow database (RAI-4, Reference 2), GNF provided additional data for NSF and Zry-2 channels operated under similar conditions. Based upon the shadow corrosion

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

- 9 - database, PNNL concluded that NSF channels exhibits lower and more controlled shadow bow than Zry-2 channels. In addition, PNNL reviewed the shadow corrosion-induced bow model and found it acceptable.

The database supporting the shadow corrosion-induced bow performance and model is limited. As described in Section 3.4, ongoing and future NSF LUC programs are expected to provide a significant amount of new data to confirm performance and validate models. A condition on the staff’s approval has been developed to document and report the NSF LUC data as it is collected (see Section 5.0 below).

Based upon the shadow corrosion bow database, ongoing and future LUC data collection, and PNNL’s assessment, the staff finds NSF’s shadow corrosion-induced bow performance and models acceptable. Revisions to the bow model based upon future data collection are allowed under the provisions described in Section 5.0 below.

NSF Corrosion: As with any in-reactor material the first and foremost performance requirement is that the material withstands corrosion to the extent that it maintains structural integrity, and thus, maintains its ability to perform its design requirements. For channels, maintaining structural integrity is the only corrosion performance requirement, which in practice means that the component must maintain a minimum thickness of metal.

In Section 2.10 of NEDE-33798P (Reference 1), GNF describes the NSF corrosion model and supporting database. PNNL’s technical assessment of NSF’s corrosion model and supporting database is provided in Section 5.4 of the TER (Reference 4). In response to an RAI regarding the potential effect of alloying composition (RAI-1a, RAI-1b, Reference 2), GNF cited several technical papers investigating alloying composition on corrosion rates. GNF concluded that the significant margin between the measured corrosion and the design value would likely account for all variation in nominal corrosion caused by variation in ingot chemistry or from other variables such as location on the channels.

In response to an RAI regarding the NSF corrosion database (RAI-3, Reference 2), GNF provided a comparison of corrosion and hydrogen content for Zry-2, Zry-4, and NSF channels. Based upon a review of the additional data, PNNL found the oxidation performance of NSF channels acceptable.

GNF has requested approval of NSF channels with and without a pre-oxidized surface finish, which has been done in the past with Zry-4 channels. PNNL concluded that the pre-oxidized surface finish will have no detrimental impact on corrosion performance.

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

- 10 -

In response to an RAI regarding the high temperature corrosion under accident conditions (RAI-3f, Reference 2), GNF provided high temperature steam oxidation data for NSF channel material (with and without pre-oxidized surface finish) along with Zry-2 channel material. The test results show acceptable performance for the NSF material. Note that the data demonstrates the applicability of the CP correlation to NSF material. Based upon the data presented in response to RAI-3f, PNNL found the high temperature corrosion performance of NSF acceptable.

Based upon the information presented in the TR, in response to RAI-1 and RAI-3, and PNNL’s assessment, the staff finds the corrosion performance of NSF acceptable.

Calculating CPR with NSF Channels: In Section 3 of NEDE-33798P (Reference 1), GNF describes the current method for calculating CPR, specifically highlighting the dependence of R-factor on channel bow. This section is provided for information only and is intended to provide clarification on the previously approved method for establishing dependence of bundle R-factors and CPRs on channel bow. As described in Section 3.1.5, the current procedure for determining BOWAVE considers effects of initial as-manufactured channel bow and fluence-induced bow. For cores with NSF channels, a revised value of BOWAVE was requested.

PNNL’s technical assessment of NSF’s BOWAVE and core-average, cell-average bow (CACABO) is provided in Section 4 of the TER (Reference 4). In response to an RAI regarding further information on CACABO calculations (RAI-6, Reference 2), GNF provided sample calculations over a broad range of operating conditions. In addition, GNF provided sample calculations for channel bow and CACABO in response to RAI-7 (Reference 2).

In response to an RAI regarding uncertainties applied to R-factor calculations (RAI-7b, Reference 2), GNF responded that the same uncertainties used for Zry-2 channels, which generally show greater bow and greater scatter in the magnitude of bow, will be applied to NSF channels. PNNL found this approach acceptable.

Based upon the information provided in the TR, response to RAI-6 and RAI-7, and PNNL’s assessment, the staff finds the approach for calculating CPR with NSF channels acceptable. Revisions to the R-factor uncertainty based upon NSF channel distortion measurements are allowed under the provisions described in Section 5.0 below.

3.6 Range of Applicability

In Section 4 of NEDE-33798P (Reference 1), GNF states that upon approval of this TR, NSF may be incorporated into GNF fuel designs in channels by inclusion in the GESTAR new fuel compliance reports for a specific fuel design, as supported by appropriate analysis using the properties described herein. This section did not systematically identify limitations on design and/or operation based upon the extent of data or experience. In response to an RAI regarding the range of applicability (RAI-4,

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

- 11 -

Reference 2), GNF proposed limitation on residence time, channel fluence, fuel burnup, and ECBE.

During subsequent meetings between GNF, PNNL, and NRC, the response to RAI-2 was discussed and it was determined that further limitations were required. Based upon a review of the NSF experience and database, the following limitations were identified: 1. The fully recrystallized Zr-Nb-Se-Fe alloy channel material, referred to as NSF, is restricted to the alloy composition range provided in Table 2-2 of NEDE-33798P. 2. The fully recrystallized NSF alloy channel material is approved for batch application to BWR channel designs based on currently approved design methodologies. The GESTAR II compliance report for each fuel product line describes the channels and confirms that the design meets the requirements for mechanical design, and seismic and LOCA conditions. 3. The fully recrystallized NSF alloy channel material is approved for batch application to BWR/2, BWR/3, BWR/4, BWR/5, and BWR/6 designs. NSF is also approved for batch application to ABWR and ESBWR designs. 4. The lifetime of NSF channels is restricted to the following limitations. Any fuel channel projected to exceed any of these limitations during the upcoming reload cycle shall not be loaded into the reactor, except as allowed in accordance with GNF’s approved lead use program. Extended life LUC shall be limited to less than 2 percent of the channels in the core. The extended life LUC program will be exclusive of the 2 percent provision of GESTAR II for the testing of new design features in lead use assemblies. The notification and inspection requirements of this program will be consistent with the approved GESTAR II lead use program. a. Residence time shall not exceed 8 years. b. End of Life (EOL) fast fluence (>1.0 MeV) shall not exceed 1.2E22 n/cm2 (channel average). NRC accepts the use of the surrogate 70 GWd/MTU peak pellet burnup limit to satisfy this fluence limit. Future changes in fuel assembly lattice design, fuel rod design, and/or fuel loading patterns which may invalidate this relationship needs to be evaluated and reported to the NRC. c. EOL ECBE shall not exceed 55,000 inch-days except to suppress power for unanticipated, emergent operational issues. If a channel exceeds 55,000 inch-days because of suppressing power for unanticipated, emergent operational issues, it shall be considered a LUC if reinserted in a following cycle. d. Channels shall not be re-used on different assemblies.

3.7 Incorporation of NSF into GESTAR II

By letter dated March 24, 2015 (Reference 6), GNF requested staff approval to incorporate NSF into GESTAR II. Change pages to GESTAR II (NEDE-24011-P-A) contain a brief description of NSF along with reference to NEDE-33798P. The approval date and “-A” version of the NSF TR will be updated in the future. The staff has reviewed these changes and finds them acceptable.

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

- 12 -

4.0 CONCLUSION

By letter dated February 13, 2013 (Reference 1), as supplemented by a letter dated June 10, 2015 (Reference 2), GNF requested review and approval of an advanced zirconium alloy, NSF, for application to existing BWR fuel channel designs. Recent operating experience has shown that channel distortion and associated control blade interference continues to be a major problem in the U.S. BWR commercial fleet. The goal of introducing NSF channel material is to resolve this issue. In addition, GNF requested a minor modification of the core-average, cell-average bow input to the channel-bow dependent CPR calculations.

In 2013, the NRC staff approved GNF’s expanded LUC program for the NSF zirconium alloy (Reference 3). The purpose of the expanded LUC program was to allow greater numbers of NSF channels to be exposed to varying in-reactor operating strategies, nuclear conditions, and water chemistry, in order to gain experience and gather data for batch application. The data being collected in the NSF expanded LUC program provides confirmation of NSF channel performance and data to validate performance models. The NRC staff has completed its review of NEDE-33798P and finds it acceptable. Licensees referencing NEDE-33798P will need to comply with the conditions listed in Section 5.0 below.

With regard to the use of NSF channels, the staff has concluded, based on the considerations discussed above, that: (1) there is reasonable assurance that the health and safety of the public will not be endangered by operation in the proposed manner, (2) such activities will be conducted in compliance with the commission’s regulations, and (3) issuance of this safety evaluation will not be inimical to the common defense and security or to the health and safety of the public.

5.0 CONDITIONS AND LIMITATIONS

Licensees referencing NEDE-33798P must ensure compliance with the following conditions and limitations:

1. The range of applicability of NSF channels is limited to those items described in Section 3.6 of this safety evaluation (SE). 2. The expanded NSF LUC program monitoring and inspection plan, detailed in Section 3.2 of the staff’s SE (MFN 12-074 Supplement 2-A), must be completed. 3. The quantity of LUCs, exposures achieved, and post-irradiation examinations and data collection shall be consistent with GNF’s response to RAI-9 and RAI-10 of NEDE-33798P. 4. To ensure continued in-reactor performance and applicability of NSF models, GNF must provide an annual report, addressed to the Director, Division of Safety Systems, Office of Nuclear Reactor Regulations, documenting the ongoing experience with the enhanced NSF LUC program, post-irradiation examinations and data collection, and validation of NSF models. At a minimum, the annual report must contain the following information:

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

- 13 -

a. Plot of NSF channel irradiation database, expressed as ECBE versus exposure. b. Plot of measured channel growth versus fast neutron fluence data, along with NSF growth model predictions. c. Plot of measured channel bulge versus exposure data. d. Plot of measured channel bulge data versus NSF channel bulge model predictions. e. Plot of measured channel distortion (total) versus exposure data, segregating low and high ECBE data. f. Plot of inferred shadow corrosion bow versus ECBE data, along with NSF shadow bow model predictions. Data collected within the prior annual period should be clearly delineated. The annual report is no longer required once all of the NSF enhanced LUCs have achieved the program EOL exposure and data has been reported. Note that item 6 below may necessitate a future report beyond the LUC program lifetime. 5. Based upon post-irradiation examinations and data collection, as described in items 2, 3, and 4, above, GNF may alter the NSF channel growth model, bow and bulge models, and shadow corrosion model to achieve an improved fit to the database. These channel distortion models will be a nominal fit to the data. The model uncertainty is defined as a standard deviation calculated using standard mathematical formulas. Any data eliminated from the model calibration should be identified and justified. Any changes to these models must be documented within the annual report described in item 4 above. 6. BWR channel distortion – control blade interference counter measures, including fuel management guidelines and augmented monitoring and inspection programs as described in MFN 10-245 (most recent version), will continue to be applied for cores containing NSF channels. These counter measures may be eliminated once a full core of NSF channels has experienced no observations of control blade-to-channel interference (e.g., slow to settle, no settle, delayed scram) for 3 consecutive years within an S-lattice design. Elimination of the counter measures should be documented in the annual report similar to that described in item 4 above. 7. Any future change in the R-factor uncertainty based upon incorporation of NSF channel distortion measurements must be justified and documented in the annual report similar to that described in item 4 above.

6.0 REFERENCES

1. GNF Letter MFN 13-008, “Application of NSF to GNF Fuel Channel Designs,” NEDC-33798P, February 13, 2013, ADAMS Accession No. ML13045A456. 2. GNF Letter MFN 15-040, “Response to Request for Additional Information Regarding Review of Licensing Topical Report NEDE-33798P, ‘Application of NSF to GNF Fuel Channel Designs’ (TAC No. MF0742),” June 10, 2015, ADAMS Accession No. ML15161A508. 3. GNF Letter MFN 12-074, Supplement 2-A, “Enhanced LUC Program for NSF Channels,” April 15, 2013, ADAMS Accession No. ML13106A067.

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

- 14 -

4. PNNL Technical Evaluation Report, “NEDE-33798P, Revision 0, Application of NSF to GNF Fuel Channel Designs, February 2013,” ADAMS Accession No. ML15211A015. 5. NUREG-0800, “Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants” (SRP), Section 4.2, “Fuel System Design,” Revision 3, March 2007, ADAMS Accession No. ML070740002. 6. GNF Letter MFN 13-008, Supplement 2, “Incorporation of the Approved Topical Report NEDE-33798P, ‘Application of NSF to GNF Fuel Channel Designs’, into GESTAR II,” March 24, 2015, ADAMS Accession No. ML15083A301.

Principal Contributor: Paul M. Clifford

Date: September 1, 2015

NEDO-33798-A Revision 1 Non-Proprietary Information - Class I (Public)

Comment Resolution Summary for Draft Safety Evaluation for NEDE-33798P, “Application of NSF to GNF Fuel Channel Designs”

Location in Comment NRC Disposition MFN 15-071 Markup Section 3.3 Page 6 (2nd to last paragraph in Section 3.3 Comment accepted. NSF Material GNF suggests the following changes: Change incorporated in Properties “The NSF material with and without the final SE. pre-oxidation film behaved better than Zry-2” Section 3.4 Page 6 (2nd paragraph in Section 3.4) Comment accepted. NSF GNF suggests the following changes: Change incorporated in Operating “In addition, the enhanced LUC program final SE. Experience includes a post-irradiation plan which is designed to gather data during subsequent reload cycles to identify negative trends and confirm expected performance.”

Attachment NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

TABLE OF CONTENTS Page

Acronyms and Abbreviations ...... vii Executive Summary ...... viii Revision Summary ...... ix 1.0 Introduction ...... 1-1 2.0 NSF Material Microstructure and Properties ...... 2-1 2.1 Historical Perspective on Zr-Sn-Nb-Fe alloys ...... 2-1 2.2 Composition and Microstructure of NSF...... 2-2 2.3 Elastic Properties ...... 2-5 2.4 Thermal Expansion ...... 2-6 2.5 Specific Heat ...... 2-7 2.6 Thermal Conductivity ...... 2-8 2.7 Density of NSF ...... 2-9 2.8 Plastic Properties...... 2-9 2.9 Thermal and Irradiation Creep ...... 2-12 2.10 Corrosion ...... 2-15 2.11 Irradiation Growth ...... 2-17 2.12 Fatigue ...... 2-18 2.13 Stress Rupture ...... 2-19 2.14 Hydrogen Effects ...... 2-20 2.15 Nuclear Properties...... 2-21 3.0 Calculating CPR with NSF Channels ...... 3-1 3.1 GNF Methodology for Critical Power Ratio Calculations ...... 3-1 3.2 NSF Application to LHGR Calculation ...... 3-4 4.0 Applicability ...... 4-1 5.0 Summary ...... 5-1 6.0 References ...... 6-1 APPENDIX A Calculating Core-Average, Cell-Average Bow for NSF Channels ...... A-1 APPENDIX B GNF Responses to NRC RAIs on NEDO-33798 Revision 0 ...... B-1

iii NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

LIST OF TABLES

Table Title Page

Table 1-1 Material Properties for Designing Channels ...... 1-2 Table 2-1 Nominal Chemical Weight Percent Composition of Zr-Nb-Sn-Fe Alloys ...... 2-22 Table 2-2 Composition Range for NSF ...... 2-23 Table 2-3 Comparison of Typical Texture Values for NSF Compared to Zircaloy-2 ...... 2-23 Table 2-4 Constants A and B for calculating the Young’s Modulus for Zircaloy-2, Zircaloy-4 and NSF ...... 2-24 Table 2-5 Constants A and B for Calculating the Poisson’s Ratio for Zircaloy-2, Zircaloy-4 and NSF ...... 2-24 Table 2-6 Specific Heat of the Constituent Elements in Zircaloy-2 and NSF (Reference 41) ...... 2-25 Table 2-7 Measured Thermal Diffusivity and Calculated Thermal Conductivity of Zircaloy-2 and NSF ...... 2-26 Table 2-8 Average Measured Yield and Ultimate Strength of NSF Channel Material in the Unirradiated and Irradiated (~9.5 x 1021 n/cm2) Condition ...... 2-27 Table 2-9 Design Yield and Ultimate Strength of NSF Channel Material ([[ ]]in/in/min) ...... 2-28 Table 2-10 Rockwell B Hardness Measurements NSF-to-NSF Weld Compared to NSF in the Fully Recrystallized State ...... 2-28

iv NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

LIST OF FIGURES

Figure Title Page

Figure 1-1 Schematic of a Typical BWR Fuel Bundle Assembly ...... 1-3 Figure 1-2 Schematic of a BWR Control Cell ...... 1-4 Figure 2-1 Irradiation Growth of E635 (Reference 15) Compared to Zircaloy-2 (Reference 42) ...... 2-29 Figure 2-2 Hexagonal Close-Packed (HCP) Crystal Structure of Zirconium ...... 2-30 Figure 2-3 Recrystallized Grain Structure ...... 2-31 Figure 2-4 TEM Image of NSF Second Phase Particles ...... 2-32 Figure 2-5 Widmanstätten Microstructure of a NSF-to-NSF Weld...... 2-32 Figure 2-6 Thermal Expansion (dL/L=i/io) of NSF and Zircaloy-2 ...... 2-33 Figure 2-7 Thermal Conductivity of NSF ...... 2-34 Figure 2-8 Measured Strength of NSF (yield and UTS) Compared to the Predicted Strength Using the NSF Plasticity Model ...... 2-35 Figure 2-9 Measured Strength of NSF (yield and UTS) Compared to the Design Strength Using the NSF Plasticity Model ...... 2-36 Figure 2-10 Schematic of Force and Deflection on a Channel Face ...... 2-37 Figure 2-11 Creep Bulge Measurement for NSF and Zircaloy-2 Channels taken at the 40 in. Elevation ...... 2-38 Figure 2-12 Upper bound Design Value for Corrosion and Recent Hot Cell Oxide Thickness Data for Zircaloy-2 and NSF ...... 2-39 Figure 2-13 GNF Channel Growth Data and NSF Irradiation Growth Data ...... 2-40 Figure 2-14 Irradiation Growth Data on a Series of Zr-Nb-Sn-Fe Alloys ...... 2-41 Figure 2-15 Irradiation Growth Data on a Series of Zr-1Nb-1Sn Alloys ...... 2-42 Figure 2-16 Fatigue Data for Zircaloy Used in Fatigue Evaluations for Cladding (References 26, 37, and 38) ...... 2-43 Figure 2-17 NSF and Zircaloy-2 Fatigue Data ...... 2-44 Figure 2-18 Comparison of Zircaloy-2 and NSF Tested Using Similar Applied Stresses ...... 2-45 Figure 2-19 Comparison of Irradiated Zircaloy-2 Tested at 550°F(288°C) and Irradiated NSF Tested at 752°F(400°C) ...... 2-46 Figure 2-20 Measured Hydrogen Pickup in a NSF Channel ...... 2-47 Figure 2-21 Photomicrograph of Precipitated Hydrides ...... 2-48 Figure 3-1 Calculated Values of CACABO for NSF Cores ...... 3-5 Figure 3-2 Absolute MFLCPR Difference ...... 3-6 Figure 3-3 Uncertainty in the Core-Average, Cell-Average Bow as a Function of Cycle Exposure ...... 3-7 Figure A-1 Channel Face Numbering for the Bow Model ...... A-10

v NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Figure A-2 Calculation of Segment Edge Bow ...... A-10 Figure A-3 Fluence Induced Bow ...... A-11 Figure A-4 Channel Bow Calculation ...... A-11 Figure A-5 Range of ECBE and Exposure in the ...... A-12 Figure A-6 Measured Bow Plotted as a Function of Predicted Fluence Bow for NSF Channels ...... A-13 Figure A-7 Residual (Measured – Predicted) of Fluence Gradient-Induced Bow Data ...... A-14 Figure A-8 Residual (Measured – Predicted Fluence Bow) of Fluence Gradient-Induced Bow Data ...... A-15 Figure A-9 Inferred Shadow Bow (Measured – Predicted Fluence Bow) as a Function of ECBE for the NSF Database ...... A-16

vi NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

ACRONYMS AND ABBREVIATIONS

Term Definition ASTM American Society for Testing and Materials BCC Body Centered Cubic BOWAVE Core-Average, Cell-Average Bow used in CPR Methodology BWR Boiling Water Reactor CACABO Core-Average, Cell-Average Bow calculated in 3D core simulator CPR Critical Power Ratio ECBE Effective Control Blade Exposure EPRI Electric Power Research Institute GNF Global Nuclear Fuel HCP Hexagonal Close-Packed HPUF Hydrogen Pickup Fraction ISTC International Science and Technology Center KAPL Knolls Atomic Power Laboratory LHGR Linear Heat Generation Rate LOCA Loss-of-Coolant Accident LTR Licensing Topical Report LWR Light Water Reactor MFLCPR Maximum Fraction of Limiting Critical Power Ratio NRC Nuclear Regulatory Commission NSF Zr-Sn-Nb-Fe Alloy R-Factor Weighted rod local power peaking for critical power calculations SLMCPR Safety Limit Minimum Critical Power Ratio SPP Second Phase Particles TEM Transmission Electron Microscopy TMOL Thermal-Mechanical Operating Limit UTS Ultimate Tensile Strength

vii NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

EXECUTIVE SUMMARY

The purpose of this licensing topical report (LTR) is to request approval to use NSF, a fully recrystallized Zr-Sn-Nb-Fe alloy, for channel components within Global Nuclear Fuel (GNF) fuel assemblies used in boiling water reactors (BWRs). The material properties of NSF that are necessary to support licensing of channels are provided in Section 2. [[

]] In addition to providing the material properties that are used to ensure NSF channels meet all design requirements, GNF is requesting Nuclear Regulatory Commission (NRC) approval of a minor modification of the core-average, cell-average bow input to the channel-bow dependent critical power ratio (CPR) calculations. The bow inputs are currently based on performance of Zircaloy channels and need updating because of the improved channel bow characteristics of NSF channels compared to Zircaloy channels. The CPR methodology is described for information in Section 3; this section includes a description of how the core-average, cell- average bow is calculated and how it is applied in the R-factor methodology to account for channel bow in the CPR calculation. [[

]] Thus, GNF is recommending that once a plant has transitioned to NSF channels, the core-average, cell-average bow will be [[ ]].

viii NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

REVISION SUMMARY

Revision 1: 1. Created '-A' version by adding the NRC's Final Safety Evaluation (Reference 50), GNF's responses to the NRC's Requests for Additional Information (RAIs) (Reference 51), and GNF correspondence regarding corrected pages (Reference 52). 2. Revised Table 1-1 and Section 5.0 consistent with GNF’s letter MFN 13-008 Sup 1 dated June 12, 2013 (Reference 52). 3. Added References 50, 51, and 52.

ix NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

1.0 INTRODUCTION

The purpose of this licensing topical report (LTR) is to request approval to use NSF, a Zr-Sn-Nb- Fe alloy, for channel components within Global Nuclear Fuel (GNF) fuel assemblies used in boiling water reactors (BWRs). NSF will be discussed in terms of the performance requirements, described in References 1 and 2, for BWR fuel channels. In addition, an evaluation of how NSF will affect the calculation of the critical power ratio (CPR) and the linear heat generation rate (LHGR) is included. Because the GNF methodology for CPR has a channel- bow dependence, GNF is also requesting NRC approval of a minor modification to the bow assessment and application approach for use with NSF channels. A schematic of a typical BWR fuel assembly is shown in Figure 1-1. It consists of a fuel channel placed over a fuel bundle that contains an array of ~100 fuel rods, one or more water rods, spacers distributed axially, and an upper and lower tie plate. BWR fuel assemblies are placed within a control cell that consists of four fuel assemblies in a square array defined by the top guide structure (see Figure 1-2). A cruciform-shaped control blade moves between the fuel assemblies to control reactivity. The fuel channel under normal operating conditions (1) provides a guide for control blade insertion, (2) directs reactor coolant flow over the fuel rods effectively defining the flow envelope, and (3) controls the coolant flow leakage at the channel/lower tie plate interface. In addition, for transients and accident conditions, the channel (1) provides the structural stiffness for the fuel bundle to withstand fuel drop and seismic/loss-of-coolant accident (LOCA) loads and (2) transmits fuel assembly seismic loads to the top guide and fuel support. To perform these functions the channel must: 1. Maintain shape – avoid excessive channel distortion 2. Maintain structural integrity – avoid failure; avoid excessive metal thinning from corrosion To ensure that channels perform their functions, evaluations that depend on the material properties are completed during the design phase. These evaluations are listed in Table 1-1 with reference to the governing licensing topical report and the material properties used in the evaluations. The NSF material properties documented in Section 2 are used to confirm fuel channel design adequacy as defined in References 1 and 2. The effect of NSF channels on CPR and LHGR calculations is provided in Section 3.

1-1 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Table 1-1 Material Properties for Designing Channels

Design Evaluation Governing LTR Material Properties

[[ Young’s Modulus, Poisson’s Ratio, NEDE-21354-P Yield Strength, Ultimate Tensile Strength

Young’s Modulus, Poisson’s Ratio, NEDE-21354-P Coefficient of Thermal Expansion

Young’s Modulus, Poisson’s Ratio, Coefficient of Thermal Expansion ([[ ]]), NEDE-21354-P Fatigue, Stress Rupture, Metal Thinning from Corrosion ([[ ]])

Young’s Modulus, Poisson’s Ratio, NEDE-21354-P Stress/Strain Response

NEDE-21354-P Young’s Modulus, Poisson’s Ratio

Young’s Modulus, Poisson’s Ratio, NEDE-21354-P Coefficient of Thermal Expansion, Stress/Strain Response, Fatigue

Young’s Modulus, Yield Strength, NEDE-21175-3-P-A ]] Ultimate Tensile Strength

1-2 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Figure 1-1 Schematic of a Typical BWR Fuel Bundle Assembly

1-3 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Figure 1-2 Schematic of a BWR Control Cell Figure Note: The figure also shows the relative distortion caused by channel bow and bulge. Channel bow is generally greatest at the center elevation while bulge is larger near the bottom of the core because the pressure drop across the channel wall decreases as the elevation increases.

1-4 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

2.0 NSF MATERIAL MICROSTRUCTURE AND PROPERTIES

2.1 HISTORICAL PERSPECTIVE ON ZR-SN-NB-FE ALLOYS

In 1955, the groundbreaking work on the use of zirconium alloys in light water reactors (LWRs) within Admiral Rickover’s Naval Reactors program became public knowledge (References 3 and 4). The work within the Naval Reactors program culminated with the development of the Zircaloy series alloys (Zircaloy-2 and Zircaloy-4). After these publications, work on other zirconium alloys began throughout the world. In 1958 at the 2nd International Conference on the Peaceful Uses of Atomic Energy, the Russian industry published a series of papers on Zr-Nb alloys (References 5 and 6). Soon afterwards the British and Canadians published work on Zr-Nb alloys (References 7 and 8). This work culminated with the development of alloys E110 (Zr-1%Nb) and E125 (Zr-2.5%Nb) that were used in the Russian VVWR and RMBK reactors for cladding and pressure tube applications and the use of Zr-2.5%Nb alloys for pressure tubes in the CANDU reactors. In the late 1960s and early 1970s, work continued on these Zr-Nb alloys but with the additions of Sn and Fe. While Kass (Reference 9) found little benefit of Nb additions to Zircaloy-4 when considering only out-of-reactor corrosion, Amaev et al. found encouraging corrosion performance in-reactor for an alloy designated E635 (Zr-1%Sn-1%Nb-0.5%Fe, Reference 10). The Amaev work led Sabol and McDonald to re-evaluate the Kass work on Nb additions to Zircaloy-4 (Reference 11). They found out-of-reactor corrosion performance improved with the Nb addition contrary to Kass’s conclusion, and this led eventually to the development of ZIRLOTM (Zr-1%Sn-1%Nb-0.1%Fe, References 12-14). Sabol reported that in a PWR environment the in-reactor corrosion rate of ZIRLOTM was ~60% of Zircaloy-4, the irradiation growth rate was ~50% of Zircaloy-4 and the creep rate was ~80% of Zircaloy-4. In addition, the hydrogen pickup characteristics of ZIRLOTM were similar to Zircaloy-4 (Reference 13). The Russian industry continued to develop knowledge on the performance of E635 and documented its performance benefits in 1996 (Reference 15). By then the composition of E635 had changed slightly from the 1970 version to have higher Sn and lower Fe with the Fe now being more similar to NSF (see Table 2-1 in Section 2.2). Nikulina et al. reported similar results for E635 as Sabol did for ZIRLOTM. Autoclave corrosion performance of E635 was better than Zircaloy-4 in 360°C water and significantly better in 500oC steam but slightly worse in 400°C steam. In-reactor corrosion of E635 under both the boiling RBMK conditions and the non- boiling VVER conditions was comparable to what is typically observed for Zircaloy-2 and Zircaloy-4 in normal BWR conditions. The out-of-reactor thermal creep rate of E635 was also found to be less than Zircaloy-4 at 385°C. The in-reactor creep rate was also lower but measured at higher temperatures than applicable to BWR conditions. The main performance improvement suggested for E635 was a dramatic decrease in irradiation growth compared to Zircaloy-4 and Zircaloy-2; it was found to be resistant to breakaway growth (see Figure 2-1). Although, the irradiation temperature was higher than typical for BWR conditions, this result supported the use of NSF as a channel material where fluence-gradient induced bow was a concern.

2-1 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

When Nikulina et al. published their work on E635 in 1996, GE had been evaluating the use of Zr-Nb based alloys in BWR conditions for approximately 15 years, and soon after this publication, GE initiated a collaborative project with the Russian industry through International Science and Technology Center (ISTC) Project #597 (Reference 16). The objective of this collaborative work was to evaluate the irradiation growth performance of NSF compared to E635 in the BOR-60 fast reactor in Dimitrovgrad, Russia. The samples were evaluated after reaching a fluence of 11 x 1021 n/cm2 (Reference 17). The samples were reinserted and irradiated in an Electric Power Research Institute (EPRI) program to a fluence of 27.8 x 1021 n/cm2 – approximately twice the end-of-life fluence experienced in a BWR (Reference 18). The results of this irradiation growth test program confirmed that NSF was resistant to the breakaway growth that occurs in Zircaloys. In parallel to this irradiation growth program with the Russian industry, GNF began inserting NSF channels into operating BWRs in 2002. As of November 2012, 35 NSF channels have been inserted into operating BWRs in seven different lead-use- channel programs. Eleven NSF channels have operated to end-of-life conditions. The data from these lead-use-channel programs and the irradiation growth program will be discussed in later sections.

2.2 COMPOSITION AND MICROSTRUCTURE OF NSF 2.2.1 Composition NSF is a zirconium alloy with additions of niobium, tin and iron. The nominal composition by weight (wt %) is: Zr (base) with 1.0% Nb, 1.0% Sn and 0.35% Fe, and shares similarity with two other alloys in the Zr-Nb-Sn-Fe family with operational experience in LWRs, namely ZIRLOTM and E635 (see Table 2-1). 2.2.2 Zirconium Zirconium is a metal with a hexagonal close-packed (HCP) crystal structure at room temperature – referred to herein as the  phase (see Figure 2-2). The lattice parameters for zirconium at room temperature are: a = 0.323 nm and c = 0.515 nm where the a-direction lies in the hexagonal close-packed basal plane and the c-direction is perpendicular to the basal plane and parallel to the basal pole (Reference 3). When the temperature increases above 863°C, alpha zirconium goes through an allotropic phase transformation where the crystal structure changes from the  phase to a body centered cubic (BCC)  phase. With further increase in temperature, zirconium melts at 1852°C. Similar to other HCP metals, the mechanical properties of zirconium and its alloys are dependent on the orientation of the crystal. This becomes important after an operation like rolling where deformation causes the grains in polycrystalline zirconium alloys to become preferentially oriented or textured. A recrystallization anneal after rolling, like that used for NSF, causes a rotation of the lattice around the basal pole but does not affect the basal pole distribution. The texture is characterized by the Kearns factors, which represent the relative fraction of crystallographic poles in the longitudinal (fL), transverse (fT) and normal (fN) orientation. For a

2-2 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

rolled zirconium alloy in the fully recrystallized state, typical Kearns factors for basal poles are [[ ]]. 2.2.3 NSF Alloy When alloying elements are added, they either go into a solid solution or precipitate out as a second phase depending on the concentration and temperature. The alloying elements in NSF, like the Zircaloys, result in a two-phase microstructure. In NSF, the Sn1, O and portions of Nb2 will be in solid solution with the -Zr matrix at operating temperatures (288°C). While the remainder of the Nb and essentially all of the Fe combine with Zr to create small second phase particles (SPPs) that are distributed uniformly within an  grain structure. The SPPs in NSF have been identified to be of the Zr(Nb,Fe)2 type at low and intermediate temperatures. The specific concentrations of the major alloying elements are controlled during the ingot melting process. The range of alloy composition for NSF is provided in Table 2-2. [[

1 Based on the phase diagram for Zr-Sn, as much as ~5%Sn (~0.04 mole fraction) can be in solution with Zr at ~288°C (Reference 19). 2 Based on the phase diagrams for Zr-Nb, approximately 0.3 wt% Nb is in solution with Zr at the operating temperatures (References 20 and 21).

2-3 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

]] Thus, it is concluded that the properties of NSF, which control channel performance, are insensitive to the range of alloy content defined in Table 2-2. 2.2.4 Microstructure The standard microstructure of NSF (except in weld-zones) in channels is a fully recrystallized grain structure with uniform distribution of SPPs. Because the microstructure is developed in the channel strip process, the microstructure will be discussed in terms of that expected for a channel strip. The normal manufacturing process to produce channel strip begins with an ingot that is triple melted for homogeneity. The ingot is then forged, hot rolled, and beta quenched. The beta quench dissolves all of the SPP’s into solid solution at temperature to precipitate them into finer, more dispersed particles upon cooling. After beta quenching, the final thickness of the strip is then produced by a sequence of hot and cold rolling to final size with intermediate and final anneals. The cumulative annealing after beta quenching determines the SPP size while the final anneal fully recrystallizes the grain structure. The current process used to manufacture NSF strip for use in channels was originally developed for Zircaloy-2 and Zircaloy-4 to optimize the SPP size and distribution for improved corrosion resistance in BWR environments. Subsequent processing of the strip into channels has no effect on the microstructure except in the seam welds, which is discussed in Section 2.2.4.3. The corrosion performance of NSF in the fully recrystallized state will be discussed in Section 2.10.

2-4 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

2.2.4.1 Grain Size Because the strip processing of NSF is based on Zircaloy-2 and Zircaloy-4 technology, the final products are in the recrystallized annealed condition and the average grain size in NSF strip is similar to these alloys. The similarity of Zircaloy-4 (Figure 2-3a) and NSF (Figure 2-3b) is shown below. The grain size is approximately [[ ]] in diameter. 2.2.4.2 Second Phase Particles

The second phase particles in NSF are intermetallics with a composition of Zr(Nb,Fe)2, which can be observed using modern electron microscopy techniques. In Figure 2-4 below, the SPP’s are shown as dark, solid circles by transmission electron microscopy. 2.2.4.3 Weld Microstructure Welding is an important joining process in manufacturing channels. For example, channel boxes are produced from two strips that are bent and seam welded full length along two opposite sides of the square box. Because the metal is melted and rapidly cooled during welding, the microstructure has a Widmanstätten grain structure (see Figure 2-5) rather than an equiaxed grain structure typical of a fully recrystallized material. In addition, the melting and solidification randomizes the texture in the weld zone. The performance of NSF with the weld microstructure will be discussed as needed in subsequent sections. 2.2.5 Texture GNF has developed a range of acceptable crystallographic texture values based on many years of performance in-reactor. This acceptable range of texture values serves as a foundation for comparison when changes in alloy elemental concentrations or material processing are made. This ensures control of texture in the manufacturing process, which can often be influenced by processing changes such as hot and cold rolling of strip more than alloy changes. Processing developed for NSF is similar to Zircaloy-2, and texture measurements taken using X-Ray diffraction techniques of the basal crystallographic pole have provided statistically similar results to Zircaloy-2 as shown in Table 2-3.

For material in the weld-zone, the texture is random and the Kearns factors are fL= fT = fN = 0.33.

