ML19052A572

Final Safety Evaluation for Framatome, Inc. Topical Report ANP-10332P, Revision 0, AURORA-B: an Evaluation Model for Boiling Water Reactors; Application to Loss of Coolant Accident Scenarios (Non-Proprietary Version)

ML19052A572
Person / Time
Site:	PROJ0728, 99902041
Issue date:	03/26/2019
From:	John Lehning NRC/NRR/DSS/SNPB
To:	Peters G Framatome
Rowley J, NRR/DLP, 415-4053
Shared Package
ML19057A551 -Pkg.	List: ML19052A572 ML19057A555
References
CAC MF3829, EPID: L-2014-TOP-0004 ANP-10332P, Rev 0
Download: ML19052A572 (186)

Text

FINAL SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION TOPICAL REPORT ANP-10332P, REVISION 0 AURORA-B: AN EVALUATION MODEL FOR BOILING WATER REACTORS; APPLICATION TO LOSS OF COOLANT ACCIDENT SCENARIOS FRAMATOME, INC.

PROJECT NO.: 728/DOCKET NO. 99902041 Enclosure 2

Table of Contents

1.0 INTRODUCTION AND BACKGROUND

....................................................................... 1.1 The AURORA-B Code System and Evaluation Models............................................. 1.2 Scope of Application and NRC Staff Review ............................................................. 1.3 Evolution of Appendix K LOCA Methodologies .........................................................

2.0 REGULATORY EVALUATION

...................................................................................... 2.1 Applicable Regulatory Requirements ........................................................................ 2.2 Relevant Regulatory Guidance ................................................................................

3.0 TECHNICAL EVALUATION

........................................................................................ 3.1 Overview of BWR LOCA Event ............................................................................... 3.1.1 BWR Large-Break LOCA ................................................................................. 3.1.2 BWR Small-Break LOCA .................................................................................. 3.1.3 BWR Intermediate-Break LOCA ....................................................................... 3.2 Identification and Ranking of Relevant Phenomena ................................................ 3.3 Evaluation Model Development ............................................................................... 3.3.1 Overview of S-RELAP5 Models and Correlations ............................................ 3.3.2 Overview of RODEX4 Kernel Models and Correlations ................................... 3.3.3 Coupling of S-RELAP5 and RODEX4 Kernel ................................................... 3.3.4 Modeling Options and Nodalization .................................................................. 3.3.5 Application Framework ..................................................................................... 3.4 Assessment and Validation of Evaluation Model ..................................................... 3.4.1 Foundation Methodology Assessments ........................................................... 3.4.2 Component Effects Tests ................................................................................. 3.4.3 Separate Effects Tests ..................................................................................... 3.4.4 Integral Effects Tests ...................................................................................... - 107 -

3.4.5 Impact of ['''''''''''''''''] Code Version ............................................................... - 120 -

3.4.6 Summary of Assessment and Validation ........................................................ - 122 -

3.5 Sensitivity Studies ................................................................................................. - 127 -

3.5.1 Timestep Length ............................................................................................. - 128 -

3.5.2 Hot Channel Axial Nodalization ...................................................................... - 128 -

3.5.3 Core Axial Power Shape ................................................................................ - 129 -

3.5.4 Core Radial Power Shape .............................................................................. - 130 -

3.5.5 Fuel Channel Grouping / Parallel Channel Flow ............................................ - 131 -

3.5.6 Break and ECCS Injection Nodalization ......................................................... - 135 -

3.6 Demonstration Analyses ........................................................................................ - 135 -

3.6.1 BWR/4 and BWR/6 Demonstration Cases ..................................................... - 136 -

3.6.2 NRC Staff Audit of Demonstration Case Files ................................................ - 139 -

3.6.3 Impact of [''''''''''''''''] Code Version ['''' ''''''''' ''''''''''''''''''''''''] ....................... - 155 -

3.6.4 AURORA-B Comparison Cases Versus EXEM BWR-2000 Model ................ - 156 -

4.0 ADMINISTRATIVE REQUIREMENTS ...................................................................... - 158 -

4.1 Documentation ...................................................................................................... - 158 -

4.2 Quality Assurance Plan ......................................................................................... - 160 -

4.3 Update Process ..................................................................................................... - 161 -

5.0 LIMITATIONS AND CONDITIONS ........................................................................... - 163 -

6.0 COMPLIANCE

SUMMARY

....................................................................................... - 167 -

6.1 Conformance with Relevant Regulatory Guidance ................................................ - 168 -

6.2 Compliance with Applicable Regulatory Requirements ......................................... - 168 -

6.2.1 Conformance to Appendix K to 10 CFR 50 .................................................... - 168 -

6.2.2 Compliance with Relevant Criteria from 10 CFR 50.46 .................................. - 179 -

6.2.3 Compliance with General Design Criterion 35................................................ - 179 -

6.2.4 Proposed Rule 10 CFR 50.46c ...................................................................... - 179 -

7.0 CONCLUSION

.......................................................................................................... - 180 -

8.0 REFERENCES

.......................................................................................................... - 180 -

1.0 INTRODUCTION AND BACKGROUND

This safety evaluation (SE) documents the NRC staffs review of a Framatome Inc. (Framatome, formerly AREVA Inc.) topical report (TR) that describes a methodology for analyzing the complete spectrum of postulated loss-of-coolant accidents (LOCAs) for certain types of boiling-water reactors (BWRs). The methodology proposed by Framatome is intended to conform to conservative requirements prescribed in Appendix K to Title 10 of the Code of Federal Regulations Part 50 (10 CFR 50).

The methodology proposed by Framatome is contained in TR ANP-10332P, Revision 0, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Loss of Coolant Accident Scenarios (Reference 1), which was submitted in March 2014. The NRC staff initiated its review of this report in August 2016, held an audit on May 16-18, 2017 (Reference 2), and issued requests for additional information (RAIs) on October 24, 2017 (Reference 3), and January 19, 2018 (Reference 4). Framatome formally responded to the NRC staffs RAIs in a submittal dated April 30, 2018 (Reference 5).

The NRC staff completed its draft SE on August 31, 2018 (Reference 65). In an effort to address certain limitations and conditions imposed by the draft SE, Framatome submitted an updated response to the NRC staffs RAIs on October 31, 2018 (Reference 66). The NRC staff has reviewed the updated response and, as appropriate, has modified the final SE in response.

At the time ANP-10332P was submitted, Framatome went by the name AREVA Inc. (AREVA).

In January 2018, AREVA changed its name to Framatome. Since the name change occurred during the course of the review, the AREVA name is used in the incoming submittal, the NRC staffs RAIs, and other relevant documents. For simplicity, with the exception of the references listed in Section 8.0, this SE generally uses the name Framatome, even when describing events and circumstances that occurred prior to the name change.

The NRC staff conducted its review of ANP-10332P with the assistance of consultants from Brookhaven National Laboratory, who provided a technical evaluation report (Reference 35)

(withheld for proprietary reasons) that has served as a reference source in the development of this SE. The Brookhaven technical evaluation report focuses in particular on the assessment and validation of the AURORA-B LOCA evaluation model, which is covered primarily in Section 3.4 of this SE.

1.1 The AURORA-B Code System and Evaluation Models The name AURORA-B refers to a code system developed by Framatome to perform safety analysis for BWRs for a wide variety of events specified in Chapter 15 of the Standard Review Plan (SRP) (Reference 6), including anticipated operational occurrences and accidents. In full generality, the AURORA-B code system is composed of three subsidiary components:

• S-RELAP5, a system-level thermal-hydraulic code that is the primary component,

• a subset of routines from the RODEX4 code intended for the prediction of fuel thermal-mechanical performance under transient conditions, and

• MB2-K, a neutron kinetics code developed from MICROBURN-B2.

For each proposed application of AURORA-B to a defined subset of the safety analysis required for a BWR, Framatome has submitted a TR describing relevant aspects of the AURORA-B code system along with the overall framework of assumptions, required code modeling options, initial conditions, figures of merit, and so forth. The collective union of these subjects, as described in summary detail in the associated TR, constitutes an application-specific evaluation model. It should be emphasized that, while the same basic code system may support a number of distinct applications, for each application necessitating a distinct framework of assumptions, modeling options, system nodalizations, figures of merit, etc., a unique evaluation model is typically defined. As noted above, the present review pertains to the AURORA-B evaluation model for analyzing BWR LOCA events in accordance with the methodology prescribed in Appendix K to 10 CFR 50. Thus, this SE will refer to the methodology described in ANP-10332P as the AURORA-B LOCA evaluation model.

Of the three AURORA-B subsidiary components described above, the LOCA evaluation model makes use of only two: S-RELAP5 and the subset of routines from RODEX4 (hereafter referred to as the RODEX4 kernel). ANP-10332P states that a three-dimensional (3D) neutron kinetics method is unnecessary for analyzing the LOCA event, and, in lieu of MB2-K, the AURORA-B LOCA evaluation model relies upon the S-RELAP5 point kinetics model. Consequently, the present review does not cover the MB2-K code. Furthermore, inasmuch as the RODEX4 kernel has been integrated into S-RELAP5, it is accurate to consider the AURORA-B LOCA evaluation model an S-RELAP5-based methodology.

Although the codes supplying initial conditions are not strictly part of the AURORA-B LOCA evaluation model, the evaluation model depends on receiving accurate inputs from these models to function properly; for example:

• Nodal core simulator and lattice physics codes are used to initialize core thermal-hydraulic conditions for steady-state S-RELAP5 simulations and provide inputs to the S-RELAP5 point kinetics model.

o Framatomes current nodal core simulator code is MICROBURN-B2, and its current lattice physics code is CASMO-4. These codes provide ['''''' ''' '''''''''''''''''''''''' ''''''

''''''''''''''''''''''''''''''''''' ''''''''''''''''''' ''''' ''''''''''''''' ''''''''''''''''''''''''] necessary to initialize S-RELAP5 simulations.

o The XCOBRA code may also be used [' ''''''''''''''''''' '''''''''''''''''''''' ''''''

''''''''''''''''''''''''''''''''' '''''''''''''''''] for initializing S-RELAP5. XCOBRA [''''''' '' ''''''''''''

''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''']

• The full version of the RODEX4 code provides inputs to AURORA-B concerning fuel rod material properties, geometry, and long-term effects associated with burnup.

In response to RAI 44, Framatome clarified that core simulator and lattice physics methods used with the AURORA-B LOCA evaluation model must have received prior approval by the NRC staff. The NRC staff considers the use of approved core simulator methods (e.g., MICROBURN-B2) appropriate for assuring the quality of inputs to the AURORA-B LOCA evaluation model. In addition, since [''''' '''' ''''''''''''''''''''''''' ''''''''''''''' ''''''''''''''''''''''] from the

lattice physics code (e.g., CASMO-4) are also required inputs to the AURORA-B LOCA evaluation model, such inputs must likewise be taken from an approved method. The NRC staffs position on the provision of inputs to the AURORA-B LOCA evaluation model from an approved core simulator and lattice physics methodology is captured in Section 5.0 of this SE as Limitation and Condition 1.

Furthermore, because (1) ANP-10332P credits RODEX4 [''''''''''''''''''''''' '''''''''''' '''' '''''''''''''''''''

'''''''''''''''''''''' '''''' '' ''''''''' '' '''' '''''''''''''''''''''''''''' ''' '''''''''' '''''''''''] used in the AURORA-B LOCA evaluation model, and (2) the kernel of fuel thermal-mechanical performance routines in AURORA-B is based on RODEX4 models, the NRC staffs acceptance of the AURORA-B LOCA evaluation model is also contingent upon Framatome using the stand-alone version of the RODEX4 code, in accordance with an approved methodology, to supply the initial conditions necessary for modeling fuel rods in the AURORA-B LOCA evaluation model.

Framatome confirmed its acceptance of this position in response to RAI 12. Use of the full, stand-alone version of RODEX4 in accordance with an approved methodology to supply fuel performance inputs is captured in Section 5.0 of this SE as Limitation and Condition 2.

1.2 Scope of Application and NRC Staff Review As will be described in greater detail in Section 3.3 of this SE, a number of aspects of the AURORA-B code system and its component codes are familiar to the NRC staff in connection with previous reviews. While the NRC staff performed a comprehensive review of the AURORA-B LOCA evaluation model described in ANP-10332P and its numerous references, the review concentrated especially on unique aspects of the AURORA-B LOCA evaluation model that had not been previously reviewed and approved for similar applications. In light of the NRC staffs partial reliance on past reviews for similar applications, a discussion of previous NRC reviews associated with the AURORA-B code system and its subsidiary components is in order.

In addition to the AURORA-B LOCA evaluation model reviewed herein, Framatome has recently received approval for two other application-specific AURORA-B evaluation models:

• ANP-10300P, Revision 0, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Transient and Accident Scenarios (Reference 7), describes an AURORA-B evaluation model applicable to many anticipated operational occurrences and several accidents described in Chapter 15 of the SRP (Reference 6). This evaluation model uses a best-estimate approach with a non-parametric statistical method for quantifying uncertainties. The NRC staff issued its SE for ANP-10300P (Reference 8) on January 5, 2018.

• ANP-10333P, Revision 0, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Control Rod Drop Accident (CRDA) (Reference 9), describes an AURORA-B evaluation model applicable to the control rod drop accident. The calculational uncertainty is determined according to a simplified statistical methodology.

The NRC staff issued its SE for ANP-10333P (Reference 10) on March 15, 2018.

While S-RELAP5 and a kernel of routines from RODEX4 have been incorporated into the AURORA-B code system, versions of these codes also exist as independent entities that have

previously been reviewed by the NRC staff and approved for specific applications. Table 1 summarizes past reviews of TRs relevant to the current review.1 As stated in ANP-10332P, the AURORA-B LOCA evaluation model is applicable exclusively to BWR designs that incorporate jet pumps internal to the reactor vessel. Among other design functions, internal jet pumps allow reflooding of the reactor core to approximately two-thirds height following the limiting postulated LOCA. Internal jet pumps were first introduced by the General Electric Company (GE) in the BWR Type 3 (BWR/3) design, and have been included in subsequent operating BWR designs through the BWR Type 6 (BWR/6). The BWR Type 2 (BWR/2) design, which lacks jet pumps, is thus outside the scope of ANP-10332P. Similarly, application of the AURORA-B LOCA evaluation model [ ''''''''''''''''''''''''''''''''''''''''''''' '''''''''''''

''''''''''''''''''''''''' '' '''''''''''''''''''''''' ' ''''''''''' ' ''''' ''''''''''''''''''''''''''''' ' '''' '''''''] the scope of the current review.

Table 1: Previous NRC Staff Reviews Pertaining to Subsidiary Components of AURORA-B Topical Date of NRC Description Report Safety Evaluation EMF-2103P, Realistic large-break LOCA methodology for pressurized-6/17/16 Revision 3 water reactors (PWRs) using S-RELAP5 PWR small-break LOCA evaluation model using EMF-2328P 3/15/01 S-RELAP5 Standard Review Plan Chapter 15 non-LOCA methodology EMF-2310P 5/11/01 for PWRs using S-RELAP5 Realistic thermal-mechanical fuel rod methodology for BAW-10247P 2/12/08 BWRs using RODEX4 Evaluation and validation of CASMO-4/MICROBURN-B2 EMF-2158P 10/18/99 for application to BWRs Across the BWR/3-6 designs within the scope of ANP-10332P, the plant response to postulated LOCA events tends to proceed in a fairly similar manner. However, design differences in the emergency core cooling system (ECCS) and recirculation system are relevant; in particular, incorporation of the high-pressure core spray system and slightly reduced recirculation system piping sizes result in enhanced core cooling performance for later BWR designs (Reference 55).

As stated in ANP-10332P, the AURORA-B LOCA evaluation model is intended to support the licensing basis for each BWR/3-6 plant that does not significantly depart from the SRP. The NRC staff requested in RAI 1 (Reference 3) that Framatome clarify the intent of this statement.

Framatome responded that, while defining a significant departure is inherently subjective, the AURORA-B LOCA evaluation model is [''''''''''''''''' '' '''''''''''''''''''' ''''''''''''''''' '''''''''''' ''''

' ''''''''''']. Furthermore, Framatome observed that, inasmuch as a licensees initial implementation of the AURORA-B LOCA evaluation model is envisioned as necessitating a license amendment request, the NRC staff would have opportunity to review such applicability justifications. The NRC staff agrees that plant-specific implementation reviews are an appropriate vehicle for ensuring sufficient consistency of plant licensing bases with the SRP to support application of the AURORA-B LOCA evaluation model.

1 In instances where TRs have multiple revisions or supplements, for brevity, only the most relevant is reflected.

As stated in ANP-10332P, the AURORA-B LOCA evaluation model is not intended for containment analysis; in fact, an atmospheric boundary condition is assumed in ANP-10332P in lieu of a mechanistic containment model. Framatome suggested, however, that the mass and energy output from AURORA-B may be used as an input to other downstream analyses that are not specifically described in the TR. Because specific methods for determining a limiting mass and energy release are not described in ANP-10332P (e.g., procedure for determining limiting break location, selection of discharge coefficients), the NRC staff considered such applications beyond the scope of the current review.

Similar to a condition imposed previously on implementation of ANP-10300P, the NRC staff determined that the AURORA-B LOCA evaluation model may not be used to perform analyses that result in any of its constituent components or supporting codes (i.e., S-RELAP5, RODEX4 kernel, RODEX4, core simulator and lattice physics codes) being operated outside limits of approval documented in their respective TRs, SEs, code manuals, and plant-specific licensing applications. This position is documented as Limitation and Condition 3 in Section 5.0 of this SE.

1.3 Evolution of Appendix K LOCA Methodologies To place the NRC staffs review of the AURORA-B LOCA evaluation model into perspective, it is worth reflecting upon the conservatisms traditionally associated with Appendix K LOCA methods. According to NUREG-1230, which compiles research supporting the 1988 revision to 10 CFR 50.46 to permit realistic LOCA analysis methods, the conservative margin in Appendix K-based methods has traditionally been expected to be quite substantial, with estimates ranging from several hundred to one thousand degrees Fahrenheit (°F)

(Reference 13).

The conservatism of previously approved Appendix K LOCA evaluation models has typically arisen from two distinct sources: (1) requirements imposed directly in Appendix K to 10 CFR 50 and (2) discretionary conservatisms not specifically required by Appendix K that were adopted as practical simplifications, for example, due to computational or state-of-knowledge limitations.

In fact, a limited set of key models is directly prescribed in Appendix K as required features, such as the Baker-Just correlation for the metal-water reaction, the Moody model for two-phase critical flow, and the proposed 1971 American Nuclear Society 5.1 standard with a multiplier of 1.2 for decay heat. Whereas, in most other areas of the analysis, including those for which Appendix K defines certain features as acceptable (but not required), vendors are permitted to propose other models they can demonstrate to be acceptable.

Framatomes AURORA-B LOCA evaluation model relies upon a modified version of a thermal-hydraulic code originally developed for best-estimate analysis of PWRs (i.e., S-RELAP5) to perform BWR LOCA calculations according to Appendix K. While all required Appendix K conservatisms would be maintained, the increased sophistication of the AURORA-B LOCA evaluation model would overcome some limitations associated with previous generations of LOCA evaluation models through the use of more complex models and techniques (e.g., a [''''''''''''''''''''''''''''''''''' '''''' ''''''''''''''''''''''''''], non-equilibrium thermodynamic modeling, etc.). In addition to reducing conservatism, experience shows that the increased complexity of modern analytical methods may result in increased susceptibility to variations in code output from nominal input changes. On the other hand, in light of NRC staff comments on previous reviews, Framatome has proposed an ['''''''''''''''''' ''''''''''''''''''''''''' '''''''''''''' ''''''''''''''''''''

'''''''''''' ''''''''''''''''' ''''''' '''''''2,3 ' '''''''' '''' '''''''''''''''''''' '''''''''''''''''' '''''''''''''''''''''

'''''''''''''' '' ''' '' '''' '''''''' '''''' ''''''''''''''''''''''''''''''''''' ''''' ''''''''''''].

The net effect of these changes is not obvious a priori. Therefore, this SE (e.g., in Section 3.6) will attempt to illustrate the impacts of these countervailing effects by providing an indication of the overall conservatism of the AURORA-B LOCA evaluation model.

2.0 REGULATORY EVALUATION

2.1 Applicable Regulatory Requirements The AURORA-B LOCA evaluation model is a methodology applicable to BWR/3-6 plants for demonstrating compliance with specific requirements in 10 CFR 50.46 concerning the core cooling performance of light-water reactors during a postulated LOCA event. Among the requirements of 10 CFR 50.46, several are particularly worthy of mention:

• Each light-water power reactor licensee using uranium oxide fuel clad with certain types of zirconium alloy4 must perform analysis of core cooling performance under postulated LOCA conditions using an acceptable evaluation model.

• An acceptable LOCA evaluation model must be used that either applies realistic methods with an explicit accounting for uncertainties or follows the prescriptive, conservative requirements of Appendix K to 10 CFR 50.

• Core cooling performance must be analyzed for a number of postulated LOCAs of different sizes, locations, and other characteristics to ensure that the most severe event is calculated.

• According to 10 CFR 50.46(b), for the most severe postulated LOCA event, the calculated performance of the ECCS must demonstrate that o the peak cladding temperature does not exceed 2200 °F, o the calculated local cladding oxidation does not exceed 17 percent of the cladding thickness prior to oxidation, o the total amount of hydrogen generated from the metal-water reaction is limited to 1 percent of the hypothesized theoretical maximum, 2 The [proposed downflow restriction is similar in concept to restrictions on top-down cooling imposed in LOCA analyses for some licensees that reference Framatomes existing] LOCA evaluation model for BWRs (e.g., Reference 60).

3 According to EMF-2361(P)(A) (Reference 16), this pressure is [approximately 100-120 psig].

4 10 CFR 50.46(a)(1)(i) strictly applies to each boiling or pressurized light-water reactor fueled with uranium oxide pellets within cylindrical zircaloy or ZIRLO cladding. However, by means of exemptions, the requirements of 10 CFR 50.46 are currently in effect for all operating (light-water) power reactors, regardless of the specific (zirconium-based) cladding alloy in use.

o the reactor core is maintained in a coolable geometry, and o sufficient long-term core cooling is provided to maintain an acceptable core temperature.

In lieu of developing best-estimate physical models and explicitly accounting for uncertainty, Framatome has proposed the AURORA-B-based methodology in ANP-10332P as a LOCA evaluation model for performing calculations to demonstrate compliance with 10 CFR 50.46 according to the prescriptive, conservative requirements of Appendix K to 10 CFR 50. In particular, as described in Section 4.4 of ANP-10332P, the AURORA-B LOCA evaluation model provides calculated results for three figures of merit (peak cladding temperature, maximum local cladding oxidation, and core-wide cladding oxidation) for comparison with the quantified regulatory limits in 10 CFR 50.46(b)(1)-(3). As the Commission stated in its opinion regarding the rulemaking hearing on ECCS acceptance criteria (Reference 14), the purpose of the first two criteria (peak cladding temperature and maximum local cladding oxidation) is to ensure that the fuel cladding would remain sufficiently intact to retain the fuel pellets in a rod array configuration that constitutes a coolable geometry. Thus, the AURORA-B LOCA evaluation model is further capable of determining compliance with the qualitative requirement to maintain the reactor core in a coolable geometry.

With regard to the final criterion from 10 CFR 50.46(b), ANP-10332P states that demonstration of adequate long-term core cooling is beyond the scope of the AURORA-B LOCA evaluation model. ANP-10332P further indicates that adequate long-term core cooling is demonstrated for BWRs with jet pumps (i.e., BWR/3-6) via reference to a generic analysis demonstrating that, with two-thirds core coverage, the upper third of the core can be adequately cooled by a combination of core spray and steam cooling. Such demonstration has been previously provided, for instance, in GE TR NEDO-20566A (Reference 15). The NRC staff has specified in Limitation and Condition 4 that ['' ''''''''''''''''''''''' ''''''''''' ''''''''''''''''''''' ''''''''''' ' '''' ''''''''''''''''''' ''

'''''''''''''''''''' '' '''''''''''''''''' ']

Appendix K to 10 CFR 50 consists of two parts, the first of which specifies required and acceptable features of LOCA evaluation models, and the second of which specifies documentation required for LOCA evaluation models. In particular, Appendix K specifies requirements for modeling significant physical phenomena throughout all phases of the LOCA event, including relevant heat sources, fuel rod performance, and thermal-hydraulic behavior.

Conformance to Appendix K is fundamental to the acceptability of the AURORA-B LOCA evaluation model; as a result, this topic will be addressed at various junctures throughout this SE and summarized in Section 6.2.1.

An additional, fundamental regulatory requirement relevant to ECCS performance that is generally included in the licensing bases of operating power reactors is General Design Criterion (GDC) 35 from Appendix A to 10 CFR 50.5 The GDC of Appendix A to 10 CFR 50 were finalized in 1971 (approximately three years prior to the issuance of 10 CFR 50.46 and Appendix K), and they outline criteria for the design of nuclear power plants in broad, qualitative 5

Alternatively, some plants may be licensed to a proposed draft or plant-specific criterion with a similar intent.

terms. In particular, GDC 35 requires abundant core cooling sufficient to (1) prevent fuel and cladding damage that could interfere with effective core cooling and (2) limit the metal-water reaction on the fuel cladding to negligible amounts. GDC 35 further requires suitable redundancy of the ECCS, such that it can accomplish its design functions, assuming a single failure, irrespective of whether its electrical power is supplied from offsite or onsite sources.

Finally, Sections 2.5.1 and 4.4 of ANP-10332P indicate that the AURORA-B LOCA evaluation model may be applied, without substantial modification, for demonstrating compliance with proposed rule 10 CFR 50.46c (Reference 61). Specifically, Framatome suggested that evaluation model revisions would not be necessary; rather, only the limiting values for figures of merit (e.g., peak cladding temperature, cladding oxidation) may require revision to address issues such as cladding embrittlement via hydrogen uptake. However, the NRC staff noted that (1) proposed rule 10 CFR 50.46c has not been finalized and (2) no basis for compliance with the proposed rule and associated draft guidance documents was provided in ANP-10332 to support Framatomes position. Therefore, the NRC staff could not effectively review the AURORA-B LOCA evaluation model relative to its acceptability for determining compliance with proposed rule 10 CFR 50.46c. Limitation and Condition 5 emphasizes that the conclusions of this SE apply only to the use of the AURORA-B LOCA evaluation model for the purpose of demonstrating compliance against currently existing regulatory requirements (i.e., as of December 31, 2018); in particular, this SE does not confer approval for demonstrating compliance with proposed rule 10 CFR 50.46c.

2.2 Relevant Regulatory Guidance In its review of the AURORA-B LOCA evaluation model, the NRC staff consulted regulatory guidance relevant to (1) the LOCA event and (2) the development and assessment of analytical evaluation models. Guidance on these topics is provided in both the SRP, which assists the NRC staff in performing consistent and predictable regulatory reviews, and in regulatory guides that provide licensees and vendors recommended approaches for complying with NRC regulations.

The most relevant chapters of the SRP consulted by the NRC staff during its review of ANP-10332P include

• Chapter 15.0, Revision 3, Introduction - Transient and Accident Analysis,

• Chapter 15.0.2, Revision 0, Review of Transient and Accident Analysis Methods, and

• Chapter 15.6.5, Revision 3, Loss-of-Coolant Accidents Resulting from Spectrum of Postulated Piping Breaks within the Reactor Coolant Pressure Boundary.

Chapters 15.0 and 15.6.5 of the SRP present guidelines for the review of safety analyses and corresponding analysis methods, affirming the regulations discussed above and generally providing additional recommendations regarding how compliance may be demonstrated.

Additional key guidelines from these chapters relevant to the LOCA event that are not specifically mentioned above include

• allowing credit only for safety-related equipment,

• carrying out analyses until the core has been recovered with a two-phase mixture and cladding temperatures have been reduced to near saturation conditions,

• presenting appropriate analyses to support any credit taken for control rod insertion, and

• demonstrating consistency of calculated results with previous calculations performed by the NRC staff or vendors.

In addition, Chapter 15.0.2 of the SRP identifies key review areas for transient and accident analysis methods that are relevant to the current review, namely:

• the documentation for the evaluation model,

• the evaluation model itself, which typically consists of computer codes and the complete framework of inputs and assumptions necessary to perform licensing-basis analyses,

• the identification of the accident or transient sequence, the ranking of relevant phenomena, and determination of modeling requirements,

• the assessment of the evaluation model against applicable experimental data or exact mathematical solutions,

• assurance that calculational uncertainties are bounded, either through an explicit uncertainty analysis (for best-estimate methods), or through the use of a prescriptive method with adequate built-in conservatism (i.e., Appendix K to 10 CFR 50), and

• the quality assurance plan for the evaluation model.

The topical structure of this SE derives from these key review areas from Chapter 15.0.2 of the SRP. However, the sequence has been modified to present information in the most logical and comprehensible order possible.

Relevant regulatory guides consulted by the NRC staff in its review of ANP-10332P include

• Regulatory Guide (RG) 1.157, May 1989, Best-Estimate Calculations of Emergency Core Cooling System Performance (Reference 18)

• RG 1.203, December 2005, Transient and Accident Analysis Methods (Reference 19)

RG 1.157 provides guidance related to best-estimate analysis of the LOCA event in a format that closely follows Appendix K. Because the AURORA-B LOCA evaluation model is an Appendix K method, RG 1.157 was not directly used in the review.

On the other hand, ANP-10332P identifies that the evaluation model development and assessment process (EMDAP) described in RG 1.203 was followed in developing the AURORA-B LOCA evaluation model. The EMDAP consists of four elements, each with a number of subsidiary steps. Framatomes rationale that the AURORA-B LOCA evaluation model complies with each element in the EMDAP is documented in ANP-10332P. Inasmuch as

ANP-10332P is further structured in accordance with the EMDAP, this process is diagrammed in Figure 1 below, which has been adapted from RG 1.203.

Another relevant aspect of RG 1.203 that deserves emphasis is the graded approach recommended in Appendix B therein for reviewing a proposed new application for a code system that has previously been reviewed in a different context. As discussed above, the AURORA-B code system and its component parts have been the subject of a number of past NRC staff reviews, which has permitted simplification of the present review with focus on novel features and application-specific topics, as discussed further in Section 3.3 of this SE.

Figure 1: The Evaluation Model Development and Assessment Process6 6

Abbreviations in figure:

EM evaluation model IET integral effects test SET separate effects test

RG 1.203 further allows that the degree of assessment required for the evaluation model may be reduced if the documented degree of conservatism is large or if the new model can be shown to give more conservative results than the previous model. Inasmuch as the AURORA-B LOCA evaluation model was developed to conform to the conservative requirements of Appendix K rather than as a best-estimate model, a reduction in the degree of rigor when assessing against the EMDAP criteria is justified. A comparison of the results of the proposed AURORA-B LOCA evaluation model with those of the currently approved evaluation model (EXEM BWR-2000 (Reference 16)), discussed further in Section 3.6.4 of this SE, ['''''''''''''''''''''''

'''' '''''''''''''''''''''''].

3.0 TECHNICAL EVALUATION

3.1 Overview of BWR LOCA Event A LOCA event involves a pressure boundary breach that results in an uncontrolled blowdown of reactor coolant at a rate in excess of the normal makeup capacity. According to 10 CFR 50.46, the maximum postulated size of a LOCA is the double-ended rupture of the largest pipe in the reactor coolant system. Consistent with the nomenclature of Appendix K to 10 CFR 50, a LOCA event is traditionally divided into three phases; for a BWR these phases are typically defined as follows:

• blowdown, which begins when a pressure boundary breach results in a loss of reactor coolant and uncontrolled reactor depressurization,

• refill, which begins when the reactor pressure has ['''''''''''''''''' ''''''''''''''''''' ' '''''''

'''''''''''''''''''''''''' ''''''''' ''''''''''''' ''''''' '''''''''''' ''' '''''''''''' ''''''''''' '''' '''''''''''''''''''' ''''''''],

and

• reflood, which begins ['''''''' '''''''''''''''''''''''''''' ''''''''' '''''''''''' ''''''' '''''''''' '' '''''''''''

''''' '''' '''''''''''''''''''''''' '''''''''''''''' ''''''''' '''''''''''''''''' ''''''''''''''' '].7 In retrospect, division of the LOCA event into these sequential, neatly defined phases appears more appropriate to (1) the phenomenology of the PWR LOCA event and (2) the analytical predictions of simplified LOCA evaluation models from the early 1970s, than to the physical behavior expected during a BWR LOCA event. In particular, owing to a number of factors, such as the channeled fuel design with engineered bypass leakage paths, injection of emergency core coolant inside the core shroud, and the countercurrent flow limitation, coolant addition into core fuel bundles would occur for domestic, operating BWRs prior to the complete refilling of the lower plenum.8 Hence, application of the above definitions in the present SE is intended in this spirit.

The sequence of events for a given LOCA scenario depends upon the location of the pressure boundary breach and the rate at which reactor coolant is lost. In a BWR, piping ruptures may be postulated on various systems, such as recirculation, feedwater, main steam, and ECCS.

Previous analyses typically indicate that ruptures on recirculation piping are among the most 7 As noted above, the BWR/3-6 plants to which the AURORA-B LOCA evaluation model is applicable all have jet pumps that facilitate reflooding the reactor core to at least two-thirds core height.

8 Subsequent to the promulgation of Appendix K, such behavior was demonstrated, for example, by integral testing performed in the Steam Sector Test Facility.

severe postulated events; hence, this scenario is selected as an example for this introductory discussion. With the understanding that the spectrum of postulated LOCA events encompasses conditions other than piping ruptures (e.g., stuck-open, unisolable valves), as a convenience, the spectrum of BWR LOCA events is described conventionally below according to break size categories.

3.1.1 BWR Large-Break LOCA A double-ended guillotine break on recirculation system piping (e.g., with a maximum total flow area on the order of 5-7 square feet (ft2)) is typically among the most limiting LOCA events postulated for a BWR. This hypothetical event results in a rapid discharge of coolant through both ends of the ruptured pipe, which can depressurize the reactor from over 1000 pounds per square inch absolute (psia) to less than 100 psia in less than a minute. The analyzed set of large-break LOCA events encompasses large split breaks and double-ended guillotine breaks with an appropriate range of discharge coefficients.9 The rapid depressurization of the reactor following a large-break LOCA results in immediate flashing of saturated liquid in the core. This rapid vapor formation, along with the insertion of control rods, terminates the fission chain reaction within seconds of the piping rupture. As the reactor pressure continues to fall, the slightly cooler water in the lower plenum begins to flash, resulting in an updraft of two-phase flow through the reactor core. The consequent reactor power decrease, coupled with high core flow rates and decreasing saturation temperatures induced by the rapid depressurization, initially results in a reduction in fuel cladding temperature. This decrease, however, is short-lived; as the pressure and liquid fraction in the core continue to decrease, heat transfer to the two-phase coolant mixture becomes less efficient, and the fuel cladding begins to heat up. Under the limiting single-failure assumption, heatup of the fuel cladding may continue for a few of minutes in high-powered fuel bundles until it is turned around, typically by the injection of coolant from low-pressure ECCS pumps.

Further discussion and example traces for reactor pressure, cladding temperature, and other key parameters for a BWR large-break LOCA may be found in Sections 7.7.4 and 7.7.6 of ANP-10332P. Section 4.2.5 of NUREG-1230 (Reference 13) provides additional discussion of interest.

3.1.2 BWR Small-Break LOCA Small split breaks could result in limiting conditions for some BWRs, particularly in the event that a top-peaked axial power distribution is assumed. Compared to the large-break LOCA event discussed above, the greatly reduced flow area for a small-break LOCA (e.g., on the order of 0.1 ft2) results in the reactor blowing down over a significantly expanded time scale.

For the small-break LOCA event, the timings of the reactor scram and isolation of the main steam system, among other events, depend upon offsite power availability. In the case that a loss of offsite power is assumed, reactor scram and isolation will occur early in the event due to the loss of normal power. With offsite power available, reactor scram and isolation signals would eventually occur after the reactor protection system senses abnormal conditions (e.g., low reactor water level). Unlike the large-break LOCA event discussed above, a 9 Although the event timescale for such breaks would be mildly dilated due to a reduction in flow area (or effective flow area), at the level of the present discussion, the overall system response is similar.

significant reactor depressurization will not immediately occur for the small-break LOCA. In fact, because the rate of energy removal from a break on the order of 0.1 ft2 may be less than the post-trip decay heat power, isolation of the main steam system may cause an increase in reactor pressure sufficient to lift safety/relief valves that vent to the suppression pool.

Even if the reactor does not significantly depressurize initially, the loss of coolant from the break will result in a decreasing trend for reactor vessel water level. This decreasing level trend will eventually result in the actuation of a high-pressure ECCS pump that, if available, would typically be capable of mitigating such an event without an extensive cladding temperature excursion. Hence, the single failure assumed for a small-break LOCA generally involves failure of the high-pressure ECCS pump, which results in a further level decrease and actuation of the automatic depressurization system after the expiration of a delay period (e.g., two minutes).

Depending upon the size of the break, the continued loss of reactor coolant during the delay period prior to actuation of the automatic depressurization system may result in partial uncovery and heatup of the reactor core. This initial core temperature rise tends to be mitigated by a significant flashing-induced updraft of the two-phase mixture in the vessel in response to the opening of safety/relief valves by the automatic depressurization system. However, similar to the case of the large-break LOCA scenario discussed above, the pressure and liquid fraction in the reactor vessel continue to fall; as the rate of flashing slows, heat transfer becomes less effective, and a renewed core heatup occurs. As in the large-break LOCA scenario discussed above, core heatup in the limiting small-break LOCA scenario is typically turned around by the injection of coolant from low-pressure ECCS pumps.

Further discussion and example traces for reactor pressure, cladding temperature, and other key parameters may be found in Sections 7.7.4 and 7.7.6 of ANP-10332P.

3.1.3 BWR Intermediate-Break LOCA The plant response to piping ruptures in a size range between the small- and large-break LOCA scenarios discussed above may be governed by behaviors from either or both of the preceding sections. Hence, additional phenomenological discussion for intermediate breaks is not necessary. The results from the demonstration case scenarios in ANP-10332P generally showed intermediate-break LOCA cladding temperatures below those of the limiting large and small breaks; however, in many cases peak cladding temperatures were of a comparable magnitude. Hence, it is clear that the intermediate region must be analyzed to assure that the globally limiting result has been calculated.

3.2 Identification and Ranking of Relevant Phenomena According to the EMDAP shown above in Figure 1, prior to undertaking development of the evaluation model itself, phenomena relevant to the event of interest should be ranked according to their expected importance relative to the defined figures of merit. Introductory discussion in foregoing sections of this SE has already established

• accident scenarios and plant types within the intended scope of application (i.e., the entire spectrum of postulated LOCA events for BWR/3-6 plants), and

• figures of merit (i.e., peak cladding temperature, maximum local cladding oxidation, and core-wide cladding oxidation).

Prior to ranking the phenomena relevant to the scenarios of interest, the EMDAP calls for consideration of the systems, components, phases, geometries, fields, and processes that should be included in the model to accomplish the intended objective. Framatome addresses this topic in Section 4.5 of ANP-10332P, as summarized below in Table 2.

Table 2: Summary of Modeling Requirements for the AURORA-B LOCA Evaluation Model Model Feature AURORA-B Modeling Capability Systems, The BWR primary system and connected components, including the reactor subsystems and core and vessel internal geometry and structures, the recirculation system, modules the ECCS, safety/relief valves, reactor protection system, etc.

Constituent Fluids Water, noncondensable gas Fluid Phases Liquid and vapor phases of water, noncondensable gas Both wet-wall and dry-wall vertical flow regimes are necessary. In addition, Two-Phase Flow wet-wall horizontal flow configurations with stratification must be considered.

Configurations The specific flow regimes in S-RELAP5 are documented in the code theory manual (Reference 11) and are similar to those of the RELAP5 code.

S-RELAP5 uses a six-equation, two-fluid formulation for modeling two-phase flow, which includes field equations for mass, momentum, and thermal Field Quantities energy for each fluid phase (i.e., liquid and vapor). In addition, thermal conduction equations are provided to model the temperature of heat structures (e.g., fuel rods and structural metal).

Numerous process models exist to support the calculation of mass, Transport momentum, and energy transport in S-RELAP5. These processes are Processes summarized in Section 4.5 of ANP-10332P, and discussed further in Section 3.3.1.2 of this SE.

The modeling capabilities discussed above for the AURORA-B LOCA evaluation model do not capture phenomena associated with post-LOCA debris, including its transport, accumulation, and potential for restricting coolant flow to the reactor core. The evaluation model in fact assumes the coolant to be pure water. Consequently, this SE does not review the capability of the AURORA-B LOCA evaluation model for simulating the potential impacts of post-LOCA debris.

As observed in RG 1.203, all phenomena and processes that occur during a transient or accident scenario do not have equal influence on the determination of the intended figures of merit. As such, a systematic method for identifying relevant phenomena and processes and ranking them according to their expected influence promotes the effectiveness and rigor of the code development and assessment processes. To this end, as Framatome has done in support of ANP-10332P, a group of knowledgeable individuals is empaneled to use its collective judgment to identify and qualitatively rank phenomena according to their expected importance (e.g., high, medium, low) for the event of interest. The results of this process are generally tabulated, culminating in a phenomenon identification and ranking table (PIRT).

Framatome presented PIRT results for the AURORA-B LOCA evaluation model in Table 4-1 of ANP-10332P. The PIRT results are intended to characterize BWR/3-6 reactors with internal jet pumps. In accordance with RG 1.203, Framatome has broken out its PIRT results according to both space (i.e., specific systems and components, such as the reactor core, core bypass region, and recirculation system) and time (i.e., the three defined phases of the LOCA event,

namely, blowdown, refill, and reflood). Condensed results for Framatomes PIRT exercise showing the maximum ranking for each phenomenon across all three phases of the LOCA event are provided below in Table 3.

Table 3: Condensed PIRT Results for the AURORA-B LOCA Evaluation Model Core Initial Conditions Power-Related Phenomena

'''''''' '''''''''''' ''''''''''''''''''''' '''' 10 Neutronics-Related Phenomena Fuel-Related Phenomena Thermal-Hydraulic Related Fluid Phenomena Thermal-Hydraulic Flow-Related Phenomena Thermal-Hydraulic Heat Transfer Related Phenomena 10 Note that Framatome ranked core power distribution [' ''' ''''''''''''''''''''' '''''''''''''''' ''' ''''''' '''''''' '''

'''' ''''''''''''''].