2.3 ELASTIC PROPERTIES The elastic properties of solids with metallic bonding are directly related to the stiffness of the interatomic bonds. Although ultimately metallic bonding follows quantum mechanics, a simplified spring model representing the interatomic bonds may be used to predict elastic properties. With such a spring model, the elastic modulus is estimated with the following equation (Reference 25):

(2-1)

2-5 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Where 1 with n being the coordination number (or the number of closest neighbors), q is the charge of the metal ion, is the permittivity of vacuum, and is the nominal bond length. In dilute alloys such as Zircaloy-2, Zircaloy-4 and NSF, where zirconium makes up approximately 97.5 % of the alloy, the charge of a zirconium ion and the bond length of the zirconium lattice controls the elastic properties, and the difference between Zircaloy and NSF is insignificant. Thus, it is reasonable to apply the elastic properties (modulus and Poisson’s Ratio) of Zircaloy to NSF. The Young’s modulus of these zirconium alloys as a function of temperature and direction is given by:

(2-2)

2 Where Ei is the Young’s modulus in lb/in , T is the temperature in °F, A and B are constants given in Table 2-4, and i is the direction (1-through-thickness, 2-transverse or 3-rolling).

Similar to the Young’s modulus, the Poisson’s Ratio (ij) is a linear function of temperature such that

ν (2-3)

Where T is temperature in oF, A and B are constants and i and j are directions. The values of A and B are given in Table 2-5.

2.4 THERMAL EXPANSION Thermal properties, such as thermal expansion, can be temperature and texture dependent. As discussed in Section 2.2.5, texture for NSF is similar to that for Zircaloys, when manufactured using equivalent process sequences; therefore, no changes in thermal expansion occur due to texture. Thermal expansion properties are not sensitive to small changes in composition, such that Zircaloy-2 and Zircaloy-4 are properly treated without differentiation. However, thermal expansion can be composition dependent if the compositional change is large. [[

]]. In Figure 2-6, the thermal expansion of NSF as a function of temperature is shown together with that for Zircaloy-2. For each alloy, measurements were obtained from channel strip materials along the transverse and longitudinal directions. Both materials were tested in the recrystallized state. The figure also shows the thermal expansion design lines along these directions for recrystallized Zircaloys. For recrystallized Zircaloys and NSF, the thermal expansion is given by

2-6 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

i  (2-4) in / in i  T 68 F io where

Δi=change in the i dimension io=i dimension at ambient temperature (68°F) i=directional index, L, t, or r, where L=longitudinal (axial) direction, t=transverse (circumferential) direction r=radial (normal) direction

- αi=coefficient of thermal expansion in the i direction (°F 1). [[ ]]

where

fi=basal pole texture factor in the i direction, T=temperature (°F) The design lines are consistent with that for cladding (Reference 26) but are shown in Figure 2-6 with basal pole texture parameters (see Table 2-3) of [[ ]] appropriate for channel strip along, respectively, the longitudinal and transverse directions. [[

]]

2.5 SPECIFIC HEAT The specific heat (or heat capacity per unit mass) is the slope of the function between absorbed energy and temperature (i.e., Q/T on a per mass basis); typical units are J/kg-K. The specific heat of a material is mostly a function of the vibrational response of atoms (Reference 27), and when elements are combined the specific heat is modeled as the weighted sum of the constituents (Reference 28):

(2-5)

Where (Cp)A and (Cp)B are the specific heats of the constituents and x and y are the mass fractions.

2-7 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

With this relationship the specific heat of Zircaloy-2 and NSF may be assessed. Table 2-6 provides the specific heats of the constituents and the relative mass fractions of Zircaloy-2 and NSF. Using Equation (2-5), the specific heat of Zircaloy-2 is calculated to be 280.0 J/kg-K at 25°C and NSF is calculated to be 280.3 J/kg-K at 25°C. These values are within the expected measurement variability, and thus, the specific heat used by GNF for [[ ]]. The GNF model for specific heat of the alpha phase [[ ]] has the standard temperature dependence.

(J/kg·K) (2-6)

Where T is the temperature in K and A, B and C are constants with values of [[ ]], respectively. For the temperature range [[ ]], Zircaloy is in the  +  phase regime and the specific heat is given by:

(J/kg·K) (2-7)

Where [[ ]]. At temperatures greater than 1248.6 K, Zircaloy is in the beta phase and the specific heat is given by

[[ ]] (2-8)

2.6 THERMAL CONDUCTIVITY

The thermal conductivity is related to specific heat in the following way (Reference 29):

(2-9)

2 where  is the density, and DT is the thermal diffusivity, which typically has units of cm /s. Given the measured density (Section 2.7) and the specific heat of NSF are equivalent to Zircaloy, any potential difference in thermal conductivity would be related to differences in thermal diffusivity. The thermal diffusivity of NSF was measured and found to be equivalent to Zircaloy-2. A summary of those measurements and the subsequent calculated thermal conductivities are provided in Table 2-7.

2-8 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

The GNF model used for the thermal conductivity of [[ ]]. The form of the model is as follows

(2-10)

where K is the thermal conductivity in W/m-K, T is the temperature in K, and C1, C2, C3, and C4 are constants with values, [[ ]], respectively. Figure 2-7 plotted the data in Table 2-7 as a function of temperature and compared to the model in Equation (2-10) and historical data from a Knolls Atomic Power Laboratory (KAPL) report (Reference 30). The model used by GNF predicts both the historical and current data well.

2.7 DENSITY OF NSF The density of NSF was measured and found to be [[ ]].

2.8 PLASTIC PROPERTIES The yield strength and UTS are material properties that represent a material’s resistance to plastic deformation and failure by a plastic instability. The yield strength represents the stress where significant plastic deformation begins to occur, and the ultimate strength represents the stress needed to cause a plastic instability leading to failure. The tensile properties of NSF were measured at [[ ]]in the unirradiated condition and at [[ ]]in the irradiated condition (See Table 2-8). Because the strength of zirconium alloys, like NSF, decreases with increasing temperature and increases with irradiation, channels are designed for the most limiting condition assuming the material is unirradiated and the temperature is 550°F. Effectively, the most limiting condition is when the fuel is fresh and the reactor has just started. [[

]]. These measured plastic tensile properties are used to develop a model for the plastic deformation based on the following relationship:

(2-11)

where:

T is the true stress in ksi, K is the strength coefficient (ksi-sm),

is the true plastic strain (in/in), is the strain rate in (in/in-s),

2-9 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

n is the strain hardening exponent (dimensionless), m is the strain rate sensitivity exponent (dimensionless) The parameters K, n and m are material, direction, temperature, and fluence dependent parameters. [[

]]

where  is the fast flux in n/cm2-s (> 1.0 MeV) and t is time in seconds such that t is fluence in n/cm2 (> 1.0 MeV). [[

2-10 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

]]

The parameters in the relations above were determined by fitting the model to measured data3. When comparing the model predictions to measured engineering strength values (yield and UTS), the true stress is converted to engineering stress using the standard relationships:

/ (2-21)

and

(2-22)

The predicted yield strengths and UTS are compared to the measured values in Figure 2-8. The predicted yield strength is calculated at a plastic strain of [[ ]], and the UTS is predicted for the plastic strains reported in Table 2-8. 2.8.1 The Model for Design A design value for a material property is generally defined as a backoff from a nominal value. In some cases the backoff is defined in statistical terms. For example, in Reference 1, the yield strength and ultimate tensile strength used in channel design were defined as a 95/95 tolerance limit relative to the available data for unirradiated Zircaloy-4 channels at 550°F. In this case, the standard deviation was likely based on a large amount of test data collected to ensure the purchased material met the specified strength. [[

3 The relationship for R, V and m are equivalent to those documented in Reference 26, which provides the plasticity model for Zircaloy-2 and Ziron cladding.

2-11 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

]]

giving (2-24)

At the most limiting conditions [[ ]], the design yield strength is given as [[ ]]. The calculated design yield strengths and UTS are compared to the measured values in Figure 2-9. In practice, the design strength will be controlled in the specification of the material where it is defined as the minimum allowable strength. Testing of each lot documents the objective evidence that the strength of the material is greater than the design value. 2.8.2 Plastic Properties of Weld Metal The plastic deformation of weld metal was investigated by hardness testing rather than tensile testing. Typical of zirconium alloys, the hardness of the NSF-to-NSF weld microstructure was found to be greater than the hardness of NSF in the fully recrystallized state (see Table 2-10). The implication of the greater hardness is higher yield strength. Thus, it is conservative to assume the plastic deformation properties of fully recrystallized NSF for NSF with a weld microstructure.

2.9 THERMAL AND IRRADIATION CREEP Thermal creep is the time-dependent permanent deformation of a material that has an applied stress below the yield strength (which is considered the onset of time-independent deformation). If a component is loaded to a constant strain, creep has the effect of decreasing the applied stress when elastic strain is converted to permanent strain. If the component is loaded to a constant stress, permanent deformation increases over time. In traditional thermal creep, at intermediate stress levels this time-dependent permanent deformation occurs mainly by dislocation motion that results from stress-driven diffusion (a time-dependent process). Typically the strain rates of these dislocation-creep mechanisms are found to be a function of the applied stress to a power ( ∝). At lower stresses, thermal creep may also occur without dislocation motion by diffusion through grains (Nabarro-Herring creep) or around grain boundaries (Coble creep), which changes the shape of the grain – resulting in permanent deformation. The strain rates of

2-12 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

these mechanisms are linearly proportional to stress ( ∝). At high stress, creep transitions to plasticity where the mechanism is stress-driven motion of dislocations that is time-independent. The addition of a neutron flux and the accumulated damage it causes within a nuclear reactor increases the complexity of creep deformation although the general trend in mechanisms is maintained. At high stress, permanent deformation will result from time-independent motion of dislocations, and the response will follow a plasticity model as described in Section 2.8. The strain rates are such that flux does not enhance creep but the damage from the neutron flux (i.e., fluence) retards irradiation creep. At low stresses the creep rates are low enough that flux enhances creep. In addition, the creep rate is approximately linear with stress (though the mechanism is likely a flux-enhanced, stress-induced irradiation growth mechanism and not related to Nabarro-Herring or Coble creep, Reference 31). Irradiation creep at intermediate stress levels is less well understood. The general consensus appears to be that dislocation-creep mechanisms similar to thermal creep are active and the irradiation creep rate is modeled as function of stress raised to a power (Reference 31). However, the effect of flux and fluence is not fully understood. GNF models the strain rates at intermediate stresses such that fluence retards creep. Most analyses for channel components are characterized by low stress, as is discussed below. 2.9.1 Creep of NSF The thermal creep of NSF or NSF-type alloys at stresses where dislocation creep dominates is lower than Zircaloy (References 15 and 32). Thus in-reactor, at stress levels where deformation is controlled by dislocation motion, the creep of NSF would likely be less than Zircaloy. Data for ZIRLOTM (Reference 14) and a NSF-like alloy (Reference 33) support this conclusion. [[

]]. Creep bulge in channels occurs because of the differential pressure between the inside and outside of the bundle (see schematics in Figure 1-2 and Figure 2-10). The pressure drop is effectively a constant stress on the channel face that induces an elastic bulge that over time results in permanent strain (i.e., creep bulge). [[

]] The stress in the channel that controls the creep bulge may be estimated at the maximum deflection () as

2-13 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

∆ (2-25)

Given typical pressure drop [[

]]. 2.9.2 GNF Creep Models for NSF The models used in design to predict NSF creep deformation of channel components follow those used for cladding in the recrystallized state, which are documented in Reference 26. [[

2-14 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

]]

2.10 CORROSION As with any in-reactor material the first and foremost performance requirement is that the material withstands corrosion to the extent that it maintains structural integrity, and thus, maintains its ability to perform its design requirements. For channels, maintaining structural integrity is the only corrosion performance requirement, which in practice means that the component must maintain a minimum thickness of metal. The design basis for channel metal thinning due to corrosion takes into consideration the differences in densities for the metal and oxide. The metal loss from corrosion is defined as

2-15 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

(2-31) .

Where Ztot is the total zirconium oxide thickness in microns for corrosion on one surface, the factor 2 accounts for two-sided corrosion that occurs on channel components, and the factor 0.63 accounts for the density differences between the oxide and the metal. The total oxide thickness is given by the relationship

(2-32) where: Zth = thermally activated component of the oxidation (temperature dependent), and Zirr = the irradiation enhanced component of the oxidation (temperature independent). For corrosion of Zircaloy-2 and Zircaloy-4, the upper bound thermally activated component is given by:

[[ ]] where T is the temperature in degrees Celsius and t is the hot core residence time in years. The upper bound irradiation enhanced component at 288°C for corrosion of Zircaloy-2 is given by:

[[ ]]

The variation of the total zirconium oxide thickness, Ztot, with in-reactor operating time for Zircaloy-2 is shown in Figure 2-12. The figure includes recent hot cell oxide data for both Zircaloy-2 and NSF, which indicates significant margin between the measured corrosion and the design value for both alloys. Because of this significant margin, [[ ]] 2.10.1 Effect of Pre-Oxidation [[

2-16 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

]]

2.11 IRRADIATION GROWTH Irradiation growth is a unique deformation mechanism in the realm of material performance in that growth strain due to neutron irradiation occurs absent an applied stress. Because of this, irradiation growth is also known as stress-free growth. Irradiation growth is mainly attributed to the anisotropic redistribution of irradiation-induced vacancies and interstitials into dislocation loops on preferred crystallographic planes. Because the interstitial loops cause a positive growth strain and collapsed-vacancy loops cause a contraction, the volume change of irradiation growth is essentially zero (Reference 34). Models for predicting irradiation growth strain () are a function of fluence (t), which is a measure of the amount of irradiation-induced vacancies and interstitials, and the Kearns texture factor (f), which is the fraction of basal poles in a given direction (z - longitudinal, transverse, normal or weld (i.e., beta-quenched)):

(2-35)

In this model, f(t) is an empirical function that captures the phenomenological observations that the irradiation growth rate is initially relatively high, then decreases to almost zero and then increases again to a relatively high rate. The initial growth strain occurs when a-type dislocation loops form; these a-type dislocation loops saturate at low fluences (~1 to 1.5 x 1021 n/cm2), which causes the growth rate to decrease to nearly zero; when the fluence reaches the breakaway fluence (~5 - 8 x 1021 n/cm2), c-type dislocations start to form and the growth rate increases (Reference 35). The third term, f(T), is an empirical function that provides a temperature dependence when calculating irradiation growth at temperature other than 288°C. At 288°C, the value of this function is 1.0. Because irradiation growth of channels occurs at 288°C, this temperature dependence is not considered herein. [[

]]. This model is shown in Figure 2-13 compared to growth data

2-17 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

for Zircaloy-2. Actual growth measurements from irradiated Zircaloy-2 channels are also shown. The channel growth data is consistent with the observation of breakaway growth. In contrast to Zircaloy, [[

]]. Recently, EPRI has reported irradiation growth from a series of alloys supporting the conclusion that [[

]] These observations were consistent with those reported by Shishov et al. (Reference 22), who found a similar correlation of resistance to irradiation growth and less frequent c-type dislocations for Zr-Nb-Sn-Fe alloys containing a range of Fe (0.35 – 0.65%) and a range of Nb (0.7 – 2.0%).

2.12 FATIGUE The fatigue behavior of Zircaloy is well characterized in References 37 and 38 (see Figure 2-16), and was previously described to, and approved by, the NRC in Reference 26. The relationship used to describe the data follows the relationship  = bNk; where two relationships are used, one for low cycle and one for high cycle fatigue failure. The total (elastic plus plastic) strain amplitude, ε, for a cyclic loading life of N cycles for Zircaloy is as follows: [[

]]

2-18 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

To evaluate the fatigue performance of NSF, an hourglass shaped fatigue specimen that GNF developed for sheet material in the 1980’s was utilized. For comparison purposes, the fatigue testing included both unirradiated and irradiated NSF and Zircaloy-2 channel materials. The irradiated channel coupons were [[

]]

2.13 STRESS RUPTURE Stress rupture characterizes the time to creep failure due to an applied load on a component. For channels, the differential pressure across the channel wall induces a “hoop” stress in the channel face (see Figure 2-10). Thermal and irradiation creep under the induced stresses result in channel face distortion or channel bulge. GNF design criteria for channels include the long term cumulative effect of creep considering a conservative estimate of time at an applied stress over the operating life history of the channel. The relationship between applied stress, temperature, fluence and time to failure shown in Equation (2-40) is derived from open literature data as well as GNF test data and is considered conservative when true stress levels remain below 90% of the ultimate strength of the material. To validate NSF performance relative to other Zircaloys, testing was performed on channel material in the irradiated and unirradiated conditions at temperatures consistent with normal operations (288C) and transient temperatures (400°C). [[

]] The test applies a static load and constant temperature to the test specimen until stress rupture was reached, determining the time to failure due to stress rupture.

2-19 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Equation (2-40) below, provided in Reference 1 has been used to conservatively predict time to failure for Zircaloy channels. This predictive equation utilizes a Larson-Miller parameter approach to determine the relationship between applied stress, temperature and fluence and time to failure. [[

]] The creep performance of NSF and similarly processed Zircaloy has been discussed previously where testing indicates unirradiated NSF has a slightly higher strength than unirradiated Zircaloy-2. [[

]] The data also show that both alloys are conservatively bound by Equation (2-40). Similarly, testing on irradiated Zircaloy-2 and NSF, shown in Figure 2-19, at different temperatures and fluences demonstrates Equation (2-40) conservatively predicts time to failure for either alloy. [[ ]].

2.14 HYDROGEN EFFECTS Corrosion is the major source of hydrogen absorbed by zirconium alloys. At certain levels, this absorbed hydrogen may degrade mechanical properties. The level of hydrogen absorbed

(measured in wppm) is directly proportional to the corrosion thickness (Ztot) and the hydrogen pickup fraction (HPUF) while inversely proportional to the thickness of the component (tm). The hydrogen pickup in an NSF channel was measured and compared to a Zircaloy-2 channel that operated in symmetric locations throughout life (see Figure 2-20). The measured hydrogen (absorbed) in wppm is plotted relative to the hydrogen generated from corrosion in wppm using the metal as the mass basis for the normalization. [[

]].

2-20 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

The concern with absorbed hydrogen in zirconium alloys is the potential detrimental effect on mechanical properties. In this respect, NSF is expected to be similar to other zirconium alloys. Hydrogen may detrimentally affect mechanical properties when the hydrogen concentration increases above the solubility limit and hydrides form. However, hydride embrittlement depends on the temperature (Reference 20) and on whether failure is caused by a plastic instability or by crack propagation; there is less of an effect on failure by a plastic instability than crack propagation (References 39 and 40). Yunchang and Koss (Reference 39) investigated the microstructural effects of hydrides on failure by a plastic instability. They found that hydrides promote ductile void growth and coalescence but have little effect on the onset of void formation, which occurs after the instability at the UTS. Similarly, Kuroda et al. (Reference 40) found the burst pressure in a fuel rod was independent of hydrogen (up to 461 wppm) when the hydrides were uniformly distributed and oriented perpendicular to the radial direction (or perpendicular to the through-thickness direction as is the case in a channel component – see Figure 2-21). Hydrides do potentially have a larger effect when failure is caused by crack propagation. In this case, a propagating crack will seek the path of least resistance. As zirconium hydrides are generally more brittle than the Zr matrix (especially at temperatures below 100°C, Reference 20), a propagating crack will tend to jump from hydride to hydride. The result is that the effective toughness of a zirconium alloy with hydrides may decrease with increasing hydrogen (Reference 20). This is mainly an issue when the hydrides are oriented parallel to the failure plane, which is not the case in channels where the hydrides are oriented perpendicular to failure in the through-thickness direction. 2.14.1 Hydrogen Considerations in Design of NSF Components [[

]].

2.15 NUCLEAR PROPERTIES The nuclear properties of NSF have been assessed. The affect can be described as an effect on the infinite lattice eigenvalue determined by lattice physics calculations due to the differences in alloy isotopic content. The difference in infinite lattice eigenvalue between Zircaloy-4 and NSF was found to be of order [[

]]

2-21 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Table 2-1 Nominal Chemical Weight Percent Composition of Zr-Nb-Sn-Fe Alloys

E635 E635 NSF ZIRLOTM Circa 1970 Circa 1996

Nb 1.0 1.0 1.0 1.0

Sn 1.0 1.0 1.0 1.25

Fe 0.35 0.1 0.5 0.37

O 0.12 0.14 0.04 0.06

2-22 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Table 2-2 Composition Range for NSF

Composition Range

Nb [[

O ]]

Table 2-3 Comparison of Typical Texture Values for NSF Compared to Zircaloy-2

Alloy/Sample fN fL fT Total f

Zircaloy–2/Typical [[

NSF/Typical ]]

Note: This is a comparison of Typical Texture Values for NSF Compared to Zircaloy-2 where fN, fL, and fT are the relative fraction of basal poles in the normal, longitudinal and transverse directions.

2-23 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Table 2-4 Constants A and B for calculating the Young’s Modulus for Zircaloy-2, Zircaloy-4 and NSF

Young’s A B Modulus

E1 [[

E ]]

Note: 1 = through-thickness direction, 2 = transverse direction, 3 = longitudinal direction, and  = weld metal or Beta-Quenched

Table 2-5 Constants A and B for Calculating the Poisson’s Ratio for Zircaloy-2, Zircaloy-4 and NSF

Poisson’s A B Ratio

12 [[

13

21

23

31

32

 ]] Note: 1 = through-thickness direction, 2 = transverse direction, 3 = longitudinal direction, and  = weld metal or Beta-Quenched

2-24 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Table 2-6 Specific Heat of the Constituent Elements in Zircaloy-2 and NSF (Reference 41)

Specific Heat Specific Heat Zircaloy-2 NSF Element at 25°C at 25°C Mass Mass cal/g-K J/kg-K Fraction Fraction

Zr 0.067 280.5 .9818 .9765

Nb 0.064 268.0 - .01

Sn 0.051 213.5 .015 .01

Fe 0.106 443.8 .0015 .0035

Ni 0.106 443.8 .0007 -

Cr 0.107 448.0 .001 -

2-25 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Table 2-7 Measured Thermal Diffusivity and Calculated Thermal Conductivity of Zircaloy-2 and NSF Measured Calculated Heat Temp Thermal Density Thermal Material Capacity (°C) Diffusivity (g/cm3) Conductivity (J/kg-°K) (cm2/s) (W/m-°K) Zr-2 23 [[ Zr-2 100 Zr-2 175 Zr-2 250 NSF 23 NSF 100 NSF 175 NSF 250 ]]

2-26 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Table 2-8 Average Measured Yield and Ultimate Strength of NSF Channel Material in the Unirradiated and Irradiated (~9.5 x 1021 n/cm2) Condition Uniform 0.2% Strain Test Elongation (%) Irradiation offset UTS Orientation Rate Temp. Condition YS (ksi) Elastic Plastic (1/min) (°F) (ksi) +Plastic Only

[[

]] [[

]]

2-27 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Table 2-9 Design Yield and Ultimate Strength of NSF Channel Material ([[ ]]in/in/min)

[[

Yield Strength (ksi)

UTS (ksi) ]]

Table 2-10 Rockwell B Hardness Measurements NSF-to-NSF Weld Compared to NSF in the Fully Recrystallized State

Hardness Microstructure (Rockwell B)

Weld [[

Weld

Recrystallized

Recrystallized ]]

2-28 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

1 E635: Nikunlina et al. (Irradiation Temperature = 330-350 C) 0.9 Zr-2: Mahmood et al. (Irradiaiton Temperature = 290 C)

0.8

0.7

0.6

0.5

0.4 Growth (%)

0.3

0.2

0.1

0 0.00E+00 5.00E+21 1.00E+22 1.50E+22 2.00E+22 BWR Fluence (n/cm^2)

Figure 2-1 Irradiation Growth of E635 (Reference 15) Compared to Zircaloy-2 (Reference 42) Figure Note: In each case the fast fluence was converted to dpa and then BWR fluence assuming a 40% average void fraction. The conversion factors are provided in Reference 43.

2-29 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Figure 2-2 Hexagonal Close-Packed (HCP) Crystal Structure of Zirconium

2-30 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

(a) Zircaloy-4 at 548x

[[

]]

(b) NSF at 500x

Figure 2-3 Recrystallized Grain Structure

2-31 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 2-4 TEM Image of NSF Second Phase Particles

[[

]]

Figure 2-5 Widmanstätten Microstructure of a NSF-to-NSF Weld

2-32 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 2-6 Thermal Expansion (dL/L=Δi/io) of NSF and Zircaloy-2

Figure Note: Thermal expansion (dL/L=Δi/io) of NSF and Zircaloy-2 obtained on channel strip materials in the transverse (Trans) and longitudinal (Long) directions.

2-33 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 2-7 Thermal Conductivity of NSF Figure Note: The thermal conductivity of NSF is compared to the thermal conductivity of Zircaloy-2, the calculated thermal conductivity (Equation (2-10) and historical data on Zircaloy (Reference 30). The uncertainty on the GNF thermal conductivity data is shown as [[ ]].

2-34 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 2-8 Measured Strength of NSF (yield and UTS) Compared to the Predicted Strength Using the NSF Plasticity Model

2-35 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 2-9 Measured Strength of NSF (yield and UTS) Compared to the Design Strength Using the NSF Plasticity Model

2-36 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Figure 2-10 Schematic of Force and Deflection on a Channel Face

2-37 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 2-11 Creep Bulge Measurement for NSF and Zircaloy-2 Channels taken at the 40 in. Elevation Figure Note: Creep bulge measurement for NSF and Zircaloy-2 channels taken at the 40 inch elevation, where typically the bulge is a maximum.

2-38 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 2-12 Upper bound Design Value for Corrosion and Recent Hot Cell Oxide Thickness Data for Zircaloy-2 and NSF Figure Note: Uncertainty bars represent one standard deviation.

2-39 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 2-13 GNF Channel Growth Data and NSF Irradiation Growth Data Figure Note: GNF channel growth data and NSF irradiation growth data collected at BOR-60 plotted relative to open literature data for Zircaloy-2. The fluence of the data collected at BOR-60 (Reference 18) and in a PWR (Reference 42) has been corrected for the BWR neutron energy distribution at 40% void (See Reference 43).

2-40 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] ]] Figure 2-14 Irradiation Growth Data on a Series of Zr-Nb-Sn-Fe Alloys Figure Note: Irradiation growth data on a series of Zr-Nb-Sn-Fe alloys that were irradiated at BOR-60 and reported in Reference 18. The BOR-60 fluences have been corrected to be equivalent to the BWR neutron energy distribution at a 40% void fraction (See Reference 43).

2-41 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 2-15 Irradiation Growth Data on a Series of Zr-1Nb-1Sn Alloys Figure Note: Irradiation growth data on a series of Zr-1Nb-1Sn alloys with varying amounts of Fe that were irradiated at BOR-60 and reported in Reference 18. The BOR-60 fluences have been corrected to be equivalent to the BWR neutron energy distribution at a 40% void fraction (See Reference 43).

2-42 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] Figure 2-16 Fatigue Data for Zircaloy Used in Fatigue Evaluations for Cladding (References 26, 37, and 38)

2-43 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] Figure 2-17 NSF and Zircaloy-2 Fatigue Data Figure Note: NSF and Zircaloy-2 fatigue data generated with an hourglass shaped fatigue sample plotted relative to the NRC accepted mean and upper/lower 95/95 boundaries for Zircaloy cladding (Reference 26).

2-44 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 2-18 Comparison of Zircaloy-2 and NSF Tested Using Similar Applied Stresses Figure Note: Comparison of Zircaloy-2 and NSF tested using similar applied stresses at 752°F (400°C) compared to the stress rupture predictive equation.

2-45 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 2-19 Comparison of Irradiated Zircaloy-2 Tested at 550°F(288°C) and Irradiated NSF Tested at 752°F(400°C) Figure Note: Comparison of irradiated Zircaloy-2 tested at 550°F (288°C) and irradiated NSF tested at 752°F (400°C) relative to their respective predictive equation for stress rupture.

2-46 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 2-20 Measured Hydrogen Pickup in a NSF Channel Figure Note: Measured hydrogen pickup in a NSF channel compared to a Zircaloy-2 channel operated in symmetric locations. The mass basis for normalizing the calculated hydrogen generated in wppm is the metal. [[

]]

2-47 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 2-21 Photomicrograph of Precipitated Hydrides Figure Note: Photomicrograph of precipitated hydrides oriented perpendicular to the through thickness direction in a NSF channel component (152 wppm H).

2-48 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

3.0 CALCULATING CPR WITH NSF CHANNELS 3.1.1 GNF Methodology for Critical Power Ratio Calculations This section describes the current method for calculating CPR, specifically highlighting the dependence of R-factor on channel bow. This section is provided for informational purposes only and is intended to provide clarification on the method for establishing dependence of bundle R-factors and CPRs on channel bow as documented in Reference 44. 3.1.2 Critical Power Ratio and R-Factor Methodology The method for calculating CPR and bundle R-factors is described in NEDC-32505P Revision 1 (Reference 45) for GE11, GE12, and GE13. This method is identical to that described in NEDC- 33292P Revision 3 (Reference 46) for GNF2 fuel and NEDC-32851P Revision 5 for GE14 fuel (Reference 47). In these documents, it is stated that the effects of channel bow are accounted for on an individual fuel rod power basis. 3.1.3 R-Factor Calculational Process As described in Reference 45, the steps used in the R-factor calculational process are as follows: 1. Obtain relative 2D rod-by-rod power distributions from the lattice physics code. These values are a function of lattice nuclear design, average exposure, void fraction and control state. 2. [[

]] 3. Calculate an R-factor for each individual fuel rod. [[ ]] 4. The bundle R-factor is the maximum value from all individual rod R-factors. 5. Repeat these calculations for each desired bundle average exposure, control fraction and channel bow. [[

3-1 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

]] As demonstrated in this clarification, the amount of channel bow serves only as an input to these calculations; therefore, implementation of NSF will have no effect on the process to calculate either CPR or bundle R-factors. No change to process related to either CPR or R-factor calculations is intended or required based on this LTR approval. The only change being recommended [[ ]] 3.1.4 Current Procedure for Assigning Bow-Specific R-factors [[

]]

3-2 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

The shadow-corrosion bow phenomenon was not recognized until after the development of the current process. However, because shadow-corrosion bow results in the channel bowing back towards the control blade, it generally reduces the total amount of bow, which on average is away from the blade. In extreme cases, shadow-corrosion bow can result in the water gap decreasing on the blade side of individual bundles and increasing on the opposite side. [[

]] 3.1.5 Bow-Specific R-Factor Calculations Procedure for NSF Channels As described above, the current procedure for determining BOWAVE considers effects of initial as-manufactured channel bow and fluence-induced bow. [[ ]] For clarity, the variable CACABO will be used to represent the core-average, cell-average bow as calculated by the 3-D core simulator. [[ ]] The models for calculating channel bow and CACABO that are used in the 3-D core simulator are provided for information in Appendix A. [[

3-3 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

]] versus using the calculated value of CACABO from the 3D core simulator has been quantified. The absolute difference in the Maximum Fraction of Limiting CPR (MFLCPR) for a given plant cycle as a function of exposure is used to demonstrate this effect. Figure 3-2 illustrates this difference for [[ ]] [[

]] The CACABO uncertainty component that feeds into the safety limit minimum CPR (SLMCPR) is plotted in Figure 3-3 versus cycle exposure for [[ ]]. This has been determined using the method described in Appendix A. As seen from the figure, the uncertainty is well below [[ ]]. A [[ ]] is used for the roll-up of the bow uncertainty in the SLMCPR calculations (References 44 and A-2). This [[ ]] will continue to be used as [[ ]]for NSF channels for the roll-up of the uncertainty in the SLMCPR calculations. No credit is taken for the [[ ]] observed in the figure for the NSF channels.

3.2 NSF APPLICATION TO LHGR CALCULATION

[[

]]

3-4 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 3-1 Calculated Values of CACABO for NSF Cores

3-5 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 3-2 Absolute MFLCPR Difference Figure Note: [[ ]]

3-6 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure 3-3 Uncertainty in the Core-Average, Cell-Average Bow as a Function of Cycle Exposure

3-7 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

4.0 APPLICABILITY

Upon approval of this licensing topical report, NSF may be incorporated into GNF fuel designs in channels by inclusion in the GESTAR new fuel compliance reports for a specific fuel design, as supported by appropriate analysis using the properties described herein. As outlined in this LTR, NSF channels have channel bow performance characteristics that are distinctly different than those observed for Zircaloy channels used in CPR evaluations. It is therefore inappropriate to continue using only the Zircaloy-specific, [[

]]

4-1 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

5.0 SUMMARY

This licensing topical report documents the material properties of NSF for application in channels as part of GNF fuel designs. These properties are used in evaluating the adequacy of components designed with NSF. In general, the properties of NSF are the same as Zircaloy. Where the properties differ, data supported material design values are provided. In addition, this licensing topical report documents recommendations related to the bow assessment and application approach in the CPR calculation for NSF channels.

5-1 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

6.0 REFERENCES

1. General Electric Company, “BWR Fuel Channel Mechanical Design and Deflection,” NEDE-21354-P, September 1976.

2. General Electric Company, “BWR Fuel Assembly Evaluation of Combined Safe Shutdown (SSE) and Loss-of-Coolant Accident (LOCA) Loadings (Amendment 3),” NEDE-21175-3-P-A, October 1984.

3. McGraw-Hill, “Metallurgy of Zirconium,” Lustman, B. and Kerze, F., 1955.

4. Proceedings of the 1st International Conference on the Peaceful Uses of Atomic Energy, “Aqueous Corrosion of Zirconium and Its Alloys at Elevated Temperatures,” Volume 9, United Nations, New York, and IAEA, Vienna, Thomas, D.E., p.407, 1956.

5. Proceedings of the 2nd U.N. International Conference on Peaceful Uses of Atomic Energy, “Mechanical Properties and Corrosion Resistance of Zirconium and its Alloys in Water, Steam and Gases at Elevated Temperature,” Volume 5, Geneva, CH, Ambartsumyan, R.S. et al., P/2044, 1958.

6. Proceedings of the 2nd U.N. International Conference on Peaceful Uses of Atomic Energy, “Structure and Properties of Zirconium Alloys,” Volume 5, Geneva, CH, Ivanov, O.S. and Grigorovich, V.K., P/2046, 1958.

7. UKAEA Harwell, “The Oxidation and Corrosion of Zirconium and Its Alloys: XV Further Studies of Zirconium-Niobium Alloys,” Report AERE-R 4134, Cox, B., 1962.

8. Atomic Energy of Canada, Ltd., “Corrosion and Hydriding Behaviour of Some Zr 2.5wt% Nb Alloys in Water, Steam and Various Gases at High Temperature,” Chalk River, Dalgaard, S.B., AECL-1513, 1962.

9. Westinghouse Electric Corp., Bettis Atomic Power Laboratory, “The Corrosion and Hydrogen Absorption Properties of Zircaloy-4 Alloys Containing Additions of Niobium,” WAPD-TM-647, Kass, S., 1967.

10. Proceedings of the 4th UN International Conference on Peaceful Uses of Atomic Energy, “Corrosion Behavior of Zirconium Alloys in Boiling Water Under Irradiation,”

6-1 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Volume 10, United Nations, New York, and IAEA, Vienna, Amaev, A.D., et al., p. 537, 1972.

11. Nuclear Science and Engineering, “The Effect of Niobium Additions on the Corrosion Behavior of Zircaloy-4,” Volume 63, Sabol, G.P. and McDonald, G.G., p. 83, 1977.

12. 10th International Symposium Symposium on Zirconium in the Nuclear Industry, “In- Reactor Corrosion Performance of ZIRLOTM and Zircaloy-4,” ASTM STP 1245, Sabol, G.P., et al., p. 724, 1994.

13. Proceedings, ANS International Topical Meeting on Light Water Reactor Fuel Performance, “In-Reactor Fuel Cladding Corrosion Performance at Higher Burnups and Higher Coolant Temperatures,” Portland, OR, Sabol, G.P., et al., p. 397, 1997.

th TM 14. 14 International Symposium on Zirconium in the Nuclear Industry, “ZIRLO – An Alloy Development Success,” ASTM STP 1467, Sabol, G.P., p. 3, 2005.

15. 11th International Symposium on Zirconium in the Nuclear Industry, “Zirconium Alloy E635 as a Material for Fuel Rod Cladding and Other Components of VVER and RBMK Reactors,” ASTM STP 1295, Nikulina, A.V., et al., p. 785, 1996.

16. Moscow Engineering Physics Institute, “Fundamental Research of Materials Structure and Properties Changed Resulted from Irradiation by Means of Complex of Modern Physical Methods,” ISTC Project #597, October 2006.

17. 15th International Symposium on Zirconium in the Nuclear Industry, et al., “Irradiation- Induced Growth and Microstructure of Recrystallized, Cold Worked and Quenched Zircaloy-2, NSF, and E635 Alloys,” ASTM STP 1505, Kobylyansky, G. P., p. 564, 2009.

18. EPRI, “The NFIR-V Dimensional Stability Project: BOR-60 Irradiation and Growth Data,” EPRI Report 1021035, June 2010.