Core Bypass Lower Plenum Control Rod Guide Tubes

''''''''''' ''''''''11 Upper Plenum Jet Pumps Recirculation Pumps Recirculation Lines 11 As clarified in Framatomes response to RAI 6, this phenomenon is ['''''''''''''' ' ''' ''''''''''''''''''''''''''''' ''''''

' ''''''''' '''''' '''''''''''''' ''''''''''''''''''''''''''' '''''''' '' ''' '''''' ''''''''''''' '''''''''].

Downcomer Steam Line and Automatic Depressurization System Valves According to RG 1.203, PIRT results are intended to guide the uncertainty analysis and assessment of overall evaluation model adequacy. Inasmuch as the AURORA-B LOCA evaluation model is intended to conform to Appendix K, an explicit accounting of calculational uncertainties is not required. However, Framatome does use the PIRT results in its demonstration of overall evaluation model adequacy; in particular, the PIRT results form the basis for determining which phenomena and processes should be validated under the code assessment process. As shown in Table 5-1 of ANP-10332P, Framatome attempted to assess the AURORA-B LOCA evaluation models predictions for each highly ranked phenomenon using one or more validation comparisons.

As addressed in RAI 5, the NRC staffs review of Framatomes PIRT results in Table 4-1 of ANP-10332P identified several areas where additional justification was necessary regarding the general process used to arrive at the PIRT determinations:

• Framatome provided [ ''''''''''' ''''''''''' '' '' ''''''''''''''' '''''''''''''''' ''''''''''''''' '''''''

'''' ''''' ''''''''''''''''''''''' '''''''''''' '' '''''''''''''''''' ' ''''''''''''''''''''] break locations and sizes.

• Framatome applied [' '''''''' '''''''''''''' ' ' ''''''''''''''''' ''''''' ''''''''''' '''''''''' '' ''''''''

'''''''''''''''''''''''''' ''''''''''''''''''' ''''''' '''''''''''''''''''''''''''''''']

Framatome responded to RAI 5 by affirming that the PIRT [''''''''''''''''' ' ''''''''''''''''''''''''''

''''''' ''''''''''''''' ''''''''''''' ''''''''''''' ]. Framatome further stated ['''' '' '''''''''''''''' ''''''''''' ''''''''''

' ''''''''''''''''''' '''''''''' ''''' ''''''''''''' ''''''''''' '''''''''''''' '''''''''''''''''] in the PIRT results presented in ANP-10332P. Additionally, in some cases, Framatome ['''''''''''''' ''''''''' ' '''''''' ''''''''''

'''''''''''''' ' ''''''''''''''''''''' ''''' ''''''''''' '''''''''''' ''''''''''''' ''''''' '''' ''''''''''' ''''''' '''''''''''''].

While the NRC staff considered Framatomes response appropriate, verification of the final, composite PIRT results presented in ANP-10332P was hindered by ['' ''''' ' '''''''''''''''''''''''''''''''''

' ''''' '''''''''''''''''''''''' '''''''''' '' '' ''''''''''''' ''''''''''''''''' ''''''''''' '''' ''''''''''''''' '''''''''' '''''''''''''].

However, based on comparisons of Framatomes results against other relevant PIRT evaluations for the BWR LOCA event, the NRC staff ultimately concluded that, in general, Framatomes PIRT process has satisfactorily considered the range of postulated breaks and relevant plant design features.

That being said, as pointed out in RAI 7, the NRC staffs detailed review of Framatomes PIRT results identified several potential exceptions to the above generalization, wherein the phenomenon rankings determined by Framatome did not appear to have been sufficiently justified. The additional information provided in response to RAI 7 generally addressed the NRC staffs concerns. However, the NRC staff did not agree with several aspects of

Framatomes response, [''''''''''''''''' ''' ''''''''''''''''''''' '''''''''''''''''''''' ' ''' '''' '''''''''''''' '''''''''''''''

'''''''''''''''''''''''''' ' '' ''''''''''' ''' ''''''' ''''''''' '''' ' '''''''' ]. While the NRC staff does not agree with Framatomes rankings for these PIRT items, the NRC staff ultimately concluded that these rankings do not adversely affect the AURORA-B LOCA evaluation model. In particular, as discussed further below, the NRC staff reviewed Framatomes PIRT results only to the degree required to assure adequate validation of the conservative Appendix-K-based method under review. Considering available evidence that the AURORA-B LOCA evaluation model contains reasonable modeling capability for [''''''''''''''''''''''''], as well as the demonstrated conservatism of the AURORA-B LOCA evaluation model (as discussed further below in Section 3.6), the NRC staff concluded that the concerns underlying RAI 7 have been addressed.

Framatome specified in Section 4.6 of ANP-10332P that highly ranked PIRT items must be validated against experimental data. While in agreement with this statement, the NRC staff does not agree in general that validation is superfluous for PIRT items that have not been highly ranked. In particular, Appendix K to 10 CFR 50 prescribes in a number of places that certain phenomenological models shall be compared against adequate experimental data.

Furthermore, Appendix K requires that, to the extent practicable, predictions of the evaluation model, or portions thereof, shall be compared with applicable experimental information.

Although validation of low-ranked PIRT items is unnecessary due to their insignificant impact on the event under consideration, due to the irreducible degree of variability and residual subjectivity inherent in the PIRT process, the NRC staff concluded that the need for validating medium-ranked PIRT items cannot be precluded based on the arguments put forward in ANP-10332P. Therefore, the NRC staff addressed RAI 8 to Framatome regarding this topic.

Framatome responded by stating that the AURORA-B LOCA evaluation model ['''''''''''''''

'''''''''''''''''''' ''''''''''''''''''''''''''''''''' ''''''''''''''''''']. Framatome noted that [''''''' ''''''''''''''''''''''''''''''''' '''''''

'''''''''''''''''''''''''''''''''' ''''' ' '''''''''''''''''''' ''''' '''''''''''' '''''''''''''''''' ''''''''''''''''']. Framatomes response further listed [' '''''''''''''''''''''''''''''''''' '''''''' '''''''' '''''''''''''''] in the evaluation model and provided additional explanation regarding 3 of the 16 phenomena.

From the NRC staffs perspective, while the rigor of assessment may be reduced relative to highly ranked phenomena, appropriate validation of medium-ranked phenomena is important to assure conservative evaluation model predictions. In evaluating the response to RAI 8, the NRC staff focused particularly on several phenomena ranked as medium by Framatome that could arguably have significant influence on the calculated figures of merit:

• Framatome ['''''''''' ' ''''''''''''' ''''''''''' ''''''''' ''''''''''''''''''''' '''' ''' '''''''''''''''''''''''''''' '''

'' '''''''''''''''' '''''''''''''''''''''''''''' ''''' '' ''''''''''''' '''''''''''''''''''' ''''''''''' ' ']

• Framatome [''''''''''' ' ''''''''''''' '''''''''''''' '''''''''''''''''''''' ''''''''''''''''''' '''' ''' '''''''''''''''''''''

''''''''' ''''''''''''' '''''''''''' '' ''''''''''''''''''''' ''''''' ''''' ''''''''''''''''''], the NRC staff concluded that Framatomes overall treatment of these items is appropriate as follows:

o S-RELAP5 uses a point-kinetics model that is implemented in a manner similar to RELAP5/MOD2 and /MOD3, and which has been reviewed by the NRC staff in conjunction with previous S-RELAP5 applications.

o Conservative ['''''''''''''''' '''''''''''''''''''''''' '' ' '''''' ' ''''''''''''''''''' ''''''' ''''''''''''''''''' '''''

''''''''''''''''''' ''''' ''''''''''' '''''''''''''' '''''''''''''''''].

o The time required for ['''''''''' '' '''''''''''''''' '''''''''''''''' ''''''''''' ' '''''''''''''''''''''''

'''''''''''''''''''''''''] from the plant technical specifications.

In considering Framatomes response to RAI 8, the NRC staff concluded ['''' ''''''''''''''''' ''''''

''' '''' '''''''''''''''' ' '''''''''''''''''''''''''''''''''' ''''''''''] are effectively mitigated for the AURORA-B LOCA evaluation model by several factors, including (1) the ['''''''''''''''''''''''''' '''' '''''''''''''''''''

''''''''''''' '''''''''''''''''''''''''], (2) the ['''''''''''''''''''''''''' ''''''''''''''''' '''''''''''''''''' ''''''''' ''''''''''''''''''''''] in Framatomes PIRT (as reflected in the [''''''' '''''''''''''''''''' ''''' '''''''''''''''' ''''''''' ' ''''''''' '],

(3) the NRC staffs previous review of Framatomes methods for modeling certain phenomena

([''''' '''''''''' ''''''''''' ''' '''''''''' ''''''']), and (4) ['''''''''''' ''''''''''''''''' '''''''''''''' ''''''''''' '''''''''''''''''

''''''''''''''''' ''''''' '''''' '''' ']. Therefore, while the NRC staff does not agree with the vendors determination of a medium rank for several phenomena, the mitigation described above ensures that the evaluation model treats such phenomena with [''''''''''''''''''''''

''''''''''''''''''''''''].

In an overall sense, although the level of detail, justification, and documentation associated with Framatomes PIRT effort for the AURORA-B LOCA evaluation model is less than expected of best-estimate methods, the NRC staff ultimately found the information provided adequate to support the assessment and validation of the conservative Appendix K-based method under review. Having reached this determination, the NRC staff ['' ''' '''''''' ''''''''''''''''''''

'''''''''''''''''''''' '''''''''''''' '''''''''''''''''' ''''''''''''''''''''''''''' ''''''''''''] for the following reasons:

• Although phenomenon identification and ranking is a fundamental step in the evaluation model development process, the PIRT results themselves are not part of the evaluation model exercised to perform safety analyses.

• Although the phenomenon rankings reflect an informed opinion of the influence of each phenomenon across a range of plant types and scenarios, there is no expectation that a single set of rankings can precisely represent all postulated LOCA events at all reactors within the defined scope of application.

• Although the phenomenon rankings generated by different PIRT panels for a given event exhibit similar high-level tendencies, at a detailed level, there is an irreducible element of residual subjectivity in assigning phenomenon rankings.

Therefore, the NRC staff reviewed the PIRT rankings only to the extent necessary to support the validation of the models and correlations in the conservative, Appendix K-based AURORA-B

LOCA evaluation model described in ANP-10332P. Outside of this limited purpose, as stated in Limitation and Condition 6 in Section 5.0 of this SE, the PIRT results shown in Table 4-1 of ANP-10332P are not explicitly approved as part of this SE.

3.3 Evaluation Model Development As described in RG 1.203, evaluation model development involves establishing the structure of the supporting code system, temporally and spatially coupling all constituent codes, deriving field equations, selecting compatible closure relationships, applying appropriate numerical solution techniques, and developing control logic and additional supporting capabilities.

Section 6.0 of ANP-10332P provides a description of the development of the AURORA-B LOCA evaluation model using the EMDAP format of RG 1.203.

Section 1.1 of this SE introduced the constituent parts of the AURORA-B code system, of which S-RELAP5 and RODEX4 are germane for the LOCA evaluation model. The basic structure of the AURORA-B code system is illustrated schematically below in Figure 2, which is adapted from the S-RELAP5 code theory manual (Reference 11).

Figure 2: Schematic of AURORA-B Code Structure Previous NRC staff reviews of immediate relevance to the AURORA-B code system and its constituent codes have been noted above in Section 1.2. These previous reviews, as well as additional background information that will be discussed further below regarding the historical

development of the S-RELAP5 code, have already established the fundamental soundness of many aspects of the AURORA-B code system for performing thermal-hydraulic calculations for demonstrating reactor safety. Therefore, while this SE will attempt to touch upon all fundamental aspects of the AURORA-B code system, the main thrust of the NRC staffs present review will be on (1) new or modified parts of the evaluation model and (2) existing models especially significant to the prediction of the figures of merit for a BWR LOCA event.

3.3.1 Overview of S-RELAP5 Models and Correlations As described in Framatomes S-RELAP5 code theory manual (Reference 11), the ancestry of the current version of the S-RELAP5 code in the AURORA-B LOCA evaluation model can be traced back to two versions of the NRC-sponsored RELAP5 code (MOD2 and MOD3). The immediate forerunner to S-RELAP5 was the ANF-RELAP code, which was ['''''''''''''''''''' '''

''''''''''''''''''' '''' '''''''''''''''''''' '''''''''''''' '''''''' ''''''''''''''''''''']. The ANF-RELAP code was based on the [''''''''''''''''''''''''''''''''''' ''''''''' '''' ' '''''''''''''' ''''''''''''''' '''''''''''''''''''''''''' ''''''''''''''''''''']. Siemens Power Corporation subsequently developed the ANF-RELAP code into S-RELAP5, and, around the turn of the century, submitted S-RELAP5-based methods for PWR LOCA and non-LOCA thermal-hydraulic safety analyses (References 21, 22, 23). The effort to develop S-RELAP5 involved a significant number of improvements to the code structure to [''''''''''''''''' '''''''''''''''''''''''''

'''' ''''''''''''''''''''''''''''''']. A number of enhancements to physical modeling capabilities were also made, including the adaptation of some new models from RELAP5/MOD3.

According to ANP-10332P, development of the S-RELAP5 code version employed in AURORA-B began with the version used in the Realistic Large Break LOCA methodology in EMF-2103(P)(A) (Reference 25). Numerous changes were necessary to transform this version of S-RELAP5 into a code capable of performing safety analyses for BWRs under an AURORA-B evaluation model. As described in ANP-10300P, by 2009, Framatome developed a version of S-RELAP5 capable of analyzing BWR non-LOCA events. Significant code modifications associated with this effort are summarized below:

• [''''''''' '''' ' ''''''''' ''''''''''] * ['''''''''' '''' ''' '''''''''''''' ''''' ''''''''']

• ['''''''''' '''''''''' '''''''''''''''''''''''''''''''''''' '''''' * ['''''''''''''''' ''''''''''''''''' '''''' ''''''''']

''''''''''' ''''''''''']

• [''''''''''' ''''''' ''''''''' ''''''''''' '''''''''''''''''''''''] * ['''''''''''''''' '''''''' '''''''''' ''''''''''''''''''' ''''''']

• [''''''''''' '''''''''''''''''' ''''''''' '' ''''''''' ''' * [''''''''''''' ''''''''''''''''''' '''''' ''''' ''''''''

'''''''''''''''''''''''''''''''''''''''' ''''''''''''''] ''''''''''''''' '''''''''''']

The NRC staff has previously found these changes to S-RELAP5 acceptable, as applied to the analysis of anticipated operational occurrences and certain non-LOCA accidents for BWRs (Reference 8). In the present review, the NRC staff has reconsidered these models, where appropriate, to ensure model applicability and assessment adequacy under LOCA conditions.

Further modifications were necessary to support modeling the BWR LOCA event according to Appendix K to 10 CFR 50. Listed below is a selection of relevant changes Framatome has implemented in the [''''''''''''''''] version of S-RELAP5, as discussed in Section 8.2 of ANP-10332P, the S-RELAP5 code theory manual (Reference 11), and Framatomes response to RAI 14 (Reference 5):

1. ['''''''''''' '''''''''''''''''''' '''''''''''''' '''''']

2. [''''''''' '''''''''''''''''''''''' ''''''''''' ''''' '''' ''''''''' '''''''''' ''''''''']

3. ['''''''''' ''''''''''''''' ''''' '''''''''''''' ''''''''''''''''''''' ' '''''''''''''''''''''' ''''''''']

4. ['''''''''''''' ''''''''''''''''''''''' '''''''''''''''''' ''''''''''''''' ''''''' ''''''''''' ''''''''''''''''''''''''''' ' ''''''''''''''''''' ''''

'''' '''''''''' ''''''''''']

5. ['''''''''''''''' '' ''' '''' '''''''''''''' ''''''''''''''''' ''''' '''' ''''''''''''''''' '''''''''''''' '''''' '''' ''

'''''''''''''''' ''''''''''' ' '''''''''''''''''''']

6. [''''''''''' ''''' ' '''' ''' ''''''''''''''' ''''' '''''''''''''''' '''''''''''' ' ''''''''''''''''''''' '''''' ''''''''''''''''''' ]

7. ['''''''''''' ''''''''''''''''''''''' '''''''''' '''''' '''''''''''''''' '''''''''' '''''''''''''''''''''''']

8. [''''''''' ''''''' ''''''''''' '''''''''''''''''''''''' ' ''''''' ''''''''''' ''''''''''' '''''''''''']

9. [''''''''' '''''''''''''''''''''''''''''''''''''''''''' ''''''''''''''''''' '' '''''''''''''''''' ''''' '''''''''''' '''''''''''''''''''''''''' '''''''']

10. [''''''''''' '''''''' ' '''''''''''' '''''''''' '''''''''' ' '' '''''''''''''''''''' ''''' ' '' '''''''''''''''''''''''''''''

''''' '''''''''''''''']

11. [''''''''' ' '''''''''' ' '''''''''''''''''''''''' ' '''''''''''''''''''' ' ''''''''''''''''''' ''''''''''''' '' ''''''''''''''''']

12. [''''''''''''''' '''''''''''''''''' '''' ''''''''''''''' '' '' '''''''''']

13. [''''''''''''''''' ''''''''''''' ''''' ''''''''''''' '''''''''''''''''' ' '''''''' '''''''''''''''''' '''''''''''''' '''''''

'''''''''''''''''''''''''' '''''''''''''']

14. ['''''''''''''''''' ''''''''''''' ''''''''''' ''''' '''''''''' '''''''''''''''''' '''''''' '''' '''''''''''''''''''']

15. ['''''''''''''''' ''''''''''''''''' '''''' ''''''''''''' ''''' ''''' ''''''''''''' '''''''''' ''''''''''''' ''''''''' ' '''''''' '

'''' ''''''''''''''''' ''''' '''''''''''''''''''''''''''''''''''''''''''' ''''''''''''''''''''''''''''']

According to Chapter 15 of the SRP and RG 1.203, an analytical evaluation model submitted for NRC review should be frozen (i.e., placed in a controlled, unchangeable state) during the NRC staffs review period; however, in the audit held on May 16-18, 2017, Framatome informed the NRC staff of its decision to update the S-RELAP5 code version used in the AURORA-B LOCA evaluation model ['''''' ''''''''''''''' ' ''''''''''''''''''] during the review. This decision was motivated

[' ''''' ' '' '''''''''''''''' ' '''''''' '''''''''' ''''''''''''' '''''''''''''''''''' ' ' ''''''''' '''''''''''''

''''''''''''''''' '''''''''']:

16. [''''''''''''''' ''''''' ''''''''''' '''''''' ''''''''' ''''' '''''']

17. [''''''''''''''' '''''''''''''' '''''''''''''''''''' '' ''''' '''''''''''''' ''''''''''''''''' ''' '''' ''' ''''''''''''''' ''''''''

''''''''''''' ''''''''' ''''''''''''''''''''' '''' '''' ''''''']

As alluded to in RAI 28, the NRC staff viewed Framatomes request for approval of the AURORA-B LOCA evaluation model ['''''''' ''' ''''''''''''''''' ''''''''''''' ' '''''''''''''''''''] as a deviation from regulatory guidance, since the validation assessments, demonstration analyses, and supporting descriptions in ANP-10332P are all specific to ['' '''''''''''''''' '''''''''''' ''''''''''''''''''''''].

Therefore, in RAIs 15, 28, 51, and 126, the NRC staff requested that Framatome provide additional descriptions and justification for ['''''''''' '''''''''''''''' ''''''''''''''''''''''' ' ''' '''''''''''''''''' ''''''

''''''''''''' ' ''''' ' ''''''''''''''''' '''''''''''''''''' '''''''''''''''''''''''' ''''' ''''''''''''''''''''''''' ''''''''''''''].

While Framatome was responsive to these requests, the level of detail in the RAI responses was less than that in the original TR descriptions. Consequently, the NRC staff found it necessary to rely upon a mixture of information and results from ['''''' ''' ''''''''''''''''' '''''

'''''''''''''''' ''''''''''''''' ' ''''''''''''''''''''''] in order to complete the review. The impacts of this deviation

from regulatory guidance will be assessed in later sections of this SE covering the evaluation model assessment, sensitivity studies, and demonstration analyses.

The discussion below concerning S-RELAP5 and the RODEX4 kernel is intended to serve as a general review of code modeling practices, with a focus on new models and other models of significance to the LOCA event. While both the S-RELAP5 field equations and closure relations will be discussed below, the NRC staffs review focus was mainly on the latter. As discussed further below, although S-RELAP5 ['''''' '' ''''''' ''''''' '''''''''' ' ''''''''''''''''''''''' ''''''''], the NRC staff expects the basic S-RELAP5 field equations and numerical methods to be sufficiently general to extend to this application. On the other hand, a focus on closure relations is justified in light of their semi-empirical nature and more-restrictive applicability ranges.

3.3.1.1 S-RELAP5 Field Equations The S-RELAP5 code employs non-equilibrium, non-homogeneous, two-fluid field equations for determining the thermal-hydraulic behavior of the liquid and vapor phases present in the reactor system (Reference 11). The two-fluid formulation, which is commonly used in modern thermal-hydraulic system codes for reactor safety analysis, involves conservation equations for mass, momentum, and energy for each phase, which results in a total of six field equations. In addition, S-RELAP5 incorporates field equations for noncondensable gas and boric acid.12 The S-RELAP5 field equations and their numerical solution methods have been subject to NRC staff review on numerous occasions, as attested to above in Section 1.2. Framatomes S-RELAP5 code theory manual further emphasizes the similarity of the formulation of the two-fluid field equations in S-RELAP5 with those in a number of other codes, including RELAP5/MOD2, RELAP5/MOD3, TRAC-PF1/MOD1, and COBRA/TRAC (Reference 11). While acknowledging this point, the NRC staff found that past critiques by the Advisory Committee on Reactor Safeguards of the documentational rigor supporting the derivation of the field equations (References 32 and 33) generally remain unaddressed in the present version of the S-RELAP5 code theory manual (Reference 11). As discussed further below in Section 4.1, the NRC staff found room for improvement in the S-RELAP5 code theory manual; however, considering past approvals discussed above (in particular the NRC staffs 2001 review of EMF-2328, Revision 0 (Reference 34)), additional discussion of this topic in RAI responses in the 2003 review of EMF-2103, Revision 013 (Reference 25), as well as continued indication that the S-RELAP5 field equations are performing adequately over a wide range of PWR and BWR applications, the NRC staff found that the existing evidence firmly supports the formulation of the S-RELAP5 field equations. Section 3.4 of this SE describes the code assessment comparisons presented in ANP-10332P, which offer further evidence of adequacy.

3.3.1.2 S-RELAP5 Closure Relations and Process Models Owing to the temporal and spatial averaging necessary to support derivation of the two-fluid field equations, as well as the relatively coarse noding implemented in thermal-hydraulic system 12 Boric acid tracking is not a part of the AURORA-B LOCA evaluation model, and the S-RELAP5 modeling of this phenomenon was not assessed in the present review. Modeling of noncondensable gas

[' ''''''''''''' '' ''''''''''''''''''''''''' '''''''' '''''''''''''''''''' '''''''''' ''''''''''''''''''' ' ''''''''''''''''' '''''''''

''''''' ''''''''''''''''''''''''].

13 See in particular the responses to RAIs 45-52.

codes such as S-RELAP5, detailed modeling of physical phenomena occurring at the vapor-liquid interface and near walls is not possible. The two-fluid field equations further do not contain models for all physical behavior and processes of relevance to reactor safety analysis.

Hence, many relevant behaviors and processes must be modeled using closure relations and process models that typically have semi-empirical bases and comparatively limited ranges of application. Furthermore, due to the complexity of two-phase flow and the significant differences in physical behavior associated with different flow patterns (e.g., bubbly, slug, annular-mist), different closure relations must typically be defined for each specific flow pattern.

The situation is complicated further by the need to assure smooth transitions between the closure relations associated with different two-phase flow regimes to promote code stability.

Proper selection of closure relations and process models is vital to the accuracy of a codes predictions. As alluded to above, while all currently approved applications of the S-RELAP5 code rely on the same basic field equations, the closure relations and process models are generally application-specific. For example, S-RELAP5 contains multiple models for addressing many phenomena, such as SE, heat transfer in a rod array during core reflood, and choked flow. The analyst is required to select the S-RELAP5 model appropriate to a given application, depending on whether the analysis is to be performed for a BWR or PWR, for a LOCA14 or non-LOCA event, and under an Appendix K or best-estimate evaluation model. As discussed further below in Sections 3.3.4 and 4.1, the analyst is aided in making appropriate selections by a modeling guidelines document (Reference 26) [''''' '''''''''''''''''''' ''''''''''''''].

Given the large number of closure relations and process models used in thermal-hydraulic system codes such as S-RELAP5, the discussion below organizes these relationships into broad categories. A general description of the closure relations and process models in each category is summarized in tabular form. Additional discussion is provided in supplementary text where appropriate. Detailed review is generally not provided where previous, applicable reviews have already addressed particular code models. Alternative S-RELAP5 code options to those defined by the AURORA-B LOCA evaluation model are omitted from discussion.

14 For PWRs, a further distinction exists between the modeling options approved for small- and large-break LOCA analysis.

3.3.1.2.1 Hydrodynamic Closure Relations and Process Models Model S-RELAP5 Model for BWR LOCA NRC Staff Assessment

['''''''''''''''''''''' '''''' ''''''''''' '''''''' '''' '''''''''''''''

General approach consistent Flow map ''''''' ''''''''''''''' '''''''''' ''''''''''''''''''' ' ''''''''''''

with previous reviews.

'''''''''''' '''''''''''''''' ''''''''''''''''''''''' ''''''''''''''''' ]

['''''''''''''''''''' '''''''' '''''''''''''''''''''''''''''''''''''''' General approach consistent

'''''''''''''''''''''' '''''''''' ''''''''''' '''' '' ''''''''' ' with previous reviews. See

'''''''' ''''''' ''''''''''''''''' '''''''''''''''' '''''''''' Section 3.3.1.2.2 for discussion Interphase

'''''''''''''''' ''''''''' ''''''''''' ''''''''' ''''''''''' concerning modification to Friction

''''''' '''''' '''' ''''''''''''' '''''' '''''' '''''''''''''' [''''''' ''''''''''''' ''''''''''' '''''''' '

''''''' ''''''''''''''''' '''''''''''''''''''''' '''''''''''' '''''' '''''], which is of

''''''''''''''' ''' ] significance to LOCA event.

[''''''''''''''''''''' ''''' ''''''''''''''''''''''''''''''''''''''''''''

Interphase

'''''''''' '''''''' ''''' ''''''''''''''' ''''''''''''' '''''''''''' Discussed below.

Heat Transfer

''''''''''''''''''' '''''''''']

['''''''''''''''' ''''''''''' '''''''''''''''''''''''''''''''' '''' ''''

General approach consistent Wall Friction ''''''''''''''''''''''''''''' ' ''''''''''''''''''' ''''''''''''''''

with previous staff reviews.

''''''''''''''' ' '''''''''''''' ''''''''''''''']

Use of Moody model for two-phase conditions conforms to Moody critical flow model used for two-phase Appendix K requirements.

break flow.

Choked Flow Homogeneous equilibrium Homogeneous equilibrium model used for model commonly used for vapor single-phase vapor flow.

flow.15 Discussed further below and in Section 3.3.4.1.3.

General approach relies on

['''''''''''''' '''''''' ' '''''' ''''''''''''' commonly accepted modeling Countercurrent

''''''''''''''''''''''''''''''''''''' ''''''''''''''''''''''''''' '''''''''''''' '''' practices. Discussed Flow Limitation

'''''' '''''''''''''''''] subsequently in Section 3.3.4.2.2.16 15 Note that this model is essentially [''''''''''''''''''' '''''''''''''''''''' ''' '''''''' '''''''''''''''''''''''].

16 As noted above, while not directly associated ['''''' ''' ''''''''''''''''''''''''''' ''''' '''''''''''''''''' '''''''''''''''''''' '''''

''''''''''''''''''' ''''''''''''''''' '''''''''''''''' ' ''''''''''''''''' '''''''''''''''''''''''''' ''''''''''''''' ''''''''''''''''' ''''' ''''''''''''']

Interphase Heat Transfer Several minor changes relevant to the BWR LOCA event have been made with respect to modeling interphase heat transfer.17 These changes are discussed further in Section 3.4.10 of the S-RELAP5 code theory manual (Reference 11):

• A simplified model was added to ['''''''''''''''' ' '''''''''''' ' '' ''''''' ''''' '''''' '''''

'''''''''' ''''''''''''' ''''''' ''''''''''''''''''''''' ''''''']

• A spray [''''''''''''' '''''''''''''''''''''''''' ''''''''' '''' '''''''''' '''''''''''' '' '' ''''''''''''''''''' ''''''''''''''

'''''''' ''''''''' ''''''''''''''''''''].

• Framatome implemented changes to augment interphase heat transfer [' ''''''''''''''

'''''''''''''''''''''' '''''' '' '''''''''''''''''''' '''''''''''' ''''''''''''' ' '''''''''''''''''''']

The impacts of these changes will be considered in the assessment and validation effort described in Section 3.4 of this SE. In particular, the ['''''''' '''''''''''''' ''''''''''''''''''' '' '

''''''''''''''''''''' ''''''''''''' ' ''' '''''''''' ''''''''''''''' ''''''''''''''''''''''' ''''''''''''''''' '''''''''].

Choked Flow As required by Appendix K, the Moody critical flow model is applied at potential choking planes where two-phase flow may occur; in particular, [''''''' '''''''''''''''' ''''''''''' '' '''''''''' '''''''''''''' ''''

' '''''''' ''''''']. In response to RAI 25, Framatome confirmed that the ['''''''''' ''''''''' '' '

'''''''''''''' ''' ''''''''' '''''''''''] for all recirculation line breaks (i.e., both discharge and suction breaks). However, at potential choking planes where two-phase flow is not expected, for example the single-phase-vapor choked flow that may occur through safety/relief valves, particularly during small-break LOCA scenarios, Framatome ['' ''' ''' '''''''''''''''''''''''''''''''

'''''''''''''''''''''''''''' ''''']

17 Note that the original impetus for some of these changes ['''''' ' ''''''''''''''''''''''' ''''' ''''''''''''''''' ''''''''

'''''''''''''''''''' ''''''''''].

3.3.1.2.2 Heat Transfer and Heat Structure Models Model S-RELAP5 Model for BWR LOCA NRC Staff Assessment Critical Heat Flux 2006 Groeneveld CHF Lookup Table, Discussed below.

with eight correction factors.

Transition Boiling McDonough, Milich, King correlation Acceptable per Appendix K.

Minimum Stable Groeneveld-Stewart correlation Discussed below.

Film Boiling Temperature (Tmin)

Film Boiling - Heat Modified Bromley correlation, with Bromley correlation in common Transfer to Liquid attenuation adjusted as explained in use. Attenuation factor is response to RAI 15. discussed further below.

Vapor Convection McEligot correlation (heated walls) Discussed below.

Sleicher-Rouse correlation (unheated walls).

Drucker-Dhir two-phase turbulent enhancement (dry-wall form) applied in dispersed flow film boiling regime.

Radiation Heat Sun, Gonzales, and Tien model for Discussed below.

Transfer wall-to-fluid.

RELAP5/MOD3.3 model for wall-to-wall.

Condensation Wall condensation modeling Generally consistent with previous generally similar to previous reviews. reviews. Spray droplet Spray droplet condensation model in condensation model was discussed upper plenum. above.

Heat Conduction Heat structure modeling is highly Consistent with previous reviews.

consistent with RELAP5/MOD2 and

/MOD3.

Reflood [ ''''''''''''''''''''''''''''''''' '''''''''''''''''''' ['' ''''''''''''''''''''' ''''''''''' ''''''''''''''

Solution/Quench '''''''''''' '''' '''''''''''''''''''' ''''''''''''''''' '''''''' '''''''''' '''''''''''''''''' '' '''''''''''

Front Modeling ''''''''''''''' '' ''''''' '''''''''''''''''''' '''''''''' ''''' ''''''''''''''''' '''' '' '''''''''''''''''''

''''''''''' '''''''''' ''''''' ''' '''''''''''''' '' '''''''''''''''']

'''''''''''']

The overall heat transfer logic used for the AURORA-B LOCA evaluation model is provided in Figure 6-16 of ANP-10332P, which was reproduced directly from the S-RELAP5 code theory manual. The basic heat transfer logic therein has been previously reviewed and accepted for other applications, and the NRC staff likewise finds it acceptable for the AURORA-B LOCA evaluation model.18 18 Note, however, that Framatome implemented [ ''''''''''''' ' ''' '''''''''''''''' ''''''''''''' '''''''''''''' ''

''''''''' '''''''''''' ''''''''''''].

Critical Heat Flux As discussed in Section 6.4.10 of ANP-10332P, the AURORA-B LOCA evaluation model uses the 2006 version of the Groeneveld CHF lookup table to compute the CHF according to the thermal-hydraulic conditions determined by the S-RELAP5 code at each simulated fuel rod node. Use of this lookup table method in lieu of fuel-specific correlations is necessary for the LOCA event to permit consideration of a wide range of thermal-hydraulic conditions that extends well beyond those typically considered by fuel-specific correlations that are primarily intended for normal operation.

Framatome intends [ '''''' ''' ''''''' ''''''''''''''''' ' '' '''''''''''''''''''''''' '''''''''' ''''''' ''''

''''''''''''''''''''''''''' '''']

In RAI 16, the NRC staff requested additional information concerning Framatomes implementation of the 2006 Groeneveld CHF look-up table, including

• justification for the fuel-specific correction factors Framatome proposes to apply to the Groeneveld CHF lookup table, and example values for a specific fuel design,

• clarification of a possible error identified in ANP-10332P and its Reference 43, and

• explanation of the process by which the AURORA-B LOCA evaluation model computes the inputs to the CHF lookup table and [''''''''''''''''''''' '' '' '''''''''''''''''' ''''''''''''''''

'''''''''''''''''''''' '''''''''''''''].

Framatomes response to RAI 16 explained that the fuel-specific CHF correction factors were derived from Groeneveld (Reference 46) and Lee (Reference 47). The response further provided correction factor values for the ATRIUM 10 fuel design as a function of time during one of the demonstration case analyses included in Section 7.7 of ANP-10332P. Framatomes response also acknowledged that Reference 43 to ANP-10332P is not self-consistent. Finally, Framatomes response attempted to clarify the process used to compute the conditions for determining the CHF. [' ''''''''''''''''''''' '''''' '''''''''''''''''''''''' ''''''''''''''''''' '''' '''''' ''''''''' '''' ''

'''''''''''''''''''' ''''''' '' ''''''''' '''''''''''''''''''' '''' ''''''' ]

Framatome selected fuel-specific CHF correction factors from different sources. In some cases, Framatome did not offer a clear rationale for doing so. Thus, the NRC staff found it challenging to assess the combined impact of the different correction factors on the accuracy of the modified

output from the lookup table, and additional considerations were necessary. The NRC staff considered

• the ['''''''''''''''''''''''''' ''''''''''''''' '''''''''''''' '''''' ''''' '''' ' '''''''''' ] the CHF lookup table output,

• the individual justifications for certain correction factors provided in the supporting references, including References 46 and 47, and

• the common usage of a number of the correction factors [' '''''''' ''''''''''''' '''''''' '''''''''

'''''''' ''''''''''''''''''''''''''''''''''''' '''''''''''''''''''''' ''' ''''' ''''''''''''''''''''''''''' '''''''''''''''''' ''''].

Based on these additional considerations, the NRC staff generally found the correction factors proposed by Framatome acceptable.

However, an exception to this general statement is that the NRC staff did not agree with the correction factor proposed for [''''''''''''''''''''''''' ''''' '''''' '''' ''''' ''''' ''''' '''''''' '''''''' ''''''''''

'''''' '''' '''''''''''''''''''] As a result, the NRC staff concluded ['''' '''' '''''''''''''''''''''' '''''' ''''' '''

'''''''''''''''''' ''''''' ''''''' ''''] may not be credited when determining CHF values according to the

['''''''''''''''''''''' ''''''' '''''''''' '''''']. This position is documented as Limitation and Condition 7 in Section 5.0 of this SE. In addition, the NRC staff concluded [''''' ''' '''''''' '''' '''''''''''''''''''''''

''''''' ' '''''''''''''''''' '''''''']. In its updated RAI response (Reference 66), Framatome [''''''''''''''

''' ' ''' '' '' ''''''''' ''' '''''''''''''''''''' ''''''' '''''''' ' '''''] for analyses performed with the AURORA-B LOCA evaluation model.

The NRC staffs review also indicated that the acknowledged inconsistency in Reference 43 to ANP-10332P can most readily be explained by interpreting the document as showing a nonconservative prediction (i.e., overprediction of the CHF) by ['' '''''''''''''''''''''' '''''' '''''''''''''

'''''''' ''''''''' ''''''''''' '''''''''''''''' '' ''''''' '].19 Consequently, the NRC staff concluded that the statement in ANP-10332P ['''' '''''''''''''''''''''''''''' ''''''''''' '''''''' ''''''''''''''''''''''''''' ''' '''''''''''' ''''''''''

'' '''''''''''''''''''''''' ] is quite likely in error. However, it should be noted that ['''''''''''''''''''''''''''''''''''''

'''''''''''''''''''''''''' '''''''''''''''''' '' ''''''' ''''''''''''''''' '''''''''''''''''''''''''] Furthermore, Framatome provided additional data in its response to RAI 16 showing [ '''''''''''''' ''''''''''''''''''''' ''''''''''' ''' '''''''''''''''''''''

''''''''''''].

19 Note that the phrase [''''''''' '''''''''''''''''''''' ''''''''' '''''''''''''''''''''' ''''''''''''''''' ''' ''''''''''''''''' '' ''''''

'''''''''''''''''''''' '''''''''''''' ''' ''''''''''''' ''''' '''''''''' ' ''''''''' ''].

In its updated response to RAI 16 (Reference 66), Framatome provided additional information concerning an error in [''' '''''''''''''' ''''''' '''''' '''''''''''''' ''''''''''] comparisons discussed below in Section 3.4.3.7. Framatome stated that the error in ANP-10332P [''''''''''''''''' '

'''''''''''''''''' ' '''''''''''']. Nevertheless, in light of the scaling and applicability issues discussed further in Section 3.4.3.7, the information in Framatomes updated response to RAI 16

['''''''''''''''''''''' ''' '''''''''''''' '''''''' '''''' ''''''''''''''''''''''] did not substantially alter the NRC staffs judgment concerning ['' '''''''''''''''''''''''''''''''''' '' '''''' '''''''''' ''''''' ''''''''''''''''''']

Finally, with regard [ ''''''''''''''' '' ''''''''''''''''''''''' '''''' ''' '''''''''''''''''' ''''''' ''''''''''''''''''' '''

''''''''''''''''' ''''''' '''' ''''''''''''''''' '''''''''' '''' ''''''''''''' '''''' '''']

[''''' ''''''''''''''' '''''''' ' ''''''' '''''''''''''''' '''''''''' ''' ''''''''' ''''''' '''' ''''''

'''' '''''''''''''' ''' ''''''''''''' '''''''''' ''''''''''''' ''''''''''''' '''''''''''' ''''''' '''''''''''20 ''''

'''''''''''''' '''''''''' '' ''''''''''''' ''' '''''''' ''''''''''''''''''' ''''''''''''''''']

and

[''''''' '''''' ''''''''''''' ''''''''''''''''' ''''''''''''' '' '' '''''' ' ''''' '''''''''''''''''' '''''''''' ''

''''''''''''''' ''''' '''''''''''' '' ''''''''''' ' '''''''''''''''''''''' '''''''''''''''''''''''''''''''''''''''']

While a definitive determination could not be made based upon the descriptions contained in Framatomes response to RAI 16, ANP-10332P, and the S-RELAP5 code theory manual (Reference 11), considering all of the relevant information discussed above, it appears that the AURORA-B LOCA evaluation model [''''' '' '''''''' '''''''''''''''''''''' '''''''''''''' ''''''' '''''''''''''' '

''''''''''''''''''' ''''''''''''' '''''''']

20 Note that the [''''''''''''''''' ''''''''' '' ''''''''''' '''''''''''' ''''' ' ''''''''''''''' ' '''' '' '''

''''''''''''''''''''''''''' '''''''''''''''' '''''''''''''''''''''' ''''''''''''' '''']

The NRC staff finally notes that ['''' '''''''''''''''''' '''''''''''''''' ''''''''' '''' '''''''''''''' ' ''''''''''''''''''' ''''

'''''''''''' ' ' '''''''''''''''''''''''' ' '' ''''''''''''''''''''''' ''''''''''' '']

Framatomes updated response to RAI 16 dated October 31, 2018 (Reference 66), included a modest extension of its discussion concerning ['''' '''''''''''''''''''' ''''''''''''''''' ''' ''''''' '''''''''''''' ''''

'''''''''''''''''''''''''''''''] However, the NRC staffs review of the updated material did not fundamentally alter the judgment rendered above.

In conclusion, while ultimately agreeing with Framatomes approach for computing CHF, the NRC staffs review of Framatomes response to RAI 16 found that [''' '''''''''' '''''''''''''''''''''''''''''''''

''' '''''''''''''''' ' '''' '''''' '' ''''' ''''' ''''''''''''''' ''''''''''''''''' '''''' ' ''''''''''' ''''''''''''''''''']. In particular, this determination contributed to the NRC staffs decision to impose Limitation and Condition 7.

Minimum Stable Film Boiling Temperature As discussed in Section 6.4.12 of ANP-10332P, the AURORA-B LOCA evaluation model determines the minimum stable film boiling temperature according to the Groeneveld-Stewart correlation, subject to the restriction on rewetting imposed by Appendix K. Namely, during the blowdown phase of the LOCA event, Appendix K effectively establishes the minimum stable film boiling temperature as 300 °F greater than the saturation temperature. Should the cladding temperature exceed this threshold, Appendix K forbids reestablishment of transition boiling for the remainder of the blowdown phase. Hence, the AURORA-B LOCA evaluation model does not use the Groeneveld-Stewart correlation under such circumstances to justify cladding rewet.