19. Computer Coupling of Phase Diagrams and Thermochemistry, “The Zr-Sn Binary System: New Experimental Results and Thermodymic Assessment,” Volume 32, Perez, R.J. et al., pp. 593-601 (2008).

6-2 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

20. International Atomic Energy Agency, “The Metallurgy of Zirconium,” Vienna, Douglass, D.L., 1971.

21. Journal of Nuclear Materials, “Effect of β Phase, Precipitate and Nb-Concentration in Matrix on Corrosion and Oxide Characteristics of Zr-xNb Alloys,” Volume 317, Jeong, Y.H., et al., pp.1-12, 2003.

22. 14th International Symposium on Zirconium in the Nuclear Industry, “Influence of Structure – Phase State of Nb containing Zr Alloys on Irradiation-Induced Growth,” ASTM STP 1467, Shishov, V.N. et al., p. 666, 2005.

23. 11th International Symposium on Zirconium in the Nuclear Industry, “Influence of Processing Variables and Alloy Chemistry on the Corrosion Behavior of ZIRLO Nuclear Fuel Cladding,” ASTM STP 1295, Comstock, R.J., Schoenberger, G. and Sabol, G.P., p. 710, 1996.

24. 13th International Symposium on Zirconium in the Nuclear Industry, “Alternative Zr Alloys with Irradiation Resistant Precipitates for High Burnup BWR Applications,” ASTM STP 1423, Garzarolli, F., et al., p. 119, 2002.

25. Pergamon Press, “Engineering Materials 1: An Introduction to their Properties and Applications,” Chapter 6, Ashby and Jones, 1980.

26. Global Nuclear Fuel, “Licensing Topical Report: The Prime Model for Analysis of Fuel Rod Thermal – Mechanical Performance; Part 1 – Technical Bases,” NEDE-33256P-A, Revision 2, September 2010.

27. Hemisphere Publishing Corporation, “Introduction to Metallurgical Thermodynamics,” 2nd Edition, Gaskell, D.R., pp. 114-122, 1981.

28. John Wiley and Sons Inc., “Thermodynamics of Solids,” Second Edition, Swalin, R.A., p. 83, 1972.

29. John Wiley and Sons, Inc. “Materials Science and Engineering: An Introduction,” 4th Edition, Callister, W.D., pp. 657, (1997).

6-3 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

30. Knolls Atomic Power Laboratory, “Application of the Ewing Equation for Calculating Thermal Conductivity from Electrical Conductivity,” KAPL-2146, April 7, 1961.

31. Franklin, D.G., Lucas, G.E., and Bement, A.L., “Creep of Zirconium Alloys in Nuclear Reactors,” STP 815, ASTM (1983).

32. Nuclear Engineering and Design, “Transitions in Creep Mechanisms and Creep Anisotropy in Zr-1Nb-1Sn-0.2Fe Sheet,” Volume 156, Murty, K.L. et al., p. 359, 1995.

33. Journal of Nuclear Materials, “Irradiation-Induced Microstructural Changes in Zr – 1% Sn – 1% nb – 0.4% Fe,” Volume 238, Nikulina, A.V. et al., p 205, 1996.

34. 6th International Symposium on Zirconium in the Nuclear Industry, “Irradiation Growth of Zircaloy (LWBR Development Program),” ASTM STP 824, Willard, H.J., p. 452, 1984.

35. 15th International Symposium on Zirconium in the Nuclear Industry, “In-Reactor Deformation of Zirconium Alloy Components,” ASTM STP 1505, Holt, R.A., p. 3, 2009.

36. EPRI, “The NFIR-V Dimensional Stability Project: Post-Irradiation Examination Results on Material Variants Irradiated in BOR-60,” EPRI Report 1022905, December 2011.

37. Nuclear Science and Engineering, "Fatigue Design Basis for Zircaloy Components," Volume 20, O’Donnell, W.J. and Langer, B.F., p.1, 1964.

38. Journal of Nuclear Materials, "Low Cycle Fatigue Properties of Zircaloy Cladding," Volume 56, Pettersson, K., pp. 91-102, 1975.

39. Metallurgical Transaction. A, “The Influence of Multiaxial States of Stress on the Hydrogen Embrittlement of Zirconium Alloy Sheet,” Volume 16A, F. Yunchang and D. Koss, p. 675 (1985).

40. Nuclear Engineering and Design, “Analysis of the Fracture Behavior of Hydrided Fuel Cladding by Fracture Mechanics,” Volume 203, M. Kuroda, et al. p. 185, (2001).

41. CRC Press, “CRC Handbook of Chemistry and Physics,” 71st Edition, pp. 5-68 – 5-69, (1990 – 1991).

6-4 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

42. Zirconium in the Nuclear Industry: Twelfth International Symposium, "Post-Irradiation Characterization of Ultra-High-Fluence Zircaloy-2 Plate," ASTM STP1354, Mahmood, S. et al. p. 139, 1998.

43. EPRI, “NFIR Dimensional Stability Project: A Method for Transposing Test Reactor Irradiation Data for PWR and BWR Applications,” EPRI Report 1019098, October 2009.

44. Letter and attachments to Robert C. Jones, Chief Reactor Systems Branch - USNRC, ‘‘Fuel Channel Bow Assessment,” November 15, 1989, JSC89115, MFN086-89.

45. GE Nuclear Energy, “R-Factor Calculation Method for GE11, GE12 and GE13 Fuel,” NEDC-32505P-A, Revision 1, July 1999.

46. Global Nuclear Fuel, “GEXL17 Correlation for GNF2 Fuel,” NEDC-33292P, Revision 3, April 2009.

47. Global Nuclear Fuel,“GEXL14 Correlation for GE14 Fuel,” NEDC-32851P-A, Revision 5, April 2011.

48. Global Nuclear Fuel, “The PRIME Model for Analysis of Fuel Rod Thermal – Mechanical Performance Part 3 – Application Methodology Licensing Topical Report,” NEDC-33258P-A, Revision 1, September 2010.

49. Global Nuclear Fuel, “Applicability of GE Methods to Expanded Operating Domains Licensing Topical Report,” NEDC-33173P-A, Revision 1, September 2010.

50. Letter from Mirela Gavrilas (NRC) to Jerald G. Head (GEH), Subject: Final Safety Evaluation for Global Nuclear Fuel - Americas, LLC Licensing Topical Report NEDE 33798P, "Application of NSF to GNF Fuel Channel Designs" (TAC No. MF0742), September 1, 2015.

51. Letter from Brian R. Moore (GNF) to NRC Document Control Desk, Subject: Response to Request for Additional Information Regarding Review of Licensing Topical Report NEDE 33798P, “Application of NSF to GNF Fuel Channel Designs” (TAC No. MF0742), MFN 15-040, June 10, 2015.

6-5 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

52. Letter from Brian R. Moore (GNF) to NRC Document Control Desk, Subject: Modified Pages for NEDE-33798P, “Application of NSF to GNF Fuel Channel Designs”, MFN 13-008 Sup 1, June 12, 2013.

6-6 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

APPENDIX A CALCULATING CORE-AVERAGE, CELL-AVERAGE BOW FOR NSF CHANNELS

A-1 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

A.1 CORE-AVERAGE CELL-AVERAGE BOW (CACABO) This appendix describes the methodology utilized to calculate CACABO for a plant cycle and presents models validating the approach. The sections are divided as described below: 1. Section A.2 – Description and validation of channel bow calculation methodology on an individual bundle basis for NSF, including, a. Initial manufactured channel bow, described in Section A.2.1 b. Fluence-induced bow, described in Section A.2.2 c. Shadow-corrosion induced bow, described in Section A.2.3 d. Qualification of the individual channel bow models for NSF, Section A.2.4 2. Section A.3 - Description of cell-average bow calculation methodology, 3. Section A.4 - Description of core-average, cell-average bow calculation methodology.

A.2 INDIVIDUAL CHANNEL BOW CALCULATION APPROACH The total bow of an individual channel results from the cumulative effect of the initial manufactured bow, fluence induced bow, and shadow bow.

A.2.1 INITIAL MANUFACTURED CHANNEL BOW During the manufacturing process, some nominal amount of channel bow is inherently introduced into the finished product. For this reason, channels with no irradiation history (beginning of life channels) still are characterized by some amount of channel bow. [[

]]

A.2.2 FLUENCE-INDUCED BOW CALCULATION Fast fluence (E>1MeV) gradient induced bow results from differential growth of channel material on opposite channel faces. The growth is towards the longer side to maintain a minimal stress condition. The steps to calculate fluence-induced bow of a channel can be summarized as follows: 1. Define the accumulated channel wall fluence, as described in Section A.2.2.1, 2. Calculate the irradiation growth strain of the channel wall segments due to the presence of fast fluence, as described in Section A.2.2.2. The growth strain is given in a dimensionless form (in/in), where the denominator is the length over which the strain is applicable. 3. With the growth strain known over a given length (segment) and for all sides of the channel, it is then possible to calculate the fluence-induced bow of this channel, as described in Section A.2.2.3.

A-2 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

A.2.2.1 ACCUMULATION OF CHANNEL FLUENCE GNF lattice-physics codes calculate the fraction of fast flux (flux above 1MeV) as a function of void and exposure for each lattice in the core. GNF core simulators take these values and calculate the channel surface-center flux above 1MeV on a channel surface, node specific basis. When performing this assessment, the channel numbering (n=1,..,4) is with respect to control blade position, as shown in Figure A-1. Using this numbering scheme, the channel surface, node specific flux (FLUXS(k,i,j,n)) is tracked as a function of the node center fast flux, node width, water gap size, diffusion coefficients, and discontinuity factors. These channel surface fluxes are used to track channel surface fluence over a cycle on a nodal basis as follows:

, , , , , , ∗ , , , (A-1) where: ,,, Is the surface-center fluence (n/cm2) 2 ,,, Is the surface-center fluence from prior exposure step (n/cm ) DT Is the time in seconds consistent with the exposure increment FLUXSk,i,j,n Is the channel surface, node specific flux (n/cm2-s)

A.2.2.2 IRRADIATION GROWTH STRAIN

Irradiation growth strain is defined as a function of channel fluence. It is expressed as  n for face n in the following section when describing the channel bow calculation process.

A.2.2.3 EVALUATION OF FLUENCE CHANNEL BOW The evaluation of the bow of a single channel is performed in a step-wise fashion for each channel surface pair (n=(1,3) and n=(2,4)), as outlined below: 1. [[

A-3 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

]]

A.2.2.4 CALCULATION OF DEFLECTION OF A CHANNEL SEGMENT For a single segment, as shown in Figure A-3, the amount of fluence induced channel bow can be defined as by Equation (A-4). [[

A-4 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

]]

A.2.2.5 CALCULATION OF CHANNEL BOW AS A FUNCTION OF AXIAL HEIGHT The fluence induced bow of a channel at a given axial height is a function [[

A-5 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

]]

A.2.3 SHADOW-CORROSION INDUCED BOW CALCULATION Shadow corrosion is an enhanced irradiation corrosion mechanism that occurs on zirconium alloys when a dissimilar material (such as a control blade) is near the zirconium surface (such as a BWR channel) and the water chemistry is oxygenated. When a fuel bundle is controlled early in life, the increased corrosion on the blade side relative to the non-blade side results in a

A-6 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public) difference in hydrogen absorbed in channel material. Hydrogen is absorbed into the metal as part of the corrosion process and causes a volume change resulting in channel bow. Because direct measurement of shadow corrosion-induced bow is only possible when the fluence gradient is zero, shadow bow is generally observed by accounting for the fluence gradient induced bow. After accounting for fluence bow in the data, the end-of-life channel bow correlates well with the total effective control blade exposure (ECBE). Prediction of shadow bow involves two parts. The first part is calculation of ECBE for each channel (this is a measure of the susceptibility to shadow-corrosion induced bow). The second part is using an empirical correlation to convert the ECBE to a corresponding shadow-induced bow. Both parts are described in more detail in the following sections.

A.2.3.1 ACCUMULATION OF CONTROLLED TIME ECBE is the summation of insertion length in inches times the length of control period in days, weighted by a weight factor, as described in the equation below. [[

]]

A.2.3.2 EVALUATION OF SHADOW CORROSION-INDUCED BOW

[[

A-7 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

]] [[

]]

A.2.4 QUALIFICATION OF FLUENCE AND SHADOW BOW MODELS FOR NSF The purpose of this section is to provide the evidence necessary to qualify the channel bow models that will be used in calculating the core-average, cell-average bow, which is an input to calculate the CPR. [[ ]] The relative range of exposures and ECBE are provided in Figure A-5. The NSF channel data reasonably represents the operating conditions experienced by channels over their lifetimes. The evaluation of the channel bow data is complex because two independent mechanisms cause bow. The evaluation approach taken is to first qualify the predicted bow when only fluence gradient-induced bow is active [[ ]]and then evaluate shadow corrosion-induced bow as a function of ECBE after subtracting the effect of fluence gradient-induced bow. The measured bow is plotted versus the predicted fluence gradient-induced bow (Figure A-6) using the model described in Section A.2.2. The line through the data represents the one-to-one relationship between measured and predicted bow. [[

]]

A-8 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

When the shadow corrosion-induced bow mechanism was being recognized, one of the most important observations was that there was a positive bias in the residuals of the fluence gradient- induced bow predictions when channels experienced early-life control (Reference A-1). [[

]] This bias in the residual is now considered the inferred shadow bow and is found to be correlated to ECBE (Figure A-9). The model described in Section A.2.3 is also provided for comparison to the data.

A-9 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

2 1

1 1 3 4 2 2 control blade 4 3 channel orientation number 3 4 2 4 4 3 3 1 channel face number 1 2

Figure A-1 Channel Face Numbering for the Bow Model

[[

]]

Figure A-2 Calculation of Segment Edge Bow

A-10 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure A-3 Fluence Induced Bow

[[

]]

Figure A-4 Channel Bow Calculation

A-11 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] Figure A-5 Range of ECBE and Exposure in the Channel Distortion Database for NSF Channels

A-12 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure A-6 Measured Bow Plotted as a Function of Predicted Fluence Bow for NSF Channels

A-13 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] Figure A-7 Residual (Measured – Predicted) of Fluence Gradient-Induced Bow Data Figure Note: Residual (measured – predicted) of fluence gradient-induced bow data [[ ]] for the NSF channel database.

A-14 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] Figure A-8 Residual (Measured – Predicted Fluence Bow) of Fluence Gradient-Induced Bow Data Figure Note: Residual (measured – predicted fluence bow) of fluence gradient-induced bow data [[ ]]and data susceptible to shadow corrosion-induced bow [[ ]] for the NSF channel database. The data available for NSF [[ ]].

A-15 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure A-9 Inferred Shadow Bow (Measured – Predicted Fluence Bow) as a Function of ECBE for the NSF Database Figure Note: Only data from channels with exposures greater than [[ ]] are included because this represents the exposure where shadow bow has fully accumulated.

A-16 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

A.3 CELL-AVERAGE BOW CALCULATION

The mean bow based on manufacturing, fluence, and shadow corrosion for a given channel at core position (i,j) is obtained using the following formulas:

[[

]]

This method changes the cell-average bow by including a [[

]]

A-17 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

A.4 CORE-AVERAGE CELL-AVERAGE BOW AND UNCERTAINTY CALCULATION The cell-average bows are combined to determine the core-average, cell-average bow (CACABO) and the uncertainty as shown in Equations (A-23) and (A-24). [[

]]

A-18 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

A.5 REFERENCES

A-1 16th International Symposium on Zirconium in the Nuclear Industry, “Shadow Corrosion-Induced Bow of Zircaloy-2 Channels,” ASTM STP 1529, Mahmood, S.T., et al., p. 954, 2011. A-2 GE Nuclear Energy, “Methodology and Uncertainties for Safety Limit MCPR Evaluations,” NEDC-32601-P-A, August 1999.

A-19 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

APPENDIX B GNF RESPONSES TO NRC RAIS ON NEDO-33798 REVISION 0

B-1 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

RAI-1 NSF Alloy Composition Section 2.2 of NEDE-33798P defines the nominal composition and allowable ranges of major alloying elements. Current practice is to define nominal along with a tight range for manufacturing tolerance. While the American Society for Testing and Materials allows a broader range, it also comes with required periodic testing. Further justification is required for the proposed ranges.

Response: The range in composition of NSF defined in the Licensing Topical Report (LTR) is analogous to the composition ranges defined for Zircaloy-2 (Zry-2) and Zircaloy-4 (Zry-4) in ASTM B352/352M-11. As suggested above, the standard practice is to tighten the chemistry ranges when purchasing a material. But chemistry is not the only material specification. Typically, there is a microstructure requirement (fully recrystallized in the case of NSF), mechanical property requirements and corrosion performance requirements. Similar to the requirements in ASTM B352/352M-11 and as defined in 10 CFR 50 Appendix B, for the material to be accepted, the supplier is required to provide test results in quality documents that provide objective evidence that the material lot meets all requirements. a) With respect to corrosion rates, a similar Zr-Nb-Sn-Fe alloy shows dramatic differences between 1.2 percent - 0.6 percent Sn. This is contrary to a portion of your justification. Please describe the influence of Sn on the expected corrosion of NSF channels within the proposed ranges.

Response: For NSF, Global Nuclear Fuel (GNF) considers the range of Sn to be from [[ ]] Sn. It is expected that alloy chemistry will be one source of variability in corrosion performance. Based on the data presented and the references provided in the LTR, the variability in corrosion rates with differences in Sn is acceptable. Besides the corrosion data in Figure 2-12 (updated in response to RAI-3(d)), Nikulina et al. (LTR Reference 15) found the corrosion performance was adequate for E635 with Sn levels between 1.2% and 1.3%. In LTR Reference 23, Comstock et al. found Pressurized Water Reactor (PWR) corrosion improved in ZIRLO when the Sn decreased from 1.2% to 0.97%. The suggestion of a strong negative effect on corrosion as a function of Sn in the [[ ]] range cannot be evaluated without the specific reference information. However, it is also important to note that the impact of corrosion on channel performance is on metal thinning that is accounted for in the mechanical design process (LTR Reference 1). As discussed in Section 2.10, metal thinning accounts for corrosion on both sides of the channel over its entire length. When accounting for the effect of metal thinning in the channel design

B-2 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

process, it is assumed that the channel thickness is decreased over the entire channel length by the design curve provided in Figure 2-12. Thus, it is more appropriate to think of the corrosion thickness value in Equations 2-31 and 2-32 as nominal values rather than maximum values in a distribution. The implication of this is that the average corrosion thickness must remain below the design curve rather than every local area remaining below the design curve. With this understanding and given the significant margin between the measured nominal corrosion value of NSF and the design value (see updated figure in response to RAI-3(d)), it is expected that this margin will cover all variation in nominal corrosion on future NSF channels caused by variation in ingot chemistry or from other variables such as location on the channels.

b) Section 2.2.3 provides no discussion of the potential effect of alloying composition on shadow corrosion. On a similar note, there is no discussion on hydrogen uptake. Please describe the influence of alloying composition on hydrogen uptake and shadow corrosion induced bow.