The NRC staff identified several areas where additional information was necessary to complete its review and addressed RAIs 19 and 20 to the topics of (1) assessment of the Groeneveld-Stewart correlation in the specific pressure range of interest to the BWR LOCA event post-blowdown, (2) the difference in behavior at low pressure observed in the two datasets Framatome used to assess the Groeneveld-Stewart correlation (i.e., Groeneveld and Stewart, Winfrith), and (3) the apparent variation in accuracy of predictions of the Groeneveld-Stewart correlation, depending upon the degree of liquid subcooling.

Framatome responded to these RAIs by providing a plot to illustrate the performance of the Groeneveld-Stewart correlation against data in the range of interest for the BWR LOCA event.

Framatomes responses further discussed several differences in procedure and setup for the tests Framatome used to assess the performance of the Groeneveld-Stewart correlation that could have contributed to differences in behavior between the datasets. Framatome further emphasized the conservatism of the Groeneveld-Stewart correlation, noting in particular how it was developed from testing with Inconel 600 as opposed to Zircaloy or zirconium oxide, for which Peterson and Bajorek observed increased Tmin values (Reference 39). In view of this

demonstrated conservatism, the NRC staff considered Framatomes responses to the first two issues addressed in RAIs 19 and 20 satisfactory.

Regarding the third item, the NRC staff questioned whether the Groeneveld-Stewart correlation would overpredict Tmin for subcooled conditions at temperatures of interest to the LOCA event (e.g., 400-480 °C/750-900 °F, in Figure 6-22 of ANP-10332P). Framatome performed a sensitivity study to illustrate the influence of subcooling on the predictions of the correlation.

The study considered two sensitivity cases, [''''''' '''''''''' ''''''''''''''''''''' '''''''' ' ' ''''''''

''' ''' ''' ''''''''''' '''''''' '' '''''''''''''''''' ''''''''''''''''' '' ''''''' ']. The first case included both the saturated and subcooled terms of the Groeneveld-Stewart correlation, and the second case biased the calculated Tmin downward by neglecting the contribution from the subcooled term.

The results of the sensitivity calculations ['''''''''''''''''' '''''''''''' ''''''''''''''' ' ''' '''''''''''' ''''''''

' '''''''' '''''''' ''''''''''''''''' '''''''''''''''''''' ' ''''''''''' ''''' '''''], the NRC staff found Framatomes response acceptable.

While the NRC staff considered the above issues to have been acceptably addressed, during its review of the sensitivity study described in Framatomes response to RAI 20, the NRC staff

['''''''''''''''' '''''' '''''' ''''''''''''''''' '''''''''''''''''''''''''' '''' '''''''''''''''''' '''''''''''''''''' ' ''''''''' '

'''''''' ''''''''' ''''''''''' '''''''''''' '' '''''''''' ''''''''''''''''' ] As a means to probe the issue further, the NRC staff assessed results from independent calculations with the NRCs thermal-hydraulic code, TRACE, for small- and large-break BWR LOCA scenarios. The calculation of Tmin in TRACE [''''' '''''' ''' ''''''''''''''''''''''''''''''''''''''''' ''''''''''''''''''''], but further applies a minimum value of 725 Kelvin (K) (i.e., approximately 845 °F) within 15 centimeters of a quench front21 (Reference 44). The Tmin values predicted by TRACE for the hot node in both the small- and large-break cases at the time of quenching were found to be 845 °F. [''''''''''''' '' '''''''''''''''''''''

''''''''''''''''''''''' '''''''''''''''''' ''''''''' '' '''''''''''''' ], the NRC staff remained concerned with the inconsistencies in the reported predictions of S-RELAP5 that were not adequately explained in any of the documents reviewed or audited by the NRC staff.

21 The use of [''' ' ' ' ''''''''''''''''' ''''''''''''''''''' ''''''''' ''''''' '''''''' ' ''''''''''''''''''''' ''''''''''''' '''''] is stated to be an empirical correction intended to improve agreement with test data.

In its updated RAI response dated October 31, 2018 (Reference 66), Framatome provided additional information regarding the proper interpretation of the relevant S-RELAP5 output. The updated response to RAI 20 identified the apparent discrepancy as an effect of [''' '''''''''''''''''''''

''''''''''''''' ''''' ''''' ' '''''''''''''''''' ' ''''''''''''''''''''''' '''''''''''] The NRC staff considered the information in Framatomes updated response to RAI 20 as providing a definitive interpretation of the relevant S-RELAP5 output. However, as noted further below in Section 4.1, the NRC staff concluded that enhanced discussion in supporting code documentation regarding the proper interpretation of the S-RELAP5 output ['' ''''' ''''''''''''''''''' '''''''''' '' ''''''''''''''''

'''''''''''''''''''''''''''' ''''''''''''''' '''' '''''''' '''''''''''] may help avoid future misunderstanding.

Film Boiling Stable film boiling, which is typically associated with the inverted annular flow pattern, is defined by Framatome as occurring at void fractions less than 0.6. Dispersed flow film boiling is defined as occurring at void fractions between 0.9 and 1.0. No heat transfer to liquid is credited in the dispersed flow film boiling region. [ ''''''''''''' '' ''''''''' '''' ''''''''''''''''''' ''''''''''''' ''' ''''

''''' ' '''''''''''''''''''' ''''' ''''''''''''''''''' '''''' '''''''''''].

Framatome calculates film boiling heat transfer to liquid using the modified Bromley correlation.

In S-RELAP5, the modified Bromley correlation is formulated using the Taylor characteristic length, as opposed to the Helmholtz criterion used in some other applications. Framatome noted that the Taylor characteristic length was also implemented in RELAP5. Figure 6-24 in ANP-10332P illustrates minor differences in the heat transfer coefficients calculated using the two criteria, but ultimately concludes that both criteria have been adequately validated by test data. The NRC staff found the proposed approach reasonable, and notes that the adequacy of Framatomes post-dryout heat transfer assessments will be evaluated below in Section 3.4.

The modified Bromley correlation has a long history of wide application for computing film boiling heat transfer to liquid. [''''''''''''''''''' ' '''''''''''''''''' ' ''''''''''''''''' ''''' '' '''''''''''''''''''''''''

''''''''''''''''''''''''''''''' '''''' ''''''''''''''''''''''' ''''] The NRC staff found Framatomes approach reasonable; in particular, as discussed subsequently, the [''''''''''''''] version of S-RELAP5 had

[''''''''''' ''''''''''''''''''' '' '''''' ''''''''''''''' ' ''''''''''''' ''''''''''' ' '''''''''''' ' ''''''''''''''''''''''

''''''''' '''''''''''''''' '' '''''''''''''''' ''' '''''''''''''''''''''''''' ''''''' ' '''''''''''''' '''' '''''''''''''''''''' '''''''' ']

As discussed further in response to RAIs 15 and 51, during the NRC staffs review of ANP-10332P, Framatome [''''''''''''''' '' ''''''''' '''''''''''' ''''''''''' '''''''' '''''''' ''''''''''' ' ''''' ''''

'''''''''''''''''''''''''' '''''' ''''''''''''''' '''''''''''''''''' ''''''''''' '''''''''''''''''''''' '''''''''''''''''''''''''' '''''' ''''''''''] Per Section 3.2 of the S-RELAP5 code theory manual (Reference 11), ['''''''''''''''''' '''''''

'''''''''''''''''''''' ''' '''''''''' '''''''''' '''''''']. While the NRC staff found Framatomes ['''''''''''''''''''''

''' '''''''''' ''''''''' '''''''''] rather empirical, it appeared to contribute to improved agreement with experimental results and to result in a slight increase in the conservatism of the predictions.

[''''''''''''''''''''''''''''' ''''' '''''''' '''''''''''''''''''' ''''''' ' ''' '''''''''''''' '' ''''''''''''''''''' ''''''''''''''''''''''' '''''

''''''''''''' ''''''''''''' ''''''''''' '''''''' '''''''''']

Framatome attempted to resolve this information gap in its updated response to RAI 15, dated October 31, 2018 (Reference 66). Framatomes updated response to RAI 15 provided

['''''''''''''''''''' '''''''''''''''''' '''''''''''' '''' ''''''''''''''' '' ''''' '''''''''''''''' ' '' '''''''''' '''''''''' ''''''' '

'''''''''' '''''''''' ''''''''' ' ''''' ''''' '''''' ''''''].

Vapor Convection The [''''''''''''''' ''''''''''''''''' ''''''' '''''''''''''' '''''' ''''''''''''' ''''''' '''' ''''' ' '''''''''' ''''''''''''''''''''

' '' ''''''''''''''''''''''''' ''''''''''' '''''''''''' '''' '''''''''''''''''''' ''''' '''' ''''''''''' '''''']. In the dispersed flow film boiling regime, [' ''''''''''''''''' ' ''''' ''''''''''''' ' '' '''''' ''''''''''''''' ''''''''''''''''''' ' '''

''''''''''''''''' ]. Framatomes rationale [f' '''''''''''''''' ''' ''''''''''''''''' ' ''''''''''''' '' '''''''''''''''''''

''''' ''''''''''''' ' '''''''' '''''' ''''''' '' '''''''''''''''''''''''''''''''''''''' '''''' '''''''''' '''''''''''], as discussed in Section 7.3.11 of ANP-10332P. Furthermore, a two-phase turbulent heat transfer enhancement is added in the dispersed flow film boiling regime based on the formulation by Drucker and Dhir.

When the [''''''''''''''''' '''''''''''''''''' ''''''''''''' ' '' '''''''''''''''''''''''''''' ''''''' '''''''''''''''''''' '''''' ''''''''''''

''''''''''''''''' '''''' '''''''''''' '''''''''''''''''''''''''''''''' ''].

Regarding the use of the Drucker-Dhir correlation, in RAI 124, the NRC staff ['''''''''''''''''''' '''''

'''''''''''''' '''''''''''''''''''''']. The NRC staff further requested that Framatome clarify the

implementation of this correlation, ['''''''' '''''' '' ''''''''''''' ' ''''''''''''''' ''''''''' '' ''''' ''''''''''''

'''''''''''''''''''''''' ''''''']. Framatome responded to RAI 124 by stating [''' '' ''''''''''''''''' '''''''''' '

' ''''''''''' '' ''' '] Subject to further discussion of the film boiling/steam cooling transition criterion below in Section 3.6.2.3, the NRC staff generally found Framatomes response to RAI 124 reasonable.

Radiation Heat Transfer The basic S-RELAP5 code models for calculating wall-to-fluid and wall-to-wall radiation heat transfer are similar to those the NRC staff reviewed for previously approved PWR Realistic Large Break LOCA evaluation models (References 25 and 54). However, as a result of heat sinks afforded by the channeled fuel bundles surrounded by a bypass region, as well as the incorporation of water rods in many modern fuel designs, radiation heat transfer generally plays a more significant role in the BWR LOCA event.

S-RELAP5 models wall-to-fluid radiation heat transfer based on a model developed by Sun, Gonzales, and Tien (Reference 11). In this model, the vapor-droplet mixture is treated as an optically thin medium, and the wall, droplets, and vapor are regarded as three nodes in an electrical network analogy. Wall-to-fluid radiation is ['''''''''''''''''''' ''''''''''''''''''''''''''' '''''''''' '''' '

'' ''''''''' ''''' '' '''''' ''''''''''' ''''''''' '''''' '''''']. The general modeling approach for wall-to-fluid radiation heat transfer was included in previously approved Revisions 0 and 3 of the PWR Realistic Large Break LOCA evaluation model (References 25 and 54).

Wall-to-wall radiation is based upon a simplified view-factor approach that considers whether a line of sight or reflection path exists between the two surfaces of interest. Framatome noted that the S-RELAP5 modeling approach for wall-to-wall radiation is based upon that of RELAP5/MOD3.3. The NRC staffs review found that the code theory manual description regarding wall-to-wall radiation heat transfer in S-RELAP5 is essentially identical to those of both RELAP5/MOD3.3 (Reference 48) and RELAP5-3D (Reference 49). The S-RELAP5 modeling approach for wall-to-wall radiation heat transfer was previously reviewed as part of Revision 3 of the approved PWR Realistic Large Break LOCA evaluation model (Reference 54).

Based on its review, the NRC staff obtained confidence in the basic S-RELAP5 code models used to determine radiation heat transfer. However, as appropriate for an Appendix-K-based evaluation model, Framatomes implementation of radiation heat transfer modeling in the AURORA-B LOCA evaluation model incorporates a number of conservatisms. These conservatisms include the [''''''''''''' '' '''''''''''' ''' ''''''''''''''' ' ''''''''''''''''''''' '''''' '''''''''''' '''''

'''''''' '''''''''''''' ''''''''' ''''' ''''''''''' '' ''' ''''''''''''' '''''] discussed below in Section 3.3.4.2.1.

Reflood Solution The NRC staff reviewed ['''' ''''''''''''''''''''''' ''''''''' ' '''''''''''''''''''''' '''''''''''''''' '' ''''''''''''' ''''''''''''''

'''''' ''''''''''''''''' '''''''''''' ''' '''''''''''''''].

Regarding the ['''''''''''''''''''''''''' ''''''''' ''' '''''''''''''' ''''''' '''''''''' ''' ''''''''''''''''''' ''''''''''' '''''''

''' ] However, for the AURORA-B LOCA evaluation model, Framatome ['''''''''''''''

'''''''''']. The NRC staff considers this change appropriate, ['''''''' ''''''''''''' '''''' ' ' ''''''''

''''''''''''''''''' ''''''''''''''' '''''''''']

Regarding ['''''''''''''''''' ''''''''''''''''' '' ''''' '''''''''''''''''''' '' '' ''''''''' ''''''''''''''''' '''' '''''''''''' '''''

'''''''''''''''''''''' '''''''''' '''' ''''''''''''''''''''''''''' ' '''''''''' '' '' '''''''''''''''' '''''''' ]. Framatome attempted to resolve this information gap in its updated response to RAI 15, dated October 31, 2018 (Reference 66). Framatomes updated response to RAI 15 provided ['''''''''''''''' '''''''''''''''''''

'''''''''''''''''' ''''''''' '''''''' '''''''''''''''''] as discussed further in Section 3.6.3, the NRC staff imposed Limitation and Condition 25, [''''''''' ''''''''''''''''' '''''' '''''''''''''''''''''''''' '''''''''''''''''' '' ''

''''''''''''''''].

3.3.1.2.3 Component and Control Logic Models Model S-RELAP5 Model for BWR LOCA NRC Staff Assessment Centrifugal Similar to RELAP5/MOD2 and /MOD3. Consistent with previous reviews.

Pump EPRI two-phase degradation model.

Jet Pump Modeled as streamtube using coupled Generally consistent with previous one-dimensional equations. Model reviews with minor modifications.

based on steady-flow energy equation with inclusion of loss and pump terms.

Control and Trip Similar to RELAP5/MOD2 and /MOD3. Consistent with previous reviews.

Logic Jet Pumps Prior to incorporation into the AURORA-B code system, S-RELAP5 contained a jet mixer model that could be used for BWR jet pumps. However, to support AURORA-B applications, Framatome [''''''''''' ''''''''''''''''' '''''''' '''''''''' '''' '''' '''''''''''''''' ' '' ''''''' ''''' ' ''

''''''''''''''''''''''''''' '''''''''''''''] Furthermore, Framatome performed additional validation comparisons, which the NRC staff has reviewed below in Section 3.4.2.2.

3.3.1.2.4 Reactor Core Power Models Model S-RELAP5 Model for BWR LOCA NRC Staff Assessment Point Kinetics Similar implementation to Generally consistent with previous RELAP5/MOD2 and /MOD3. reviews, discussed below.

Decay Heat Fission product decay heat from Discussed subsequently in proposed American Nuclear Society Section 3.3.4.1.1.

Standard 5.1, October 1971, with multiplier of 1.2.

Actinide decay heat from American Nuclear Society Standard 5.1, August 1979.

['''' ' '' '''''''''''''' '''''''' '''''

''''''''''''']

Metal-Water Baker-Just equation As per Appendix K, Baker-Just Reaction correlation is applied, and reaction assumed not steam limited.

Discussed below.

Point Kinetics The AURORA-B LOCA evaluation model relies upon the point kinetics model to determine fission heat. Use of point kinetics modeling is typical for the LOCA event, and the NRC staff considers this approach appropriate since (1) postulated LOCA events are not predominately governed by coupled neutronic-thermal-hydraulic behavior and (2) the spatial dependence of neutronic behavior is not significant.

Framatome stated that approved lattice physics/core simulator methods (e.g., CASMO-4/MICROBURN-B2 (Reference 40)) will be used to provide input to S-RELAP5 to support implementation of the point kinetics model for fission power as well as the determination of decay heat. Regarding the latter, although, as required, decay heat from fission products is determined according to the proposed American Nuclear Society Standard 5.1 from October 1971, Appendix K does not prescribe a method for determining decay heat from actinides. As an input to determining the decay heat contribution from actinides, Framatome would calculate the capture-to-fission ratio using approved lattice physics/core simulator methods.

In response to RAI 44, Framatome summarized inputs necessary to support the point kinetics model implemented in S-RELAP5. The response further stated that bounding neutronic parameters have been established that result in ['''''''''''' '''''''''' ''''' '''''''''''''''' ''''' '''''''''''''''''''

'''''''''''''''''''' '''' '' ''''''''' ''''''' '''''''''''''' '' '''''''''' ''''' '''''''''' ' '''''''' '''''''''''''''' '''''''''] The

topic of licensee verification of the appropriateness of plant-specific analysis parameters is further addressed at a general level in Limitation and Condition 22. Namely, [ ''''''''''''' ''

''''''''''' ' '''''''''''''' '''''''''''''''''''''''''' ''''''''' ''''''''''''''' ' ''''' ''''''''''''''''''''''' ''''''''''''''''''''']

Metal-Water Reaction During the NRC staffs review of the AURORA-B LOCA evaluation model, Framatome informed the NRC staff of an error identified in S-RELAP5, wherein the computation of the percentage of cladding thickness oxidized by the metal-water reaction did not account for cladding swelling and rupture.22 In RAI 114, the NRC staff requested that Framatome describe the error, discuss its implications regarding the adequacy of the validation and continuity of assessment procedures for the AURORA-B LOCA evaluation model, and discuss the necessity of documentation updates (e.g., analyses in ANP-10332P, S-RELAP5 code theory manual).

Framatome responded that the error in S-RELAP5 involved not accounting for strain-induced changes in cladding geometry when calculating oxidation. By contrast, 10 CFR 50.46(b)(2) requires the determination of the local oxidation thickness to account for the thinning of the cladding wall that occurs as the cladding swells. Furthermore, according to Section I.B of Appendix K, the computation of oxidation must account for the expanded surface area that results when the cladding swells. Framatomes response indicated that the S-RELAP5 error was originally associated with the UDEC93 code version, which well predates development of the AURORA-B code system. In response to the NRC staffs concern that lack of specificity in the S-RELAP5 code theory manual may have contributed to the persistence of the error, Framatome included in its response to RAI 114 an excerpt containing code theory manual modifications intended to clarify the corrected methodology for computing cladding oxidation.

The NRC staff considered the inclusion of additional detail in the code theory manual appropriate to promote adequate understanding of the functionality of the S-RELAP5 model.

However, the NRC staffs review of Framatomes response to RAI 114 identified issues with the revised code theory manual equations used to model the metal-water reaction. These issues included apparently inconsistent definitions of variables ['' ' ''''' '''''''''''''''' '''''''''' ''

''''''''''''''''] and suspected errors in Equations 7.261 and 7.262. In particular, Equation 7.262 in Framatomes original RAI response appeared algebraically inconsistent with the formulations described in the theory manuals of TRACE (Reference 44) and RELAP5-3D (Reference 49).

In its updated response to RAI 114 dated October 31, 2018 (Reference 66), Framatome provided corrections to Equations 7.261 and 7.262 that were intended to address the errors identified by the NRC staff. The NRC staffs review of the updated response found the errors in the original response to RAI 114 to have been substantially corrected. However, the NRC staff made the following further observations concerning the updated response to RAI 114:

22 Although the ['''''''''''''''''''' '''''''' '''''' '''' ''''''''' ''''''''' ''''''''''''''''''' ''''''''''''''''''''' ''''''''''''' ''''''''''''''' '

] ] the proposed AURORA-B LOCA evaluation model.

• The issue identified above concerning inconsistent definitions of variables [' ' ''''

''''''''''''' ] remains. For example, excerpted text from the code theory manual contained in the updated response to RAI 114 appears to characterize Equation 7.261 incorrectly ['' ''''''''' ''''' '''''''''''''''''' ''''''' '''''''''''''''''''''' '' ''''''' '''''''''''''''''''' ''''''''

'''''''''''''''''' '' '''''' ''''''''' '''''''''''''''''' ''''''''' '''''''''''' ''''''''''''''' ''''''''''''''' ''''']

• A typographical error remains in the updated versions of Equations 7.262a and 7.262b

['''' '''''''' '' ' ''''''''''''' ''''''''''''''''' ' ''''' ' '''''''''''''' ''''''''''' ''''' '''''''' '''''''''

''''''''''''''' '''''''''''''''''' '' ''''''''''''''' '''' '''''''''']

• Usage of variables with parallel nomenclature to represent different concepts persists.

[''' '''''''''''' ' ''''' '''''''' ''''''''''''' ''''' ''''' '''''''''''' ' ''' '''''' ' ''''''''''''''' ''''''

''''''''''''' ''''''''''' ''''''''''''' ']

Despite lingering issues with code documentation, the NRC staff considered Framatomes updated response to RAI 114 to reflect substantially accurate modeling of cladding oxidation during a LOCA event where cladding deformation may occur. The NRC staffs confidence in this conclusion was bolstered by a point comparison the staff performed using the TRACE code that predicted values of local cladding oxidation comparable to the corrected version of the S-RELAP5 code. The NRC staffs confirmatory calculation is described in Section 3.6.2.5 of this SE.

Framatomes response to RAI 114 further noted [''''' '' ''''' ''''''''''''''''''''''' '''''''''''''''''' '

''''''''''''''''' ''''''''''' '' '''''''''''''''''''''' '''' ''''''''' ' '''''''' '''' '''''''''''''' '''''''''''''''' ''''' '''] In its updated response to RAI 114 dated October 31, 2018 (Reference 66), Framatome stated that future plant-specific calculations using the AURORA-B LOCA evaluation model will be performed with a version of S-RELAP5 that correctly accounts for strain-induced changes in geometry when calculating cladding oxidation.

3.3.1.2.5 Fuel Rod Transient Behavior Model S-RELAP5 Model for BWR LOCA NRC Staff Assessment Fuel Rod Transient RODEX4 kernel Consistent with RODEX4 stand-Thermal- alone code, as discussed Mechanical Model subsequently in Section 3.3.2.

Fuel Rod Swelling Model derived from NUREG-0630 Discussed subsequently in and Rupture (Reference 29). Section 3.3.4.1.2.

Fuel Relocation Not Included Discussed below.

Fuel Relocation ANP-10332P discusses the potential for fuel pellets exposed to high burnup to experience fragmentation, with the resulting fragments having the potential to relocate during the LOCA event into space opened up within a swelled region of fuel cladding. Framatome [''''''''''''''' '''''

''''''''''''''''''' '''''''''''''''''''''''''''''''''''''''' '' ''''''''''' ''' ''''''''''''''' '''' '' '''''''''' ''''''''' '''''].23 During its initial review of ANP-10332P, the NRC staff identified questions [''''''''''''''''''''''' '''''''''''''''''''''''''''''

' '' '''''''''''''''''''' ' ''' ''''''''''''''''''''' ''''''''''' ''''''''''''''''' '''''''''''']

Framatome responded to RAI 10 by ['''''''''''''''''' '' '''''''''' ''''''''''''''' ''''''''''' '''''''''''''' '

'''' '''' '''''''''''''''''' ''''''''' '''' '' '' ' ''''''''''''''''''''' ''''''' ''']

On the strength of this demonstration, as well as other conservatisms inherent in the methodology, the NRC staff agrees that Framatome has provided sufficient information to demonstrate that ['''' ''''''''''''''''' '''' '''''''''''''''' ' ''' ''''''''''''''''''] to assure that the existing criteria of 10 CFR 50.46 are satisfied for the Appendix K-based AURORA-B LOCA evaluation model. ['''' '''''' '''''' '''''''''' '''''''' '''' '' ''''''''''''''''' '''''''' '''''''''' ''''''''''''''' ''''''''''''''''''''''''''''''

'''''''''''''' ''''''''''']

23 A reference containing further information concerning these facilities and fuel relocation phenomena in general is NUREG-2121 (Reference 56).

3.3.1.3 S-RELAP5 Numerical Methods and Code Stability A semi-implicit numerical solution scheme is used to solve the hydrodynamic field equations, which is similar to that used in RELAP5/MOD2 and /MOD3. However, as noted above, S-RELAP5 uses algebraic manipulation as opposed to the Gaussian elimination approach for solving the finite-difference equations. The NRC staffs review found the general approach used in S-RELAP5 for the AURORA-B LOCA evaluation model to be consistent with that evaluated in previous applicable reviews, one of which is documented in Section 4.2 of the NRC staffs evaluation of EMF-2328, Revision 0 (Reference 34).

The NRC staffs review of ANP-10332P generally did not find concerns with code stability.

However, discussions during the May 16-18, 2017, audit did identify the potential for premature termination of some cases, ['''''''''''''''''' '''''' '' ''''''''''' '''' ''''''''''''''''''''''''''''' ''' ''''' ''

''''''''''''' '''''''''' '''''''''''''''' ''''''' ''''' ''''''''''''''''''] These issues, along with a non-representative practice Framatome employed to mitigate them, are discussed further in connection with RAIs 30 and 31 in Section 3.3.5.7, below, which outlines the NRC staffs expectations regarding acceptable analysis termination criteria.

3.3.2 Overview of RODEX4 Kernel Models and Correlations The S-RELAP5 code incorporates several fuel rod modeling options, including RODEX2, RODEX3A, COPERNIC, and RODEX4; selection of the proper option is evaluation-model specific. As noted above, the AURORA-B LOCA evaluation model requires selection of the RODEX4 kernel.

Framatomes general strategy for modeling fuel behavior across various S-RELAP5-based evaluation models, [''''''''' '''' ''' '''''''''' '''' ''''''''''''''''' '''''''''''''''''''''''''''''''''''''''''''' '''''''''''' '

'''''''''''''''''' '''''''''''''' '''''' '''''''''' '''''''''' ''' '''''' '''''''''''''''''''''']

According to the S-RELAP5 code theory manual (Reference 11), [''' '''''''''''''''''''''' '''''''''''''''

'''''''''''''''''''''''' ' '''''''''''''''''''''''''''''''''''''''' ''''''''''''''''''''''''' '''''''''''''''''' ''' ''''''''''''''''' ''''''''''] of the RODEX4 kernel include

• ['''''''''''''' '''''''''''''''''' ''''''' '''' '''''''''''''''']

• [''''''''' ''''''''''''''''''''''''' ''''''''''''''']

• ['''' ''''''''''''''' ''''']

• ['''' ''''''' ''''' ''''''''''''''''''''''''']

• ['''' ''''' ''''''''''''''''' ''''''''''' ''''''''''''''''''''''' '''' ''''' '''''''''''''''']

The [''''''''''''''''''' ''''''''''''''' '''''''''''''''''''''''''''''''''''''''' ''''''''''''''''''''''''' '''' '''' ''''''''''''''' ''''' '' ''''''''''''''

''''''''''] of the RODEX4 kernel include

• [''''''''' '''''''''''''''''''''''' '''' '''''''''''''''''''''''' '''''''''''''''' '''' '''''''']

• [''''''''''''''' '''''''''''''''''''' ''''' '''''''''''''' '''''''''' ''''''''''''''''''' '''' '''''''''']

• ['''''''''''' ''''' '''''''''''''' ''''' ''''' ''' ''''''''''''']

• [''''''''''''''' '''''''''''''' '''''''''''''''''']24 From its review of the S-RELAP5 code theory manual, the NRC staff ['''''''''''''''''''''' '''' '''

'''''''''''''''''''' ' ' ''''''''''''' ''''''''''''''''''' ''''' '' ''' ''''''''''' ' ''' ''''''''''''''' ''''''], which was previously approved by the NRC staff for use in accordance with the methodology described in BAW-10247P (Reference 27). In response to RAI 14.c, Framatome confirmed this understanding, noting that ['' '''''''''' '''''''''''''''' ''''''''''''''' ''''''' ''' ''''''''''''''' '''''''''' ''''

''''''''''''' '' '''''''''''''''''''''''' ''''''''''''''' '''''']. Framatomes response further clarified that, since the time of its approval by the NRC staff, no changes have been implemented in RODEX4 that would affect the fuel rod physical models used for the AURORA-B LOCA evaluation model.

In RAI 11, the NRC staff requested that Framatome clarify how the fuel pellet thermal conductivity degradation issue discussed in Information Notice 2009-23 was addressed by the models used in RODEX4 and the broader AURORA-B code system. Framatome responded that the RODEX4 code [''''''''''''''''''' ''''' '''''''''''''' ''''''''''''''''''''''' ''''''' ''''''''''''''' '''''''''''''''''''''

'' ''''''' ''''' ''''''''''''''''''''' '''' ''' '''''''''''''''' ''''''''' ''''''' ''''''''''''' '''''''''''] and concluded that the RODEX4 modeling of thermal conductivity was acceptable.

In RAIs 13.b and 13.c, the NRC staff requested that Framatome address how the calculation of effective gap conductance in the AURORA-B LOCA evaluation model accounts for the effects of dimensional changes of the fuel and cladding, as well as non-uniformity in the fuel-cladding gap in the axial and azimuthal directions. Framatome responded [''''' ''' ''''' ''''''''''''''''''''''''

'''' ''''''''''' '''''''''' '' ''''''''''''''''''' '' ' ''''''''''''''''''''''''] The NRC staff found these aspects of the gap-conductance calculation reasonable.

Section 6.3.8 of ANP-10332P states that Framatome intends to [''''''''''''' '' '' ''' '''' ''''''

'''''''''''''''' ''''''''''' ''' ]. However, as clarified in Information Notice 98-29, when assessing compliance with the regulatory limit for maximum local oxidation, any oxidation during the operating cycle prior to the LOCA event must be included. Inasmuch as ANP-10332P does not describe how Framatome will determine the pre-transient oxidation, the NRC staff issued 24 As noted above, [''''''''''''''''''' '''''''''''''''' ' ''''''''''''' '''''''''''''''''''' ''' ''' ' '''' ''' '''''''''''''''' ''''''''].

RAI 42 to request that Framatome describe and justify its methods for determining pre-transient oxidation and clarify whether this method is considered a part of the AURORA-B methodology.

Framatome responded to RAI 42 [ ''' ''''''''''''''''''' ''''' '' '''''''''''''''''' ' '''''''''''''''''''''''

''''''''' ''''''''''''' '''''' '''''''''''''''' ''''''''''' ''''''''' ''' '''''''''''''''''' '''''''' ' ''''''''''''''''''''''''' ''''''''''''''''''25

''''''''''''''''''''' '' ''''''''''''''' ''''''' ''''''''''''''' '''''''''''''''' ' '''' '''''''''''''''''''' ''''''''''''''''''' '' ']

The NRC staff considers Framatomes practices for modeling pre-transient oxidation acceptable. ['''' '''''''''''''''' ''''''''''' ''' ''''''''''''''''''' ''''''' '''''''''''''''' '' ''''''''''''''''''''''''' ''''''''''''''''''

''''''''''''''''''''' '''''''''''''''' ''''''''''''''''''''''''' '' ''''''''''''''''' ''''''''''''''''''''] In responding to RAI 42, Framatome further committed to updating the approved version of the TR to reflect its actual practice regarding the modeling of pre-event oxidation.

3.3.3 Coupling of S-RELAP5 and RODEX4 Kernel The basic coupling scheme employed in the AURORA-B LOCA evaluation model has been previously reviewed by the NRC staff in conjunction with AURORA-B-based methods in ANP-10300P (Reference 7) and ANP-10333P (Reference 9). Furthermore, the practice of coupling thermal-hydraulic and fuel thermal-mechanical codes has long been associated with S-RELAP5. For example, a similar coupling scheme to that shown above in Figure 2 was used in the original version of S-RELAP5 submitted for review in 2000, wherein transient models from RODEX2 were brought into S-RELAP5 for the analysis of PWR events.

As illustrated in Section 3.3.5 of the NRC staffs SE on ANP-10300P (Reference 8), the S-RELAP5 thermal-hydraulic module and the RODEX4 transient fuel-performance kernel within the AURORA-B code system need not use identical timestep sizes. The NRC staffs review of ANP-10332P identified that the [''''''''''' ''''''''''''' ''''''''''''' '''''''''''''' '''''''' '''''''''''''''''] of AURORA-B was not defined. The NRC staff requested that Framatome clarify this issue in RAI 13.a. Framatome responded that the AURORA-B LOCA evaluation model [''' ''''

''''''''''''''''''''''''''''''''''' '''''''''']. The NRC staff considered this response acceptable because it

['''''''''''''''''''''' '''''''' ''''''''''''''''''''''' '' '' '''''''''''''''' ''' '''''''''''''''''' ''''''''' '''''' '''''''''''''''

''''''''''''''''''''''' ' ''' '''''' '''''' '' '''''''' ' '''''' '''''''''''''''''''''''''''''' ''''''''].

25 Framatomes process for performing plant-specific calculations, ['''''''' '''''''''''''' '''''''''' '''''

'''''''''''''''''''''''''''' '''''''''''''''''''''''' ''' ''''''''''''''' '], is discussed further below in Section 3.3.5.1.

3.3.4 Modeling Options and Nodalization Section 6.3.7 of ANP-10332P describes the basic modeling and nodalization practices that Framatome has proposed to use with the AURORA-B LOCA evaluation model. Framatome stated that many modeling practices used for the AURORA-B LOCA evaluation model derive from those used for the AURORA-B AOO evaluation model. However, Framatome has made a number of modeling and nodalization changes to support simulation of the LOCA event.

Analysis of the LOCA event brings into consideration a number of different phenomena that do not arise in the course of normal operations and anticipated transients. The basic S-RELAP5 code models for simulating such phenomena have already been discussed (e.g., post-dryout heat transfer, metal-water reaction, and reflood/quench front modeling). The discussion below concentrates upon modeling practices necessary to satisfy Appendix K to 10 CFR 50 and, more broadly, the evaluation model described in ANP-10332P.

The NRC staffs review observed that the S-RELAP5 code has numerous options for modeling various physical processes, which may be activated or deactivated by the analyst. The NRC staff identified that ANP-10332P does not identify the full set of code modeling options that must be activated when S-RELAP5 is used in support of the AURORA-B LOCA evaluation model.

Neither are all the required modeling options fully specificed in the S-RELAP5 code theory and users manuals (References 11 and 12), ['''''''''' ''''''''''''''''' ''''''''''''''''''''''' ' '''''''''''''''''''''''''''''''''

''''''' ''''' ''''''' ''''''''''''''''' ''''''''''''''''' ''''''''''' ''''''''''''''''' ''''''''''' ''''''''''''''' ] Therefore, the NRC staff issued RAI 27 requesting, in part, that Framatome identify the full set of necessary code modeling options for S-RELAP5 simulations performed under the AURORA-B LOCA evaluation model.

Framatome responded to RAI 27 by [''''' '''''''''' '' ''''''' ''''''''''''''''''' ''' '''''' ''''''' '''''''''''''''

'''''''' ''''''''''''''' '''' '' ''''''''''''''' ''''''''''' ' '''' ' '''' '''''''''''''''''' ''''' ''''''''''''''],26 the NRC staff could not conclude [''''' '' '''''''' ''''''' '' '' '''''' ''''''''''''''''''''' '''''''''''''' '''' ''''''''''''''' ' '''''''

'''''''''''''''' '''''''''''''''''''''''''''''' ' '''' '''''''''''''''''''''''' ''''''''''''''''' ' '' '''''' ''''''''' '''''''''] As discussed further in Section 4.3, substantial changes to the methodology or results of the evaluation model must be avoided to assure that the basis for the NRC staffs conclusions regarding the acceptability of the AURORA-B LOCA evaluation model is preserved.

3.3.4.1 Modeling Options Relevant to Appendix K Requirements Sections 3.3.1 and 3.3.2 above discuss significant closure models in the S-RELAP5 code and RODEX4 kernel. In some cases, as has already been discussed, these closure models directly address Appendix K requirements. However, in other cases, the requirements of Appendix K 26 Note that, in some cases, the NRC staffs agreement is contingent upon the satisfaction of limitations and conditions specified in Section 5.0 of this SE.

focus on input choices that do not directly concern code models and correlations. This section of the SE is intended to address selected input choices of particular relevance to the conformance of the AURORA-B LOCA evaluation model to required and acceptable features specified in Appendix K to 10 CFR 50. A summary of the conformance of the AURORA-B LOCA evaluation model to the full set of Appendix K requirements is provided in tabular form in Section 6.2.1 of this SE.

3.3.4.1.1 Heat Sources Appendix K defines the heat sources to be considered for BWRs during a LOCA event as the initial stored energy in the fuel, fission heat, decay heat from fission products and actinides, the metal-water reaction from cladding oxidation, and heat transfer from reactor internals. Key issues examined during the NRC staffs review of LOCA heat sources are described below.

Axial Power Profiles Framatome proposed that the AURORA-B LOCA evaluation model [''''''' ''''''''''''''''' ''''' ''''

''''''''''''''' '''' ''''''''''''''''''''' '''' ''''''''''''''''''''' '''''''''''' ''''''' '' ''''''''''''''''''''''' '''''''' ' ']

Furthermore, during the May 2017 audit, ['''''''''''''''''''''' ''''''''' '' ''''''''''''''' ' '' ''''''''''''''''''''''''

'''''''''''''''''''''''''' ' ' '''''''' ''''''' '] within the scope of the AURORA-B LOCA evaluation model.

Framatome responded [''''' '''''''' '' ''''''''''''''''' '''''''''''''''''''' ' '' ''''''' ' ''''''' '''''' ' ''

''''''''''''' '''''' '''' ''' '''''''''' ''''''''''''' ''''''''''' ''''''' ''''''''' '''''''''''''''']. Framatomes response to RAI 38 [''''''''''' ' '''''''''''''''''''' ''''''''''' '''''''''''''''' ''''''''''''''''''' ''''''''''' '' ''''''''''''''''' '''''''''''''''' '''''

'''''' ' '''''''''''''''' '''''''''''''''''' '''''''''''''''] Framatomes response further ['''''' '''''''''''''''''''' ''''''''

''''''''''''''''''' ''''''''''''''''''' ''''''''' '''''''''' ''''''''''''''''' '''''' ''' '''''''''''''''''''''' ' '' '''''''' '''''''''''''' '']

The NRC staff ['''''''' '' '''''''''''' ''''''''''''' '''''' ' ''''''''''''''''''''''''''''' ''''''''''''''''''' ''''' '

''''''''''''''''''''''''''''' ''''''''''''''''''''''' '''''' '' '''''''''''''''' ''''''''''''''''' '''''''''''''''' ' ''''''''''''''''''' 27 '''' '''''

'''''''''''''''''''''' ''''''''''''''''''''''']

Meanwhile, as this information came to light, the NRC staff [''''''''' ' '''''''''' '''''''''''''' ''''''''''''''''

'''''''''''] In response to RAI 111, Framatome stated [''''' '''' '''' '''''''''' ''''' ''''''''''

''''''''''''' ''''''''' '''''''''''' ''''''' '''' ''''''' ' '''''''''''''' '' '''''''''''''''''''].28 In particular, [''

'''''''' '''' ''''''''''' '''''''' '''' '''''''''''''''''''''''''' ''''''''''' ].29 Framatome stated that, [''''''''

'''''''''''''' '''''''''''''''''''' '''''''''''' '''''''''''''' ''' ''''''''''''''''''''''''''''' ''''''''']. Framatomes response to RAI 111 provided results for a BWR/4 demonstration case that was re-run using the [''''''''''''''']

version of S-RELAP5. For the sample case considered in the response, ['' ''''''' ''''''''''''''''

''''''''''''''''''' '''''''' ''''' '''''''''''' ''''''''''''''' '' ''''''' ''''''''''''''' ''''''''''''''''''''' '''' '']

Based upon the discussion above, the NRC staff ['''''''''''''''''' '''' ''''''''''''''' '''''''' '' '''''''''''''''''''''

''''''''''''''''''''''']. However, due to concerns remaining with the validation of the AURORA-B LOCA evaluation model that will be discussed subsequently in further detail, particularly those associated with (1) the parallel channel flow phenomenon, (2) downflowing quench fronts, and (3) insufficient agreement in certain integral effects testing comparisons ['''''''' ''''''''''''''''''''''''

''''''''''''''''''''''' ''''''''''''''''''''''''' '''''''''''''''' ''''''' '''''''''''''''''''' '''''''''''' ''''' ''''''''''''], the NRC staff lacked assurance in the capability of the AURORA-B LOCA evaluation model to simulate hot-channel downflow in a reasonable manner. Furthermore, in several important assessment and sensitivity analyses [''''''''''''''''' '''''''''''' '''''''''''''''''''' '''''''''''''''''''''' '' '''''''''''''''''' '''''''''''''''' '

''''''' ' ''''''' '''''''''''''] This position is documented as Limitation and Condition 8 in Section 5.0 of this SE. ['''''' '''' '''''''''''''''''''' ' '''''''''' '' '''''''''''''''''''''''''''' ' ''''''''''''''' '''''' ''''''''''''''

27 See further discussion on this topic in Section 3.6.

28 This modeling practice is discussed further below in Section 3.3.5.7.

29 Whereas, the recirculation discharge isolation valves in the BWR/3-4 design are signaled to fully close post-LOCA (1) in the intact recirculation loop for plants with loop-select logic, and (2) in both recirculation loops for plants without loop-select logic.

''''''''''''''''''' ''''''''''''' ''''''''''''' '''''''''''''''''' '' ''30 '''''''''''''''''''''''''''' ''''''''''''''''''' ''''' '''''''''''''''''''

''''''''''''''''''''''''' '''' ' ''' '''''''' ']

A final note regarding axial peaking factor elevation is that the NRC staff makes no conclusion regarding whether a value of ['' ''''' '''''''''] is the most limiting top-peaked core power profile for all BWRs at present or in the future. As specified in Limitation and Condition 22, it remains the responsibility of each licensee to ensure that its safety analyses envelop allowable operating conditions, and furthermore, that its operations remain within analyzed conditions.

In response to RAI 29.k, Framatome further [''''''''' '''' '' ''''''''''''''''''''''''' '''''''''' '''''''''''''''''''

''''''''''''''''''''''''''''''''''' '''''''''' ''''''''''' ''''''''' '''''''''''''''''''' ''''''''''''''''''''''''''''' '''''''''''''''''' ''' ] In this regard, the NRC staff observed that 10 CFR 50.36 requires the establishment of a technical specification limiting condition for operation for

[a] process variable, design feature, or operating restriction that is an initial condition of a design basis accident or transient analysis that either assumes the failure of or presents a challenge to the integrity of a fission product barrier.

Whereas, [''''''' ''''''''''''''' '''''''''] limits do not typically exist in BWR technical specifications.

[''''''' ''''''' ''' '' ' ''''''''' '''''''''''''' '''''''' '''' '''''''''''''''''' ''''''''''''''''''' ''''' '' '''''''''''''''''''''

''''''''' '''''''''''''''''''' ''''''''' ' ''' '''''''''''''''''''''''''' ''''''' ' ''''''''''''''' ''''''''''''''''''''''''], Framatomes response to RAI 29.k did not recognize that such credit may require plant-specific implementation of a technical specification limiting condition for operation, along with one or more surveillance requirements to verify satisfaction of the limit. Framatomes updated response to RAI 29.k dated October 31, 2018 (Reference 66), further suggested an alternative of [''''''''''''''' ''''''''''''''''' '''''''''''' '''''''''''''''''''' ' '''''''''''''''''' ''''''''''''''''''''''''''''' '''''' '''''''''''''''''''''''

'' '''''''''''' ' '''''''''''' '''''''' '''''''''''' ''''']. However, Framatome did not provide adequate rationale that this alternative approach would satisfy 10 CFR 50.36, ['''''' ''''''''''''' '''' ''''''''''''''

''''''''''''''' '''''''''''''''''''''' ''' '''''''''''''''''' ''''''' ' ' '''''''''''''' ' '''''''' ''''''''''''' '''''''''] Hence, the NRC staff imposed Limitation and Condition 9 in Section 5.0 of this SE, stipulating that analyses with the AURORA-B LOCA evaluation model may not credit a limit on ['''''''' ''''''''''''''

''''''''''], absent a plant-specific determination from the NRC staff that such credit is consistent with 10 CFR 50.36. Lacking such a determination, [''' ''''''' '''''''''''''' '''''''''' '' '' ''' ''''''''''

'''''''''''''''' ' '''''''''''' '''''''' ''''''''''''''''''''''']

Fuel Stored Energy Input for determining fuel stored energy was obtained ['' ''''''''''''''''''''' '''''''''''''''''' '''''''''''''''''''''

30 Although note that Limitation and Condition 23 concerns, in part, ['''' ''''''''''''''''' '''''''''' '''''' '''''''''''

' ''' '''''' ''''''''''''''' '' ''' '''''' ''''''''''''' ''''''''''''''''''''''''''']

'''''''''''''''''''''' ']

The conservative ['''''''''' ''' '''''''''' '''''''''''' '''' ''''''''' ''''''''' ''''''''''''''''''' ''''' ' ''''''''''''''

''''''''''''''''''''''' ''' '''''''''' '''''''''''''']. However, once the LOCA [''''''''''''''' ' ''''''''''''' '''

'''''''''''''' '''''''''''''''''''' ' ''''''''''''''''' ' ' ''''''''''''''' ''''']. In RAI 13.d, the NRC staff requested that Framatome discuss the impacts of this variation in thermal conductivity on the timing of the heat transfer to the fuel cladding and its ultimate effect on the conservatism of the calculated results.

Framatome responded [' '''''''''''''''''''''' '''''''''''''''''''' '''''' ''''' ''''''''''' ' ''''''''''''''''' '''''''

'''''''''''''' '''''''''''''''''''' '] The NRC staff concurred [''''' '''' ''''''''''''' '''''''''''''''''''' '''''''''''''''

''''''''''''''''''''''''' '''''''''']

In its updated response to RAI 13.d, dated October 31, 2018 (Reference 66), Framatome clarified that [''' '''''' '''''''''''''''' ' '' ''''''''''''''''''' '''''''''''''''''' ''''' '''' ' '''''''''''''''''''''' ''''''''''''''''

''''''''''''''''''''''' ''''''''' ''''''''''''''''''' '''''''''''''] Therefore, the NRC staff found Framatomes response to RAI 13.d (as updated) to be acceptable. However, in accordance with Limitation and Condition 22 in Section 5.0 of this SE, licensees remain responsible for assuring the applicability of all plant-specific inputs used in their analyses.

Making a conservative determination of fuel stored energy also requires consideration of the limiting power history. In Section 6.4.2 of ANP-10332P, Framatome cited ['''''''''''''''''''''

'''''''''''''''''''' ' '''''''' ' ''''''' ''''' ''''''' ' '' '''''''''''''''' '''''''] as part of the basis for [''

'''''''''''''''''''''''] of its power history determination. In RAI 101, the NRC staff questioned the applicability [ ''''''''' ''''''''''''''''''''''''] to the analysis of modern BWR fuel and core designs ['''''

''' '''''''''''''''' ''''''']. The NRC staff particularly emphasized the potential for the recent reconsideration of thermal conductivity degradation to impact the determination of the limiting power history. [''''''''''''''''''''''' ''''''''''''''''''' ''''' ''''''''' '''''''''' '''''''''''''''''''''''' '''' ''''''''''''''''' ''''

'''''''''''''''''''''''] Framatome stated that the ['''''''''' '''''''''' ' '''''' ''' ''''''''''' ' '' '''''''' ''''''''''''

'' ''''''''''''''''' ''''''' ''''''''''''''' '' ''''''''''''' ''''''''''''' ' ''''''''' ''''''''' '''''''] Framatome then stated

that [''''''''''' '''''''''''''' ' '''''''''''''''''''''''' ''''''''''' '' ''''''''''''''''' ''''''''''''' '''''''' ''''''''' '''''''''''''''''''

''' ''''''''''''' ' '''''''''''''''' ''''' ' '''' '''''''''' '''''''''' ''''''''''].

The NRC staffs review found that, while the objective of using a bounding power history is acceptable, Framatomes RAI response did not completely justify that the process is capable of ensuring the selection of a bounding power history. In particular, it is not clear how new, limiting power histories are generated, and when they are considered necessary [ '''' '''

' ' '''''''''''' '''''''''''''''''''''' ''''''''''''''' ''''' ''''''''' '' '''''''''''''' '''''''''''''' ''''] Furthermore, in Section 6.2.1 of ANP-10332P, Framatome [''''''''''' ''''' ''''''''''''' ''''''''''''''' ''''' ''' '''''''''''''''''

''' '''''''' '''''' ''''''' ''''' '' '' '' '''''''''''']

In its updated response to RAI 46 dated October 31, 2018 (Reference 66), Framatome provided additional clarification regarding the [''''''''''''''''''''''' ''''''''''''''''''' '''''''''' '''''''''''''''' '''''''''''''''''''''

''''''' '''''''''' '' '' '''''''' '''''' ''''''''''''''''' ''''''''''''' '''''''''' ''''''''''''''' '' ''' ]

The NRC staff found that the information provided in response to RAI 46, as updated in the submittal dated October 31, 2018, adequately addresses the NRC staffs concerns regarding the determination of the limiting fuel rod stored energy. [ ''''''''''''''''''' ''''''''''''''''''''''''' '''''''''''''''''

''''''''''''''''''' '' ''''''''''' ''''''''''''' ''''''''''']

Decay Heat As already noted, Framatomes approach conforms to the proposed 1971 American Nuclear Society 5.1 standard for fission product decay heat with a multiplier of 1.2, as stipulated in Appendix K. Framatome accounts for actinide decay heat using a model ['''''''''''''''''''' ''''' ''''

'''''' ''''''''''''''''' '''''''''''' '''''''''''''' '' ''''''''''''''']. While the proposed 1971 American Nuclear Society 5.1 standard called for consideration of actinide decay heat, it did not specify a particular calculational method. As such, the NRC staff finds [''''' ''' '''''''' ''''''''''''''''

''''''''''''''' '''''''''''''' '' ''''''''''''''''' ''''''''''''''''''''] and acceptable for this purpose.

According to Paragraph I.A.4 under Appendix K, some fraction of the gamma energy from radioactive decay is permitted to be deposited outside the fuel, if justified by a suitable calculation. Section 6.2.5 of ANP-10332P states that the assumed distribution of gamma energy between the fuel and coolant in S-RELAP5 will be determined by Framatomes BWR fuel management reactor physics code used for reload licensing analysis. The NRC staffs audit

of several demonstration case decks identified ['''' '' '''''''''''''' '' ''''''''''' '''''''''''''''''''' '

'''' '''''' '''' ''''''''''''''' '' ''''''''''''''''''''' '''''''''' ' '' ''']. The NRC staff requested in RAI 23 that Framatome provide justification for its approach, particularly under LOCA conditions where the core may be completely voided.

In response to RAI 23, Framatome stated [''''' '' ''''''''''''''''''''''''' ''''' '' ''''''''''''' ' '''''''' '''''''

' '''''''''''''''''''' '''''''''''' ' '' ''''' '' '''''''''''''' '''''' ' ''''''''''''''' '''''''''''''''''] performed using the CASMO-4 lattice physics code. Framatome stated that its calculations ['''''' '' ''''''''''''

'''''' '' '''''''''''''' '''''''''' '''' '''' '] for ATRIUM 10 and ATRIUM 11 fuel.

The NRC staff found Framatomes response to RAI 23 reasonable overall; however, one weakness was that ['' ''''''''''''''''''' '' '''''''''''''' '''''''''''' ''''''''''''''' ''''' '''''''''''''' ' '''''''''''''

'''''''''''''''''' '''''''''''''''''' ' ''' '''''''''''''''' ''''''''''' '''''' '''''''''''''' ''''' '''''''''''''''''' '''''''''''''''']

In its updated response to RAI 23 dated October 31, 2018 (Reference 66), Framatome provided additional ['''''''''''''''' '''''''''' ' '''''''' '''' ''' '''' ''''''''''''''''' '''''''''''''' ''''''''''''''''''''''''' ''''''''''''''''

'''''''''''''' ''''''''''] However, to ensure adequate conservatism in future plant-specific analyses, absent NRC staff approval for higher values, Limitation and Condition 10 restricts credit for gamma energy deposition outside of a fuel rod to [''' ''''''''''''' ''' ''''''''''''''''''''' '''''''''' ''''''']

3.3.4.1.2 Fuel and Cladding Performance Framatome noted in ANP-10332P that swelling and rupture of fuel rod cladding may occur when cladding temperatures are high and the rod internal pressure exceeds the pressure in the coolant channel. Hence, consideration of cladding swelling and rupture is necessary during a LOCA event. Sections 6.2.9 and 6.4.5 of ANP-10332P state that the fuel rod swelling and rupture models implemented in S-RELAP5 are a modified version of those established in NUREG-0630 (Reference 29). Framatome stated that its implementation of these swelling and rupture models was originally approved by the NRC staff in an Exxon Nuclear Company TR from November 1982 (XN-NF-82-07(P)(A), Revision 1 (Reference 30)), and that these models are currently in use with Framatomes existing Appendix K-based LOCA methodology (Reference 16). However, in view of the significant industrywide evolution of fuel designs since the establishment of these models, the NRC staff questioned the models applicability to modern fuel types to which Framatome would apply the AURORA-B evaluation model. The NRC staff

further observed that the AURORA-B LOCA evaluation model uses the S-RELAP5 code to calculate swelling and rupture, as opposed to the HUXY/BULGEX codes in XN-NF-82-07(P)(A),

Revision 1. Consequently, the NRC staff issued RAI 109 to request that Framatome provide justification for the swelling and rupture models used in the AURORA-B LOCA evaluation model.

Framatomes original response to RAI 109 ['''''''''''''' ' ''''''''''''''''''' '''' '''' ''''''''''' ''''''''

''''' '''''''''''''''''']31 Framatome further stated ['''' '' '''''''''' '''''''''''''' '''''''''''' ''''''''''''''' '''''

'''''''''''''''''' ''''''''''''' '''''']. Performing thermal-hydraulic analysis ['' ''''' ''''''''''''''''

''''' ''''''' '''''''''' ''''''' '''''' '''''''''''''''''' ''''' '''' '''''] Thus, Framatome concluded that the AURORA-B LOCA method is consistent with the methodology in XN-NF-82-07(P)(A),

Revision 1, as well as the currently approved EXEM BWR-2000 evaluation model documented in EMF-2361(P)(A) (Reference 16).

The NRC staffs review of Framatomes response found it generally reasonable that the currently approved model should continue to apply to modern fuel rods clad with Zircaloy-2.

The NRC staff ['''''''''' ''''''''' '''' '' ''''''''''' ''''''''''''''''''' ''''''''''''''''''''''''' '' '''''''''''' ''''''''''''

''''''''' ''' ''''''''''''''''''''''''''''''''''''''' ']. This comparison increases confidence that application of the swelling and rupture model to modern fuel designed by Framatome should produce expected results. The NRC staff also generally agreed [''''' '''' ''' ''''''' ''''''''''''''''' ''''

'''''''''''''''''''''''''''''''''''' ''''' ''''''''''''''''''''''''']. However, one code-specific aspect not addressed in response to RAI 109 is ['' '''''''''''' '' '''''''''''''''''' '''''''' ''''''' '''''''' '' '''''''''''''''''''''' ''''''''

''''' ' '''''''''''''''''''''''] In particular, XN-NF-82-07(P)(A), Revision 1, discusses ['''''''''''''''''''''' ' '

''''''''''''''''''''''''''''''''''] Neither Framatomes response to RAI 109 nor the S-RELAP5 code theory manual appear to address this point.

Framatome provided additional information on the temperature ramp rate calculation in S-RELAP5 in its updated RAI response dated October 31, 2018 (Reference 66). In particular, Framatome explained ['' ''''''''''''''''''''''''''' ''''''''''''''' '''''' '''''''''''''''''''' ' '''''''''''''''''''''' '''

31 The NRC staffs understanding was confirmed in Framatomes updated RAI response dated October 31, 2018 (Reference 66).

'''''''''''''''''' ''''''''''' '''''''''''''''''''''''''''''''''''''''''''''''''''' '] However, the additional information did not definitively resolve the issue because Framatomes means of time averaging appeared arbitrary and potentially subject to undesirable variation, the impacts of which were not assessed.

Therefore, the NRC staff designated Limitation and Condition 11 for plant-specific applications implementing the AURORA-B LOCA evaluation model to provide justification for the time period over which the temperature ramp rate used to determine the swelling and rupture of cladding is averaged.

In RAI 13, the NRC staff further questioned Framatomes modeling of cladding rupture. In response, Framatome stated [''''' ''' '''''' '''''''''''''''' '''''''''''''''''''''' ' ''''''''''''''''''' '''''' ' ''''''

''''''''''''''''''''' ' '' ''' '''' '''''''''' '''''''' ''''''''''''''']. Framatomes response to RAI 13.c states

['''' '''''''''''''' '''''' ''''''' ''''''' '''''''''''''' '''''''''' '''''''''' '' ''''''''' '''' ''' ''''''''''''''''''''' ' ''

''''''''''''''''''''''''''''' ''''''''' ''''''''''''''''''' ''''''']. However, the response to RAI 13.c goes on [ ''''''''

''' '''''''''''''''''''' '] A discussion of how ['''''''''''''''''' ''''''''''''''''' ''''''''''''''''''''''' '''''''''''''''''' '

''''''''''''''''''''''''''''' ' ''''''''''''''''''' ''''''''''''''''''''''''''' ' '' '''''''''''''''''''' ''''''''' ''''].

In its updated response to RAI 13.c, dated October 31, 2018 (Reference 66), Framatome discussed the ['''''''''''''''''''''''''' ''''''''''''''''' '' ''''''''''' ''''''''' ''''' '''''''''''''''' ''''''''''' ''''''' '''''''''''''''''

' ''''''''''''''''''''''']

In RAI 110, the NRC staff requested that Framatome further describe and justify the modeling of flow blockage due to swelled or ruptured cladding, as described in Section 6.4.5 of ANP-10332P. In particular, the NRC staff ['''''''' '''' ' '''''''''' ''''''''''''''' ''''''''''''''''''' ''

'''''''''''''''' ''''''''''' '''''' ''''''''''''''''''''''''''''' ''''''''''''''' ' '' ''''''''''''''''''' ''''''' '''''''].

Framatomes response to RAI 110 indicated [''' ' ''''''''''''''''' '''' ''''''' '''' ''''''''''''''''''''''''

'''''''''''''''''' ''''' ''''' ''''''''''' '''''''''''''].32 As an example, Framatome stated ['''''' '''''''''''''''''''''''''

' '' ''''''' ''''''''' '''''''''' '''' '''' ''''''''''''''''' ' ''''''''''''''''''' '''''''''''''' ' ]

The NRC staffs review of Framatomes response to RAI 110 determined the following:

• [''''''''''''''''''''''''''' ''''' ''' '''''''''''''''''''''''''' '''''''''''''' '' ''''''''''''''' ''''''''''''' ''''''''''''''''''' '

' '''''''' '' ''''''' ''''''''''' '''']

• [''' '''''''''''''' '''''' ''''''''''''''''' ''''''''' '''''''''''' ' ''''''''''' '''''''' '' '' '''''''''''''''''' ''''

'''' '''''''''' ''''''''''']

• [''''' ''''''''''''' ' ''''''''''''''' '''''''''''''''''' ' '''''''''''''''''' '''''''''' '''''' '''' '' ''''''''''''''''']

• ['''''''''''''''''''''''' ''''''''''''''''''''' ''''''''' '''''''''''''' '''' '''''''''''''''' ''''''''''''''' ''''' ''''''''''''''''''

''''''''''' '''''' ''''''''''' ''''''''' ''''''''' ''' ''''''''''''''''''' '''''''''' ''']

In its updated response to RAI 110 dated October 31, 2018 (Reference 66), Framatome provided the following additional information in response to the NRC staffs concerns:

• ['''''''''''''' '''''''''''''''''''' '''''''''''''''''' ''''''''''''' '' '''''''''' '''''''''''''' ''' ''''' ''''''''''''''''

'''''']

• [''''''''''''''''''''''''' '''''''''''''''' ''''''''''' '''''''''''''''''''' ' ''''''''''''''''''' ''''''' '''''''''''''''''''''''''''' ''

'''''''''''''''''' '''''''' '''''''''''''''' '''''' ''''''' '''''''''''''''' '''''''''' '''''''''''''''''''']

In light of the sensitivity study performed by Framatome, as well as the other evidence presented in the original and updated responses to RAI 110, the NRC staff considered the issue resolved. As noted by Framatome, [''''''''''''''''' ' ' ''''' ' ''''''' '' ''''''''''' ''''''''''''''''''''''''''''

32 Framatomes response noted ['''' '''''' ''''''' '''''''''''''''''''''''''' ''''''''''' '' '''''''''''''''' ''''''''''''''''' '''''''

''''''''''''''''' '''''''''''' '' ''' '''''''''''']

'''''''''''' ' ''''''''''''' '''''' ''''' '''''''''''''''' '''''''''' '' ''''''' '''''''' '''] evaluated subsequently in Sections 3.6.2.4 and 3.6.2.7.

3.3.4.1.3 Blowdown Phenomena Following the initial prediction of CHF exceedance at a given location in the core, Appendix K stipulates that use of nucleate boiling heat transfer correlations be discontinued at that location for the remainder of the blowdown phase, even if local conditions would apparently justify rewetting. As described above, the AURORA-B LOCA evaluation model would determine the CHF according to the 2006 version of the CHF tables compiled by Groeneveld (Reference 17).

['''''''''''''''''''' ''''''' '''''''''''''''''''''''''' ''' ''''''''''''''' ' '''''''''''''' ''''' '''''''''''''''' '''''''' '' ''''''''''''''''''

'''''''''''''''''''''''' ''''''' '''''' '''''''''''' '' '''''''''''''''''' '' ''''''''' ''''''''''''''''' '''''''''' '''''''''''''''''' ']

In RAI 17, the NRC staff requested additional information concerning Framatomes approach, with focus on the following issues:

• In Section 6.4.9 of ANP-10332P (and as reiterated in response to RAI 14), Framatome stated ['''' '''''''''''''''' '''''''''''''''''''''' ' ''' '''''''''' ''''''''''''''''''''' ' ''''''''''''''''' '''''''''

'''' '''''''''''''''''''''''''] However, the NRC staff questioned this interpretation because

['' ''''''' ' ' '''''' ''''''''''''' '''''''''''''' ''''' ''' ''''''''' ''''''''''''''' ''''''''' '''''''''''''''''''' '''''

'''''''''''''''''''''''''' ''''''''''''''''''''''' ''''''''''''''''''' '''''''''' ''''''' '''' ]. As such, the NRC staff further questioned whether Framatomes proposed application of ['''' '''''''''' ''''''''''''''''''''

''''' '''''''''''''''' ''''''''''' ' ''' '''''''''''''''''' ]

• The NRC staff further questioned Framatomes basis for ['''''''''''''''''' ' '''''''''''' ''' ''''''

''''''''''''''''''''''''''' ''''''' '''''''''''''''''''] in light of the Commissions opinion regarding the rulemaking hearing on ECCS acceptance criteria (Reference 14). [ '''''''''''''''' '''

'' ' ''''''''''''''''' '' '''' '''''''''''''''''''' '' '''''''''''''''' ' '''''''''''''']

• The NRC staff [''''' '''''''''''''''' '''' '''''''''''''''''''' '''''''''''' '''''''''''''''' '' '''''''''''' '

'''''''''''''''''' '''''''''''''''' '' ' '''''''''''''''''''' ''''''''''']

Framatomes response to RAI 17 [''''''''''''''' ''''''''''''''' '''' ''''''''''''''' '''''''' ' '''''''''''''''''''''''''''''''

''''''''''''''' ''''''''''''''''''''' ' ''''' ''''''''''' ' '''''''''''' ' ''''''''''''''' ''''''''']. However, in the course of addressing the NRC staffs questions, Framatome recognized that ['''' '''''''''' ''''''''''''''''''' '

''''''']. Therefore, in response to RAI 17, Framatome proposed a new approach [''''''''

'''''''''''''''''''''' ''''''''''''''''' '''''''''''''''''''' ' ' '''''''''''''''' ' '''''''''''''''' '''' ''''''''''''' '' ''''''' ''33 ''''

'''''''''''''''''''''' '''''''''''''''''''''''' '''''''''''''''''''' '''''''' ''''''''''''' ''''''''''' ''''''''''''''''''']. Framatome stated that

[' ''''''''''''''' '''''''''''''''''' ' '''''''''''''' '''''' '''' ''''''''''''''''''''' ''''''''' ' ' '''''''''''''''''''' ''''' ''''''''

'''' ''''''' '''''''''' ''''''''''''''''''' ''' ''''''']. Framatomes response [''''''''' '''''''''''''' '''''

'''''''''''''''' ''']

The NRC staff reviewed Framatomes response to RAI 17 and found that

• [''''' '''''''' '''' ''''''' ''''' '''' ''''''''''' ''''' '''''''''''''''' '' '''''''''''''''''''''' '''''''''''''' ''''''''''''

''''''''''''''''''' ' '''''''''''''''''''''''' '''''''' ''''''''' '''''''''''''''']. Furthermore, the NRC staff does not agree with Framatomes use of ['' ''''''''''''''''' '''' ''''''''''' '''''''''''''''''''' ''''''''''''''''''''''' '

''''''''' '''''''' ''''''' ' ]. The NRC staff also disagreed with [''''''''''''''''''''''''''

'''''''''''''''''' ''''' '' ''''''''''''' '''''''''' '''''''''''''' ''' ''''''''''''''' ''''' ''''''''''' '''''''''''''''] because the physical phenomena considered in each are distinct. ['''''''''''''''''''''' ' '''''''''''''''''''

''''''''''''''''''''' '''''''''''''''''''''''''' ' ] Thus, these two pieces of evidence do not confirm each other, and neither is applicable to post-LOCA dryout for a BWR.

• Justification was not provided for ['''''''''''''''' '' ''''''''' ''''''''''''''''''''' '''''''''' ''''' ''''''''''''''

'''''''' '''''''''' '''''''']

• Framatome did not adequately address the Commissions opinion regarding [''

'''''''''''''''''''''''' '''''' ''' ''''''''''' ' ''''''''''''''''''' '''''''' ''''' '' ''''''''' ]

• Available evidence from a historical review of precedent applications does not support Framatomes position [' '''''''''''''''' '''' ''''''''''''''''''' '''''''' ''' ''''''' ''''''' '' '

33 Note that Framatome cited the fourth edition of this reference, whereas the NRC staff referenced the fifth edition available in the NRC library.

'''''''''''''''''' ''''''''''''''''''''''''''''' '''''''''''''' ''' '''''''''''''''''''' ]

• The duration of ['' ''''''''''''''''' ''''''' '''''''''''''''''''''''''' ' '' ''''''''''''' '''''''''''''''''

''''''''''''' '' ']

Based upon its review of the response to RAI 17, the NRC staff did not agree with Framatomes proposals to ['''''''''''' '' ''''''''''''''''''' ''''''''''' ' '''''''''' '''''''''''''' ''''''''''' ''''' '''''

''''''''''''' ''' '''''''''''''''''' '']. Rather, as documented in Limitation and Condition 12, the NRC staff concluded that the AURORA-B LOCA evaluation model must impose the Appendix K lockout on return to nucleate boiling [' ' ''''''' '''''''' ''''''' ''' ''''''''''''''''''''' ''''''''' ''''''' ''''' ''''''''''''''''''

'''' '' ''''''' ''''' ''''''' ''''''''''''''''''' '''''']

Appendix K further requires progressive reduction of the discharge coefficient used with the Moody critical flow model to ensure that the maximum cladding temperature has been achieved.

While variations in split break flow area are considered down through the entire range of potentially limiting postulated small-break scenarios, with regard to the discharge coefficient applied to double-ended guillotine breaks, Framatome proposed to ['''''''''''' ''''' ''''''''''''''''''''''''

'''''''''''' '''''''' ''''''''''' '''''''''''''''''''''''''''' ''''''' ''''''''''''''''''' ''''''''''' '''' ''''''''']

Framatomes response stated that the calculations in Section 7.3.16 of ANP-10332P (e.g., Figures 7-159 and 7-160) were performed with ['''''''''''''''''''' ''''''''''''''''''' '''''''''''''''''''' ''].

Visual inspection of Figures 7-159 and 7-160 from ANP-10332P indicates that ['''''''''''''''''''''''

''''''''''''''''''''' ''' '''''' ''''' ''''''''''''''''''''' '''''''''''''''''''''''''''' ''''''''''''''''' ''''''']

Also in RAI 24, the NRC staff questioned whether Framatome intends to [''''''' '''''''''''''''''''''''''

'''''''''' ' '''''''' ''''''''''' ''''''''''''' ''''''' '''''''' '''' '''''''''']. Framatome responded ['''''

'''''''''''''''''''' ' '''''''''' ''''''''''''''''' '''''''''''' ] Therefore, the NRC staff considers Framatomes approach acceptable.

In RAI 9, the NRC staff requested that Framatome address recirculation pump pressure drop, including the assessment of sensitivity to pump operation or locking discussed in Item II.3 from Appendix K to 10 CFR 50. In response, Framatome characterized the S-RELAP5 pump model as a modified version of the RELAP5/MOD3 model that uses two-phase pump performance data from the Electric Power Research Institute. Framatome stated that the AURORA-B LOCA evaluation model ['''''''''''''''''' '''' ''' '''''''''''''''''''''''' ''''''''' '' ''''''''''''''' ' ''''''''''''''' '''''''' '''''

'''''''' '' ''''''''''''''''''' '''''''''''' ''''' ']

The NRC staff considered Framatomes response plausible, but noted that limited evidence

['''''' '''' ''''''''''''''''' '''''''''' '''' ''''''' ''' '''' ''''''''''''''''' ' '' '''''''''''''''''] was presented to justify that ['''''''''''''' ' ''''''''''''''''''''''' '''''''''''''''' ''''''' ' ' '''''''''''''''''' ''''' ''''''''''''''' ''''''''''''''''''''''

''''''''''' ''''' ' '' '''''' ''''''''''''' ''''''''''' ''''''''''''] Therefore, as a check on the reasonability of the response provided by Framatome, the NRC staff performed a set of independent sensitivity cases using the TRACE thermal-hydraulic code to examine ['' ''''''''''' '''''''' '''''''''''''''''

'''''''''' ''' ''''''''''''''''' '''''''''''''''' '' '''''''''''''''' '''''''''''' '' '''''''''''''' ''''''''] The results of the sensitivity calculations generally supported Framatomes position, ['''''''''''''''''' '' ''''''''''''

''''''''''' '''''''''' '''''''''' '''''''' '''''''''''' '''''' ''''''''''''' ''' '''''''''''''''''''''''' ''''''''''' ''''' ''''''''''''''''''']

However, the NRC staff also observed that the impact may be more complex than depicted in Framatomes response to RAI 9; in particular, [''' '''''''' '''' ''''''' '''''''''''''''''' '''' ''''''''''

'''''''''''''' '''''''' '''''''''''''''''''' '''''''''' ''''''' ''''''''''' ''''''''''' ' '''''''' ' '''''''''' '''''' ''''''''''''' '''''].

Based upon the discussion above, [''''' ' '''''''''''''''''''''''''''' '''' '''''''''''''''''''''''''''' '''''''''''''' '

'''''''''''''''''' ' ''' ''''''' ''''''' ''''''' '''' '''''''''''''''''''''' '''''''''''''' '''''''''''''''''''] for the AURORA-B LOCA evaluation model.

3.3.4.1.4 Refill and Reflood Phenomena As noted above, in the BWR LOCA event, the refill and reflood phases tend to overlap (i.e., reflooding of the core begins prior to the complete refilling of the lower plenum). This behavior occurs as the result of several factors, including one or more core spray systems (and in the BWR/5-6 design, also the low-pressure coolant injection system) that inject inside the core shroud, the channeled core design with engineered bypass leakage paths, and the countercurrent flow limitation at the fuel channel side-entry orifice. Although a qualitative physical understanding of the BWR LOCA event existed at the time Appendix K was developed, a more detailed quantitative understanding was not obtained until the completion of integral testing for the LOCA event, largely over the subsequent decade. Nevertheless, prescriptive lockouts on nucleate and transition boiling are imposed by Appendix K and are tied to the definitions of blowdown, refill, and reflood that apply to all light-water reactors.

Inasmuch as the complex phenomenology associated with the BWR LOCA event is not fully reflected in the prescriptive definitions of Appendix K, the NRC staff issued RAI 18 to request

that Framatome clarify how the heat transfer lockouts have been implemented and justify that the AURORA-B LOCA evaluation model conforms to Appendix K.

Framatomes response clarified that the Appendix K heat transfer lockouts ['' ''''''''''''''''''' ''

'''''''' '''''''''' ''' '''''' '''''''''''''']

On the other hand, in the case of the ['''''''''''''' '''''''''''''''''''' ''''''''' '''''' ''''''''''''' ' '' ''''''''

''''''''''''''''''''''''''' ''''' ''''''' ''''''''''''''' '''''''' ]

While the issue observed by the NRC staff resulted [ '''''' ''''''''''''' ''''''''''''''''' '''''''' ' '''

'''''''''''''''''''''''''''''' ''' '''''''''''''].34 Therefore, the NRC staff specified in Limitation and Condition 13 that the AURORA-B LOCA evaluation model may not release the Appendix K lockouts on heat transfer until [''''''''''''''' ''''' ''' '''''' '''''''''' ''' '''''''''''''' ''''''''''].

3.3.4.2 Additional Relevant Modeling Features Beyond the modeling options required by Appendix K, ANP-10332P specifies additional features that are essential for calculations performed under the AURORA-B LOCA evaluation model. This section of the NRC staffs SE describes several such modeling features the NRC staff deems worthy of discussion.

3.3.4.2.1 Hot Channel Model As described in Section 6.3.8 of ANP-10332P, the AURORA-B LOCA evaluation model incorporates a hot channel model that is intended to provide a conservative representation of the most limiting fuel assembly in the reactor core with respect to the criteria of 10 CFR 50.46.35 Defined hot channels ['' '''''''''''''''' '''' ''''''' ''''''' '''''''' ''' ''''''''''''''''''' '''''' ''''' '''

'''''''''''''''''''''' ''''' ]

The hot channel model also incorporates ['''''''''''''''''''''''''' '''''''''''''''' ''''''''''''''''' ''''' ''''''''''''''

34 Examples of plant-specific delay times are discussed in response to RAI 87.

35 Note that the Appendix K requirement for modeling the hot region of the core during blowdown using a region no greater than the size of one fuel assembly strictly applies only to PWRs.

'''''''''''''' '''''''''''''''' '' '''''''''''''''''''''''''' ]

Another salient feature of the hot channel model in the AURORA-B LOCA evaluation model,

[''''''''' '''''''' ''' '''''''''' ''''''' ''''''''''''''' ' ''' '''''''''''''''''' '''''''''''''''''' '''''''''''''''''

''''''' '''''''''''']

Because the AURORA-B LOCA evaluation model nodalizes [''' ''''''''''''' ''''' '''' ''''''' '''''''

''''''' '''''''' '''''' ' ''''''''''''''''' '''' ''' ']. During the NRC staffs review, Framatome stated

[' ''''''''''''''''' '''' ' '''''''''' '' '''''''''''' '''''' ' ' ''''''' '''''''''''''''''''''''''''' ''''''''''''''''''''''''' '''''

''''''''''''''''''''' ''''''''''''' ''' ''''''''''''''''' '''''''' ''''''''''''''''''''''' ''''''''''''''''''''''' ''''''''' '''''''''''] The NRC staff designated this position as Limitation and Condition 14.

In RAI 39, the NRC staff requested that Framatome explain how the hot channel model

[''''''''''''' ''''''''''''''''''''''''' ''''' ''''''''''' '' '''''''''''''''''''' ''''''''''''''' '''''' '''''''''''' '''''' '' '''' '''].

Framatomes response ['''''''''''''''''' '''' '''' ''''''' ''''''''''''''''''' ''' '''''''' '''' ''''''''''''''''' '''''

'''''''''''''''''''''' '' ''''''' ''''''''' '' ''''''' '''' ''''''''''''''''''] Framatomes response further clarified

['''' '''' ''''' ''''''''' ' '''''''''''''''''' ' ' '''''''' ''''''''' '''''' '''' ''''''''''''''''''''''''''''' ''''''''''''''''''''

''''''' ''' '' '' '''''''''''''''' ''''''''''''''''' '''''''' '' ''''''''''''''''''''''''''''' ' '] The NRC staff found Framatomes response to RAI 39 acceptable because it provided ['''''''''''''''' ''''''''''''''''''''''''

''''''''''''''' ''' ''''''''''''''''''''''''' '''''''''''''''''''''' '' ''''' '''''''''''' ''''''' ''' '''' ] that the NRC staff found necessary to supplement the discussion in Section 6.3.8 of ANP-10332P.

Framatome proposed [''''' ''' ''''''''''''''''''''''' ''''''''''''''' '''''''''''''''' ''''''' ''''''''''''''' ''''''' '

''''''''''''''' '''' ' '''''''''''''''''' '''''''''''''''''' ''''''''''''''''''' '''''' ' ' ''' '''''''''' '''''''''''']. The NRC staff issued RAI 40 on this topic, requesting that Framatome justify the alternative approach, particularly with respect to its validation against experimental data. [''''''''''''''''''''' ''''''''''''''''''''''

'''''''''''' ''''''''''''''''''''' '''''''''' ''''''' '''''' ''''''''''''''' '''' '''''''''''''''''''''''' ]

In its updated response to RAI 40 dated October 31, 2018 (Reference 66), Framatome further proposed [''''''''''''''' '''''''''''''''''''' ''''''''' ''''''''''''''''' ''''''''' ''''''''''''''''''''''''''''''''''''''''''''''''''' ''''''''' ''''''

''''''''' ]. However, in addition to the issues discussed above, the NRC staff did not agree that Framatome had adequately justified that the proposed approach to [''''''' ''''''''''''''''

'' ''''''''''''' ''''''''] would satisfy the requirements of 10 CFR 50.36 concerning technical specification limiting conditions for operation and surveillance requirements. From the NRC staffs perspective, the foregoing discussion in Section 3.3.4.1.1 concerning ['' ''''''''''''''''''

'''''''' ' '''' '''''''''' ''''''' '''''''''''''' '''''''] applies equally to Framatomes proposed approach to ['''''''' '' '''''''''''' ''''''''''' ''''''''''''''' ' '''''''' ''''''''''' ''''''''''' '''''''''''''''''''''' ''' ''''''''''''''''''

'''' ''''''''''''' ''''''''''''' '''''' '' ' '''''''''''' ''''''''' '''''''''' ''''''''''''']

Consequently, the NRC staff concluded that the AURORA-B LOCA evaluation model must generally [''''''''''''''' ''''' ''' '''''' ''''' ''''''''''''''''''''''''' ''''''''''''''''''''''' '' '' '' '' ''' '''''''

''''' ''''''' ''''' '''' '''''''''''''''''' ''' ''] in a manner consistent with this paragraph as Limitation and Condition 15 in Section 5.0 of this SE.

Defined hot channels are expected to provide the limiting results for both the peak cladding temperature and maximum local oxidation figures of merit. In many cases, ['''''''''''''''''''''''''

' ''''''''''''''''''''''''''''''' ''' '''''''' ''''']. The results of the demonstration cases included in ANP-10332P appear to support this view. [''''''''''''''''' '''''''''''''''''''''''' '''''''' ''''''' '''''''''''''''''''''

''''''''''''' '''' '''''''''''''' '''''' '''''''''''''''''''''''''' ''''''''''''''''''''' '' '''' '''''''''''''''' ''''''''''''''''' ''''''],

which the NRC staff finds acceptable. However, details concerning the implementation of this approach were not discussed in ANP-10332P. In its updated response to RAI 29.e dated October 31, 2018 (Reference 66), Framatome described additional approaches [ ''''''''''''''''

''''''''''''''''''''''''''''''''' '''''']. The NRC staff finds the three proposed approaches acceptable [

'''''''''' '''''''']

3.3.4.2.2 Countercurrent Flow Limitation ANP-10332P ranks the countercurrent flow limitation as [' '''''''''''''''''''''''''''' '''''' ''''''''''''''''''''''''

''' '''''''''' ''''''''].36 However, as noted in response to RAI 100, Framatome stated ['''' ''

'''''''''''''''''''''''''' ''''' ''''''''''''''' '''''''' ''''''''''']

The response to RAI 100 further describes the countercurrent flow limitation correlation parameters ['''''' '' '''''''''''' '''''' ''''''''''''''''''''''''']. In particular, the NRC staff found the correlation parameters Framatome used [' ''' '''''''''' ''''''' '''''''''''''''''' ''''' ''''''''''''''''''''''''''

''''''''''''''' '''''' '''''''' ''''' ''''''' '''''''''''''''''' ' ''' '''''''' '''''''] The parameters used in the

['''''''''''''' '''''' ''''''''' ''''''''''' '''''''''''''''''' ''''''''''' ''''''' '''''''''' '''''''''''''''''''''''' ' '''' '''''

''''''''''''''''''''''''''' '''''''' ' '' ''''''' '''''''''''''''''''' ''']. While the NRC staff ['''' ''' '''''''''''' ''

'''''''''''''''''''' ''''''''''''''''''''''''''' ''''''''''''''''''''''''''''''''' '''' '''''''''''''''''' '''''''''''''''''''''' '' '' ''''''''''''''''''' '],

the NRC staff agreed with Framatomes position [''' ''' ''''' ''''''''''''''''''' '' '''''''''' ''''

'''''''' '''''''''''''''' '''' ' '''''''''''''''''''' ' '''''''''''''''' ]

Framatome proposed in ANP-10332P [''' '''' ' '''''''''''''''''' '''''''''''''''''''''''''''' '''''' ''''''''''''''''

'''''''''''''''''''''' '''''''''''''''''''' '''''''''''''''''''' ' ''''''''''''''''''''' '' '''''''''''''''''''''''''''''' '''' ''''''''''''''].

However, the NRC staff had concerns with the proposed approach, [''''''''''''''''''''' ''''''''''''''

''''''''''''''''' ''''''''''''''''''''''''''''''' '''''' ''''''''''''''''''' ''''''''''''''''''''''' '' ''''''''''''''''''''' ''''''''''' ' '''' '''''''''''''']

Beyond this, more generally, ['' '''''''''''''''' '' ''''''''''''''''''''''''''''' ''''' '''''''''''''''''' ''''''''''''''''''' ' '''''''''

''''''''''''''''' ''''''''''''''''''''''''' '''' ' ']. As a result, the NRC staff concluded that, absent future NRC staff approval for approaches using [''''''''''''''''''''''''''''''' '''''' '''''''''''''''''''' ''''

36 Note that the PIRT in ANP-10332P ['''''''''''' '''''''''''''' '''' '''''''''''' '''''''' '' ''''''''''''''''''''''''''''' ''''

'''''''''''''''''' '''''' '''''''' '''''''''' ''''' ''''''''''' '''' ''''''''' '' ''' '''''''' '''''''].

'''''''''' ''''''''''''''' ' '''''''''''''''''''''' ''''''''' '''''''' '''''''''''''''''''''' ''''''']. To address these concerns, in responding to RAI 21, Framatome withdrew its request to use ['''''''''''''''''''''''''' ''''' '''''''''''''''

'''''''''''''''''''''' ' ''''''''''''''''''''' ''' ''''''''''''''''''''''''''''''' ''''' ''''''''''''''].

3.3.4.2.3 Bypass Leakage Flowpaths Engineered bypass leakage flowpaths (i.e., particularly those between the fuel bundles, core bypass region, lower plenum, and guide tubes) have a significant impact on the distribution of liquid within the reactor vessel during a BWR LOCA event, ['''''''''''''''''''''' '''''''' '' '''' '''''

''''''''''''''''''' '''' '''''''''''''''' ' ''''''''''''''''''''''''], which prompted the NRC staff to issue RAI 33.