Response: Shadow corrosion and hydrogen uptake are important because they are the key variables in shadow corrosion-induced bow. Hydrogen uptake is discussed in more detail in the response to RAI-3. An updated version of the shadow bow data for NSF is provided in the response to RAI-9. The GNF plan is to evaluate the variability in shadow bow directly (and thus shadow corrosion and hydrogen pickup indirectly) by measuring channel distortion in the NSF channels that are part of the 8% Lead Use Channel (LUC) program (See Reference 10-1 of this RAI). Specifically, approximately 225 channels will be measured as part of that program and will adequately quantify the variability in shadow bow of NSF (including the effects of variation in alloy chemistry and processing). These measurements are part of the future Post Irradiation Surveillance plan discussed in the response to RAI-10(a).

c) According to Table 2-1, the nominal composition of NSF includes 1.0 percent Sn, and 0.12 percent O. GNF provided strength data for a nominal composition. In the Licensing Topical report (LTR) for Ziron cladding, GNF indicated that the YS and UTS for Ziron cladding was likely [[ ]] Table 2-2 of NEDE-33798P lists an allowable minimum content of [[ ]] Sn and [[ ]] O.

i. Describe the impact of these minimum ranges of Sn and O on TS and UTS.

ii. Describe the impact of these minimum ranges of Sn and O on creep rate and channel bulge calculations.

iii. Describe the impact of these minimum ranges of Sn and O on maximum channel distortion predictions (bow and bulge). iv. Identify any post-irradiated data from NSF lead channels at these lower ranges.

B-3 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Response: i. As stated in the LTR in Section 2.8.1, the design strength of NSF or any zirconium alloy is actually controlled in the material specification that requires objective evidence that the material lots meet the strength requirements. This accounts for any variability in the material strength from variation in chemistry. ii. There is open literature evidence (Nikulina et al., LTR Reference 15) that the in-reactor creep rate of NSF-like alloys is lower than other Zirconium (Zr) alloys. However, the test conditions considered by Nikulina et al. were more similar to fuel rod performance in PWR-type conditions (~330oC and ~120 MPa/~17 ksi) than in Boiling Water Reactor (BWR) channels. While stress concentration sites exist in the channel and can cause the stress levels to be in the [[ ]] psi range, the uniform tensile stress due to the pressure drop across the channel face is calculated to be less than [[ ]] psi for a 10 psi pressure drop. GNF has observed [[

]] This point was made in the LTR in Section 2.9.1, where it was discussed that the stress levels in the channel are in the [[

]]. The conclusion from this analysis is that the variation in Sn and O within the range of NSF will have no significant effect on channel bulge calculations. iii. The effect of variability in Sn and O content on bulge was discussed above for creep deformation. However, there is also an elastic component for bulge. The case is made in Section 2.3 of the LTR that the elastic properties of NSF will not vary significantly with alloy content. Thus, both creep and elastic bulge are not affected by variation in Sn and O content. The impact of variability in Sn and O on channel bow is expected to be [[ ]]. The effect of Sn on irradiation growth that causes fluence bow was discussed in Section 2.11. Variation in Sn has [[ ]]. In alloys such as NSF, the role of Sn may be to [[ ]]. Oxygen is not known to have a significant impact on irradiation growth in zirconium alloys. The impact of Sn and O on shadow corrosion, which causes shadow bow, is discussed above in part (b) of this Request for Additional Information (RAI), and will effectively be evaluated by channel bow measurements.

B-4 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Ultimately, the purpose of the 8% LUC program (MFN 12-074 Supplement 2-A – Reference 1-1) is to explicitly address the potential effects of material and fabrication variability on performance. After these partial reloads are discharged, GNF will perform inspections to measure both channel bulge and bow to better quantify the effects of not only varying chemistry but also variations in vendor processing. iv. The variation in chemistry for the discharged NSF channels is limited because these were the first channels manufactured. Specifically the range of Sn in these channels is from [[ ]]wt%. When considering all the NSF channels manufactured from 2000 to 2014, the Sn has varied from [[ ]]wt%. As discussed above the purpose of the 8% LUC program is to address potential variations in chemistry. Reference 1-1 Letter from A.A. Lingenfelter (GNF) to Document Control Desk (US NRC), Subject: Accepted Version of Enhanced Lead Use Channel (LUC) Program for NSF Fuel Bundle Channels, MFN 12-074 Supplement 2-A, April 15, 2013.

B-5 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

RAI-2 Range of Applicability Section 4 of NEDE-33798P is titled “Applicability.” Besides the discussion of allowable range of alloying elements in Section 2.2.3 of NEDE-33798P, there is no attempt to define a range of applicability of NSF material to GNF channel designs. Are further limitations necessary based on the extent of in-reactor experience and empirical database? For example, is a limit on residence time, neutron fluence (or equivalent fuel burnup), and/or effective control blade exposure (ECBE) necessary?

Response [[

]], it is more appropriate to consider a fluence limit for NSF channels. Based on the available irradiation growth data that shows NSF is resistant to breakaway growth for fluences up to 2.2E22 n/cm2, a reasonable fluence limit for NSF channels is 2.0E22 n/cm2 (calculated as the average for the channel). Using the correlation between fluence and exposure reported in the response to RAI-5, this fluence corresponds to a bundle exposure of ~[[ ]]; thus, the expected average channel fluence at [[ ]]. Therefore, the applicability range of the NSF channel will be based on only residence time and will specifically be limited to [[ ]] years. As of March 2015, four NSF channels have operated for four two-year cycles. GNF recognizes the potential value of extending the current residence time limits or to re-use channels. Any potential re-use of NSF channels or operation of NSF channels beyond the current residence time limits would follow the current GESTAR requirements for lead-test assemblies. In addition, if a channel accumulates more than [[ ]] inch-days of Effective Control Blade Exposure (ECBE) during a cycle, it will be treated as a lead-test assembly if reinserted in a rodded location (i.e., a cell with a control rod). The lead test limitations would apply until sufficient evidence is collected to support the extension of the applicability range of NSF channels, which would be documented and submitted to the NRC for information.

B-6 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

RAI-3 Hydrogen Pickup and Corrosion Figures 2-12 and 2-20 of NEDC-33798P provide measured oxide thickness and absorbed hydrogen for NSF and Zry-2 channels. a) Does GNF have experience with Zry-4 channels? If so, describe the relationship between Zry-4 and NSF channels with respect to oxidation and hydrogen absorption.

Response: Yes, GNF does have experience with Zircaloy-4 channels. Zircaloy-4 was used as the channel material prior to transition to Zircaloy-2 in 1990s. Zircaloy-4 has also been used in some more recent applications as an interim measure to mitigate the shadow-corrosion induced channel bow issue. GNF’s experience with Zircaloy-4 channels is found in paper #8078 in the 2008 Water Reactor Fuel Performance Meeting (WRFPM) in Seoul, Korea. The main advantage of Zircaloy-4 (relative to Zircaloy-2) is the low hydrogen pickup fraction, typically around 10% (see, for example, Miyashita et al. in the 2007 Light Water Reactor Fuel Performance Meeting (LWRFPM) in San Francisco). A comparison of NSF corrosion relative to Zircaloy-4 (and Zircaloy-2) has been provided in GNF’s paper at the 2013 LWRFPM (paper #8465). The tabulated results are reproduced below in Table 3a-1. (Note the average results for NSF and Zircaloy-2 are shown in Figure 2-12 and also in RAI-3(d). Table 3a-1 also includes hydrogen data that are shown in Figure 2-20 of the LTR (for NSF and Zircaloy-2). It should be noted that the Zircaloy-4 data were obtained from a channel that was discharged after 4.3 years with a fuel assembly burnup of 33.4 GWd/MTU, and are being compared to NSF data with ~49 GWd/MTU burnup after 5.83 years. It can be seen from the data that in-reactor corrosion performance of the NSF channel is superior to Zircaloy-4 that has seen less time and neutron exposure. The hydrogen pickup fraction of NSF is [[ ]]. For NSF, the data below correspond to an average pickup fraction of ~[[ ]]%, and, when additional data is included, an average pickup fraction of ~[[ ]]% is obtained (see also RAI-3(d)). These values are [[ ]] or ~10% reported by Miyashita.

B-7 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Table 3a-1

Non-Blade Side Blade Side Channel Material Outer Inner Metal Outer Inner Metal (GWd/MTU) Elevation Surface Surface Hydrogen Surface Surface Hydrogen [inch-days] (in) (m) (m) (ppm) (m) (mm) (ppm) Zry-2 120 10.2 14.6 [[ 9.1 16.7 [[ (49.2) 90 9.5 14.5 13.4 18.1 [4,222] 55 9.5 14.4 9.1 17.2 20 9.9 5.9 4.7 10.7 Average 10 12 9 16

NSF 120 22.8 36.4 19.1 34.4 (49.1) 90 22.6 34.9 24.7 38.4 [4,222] 55 27.6 33.2 30.7 27.7 20 11.7 18.6 19.7 18.3 Average 21 31 24 30

Zry-4 120 33 44 35 46 (33.4) 90 32 53 43 60 [30,896] 55 37 49 42 51 20 32 38 43 35 Average 34 46 ]] 41 48 ]]

b) Figure 2-12 shows a single data point and one standard deviation for NSF corrosion. The NSF data point sits just above the maximum corrosion data point for Zry-2. Beneath the NSF data point is a Zry-2 data point, i.e., at the same exposure time (~5.83 years).

i. Describe the basis (e.g., local maximum, average) of the Zry2 Design Upper Bound curve, which is given by equations 2-32, 2-33 and 2-34 in the LTR.

ii. Describe the relationship (i.e., irradiation conditions, fluence) between the NSF data point and the Zry-2 data points around ~5.83 years.

Response: i. Equations 2-31, 2-32, 2-33 and 2-34 together form the basis for defining the design upper bound line in Figure 2-12. As stated in Section 2.10 of the submitted LTR, the design upper bound line for Zircaloy-2 is used in GNF’s channel mechanical design process to assess the maximum amount of metal thinning due to nominal corrosion in order to preserve mechanical integrity of the channel. Although the equations are written in the form of upper bound corrosion as a function of time, the channel design in practice considers only the maximum residence time ([[ ]]). The intent of the equations is therefore not focused on predicting the corrosion oxide thickness, but to provide the expected maximum amount of metal thinning that needs to be accounted for

B-8 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

due to corrosion up to end of life. In Section 2.10, Equations 2-31 to 2-34 are stated for completeness to show how the maximum value at any given time can be derived. The equations merely reflect the fact that irradiation effect (Equation 2-34) dominates and that there is a minor contribution due to thermal exposure. Given the purpose of the design upper bound line, the reason for Figure 2-12 is not to demonstrate any agreement between the line and measured data points; instead, the purpose is to demonstrate oxide thickness measurements fall below the design line, or more strictly, the value at maximum residence time. ii. The NSF data point in Figure 2-12 represents the average, and associated standard deviation, of all oxide measurements taken from one NSF channel that had operated for three two-year cycles in a US plant reaching ~49 GWd/MTU fuel assembly burnup in 5.83 years. During its operation, the channel had an ECBE of 4,222 inch-days, which is relatively low, such that significant corrosion enhancement due to the shadow corrosion mechanism is not expected. Coupons from four axial elevations (20, 55, 90 and 120 inches) of the NSF channel were measured as part of the hot cell examination. The Zircaloy-2 data point in Figure 2-12 represents the corresponding measurements made on a Zircaloy-2 channel that operated in symmetric core locations to the NSF channel during all three cycles of operations. The residence time, fuel assembly burnup and ECBE therefore match those for the NSF channel. For hot cell examination, coupons from the same four axial elevations (20, 55, 90 and 120 inches) were measured. For each examined axial elevation, the fluence level for Zircaloy-2 therefore matches that for NSF. See also RAI-3(a) for a detailed breakdown of measured oxide thickness.

c) Figure 2-20 shows hydrogen absorption as a function of hydrogen generated for a NSF channel and a Zry-2 channel operated in symmetric core locations.

i. Describe the higher hydrogen generated in the NSF channel.

ii. Describe how hydrogen absorption is affected by duty, fluence, and water chemistry (e.g., HWC, NMCA, OLNM).

iii. Identify any differences between operations of these two channels (e.g., ECBE).

Response: i. In Figure 2-20 (and in an updated version in Figure 3d-2 below), the “Hydrogen Generated” parameter is the theoretical potential value of the hydrogen generated from the corrosion reaction with water on both inside and outside channel surfaces expressed as a hydrogen content in the remaining metal, assuming it is uniformly distributed in the channel wall. In other words, the parameter represents the potential hydrogen content if the material were to have a 100% hydrogen pickup value. The parameter is related to the combined (inside and outside) oxide thickness and the metal thickness. Figure 2-20 in the LTR shows data for a pair of NSF and Zircaloy-2 channels that have

B-9 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

the same mechanical (dimensional) design and operated for 5.83 years (three two-year cycles) in symmetric core locations and hence have the same operational history. The higher values of “Hydrogen Generated” for NSF merely reflect the higher corrosion rate of NSF compared with Zircaloy-2, as shown by the pair of points at 5.83 years in Figure 2-12. Therefore, this is best characterized as higher hydrogen generated by the NSF rather than absorbed in the NSF. ii. For non-heat transfer zirconium-alloy components, time and fast neutron fluence are usually considered to be the two primary factors that could affect hydrogen absorption in a given alloy. Duty, in the context of how a channeled assembly is operated, therefore is a potential factor in affecting time to achieve a particular level of fluence. This is true mostly for Zircaloy-2 for which there can be considerable variation in hydrogen pickup fraction at moderate to high fluence levels or residence times. In contrast, Zircaloy-4 typically shows low hydrogen pickup fraction regardless of fluence or time; however, the corrosion is considerably greater than Zircaloy-2. For NSF, [[ ]]. Figure 2-20 in the submitted LTR provides a direct comparison between NSF and Zircaloy-2. The figure shows that NSF had greater amounts of corrosion, as reflected by higher “Hydrogen Generated” values; however, [[ ]]. As discussed below in RAI-3(d), there has been another set of hot cell data. The differentiating feature for the second set of hot cell data is the very high control blade exposure experienced by the channel (ECBE 51,262 inch-days). As shown in the measured versus expected hydrogen plot in RAI-3(d), data for NSF from both hot cell campaigns form one population, sharing a common slope reflecting [[ ]]. As control blade exposure is the main duty parameter of interest for channels, the updated results presented in RAI-3(d) show that hydrogen pickup of NSF is not sensitive to variations in this key duty parameter. Under BWR conditions, available information on Zircaloy-2 does not show appreciable variations that can be assigned to variations in water chemistry (e.g., Hydrogen Water Chemistry (HWC), Noble Metals Chemical Addition (NMCA) or On-Line NobleChemTM (OLNC)). Specific to NSF, all channels that were deployed in the United States (US) under LUC or expanded LUC programs have operated under HWC with zinc injection and with either NMCA or OLNC water chemistry conditions. Difference in hydrogen pickup behavior of NSF due to NMCA versus OLNC is not expected and is supported by hot cell examination results. As discussed in RAI-3(d), there have been two hot cell campaigns. The first is for a channel that operated under NMCA for all three cycles. The second is for a channel that operated under NMCA in the first cycle and then operated under OLNC in the second and third cycles. The hot

B-10 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

cell results after three cycles of operation showed [[ ]]. iii. There is no appreciable difference between the two channels, both operated for 5.83 years (three two-year cycles) in symmetric core locations reaching ~49 GWd/MTU with an ECBE of 4,222 inch-days. d) Please provide any data for NSF oxidation and hydrogen uptake taken since this topical report was submitted.

Response: Since the submittal of the LTR, there has been another set of hot cell examination providing corrosion and hydrogen uptake data for NSF channels. The first set of data (shown in the submitted LTR) is from the 65 mil portion of a NSF channel that operated for three two-year cycles (5.83 years) reaching ~49 GWd/MTU burnup with a relatively low ECBE value of 4,222 inch-days. The newer, second set is from the 75 mil portion of a NSF channel that operated also for three two-year cycles (5.53 years) reaching ~40 GWd/MTU burnup with a very high ECBE value of 51,262 inch-days. Updated Figures 2-12 and 2-20 are presented below. The data in Figure 3d-1 are tabulated in Table 3d-1 with the channel exposure and ECBE value of each point. Note that data from the companion Zircaloy-2 channel that operated in the symmetric core location to the NSF channel was collected only during the first hot cell campaign (5.83 years of operation). For the hydrogen pickup plot (Figure 3d-2), the theoretical hydrogen generated values takes into consideration the oxide thickness on both inside and outside surfaces as well as the channel thickness – see also RAI-3(c)(i). Compared with Figure 2-20 in the submitted LTR, it can be seen from the revised plot that inclusion of additional data for NSF yielded essentially the same average pickup fraction (slope) for NSF.

B-11 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] Figure 3d-1 Upper Bound Design Value for Corrosion and Recent Hot Cell Oxide Thickness Data for Zircaloy-2 and NSF (Uncertainty bars represent one standard deviation) [[

]]

[[ ]] Figure 3d-2 Measured Versus Expected Hydrogen Pickup (There are two sets data for NSF and one set for Zircaloy-2)

B-12 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Table 3d-1 Operational Conditions of Measured Channel Oxide Thickness in Figure 3d-1

Assembly Average Time ECBE StDev Channel Exposure Oxide Material (GWd/MTU) year inch-days microns 39.7 4.07 8,133 [[ 36.5 4.07 10,721 39.6 4.07 21,916 40.7 4.07 16,216 48.2 5.16 14,025 47.9 5.16 0 Zircaloy-2 48.1 5.70 27,112 48.1 5.70 27,112 19.7 1.86 34,105 43.9 6.58 17,795 42.9 6.58 712 49.3 5.83 4,222

49.2 5.83 4,222 NSF 40.3 5.53 51,262 ]] e) In MFN 12-074, GNF states, “The measured oxide thickness of NSF after [[ ]] For comparison, the measured oxide thickness of Zircaloy-2 channels after [[ ]].” The statement appears to refer to the data expressed in the following figure and table:

B-13 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]]

Figure. MFN 12-134, p. 77 (78 of 82 in pdf) (Table I., Paul E. Cantonwine, Yang-Pi Lin, Dan R. Lutz, David W. White, and Kevin L. Ledford, “BWR Corrosion Experience on NSF Channels,” Paper 8465, Topfuel 2013, Charlotte, September 15-19, 2013.) i. The oxidation rate for NSF seems to be a factor of 2 or 3 times that of Zry-2. Are these results obtained under the same irradiation conditions and duty level? What is the expected corrosion thickness after 8 years of operation? ii. GNF measurements indicate that the inside oxide thickness is greater than the outside thickness. In Equation 2-31 (LTR), what corrosion thickness is used – inside or outside?