Framatomes response to RAI 33 describes and diagrams the bypass leakage flowpaths modeled in the AURORA-B LOCA evaluation model. Framatome further characterized the general significance of these leakage paths in its response. The NRC staff found the vendors characterization consistent with the demonstration case input decks provided for audit purposes. Framatomes response [''''''''''''''''' ''''' '''''''''''''''' '''''''''''' '''''''''''''' '''''''''''''''''' ''''

''''''''''''''' ' '''' ''''''''''''''''''''''' '''''''''' ''''''''''''''''''''' ''''''''''']. Framatomes response further briefly discussed the modeling of hot wall effects ['''''' '''''''''''' ''''' ''''''''' '''''''' '''''''''''''''''' ''''' '''''

'' '''''''''''''''''''' ''''''']

The NRC staffs review of Framatomes response to RAI 33 found the response generally acceptable, as it provided the requested information and appeared to include modeling of all significant leakage paths. However, [''''''''''''''''''' '' ''' '''' '''''''''], the NRC staff concluded that Framatomes response did not provide an adequate description of and justification for ['''

''''''''''''''''''''''']. Furthermore, [ ''''' '''' '''''''''''''' '' '''''''''''''''''''''''''''' '''''''''''''''' ''''''''''''''''''

''''''''''''''''], the NRC staff questioned ['' ''''''''''''' '''''''''''''''''''' '' ''''''''''''''''' '''''''''''''''''''''''''

'''''''''''''''''''''' ''''''' ''''''''''''''''''' '' '''' ''''''''' ' '''''''''''''' ''''''''''''''''''''''''''' ''''''''''''''' ''''''']

The NRC staffs audit of the S-RELAP5 code theory manual (Reference 11) and draft modeling guidelines for the AURORA-B LOCA evaluation model (Reference 26) could not locate the relevant information.

In its updated response to RAI 33 dated October 31, 2018 (Reference 66), Framatome provided additional description ['''''''''''''''''''''''' ''' '''''''''''''''' ''' '' '''' ''''''''' '''''''''''''''''''

'' '''''''''''''''''''''''''' '''''''''' ]

The NRC staffs review of the updated response to RAI 33 found that Framatome had generally provided sufficient information concerning its proposed methodology ['' '''''''''''''''''''' '' ''''

''''''''' ' ''''''''''''' '''''''' '''' ''''''''''''' '''''''''''' ''''' '''''''''''' '''''''''' ''''''''''''''' '''''''''''''''''''] While the response does not adequately demonstrate [''''' ''''''''''''''''''''''''' '''''''''''' '''''''''''''' ''''''''''''

'''''''''''''''' '''''''''''''''' '''''''' '''''''''' ''''''''''' '' ''''''''''''''''''''' '''''''''' ''''' ''''''''''], the NRC staff concludes that the potential ['' '''''' '''''''''''''''''] is adequately accounted for by Limitation and Condition 22 in Section 5.0 of this evaluation, which stipulates that licensees implementing the AURORA-B LOCA evaluation model have the responsibility for verifying the applicability of all plant-specific design parameters.

3.3.4.2.4 Noncondensable Gas Intrusion As discussed in Framatomes response to RAI 34, the AURORA-B LOCA evaluation model

['''''''''''''''''' '''''''''''''''''''''''''''''''''''' '''''' '''''''''' '''''''''''''''''' ''''''''''''' ' '''''''''''''''''''''' ''''''''''''''''' '

'''''' ''''''''''''''''''] However, Framatomes response to RAI 34 stated ['''' '' '''''''''''''''''' '''''''''''''

'''''''''''''''''''''''''''''' ''''' '''''''''''''' ''''' ''''''' ''' ''''' '''''''' ]

While the calculations performed by Framatome [''''''''''''' ''''''''''''''' ''' '' ''''''''''''''''''''''''' '''''''''''

'' '''''''''' ''''''''''''''] The NRC staff performed two sets of confirmatory calculations with the TRACE code for a spectrum of break sizes for a BWR sample plant, [''''''''''''''''''''' '''''''''''''' ''

''''''''''''''''''''' '''''' ''''''''''''''''''''''''''''''' ''''' '''''''''''''' '''''''''''''''''''''''''''''' ''''''''''''''''''''''''''''']. In both cases, a significant presence of noncondensable gas in the reactor core was not observed. Ultimately, Appendix K has determined that a minimum containment pressure is limiting, and the NRC staff finds it appropriate to model noncondensable gas in a manner consistent with this position.

3.3.4.2.5 Safety/Relief Valves In its review of the BWR/6 small-break demonstration analysis presented in Section 7.7.6 of ANP-10332P, the NRC staff questioned why perturbations associated with the actuation of safety/relief valves were not apparent on the calculated trace of upper plenum pressure. A nearly constant pressure trace (after reaching the presumed safety/relief valve lift setpoint) was likewise observed in the corresponding BWR/4 small-break demonstration case. The NRC staff addressed these observations to Framatome in RAI 89. Framatome responded ['''' ''

''''''''''''''''''''''''''''''''''''' ''''''''''''''''''' ''''''''''']

3.3.4.3 Nodalization Computerized evaluation models require a nodalized representation of the reactor system (i.e., division of the spatial domain into discrete computational volumes and junctions) to permit numerical simulation of a given condition or event. In accordance with applicable regulatory guidance (e.g., Chapter 15.0.2 of the SRP, RG 1.203), the nodalization used for plant-specific safety analyses should be

• consistent with the nodalization used for the assessment and validation of the evaluation model, as well as the demonstration analyses,

• consistent with requirements imposed in the code theory manual and users manual to maintain code stability and proper numerical performance,

• sufficient to determine all phenomena of interest at an appropriate resolution, and

• assessed using appropriate sensitivity studies.

The nodalization Framatome proposed for the AURORA-B LOCA evaluation model is described in Section 6.3.7 of ANP-10332P.37 The AURORA-B LOCA evaluation model [''''''''''''''''''' '

''''''''''''''''''''''''' ' '' ''''''''''''''''''''''''' '''''''''' ''''''''''''''''' '''''''' ''''''''''''' ']

According to ANP-10332P, the nodalization used for the ['''''''' ''''''''''''''''' '''' '''''' ''''

''''''''''''''''''''' ''''''''''' ' '' ' ] for the AURORA-B LOCA evaluation model.

Excepting isolated instances specifically identified in this SE, the NRC staff generally found that Framatomes proposed nodalization scheme conforms to applicable regulatory guidance concerning nodalization. The nodalization scheme appears sufficiently detailed and relies on modeling practices [''''' ''''''''''''' '''''''''''''''''' ''''''''''''''''' ''''' '''''''' '''''''' ''''''''''''

''''''''''''''''' '''''''''''''''''' ' '' '''''' '''''' ''''''''''''''''''''''' '''' '''''''''''''''''''''''''''' ''''''''''''''''''''''] Several aspects of the nodalization of particular interest to the present review are discussed in succeeding sections of this SE, including the nodalization used for modeling breaks, connected systems, and the reactor core.

37 See in particular Figures 6-3 through 6-9 and Tables 6-5 and 6-6.

Consistency of the proposed nodalization scheme with the assessment and validation will be discussed further below in Section 3.4. Framatomes nodalization sensitivity studies will be discussed further below in Section 3.5.

3.3.4.3.1 Nodalization for Recirculation Line Breaks Framatome provided various nodalization diagrams for the reactor vessel and recirculation system in Section 6.3.7 of ANP-10332P. Specific locations of [''''''''''''''''''''' '''''''''' '''''''''''''' ''''''

'''''''''''''' '' '''''''''''''''''''''''' ''''''''''''''''''' '''' '''' '' ''''''''''''''''' '''''''''''''''']. The NRC staff found these break locations reasonable and noted ['''' '''''''''''''''''''''''' '''' '''''''''''''''' '' '''''''''''''''''''

''''' ''''''''''''''''''''''''' ''' '''''''''''''''' ''''''''''''''' ''' '''''''''''''''''''''''''' ]. Section 3.5.6 of this SE discusses sensitivity studies associated with nodalization near the break location.

3.3.4.3.2 Nodalization for Connected Systems The nodal structure for connected systems (e.g., ECCS, main steam, feedwater), was generally not adequately covered in ANP-10332P, and this topic became the subject of RAI 35. In response to RAI 35, Framatome stated ['''' '' ''''''''''''''''''' '''' '''''''''' ''''''''''''''' '''' '''''''''''''''''

''''''''''''' ''''''''' '' ''''''''' '''''''''] Framatome further identified that ['''''''''''''''' '''''''''''' '' '

'''''''' ''' '''''''''' ''''''''''''''''''''' ''''' ''''''' '''''''' ''''''''''''' '''''''' ''''''''''''''''' ''''''''''''' '' '''''''' '].

The NRC staff considered the overall modeling approach for connected systems proposed by Framatome to be reasonable. However, the response to RAI 35 did not provide adequate justification for the modeling of breaks on connected systems. In particular, [''' '''''''''''''''''''''

'''''''''''''''''] In particular, Framatome originally intended [ '''''''' ''''''''' ''''''''''''''''''''''''''''''' ''''

'''''''''''''''''''''''' '''''''''''''' '''' ''''''''''''''''' ''''''' '''''''' '''''''' '' ''''''''''''''''''''']. The NRC staffs audit further found that the draft modeling guidelines document (Reference 26) includes no guidance to support the use of consistent and appropriate nodalization for non-recirculation break scenarios.

In its updated response to RAI 35 dated October 31, 2018 (Reference 66), Framatome described a revised approach with additional detail concerning the proposed nodalization for non-recirculation system breaks. ['''' '''''''''''''''' ''''''''''''''''' '' '''''''''''''''''''''''''' '''''''''''''' ' ' ''''''

'''''''''''''''''''''''' '''']

Framatomes updated response to RAI 35 substantially addressed the NRC staffs concerns with the nodalization of non-recirculation system breaks. While ['''''''''''''''''''''''''' '''''''''''''''''''

'''''''''] somewhat limited in scope, they provided reasonable evidence that significant sensitivities to nodalization near the break location should not be expected for non-recirculation system breaks. [''''''''' '''''''''''''''''''''''''' ''''''''''''''''''' ''' '' '''''''''''''''' ''''''''''''' '''''''' '' ''

''' '''''''' '' '''''''''' '''' '' ''''''''''''''' '''''''''' '''''''''''''''''''''''''''''' ' ' '''''''''''''''''' '''''''' ' '''''''']

As expected, [' '''''''' ''''''''] in the AURORA-B LOCA evaluation model ['' ''''''''''''''''''''

'''''''''''''''''''''' '''''''''''''''''''' ''''''''''''''''' '''''''''''''' ''''''''''''''''''''''''' '''' '''''''''' '''''' ''''''''''''' '''''''''''].

3.3.4.3.3 Nodalization for Modeling Parallel Channel Flow In contrast [ '' '''''''''''''''''''''''''''''''' '''''''''''''''''''''''''] used in Framatomes current evaluation model for analyzing the BWR LOCA event (EXEM BWR-2000 (Reference 16)), the proposed AURORA-B LOCA evaluation model described in ANP-10332P simulates the reactor core using

[' ''' '''''''''''''' '''''''''''' '''' ' ' '''''''''''' ''''' '''''''''''''''' '''' ''''''' ''''''''' ']. This nodalization allows the LOCA analysis to reflect realistic differences in channel flow regime. Multi-channel phenomena occur throughout the spectrum of LOCA events, but are particularly notable during the blowdown phase of large- and intermediate-break scenarios that involve rapid flashing of the liquid in the reactor vessel lower plenum. As observed experimentally in the SSTF, updraft of vapor through some fuel channels in the reactor core, in conjunction with the countercurrent flow limitation at fuel channel side-entry orifices and upper tie plates, was an important driving factor in the observed parallel channel flow behavior (Reference 36). In particular, depending upon channel power, location, and other factors, the fuel channels in a BWR core have the potential to enter one of three possible parallel flow regimes:

• co-current upflow of vapor and liquid, which tended to be exhibited by high-power channels,

• countercurrent flow, which tended to be exhibited by average channels, or

• liquid downflow, which tended to be exhibited by non-limiting, low-power peripheral channels.

The testing further revealed that fuel channels may transition from one flow regime to another, depending upon the rate of flashing in the lower plenum and other factors.

Accurate or conservative prediction of the parallel channel flow phenomenon is important because fuel channel flow regime may strongly influence bundle thermal-hydraulic conditions and, hence, the calculation of cladding temperature and oxidation. While the basic phenomena governing parallel channel flow behavior during a BWR LOCA event are known, a reliable means of predicting the dynamic behavior of specific fuel channels does not exist. Furthermore, considering the range of possible conditions, it is not obvious a priori whether more limiting results will be achieved by a high-powered channel entering the co-current upflow regime or countercurrent flow regime.38 While the [''''''''''''''''' ''''''' ''''''''''''''''] in the AURORA-B LOCA evaluation model offers the potential for improved analytical realism, the NRC staff [''''''''''''''' '''' '' '''' '''''''''''''

'''''''''''''''']. The impact of this observation, coupled with NRC staff questions concerning other aspects of Framatomes modeling approach ['''''' ''' ''''''''''''''''' '''''''''''''''''''''''''''''''' ''''''''''''''''''

''' ''''''''' '''''''''''''], caused the NRC staff to more broadly question S-RELAP5s capability to model parallel channel flow accurately. These concerns are described further below in Section 3.5.5, along with the NRC staffs assessment of Framatomes ['''''''''''''' ''''''''''''''''

''''''''''''''''' '''''''''''].

3.3.5 Application Framework In ANP-10332P, Framatome concentrated on describing the AURORA-B code system and providing justification for its physical models and correlations according to the EMDAP paradigm. Although the S-RELAP5 code represents the foundation of the AURORA-B LOCA evaluation model, as noted above, it is the entire evaluation model that is under review in this SE, which further involves the framework of inputs, assumptions, code modeling options, boundary conditions, nodalizations, analysis procedures, and so forth.

In particular, the following sections of this SE describe steps of practical importance in performing plant-specific safety analyses using the AURORA-B LOCA evaluation model.

Because these aspects of the evaluation model, referred to herein as the application framework, are not systematically described in ANP-10332P, the NRC staff asked a series of RAIs to bring the topic into clearer focus. Important points from these interactions are summarized below.

3.3.5.1 Plant Safety Analysis Process In response to RAI 116, Framatome outlined the process for performing plant-specific analysis using the AURORA-B LOCA evaluation model, as summarized below in Table 4.

38 For example, if the channel holds up significant water, the countercurrent flow regime may provide superior cooling capacity. On the other hand, if very little water is held up and stagnant flow conditions exist in the upper part of the bundle, the countercurrent regime may be disadvantageous.

Table 4: Plant-Specific Safety Analysis Process for the AURORA-B LOCA Evaluation Model Step Description In cooperation with the plant licensee, Framatome uses a variety of sources to compile a database of information required to complete the analysis. The Database database includes items such as plant data, equipment characteristics and Setup setpoints, allowable operating domains (e.g., power and flow limits, equipment-out-of-service considerations), and operating limits (e.g.,

['''''''''''''''''''' '''''''''''''' '''' ''''''''''''''''''''''''''''' ''''''' ' '''''''''''''''' '''''''''''''''''']).

Power History A limiting power history [' ''''''''''''''''''' '''' '''''''''''''' ''' '''''''''''''''''' '''''''''

Analysis ''''''''''' ''''''' ''' ''''''''''''''''''''' ''''''''''''''' ''''''''']

Various break sizes and locations, single-failures, and operating conditions Break are analyzed to determine the most limiting LOCA scenario for each plant.

Spectrum

['''' '''''''''''''' ' ''''''''''''''''''''' ''''''' ''' ''''''''''''''''''' '''' ''''' ''''''''''''''''''

Analysis

''''''''' ''''''''''''' '''''' '''''''''''''''' ' ''''''''''']

['''' ''' ''''''''''' '''''''''''''''''''''' ''''''''''''''''''''''' ''''''' ''' ''''''''' ''''''''''''''''''

Exposure '''''''''''''''''''' '''''''''''''''' '''''''''''' '''''''''''''''' ''''''''''''''''''' ''''''''''''''' '' '''''''' '

Study '''''''''''''''''''' '''''''''''''''''''' '' ''' ''''' ''''' '' ''''''''''''''' ''''''''' '''''''

''''''''''''''' '''''''''' ''''''''''''''''''''''' '' ''']

The calculated figures of merit are compared to the acceptance criteria specified in 10 CFR 50.46. ['' ''''''''''''''''''' '' '''''''''' ''''''''''''' '''''''''''''''''''

Evaluation

'''''']

3.3.5.2 Steady-State Initialization The NRC staff requested in RAI 43 that Framatome explain how steady-state calculations are performed in both S-RELAP5 and RODEX4 to establish the initial conditions for the transient LOCA simulation with the AURORA-B LOCA evaluation model. In response to RAI 43, Framatome discussed how steady-state calculations are performed [ '''''''''''''''''''''' ''''''''''''

'''''''''''''''''' ''''' '' '''''''''''''''''''''''' ''''''''''''''''' '''''' '''''' '''''''''''''' '''' ''''''''' ] The response further notes that steady-state calculations [''''''' ''' '' ''''''''''' ' ''' '''''''''''''''''' ''''''' ''''

'''''''''''''''''''' '''''''' ''''''''''] The NRC staff considered the initialization practices described in response to RAI 43 to be reasonable and appropriate.

3.3.5.3 Break Selection The NRC staff requested in RAI 29.a that Framatome specify which break locations will be considered within the scope of a plant-specific analysis. Framatome responded that plant-specific analyses will address breaks in both recirculation and non-recirculation systems, including ECCS, feedwater, and main steam piping. Framatome stated [' '''''''''''''''''' ''''

''''' '''''''' '''''''''''' '' ''''''''''''' ''''''''''''''''''''''''] However, the NRC staff noted [''' '

''''''''''''''' '''''''' ' ''''''''''''''''''' '] In its updated response to RAI 35 dated October 31, 2018 (Reference 66), ['''''''''''''''''''''''' ''''''''''''''''' '''''''''' '' '''''''''' '' ''''' '''''''''''''''' ''''' ''''''' ''''''''

'''''''''''' '''''''''''' '' ' ''''''''''''''''''''''].

The NRC staff requested in RAI 29.b that Framatome specify the break spectrum span and resolution for use in plant-specific analysis. Considering the results from the demonstration cases in Section 7.7 of ANP-10332P, the NRC staff questioned the sufficiency of (1) the proposed break size resolution in the intermediate range of the break spectrum and (2) the proposed minimum break size. Framatomes response to RAI 29.b provides an overview of the proposed break spectrum for plant-specific analysis, which incorporates changes in response to the NRC staffs concerns. This information, as revised in Framatomes updated RAI response submitted on October 31, 2018 (Reference 66), is summarized below in Table 5.

Table 5: Break Spectrum for the AURORA-B LOCA Evaluation Model Break Range Treatment Double-Ended '''''''''''' '''''''' ''''''''''''''''''' ''''''''''''''''''''''' ''''' '''' '''' '''' '''''

Guillotine Breaks

'''''''39 ''''' ' '''''''''''''''''' ' ''' '' ''''''''''''''''' '''''''''''''''' '''''''''''''''''

Large Split Breaks Small Split Breaks '''''''''''''' ''''' '''''' '''''''' '''''''''''''''' ''''''''''''''''''''''''' ''''' ''''''' '' '''''

Framatomes response to RAI 29.b, as revised in the submittal dated October 31, 2018 (Reference 66), provided the requested information and addressed the concerns raised by the NRC staff by proposing improvements to the break spectrum analysis procedure in ANP-10332P; consequently, the NRC staff found the response acceptable. In particular, [''

39 For reference, note that the single-sided pipe areas considered in the demonstration cases were approximately ['''''''''''' ''].

' ''''''''''''''''' ''''''' ''''' '''' ''''''''''' '] In the NRC staffs judgment, [ ''''''''' '''''''''''''''''''

''''''''''''''' '''''''''' '' ''''''''' ]

Using a size increment [ '' ' '' ''''''''''''''''' ''''''' ''''' ''''''''''' '''''''''' ' ''''' ''''''''''''''''''''

'''''''''''''''''' '''''''''''''' ''''''''''''''''''' ''''''''''''''' '''''''''''''']. Furthermore, a finer resolution ['' '''''''

'''''''''''''''''''''''''' ''''''''''''' '' ''''''''' ''''''''''''']

The NRC staff requested in RAIs 29.c and 29.d that Framatome discuss ['' '''''''''''''' '

'''''''''''''''''''' ''''''' '' '' ''''''''''''''' '''''''''' ''''''''''']. Framatome responded that [''''''''''''''''' '''''''''

'' '''''''''' '''''''''''''''' ]40 However, in its updated response to RAI 29.d dated October 31, 2018 (Reference 66),

Framatome clarified that, [' '''''''''' ''''' ''''''''''''''''''''' ''''''' '' ''''''''''''''''''''''' ''''''''' ''''''''''''''''''''

'''''''''''''' '''''''''' '' '''''''''''''''' ''''''''''''''''''''''' '''''''''' '''''']

In light of the simplified modeling approach and general lack of available data for validation, the NRC staff considered it appropriate that the AURORA-B LOCA evaluation model 40 This conservatism mainly applies [' '''''''''''''' '''''''' '''''''''' '''''''''''''''''''''' ''''''' ''''''''' ''''''''''''''''''''''''''''''

''''''''' '''''']

[''''''''''''''''''''''''''''' '''''''''''' ''''''' '' ''''''''''' '''''''''''''''' '''' '' ''''''''''' '''''']. The NRC staff did not fully agree with Framatomes response to RAI 29.c concerning [''''''''''' ''''''''''''''''' '''''''''''''''''''''

'''' '''''''''''' '''''''''''''''' ' ''' ''''''' '''''''' ''''''' ]; in particular, consideration of these factors may influence reactor pressure and water levels for some postulated break sizes.

However, the effect is not expected to be large. ['''''''''''''''''''''''''''' '''''''''''''' '' ' ''''''' ' ''''''''

''''''''''''''''''''''''' ''''']

3.3.5.4 Selection of Limiting Scenario The NRC staff requested in RAI 29.e that Framatome identify whether a single simulation can achieve limiting conditions for peak cladding temperature, maximum local oxidation, and core-wide oxidation, or whether the analysis of separate cases, potentially with different limiting single failure combinations, is necessary. The NRC staff further requested that Framatome address how the limiting oxidation results [''''' ' '''''''' '' '''''' ''' ''''''''' ' ''''''''''''''' '''''''

'' ''''''''' '''' ' '''''''''''''' '' '' '''''''''''' '''''''''''''' ''''''''''''''''''''''']. Framatome responded [''''

''''''''''''' ''''''' ''' ''''''''' ''''''''''''''''' '''''''''' '''' '''''''' ''''''''''''' '''''''''' '''''' '''''''''''' ''''''''''''''''''].

Framatome further stated that, [ '' ''''''''''' ''''''''' '' ''''''''''''''' ''''''''''''''''''''''''' '''' ''''''''''''''''

''' ' ''''''''''''''''''' ' '''''''' ']. The NRC staff found this approach [''''''''''''''''' '''' '' '''''''''

''''''''''''' '''''''''' ''''' ''''''''''], acceptable.

In RAI 29.f, the NRC staff requested that Framatome clarify how the limiting single failure for each scenario is determined. Framatome responded ['''' '''''' '''''''''' '''''''''' ' '''''''''''''''''' ''

''''''''''' '' ''''''''''''''''''' '' '''']. Framatome further indicated ['''' '' '' ' '''''''' ''''''''''''''

' ''''''''''''''''''''' ' ''''''' ''''''''''''' ' '''''''''''''''''''''' ' ''''''''''''''''''''''' ''''''''']. The NRC staff considers this approach acceptable and agrees ['''' '''''' '''' '' '''''''''''''''' ''''''''''' '''''''''''''

'''' '' '''''''''''''' ''''''''''''''''''''''''''' '' '''''''' ''''''''''' ''''''''''''''' ''''''' '' ''''''''' ' ' '''''''''''''''''].

In RAI 29.g, the NRC staff requested that Framatome discuss the key initial conditions ['''''

''''''''''''''''''' '''''' '''''' '''''''''''''''' ''''' ''''' '''''''''''''''''''''''''' '''' '''''''''''''''''''' ''''''''' ''']. Framatome responded ['''' ''''''''''''''' '''''''''''''''' '' ' '''''' ' '''''''' '''''''' ''''''''''''''' '''''''''''''''''' ''''' '''''

''''''' '' ''''''''''''''''''' ''''''] Framatome stated that the following conservative initial conditions are assumed [''''''''''''''''''' ''''''''''''''''' ' ''''' ''' '''''''''''''''''''''' ''''''''''' '''''''''''''''''''' ''''''''']:

• [' '''''''''''''' ''''''''''''''''''''' ' '''''''''''' ''''''''' ' '''''' ''''''''''' ' ''''''' '''''''' ' '''''''''''''']

• ['''''''''''''''' ''''''' '''''' ''''''''''''''''''''' ''''''' ''' '''''' ''' ''' '''' '''''''''' '''''''''' '''''''''''''''''

''''''' ''''''''''''''' '''''''''''''''''''''''''''''' ''' '''''''']

• ['' ''''''''''''''''' ''''' '''''''''''''' '''''''''''''' '' ''''''''''''''''''''']

• [''''''' '''''''''' '''''''''''' ''''''''''''''''''''''''''''''' '''''''''''''''''']

• [''''''''''''''''''''''''' ''''''''' ''''''''''''' ' '''''''''''''''' ''''''''''''''''''''''''''''''''']

• [' ''''''''''''''''''''''''''''''' ''' ''''''''''''''''''''''''' ''''''''''''''' '' ''''''''''''''''''']

The NRC staff found Framatomes proposed modeling practices to be appropriate [' ''''''''''

'''' ''''' '''''''''''''''' ''''''''''''''''''' '''''''''''''' ' '' ''''''''''''''''''''''''']. However, while ANP-10332P recognizes that [''''''''''''''''''''''''''''' ''''''''''' ''''''''''''''''' ''''''' ''''''''''''''' ' ''''''''''''''''' '''''''' '''''''''''''''''''''

'' '''''' ''''''''''''''''' ''''''''''''''''''' ' ''''''''''''''''''''''''']. Examples of extended operating domains include extended power uprates and extended flow windows, as well as operation with specific equipment out of service and at conditions applicable only at certain times in cycle, such as single-loop operation, automatic depressurization system valves out of service, feedwater heaters out of service, and feedwater temperature reduction. Plant-specific applications of the ANP-10332P methodology must ensure that the LOCA analysis addresses the entire allowed operating domain, which may require the analysis of multiple initial operating states.

Furthermore, justification that all portions of the operating domain are addressed must be included specifically in each plant-specific application of ANP-10332P, which the NRC staff has designated as Limitation and Condition 16 in Section 5.0 of this SE.

In RAI 29.h, the NRC staff requested that Framatome explain whether it is necessary to model explicitly cases with and without offsite power availability, [ ''''''''' '''''''''''''' ''''' ' '''''''

''''''''''''''''' '''' ' ''''''''''''''''''''''''''' ' ''''''''''''''' '''''''''''''''''''']. Framatome responded that

[''''''''''''''''' '''''''''''''''''' '''' '''''''''''''''' '''''''''''' '''' '' '''''''''''''''' ''''''''''' '' ''''''''''' ''''''''].

Framatome stated that the AURORA-B LOCA evaluation model ['''''''''''''''' ''''' '' '''''''''

''''''''''''''''' '' ''''''''''''''''' ' ''''''''''''''''''' '''''' '''''''''''''''''''' '''' '''''''''''''''''''' ' '] The NRC staffs review found Framatomes response acceptable ['' ''' ''''''''''''''''''' '''''''' '''''''''' '''''''''' ''''

''''''''''''''''''' ''''' '''''''''''''''''''''] In the updated response to RAI 29.h dated October 31, 2018 (Reference 66), Framatome [''''''''''''''''' '''''''''''' ' ''''''''''''''''' '''''''''''''''' ' '''''''' '''''''''''

'''''''''] Therefore, to assure satisfaction of GDC 35 (or similar plant-specific requirement)], the NRC staff designated Limitation and Condition 17 for [''''''''''''''''''' '' '''''''''''''' ' ''''''''''''''''''''

' '''''' '''''''''''''''''''''' '' ''''''' ''''''''''''' ''''''''''' ''''''' '' '''''''''''''''''' '''''''].

In RAI 29.l, the NRC staff requested that Framatome discuss the modeling of the linear heat generation rate, decay heat, and stored energy, and [ '''''''''''''' ''''' ''''''''''''''''''''' ''''

'''''''''''''''''''''''''''''' '''''''''''''''' ''''''' ''''''''''''''''''''''' ' ''''''''''' ''''''''' '' ''''' ''' ''''' '''''''''''''' ].

Framatome responded that [''' ''' ''' ' '''''''''''''' '''' ''''''' '''''''''' ''''''''''''''''' ''''''' ''

'''' ''''''' ]. Conservative assumptions [' '''''''''' ''''' ''''' '''''''' '''''''''' '''''''''''''''''

''''''''''''''''''''''' '''' ''''''''''''' ' ''' ''']. The NRC staff found Framatomes modeling approach acceptable because it conforms to Appendix K.

3.3.5.5 Component Modeling The NRC staff requested in RAI 29.i that Framatome clarify ['''''''''''''''' ''''''''' ''''' '''''' '

''''''''''''''''''' ' ''''''''''''''''''']. Framatomes response [''''''''''''''' '''' ''''''''''' '''' '''''' ' '''''''''

'''''''''' ' '' '''''''''''''''' ''''''' '''' ''''''' ''''''''''''''' '''''''' '''''''''']. The NRC staff finds this approach conservative and acceptable.

In RAI 29.j, the NRC staff requested that Framatome clarify whether credit may be taken in plant-specific analysis with the AURORA-B LOCA evaluation model for a reactor trip signal based upon high drywell pressure. Framatome originally responded [''''' ''''''''''''''''''

''''''''''' ''''''''' '''''' '''''''' ''''''''' ''''']. However, in its updated response to RAI 29.j dated October 31, 2018 (Reference 66), Framatome replaced its original response with a statement

[''''' '' ''''' ''''''''''''' '''''''''''''''' '' ''''''''''''''''''''''''''''' ''''''''''''''''''' ' ''' ''''''''''''''''''''''' ''''''''''

''''''''''''''''''''' ''''''''''''' '''''''''''''''''''''' ' '' ''''''''''''''''' ''''''' ''''''''''' ''''''''''''' ''''''''''']

The NRC staff would consider [''''''''''''' ' ''' ''''' '''''''''' '''''''''''''''''' '''''''' ''''''''''''''''''''' '''''''

''''''''] Ultimately, the updated response to RAI 29.j represents a significant revision to the evaluation model that arose at the end of the review period and was not accompanied by adequate supporting analyses or other justification. As a result, the NRC staff concluded that justification for the modeling of the high drywell pressure trip signal must be provided on a plant-specific basis, which is addressed by Limitation and Condition 18.

Section 6.2.3 of ANP-10332P notes that Framatome [''' ''' '''''''''''''''''' '''''''''' ''''' ''''' ''

'''''''''''''''''''''''''''' ''' ''''''' '''''''''''''''''''''''' '''' '''''''''''' ''''' '''''''''''''''''''''']. The NRC staff requested in RAI 45 that Framatome clarify how the rod insertion time is modeled conservatively, ['''''''''' '''' ''''''''''''' '''' '''''''''''''' ''''''' ''''''''' ''''''''' ''''''''''''''''''''' ''''' '

''''''''''''''''''''''''''''''''''''''''''' ''''''''''''' ''''''''''''''''''''' ''''''''' '''''''''' '''''''']. Alternately, the NRC staff requested ['''' '''''''''''''''''''''''' ''''''''''''''''''''''' ''''' ''' '''''''''''''''' '''''''''' '''''''' '''' '''''''''''''''''''

''''''''''' ' '' ''''''']. Framatome responded by ['''''''''''''''''''' ' '''''''''''''''' ''''''''''''''''''''' '' '

''''' '''''''''''''''''' ''''' ''''''''''''''''''''''''''''' '''''''''''''''' ' ''''''''''''''''' ''''''''''' '' ''''''''''''''''] is acceptable.

In RAI 113, the NRC staff requested that Framatome address application of the AURORA-B LOCA evaluation model to mixed core configurations. Framatome responded by ['''''''''' '''''

''''''''''''''''''''''''''' '' '''''''''''''' '' ''''''''''''' ''''''''''' '''''''''''' ''''''''''''' ] While generally considering Framatomes response appropriate, [''''''''' '''' '''''''''''''''''''''''''' ' '''' '''''''''

'''''''''''''''''''''''' '''''' ''''''''''''''''' ''''''' ] This position is specified as Limitation and Condition 19 in Section 5.0 of this SE.

3.3.5.6 Exposure Dependence Framatome mentions in ANP-10332P the ['''''''''''''''''''''''''''' ''''''''''''''''''''''''''''''''''''''' ''''''''''''''''''''''

'''''''''' '''''''''''''''' ''''''''''''''''''' '''''''' '''''''''''''''''''' ''''''' ''''''''''''''' '''''''''''' '''' '' '''''''''''''' ''' ]

Therefore, the NRC staff issued RAIs 46 and 117 to request that Framatome [''''''''''' ''''''''''''''''

'' ''''''''''''''' ''''''''' '''''''''''''''''''''''''''''''''''''''''' '''''''''''''''''']. In response, Framatome [''''''''''''''''''

'''''' '''''''''''''''' ''''''''' ']

Framatome further stated ['''' ''''''''''''''''' ''''''''''''''''''''''' '' '''''' ' ''''''''''''''''''' ''' ''''''''''''''' ''

'''''''''''' '''''''' ''''''''''''''''' ''''''''']

Framatome stated [''''' ''''' '''''''''''''''' '''' '''''' ''''' '''' ''''''' '''''''''''''''''''''''' '''''''''''''''''''''''

'''''''''''' '''''''' '''' '''''''' ''''''' ' ''''''''' ''''''''''''''''''''''''' ''''''''''''''''' '''''''''''''']

The NRC staff found the proposed approach [' ''''''''''''''''''''''' '''''''''''''''' ''''''''''''] appropriate

[''''''''''''' ''''''''' ''''''''''''''''''''''''''' ''''''''''''''' ' '''''''''''''''' '' '''' '''''''''''''''''''' ''''' '''''''''''''''' '''''

'''''''''' '''' '' '''''''' ''''''''''''' ''''''''''''''''''''''''''' ''' ''''''''' '''''''''''''''' '''''''''].

3.3.5.7 Analysis Termination Criteria Framatome stated in ANP-10332P that AURORA-B LOCA simulations may be terminated

['''''''' ''' '' '''' '''' '''''''''''''''''''' ''''' '''''''''''''''' ' '''''' ''''''''''''''''''' ''''''''''''''''''''''''''

'''''''''''''''''''' ''' '' '''' ''' ''''''''''' ''' '''''''''''''''''''' ''''' ''' '' '''''''''''''' ''''''''''''''''''' '] In RAIs 30 and 31, the NRC staff questioned the termination criteria specified in ANP-10332P because

• if defined ['' ''''''' '''''''' '' '''''''''''''''''''''' '''''''''''''''''''' '''''' '' '''''''''''' '''' '' ''''''

''''''''''''' '''''''''' ' '''''''''''''' '''''''''''''''''''''''''' '''' ''''''''''''''''' '''''' ''''''' ''''''''''''''],

• considering the potential [' '''''''''''''' '''''''''''' '''' '''''''''''''''' ' '''' '''''''''''''''''''''''' '''''''''

'''' ''''''' '''' ''' '''''''''' '' '''' '' '''''''''''''''''' ''''''''], and

• [''' ''''''''''''''''''' '''''''''''''''''''' '' ''''''''''''''''''''' ''''' '' '''''''''''''''' '''''''' ''''''''''''''''''' ''''

''''''' '' ''''''''''''' ''''''' '''''''''' ''''''''''' '''' ''''''''''''' ''''''''''''''''''].

Framatome responded to RAI 30 with a broader discussion of the concerns raised by the NRC staff. [''''''''''''''''''''''' '''''''' '''' '''''''''''''''''''''''' ' '''''''''''''''' '''''''''''''''''''''' '''' '''''''''''''''''' '''''''

''''''''''''''''''''' '' ''''''''''' ''''''''''''''''''''''' ' ' ''' '''''''''' ''''''''''''''''']

• ['' '''''''''''''''' '''''''''''''''''''''' '''''''''''''''''''''''''''' '''''''''''' '''''' ''''' '' ''''''''''''''''' ''''''''''''''''''''],

• [''' '''''''''''''''''''' '' ''' ''''''''''''''''' ' '' ''''''''''''''' '''''''''''''''''''''''' ''''''' ' '''''' '''''''''],

• ['' '''''''''''''''''''''' ''' '''''' '''''''''''''' '''''''''''''''''''''' ''''' '''' '''''' '''' ''''''''''''''''''''''

'''''''''''''''''''' '''''''''''''' ' ''''''''''' '''''''''' '' ''' '''''''''' '''''''''''''], and

• ['' ''''''''''''''''''' '' ''''''''''' ''''''''''''''' ''''''''''''''''''''' ''''''''''''' ''''' '''' ''''' '''''''''

'''''''''''''''''' '''''''''''''''''''' ''''' ' '''''''''''''''''''']41 To further support assessment [ ''''''''''''''''''''''' '''''''''''''''''' ''''''''''''''''''''''''' '''''''''''''''''''' '''''''''''''''

'''''''''''''''''].

The NRC staffs review found that Framatomes response to RAI 30 generally addressed considerations applicable to ['''' ''''''''''''''''''' '''''''''''''''''''''''' '''''''''''''''''''''' '''''''''''''''' ''''''''''''''''''

'''''''''''''''''''']

In response to RAI 31, Framatome ['''''''''''''' '''''''''''''''' '''''''''''''''' ''''' ''''''''''''' '' '''''

''''''''' '''''''''''''''''' '' ''''''''''''''' '''''' ''' '''' ' '''' ''''' '' ] As evidence, Framatome ['''''''

''''' ''''''' '''''''''''''''' ' '''' '''''''''''''''''] The NRC staff noted in particular that the demonstration cases [''''''' ' ''''''''''''''''''''''' ''''''''''''' '''''''''''''''''''''' ''''' ''''''''' ''' '''''''''''''' ' '' '''''''''''''''''''''

''''''''''''''' ''''''''''''''''' ' '' ''' ''''' ''''''''''''' ''''''''''''''''].

The NRC staff concludes ['''' ''''''''''''''''''' '''' ''' '''''''''''''''''''''' '''''''''' ''''''''''''''''''''' ''''''''' '''''''''''

''''' ' ''''''''''''''''''' '''''''''''''' '''''''''''''''' '''''''''''' ' '''''''' '''' '''''''''' ''''''''''''' ''''''''''''''''''''' ''''''''']

Therefore, the [''' ' ''''''''''''''''''''' ''''''''''' ''''''' '''''''''''''' ''''' ''' '' ''''' '''''''''''''''''''''' ''''' ''

''''''''''' ''''''''''''''''''' '''''''''' ''''' ''''''''''''''''''''''' '''], the NRC staff has incorporated Limitation and Condition 20 into Section 5.0 of this SE [' ''''''''''''' '''''''''''''''''''''''''' '''' ''''''''''''''''''' '''''''''' ''''

41 Note that in the ['''' '''''''''''''''''''' '''''''''' '''''''''''''''''''''''''''' ''''''''' '''''''''''''' ' ''' '''''' '''''' ''''''''''''' ''

'''''''' ''''''''''''''' '''''''''''''''' )) ] in independent confirmatory analysis cases performed by the NRC staff using the TRACE code, which under some circumstances occasionally predicted the reversal of a cooling trend in response to breakdown of the countercurrent flow limitation at the side-entry orifice prior to complete refilling of the lower plenum.

''''''''''''''''''''''''''''' ]

During the NRC staffs review of the AURORA-B LOCA evaluation model, Framatome revealed that certain simulations performed in ANP-10332P relied upon a non-representative practice for avoiding premature termination [''''' '''''''''''' ''''''''''''''''''''''' ''''''''''''''''' '''''''''''''''' ''''''''' ''''''''''''''

'''''''''] that was not described in either the TR or the draft modeling guidelines for the AURORA-B LOCA evaluation model (Reference 26) audited by the NRC staff. [''''''''''''''''''''''''

''''''''''''''''' ''''' ' 42 '' '''''''' '''' '''''''''''''''''''''' ''''''''''''''''''''' ''''''''''' ''''' '''''' '''''''''''''''''''

'''''''''''''' ''''''''''''']

As a result, Framatome ['''''''''''' ''''''''''' ''''''''''''''' '''''''''''''' '''''''' ''' '''''''''''''''''''''' '''''''''''''''''''

'''''''''''''''''' '''''''''' ''''''''''' '''''' '''' ''''''''''''' ''''''''''''' ''''''''''''''''' ''''''''''''''''''''']. In general, reliance on non-representative modeling practices to avoid premature termination should be discouraged. Rather, in such cases, it is preferable that vendors identify and address the specific issues responsible for the premature termination of code calculations. ['''''''''''''''' '

'''''''''''''''''''''' '''' ''''''''''''''''''''''' '''' '''''' ''' ' ''' ''''''''''']

While the NRC staff observed ['''' ''''''''''''''' ' '''''''''''''''' '''''''' ''' '''''''''''''''''''''''''' ''''''''''''''''''

'''''''''''''' '''''''''' ' ''''''''''''''''''''''''''' '' ''' ''''''''''' ''''''''''''''''''''''''''''''' ''''' '''''''''''''' ' ], because the specific procedure for implementing the non-representative modeling practice was not defined and documented, its conservatism for other plant configurations remains in doubt.

Therefore, the NRC staff concluded that, prior to implementing this practice in future plant safety analyses, the practice must be adequately defined in the AURORA-B LOCA modeling guidelines. Furthermore, this non-representative modeling practice may not be implemented in the safety analysis for any given plant without prior, plant-specific approval by the NRC staff.

Plant-specific submittals to the NRC requesting credit for this non-representative modeling practice must adequately describe the extent of its intended use and justify its conservatism.

The justification must address the potential for [''''''''''''''''''''' '''''''''' '''''''' ''''''''''''''''''''''''''''''

'''''''''''''''''''''''''' '''''''''''''''''' '''''''''''''' ''''''''''' ''''''''''''''''''''''] excessive sensitivity to timestep and 42 [''' '''''''''''' ''' '''''''''''''''''''''''''' ''''''''''''''''''' '''''''''''''''' '''''''' ''''''''''''''' ''' '''''''''''' ''' '''''''''''''''''''''''''

''''''''' '' '''''''''''' ''''' ''''' '' '''''''''''''''''''''' '' ''''''''''' ]

nodalization variations, which issue is discussed further in Section 3.5. The NRC staff designated this position as Limitation and Condition 21.

3.4 Assessment and Validation of Evaluation Model According to Chapter 15.0.2 of the SRP, assessment and validation of code models and correlations should encompass the following areas and general features:

• Assessments should cover all code models, commensurate with their importance and required fidelity.

• Models should be assessed over the entire range of conditions encountered in the transient or accident scenario.

• Assessments and validation should be performed with a single, frozen version of the evaluation model that applies the same code modeling options as would be used for plant-specific calculations.

• Comparisons of evaluation model calculations should be performed against separate-effects (i.e., local) testing as well as integral-effects (i.e., global or system-wide) testing.

• An analysis should be performed to identify and evaluate scaling distortions that may affect the assessment and validation process.

Framatome has presented information summarizing the assessment and validation of the AURORA-B LOCA evaluation model in Sections 5.4 and 7 of ANP-10332P. However, the NRC staff ultimately found the assessment and validation contained in ANP-10332P incomplete because it did not adequately justify deviating from regulatory guidance to perform the code assessment on a single, frozen version of the evaluation model [''''''' ''' '''''''' ''''''''''''''''

'''''''''''' ''''''''''''''' ' ''''''''''''''''''''''' ''''''''''''''''''']. In particular,

• Framatome elected to change the S-RELAP5 code version used by the AURORA-B LOCA evaluation model from ['''''''''''''''''] (on which ANP-10332P was based) to

['''''''''''''''''] during the NRC staffs review. By contrast, a fundamental tenet of the code assessment and validation process is that calculations submitted in support of the NRC staffs review of the evaluation model (e.g., for assessment, validation, demonstration) should all derive from the same frozen code version. While Framatome undertook a significant effort to assess and validate the [''''''''''''''''] version of S-RELAP5, as documented in its responses to a number of RAIs, including 15, 28, 51, and 126, ['''

'' '''''''''''''''''''''''' ''''' ''''''''''''''''' '''''''' ''''''' ''''''''''''''''''''''''''' ''''''' ''''''''']. Furthermore,

['''''''''''''' ''''''''''''''''''''''' ''''' ''''''''''''''''' '''''''''''''''''''''' ''''''' ''''''''''''''''' '' '' '''''''' '''''''''

'''''''''''''''''''' ''' '''''''''''' '''''''''''''''] analyses in ANP-10332P. Thus, the majority of the NRC staffs review documented in Section 3.4 of this SE pertains to assessments of the

['''''''''''''''] version of S-RELAP5. Consideration of the ['''''''''''''''] version of S-RELAP5 is discussed specifically in Section 3.4.5.

• Framatome performed a significant portion of the assessment and validation of the AURORA-B LOCA evaluation model ['''''''''' ''''''''''' '''''''''''''''''' '''''''''''' ' ''''''''''''''''

'''''''''']

The term ['''''''''''''''''''''''''''' '''' '''''' ''''''''' ''''''''''''''''' '''''''''' ' ''''''''' ''''' ''''''''

'''''''''''''' ]. In response to RAI 27, Framatome clarified [''''' ''''''' '''''''' ''''''' '''''''''''''' '''''''''''

'''''''''' ]. Framatome stated ['''' ''''''' '''''''''''''''''''''''''''''' ''''''' '' '''''''''''''' ''''''''''''' '''''' '''''''''

'''''''''''''''' '' ''''''''''''''''''''''' ' ''' '''''']. Framatome also noted in response to RAI 90 [''''

'''''''''''''''''' ''''''''''], which provide an illustration of their impacts.

The NRC staffs review of Framatomes response to RAI 27 ['''''''''''''''''' '''' '''''''''''''''''''''' ''''

''''''''''''''''' '''' '' ''''''''''''''''''' '''' ''' ''''''''''''''''' ' ]. However, the NRC staff reasoned that a sufficient body of evidence is available to support the assessment of the Appendix K-based AURORA-B LOCA evaluation model, as constituted by

• the [''''''''' ' '''''''''''''''''''''''''''''''''] assessment comparisons included in ANP-10332P,

• two integral assessments presented in Section 7.9 of ANP-10332P that incorporate a subset of the required evaluation model conservatisms,

• the demonstration analyses and timestep / nodalization sensitivity cases [''''''''''''''''''

''''''''''''''''''''''''''], and

• the response to RAI 58, [''''''''' ' ''''''''''''''''''''' ''''''''''' ' '''''''''''''' ''''''' ''''''''' '' ''''''''''

''''''''''''''' ''''''''''' '''''''''''''''''' ''''''''' '''''''''''''''''''''''''''''].

Ideally, validation of an evaluation model would provide direct evidence of the adequacy of the method. However, in view of the significant differences between the analytical methods Framatome used for code validation and those intended for plant-specific application, the NRC staff will generally refer to the assessment performed by Framatome as a validation of the S-RELAP5 code rather than of the AURORA-B LOCA evaluation model per se. While validation of S-RELAP5 in the abstract indirectly supports and provides confidence that the AURORA-B LOCA evaluation model will perform as intended, the distinction between these concepts should not be glossed over.

Regulatory guidance recognizes the inescapable truth that all testing relied upon for evaluation model assessment and validation involves scaling compromises and scoping limitations. To mitigate to the greatest extent possible the consequences of such distortions, assessment and validation should include testing at different scales and in different facilities for the phenomena of greatest importance.

The assessment and validation effort documented in ANP-10332P follows this general principle.

In Section 5.4 of ANP-10332P, Framatome stated that it relies upon four types of comparisons that consider relevant phenomena across a range of scales and test facilities, including

• foundation methodology assessments, which terminology characterizes models that may be considered to have undergone an external validation process (e.g., models required by or found acceptable in Appendix K to 10 CFR 50, supporting methodologies from TRs previously approved by the NRC staff),

• component effects tests, which are specific to the performance of specialized components such as centrifugal pumps, jet pumps, and steam separators,

• separate effects tests, which are specific to individual phenomena of importance to the event under consideration, and

• integral effects tests, which consider the overall response of the entire system (or a significant portion thereof) to a multitude of phenomena associated with the event under consideration.

Succeeding sections of this SE discuss assessments performed by Framatome in each of these categories. In assessing agreement between evaluation model predictions and test data, Framatome used the four-tiered schema described in RG 1.203. Likewise, the present SE applies these criteria, which are paraphrased below in Table 6.

Table 6: Assessment and Validation Evaluation Criteria Level of Agreement Definition Evaluation model exhibits no deficiencies in modeling a given behavior.

Excellent Major and minor phenomena and trends are correctly predicted. The calculated results agree closely with data.

Evaluation model exhibits minor deficiencies but overall provides an acceptable prediction. All major trends and phenomena are predicted Reasonable correctly. Differences between calculated values and data are greater than are deemed necessary for excellent agreement.

Evaluation model exhibits significant deficiencies and overall provides a prediction that is not acceptable. Some major trends or phenomena Minimal are not predicted correctly, and some calculated values lie considerably outside the specified or inferred uncertainty bands of the data.

Evaluation model exhibits major deficiencies and provides an unacceptable prediction of the test data because major trends are not Insufficient predicted correctly. Most calculated values lie outside the specified or inferred uncertainty bands of the data.

To demonstrate the adequacy of the AURORA-B LOCA evaluation model for the prediction of highly ranked PIRT phenomena, Framatome stated [''''' ''''''''''''''''''''''''''' ' '''''' ''''''''''''''''''''

''''''''''''''''''''' ''''' ''''' ''''' ' '''''''''''''''''''']. In general, the NRC staff considers this an appropriate standard for assessing the capability of an evaluation model.

A recurring issue in the NRC staffs review of the assessment and validation of the AURORA-B LOCA evaluation model is that Framatome typically associated a large number of phenomena with each series of tests or foundation methodology assessments, regardless of whether the evaluation models prediction of each of these phenomena could reasonably be characterized as having been validated on an individual basis from the information presented in ANP-10332P.

While this SE notes such instances of perceived EMDAP nonconformance, the NRC staff recognizes that the AURORA-B LOCA evaluation model is ultimately a conservative Appendix K evaluation model; whereas, the EMDAP was intended primarily for best-estimate models that explicitly account for modeling uncertainty. Since the AURORA-B LOCA evaluation model conforms to Appendix K, the ultimate objective of the NRC staffs review of Framatomes assessment and validation is to ensure that necessary models exist and support a conservative prediction of the defined figures of merit. Framatomes ultimate satisfaction of this objective will be assessed below in Section 3.4.6.5.

The NRC staff was assisted in its review of the AURORA-B LOCA evaluation model by consultants from Brookhaven National Laboratory, who particularly focused upon the area of assessment and validation. Further details supporting the NRC staffs review of Framatomes assessment and validation effort may be found in the technical evaluation report prepared by Brookhaven National Laboratory (Reference 35) (withheld for proprietary reasons).

3.4.1 Foundation Methodology Assessments In Sections 5.4 and 7.3 of ANP-10332P, Framatome invokes three sources of foundation methodology assessments, namely: Appendix K required and acceptable features, Framatomes approved core simulator methodology, and Framatomes approved fuel thermal-mechanical performance methodology, each of which is discussed below by the NRC staff.

3.4.1.1 Appendix K Required and Acceptable Features Models and correlations defined within Appendix K to 10 CFR 50 as being required or acceptable evaluation model features have been previously reviewed by the NRC staff and do not require an additional demonstration of their acceptability for use in an Appendix K-based LOCA evaluation model. However, the NRC staffs review of Section 5.4.1 of ANP-10332P found that Framatome associated a number of phenomena with the Appendix K foundation methodology assessment despite the specific models and correlations used in the AURORA-B LOCA evaluation model not appearing to be defined within Appendix K as required or acceptable features.43 The situation is clarified below in Table 7.

43 Note that a similar list in [''''''''''''' '''''' ''''''''''''''''''''''' '''''''''''' '''' ''' ''''''''''''''''''''''''' ''''' ' '''''''

'''''''''' ''' '''''''' '''''''' '''''''''''''''''''''].

Table 7: Evaluation of Appendix K Foundation Methodology Assessment Evaluation Model Applies an Appendix K Evaluation Model Applies Method Not Required or Acceptable Feature Specifically Defined in Appendix K Core: Decay Heat '''''''' ''''''''''' ''''''''''

Core: Metal Water Reaction Parameters ''''''''' '''''' ''''''''''''''' '''' '''''''''''''

Jet Pumps: Critical Flow at Jet Pump Nozzle ''''''' ''''''''''''''''''''''' '''''''''' ''''''''''''''''''''''''' ''''

Recirculation Lines: Critical Flow at Break ''''''''' ''''' '''''''''''

In considering Table 7, the reader should understand that the use of modeling practices not specifically defined in Appendix K is necessary in a number of areas where Appendix K does not explicitly prescribe required or acceptable features. Furthermore, the use of alternative modeling practices in lieu of acceptable features defined in Appendix K is expressly permitted by the regulation.

The AURORA-B LOCA evaluation models approach for modeling each of the phenomena in Table 7 has been described in the preceding discussion. Of these approaches, ['''''' ''''''''' '

'''''''''' ''''' '''''''''']. While requiring consideration of the remaining phenomena on the right-hand side of Table 7, [''''''''''''''' '' ''''''' '' '''''''''''''''''' '''''''' ' ''' '''''''''''''''''''''''' ''

'''''''''''' ''''''''''''''''''''' ''''''''' ''''''''''''' '' ''''''''' '''''''''''''''''''''''']. As observed in the technical evaluation report from Brookhaven National Laboratory (Reference 35), however, the phenomena on the right-hand side have generally been included in one or more assessment comparisons described in ANP-10332P. As such, the acceptability of the AURORA-B LOCA evaluation model with respect to the prediction of these phenomena will be assessed by other means.

3.4.1.2 Lattice Physics/Core Simulator Methodology The use of an approved lattice physics/core simulator methodology in support of the AURORA-B LOCA evaluation model is ['''''''''''''''''''' ''''''''''''''''' '''''' ''''''''''''''''''''''' '''''''''''''''''

'''''''' '''''''''''''' '''''''''''' ''''''' '' '''''''''''''' '''''''''''''''''' '''''''''''''''''''''''''''''''''''''''''''''''''''''' ] As discussed above in Section 1.1, the NRC staff has designated Limitation and Condition 1 for Framatome to use approved core simulator and lattice physics methods in support of the AURORA-B LOCA evaluation model.

ANP-10332P associates ['''' '''''''' ''''''' '''' '' '''''''' '''''''''''''''''''''''' ''''''''''''''' '''''''''''''''''''''''''''

''''''']. From the NRC staffs perspective, [''''''' '''''''''''''' ' ''' '''''''' ''''''''''''''''''''''''''' '''''''''''''''''

'''''''''''''''''' ''''''''''''''''' ''' '''''''''''''' ''''''' '' '''''''''''''''' ''''''''''' '''''''''''''''''''''']. As noted in the previous section, the ['''''''' '''''' ''''''''' '''''' ' '' '''''''''''''''''''' '''''''''' ''''''''''''''''' '''''''''

'''''''' '' '''''''' '''''''''''''''' ''''''''''']

3.4.1.3 Fuel Thermal-Mechanical Methodology The use of an approved fuel thermal-mechanical methodology in support of the AURORA-B LOCA evaluation model [' '''''''''''''''''''''''' '''''''''''''' ''''''''''''''''''''''''''' '''' ''''''''''''''''''''''' '''''''''''''''''''''''''''

''''''''''''' '' '''''''''''''''''''' '''''''''' ''''''''''''''''' '''''] As noted above in Section 1.1 and captured in Limitation and Condition 2, [''''''' '''' '' '''''''''''''''''''''''' ''''''''' ''''''''''''''''''''' ''''''''' ''''' ''''''''' '

'''''''''''''' '''''''''''''' ''''''' ''''''' ''''''''''''''''], the NRC staff considers it appropriate that the AURORA-B LOCA evaluation model [''''''''''''''''''' '''' '' ''''''''''''''''''''''] RODEX4 code.

The NRC staff agreed that the approved RODEX4 methodology (Reference 27) ['''' '

''''''''''''''' ''' ''''''''''' ''''' ''' '''''''''''''''''''''''' ] In fact, the methods for determining these LOCA phenomena are associated with neither the RODEX4 steady-state code [''' ''' ''''''''''

'''''''''''''' ''''''''''''''' '''''''''''''' '''''''''''''''''''''''' '''' '''''''''''''''''''']. As described above, the S-RELAP5 code [''''''''''''''''' ''''''''''''' ''''''''''''' ''''' ''''''''''''' ''''''''' '''''''''' '''''''''''''' '''''' ''''''''''''''''''''''''''''

'''''''''''''''''''''' ''''''''''''' ''''''''''''''''' ''' ''''''''''''''''''''''' ''''''''''''''''''''].

Table 8: Evaluation of Fuel Thermal-Mechanical Foundation Methodology Assessment Assessment Includes Assessment Excludes 3.4.2 Component Effects Tests In Sections 5.4 and 7.6 of ANP-10332P, Framatome cites five sources of component effects test data for assessment and validation:

• Rod Bundle Pressure Drop Tests

• Jet Pump Performance Tests

• Steam Separator Tests

• Critical Power Tests

• Countercurrent Flow Limitation Mini-Loop Tests This SE reviews each source of assessment data successively below. The NRC staff found the tests concerning jet pump performance and countercurrent flow of greatest relevance to the assessment of the AURORA-B LOCA evaluation model. Reduced emphasis is placed on other areas where a similar or identical assessment has been recently reviewed that the NRC staff finds applicable to the AURORA-B LOCA evaluation model.

3.4.2.1 Rod Bundle Pressure Drop Tests As described in Section 7.6.1 of ANP-10332P, the S-RELAP5 code was assessed against a large database of rod bundle pressure drop measurements taken at the ATLAS and KATHY test facilities to validate the [''''''''' '''''' ''''''' '''''''''''''''''''''''''''' '''''''''''''''''''''']. Framatome indicated that the experimental database includes measurements taken across a wide range of [''''

''''''''''''''''' ''''''''''' '''''''''''''' '''''''''''''''''''' '''''''''''''''''' ''''''''''''''''''' '''' '''' '''''''''''''''''].

Framatomes database of differential pressure measurements [''' '''''''' '''''''''''' ''''''''''''''''' '

'''''''''''''''''''''''''' '''''' ' '' ''''' ''''''''''], there was excellent agreement between S-RELAP5s prediction and the test measurements. The NRC staff concludes that ['' '''''''''' ''''' ''''''''''

''''''''' '''''''] to the capability of S-RELAP5 to determine the initial rod bundle pressure drop prior to the LOCA event in support of the [''''''' ''''' ' ''''''' '''''''''''''''''''''''''' '''''''''''''''''''''''] for the AURORA-B LOCA evaluation model.

3.4.2.2 Jet Pump Performance Tests As described in Section 7.6.2 of ANP-10332P, predictions of the S-RELAP5 code were compared to single-phase and two-phase test measurements using 18 different reduced-scale and prototypical jet pump assemblies to assess the following PIRT items:

• [''''''''' ''''''''''''''''''''''' '''''''''''''''''''''']

• ['''' ''''''''' '''''''''' '''''' '' ''''''''' ''''''''''']

• ['' '''''''''' ''' ''''''''' ''''''' '''''''''''''' ''''' ''''''''''''''']

The single-phase tests compared calculated and measured flow- and pressure-drop-ratios across reduced-scale and prototypical jet pump assemblies.

The two-phase tests simulated blowdown behavior for 1/6th-scale jet pump assemblies at conditions applicable to both the intact (i.e., all flows in normal, forward direction) and broken (i.e., suction flow in forward direction, drive and discharge flows in reverse) recirculation loops.

Comparisons to pressure measurements using both the homogeneous equilibrium and Moody critical flow models are included in ANP-10332P.

The majority of the jet pump performance testing ['''''' '''''''''''''''''' '''''''''''''''''' ''' '''''' ''''''''''

'''''''''''''''''' ''''''''''''' ''''''''''''' ''''''''''' ' '''''''''''''''' '] Therefore, the present review focused primarily on these issues, [''''' '''' ''''''''''''''''''''' '''''''''' ''''''''''''''''' ''' '''''''''''' ''''''' ''''''''

''''''''''' ' '' ''''''''''''''''''''''''' '''''''''''''' ] These component effects tests for BWR jet pumps supplement the more general separate effects assessment of the Marviken critical flow testing (discussed subsequently in Section 3.4.3.13).

3.4.2.3 Steam Separator Tests As described in Section 7.6.3 of ANP-10332P, predictions of the S-RELAP5 code were compared to steam carryover, steam carryunder, and pressure drop measurements for two- and three-stage steam separators44 to assess ['' ''''''' ''''' ''''''''' '''''''''''''''''''''''''' '''''''''''''''''''''].

Framatome concluded in ANP-10332P that the agreement of the S-RELAP5 predictions to the measured data for each of the three phenomena ranges from reasonable to excellent.

The NRC staff [''''''''''''''''' ''''''''''''''''' ''''''''''''''''''''''''''' '''''''''''''''''''''''' ''''''''' '''''''''''''''''''' ''''''''''''''''''''''

'''' '''''''''''''''''''''' '''''''''''''''''''''''''''''' ' ''' ''''''''''''''''''' '''''''''''' '''''''''''''''''''' ] The NRC staff ultimately agreed that the S-RELAP5 predictions provide reasonable to excellent agreement with the 44 A two-stage separator is used in BWR/2-5 designs, and a three-stage separator is used in the BWR/6 design.

measured data for all three phenomena (carryover, carryunder, and pressure drop). Although steam separator performance influences the steady-state operating condition from which most postulated events initialize, from the NRC staffs perspective, detailed modeling of steam separator behavior is not of high significance to the LOCA scenarios of concern to this SE.

3.4.2.4 Critical Power Tests Section 7.6.4 of ANP-10332P cites Section 6.5.4 of ANP-10300P (Reference 7) as providing the assessment basis for the capability of the S-RELAP5 code to address [''' ''''''''' '''''' '''''''''

''''''''' ''''''''']. Framatome concluded that the previous assessment performed for ANP-10300P remains valid for the LOCA event. Framatome added that the calculation of critical power using an approved fuel-specific correlation is performed during the steady-state initialization of the AURORA-B LOCA evaluation model ['''' '' ''' '''''''''''' '''''''''''''''''''' '' '''''''''''''''''''''''''' '''

'''''' '''''''''''''''' '''''''''''' '' ''''''''''' ] the Groeneveld CHF lookup table.

From the NRC staffs perspective, methods for predicting dryout tend to rely on semi-empirical approaches that are only valid within their qualification domain. Whereas, the thermal-hydraulic conditions (e.g., mass flow rates, void fractions, flow regimes, pressures) under which fuel-specific correlations are typically applied generally do not envelop the conditions under which incipient dryout typically occurs during the spectrum of postulated LOCA events. Thus, there is considerable uncertainty in applying such critical power test data to the assessment of a LOCA evaluation model. Nevertheless, the NRC staff considers Framatomes approach reasonable and appropriate, ['''''''''''''''' ' '' ''' ''''''''''''''''''' '''''' ''''''''''''''''''''' ''''''' '''''''''''

''''''''''' ''''''''''''''''''' ''''''''''].

3.4.2.5 Countercurrent Flow Limitation Mini-Loop Tests As described in Section 7.6.5 of ANP-10332P, predictions of the S-RELAP5 code were compared to tests performed in the Countercurrent Flow Limitation Mini-Loop Test Facility to assess prediction of the [''''''' ''''' '''''''' ''''''''''''''''''''''''''''''''' ''''' '''''''''''''''' '''''''''' '' ''''''''''].

Framatome described the Mini-Loop test facility [ '''''''' '''''''''''''''''''''' '''''''''' '' '' ''''''''''''''

'''' '''''''''''''''''' ''''' ''''''''''''' ''''''' ''''' ]

Framatome stated that the experimental data ['''' ''''''''''''''''' '''''''' '' '''''''''''''''''''

''''''''''''''''''''' '''' ''''''''''''']

The NRC staff considered the S-RELAP5 prediction of the Mini-Loop test data to be in reasonable to excellent agreement with the measured data. However, in RAIs 22 and 100, the NRC staff asked several questions to confirm the applicability of the test scaling to actual plant conditions. In response, Framatome [''''''''''''' ''' ''''''''''''' '''''' '''''''''''''' ''' ''''''' '''''''''''''''

'''''''''''''' ' ], the NRC staff found the results of this assessment adequately supportive of the AURORA-B LOCA evaluation model.

3.4.3 Separate Effects Tests Sections 5.4 and 7.3 of ANP-10332P cite 14 sources of separate effects test data ['''''''''''''''''''

'''''''''' '''''' '''''''''''''''''''''' '' '' ''''''''' '''']. The NRC staff considers validation of significant phenomena using separate effects testing to be appropriate, since it is intended to isolate the performance of individual code models against relevant test data. [''''''''''''''''' ''''''''' ''''''' '''''

'''' ''''''''''''' ''''''''' '''''''''''''''''''''']. The NRC staffs review of this issue is provided below in Section 3.4.4.

The first four separate effects test comparisons consider experiments dealing with void distribution and level swell absent post-dryout heat transfer. These assessments have been recently evaluated by the NRC staff in similar or identical form during the review of ANP-10300P (Reference 8), and will not be emphasized in the present review. The remaining ten comparisons involve LOCA phenomena including post-dryout heat transfer, critical flow, and the countercurrent flow limitation. These comparisons have not been previously reviewed and are the focus of the present evaluation.

3.4.3.1 Rod Bundle Void Tests As discussed in Section 7.3.4 of ANP-10332P, Framatome assessed S-RELAP5 against void fraction measurements in rod bundles taken at several different test facilities under a variety of conditions. The test facilities in the assessment include FRIGG2, FRIGG3, and KATHY. The FRIGG testing was performed in the late 1960s using simulated Marviken fuel elements of unique design, whereas the KATHY testing simulated a modern ATRIUM 10A fuel bundle.

Framatome stated [''''' ' ''''''''''''''''''''''' ''' '''''''''''' '''''' ''''''''''''''''' ''''''''''''' ''' '''''''

''''' ' ''''''''' ''''' '''''''''''''''' '''' '''''''''''''''''''' ''''''''''''''''''''''].

Framatomes assessment of this test data was previously evaluated by the NRC staff in Section 3.3.1.2 of its SE on ANP-10300P (Reference 8). In response to RAIs from the NRC staff during the review of ANP-10300P, Framatome [''''''''''' '''' '''''''''' '''''''''''''' ''''''''''''''

''''''''''''''''''' ''' '''''''''''''''''''''' ''' '''''], the NRC staff ultimately found AURORA-B capable of providing reasonable predictions of the void distribution in rod arrays.

With respect to the present review of ANP-10332P, the NRC staff finds that the conclusions regarding the rod bundle void test assessments originally performed for ANP-10300P generally remain valid, insofar as they are applicable for the AURORA-B LOCA evaluation model.

However, the NRC staff notes [' '''''''''''''''''''' '''''''''''''''' ' '''' ''''''''''' ''''''''''''' '' '''''''' '

'''''''' '''''''''' '' ''''''''' ''''''' '''' '''' ''''''''''''''''''''''''' ' '''''''''' ''''''''''''''''''''''' '''''''''''''''''']

Therefore, considering the conditions under which the data was taken, the NRC staff found this assessment relevant primarily to the determination of the steady-state void distribution used to initialize the LOCA transient calculation.

3.4.3.2 Christensen Void Tests As discussed in Section 7.3.5 of ANP-10332P, Framatome assessed S-RELAP5 against void fraction measurements in 7 tests performed in 1961 at Argonne National Laboratory.

Framatome stated that, while the primary purpose of the testing was to study void oscillations and stability, the experiments provide data on steady-state axial void distributions, particularly for the subcooled boiling regime. The test facility used a simplified geometry involving a heated rectangular tube constructed of stainless steel. The rectangular tube measured 1.11 x 4.44 centimeters in cross section and was 127 centimeters in height.

Framatome stated that these tests are used to assess the following PIRT items:

• [''''''''' ''''' ''''''''''''''''''' '''' ''''''''''''''''''''''' ''''''''''''''''''''''''' '''''''''']

• [''''''''''' '''''''''''''' '''''' ''''''''''''''''''''''' '''''''''''''''''''''''' ''''''''''''' '''''''''']

• [''''''''' '''''''''''''' ''''''''''''''''''''''' ''''''''''' ''''''''''''''''''''''']

The assessment of 7 Christensen void tests (a total of 112 test points) was originally performed in support of ANP-10300P. The NRC staff finds the conclusions of that review generally remain valid, insofar as they are applicable to ANP-10332P. ['''''''''''''''''' '' '''''''''''''''' ''''''''''''

''''''''''''''' ''''' ''''''''''''''''' '''''45 '''' '''''' '''''''''''' '''''''''''''''''''' ''''''' ''''''''''''''''''''' ''''

''''''''''''''''' ]

45 Note that inconsistent pressure ranges are specified in ANP-10332P; Framatomes response to RAI 103 clarifies that [''''''''''''''''' ''''] is the correct range.

3.4.3.3 Allis-Chalmers Large-Diameter Void Tests As discussed in Section 7.3.6 of ANP-10332P, Framatome assessed S-RELAP5 against 162 void fraction datapoints from tests performed by Allis-Chalmers in the mid-1960s using circular pipes of 2.9, 18, and 36 inches in diameter. ANP-10332P presents a series of plots comparing S-RELAP5 predictions against void fractions reported for these tests. The pressures examined in the tests ranged from 615 to 2015 psia, with a maximum measured void fraction of 0.69.

Framatome stated that these tests are used to assess the following PIRT items:

• [''''''' '''''' ''''''''''''''' '''' ''''''''''''''''''' '''''''''''''''''''''''' '''''''''']

• [''''''''' '''''''''''''' ''''''' '''''''''''''''''''' '''''''''''''''''''''''''' '''''''''''' ''''''''''']

• ['''''''''' '''''''''''''' '''''''''''''''''''' ''''''''''''' '''''''''''''''''''''''']

Framatomes assessment of the Allis-Chalmers testing was originally performed in support of ANP-10300P. As noted in Section 3.3.1.3 of the NRC staffs SE on ANP-10300P (Reference 8),

while some direct measurements were made in the 2.9-inch test setup, the 18- and 36-inch tests inferred the void fraction using a homogeneous two-phase flow model and measured pressure drops. The NRC staff finds that the conclusions from its SE on ANP-10300P generally remain valid, insofar as they are applicable to the present review of ANP-10332P. [''''''''''''''''

'''''''''''''''''''''' ''''''''''''''''' '''''''''''''''''''' ''' ]

3.4.3.4 GE Level Swell Test 1004-3 As discussed in Section 7.3.7 of ANP-10332P, Framatome assessed S-RELAP5 against void fraction measurements taken during GE Level Swell Test 1004-3. The test involved the blowdown of a pressurized vessel containing saturated water at a pressure of 1011 psia. This experiment was performed in the smaller of two test vessels, which was 14 feet in height and 1 foot in diameter.

Framatome stated that this test is used to assess the following PIRT items:

• [''''''''' '''''' '''''''''''''''' ''' '''''''''''''''''''''''' '''''''''''''''''''''''' ''''''''']

• [''''''''' '''''''''''''''' ''''''' ''''''''''''''''''''' ''''''''''''''''''''''''' '''''''''''' ''''''''']

• [''''''''' ''''''''''''' ''''''''''''''''''''''' '''''''''''' ''''''''''''''''''''''']

• ['''''''''' ''''''''''''''''' '''''''''''''''']

• ['''''''''''''''''''''''' ''''''' '''''''''''''''']

• [''''''''''''''''''''''''''' '''''' ''''''''''''''''''''' ''''''''''''''''''''''' '''''''''']46 Framatome stated that the assessment of GE Level Swell Test 1004-3 was originally performed for the AURORA-B AOO evaluation model (Reference 7) and remains applicable for the 46 Note that there is an inconsistency in ANP-10332P; [''' ''''''' '''''''''] items are included in Section 5.4.12 but not Section 7.3.7.

AURORA-B LOCA evaluation model. While the NRC staff generally agreed with this statement for the portion of the LOCA event covered by the test data, the NRC staff observed that ANP-10332P presents comparisons to data only through 100 seconds. At this time, the test was not complete, and the NRC staff estimated that the vessel pressure had only depressurized to [''''''''''''''''''''''''''''''' '''' '''']. In RAI 103, the NRC staff requested that Framatome extend its assessment of this test to validate more fully the range of pressures experienced during postulated LOCA events.

In response, Framatome stated that ['' ''''''''''''''''''' ''''' ''''''' '''' ''''''''''''''' '''''''''''''

''''''''''''''''''''''']; however, Framatome included additional plots of the void fraction distribution at 140 and 180 seconds, ['' '''''''' '''''''' '' ''' '''' ''''''''''''''''''' '''' '''''''''''''' ''''' '''''' '''''''''''].

The level of agreement in the void fraction comparisons in the RAI response (at 140 and 180 seconds) was slightly reduced but overall comparable to that seen in similar plots (at 40 and 100 seconds) presented ANP-10332P. In particular, [''' '''''''''''''''''''''''' '''''''''''''' '''''

''' '''''''''''''''''' '' ''''''''''''''''''' ''''''' '''' '''' '''''']; therefore, the NRC staff categorized the S-RELAP5 predictions as being in reasonable agreement to the data.

The additional data in response to RAI 103 [''''' ' ''''''''''''''''' '''''''' ''''''' '''''''''''''''

]. ]. Although the expanded pressure range also does not fully cover the possible range of pressures during the LOCA event, in the NRC staffs judgment, it does substantially characterize the regime where level swell in both the reactor core and lower plenum are most significant. Evidence supporting this statement can be seen from examining plots of the demonstration cases included in Section 7.7 of ANP-10332P. In particular, ['' '''''''''' ''''''''

'''''''''''''''''''''''''''''' ''''' '']. Therefore, the NRC staff found the application of this test to the BWR LOCA event appropriate.

The NRC staff further observed a significant scaling difference that was not recognized in ANP-10332P; namely, the hydraulic diameter of the test facility (i.e., approximately 1 foot) is significantly greater than that of the flow area through the fuel channels that comprise the core of an operating BWR (e.g., approximately 0.5 inch). In general, flowpath geometry influences two-phase flow patterns, and, hence, level swell. As a result, caution should be applied when considering the application of GE level swell test comparisons to the assessment of two-phase level within the reactor core.

3.4.3.5 THTF Mixture Level Tests As discussed in Section 7.3.8 of ANP-10332P, Framatome assessed S-RELAP5 against void fraction and temperature measurements taken from six tests performed at the Thermal-Hydraulic Test Facility (THTF). The tests considered in the assessment included the 3.09 Series 10 Tests I, J, K, M, N, and DD. Intended to simulate PWR conditions, these tests used an electrically heated 8x8 simulated fuel bundle [''''' ''''''''' '''' ''''''''''''''''' ''''''''''''''''''''''

''''''' ''' ''''''''''''''' ''''' '''''''' ''''''''' '' ''''''''''''''''' ' ''''' '''''''']. The pressures considered in the testing ranged from 581 to 1173 psia. Framatome stated that the test series investigates conditions where a two-phase mixture covers about 70-80 percent of the test bundle.

Framatomes assessment of these tests compared measured and predicted values of void fraction and heater rod temperature.

Framatome stated that these tests are used to assess the following PIRT items:

• [''''''''' ''''''' '''''''''''''''' ''''' '''''''''''''''''''' ''''''''''''''''''''' '''''''''']

• [''''''' ''''''''''''''''''''''' '''''''''' ''''''''''''''''''''' '''' ''''''''''''' '''''''''''''''''''']

• ['''''''' ''''''''''''''''''''''''']

• ['''''''' ''''' '''''''''''']

• ['''''''' '''''''''' '''''''''''''']

• [''''''' ''''''''''''''' '''''' ''''''''''''''']

Framatome concluded that predictions of void fraction are in excellent agreement with the data, with the exception of Test K, which Framatome suggested may be affected by a measurement inconsistency. Framatome concluded that the predicted temperatures generally follow the trend of the data, showing overall conservatism.

The NRC staff found the results of most comparisons to be in the reasonable-to-excellent range.

However, based upon the information provided in ANP-10332P, it was not clear to what extent

[''' ''''''''''' '' '''''''''''''''''''''''''' '''''''''''''''''''''''''''''''' ''''''''' ''''''''''''' '' ''''''''''''''''' ''''' ''''''''''''

'''''''''''''' '''''''' ' '' '''' ']. Note that this observation generally applies to all separate effects tests in ANP-10332P involving post-dryout heat transfer. Nor was a clear rationale provided by Framatome regarding the basis for ['''''''''''''''''''''''' ''''''''''''''''''''''''' ' ''''''' ''''''' ''''''''''''''''' ' '''

'''''''''' ''''''''''' ''''' '''''']. In addition, ['' '''''''''''''''''' '''''''' '''''''''''''''''''''' ' ''' '''''''''] does not encompass the full range expected during the BWR LOCA event. Furthermore, the NRC staff made several general observations on these tests:

• [''''''''''''''''''''''''''' '''''''''''''''''''''' ''' '''''''' '''''''''''' '''''' ''''' ''''''' '''''''''''''''''''' '''''''''''''''''''

'''''''' '''''''' ''''''' ''''''''''''''''''' '''''''''''''''' ''''''' '''''''''''''''''' '''''''''' '''''' ''''''' '''''''''''''']

• ['''' '''''' '''''' ''''''''''''''''''''''' ''''''''''''''''''''''''''' '''''''''''''''''''''' '''' '''''''''''''''''''' '''''''''''''''' '

''''''''''''''' ' '' ''''''''''''''''' ''''''''''''''''''' ' '''''' ' ''''''''''''' ' ''''''''''''''' ''''''''''''' ' '''''''''''''''' ]

• [''''''''' ''' ''''''''''''''''''''' '''''' ''''''''''''''' ''''''''''''''''' '' ''''''''''''''''''''''''' '''''''''''''''''' ''''

''''''''''''''''''' ''''' ''''''''''''''''''' ' '''''''''''''''''''']

Framatome briefly addressed the first two observations in response to RAIs 66 and 67.

However, the responses did not contain new information sufficient to revise the NRC staffs assessment. No information was provided in these RAI responses concerning the modeling of

['' ''''''''''' ''''''' ''' '''''''''' '''''''''''''' ''''''' '''''''']47 47 Framatomes updated response to RAI 60 discusses this topic further, as described below in Section 3.6.2.1

While Framatome did not explicitly list phenomena associated with [''''''''''' ''''''''] in its PIRT, the NRC staffs audit of the demonstration cases in ANP-10332P found that ['''''''''''' '''

'''''''''''''''] has significant potential to affect figures of merit in calculations performed using the AURORA-B LOCA evaluation model (see Section 3.6.2.1 below). However, the NRC staff judged the assessment and validation [ ''''''''' ''' '''''''''''''''] in the THTF mixture level tests and other test series included in ANP-10332P as insufficient to support licensees crediting the beneficial effects [ ''''''''''' ''''']. Therefore, as discussed further below in Section 3.6.2.1, the NRC staff imposed Limitation and Condition 23 [' '''''''' ''''''''''''''' ''''''''''''''''' '''' ''

'''''''''''''''''''''''' '''''''''''''''' '''' ''''''''''''''''' ']

3.4.3.6 TLTA Boiloff Tests The TLTA was originally designed for simulating phenomena occurring during the blowdown phase of a large-break LOCA. However, the facility was subsequently modified on numerous occasions to address additional phenomena, including those associated with (1) the refill and reflood phases of a large-break LOCA and (2) small-break LOCA scenarios. TLTA represented a 1:624 scaled-down model of a BWR/6 reactor using a single, electrically heated fuel bundle.

As discussed in Section 7.3.9 of ANP-10332P, Framatome simulated two boiloff tests, Tests 6441-6 and 6441-7. These tests were performed at pressures of 400 and 800 psia, respectively. Comparisons were made for void fraction at various locations in the test facility (e.g., fuel bundle, bypass region, downcomer) and heater rod temperature.

Framatome stated that these tests are used to assess the following PIRT items:

• ['''''''' ''''''' ''''''''''''''' ''' ''''''''''''''''''''''''' ''''''''''''''''''''''''' '''''''']

• ['''''''' ''''''''''''''''''''''' ''''''''' '''''''''''''''''''''' '''' ''''''''''''' ''''''''''''''''']

• [''''''' '''''''''''''''''''''''']

• [''''''' ''''' '''''''''']

• ['''''''' ''''''''''' '''''''''''']

• ['''''''' '''''''''''''''' '''''' '''''''''''''']

As discussed in RAI 70 and its response, [' ''''''' ''''' '''''' ''''''''' ''''' ''''''''''''''''' ' ' '''''''''''

'''''''''''''''' ''''''''''''''''' ''''' ''''''''' '''''''' ''''''''''''''''''] The response to RAI 70 addressed this issue, showing improved agreement after ['''''''''''''''''''''''''''''''' '' ''''''''''''''''''''' ' ''''' '''''''''' '''''''

''''''' '''''''''''''''''''''''' '''''''''''''''''''''''''''' ''''''' ' ''''''''''''''''''''''''''''''''] The NRC staff found this approach reasonable, [''''''''''''''''' ' '''''''''' ''''''''''''''' '''' ''''''' '''''''' '''''''''''''''' '''''' '' '''''''''

''''''''' ''''''].

From the information provided in ANP-10332P, only the [''''' ''''''''''' '''''''''''''''''''''''''''] could be directly compared to measured data on an individual basis. Framatome presented no evidence to support a direct, individual assessment of the other listed PIRT items. For example,

[''''''''''''''''' '''''''''' ' ''''''' ''''''''''' '''' '' '''''''''''''''' '''' ''' '''''''''''''''''' '''''''''] considered in the TLTA boiloff tests does not fully envelop postulated BWR LOCA events. The NRC staff also

['''''''''''''''''' ''''' '''''' '''''' '''''''' ''''''''''' ''''' '''' ''''' ''''' ''''''''''''''''' ''' '''' '''

''''''''''''''''''' ' ''''''] That being said, the NRC staff generally found the S-RELAP5 predictions of void fraction and heater rod temperature for the assessed TLTA boiloff tests to be in reasonable overall agreement with measured data over the range of comparison, after accounting for [''

''''' ''''''' ''''''''' ''''''''''].

3.4.3.7 Bennett Tube Tests As discussed in Section 7.3.10 of ANP-10332P, the Bennett tube tests involved a series of experiments designed to investigate both CHF and post-dryout heat transfer. The tests involved injection of subcooled water into a heated tube pressurized to approximately 1000 psia. The vertically oriented tube was approximately 0.5 inch in diameter and had a total length of 19 feet (i.e., 228 inches). The heated length could be varied; in the tests Framatome selected for assessment [''''' '''''''' '''''''' ''' ''''''''' ''''''''''''' '''''''''' ' ''''' '''''''''' ''''' ''''''''' ''''''''''''''''

''''' ' ''''''' ''''''''''''''''''''''' '''''''''''' ''''''''''' '''''' ''''' '''''''''''' '''''''' ''' ''''''']. A rationale for this decision, and its impact on scaling to the prototypical plant condition was not included in ANP-10332P. The tests Framatome evaluated considered low-flow and high-flow conditions.