Response: i. In Figure 2-12, the data for NSF and Zircaloy-2 at 5.83 years are obtained from channels on assemblies that operated for three two-year cycles in symmetric core locations and hence have equivalent irradiation condition and duty level – also see the response to RAI-3(b). As discussed in RAI-3(b), the Design Upper Bound line in Figure 2-12 is used in GNF’s channel mechanical design process to assess the maximum amount of metal thinning due to corrosion. [[ ]]. As discussed in the response to RAI-1, this upper bound value should be compared to the nominal measured value. Given the margin between the upper bound value and the nominal measured value of NSF corrosion, the Design Upper Bound line is adequate for metal thinning calculations. Based on extrapolation of the corrosion data in the LTR and the response to RAI-3(d), the nominal corrosion thickness of NSF at [[ ]] years would be around [[ ]] microns. Four NSF channels (two each in two plants in the US) have

B-14 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

operated for four two-year cycles. All four channels completed ~eight years of operations without known issues. ii. Equation 2-31 does not make a differentiation between the inside or outside surface. It assumed both sides to be the same, as shown by the inclusion of the factor of two in the equation. As stated in (i) above, the intent of the equation is not so much to predict the thickness of the corrosion oxide; rather, the intent is to provide a guidance to the expected maximum amount of metal thinning due to corrosion. To satisfy this purpose, the expected behavior for NSF is to show corrosion less than assumed in Equation 2-31 and associated Equations 2-32 through 2-34 – also see the response to RAI-3(b).

f) The LTR does not address high temperature corrosion for NSF channels under accident conditions.

i. Describe the expected peak channel temperature history during the limiting Boiling Water Reactor (BWR) Loss of Coolant Accident (LOCA).

ii. Describe the predicted corrosion and channel performance under these conditions. iii. Provide weight gain vs time for NSF and Zry-2 material (see Figure B-15 of NEDC- 33353P, Revision 0) for the above conditions, or 1000°C.

Response: i) During a LOCA scenario, the channel acts as a heat sink in assisting removal of decay heat from the fuel assembly. The channel temperature will therefore increase as more heat is absorbed. The increase in channel temperature will lag behind the cladding temperature of the adjacent fuel rods located at the periphery of the fuel assembly. Ultimately, the temperature of the channel would approach that of the adjacent rods facing the channel. The temperature reached by the channel will be, however, lower than the limiting fuel rod peak clad temperature, which comes from fuel rods in the interior of the fuel assembly having less effective heat removal compared with peripheral rods. ii) NSF contains >~97 % (by weight) of zirconium. The corrosion or reaction of steam will therefore be dominated by the zirconium-water reaction. Although alloying elements can modify the steam oxidation process somewhat, a major departure from typical zirconium- alloy behavior is not anticipated. From a review of the literature (NUREG/CR-6967), it is noted that steam oxidation kinetics for Zirlo, which has a similar composition as NSF (except for Fe), essentially follow the Cathcart-Pawel (CP) relationship at 1,000, 1,100 and 1,200oC. The oxidation behavior of NSF at 1,000oC as discussed in (iii) is [[

]]. For zirconium alloys at elevated temperatures, the main embrittlement mechanisms are metal thinning due to oxidation, oxygen embrittlement and hydrogen embrittlement. The

B-15 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

impact of metal thinning due to oxidation, including the effect of dissolved oxygen ahead of the oxide layer, will depend on the metal thickness. For fuel cladding, the degree of zirconium-alloy embrittlement is typically expressed in terms of Equivalent Cladding Reacted (ECR) that takes into account oxide thickness, cladding thickness and thickness of the oxygen embrittled layer. When applied to NSF channels, for a given amount of oxidation and hence oxygen embrittled layer (as predicted by the CP relationship), the overall equivalent channel reacted would be lower on account of the greater channel thickness compared with the cladding case. In addition, as discussed in (i), the peak channel temperature would be lower than the Peak Cladding Temperature (PCT). The ECR for the channel next to the fuel assembly with the limiting PCT (and hence ECR) would be lowered further compared with the limiting ECR on account of the peak temperature difference. With regard to absorbed hydrogen, the thicker channel compared with cladding means that channels will be more tolerant of hydrogen. The effect of absorbed hydrogen during the transient is typically not considered, except near 1,000oC when some zirconium alloys can develop early onset of breakaway oxidation with associated increased hydrogen absorption. iii) High temperature steam oxidation tests were obtained for Zircaloy-2 and NSF channel strip in the bare (initially unoxidized) condition and for NSF channel [[

]]. The tests were performed at 1,000oC for various times ranging from 500 seconds to 7,000 seconds. Results are compared in Figure 3f-1 to the CP relationship for times less than breakaway. [[

]].

B-16 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] Figure 3f-1 Measured Weight Gain of NSF Compared to Zircaloy-2 During a High Temperature Oxidation Test for Simulating LOCA Conditions

B-17 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

RAI-4 Channel bow, creep and oxidation a) MFN 12-134 provided significant background for the NSF channel performance. The shadow corrosion-induced bow data for NSF channels is limited relative to the Zry-2 data presented on page 75. Please provide this figure with only the NSF LUC data compared to the Zry-2 control channels that operated under similar conditions to the NSF channels.

Response: The shadow corrosion-induced bow data for NSF is compared in the figure below to counterpart Zircaloy-2 control channels that operated during the same plant and cycle and with the same control history as the NSF channels. In this small population, it is apparent [[ ]]. [[

]] Figure 4a-1 Inferred Shadow Bow of NSF LUCs Compared to the Zircaloy-2 Control Channels

B-18 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public) b) Please provide measured creep data for NSF and Zry-2 at same temperature and differential pressure.

Response: The measured creep bulge of NSF and Zircaloy-2 for the same channel design in the same plant type (i.e., similar differential pressures) is provided in Figure 2-11. The data indicate [[ ]].

c) Describe the maximum ECBE and fluence for a GNF BWR channel under normal controlled operation throughout its lifetime (i.e., not suppressed).

Response: The range of ECBE and exposure (an alternative to fluence) for the NSF channels in the GNF current database are provided in Figure A-5. This also is a reasonable representation of the range of ECBE and exposures in the Zircaloy-2 database. Maximum ECBEs can be as high as ~50,000 inch-days. More typically, the maximum ECBE value in a plant is between 20,000 and 35,000 inch-days. Discharge exposures typically range from the low 40s to the low 50s in GWd/MTU.

B-19 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

RAI-5 Channel Growth Figure 2-13 of NEDE-33798P provides NSF and Zry-2 channel growth data, as well as NSF and Zry-2 irradiation growth data from BOR-60.

a) Correlate fluence to assembly average exposure for the NSF channel data.

Response: The channel fluence plotted in Figure 2-13 is a weighted average of the axial fluence distribution. For the NSF channels, the average channel BWR fluence can be estimated as the product of [[ ]]*(exposure in GWd/MTU); this conversion factor is the average of the ratio of the NSF channel fluence to exposure for the data plotted in Figure 2-13. Thus, at 50 GWd/MTU the average BWR channel fluence is estimated as [[ ]] n/cm2 and at 55 GWd/MTU (the estimated bundle average limit) the average BWR channel fluence is [[ ]] n/cm2 or ~[[ ]] n/cm2. However, because there is an axial distribution, the region around the center of the channel will have a higher fluence than the average.

b) Describe the irradiation conditions, particularly ECBE, for the NSF and Zry-2 channel growth data.

Response: The growth data for Zircaloy-2 channels plotted in Figure 2-13 had ECBE values less than 4,500 inch-days; thus, the growth in these channels is attributed to only fluence- induced growth. For the NSF channels the data at ~4.5E21 n/cm2 and ~9.0E21 n/cm2 had ECBE values less than 4,500 inch-days; like the Zircaloy-2 channels the growth in these NSF channels is attributed to only fluence-induced growth. The NSF data at ~6.8E21 n/cm2 had ECBE values of ~51,000 inch-days indicating a high susceptibility to shadow bow. Interestingly this data shows no bias in growth suggesting the majority of the length change is due to fluence-induced growth. This is also consistent with the observation of low shadow bow in these channels as shown in Figure A-9. c) The maximum fluence reported for NSF channel data is about [[ ]] Please provide any data for NSF channels at higher fluences.

Response: Two of the four channels operated to about 9.0E21 n/cm2 were reinserted for a 4th cycle. These channels have been discharged but not yet inspected for channel distortion. One of the purposes of the 8% LUC program will be to provide more data in the higher fluence (i.e., exposure) region.

B-20 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

d) What is the expected maximum fluence for NSF channels in 6 and 8 years of operation?

Response: As discussed in the response to Section (a) of this RAI, the maximum fluence in a channel will be in the region around the mid-elevation point. Because the growth of channels includes the entire distribution of fluences in the channel, the channel fluence is an average of this axial distribution. The bundle average exposure that corresponds to the [[ ]]. Using the correlation in the response to Section (a) of the RAI, the corresponding average channel BWR fluence would be ~[[ ]] n/cm2. Again the maximum value of the fluence in this case will be higher than the average and probably ~[[ ]] n/cm2.

e) Describe the difference in growth behavior between the NSF BOR-60 data and the NSF channel data.

Response: The growth data from BOR-60 is a standard irradiation growth test where small coupons were irradiated, and the length was measured with an accurate Linear Variable Differential Transformer (LVDT)-type device. Because the BOR-60 is a fast reactor that has a different neutron energy spectrum than a BWR, the BOR-60 fluence (> 1 MeV) must be corrected to a BWR fluence (> 1 MeV). In addition, because of the size of the sample there is no fluence distribution in the sample. The channel growth data is measured in the spent fuel pool with a calibrated tape measure and a video camera. In addition, there is a significant fluence distribution that is accounted for by using a weighted average calculation (as discussed above). With these differences in sample and irradiation conditions, it is not surprising that there are differences in the absolute value of the growth. Rather the conclusion from the BOR-60 irradiation growth data is that NSF is resistant to breakaway growth. Also, while the NSF channel data is at a fluence that is near the predicted breakaway growth for Zircaloy, it is significantly lower than the comparable Zircaloy-2 channel growth data.

RAI-6 Calculating CPR with NSF Channels a) Section A.2.2 of NEDE-33798P states, “Fast fluence (E>1MeV) gradient induced bow results from differential growth of channel material on opposite channel faces.” In Section 3.1.5, GNF states, “NSF channels do not bow significantly as a function of exposure, . . .”

i. Given that the data in Figure 2-13 show much the same growth behavior for Zry-2 and NSF channels, would not the differential growth be similar, and therefore the fluence- induced bow?

B-21 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

ii. Describe the magnitude of bow that would be considered “significant.”

Response: i. The data in Figure 2-13 show a significant difference in growth behavior between Zircaloy-2 and NSF. The Zircaloy-2 data clearly shows it is susceptible to breakaway growth while NSF is resistant to breakaway growth. In addition, the channel length data of NSF is significantly lower than the average Zircaloy-2 channel data at the same average fluence. Because fluence bow is a function of the growth rate and Zircaloy-2 exhibits a larger growth rate than NSF at fluences near breakaway, greater fluence bow is expected in Zircaloy-2 channels compared to NSF. ii. “Significant” is a qualitative term that is not meant to invoke a quantitative number. Based on experience evaluating channel distortion in cells with channel-control blade interference, “significant” bow might be considered something greater than about 100 mils (as measured at the mid-elevation). b) Figure 3-1 of NEDE-33798P provides calculated values of CACABO for 12 cycles of NSF cores.

i. For the data presented in Figure 3-1, please provide the channel type (e.g., 120/75, 100/60), the irradiation conditions for the calculated cases, including cell lattice (D, C, S), core power density, cycle length (EFPD) and ECBE at the beginning and end of the cycle. As part of the response, please indicate if the cycle is first, second, or third cycle.

ii. Please provide the CACABO values for a Zry-2 channel as a function of cycle exposure for Cycle F in Figure 3-1.

iii. Please provide the CACABO values for a Zry-2 channel as a function of cycle exposure for the case in Figure 3-1 with maximum ECBE.

Response i. Table 6b-1provides the channel type, cell lattice, core power density, and cycle length for each of the cycles shown in Figure 3-1 of the LTR. Table 6b-2 shows Beginning-of- Cycle (BOC) and End-of-Cycle (EOC) ECBE values for each batch of fuel in each cycle from Figure 3-1. The LTR simulations included full cores of NSF channels (e.g., NSF channels on fresh, once-burnt, and twice-burnt fuel). Thus, the ECBE inch-days information is provided at BOC and EOC for each fuel batch.

B-22 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Table 6b-1 Selected Plant Data for Plants in Figure 3-1 of the LTR

Lattice Core Power Cycle Length Plant Cycle Channel Type Type Density [kW/L] [EFPD] Cycle A [[ D [[ [[ Cycle B D Cycle C D Cycle D D Cycle E D Cycle F D Cycle G C Cycle H C Cycle I S Cycle J S Cycle K D Cycle L ]] D ]] ]]

B-23 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Table 6b-2 ECBE at BOC and EOC for Each Fuel Batch

Fresh Fuel Batch 2nd Cycle Batch 3rd Cycle Batch 4th Cycle Batch 5th Cycle Batch 6th Cycle Batch 7th Cycle Batch

Plant BOC EOC EOC EOC EOC BOC EOC BOC EOC BOC EOC Cycle ECBE ECBE BOC ECBE ECBE BOC ECBE ECBE BOC ECBE ECBE ECBE ECBE ECBE ECBE ECBE ECBE

Cycle A [[

Cycle B

Cycle C

Cycle D

Cycle E

Cycle F

Cycle G

Cycle H

Cycle I

Cycle J

Cycle K

Cycle L ]]

B-24 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

ii. Figure 6b-1 shows the CACABO values as a function of exposure for a Zircaloy-2 core for Cycle F from Figure 3-1 in the LTR. [[

]] Figure 6b-1 CACABO Values for a Zircaloy-2 Core as a Function of Exposure for Cycle F from Figure 3-1 in the LTR

B-25 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

iii. Per Table 6b-2, Cycle C in Figure 3-1 from the LTR has the maximum ECBE value (fourth cycle batch). Figure 6b-2 shows the CACABO values as a function of exposure for a Zircaloy-2 core for Cycle C from Figure 3-1 in the LTR. [[

]] Figure 6b-2 CACABO Values for a Zircaloy-2 Core as a Function of Exposure for Cycle C from Figure 3-1 in the LTR c) Figure A-6 of NEDE-33798P provides predicted versus measured fluence bow for NSF channels with ECBE < 4500 inch-days.

i. Please provide P vs M fluence bow data for NSF channels with ECBE > 4500 inch-days and M-P as a function of burnup (ref: Figures A-7 and A-8).

ii. Please provide the same data for corresponding Zry-2 channels, i.e., Zry-2 channels with the same or similar irradiation conditions to the same burnup and ECBE as the NSF channels.

iii. In equation A-17 of NEDE-33798P, a weighting factor ‘f’ is used in conjunction with the controlled time (i.e., time for which the control blade is inserted adjacent to the channel).

B-26 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Please provide an example of how the weighting factor is used for cases where a channel is 1) controlled for two cycles, 2) controlled in the first cycle and uncontrolled in the second cycle, and 3) uncontrolled in the first cycle and controlled in the second cycle.

Response: i. An updated version of Figure A-8 is provided in the response to RAI-6(d). ii. The Measured-Predicted (M-P) data for Zircaloy-2 as a function of exposure for ECBE < 4,500 inch-days is provided on Slide 72 of MFN 12-134 (Reference 6-1). This data indicates the uncertainty around the prediction of fluence bow increases dramatically as a function of exposure for Zircaloy-2. The M-P Zircaloy-2 data for ECBE > 4,500 inch- days is plotted as inferred shadow bow vs. ECBE in RAI-9(c); inferred shadow bow is measured bow minus predicted fluence bow. iii. The weight factor f is an effective control exposure weight, shown as a histogram in Figure 6c-1. It is dependent on the residence time of the channel (note: not the bundle, since in the case of re-channeling the two numbers are different). Its value is assigned based on the following histogram: [[

]] Figure 6c-1 Weighting Factors for Calculating Effective Control Blade Exposure (ECBE) During operation, a sequence of control rods is used to control the core. Sequences are exchanged periodically in time to create a more uniform exposure distribution. During the first year, the weighting factor for the product between insertion distance and time (ECBE in units of inch-days) is [[ ]]. During the second year, the weighting factor is [[ ]]. During the third year, the factor decreases to [[ ]]. Beyond three years, ECBE does not accumulate. This calculation of ECBE weighs early-life control during the first two years much more significantly than control after two years.

B-27 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

d) Figure A-8 of NEDE-33798P provides M-P fluence versus exposure. On this figure, the data points for NSF channels with [[

]] What is the predicted and M-P fluence bow for NSF channels with ECBE > 4500 inch-days and exposure up to maximum approved burnup levels of GNF fuel assemblies?

Response: Figure A-8 included all of the data collected as part of the LUC programs prior to January 2013. Since that date, seven NSF channels have been discharged and measured for distortion. These additional verified data have been added to Figure A-8, which is shown below in Figure 6d-1. Data on higher burnup channels are planned to be measured and reported as part of the normal fuel performance update during the annual technology update meeting with the NRC. [[

]] Figure 6d-1 Redrawn Version of Figure A-8 in the NSF Channel LTR Reference 6‐1 Letter, A.A. Lingenfelter (GNF) to Document Control Desk (US NRC), “NSF Topical Report Pre-Submittal Presentation Slides,” MFN 12-134, December 20, 2012.