['''''''''' '''''''''''''''''''''''''' '' '' ''''''''' ''' ''''' ''''' ''''''' '' ''''''''''' '''''''''' ''''

'''''''''''''''' ''''''''''']

Framatome stated that these tests are used to assess the following PIRT items:

• [''''''''' '''''' ''''''''''']

• [''''''''' '''''''''''' ''''''''''']

• [''''''' ''''''''''''''''''' ''''''' ''''''''''''''']

• [''''''' ''''''''' '''''''''']

Of these PIRT items, the NRC staff could [''''''''''' ''''''''''' ''''' '''''''' ''''''''''] on an individual basis against the data that was presented in ANP-10332P. Furthermore, the simplified geometry of the Bennett tests does not directly represent heat transfer conditions in a rod bundle geometry. Also, [''' ''''' ''''''''''''''' ' '''''' ''''' ' '''''''''''''''''' '' '''''''''''''''''''''''''''''' '

'' ''''''''''''''''' ''''''''''''''''''''''''''''''''' '''''''] That being said, the NRC staff found the predictions of S-RELAP5 shown in ANP-10332P (Reference 1) to be in reasonable to excellent agreement with the test data. In particular, the high-flow test prediction was somewhat conservative, but still in reasonable agreement with the data. However, as discussed in Framatomes updated response to RAI 16 (Reference 66), the information originally submitted in ANP-10332P was subsequently found to be in error. Framatome stated that correction of the error [''''''''''''''''''

''''''''''''''''''''''''' ''''''''''''''''''' '''''''''''''''''''' '' ''''' '''''''''''''''']. Framatome committed to revising the approved version of ANP-10332P to provide corrected information. As a result of [''''

'''''''''''''''''''''''''' '''''''''''''''' ''' '''''], as well as the lack of correspondence of the test conditions to the BWR LOCA event, the error correction did not fundamentally change the NRC staffs perception of this testing.

3.4.3.8 THTF Steady-State Film Boiling Tests As discussed in Section 7.3.11 of ANP-10332P, Framatome used [''' '''''''' '''''''''''''''''''''' ''''

''''''''' ''''''' ''''''''''' ''''''''''''''''''''''' '''''' ''''''''']. In fact, the vendor used these tests ['''''''''''''''''

''''''''''''''''''''''''''' '''' '''''''''' '''''' ''''' ] Additional data from the remaining tests was used in the validation; ['''''''''''''''' ''''''' ''''''' ''''''''''''''''' ''''''''' '''''''' '''' ''''' ''''''''''''''' '''''

''''''''''''''''''''''''''''' '' ''''' ''']. Framatome also stated [''''' ''''''''''''''''''''''''''' '''''

'''''''''''''''''' ' ''' ''''' '''''''''''' ''''''''''''''''' ''''''''''''''''''''''' ' '''''''''''''''''' '''''''' '''''' '''''''''''''''''''''''''']

Framatome stated that these tests are used to assess the following PIRT items:

• [''''''''' ''''' ''''''''''''' '''' ''''''''''''''''''''''' '''''''''''''''''''''' ''''''''']

• [''''''''' '''''''''''''''''' '''''''''' ''''''''''''''''''''''''' ''' '''''''''' '''''''''''''''''']

• [''''''' '''''''''''''''''''''']

• ['''''''' ''''' ''''''''''']

• [''''''''' '''''''''' ''''''''''']

From the NRC staffs perspective, Section 7.3.11 of ANP-10332P ['''''''''''''''' ' ''''''''' '''''''''''''''''

''''''''''''''''''' ''''''''' ''''''''''''''''''''''''' ''''''''''' ' '' ''''''' ''''''''''''''''' ' '' ''''''''''''''''].

The NRC staff requested additional information ['''''''''''''''''''''' ''''''''''''''''''''''''''''' ''''''''''''''''''' '' '''

'''''''''''''''' ''''''''' ''''''' ' ''''''''''''''''' ''''' '''''''''''''''''' ''] In RAI 119, the NRC staff requested that Framatome provide Reference 84 to ANP-10332P, which describes additional steady-state film boiling data from the Combustion Engineering/Columbia [''''' ''''''''''''''''' ''''' '''''''''''''''''''''''

''''''''' ''''''''''''''' ''''''' ''''''''''''''''''''''''' '''''' ''''''''''''''' '''''''''''''''''''' ''''' ''''''''''].

In response to RAI 105, Framatome confirmed ['''' ''''''''''''''''''''''''''' ''''''''''''' ''''''' '''''' ''''''''''

'''''''''''''''''''''''''''''''' ' ''''''''']

While the above information generally appeared reasonable, the NRC staffs review of Figure R105-1 in Framatomes response to RAI 105 identified additional considerations that had not been addressed. [' '''''''''''''''' '''''''' ' ''''''''' ''''''''''' ' '''''''''' '''''''' ' ''''''''''''''''''' ' ''''' '''

'''''''''' ] However, the NRC staff did not find the explanation fully satisfactory.

While the NRC staff could not locate measured ['''''''''' '''''''''''''''''''''''''''' ' ''' '''''''' ''''''''''''''''''''

''''''''''''''''' '''''''''''''''''' '''' '''''''''''''''''' ''' '''''' ' ''''''''''' '''''''''''''' '''''' ''' ''''''''''' ]

In its updated response to RAI 105 dated October 31, 2018 (Reference 66), Framatome revised its explanation for the discrepancy between the information presented in ANP-10332P and its response to RAI 105. [''''''''''''''''''''''' '''''''''''''''' ''''' ''''''''''''''''''' '''''''''' ''''' ''' '''''''''' ' '''

''''''''''''''''' ''' ''''' ''''''' ''''''''' ''''' ''''''''''''''''' ''''''''''''''] Therefore, the NRC staff considered Framatomes response to RAI 105 acceptable.

3.4.3.9 FCTF Spray and Steam Cooling Tests As discussed in Section 7.3.12 of ANP-10332P, Framatome assessed S-RELAP5 against tests of heat transfer using a single prototypical ATRIUM 10 rod bundle geometry at nearly atmospheric pressure. The tests were conducted in Framatomes FCTF. Multiple conditions were simulated, including adiabatic heatup, downdraft, updraft, and bypass entrainment. All tests except the adiabatic heatup tests included simulated core spray. Heater rod temperature comparisons at several axial elevations were provided for five FCTF tests.

Framatome stated that these tests are used to assess the following PIRT items:

• [''''''' '''''''''''''''']

• [''''''' ''''''' '''''''''''''''' '''' '''''''''''''''''''''''' '''''''''''''''''''''''' ''''''''''']

• [''''''' ''''''''''''''''''' '''''''''' '''''''''''''''''''''' '''' '''''''''''' ''''''''''''''''''''']

- 100 -

• [''''''''' ''''''''''''''''''''''''''' '''''' '''''''''''''''' ''''''''''' '' '''''''']

• ['''''''' ''''''''''''''''''''''']

• [''''''' ''''' '''''''''''']

• ['''''''' '''''''''''' '''''''''''''']

• [''''''''' ''''''''''''''' '''''' ''''''''''''']

Based upon the comparisons presented in ANP-10332P, the NRC staff generally found the temperatures predicted by S-RELAP5 to be in reasonable, and in some cases excellent, agreement with the measured data. That being said, ['' ''''''''''''''''''' ''''''''''''' '''' '

''''''''''''''''''''''''''' '' '' '''''''''''''''''''' ''''''''''''''].

The NRC staff observed that Framatome provided an example plot showing the predicted heat transfer mode as a function of time for one FCTF test in response to RAI 72. [''''''''''''''''''

'''''''''''''''''''''' ''''' ''''''' ''''''''']. In particular, the NRC staff asked RAIs concerning the applicability of this testing to [''' ''''''''''''''''''''''' ''''''''''''''' ''''''''' ''' '''''' '''''' '''''''''''''''''

'''''''''''''''''''''''''''' '' '' '''''''''''''''' ' ''''''''''''''''''''''''''''']:

• Regarding the [''''''''''''' ''''''''''''''''''''''''''''' '''''''''''''''''''' ''''''''' '''''''''''''''''' '''' ' '''''

'''''''''''''''''''''''''' ' ''''''''' '''''''''''''''' '''''''' ''''''''''''''''''''''''''''''''''''' '''''''''''''''''''''']

• Regarding the [''''' '''''''''''' '''''''''''''''''''' ''''''''''''''''''''''''''''' ''''''''''''''''''''''''' '''''''''' '''''''''''''''

''''''''''''''' '' ''''' ''''''''''''']. The NRC staff found the connection tenuous and noted that the response neglects the larger point that assessment and validation are intended to provide quantitative comparisons of predictions against measured data. In particular, because ['''''''''''''''''''''' ''''''''' ' '''''''''''''' '''''''''''''' '''''''''''''' ''''''' '''''''''''''''''''''''' '''''

'''''''''''''''''' ''''''''''''''''''''' '' ''''''''' ''' '''''''''' '''''''''''], the NRC staff could not accept this explanation.

In RAI 111, the NRC staff questioned (1) Framatomes modeling of the water channels incorporated into many modern BWR fuel assembly designs, and (2) whether any assessment and validation cases included water channels. [''''''''''''''''''''''''' '''''''''''''''' ''''' '''''''' ' ''' ''''''

'''''''''''''''''''''''''' '''''''''''''' ''''''''''''] The NRC staff found this approach acceptable.

Framatome stated that ['''''''' ' ' '''''''''' ' ''''''' ''''''''''''''' ''''''' '''''''' '''''' '' '''' ''''

- 101 -

''''' '''''''''''''''']. In general, the NRC staff does not agree with this statement, [''''''' ''''''''''''''''''''

''''''''' ' '''''''''''''''''' ' '''''''''''''' '''''''''''''''''''''' ' '''''''''''''''''''''' '''''' '''''''''''''''''''] However, [ ''''''

'''' ''''''' '''' ''''' '' ''''''' ''''''''''''''' '''''''''''] (Limitation and Condition 8) and the need [''

'''''''''''''''] (Limitation and Condition 27), the NRC staff considered the issue adequately addressed.

In response to RAI 47, Framatome concluded that FCTF Test 79 could be used to validate an additional phenomenon, [''''''''''''' ''''''' ''''''''''''''''' '''''''''''' '''''''''''''''' '''''''''''''''' ''''''''''''''''''''''''

''''''''''''''''''''''''' '''''''''''''''' ''''''''''] However, the NRC staff did not agree with Framatomes conclusion for several key reasons. ['''''' '''''''''' '''''''''''''''' '''''''''''''' ' '''''' ''''''' '''''''''''''''' '

' ''''''''''''''''''''' '''''''''''''''''' '''' '' '''' '''''''']

Finally, the NRC staff notes several additional important observations ['''''''''''''''''''''' ''

''''''''''''''' ''''' ''''''' '''''''''''''''''' ''''' ' '''''''''''''''''''''''''' ''''''''''''' ''''''' '''' '''''''''''''''''''']:

• [''' ''''''''''' ''''''' '''''''''''''''' ' ''''''''' ''''''''''''''''' '''''''''''''''' ''''''''''' '''''' ''''' '''''''

''''''''''' ''''''''']

• [''' ''''''''''''''''''' ''' ''''' ''''''''''''''' '''''''''''''' '''''' '''''''' ' ''''''' ''' ''''''''''''' ''''''''''''''''' '

'''''''''''''''''''''''''''''''' '''''''''''''''''''''''''' '''''' ''''''''' ''''''''''''' ''''''''''' ''''' '''''''' ''''''''''' ''''''''''''']

• ['''''''''''''''''''' ''''''''''''''''''' '''' ''''''''''''''' ' '' ''''' '''''''''''' ''''''' '''''''''' '''' ''''''''''''''''''''

''''''''''''''' ''''''''' ''''''''' '''''']

• [''''' '' '''''''''' '''' ''''' '''''''' '''''' '' ''''' ''''''''' ''' ''''' ''' ''''''''''''']

['''''''''' ''''''''''''''''''''' ''''''''''''''''''' '''''''''''' '''' ''''''''''' '' '''''''' '''' '''''''''' ' ''''''''''''''''''''''''''

''''''''''''' '''''''' ''' '''''''' '''''''''' ''''''''''''' ] consideration of these issues contributed to the NRC staffs decision to impose Limitations and Conditions 8 [''''' ''''''''''''''''' '''''''''''''''''' '''

'''''''''''''''''''''''' '''''''''''''''''' ''''''''''''''''' ''''''''''''''' '' '''''' '' '''''''' '''''''''''''''' '''''''''''''''''''''''''] and 15

['''''''''''''''''''''''''''''' ''' '''''''''''''''''''''' ''''''''''''''''''' '' '''''' '''''''''''''''''''' '' '' ''''] in Section 5.0 of this SE.

- 102 -

3.4.3.10 THTF Reflood Tests As discussed in Section 7.3.13 of ANP-10332P, Framatome assessed S-RELAP5 against data from 11 THTF reflood tests. Framatome stated that these experiments simulated the reflooding of a test bundle with an initial water level at about 25-30 percent of the bundle height. Key parameters for each of these tests, including pressure, reflood rate, and linear heat generation rate, are summarized in Table 7-9 of ANP-10332P.

Framatome stated that these tests are used to assess the following PIRT items:

• [''''''''' '''''''''''''''']

• [''''''''' ''''''' '''''''''''''''' '''' ''''''''''''''''''' ''''''''''''''''''''' ''''''''''']

• [''''''''' '''''''''''''''''''''' '''''''' ''''''''''''''''''''''''' ''''' ''''''''''''' ''''''''''''''''']

• [''''''' '''''''''''''''''''''''''']

• [''''''' ''''' ''''''''''''']

• [''''''''' ''''''''' '''''''''''''']

• [''''''''' ''''''' '''''''''']

• [''''''' ''''''''''''' ''''''''' '' '''''' ''''''''']

• [''''''''' '''''''''''''''''' '''''' ''''''''''''''''']

• ['''''''' '''''''''' '''''''''''''''''''''''''''' ''''''''''']

Framatome provided a series of plots comparing S-RELAP5 predictions of the temperature at the highest measurement elevation in the bundle against test data. Framatome concluded that the predicted heater rod temperatures are generally consistent with or greater than the measured data.

Of the PIRT items Framatome associated with the THTF reflood tests, the NRC staff found that several were not relevant or could not be directly assessed using the data presented in ANP-10332P [''''''' '''''''''''''''' ''''' '''''''''''''' ''''' '''''''''''''''''''''' '''''''''''''''''''' ''''''''''''''''''''''

''''''''''''''''''''''''''' '''' ']. Regarding the data that was presented, while the influence of individual heat transfer regimes generally was not apparent, the NRC staff found that S-RELAP5 tended to predict peak heater rod temperatures in the THTF reflood tests with reasonable to excellent accuracy. However, the time to quench was generally overpredicted.

The THTF reflood tests were performed at [''''''''''''''''''' ''''''''''''''' '''''' '''''''''''''''''''''''''''''

''''''''''''''' '' ''' ''''''' '''''''''' '48 '''''''''' ''''''''''''''''' ' ' '''''''' ' '''''''''''''''''' ' ''

''''''''''''''''''''''''''''''' '''''''' ' ''] Consequently, while several assessed tests in this series appear partially relevant to BWR small-break LOCA scenarios of concern, the NRC staff considered a number of the assessed tests [''''''''''''''''''' ''''''''''' '''''''''''''''''''] of limited applicability.

48 In particular, limiting small-break LOCA scenarios for operating BWRs typically involve failure of the high-pressure coolant injection (BWR/3-4) or high-pressure core spray (BWR/5-6) systems, necessitating activation of the automatic depressurization system.

- 103 -

In response to RAIs 104 and 125, ['''''''''''''''''''''''' '''''''''''''''''' '''' ''' '''''''''''''''''' '''''' ''''' '''''''''

'''''''''''''''''''''''''' '''''''''' ' '''''''''' '''' '''''''''''''''''''' ''''''''' ''''''''''''''''''''''''' ''''''''''' ''''''''''' ' ]

3.4.3.11 FLECHT Reflood and Steam Cooling Tests As discussed in Section 7.3.14 of ANP-10332P, Framatome assessed S-RELAP5 against data from 10 tests in the FLECHT and FLECHT-SEASET series. The FLECHT and FLECHT-SEASET tests simulated the reflood phase of a PWR LOCA event. As such, they modeled bottom-up reflood and quenching on simulated Westinghouse-type fuel bundles of 15x15 and 17x17 design. Framatome selected tests from these series that included both cosine-shaped and upskewed axial profiles.

Framatome stated that these tests are used to assess the following PIRT items:

• ['''''''' '''''''''''''''']

• [''''''''' ''''''' '''''''''''''''' '''' ''''''''''''''''''''''' ''''''''''''''''''''''''' '''''''']

• ['''''''' ''''''''''''''''''' '''''''''' ''''''''''''''''''''''' '''' '''''''''''' ''''''''''''''''''''']

• [''''''''' ''''''''''''''''''''''']

• [''''''' ''''' ''''''''''']

• ['''''''' ''''''''''' ''''''''''']

• [''''''''' ''''''''' ''''''''''']

• [''''''''' ''''''''''''''' ''''''''' ' ''''' '''''''']

• ['''''''' '''''''''''''''''' '''''''' '''''''''''''']

• ['''''''' ''''''''' '''''''''''''''''''''''''''' '''''''''''']

Framatome presented a series of plots comparing S-RELAP5 predictions of heater rod temperatures against measured data.

The NRC staff found [''''' ''' ' '''''''' ''''''' '''''''''''''''''''''''' '''''''''''''''''''''' ' '''''''''''''''''

''''''''''''''''''' ''''''' '''''' ''' '''''''''''''''''' ''''' ''''''''''''''''''''' ''''']

From the NRC staffs perspective, S-RELAP5 tended to predict ['''''' '''''''''' '' '''''''''''''''''''''''''''

'' ''' '''''''''''''''' ''''''' ''''' ''''''''''''''''''''''''''''' '''''' '''''''] with reasonable accuracy. However, ['

'''''''''''''''''''''''''] overprediction of significant magnitude was observed in some of these tests at

[''''''''' ''''''' ''''''''''''''''''' '''' '' ''''' ''''''''''' ''''' ''''''''''''''''' '''''''''''''''''''''''''''''.] Whereas, for

['' ''''''''''''''''''''' ''''' '''''''''''], the NRC staff found that, while conservative, the S-RELAP5 predictions of [''''''' '''''''''''''' ''''''''''''''''''''''] were in minimal agreement with the data. ['

'''''''''''''''''''' ''''' ''''''''''''''''' '''' ]. Framatome did not provide adequate justification for [''

''''''''' ''''''''''''''''''''''' '''''''''''''''''''''' ''''''''''''''''''''' '''''''''''''' '' ''''''''''''''''''' ''''''''''' ''''''], and this topic became the subject of RAI 74.

In response to RAI 74, Framatome [''''''''''''''''''' ''''''''''''''''''' '' '''''''''''''''''''''' ''''''''''''''''''''''''''' '

- 104 -

'' ''''''''''''' '' ''''''''''''''''''''''''' '''''''''' ']

In assessing the response to RAI 74, the NRC staff consulted a previous review of Framatomes Realistic Large Break LOCA methodology for PWRs, as documented in the approved version of TR EMF-2103, Revision 3 (Reference 54). [ '''''''''''''''' ''''' ''''''' '''' ''''''''''

''''''''' ''''' '''''''''''''''''' ''' ''''''''''''''''''''''''''''' '''' '''' '' '''''''''''''''''''' '''''''''' ' '''''''''' .]

Figure 3 reveals that, ['' '' '''''''''''''''''' ''''''' '''''''''''''' ''''''''''''''''''''' '''''''' ''''''''''''''''' ' '''

'''''''''''''''''' '''''''''' '''''''''' '' ' ''''''' '' ']

In conclusion, the NRC staff found that Framatome did not identify the root cause of the minimal agreement in the S-RELAP5 comparisons against some of the FLECHT and FLECHT-SEASET tests, ['''''''''''''''''' ''''''''' '''''' ' ''''''''''''''''''''' ''''' '''''''''' ''''''''''' '''''''''''''''' '' ''''' ' ''''

''''''''''''''' ''''''''' ''''''''''''''], the NRC staff ultimately found these assessments satisfactory for an Appendix K-based LOCA evaluation model.

- 105 -

Figure 3: Illustration of the Impact of [''''''''''''''' ''''''''''' '''''''''''

''''''''''''''''''''''''''] Realistic Large Break LOCA Methodology for PWRs 3.4.3.12 CCTF Reflood Tests As discussed in Section 7.3.15 of ANP-10332P, Framatome assessed S-RELAP5 against four experiments performed in the CCTF. The CCTF was a 1:21 scaled model of a four-loop PWR used to simulate the reflood phase of a large-break LOCA. Framatome performed its assessment against Tests 54, 62, 67, and 68, presenting a series of plots comparing S-RELAP5 predictions of heater rod temperatures against measured data.

Framatome stated that the above tests are used to assess the following PIRT items:

• [''''''''' ''''''''''''''''''''']

• [''''''''' '''''' '''''''''']

• ['''''''' '''''''''' ''''''''''''']

• ['''''''' ''''''''''''''' ''''''' '''''' ''''''']

• ['''''''' ''''''''''''''''''' ''''' '''''''''''''''']

• [''''''''' '''''''''''' ''''''''''''''''''''''''' ''''''''']

From the NRC staffs perspective, the significance of most of these PIRT items could not be directly assessed on an individual basis from the information presented in ANP-10332P. In particular, the NRC staff found the information concerning the CCTF presented in Section 7.3.15 of ANP-10332P to be less detailed as compared to other presented assessments; for example, a nodalization diagram was not provided, nor was a detailed

- 106 -

discussion of the test procedure and scaling rationale. The general lack of scaling rationale is particularly of note, in light of the fact that this test was designed for PWR conditions.

While keeping the above points in mind, the NRC staff observed that S-RELAP5 generally made reasonable predictions of the overall test behavior. [ ''' ' '' '''' '''''''''''''''''''''''''' ''' '''''''

''''''' ''''''''''' '''''''''''''''''''''''' ''''''' ' '''''''''''' ]

However, a nonconservative aspect of the CCTF comparison [ '''' '' ''''''''''''''' '''''''

'''''''''''''''''''''''''''' ']49 This predictive error was neither justified nor acknowledged in ANP-10332P. Although the prediction of [''''''''' '''''''''''''''''''''''] is not expected to be a dominant factor in determining the calculated figures of merit, the NRC staff considered the CCTF assessment as reducing confidence in both the accuracy and conservatism of S-RELAP5-predicted [''''''' '''''''''''''''''''''''''']. However, in light of the limited impact on relevant figures of merit expected from the misprediction [' '''''''''''''''], along with the improved predictions of [''

''''''' ''''''''''''''''''''''] observed in other test series against which S-RELAP5 was validated, the NRC staff did not pursue the issue further.

3.4.3.13 Marviken Critical Flow Tests As discussed in Section 7.3.16 of ANP-10332P, Framatome assessed S-RELAP5 against measurements of choked flow taken at the Marviken test facility. The Marviken facility was originally constructed to operate as a prototype heavy-water BWR; however, prior to reaching operation, the plant was converted to a facility for performing reactor-scale choked flow tests. In essence, the reactor vessel served as reservoir of heated, pressurized fluid that was permitted to blow down in a series of tests involving nozzles of various size and geometry.

Framatome made comparisons against measured data from a selection of nine Marviken tests using both the homogeneous equilibrium and Moody critical flow models within S-RELAP5.

Framatome concluded that the homogeneous equilibrium model agrees well with the measured data, while the Moody model makes conservative predictions.

Framatome stated that these tests are used to assess the following PIRT items:

• ['''' '''''''''' '''''''''' '''''' '' '''''''' ''''''''''''']

• [''''''''''''''''''''''''''''' ''''''' ''''''''''''' '''''''' '''''''''']

• ['''''''''''''''''''''''' ''''''' '''''''''''''''' '''''' ' '''''''''']

The AURORA-B LOCA evaluation model uses ['' '''''''''''''''''''''''''''' ''''''''''''''''''''' '''''''''' '''''' '

49 Based upon the information submitted by Framatome, [ ''' '''''''''''''''''' ''''''' ''''''''''''''' ''''

''''''''''''''''''''' '''''''''''''''''''' '''''''''''' ''''''''' ''''''''''''''' ''''''''''''''''']

- 107 -

'''''''''''''''''''''''''' '''''''''''''''' ''''''''''''''''''''''''''''' '''' ''''''''''''' ''''''''''''' '''''' ' ''''''''''''''''' ' ''' ']

Nevertheless, because [' ''''''''' '' ''''''''''''''''''''''''' '''''''' ''' ''''''''''''''''''''''''' '''''''''''''''''''''''

'''''''' '' ''''''''''''''''''''''''' ''''''''''' ' ''' ''''''''''''''''''''''''''''''' ''''''''''''''''''''''' ''''''''''''']

Regarding the Moody model, the NRC staff observed ['''' '' ''' '''''''''' ''''''''' ''''''''''

'''''''''''''''''''''' ' '''''''''''''''' ''''''''''' ' '''''''''''''''''''''''''' ''''''''''''''''' ' ' ''''' '']

In response to RAI 47, Framatome stated that the Marviken assessment can be considered to validate ['''' ''''''''''''''''''' ''''''''''''''''''''''''''' ''''''''''''''' ''''''''''' '''' ''''' '''''''''''''''''' '''''''''''''''''''''''''''''''

''''''''''''''' '''''''''''' ''''''''''' ''''''' '''''''''''''''' ''''''' '''''''''']. However, the NRC staff did not agree with Framatomes position [''''' ''' ''''''''''''''''''''''''' '''' ''''''' ''''''''''''''''' '' ''''''''''''''''''

''''' '''''' '''''''''' ''''' '' ''''''''''''''''''' '''' '''''''''''''''''' ''''' '''''''''''''''''''''' '''''''' ''''''' '''''''''''''']

3.4.3.14 UPTF Countercurrent Flow and Entrainment Tests As discussed in Section 7.3.17 of ANP-10332P, Framatome assessed S-RELAP5 against data from two tests performed at the Upper Plenum Test Facility (UPTF). Although the UPTF was designed to simulate a German four-loop PWR, Framatome concluded that several specific tests from this facility are applicable to the BWR LOCA event, including those that assess the capability of predicting countercurrent flow and liquid entrainment.

In particular, Framatome stated that these tests are used to assess the following PIRT items:

• ['''''''' ''''''''''''''''''''''''''' '''''' '''''''''''''''''' '''''''''' '' '''''''']

• [''''''''''''''''''''''''''' ''''' '''''''''''''''''''''''' '''''''''''''''''''''''''' ''''''''''']

The NRC staff does not agree that the UPTF test configuration is applicable to BWR conditions and did not perform a detailed assessment of the comparisons of S-RELAP5 to the UPTF data.

In particular, the NRC staff concluded that predictions of ['''''''''''''''''''''''''' '''' '''' '''''

''''''''''''''''''''''] are sufficiently geometry-dependent that the agreement or disagreement of S-RELAP5 with UPTF data would neither add nor detract from the assessment of the AURORA-B LOCA evaluation model. The NRC staffs concerns regarding applicability of UPTF test data to the assessment of the AURORA-B LOCA evaluation model are noted in RAI 77, to which Framatome provided a brief response that did not substantially alter the NRC staffs conclusion.

3.4.4 Integral Effects Tests Validation of an evaluation model against integral effects test data is intended to assure that the evaluation model is capable of representing the full set of phenomena associated with a given event, with particular emphasis on the complex, multifarious interactions between them during

- 108 -

the scenario of interest. As discussed in Section 7.7 of ANP-10332P, Framatome performed integral effects test comparisons against data from three facilities, namely

• the TLTA

• the Full Integral Simulation Test (FIST) facility

• the SSTF Comparisons of S-RELAP5 code predictions to testing from each of these facilities will be discussed in succeeding sections of this SE.

Framatome listed ['' ' '''''''''''''''' ''''''' ''''''''' '''''''' '''''''''''''''''''' ''''''''''''''' ' ''''''

'''''''''' ]. From the NRC staffs perspective, as alluded to above, it is essential to distinguish validation from the mere presence of a particular phenomenon in a test. In particular, validation of a phenomenon implies quantitative comparison with sufficient detail to provide a clear picture of the accuracy of the predictions of specific code models and correlations relative to measured data characterizing the individual phenomenon of interest. Clearly, such is not the objective of integral testing, and, for the vast majority of the phenomena Framatome associated with each integral test, it is not possible to isolate the performance of one code model or correlation from numerous others, such that meaningful assessment of individual phenomena can be achieved.50 As such, this SE generally will ['' '''''' ''''''''''' ''''''''''''''''' ''''''' ''' '''' ''''''''''

'''''''''''''''''''''''''' '''''''''''''''''' ''''''''''''' ''' ' ]

However, an exception where further discussion [' ''''''''''''''''' '''''''' '' ''''''''' '''''''''' ''''''''''''''''''''''

'''''''''''''''''' 51 ''''''''''''''''''''' '''''' '''''''''''''''''''''''' ''''''''' '''' '''''' ''''''''''''' '''''''''''''

'''''''''''''''''''' '''''''''''' ''''''] As a result, the NRC staff issued RAI 47 to request that Framatome provide justification for any high- and medium-ranked phenomena that are not validated by separate effects tests.

Framatome responded to RAI 47 by briefly summarizing [''''' ''''''''''''' '''''''''''''''' ''''' '''''''

'''''' '''''''''''''''''' '''''' ''''''''''''' ''''''''''' '''''''''']. The NRC staff found Framatomes response not fully satisfactory. While, in some cases, the NRC staff was able to conclude that the additional justification provided in response to RAI 47 could serve as adequate validation, in a number of others, the NRC staff found Framatomes basis insufficient. Details concerning the NRC staffs evaluation of RAI 47 may be found in the technical evaluation report from Brookhaven National Laboratory (Reference 35). The impact of this issue will be discussed further below in the summary of Framatomes assessment and validation effort in Section 3.4.6.

50 Moreover, as discussed in Section 3.4.3, from the information presented in ANP-10332P, [''''''''''''''''''''''''

'''''''''''''''''''' ' ' '''''''''''''''' '''''''' ''''''' ''' ' ''''''''''''''''''''' ''''''' ' '''''''''' ''''''''''''''''' ''''''''''' '''''''].

51 Note that this is ['''' '''''''' '''''''''''''''''''''' '''''' '''''' '''''''''''''''''' ' ''''''''''''''' '''''''''''' ''''''''''].

- 109 -

3.4.4.1 TLTA Integral Tests The TLTA was originally designed for simulating phenomena occurring during the blowdown phase of a large-break LOCA. However, the facility was subsequently modified on numerous occasions to expand its capabilities for modeling additional phenomena, including those associated with (1) the refill and reflood phases of the LOCA event and (2) small-break LOCA scenarios. TLTA represented a 1:624 scaled-down model of a BWR/6 reactor using a single, electrically heated fuel bundle.

Framatome selected two integral tests from the TLTA series for the assessment and validation of the S-RELAP5 code, namely

• Test 6425-2 (Large-Break LOCA)

• Test 6432-1 (Small-Break LOCA)

The latter case (Test 6432-1) resulted in no heatup in both the test and the ['''''''''''''''''''''''''''''

'''''''''''''52 '''' '''' '''''''''''''''' ' '''''''''''''''''''''''''''''''' ''''''''''' '''''''''''''''''''''''' ''''''''''''''''''''''''''''' ''''''''' '''

''''''''' '''''''''''''''''' ''''''''''''] The NRC staffs evaluation of these comparisons is provided below.

3.4.4.1.1 TLTA Large-Break Test 6425-2 Test 6425-2 modeled the double-ended rupture of a recirculation line with two low-pressure coolant injection pumps inoperative. Thus, available ECCS equipment included high-pressure core spray, low-pressure core spray, and one low-pressure coolant injection pump.

Framatome presented the results of its S-RELAP5 simulation of Test 6425-2 in Section 7.7.2.1 of ANP-10332P. While S-RELAP5 calculated some measured parameters acceptably, [''''''''''

''''''''' '''''''''''''''''' ''''''''' ''''' '''''''''''''''''''' '''' '' ''''''''''''''''''''' '''''''''' '' ''''''''], the NRC staff characterized the overall S-RELAP5 prediction as demonstrating insufficient agreement with the measured test data. [''''''''''''''' '''' ''''''''''''''''' '''''' '''''''' ' '' ''''''''''''''''''''''''''' '''''''''''''''''], this judgment is rendered because the observed deviations are excessive and unrealistic. For

example,

• whereas complete heater rod quenching was observed [ '' ''''' '''''''' ''''''''''''''

''''''''''''''''' ''' '' ''''''''' '''''''''''''''''' '''''''''''''''''''' '''' ''''''''''''''''''''''''''' ''''''''' ''''''''''''''], and

• the measured peak cladding temperature of approximately 700 °F was ['''''''''''''''''''''''''''''

' '''''''''''''''''''''''''''' ''' '].

A credible explanation for such significant error is not provided in ANP-10332P. However, from included comparison plots, it is apparent that the S-RELAP5 analysis [''' '' '''''''''''''''''''' ''''''''''

''''' ''''''''''''''''''''''''''' '''' '''''''''''''''''], which hindered the NRC staffs attempts to understand the 52 As noted previously, S-RELAP5 ['''''' ''''''''''''''''' '''''' ' '''' ''''''''''''''''''''''''''''''' '''''''''] for performing assessment and validation analyses.

- 110 -

source of error in the S-RELAP5 prediction. The NRC staff issued RAIs 80 and 106 to request that Framatome explain further its predictions for this test and justify that the poor observed agreement should not undermine confidence in the capability of the AURORA-B LOCA evaluation model.

In its response, Framatome [''''''''''''''''''' ''' '''''''' '' '''''''''''''''''''' '''''''''''''''''' ' '''''''''''''''''''' '

'''' ' '''''''' '' '' '''''''' ' '''''''''''''' ' ''' ''''' '''''''''''']. Framatome stated [''''' ''

''''''''''''''''''''''''''''''''' '''' '''''''''''''''' '' ''''' '''''''''' '''''' '''''' ']. Framatome noted ['''' '''''

''''''''''''''' ''''''''''' '' ''''''''''''''''' '''''' ''' '''''' ''''''''''' ''''' '''''''''''''''''''''']53 While Framatomes responses to RAIs 80 and 106 added some clarity, the NRC staff found they did not adequately explain the insufficient agreement of S-RELAP5 with the measured results of Test 6425-2. A major source of modeling error appears present, but the specific modeling error was not identified; nor do the responses provided by Framatome appear to reflect a balanced and complete understanding of the event physics. For example,

• Framatome reasonably concluded [''''' ' '''''''''''''''''''''''''''''' '''''''' '''' ''' '''''''''

''''''''''''''' '' '''''''''' ' '''''' ''''' ' ''''''''''''''''''''''''''''' '''''''''' ''''''''''' '''''''''']

• While Framatome observed [''''' '' '''''''' ''''''''''' ''' ''' ''''''''''''''''''''''' '''''

'''''' '''''''''''''''''''' '''''' '''''''''' '' '''' '''''''''''''' ' ''''''''''''''''''''' ']

The NRC staff further observed that

• The test condition simulated by TLTA Test 6425-2 is very similar to that of FIST Test 6DBA1-B, which, as discussed below, was a tie-back test intended to benchmark the two facilities. Measured peak cladding temperatures for both tests were approximately 700 °F. [''''''''''''''''''' '''''''''''''' ''''''''''''''''' ''''''''''''''''''''''''''''' ' '''''' ''''''''''''''

53 Note that section 7.7.2.1 of ANP-10332P states S-RELAP5 [''''''''''''''' ''''''''''''''' ''''''''''''' '' ''''''''

'''''''''''''' '''''''''' '''' '''''''''''''].

- 111 -

''''''']

]

• Past assessments of TLTA Test 6425-2 with other codes have demonstrated reasonable predictions, including TRACE (Reference 58).54

• As described above in Section 3.3.4.2.2, the [''''''''''''''''''''''''''''' ''''' ''''''''''''''''' ''''''''''''''''''''''

'''''''''''''''''''''' '''''''''' ''''' '''''' ''''''''' ''' ]

In its updated response to RAI 106 dated October 31, 2018 (Reference 66), Framatome provided [' ''''''''' ''''''''''''''' '''''''''''''''''''''''' ' ''' ''''''''''''''''' ''''''''''''''''''''''' ''' ''''''''''''''''''''

'''''''''''''' ''''''''''']

• ['''''''''''''''''''''''''''''''''''''''''''' ''''' ''''''''''''' ' ''''''''''''''''''''''' '''''''''' '''' '''' '''''''' ''''' ''''''''''''''''

''''''''''''' '' '''''' '''''''''']

• [''''''''''''''''''''''''''' '''' ''''''''''''''' '''''''''''''''''' '' '''''''''''''''''' ''''''''''' ' ''''''''''''''''''''''

'''''''''''''''''''''''''' '''' '''''''''''''''''''''' ''''''''''''']

• [''' ''''''''''''''''' ''''''''' '''' '''''''''''' '''''''''''''''''''''''''' '''''''''' ' '''''''''''''''''''''' ''''''''''''''''''' ''

'''''''''''''''' '''''''''' ' ''''''' '''''''' ''''''''' ''''''''''''' '''''' '''''' '''''''''''''''''''''' '''''''''' ']

• [''''' '''''''''''''' '' '''''''''''''''''' ''''' '''' '''''''''''' ''''''''''' ''''''' ''''''''''''''''''' ''

'''''''''''''''''']

Approximate results from these sensitivity analyses (i.e., as estimated visually by the NRC staff from plots included in the updated response to RAI 106) are summarized for the 50- and 107-inch elevations in Table 9.

54 Note that this observation contrasts with subsequent discussion in Sections 3.4.4.2.2 and 3.4.4.2.3 regarding FIST Test 6SB2C, for which many codes and analysts have underpredicted the measured data.

- 112 -

Table 9: Approximate Results for TLTA 6425-2 Sensitivity Cases Peak Cladding Temperature (°F) Quench Time (s)

Case

[' '''''''''''''] [''''' '''''''''] [' '''''''''''''] [''''' ''''''''']

''''''''''''''''''''' ''''''' '''''' '''' 851 '''''

From the NRC staffs perspective, the results of Framatomes sensitivity studies indicate the following:

• [''''''''''''''' ''''''''''''' ''''''''''''''' ''' ''''''''''''' ''''''''' ''''' '' '''' '''''''''''''' '''''''''''''''' '

''''''''''' ' ''''''''' ''''' ''''''''''']

• [''''''' ''' ''''''''''''''''''''''''''''' ''''' '''' '''''' '''''''''''''''''''' ''''''''''''' ''''' '''''''''''' '''''''

'''''''''''''''''' '''' ']

• ['''''''''''''''''''''''''''' ' '' ''''''''''''''''''' ''''''' '''''''''''''' ''''''''''''''''''' ' ''' '''''''''''''''''' '''' ''''

''''''''' ''''' '' '''''''''''' '''' ''''''' '''''''''''''' ''''''''' '' ''''''''''''''''''''''' ]

Ultimately, the NRC staff considered Framatomes updated response to RAI 106 acceptable

[''''''''''''' ''''''''''''''''' '''''''''''''''' '''''''''' '''' ''' ''''''''' '''''''''''''''''''''''' '''''''''''''''''''''''' ''''''''''''''' '' '''

'''''''''''''''''''' ''''''''''''''''' ''''''''' '''''' '''''''''']. In particular, Framatomes response provided reasonable assurance that the possibility of a gross modeling error may be excluded, and offered credible grounds that the observed predictive discrepancy could be attributed to intended evaluation model conservatisms. [' '''''''''''''''' ' ''''''''''''''''''''''' ''''''''''''''' ''''''''''''''''''''''''

''''''''''''''''''' '' ''''''''''''''''''' ' '''' '' ''''''''''''''''''' ']

- 113 -

3.4.4.1.2 TLTA Small-Break LOCA Test 6432-1 (Best-Estimate Modeling)

Test 6432-1 modeled a small break (i.e., simulated size of 0.05 ft2) on the recirculation line with the high-pressure core spray system inoperative. Thus, available ECCS equipment included low-pressure core spray and three low-pressure coolant injection pumps.

Framatome presented the results of its S-RELAP5 simulation of Test 6432-1 in Section 7.7.2.2 of ANP-10332P. Overall, the NRC staff considered the agreement of the S-RELAP5 predictions to the measured test data to be reasonable. However, it should be immediately pointed out that TLTA Test 6432-1 resulted in essentially no measured heatup of the test bundle. [

[''''''''''''''''''''''''''' ' '''' ''''''''''''' ''''''' '''' '''''''''''' '''' '''''''''''''''''''''''' ''''''''''''''''

''''''' '' ''''''' '''' ''''''''].55 As a result of these factors, the NRC staff could not significantly credit Framatomes comparison to TLTA Test 6432-1 when considering the acceptability of the validation of the AURORA-B LOCA evaluation model. ['''''''''''''''''' ''''' '''' '''''''''''''''''''''' '''''''

''''''''''''''''''' ' ''' '''''''''''''' '''''''''''''']

Several anomalies were present in the measured ECCS flow rates shown in ANP-10332P

[''''''''''''''''''''''''''''' '''''''''''''''' ''''''' '''''''' ''''' '''''''''''' ''''''''' ''''''''''' '''''''''''''''' '''''''''''''''''' ''''''''

''''' '' ''''''''''''''''''''' '''''''''] The NRC staffs review of RAI 81 found the explanation concerning these flow anomalies acceptable. However, the NRC staff considered Framatomes response to RAI 81.b oversimplified and apparently nonphysical; ['' '''''''''''''''''' '' '''''' ''''

' ''''' ''''''''''''''' ' '''''''' '' '''''''''''''''' ''' ''''''' '''''''''''' ''''''].

3.4.4.1.3 TLTA Small-Break LOCA Test 6432-1 (Conservative Modeling)

In Section 7.9.2.1, Framatome described an additional simulation of TLTA Test 6432-1 that incorporated three generally conservative modeling practices from the AURORA-B LOCA evaluation model (the first two of which derive from Appendix K), namely,

• decay heat increased by 20 percent,

• Moody critical flow, and

• [''''''''''''''''''''''''' ''''''''''''''''''' ''''''''''''''''''''']

As noted above, there was essentially no measured heatup in TLTA Test 6432-1; neither was a heatup predicted in the S-RELAP5 best-estimate simulation. However, with the three above conservatisms applied, S-RELAP5 predicted heatup to a peak cladding temperature of approximately ['''' ' ''''' '''''''''''''''''''''''''' '''''''''''' ''''' ''']. In addition to various plots of cladding temperature, Framatome included plots of coolant mass in the test bundle and upper plenum. Reduced bundle mass and increased upper plenum mass were predicted in the conservative modeling prediction of Test 6432-1. Such behavior is expected ['''''''' '''''''''''''' ''

'''''''''''''''''' ''''''''''''''''' '''''''''''''''''''''''''''''] test facility.