B-28 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

RAI-7 Sample Calculation on Channel Bow a) Please provide sample plots for bow (including FLUBOW, SHADBOW, CHANBOW, BOCELL, CACABO) of Zry-2 and NSF channels as a function of burnup for channels that exhibit different amounts of ECBE and different cycles of operation. Response: Background Core-Average, Cell-Average Bow (CACABO) is a measure of the core-average channel bow and includes the contributions of  Initial manufactured channel bow  Fluence-induced channel bow  Shadow corrosion-induced channel bow Some nominal amount of channel bow is inherent to the manufacturing process; therefore, channels with no irradiation history (beginning of life channels) are still characterized by some amount of channel bow. Fluence-induced channel bow is caused by a fast fluence (E > 1MeV) gradient that results in differential growth of the channel material on opposite faces. Shadow corrosion-induced channel bow is an enhanced irradiation corrosion mechanism that occurs on zirconium alloys when a dissimilar material (such as a control blade) is near the zirconium surface (such as a BWR channel) and the water chemistry is oxygenated. Calculation of CACABO and CACABO Uncertainty [[

]]

B-29 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

To illustrate the different behavior of channel bow between Zircaloy-2 and NSF channels, nominal values of FLUBOW, SHADBOW, CHANBOW, and BOCELL are presented as part of this response. Case Study of CACABO A representative cycle design for a BWR/6 plant was chosen to provide an illustration of CACABO and its constituent contributors (including FLUBOW, SHADBOW, CHANBOW, BOCELL) to compare the behavior of Zircaloy-2 and NSF channels with different amounts of ECBE and different cycles of operation as a function of burnup. The CACABO results of Zircaloy-2 and NSF channels are shown in Figure 7a-1. Note, the bundles that contribute to the calculation of CACABO [[

]]. The representative cycle design is one with a two-year cycle length. Both Zircaloy-2 and NSF results for CACABO are equivalent at the BOC, where CACABO is primarily influenced by [[

]]. The CACABO result for Zircaloy-2 channels begins at approximately [[ ]] mils and progressively increases to approximately [[ ]] mils at EOC. The CACABO result for NSF channels begins at approximately [[ ]] mils and slightly increases to approximately [[ ]] mils at EOC. The difference between the CACABO results of Zircaloy-2 and NSF channels as the cycle exposure increases is driven by fluence and shadow corrosion-induced channel bow, which will become more evident in the examination of FLUBOW and SHADBOW, respectively.

B-30 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] Figure 7a-1 Core-Average, Cell-Average Bow as a Function of Burnup (Cycle Exposure) Five cells from the group of cells with at least one bundle that is defined as potentially limiting in bundle power were chosen to display BOCELL results of Zircaloy-2 and NSF channels. The characteristics of these five cells (ECBE, channel exposure, and channel residence time) are provided in Figure 7a-2 for both BOC and EOC. The BOCELL results of Zircaloy-2 and NSF channels are shown in Figure 7a-3 and Figure 7a-4. Figure 7a-3 shows results for cells with equivalent cell-average channel exposure and residence time but different amounts of ECBE. Figure 7a-4 shows results for cells with zero ECBE but different channel exposure and residence time. In general, the BOCELL results of the NSF channels are relatively constant as a function of cycle exposure. The BOCELL results of the Zircaloy-2 channels in Cells A and B exhibit the effects of high ECBE; whereas, the low accumulation of ECBE in Cell C is not significant. The BOCELL results of the Zircaloy-2 channels in Cells D and E exhibit break-away fluence- induced channel bow. Note, the changes in slope of the BOCELL results of the Zircaloy-2 channels (Cells A and D) are from the individual channels within the given cell either bowing (on average) toward or away from the control blade.

B-31 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] Figure 7a-2 Cell Characteristics for Select BOCELL Results [[

]] Figure 7a-3 Cell-Average Bow for Selected Cells A, B, and C as a Function of Burnup (Cycle Exposure)

B-32 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] Figure 7a-4 Cell-Average Bow for Selected Cells D and E as a Function of Burnup (Cycle Exposure) Four channels from the group of bundles that contribute to CACABO were chosen to display CHANBOW, FLUBOW, and SHADBOW results of Zircaloy-2 and NSF channels. Note, the four channels do not belong to the five cells described above but were chosen based on different amounts of ECBE and different cycles of operation. The characteristics of these four channels (ECBE, channel exposure, and channel residence time) are provided in Figure 7a-5 for both BOC and EOC. The CHANBOW, FLUBOW, and SHADBOW results of Zircaloy-2 and NSF channels are shown in Figure 7a-6 and Figure 7a-7. Figure 7a-6 shows results for channels in their first cycle of residence and different amounts of ECBE. Figure 7a-7 shows results for channels in their second cycle of residence and different amounts of ECBE. Again, in general, the bow results of the NSF channels are relatively constant as a function of channel exposure. For channels in their first cycle of residence that have yet to reach the threshold for break-away fluence-induced channel bow, the FLUBOW results for both Zircaloy-2 and NSF channels are the same as shown in Figure 7a-6. Note, in Figure 7a-7, the total channel bow for Channel Y increases initially during the cycle from the accumulation of shadow corrosion-induced bow, but then decreases from fluence-induced bow after the shadow corrosion-induced bow becomes saturated.

B-33 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] Figure 7a-5 Channel Characteristics for Select CHANBOW, FLUBOW, and SHADBOW Results In summary, NSF channels show much less sensitivity to fluence and shadow corrosion-induced channel bow than Zircaloy-2 channels, which explains the difference between the CACABO results of Zircaloy-2 and NSF channels.

B-34 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] Figure 7a-6 Total, Fluence-Induced, and Shadow Corrosion-Induced Channel Bow for Selected Channels W and X as a Function of Burnup (Channel Exposure)

B-35 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] Figure 7a-7 Total, Fluence-Induced, and Shadow Corrosion-Induced Channel Bow for Selected Channels Y and Z as a Function of Burnup (Channel Exposure)

B-36 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public) b) In MFN 12-134, on page 72, GNF compares the measured-predicted (M-P) for fluence bow for NSF and Zr-2 channels. The NSF channels seem to have bounds on M-P of fluence bow of [[ ]] In another figure, Figure A-8 from the LTR, the M-P range [[ ]]

Pages 43-45 of MFN 12-134 mentions the uncertainty of channel bow is accounted for in the bundle R-factor calculation, and GNF states, [[ ]] On page 45 of MFN 12-134, GNF states (in NEDO-32601P), [[ ]] With regard to FLN 2004-030, GNF indicates that [[ ]]. These uncertainties apparently apply to Zr-2 channels.

Please provide a statement or explanation of how the R-factor uncertainty will be developed for and applied to NSF channels.

Response The rod peaking and corresponding R-factor uncertainties in NEDC-32601P-A (Reference 7-1) were defined based on Zircaloy-2 channel bow characteristics. These uncertainties are conservative with respect to uncertainties that would result from similar studies that incorporated NSF channel bow considerations. For this reason, the continued application of these values is acceptable for assemblies with NSF channels. To ensure further conservatism, the increase in R-factor uncertainty to [[ ]]% based on Zircaloy-2 shadow bow considerations described in FLN 2004-030 (Reference 7-2) will also continue to be applied for assemblies with NSF channels. As demonstrated in Section 3.1.5 of the LTR, calculations of BOWAVE uncertainties for NSF indicate that this increase may not be necessary. Any future reduction in R-factor uncertainty back to [[ ]]% to account for an NSF based BOWAVE uncertainty less than [[ ]] mils will be supported by measured data and justified in a transmittal to the NRC. References 7-1 GE Nuclear Energy, “Methodology and Uncertainties for Safety Limit MCPR Evaluations,” NEDC-32601P-A, August 1999. 7-2 Letter, John F. Schardt (GNF-A) to U.S. Nuclear Regulatory Commission Document Control Desk with attention to Mel B. Fields (NRC), “Shadow Corrosion Effects on SLMCPR Channel Bow Uncertainty,” FLN-2004-030, November 10, 2004.

B-37 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

RAI-8 Measured and Predicted Bow With respect to the Figure (M vs P, Limerick Lead Use Channel Inspections after 3rd cycle) on page 74 of MFN 12-134 (page 75 of 82 in pdf), please provide M vs P of bow for NSF LUC channels from Hatch 2 (Cycle 20-22) and Perry and Clinton.

Response: The figure on page 74 is measured bow vs. predicted fluence bow for Zircaloy-2. This figure specifically highlights that NSF channels that experienced a fluence gradient that caused Zircaloy-2 channels to bow ~[[ ]] mil had only ~[[ ]] mil of fluence bow. The measured vs. predicted NSF channel data are plotted in Figure A-6 for only the channels with ECBE values less than 4,500 inch-days. These channels are only susceptible to fluence gradient- induced bow. In Figure A-6, data from Limerick are provided. All the channels from Clinton and Perry (except from the most recent inspection in March 2015 that is yet to be verified) and most of the channels from Hatch Unit 2 had ECBE values greater than 4,500 inch-days. Figure A-6 shows that the model for fluence bow of NSF channels is reasonably predicting the nominal measured fluence bow.

B-38 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

RAI-9 Expanded Lead Use Channel Program GNF proposed an expanded LUC program for NSF channels – MFN 12-074. The NRC approved the expanded program (March 29, 2013, ML13106A068). a) Please provide an update on the status of current NSF LUA and expanded NSF LUC programs.

Response: As background first consider the population of observations of interference in Figure 9a-1. The plant population is divided into plant types: S-Lattice, C-Lattice and D-Lattice. The important plant feature that effects channel control-blade interference is the gap distance between channel and control blade. The S-Lattice plants have the smallest gaps while the D-Lattice plants have the largest. Thus, the observations of interference followed the relative susceptibility to interference and first occurred in an S-Lattice plant, then C-Lattice, then D-Lattice. This same information is plotted per plant in Figure 9a-2. The plants listed in Figure 9a-2 provide the population of plants that have experienced channel control-blade interference with Zircaloy-2 and Zircaloy-4 channels. As part of the current NSF LUC program and the expanded 8% LUC program, NSF channels are being inserted in a significant subset of the plants that have experienced interference (See Figure 9a-3). It is important to note that 8% quantities of NSF channels will be operating in all four S-Lattice plants in the US by the fall of 2015. Since the S-Lattice plants are the most susceptible to channel control-blade interference, any unexpected performance issues with NSF channels will be observed in this leading population and would enable the development of a mitigating response to full reloads inserted in 2016 and after.

Mixed Vendor C-Lattice Plants

35 Mixed Vendor S-Lattice Plants

30 D-Lattice Plants

C-Lattice Plants 25 S-Lattice Plants 20

0 Numberof Interference Observations

Year of End-of-Cycle

Figure 9a-1 Observations of a No-Settle Condition as a Function of the Year of End-of- Cycle

B-39 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

[[

]] Figure 9a-2 Observations of Interference in Specific Plants (This provides the population of plants that have experienced channel control-blade interferences.)

[[

]] Figure 9a-3 The Plant Population Where NSF Channels Have Been or Will Be Inserted by the Fall of 2015

B-40 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Below is a table (Table 9a-1) listing the plants where NSF channels have been inserted as part of normal LUC programs or as part of the expanded 8% LUC program. In 2014, GNF inserted [[ ]] Pre-Ox NSF channels, and in 2015, another [[ ]] will be inserted. Six US plants have inserted or will insert NSF channels as part of the expanded 8% LUC program. Table 9a-1 History of NSF LUC and 8% Expanded LUC Programs [[

]] The following two tables summarize the conditions of the NSF channels that have been discharged and inspected (Table 9a-2) and the expected conditions of the NSF channels that will be discharged and inspected as part of the normal and 8% LUC programs (Table 9a-3). There will be a distribution in operating conditions with the majority of channels operating for ~six years (three two-year cycles); thus the majority of data will be collected in these conditions. In addition, the largest number of channels inspected after completion of the LUC programs will have operated in S-Lattice plants, which are most susceptible to channel interference. Operation into the 4th cycle is not universal. Where possible, NSF channels will be reinserted to operate out to near the [[ ]] residence time limit. Currently four NSF channels have operated for four two-year cycles.

B-41 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

Table 9a-2 Current NSF Database of Discharged and Inspected Channels as a Function of Plant Type, Exposure Range and Residence Time [[

]] Table 9a-3 Expected NSF Channel Database Once Channels Inserted by the Fall of 2015 are Discharged and Inspected (also as a function of plant type, exposure range and residence time) [[

]]

B-42 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public) b) Please provide the ECBE, EFPD and exposure (assembly burnup) for the channels in the following table.

Unit Type - Lattice Cycle Number of Year Year Inserted Channels Beginning Discharge [[

]]

Response: Table 9b-1

Unit Type - Cycle # of Years in Residence Exposure ECBE (in- Lattice Inserted Channels Operation Time (days) (GWd/MTU) days) [[

]]

B-43 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public) c) Update the figure below, Inferred Shadow Bow versus ECBE, with the latest data. [[

]] Figure MFN 12-134, p. 75 (76 of 82 in pdf)

Response: [[

]] Figure 9c-1

B-44 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

RAI-10 Future Surveillance / Reporting Program The enhanced LUC program was approved to expedite data collection to support batch approval of NSF channels. Depending on the response to RAI-9 above, additional NSF in-reactor data may be necessary. In similar, past situations, the NRC has accepted a surveillance program which mandates data collection, confirmation of empirically-based models or performance, reporting requirements, and, where necessary, actions to ensure safe operation. The NRC has been willing to accept this approach when large quantities of lead use prototypes will continue to lead batch application such that compensatory action would be possible to avoid safety issues. a) Please propose a surveillance program for the collection of NSF channel growth and distortion data and the confirmation of fluence gradient-induced bow and shadow corrosion-induced bow models. As part of this response, describe a process for updating models, implementing models, and reporting. The NRC is looking for assurance that existing fuel management guidelines, compensatory measures, and augmented control blade surveillances are not minimized prior to achieving high confidence NSF models. Response Monitoring Plan During Operation The expanded NSF LUC program (Reference 10-1) includes provisions to ensure that NSF LUC channels are included in the normal plant Technical Specifications (TS) scram-time testing of 10% of the control rods every 120 days (nominally). Plants that have loaded NSF channels under the 8% LUC program will continue to monitor those channels based on the provisions of the LUC program. As discussed in the response to RAI-9, the population of plants where NSF has been inserted under the 8% LUC program is representative of the plants that have experienced channel control-blade interference (including all four BWR/6, S-Lattice plants in the US). However, when reload quantities of NSF are loaded into the core, the probability of scram- time testing cells with NSF channels is greatly increased such that the specific provisions of the LUC program for added scram-time testing are no longer needed to ensure NSF channels are adequately included in the tested population. If there is a friction observation in a cell, the plant will immediately go into MFN 10-245 R6 (Reference 10-2) or MFN 08-420 R1 (Reference 10-3) testing. For plants transitioning to full core use of NSF channels, the friction monitoring provisions of References 10-2 and 10-3 will continue to apply. That is, these procedures will remain in place to manage any potential unforeseen channel distortion issues with NSF. However, the expectation is that as plants transition to 100% NSF there will be no observations of channel–control blade interference. Post-Irradiation Surveillance Plan The 8% LUC program requires that a subset [[ ]] of the expanded NSF LUC program channels be inspected visually to evaluate corrosion performance and to measure the length change. Further, after discharge, a subset [[ ]] of the expanded NSF

B-45 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public)

LUC program channels will be [[ ]]. In addition after discharge, [[ ]] of the NSF LUC channels to confirm that [[ ]]. This program will continue on channels introduced via the 8% LUC program. A total of more than 300 channels will be inserted into six plants as part of the 8% LUC program representing operation in all plant lattice types (D-, C- and S-Lattice) and both the thick and thin channel designs. In addition, there are 28 NSF channels operating in four US plants as part of the normal 2% LUC program. This population of channels represents a good cross section of variations in channel material within the specification range and operational experience and will lead the reload quantities by at least one two-year cycle. Per the inspection requirements, channel distortion measurements will be performed on approximately 225 NSF channels; the expected range of conditions in this measurement population was provided in the response to RAI-9 (see Table 9a-3). Because these lead channels are being placed in aggressive locations relative to the mechanics of channel distortion, the experience and measurements from these channels alone will provide sufficient evidence for the observed variability in the distortion of NSF channels and confirm the fluence gradient-induced bow and shadow corrosion-induced bow models. In addition to the channel visual, length, and distortion measurements, GNF will measure oxide thickness using a non-destructive, pool-side technique or alternative method on ~20 high exposure bundles (~6 and ~8-year residence time) to confirm the oxide thickness design curve in Figure 2-12. As the full reloads of NSF channels are discharged, inspections of channel distortion will only occur as is deemed necessary. Reporting GNF will provide a progress report to the NRC each year on the inspection program results for both the 8% LUC program and the requirements as described in the NSF LTR. When the post-irradiation surveillance protocol is complete, GNF will provide an information report to the NRC summarizing the results. It is anticipated that the proposed fluence gradient- induced bow and shadow corrosion-induced bow models will be confirmed and will not require modification. In the unlikely event that changes to the fluence gradient-induced bow and shadow corrosion- induced bow models are necessary, GNF will evaluate the impact on the process for calculating the Critical Power Ratio (CPR) defined in the NSF LTR and communicate the results of that evaluation to the NRC in an information letter.

B-46 NEDO-33798-A Revision 1 Non-Proprietary Information – Class I (Public) b) Similar to above, for the core-average, cell-average bow input to the channel-bow dependent critical power ratio calculation.

Response The recommendation in the NSF LTR is to use a [[ ]] for the core-average, cell-average bow for NSF channels. This recommendation is based on the observation that [[ ]] (Figure 3-1 in the LTR). These predictions are based on the channel bow models provided in Appendix A. As discussed above in the section on the Post Irradiation Inspection Plan, channel bow will be measured on approximately 225 NSF channels to confirm the channel bow models. If the channel bow data collected on NSF indicates a need to modify the channel bow models, the impact on the prediction of core-average, cell-average bow will be evaluated and communicated to the NRC. References 10-1 Letter, A.A. Lingenfelter (GNF) to Document Control Desk (US NRC), “Accepted Version of Enhanced Lead Use Channel (LUC) Program for NSF Fuel Bundle Channels,” MFN 12-074 Supplement 2-A, April 15, 2013. 10-2 Letter, D. E. Porter (GEH) to Document Control Desk (US NRC), “Update to Part 21 Reportable Condition Notification: Failure to Include Seismic Input in Channel-Control Blade Interference Customer Guidance,” MFN 10-245 R6, December 16, 2013. 10-3 Letter, D. E. Porter (GEH) to Document Control Desk (US NRC), “Update to GEH Surveillance Program for Channel-Control Blade Interference Monitoring,” MFN 08-420 R1, December 16, 2013.

B-47