55 For example, see discussion in Section 7.7.2.2 of ANP-10332P and Framatomes response to RAI 81 regarding TLTA Test 6432-1.

- 114 -

While Framatomes comparison with Test 6432-1 using conservative modeling can be considered a demonstration of the conservativism of the AURORA-B LOCA evaluation model,

['''' ' '''''''''''''''' '' ''''''''''''' '''''''''''''''''' ''''''' ''''''''''''''' '' ''''''''''''''''''' '''''''''''''''''

''''''''''''''''''''''''''' ''''' ''''''''''], the NRC staff expects the comparison to over-represent the actual conservatism of the evaluation model for performing BWR safety analyses. ['''''''''''''''''''''''' ''''''''

''''''''''''''] These issues will be revisited in greater detail below in Section 3.4.4.2.3, which concerns an equivalent small-break test conducted in a different facility (i.e., FIST Test 6SB2C).

3.4.4.2 FIST Integral Tests The FIST facility was used to perform thermal-hydraulic testing for LOCA and non-LOCA transients for BWRs. The FIST facility can be considered an upgraded version of the TLTA facility; while many of the same components were used, a number of scaling improvements were made, such as increasing the height of the jet pumps to full scale. While generally employing full-height components, FIST also used a single, electrically heated fuel bundle. This led to an overall scaling ratio of 1:624, equivalent to that of TLTA.

As discussed in Sections 7.7.3 and 7.9.2.2 of ANP-10332P, Framatome validated the S-RELAP5 code using the following tests performed at the FIST facility:

• BWR/6 Large-Break Test 6DBA1-B

• BWR/6 Small-Break Test 6SB2C

• BWR/6 Low-Pressure Coolant Injection Line Break Test 6LB1

• BWR/4 Large-Break Test 4DBA1 In general, these tests were simulated using the S-RELAP5 code [ '''''''''''''''''''''''''''''''' ''''''''''

''''' ' '''' ''''' ''''''''''''''''''' ''''' ''' '''''''''''''''''''''''''''' ''' '''''''''''''''''''''''' ''''''''''' '''''''''''''''''' '].

However, the S-RELAP5 simulation for Test 6SB2C (BWR/6 small-break) [' '''''''''''''''''''''''''''

''''''' ''''''''''''''' '''''''''''''''''''''''''''''''' '''''''''''''''''''''']. Framatome subsequently performed additional calculations for this test ['''' ''''''''''' ''''''''''''''''''''''''''' '''''''''''''''''''''''''''''] from the AURORA-B LOCA evaluation model. Successive sections of this SE assess the comparisons of S-RELAP5 predictions against measured data from these FIST tests.

3.4.4.2.1 BWR/6 Large-Break Test 6DBA1-B Test 6DBA1-B modeled the double-ended rupture of a recirculation suction line with two low-pressure coolant injection pumps inoperative. Thus, available ECCS equipment included high-pressure core spray, low-pressure core spray, and one low-pressure coolant injection pump. Note that FIST Test 6DBA1-B is considered a tie-back test to the TLTA large-break Test 6425-2 that was discussed above, in that both tests were intended to model the same scenario. The measured peak cladding temperature for FIST Test 6DBA1-B was 703 °F.

Framatome presented the results of its S-RELAP5 simulation of Test 6DBA1-B in Section 7.7.3.1 of ANP-10332P. Overall, the NRC staff considered the S-RELAP5 predictions to be in reasonable agreement with the measured test data. In particular, cladding temperatures were predicted accurately or conservatively. [ '''' '''''' ''''''''' '' ''''''''''''''''

- 115 -

'''''''''''''''' ''''''''''''''''''''''''''' ''' ''''''''''''''''''''' ''''''' ''''''''''''''''' ''''''''''''''''''''''''' ' ' ']. While some predictive error is apparent, ['''''''' ''''''''''' ''''' ''''''''''' ' '''''''' '''''''' '''''' ''' '''''''''''''''''' '''''''

''''''''''], these discrepancies did not appear significantly to influence the prediction of figures of merit.

3.4.4.2.2 BWR/6 Small-Break Test 6SB2C (Best-Estimate Modeling)

Test 6SB2C modeled a small-break rupture of a recirculation suction line with the high-pressure core spray system inoperative. Thus, available ECCS equipment included low-pressure core spray and three low-pressure coolant injection pumps. Note that FIST Test 6SB2C is considered a tie-back test to the TLTA small-break Test 6432-1 that was discussed above.

However, while TLTA Test 6432-1 did not result in significant heatup, FIST Test 6SB2C experienced a mild heatup to 925 °F.

Framatome presented the results of its S-RELAP5 simulation of Test 6SB2C in Section 7.7.3.2 of ANP-10332P. Despite having predicted trends for many parameters similarly to measured data, S-RELAP nevertheless underpredicted the measured peak cladding temperature by

[''''''''''''''''''''''''''''' ''''' ']. Accordingly, the NRC staff considered the S-RELAP5 prediction as overall providing minimal agreement with the measured test data.

As shown in ANP-10332P, the heatup of the peak node was measured as occurring ['''''''

'''''''''''''''' '''''' '''''''''' ''''''''' '''''''''''' ' '' '''''''''''''''''' '' '''''''''']

The NRC staff did not find the explanations provided in ANP-10332P regarding the underprediction of the measured cladding temperature satisfactory. [''' ''''''''''''''' '''''''''''''''''''''

'' ''''''''''''''''''' '''''''''' '''' ' ''''''''''''''''''''''''''''' '''''''''''''''''''''''''''''' '' ''''''''''''''''' '''''''''] The NRC staff issued RAI 53 to address these concerns to Framatome.

Framatomes response to RAI 53 ['' '' ''''''''''''' ''''''''''''''' '' ''''''''''''''' '''''''''' ' ''' ''''''

- 116 -

''''''''''''''' ''''''''''''''''''' '''''''''''''' '' ''''''' ''''' ''''56 ''''''''''''''''''''' '''''''''''''' '''''''''''''''''''''' '''''''''

''''''''''''''''''''''''''''''''' '''''''''''''' ]

Beyond the [''''''''''''''''''' ''''''''''''''''''''''''''''''' ' '''' ''''''' '''''''''''''' '''''''''''''''''''''''''''' ' '''''''' '''''

'''''''''' ''''''''''''''''''' '''''''''''''''''''''''''''''''''''' ]. The NRC staff generally considers this practice undesirable because it artificially burnishes the perceived level of agreement at the expense of the scope or even integrity of the comparison. In RAI 84, the NRC staff requested that Framatome provide the results of the S-RELAP5 simulation ['''' '''''''''''''' '' '''''''''''''''''

'''''''''' ''' '''''''''' ''' '''''''''''''''' '''''''''''''''''''' ''''''''' '']

3.4.4.2.3 BWR/6 Small-Break Test 6SB2C (Conservative Modeling)

['''''''''''''''' '' ''''' '''' ''''''''''''''''''''''''''''''''' '''''''''''''''' ' '''''' '''''' '''''''''' ''''''''' '''''''''''''''''''''''''''

'''''''''''''''''' '''''''''''' '''' ''''''''''''''''''''], as described in Section 7.9.2.2 of ANP-10332P, Framatome performed another simulation of the same test that incorporated three conservatisms from the AURORA-B LOCA evaluation model (the first two of which derive from Appendix K), namely

• decay heat increased by 20 percent,

• Moody critical flow, and

• [''''''''''''''''''''''' ''''''''''''''''''' ''''''''''''''''']

With these conservatisms imposed, S-RELAP5 was observed to conservatively overpredict the measured peak cladding temperature [ ''''''''''''''''''''''''' '' ' '''''''''''''''''''' '''''''''''''''''' ' ''' '

'''''' ''''''''''']

While ANP-10332P presents a more limited set of results for this conservative simulation of Test 6SB2C [''''' '' '' ''''''''''''''''''''''''''''''''' ''''''''''''''''' ''''''''''''''''''' '''''''''''''''''''], the NRC staff found the overall level of agreement reasonable, if somewhat conservative. However, the NRC staff observed that the FIST facility contains a single channel. [''''''''''''''' ''' ''''''''''''''''''''''''' ''''''''''''''''

56 The TRACE assessment (Reference 58) references the prediction of excessive drainage of liquid from the upper plenum into the test bundle as a potential cause for the underprediction. This phenomenon may be at work in the S-RELAP5 prediction as well, as suggested by Figure 7-342 of ANP-10332P.

- 117 -

'''''' '' ''''''''']. Furthermore, as explained in the response to RAI 94, Framatome ['''''''''''''

'''''''''''''''''''''' ''''''''''' ''''''''''''''''''' '''''''''']. Therefore, the NRC staff questioned whether the simulation of FIST Test 6SB2C in Section 7.9.2.2 of ANP-10332P overemphasizes the degree of conservatism that may be expected when applying the AURORA-B LOCA evaluation model for BWR safety analyses. Framatome replied to this concern in response to RAI 53, ['

'''' ''''''''''''''''''''''' '' '''''''' '''''''''''''''' ''''''''''''''''''''''''' ''''' '''''''''''''''''''''''''''' '''''''''''''''''''''''''''' '' '']

The NRC staffs review of the third simulation of FIST Test 6SB2C made the following observations:

• The underprediction of peak cladding temperature was of a similar magnitude to those of other analysts and codes.

• The conservatism in the full AURORA--B LOCA evaluation model for the prediction of BWR small-break LOCA scenarios would be increased considerably because no credit was taken in the simulation of FIST Test 6SB2C for a number of additional conservatisms. For example:

o ['' ''''''''' '''' '''''''''''''' '' '' '''''''''''''''''' ''''''''''''''''' ' '''''''''''''''''''' ''''''''

'''''''''''' '''''''''''''''' ''''''''''''''''''' '' '' '''''''''''''''''''''''''''' '''' ''''''''''''''']

o No credit was allowed for Appendix K heat transfer lockouts, the impact of which is discussed further below in Section 3.6.2.3.

o No credit was taken for evaluation model conservatisms associated with fuel rod stored energy due to the comparison against a test facility using heater rods.

o Due to the use of heater rods and the mild peak cladding temperature in FIST 6SB2C [''''' '''''''''''''''''''''''''''' '''' '''' '''''''''''''''''''''''''''' ''''''''''''''''''' '''''' ''''

'''''''''' ' ''''''''''''' '''''''''''''''' '' '' '''''''''''''''''''''''' ''''''''''' ''' '''''''''''''''''''''']

In light of the considerable conservatisms associated with the evaluation model, the full effect of which, in a number of cases, cannot be reasonably assessed via comparisons to existing integral effects tests, the NRC staffs remaining concerns with Framatomes modeling of FIST Test 6SB2C were addressed.

3.4.4.2.4 BWR/6 Low-Pressure Coolant Injection Line Break Test 6LB1 Test 6LB1 modeled complete severance of a low-pressure coolant injection line for a BWR/6 with the high-pressure core spray system inoperative. Thus, available ECCS equipment

- 118 -

included low-pressure core spray and two low-pressure coolant injection pumps. While a break on the low-pressure coolant injection line results in an additional reduction in the available ECCS flow rate (i.e., beyond that associated with a postulated single failure), the higher elevation of the break (i.e., relative to a break of similar size on a recirculation line) confers the salutary effects of an earlier transition of the break flow to vapor and full core refloodability.

Test 6LB1 experienced only a slight heatup that terminated at a peak cladding temperature of 635 °F.

Framatome presented the results of its S-RELAP5 simulation of Test 6LB1 in Section 7.7.3.3 of ANP-10332P. The S-RELAP5 results reported for Test 6LB1 generally provide accurate predictions of key test parameters. [ ''''''''''''''''' ''' ''''''''''''''''''' ''''''' ''''''''''''''''' '''''''''''''''''''''''''

''''''''''''''''''']

During the review, several errors were observed in the discussion of this test in Section 7.7.3.3 of ANP-10332P. ['''''''''''''''''' '''''''''' '''''''''''''''' '' '''''''' ''' ' '''''''' '''''''''''''''''' '''''''''''''''''

'''''''''''''''' '''''''''' '''''''''' ]. The NRC staff will verify that these errors have been corrected in the approved version of ANP-10332P.

3.4.4.2.5 BWR/4 Large-Break Test 4DBA1 According to NUREG/CR-4128 (Reference 43), FIST Test 4DBA1 modeled a double-ended guillotine rupture of a recirculation line for a BWR/4 with one low-pressure coolant injection pump inoperative. The available ECCS equipment included high-pressure coolant injection, one low-pressure core spray, and two low-pressure coolant injection pumps (Reference 43). The measured peak cladding temperature for FIST Test 4DBA1 was 960 °F.

Framatome presented the results of its S-RELAP5 simulation of Test 4DBA1 in Section 7.7.3.4 of ANP-10332P. The NRC staff found the cladding temperature comparisons in ANP-10332P to be reasonable to conservative. [''''''''''''''''' ''' ''''''' '''''' '''''''''''''''''' '''' ''''''''''''''''''''''''''''' ' '''''''

'''''''''' ''''''' ''''''''''''''' ''''''''''''''''''''''''' '' '''' '''' ' ''''''''''''''''''''''''''''''' ]

In Framatomes assessment of FIST Test 4DBA1, and to some degree in other assessments as well (e.g., TLTA Test 6425-2), the NRC staff perceived [' ''''''''' ''' ''''''''''''''''''''''' '''''''''''''''''''''''

''''''''''''''' '''''' '''' '' '''''''''''''''''''''''''''''''''''''''''' ''''' '''''''''''''''''''''''''''''''''' '''''''''''''''''''' '' ''''''']

Framatome did not fully explain the cause of this observed behavior in ANP-10332P. However, the NRC staff suspected that S-RELAP5 may have a tendency to ['''''''''''''''''''''''' '' ''''

''''''''''''''''''''''''''' ''''''''''''' '' '''''''''''''''' '' ''''' '''''''''''''''''''''' '''''''''''''''] In response to RAI 57, Framatome confirmed this suspicion and further suggested ['''' '''''''''''''' '''' '''''''

- 119 -

''''''''''''''''''''''''''''' ''''''''''''''' ' ''''''''''''''''''''']

The [''''''''''''''''''''''''' ''''''''''''''''''''''''' ''''''''''''''''' ''''''' ''''''''''''''] void fraction comparisons for FIST Test 4DBA1 in at least two locations:

• The fuel bundle, where the NRC staff observed ['''''''''''''''''''' ''''''''''''''''' ''' '''''''''' '''''

'''''''''' ' '''' ''''' ']

• The lower plenum, where the NRC staff made similar observations [''''''''''''''

'''''''''''''''''''''''''''''' '' ''''''' '''''''''''''' ''''' '''''''''''''' ''''''''''' '' ''''' ''''''''''' ''' '']. In response to RAI 86, Framatome attributed ['''' '''''''''''''''''''''''''' ' '''''''''''''''''''''''''''''''

' ''''' ''''''''''' ''' ''''''''''''''''''''''''''' '''''''''' '''' ''''' '''''''' ' '''''''''''''' ''']

As noted above, these inaccuracies [ '' '''' '''''''''''''' '''''''''''''''' ' '''''''''''''''''''''''] did not adversely impact the conservativism of the predicted figures of merit for FIST Test 4DBA1.

3.4.4.3 SSTF Integral Tests The SSTF was used to model phenomena associated with 3D liquid and vapor flows during the refill and reflood phases of a BWR LOCA event. The SSTF simulated, at nearly full-scale, a 30-degree section of a reactor core containing 58 rod bundles. The SSTF core was unheated; it merely served as a prototypical geometric obstruction for the representative steam and liquid flows injected into the test facility. Due to test facility design limitations, SSTF tests typically began at reduced pressure (e.g., 150 psia).

Testing in the SSTF was intended to complement the single-channel TLTA and FIST facilities discussed above. In particular, the scaling of both the TLTA and FIST facilities led to tall, narrow test-facility components which had the effect of constraining certain inherently 3D phenomena into a single dimension.

Two SSTF tests were modeled in S-RELAP5 for the purpose of assessment and validation,

namely,

• SSTF System Test EA3.1, Run 111 (BWR/4 reference case)

• SSTF System Test SRT-3, Run 26 (BWR/6 reference case)

- 120 -

As described in further detail in NUREG/CR-2568, each of these tests simulated a large-break LOCA with a design-specific ECCS configuration and appropriate single-failure.

For each of these tests, the S-RELAP5 model ['''' ''''''''''''''''''' ''''''''''''' ''' '''''''' ''''''''''''''''

''''''''''''' '''''''''''''''' '''''''''''''''' '''''''''' ' ''''''''' ''''''''''''' ''''''' ''''''''''' ''''' ''']

For both of these tests, Framatome compared S-RELAP5 predictions against measurements [

'''''''''''''''''''''' ''''''''''''''''''' '''''''' '''''''''''''''''''''''' ''''' '''''''''''''''' '''''''] The NRC staff found a number of the comparisons for these two SSTF tests to be reasonable; in particular, most fuel bundle differential pressure comparisons appeared reasonable. In addition the NRC staff found the lower plenum subcooling predictions discussed in response to RAI 50 in reasonable-to-excellent agreement with measured data. However, the NRC staff also found a number of the predictions of [''''''''' '''''''''''' ''''''' ''''''''''''''''''''''''' ''''' '''''''''''''''''' '''''] to be in minimal agreement with the measured data. [ '''''''''''''''''' ''' '''''''''''''''''' ''''''''''''''''''''''' ''''''''''''''''' ''''''''''''''' '' ''''''''''' '''''''''

'''''''] While a plot provided in response to RAI 47 showed better agreement ['''' ''''''''''

'''''''''''''''' ''''' ''''''' ''''''''''''''''''''' ' ''''''''''''''] in Limitation and Condition 25, in Section 5.0 of this SE. Also, ['''''''' '' '''''''''''''''''' '''''''''''''''''''' ''''''''''''''''''''''' '' '''''''''''''''''''''''''''' '''''

'''''''''''''''' ' '' ''''''''' '''''' '''''''''''''''' ''''''''''''''''''''], as noted in RAI 79, [''''''''''''''''' '''''''''''''''''''

'''''''''''''''''''''''' '' '''' ''''''''''' '''''' ']. Framatome did not provide adequate explanation in either ANP-10332P or in response to RAI 79 for the observed behavior, [''''''''''''''''''' ''''''''''''''''''''''' ''

'''''''''''''''' '''''''''''''''''''''''''' '''''''''''''''''' ''' '''''''''' '''''''''''''''']

From the NRC staffs perspective, prediction of the 3D phenomena occurring in the SSTF is challenging, even for detailed mechanistic models. In light of ['' ''''''''''''''''' ''''''''''''''''''' ''''''''''

''''''''''''''''''' ''''''''''''''''''' ''''''' ' '''''''''''''''''''''' '' ''''''''''''''''''' ''''''''''' ' ''''''''''' '''''''''''''], predictive discrepancies are not surprising. Although basic trends were captured correctly in the S-RELAP5 assessment simulations, the NRC staffs decision to [''''''''''' '' ''''''''''''''''''''

'''''''''''''''' ''''''''''''''''''' '''''''''''''' '''' ''''' '' '''''' '''''''''''''''' '''''''''''''''''''''] in Limitation and Condition 8 was influenced by the inaccuracy in ['''''''''''''''''' '''''''' '''''''''''''' '''''''''''''' ''

''''''''].

3.4.5 Impact of [''''''''''''''''] Code Version The NRC staffs evaluation of the assessment and validation for the S-RELAP5 code has focused upon the material included in ANP-10332P. As noted above, [''''''''' ''''''''''''''''' ''''''' '

''''''''''''''''''''' '''' '' ''''''''''''''] version of S-RELAP5. However, during the NRC staffs review of ANP-10332P, Framatome elected to request approval of the AURORA-B LOCA evaluation model with the [''''''''''''''''] version of S-RELAP5.

To accommodate this change to the proposed evaluation model, across several RAIs (primarily 15 and 126) the NRC staff requested that Framatome provide the results of additional

- 121 -

assessments to demonstrate adequate performance of the ['''''''''''''''''] version of S-RELAP5.

Furthermore, ['''''''' '''''''''''''''''''''''' ''''''''''''''''''''''''''''''''''''''''''''''''''''' ''''''''''''''' ''''''''''''' '''''''''''''''''' '''''''''

'''''''''''''''''''''' '''''' ''''''''''''''''' '' '' '' ''''''''''''''''' ''''''''''''''' ']

As described in response to RAI 15, Framatome [''''''''''''''''' ' ''''''''''''''''''' ' ''''''''''''''''''''''' ''

''''''''''''''''''''''''' ''''''''''''''' ']

Framatomes ['''''''''''''''''' ' ''''''''''''''''''''''''' ''''''''''''''''''' '''''' ' '''''''''' ' '' ''''''''''''''''''''''''''' ''''''''''''''

''''''''''''''''''''''''''' '''''' '' '''] No rationale was presented by Framatome to support which tests should be included or excluded from the continuity of assessment process (e.g., relative to coverage of the phenomena most relevant to prediction of LOCA figures of merit), and the NRC staff considered Framatomes selections somewhat arbitrary.

The [''''''''''''''''' ' '''''''''''''''''''''''' ''''''''' '''''''''''''''' ''''''''''' ''' ''''''''''''''' ''''''''''' ' '' '''''''''' '

''''''' ''''''''''''''''''' ' '' '''''''''''''' '''''''''''''' ' ''''''''''''''''''']. However, several relevant exceptions exist to this general statement, including the following:

• As noted in RAI 51, the NRC staff observed a conservative ['''''''''''''''''' '' '''''

'''''''' '''''''''' '''''''''''''''''''' '''' '' ''''''''''''''''''''''''''''' ' '''''''] to the NRC staffs decision to impose Limitation and Condition 25.

- 122 -

• The [''''''''''''''' '''''''' ''''''''''' '''''''''''''''''' '''''''''''''' ''''''''' '''''''''''''''''''''' ' ''''''''''''''''''

''''''''''''''' '''''''''''''''], generally resulted in improved agreement with measured data.

• Improved agreement [''''' '''''''''''''''' ''''' ''''' ''''''''''''''''' '' '' ''''''''''''''''' '''''' ''''''''''''

'''' ''''''''''' ''''''''''''''''''''''''''' ''''' ''''' ''''' ' '' ' ''''''''''''''' ''''''''''''''''' ''' ]

Based upon its review, the NRC staff found the ['''''''''''''''' ''''''''''''''''''''' ''''''''' ''''''''''''''''' '

'''''''''''''''''''' ' '''''''''''''''''' '''''''''''''''''''''' '' ''''''''''''''''''' ' '' ''''''''''''''' ''''''''''' ''''''''''''''''''''''].

However, in the NRC staffs view, a major shortcoming of these continuity of assessment results is that ['''''' ' '' ''''''''' ' '' '''''''''''''''''''''' '''' ''''''''''''''''' '''''''''''''''''''''''' ''''''''''''''

'''''''''''''''''''''''''''''] In particular, as addressed in RAI 126, the NRC staff ['''''''' ''' '''''''''''''' '

'''''' ''''''''''''].

Framatome provided the requested results in response to RAI 126; [''' ''''''''' '''

''''''''''''''''''''' '''''''''' ''''''''''''''' ' ''' '''''''''''''''''' '''''' ''''''''''''''' ] for these integral tests is summarized below in Table 10.

Table 10: Impact of ['''''''''''''''''] Code Version on Integral Effects Test Comparisons Peak Cladding Temperature (°F) Temperature Integral Test

['''''''''''''''''] ['''''''''''''''] Difference (°F)

TLTA Test 6425-2 '''''' ''''''' '''''

TLTA Test 6432-1 ' ''''''''''''''''''' ''''''''''

FIST Test 6DBA1 ''''' '''' ''

FIST Test 6SB2C ''''' '''' ''

FIST Test 6LB1 '''' '''' '''''

FIST Test 4DBA1 '''''' '''''''

Overall, the NRC staff considered the results shown in Table 10 reasonably consistent [''''' ''

'''''''''''''''''''''''' '''' '' ''''''''''''''''' ''''''' '''''''''''''] It is further worthy of remark that most BWR integral test results [''''''' '''''' ''''''''''''''' '' '''''''''''''''''''''''''' ''''' ''''''''''''''''''' '' ''''''''''''''''''''''

'''''''''' ''''''''''''''''''' ''''''''''' '''' ''''''' ''''' '' ] do not involve peak cladding temperatures near regulatory limits. High-temperature validation for BWR LOCA evaluation models typically comes from comparisons against separate effects testing [''''''' '' ''' ''''''''

'''''''''''''''''''''''''''''''''''''']

3.4.6 Summary of Assessment and Validation The foregoing discussion of the assessment and validation effort for the AURORA-B LOCA evaluation model considered these comparisons at an individual level. However, the

- 123 -

assessment of an evaluation model must ultimately be done on a holistic basis. In particular, for an Appendix-K-based evaluation model, the assessments must provide confidence of conservative prediction of the figures of merit. This section of the NRC staffs SE summarizes each set of assessment comparisons and concludes with an overall evaluation of adequacy.

3.4.6.1 Foundation Methodology Assessments Summary The foundation methodology assessments for the AURORA-B LOCA evaluation model cover models and modeling practices associated with Appendix K to 10 CFR 50, core simulator and lattice physics methods, and fuel thermal-mechanical methods. In a number of cases cited above, the NRC staff determined that some relevant PIRT items were not actually covered by the specific foundation methodology cited by Framatome. However, the NRC staff considered these PIRT items to be adequately covered by a combination of other sources, including separate effects test assessments, previous NRC staff reviews of S-RELAP5-based methods, the NRC staffs present review of ANP-10332P, and limitations and conditions specified in Section 5.0 of this SE. As a result, the NRC staff concluded that the foundation methodology assessments adequately support the AURORA-B LOCA evaluation model.

3.4.6.2 Component Effects Tests Summary Component effects tests were used for the assessment of specialized component models in S-RELAP5, namely, rod bundle pressure drop tests, jet pump performance tests, steam separator tests, critical power tests, and countercurrent flow limitation tests. Of these, the NRC staff found the jet pump and countercurrent flow limitation tests of particular relevance to the LOCA event. As documented in the discussion above, the NRC staff generally found these tests in reasonable agreement with data and concluded the component effects tests adequately support the AURORA-B LOCA evaluation model.

3.4.6.3 Separate Effects Tests Summary The separate effects test comparisons Framatome performed for the AURORA-B LOCA evaluation model ['''''''''''''''''''''''''' '''''' ''''' ' ''''''''''''' '''''''''''' ' ''''''''''' '' '''''''''''''''''' '

'''''''''''' '''' ''''''''''''''' ''''''''''''''''' '''''''''''''''''' '' ''' '''''''''''''''''''' ' '' ''''''''''''''''''''' ''''''''''']

Regarding the topic of test scaling and distortion, the NRC staff observed that a significant proportion of the separate effects assessments for the AURORA-B LOCA evaluation model

[''''''''' ' '''''''''''''''''''''''''' '''''''''''' ''''''''' ''''''''''''''''''''' '''''''' ''''''''' '''''''''' ''''''''''' '''' ''''''''''''''''

'''''''''''''''''] However, ANP-10332P generally did not adequately discuss the scaling of these PWR tests to BWR conditions. In RAI 48, the NRC staff requested that Framatome provide additional scaling rationale.

Framatomes response to RAI 48 [''''''''''''''''''' '' '''''''''''''''''''''' ''''''' '''''''''' ''''''''''''''''''''' '

''''''''''''''' '' ''' ''''' '''' ''''''''''' '''''''' ' '''''''''], specific details and rationale for scaling these individual tests to BWR conditions were not provided in the response. Framatome did,

- 124 -

however, assert that tests outside of the BWR operational range were excluded from assessment activities. Framatomes response also noted an important factor, in that modeling options associated with the AURORA-B LOCA evaluation model were applied in simulating all the PWR tests discussed above ['''''''' ' ''''''''''''''' '''' ''''''' ''''''''''''''' ''''''''''' ''''''''''''''''''

'''''' '' '''''''''''''''' ''''''''' ''''''''' '''''''''' '''''''''''''''''''''''''''' '''''''''''''''''''' ''''']

As discussed above, the NRC staff has judged some separate effects tests in ANP-10332P (i.e., UPTF countercurrent flow and entrainment tests) as inapplicable to BWRs. In addition, the NRC staff judged that some other tests, while partially applicable, may not be the best available choice for BWR LOCA conditions (e.g., Bennett tube tests). However, in light of the graded approach suggested in RG 1.203 for conservative evaluation models (see Section 2.2 above),

the NRC staff found Framatomes response to RAI 48 acceptable. In particular, the NRC staff agreed [''''' ''' ''''''''''''''''''''' ''''''''''''''''' ''''''' '' '' ''''''''''''''''''''''''' ''''''''''''''''''''''''' '''''''''''''''' '

'''''''''''' '''''''''' '''''' '''''''''''''''''' '''''''''''' '''''' ''''''' ''''''''''''''' '' '''''''''' '''''] to provide confidence that the Appendix K-based AURORA-B LOCA evaluation model will predict conservative results for BWR LOCA scenarios.

An issue that recurs in the NRC staffs evaluations of individual separate effects test series above is that the test series Framatome selected for assessment did not cover the full range of LOCA conditions. [ '''''''''''''''''''' ''''''''''' '''''' '''' ''''''''''''''''' '''''''''''''''' '' '''''''''''''''

''''''''''''''''''''' ''''''''''''''''''''' ''''''''''' ' ']. In light of the significantly restricted range of ['''''''''

'''''''''''''''''' '''' ''''''' ''''''''''''''''''''''''], the NRC staff considered the assessments for certain tests (e.g., rod bundle void tests and Bennett tube tests) as having limited applicability to the BWR LOCA event. However, the NRC staff found that many key phenomena were included in multiple test series, such that the selected assessment suite provided reasonable coverage of the BWR LOCA event domain; furthermore, the assessments generally showed reasonable agreement with test data. In a handful of cases where minimal agreement was observed

['''''''''''''''''''''''''''''''''''''''''''''''''''''' '''''''''''''''' '''''''], the results tended to be conservative and did reflect overall correct trends.

In conclusion, as the result of several fundamental issues, [''''''''''''''''' '' ''''''''''''''''''''''''' ' ''''''''

''' ''' '' '''''''''''''''' '''''''''' '''''''''''' '''' ''' '''''''''''''''''''''''''' '''' ''' '''''''], the NRC staff concluded that the separate effects testing documented in ANP-10332P does not fully satisfy EMDAP guidance. Nevertheless, as explained further below in Section 3.4.6.5, the NRC staff ultimately found Framatomes assessment sufficient for a conservative, Appendix K-based evaluation model.

3.4.6.4 Integral Effects Tests Summary The integral effects test comparisons Framatome performed for the AURORA-B LOCA evaluation model incorporated data from eight integral tests conducted at three different facilities. Framatomes selection of integral tests considered both the BWR/4 and BWR/6 plant designs. As noted above in Section 1.2, while these designs are similar, there are relevant differences that Framatome has explicitly taken into account in its validation. [''''''''

'''''''''''''''''''''''''' '' ''''''''''''' ''''''''''''' '''''''''''''''''''''''''], overall, the NRC staff found the quantity and

- 125 -

variety of tests selected for comparison appropriate, particularly for an Appendix K-based evaluation model.

A recurring issue pertaining to integral effects test assessments that deserves mention is discussed in RAI 52. In particular, over the course of reviewing the validation of the AURORA-B LOCA evaluation model against integral-effects test data, the NRC staff questioned Framatomes practice in many cases of [''''''''''' '''''''''''''''' '''''''''''''''''''''''' '''''''' '' ''''''''''

''''''''''''' '' ''''''''''''''''''''' '''''''''' '''''''''''''''''''''' ''''''''''''' ''''''''''']. The NRC staff also found that in certain cases, due to significant error in the S-RELAP5 predictions of [''''''''' ''''''''' '''''''''' '''''

''''''''''''' ''''' '''''''''''' ''''''' ''''' ''''' ''''''''''''''''''''''''''' '''''''' '''''' ''''' '''''''''''']. In response to the NRC staffs concern, Framatome stated that incorrect predictions of [''''''''''''''''' '''' ''''''''

''''''''''''''' ''''''' ' ''''''''''''''''''''''''''''''''''''' ''''''''''''''''''' ''''''''' '' '''''''''' ''''''''''']. While ultimately agreeing that the practices used by Framatome [ ''''''''''' ''''''''''''''''''''' ''''''''''''' '''''''''] in the validation comparisons correspond reasonably well with those for performing plant-specific safety analyses, the NRC staff concluded that, in general, ['''''''''' '''' ' '' '''''''''''''''''''

'''''''''''''' '''''''''''''''''''''']. In general, [''''''''''''''''' ''''''''' '''''' '''''''''''''''''''''' '''''''''''

'''''''''''''''''''' ] tends to undermine the objective of the validation exercise by providing an overly optimistic assessment of the predictive accuracy of the code being validated.

Reasonable agreement was observed in most of the integral effects tests presented in ANP-10332P. Two exceptions to this statement were noted with respect to TLTA Test 6425-2 and FIST Test 6SB2C. [ '' ''''''''' ''''''' ''''''''''''''''''''' '''''''''' '''''''''''''''''' '''''''' '''''''''''''''''''''''''''''

'''''''''' ], the NRC staff also found the prediction acceptable. Furthermore, [''''''''' ''''

'''''''''''''''' ' '''''' '' '''''''''''''''''''''] (as extended by Limitation and Condition 8), the NRC staff considered the AURORA-B LOCA evaluation model capable of making conservative predictions of relevant phenomena.

In conclusion, the NRC staff found that, subject to limitations and conditions specified in this SE, Framatome has presented sufficient evidence of the assessment of S-RELAP5 against integral effects test data. However, because (1) the NRC staff did not agree with a number of aspects of Framatomes response to RAI 47 concerning [''' '''' ' '''''''''''' '''''''''' ''''''''''' '''''''''''''''''

''''''''''''''' '''''''''''' '''' '''''''' '''''' '' ''''''''''''''''''''''''' ''''''''' '''''''''''''''''''' ''''''''''], the NRC staff concluded that the integral effects testing does not fully satisfy EMDAP guidance.

Nevertheless, as explained below in Section 3.4.6.5, the NRC staff ultimately found Framatomes assessment appropriate for a conservative, Appendix K-based evaluation model.

3.4.6.5 Assessment and Validation Conclusion Prior to concluding discussion of Framatomes assessment and validation of the AURORA-B LOCA evaluation model, several overarching summary observations should be made:

- 126 -

• A sufficiently wide range of separate and integral effects tests was assessed.

• Limited explanation was generally provided to justify the selection of specific tests for assessment, as opposed to others of equal or occasionally greater apparent applicability.

• Limited effort was made in a number of separate effects assessments to justify validation of phenomena on an individual, quantitative basis. A significant number of relevant phenomena was assessed [''''' ''''''''''' ''''''''''' '''''''''' ''''''''], which the NRC staff generally did not find effective.

• Given the reliance in ANP-10332P [' ' ''''''''''' '''''''''''''''''''''''''' ''''''], few separate effects test assessments focused on the accuracy of [''''''''''''''''' '''''''''''], which influenced the NRC staffs determination to impose Limitation and Condition 8. While

[''''''''''''''''' '''''''''''''] was incorporated in a number of BWR-specific integral effects test assessments, its individual influence in these tests and the evaluation models corresponding predictive accuracy with respect to this individual phenomenon were not well characterized in the material provided for NRC staff review.

• [''''''''''' ''''''''''''' '''' '''''''''''''''''''''''''' '''' ''''''''' ''''' '''''''''''''''''''''''' ' ''' ''''''''''' ''''''''']

could not be adequately validated based upon the information in ANP-10332P.

However, staff-imposed limitations and conditions ['''''''''''''''' '''' '''''''''''''''''' '''

''''''''''''''''' '''' '' ''''''''''''''' ''''''''''''''''' '''''' ''''''''''''''''''''''' ''''' ''''''''''''''''''''' ' ''''' ''']

address the incomplete validation [' '''''''''''''' ''''''''''''''' '''''' ''''''''''''''''''''''' '''' '''''''''

''''' ''''''''''''''''''''''''''].

• While the full set of assessments ['''' '' ''''''''''''''''''''''''''' '''''' '' ''''''''''''''' '''''''''''' '

'''''''''''''''''' ''''''''''''''' ''''''''''''''' '''' '' ''''''''''''''' ''''''' ''''''''''''''] is capable of calculating conservative results when used in the Appendix K-based AURORA-B LOCA evaluation model was generally provided (noting, however, Limitation and Condition 22).

• While evidence of consistency was observed between the nodalization used in many assessment and validation cases and that intended for plant safety analysis (e.g., nodalization for core and reactor vessel), test facility scaling differences make a direct comparison difficult. Framatome in some cases did not justify nodalization consistency or even provide nodalization diagrams for test facilities.

The NRC staff has made a point in this SE of emphasizing that Framatomes assessment and validation applies strictly to the S-RELAP5 code as opposed to the AURORA-B LOCA evaluation model. That being stated, it is reasonable to expect that ['''''''''''''''''' ' ''

''''''''''''''''''''''' ''''''''''''''''''''''] will in general tend to result in more conservative predictions. A demonstration of the impact of applying a selection of important conservatisms is provided in Framatomes response to RAI 58, which as discussed further below in Section 3.6.1, shows a peak cladding temperature [''''''''' ''''' ''''' ']. Based upon this evidence, and further evidence of reasonably expected predictions for the full evaluation model in sensitivity studies (Section 3.5) and demonstration cases (Section 3.6), the NRC staff ultimately considers the validation Framatome performed on the S-RELAP5 code as providing sufficient indication that

- 127 -

the full AURORA-B LOCA evaluation model will reliably produce conservative figures of merit, as intended by Appendix K to 10 CFR 50.

The NRC staff concluded that the assessment and validation performed for the AURORA-B LOCA evaluation model does not fully satisfy the intent of the EMDAP outlined in RG 1.203.

However, the NRC staff reasoned that the EMDAP was designed for best-estimate methods, and, as noted in RG 1.203, allowance should be made for conservative methods that satisfy the requirements of Appendix K to 10 CFR 50.57 In particular, the assessment and validation effort described in ANP-10332P provides adequate confidence that, subject to the limitations and conditions in Section 5.0 of this SE, the AURORA-B LOCA evaluation model will predict conservative figures of merit for the BWR LOCA event. Therefore, the NRC staff found the validation and assessment presented in ANP-10332P for the AURORA-B LOCA evaluation model acceptable.

3.5 Sensitivity Studies Inasmuch as the AURORA-B LOCA evaluation model is submitted as an Appendix K evaluation model, formal statistical quantification of uncertainties is not necessary to support the demonstration of compliance with 10 CFR 50.46. Nevertheless,Section II.2 of Appendix K requires the performance of sensitivity studies to provide assurance that arbitrary modeling practices do not adversely affect predicted figures of merit.

Sensitivity studies supporting the AURORA-B LOCA evaluation model are presented in Section 7.9 of ANP-10332P and consider the following perturbations:

• requested maximum timestep length

• hot channel axial nodalization

• core axial power shape

• core radial power shape

• fuel channel grouping and its impact on predicting parallel channel flow

• break and ECCS injection nodalization (see response to RAI 41)

Framatome found the results of all sensitivity studies presented in ANP-10332P acceptable, including several that exhibited ['''''''''''''''''''' '' ''''''''' ' ''''''''''''''' ''' '' '''''''''''''''' '''''''''''''''''

'''''''' ''''' '''''''''''' ''''''''' ''''''' ''''''''''''' ' '''''''''''''''''''''']. Whereas, in light of the significant variation demonstrated in certain sensitivity studies, the NRC staff disagreed with Framatomes conclusion, and issued RAIs 112 and 126 to request that Framatome address [these large variations and other concerns, ['''''' ' '' '''''''''''''''''''''' '' '''''''''''''''''''' '''''''''''''''''' ''''''''''''''''''

'''' '' ''''''''''''''''' '''''' '''''''''''' '' '''''''''''''''' ''''''''''''' ''''''''''''''''''''''].

In response to RAIs 112 and 126, Framatome reperformed a number of the sensitivity cases in ANP-10332P. The results reported in these responses showed significant improvement relative to the original analyses in ANP-10332P. [''''' ''''''''''''' '''''''''''''' '''''''''' ''''''''''''''''''' ''''''''' ''''

''''''''''''']. Details of these sensitivity studies are discussed in succeeding sections of this SE.

57 Interestingly, all integral effects tests assessed in ANP-10332P post-date the promulgation of 10 CFR 50.46 and Appendix K, as do many of the separate effects tests. Such tests were performed primarily to support an improved understanding of physical phenomena to support the development of more realistic analysis methods.

- 128 -

As it turns out, while not mentioned in ANP-10332P, many of the original sensitivity analyses for the BWR/4 case were affected by the non-representative practice described above of ['''''''''''''

'''''''''' '''''' ''' ''''''''''''''' ''''''''' '''''''''''''''''''''''' '''''''''''''''''''''''' '''''' ''' '''''''''''''''''] The evidence provided in response to RAIs 112 and 126 provides confidence that the AURORA-B LOCA evaluation model is not unduly sensitive to changes in timestep and nodalization provided that the non-representative practice [''''''''''''''''''''' '''' ''' '''''''''''''''''''''''' '''''''''''''''' '''''''''''''' ''''''''''] is not employed. However, as noted above in Section 3.3.5.7, Framatome may seek to use this non-representative modeling practice in future plant-specific safety analyses. Therefore, as part of Limitation and Condition 21, prior to permitting plant-specific application of this non-representative modeling practice, the NRC staff will require adequate justification that the excessive sensitivity displayed in the S-RELAP5 calculations using the technique involving

[''''''''''''''''''''''''''''' '''''''''''''''''''''' '''''''''''''''' ''''''''''''''''' '''''''''''] will not have an adverse effect on the conservatism of the calculated figures of merit.

3.5.1 Timestep Length Sensitivity studies for the requested maximum timestep length are described in Section 7.9.3 of ANP-10332P. The sensitivity analyses used a BWR/4 input deck and considered four permutations of large and small breaks on the recirculation system pump discharge and suction lines. Three sensitivity timesteps (ranging from 0.001 to 0.01 seconds) were considered in addition to the nominal value of 0.005 seconds. The sensitivity results in ANP-10332P revealed unacceptably large variations in peak cladding temperature [ ' ''''''''''''''''''''''''''''''' '''' ].

Thus, in RAI 112.a, the NRC staff requested that Framatome justify or revise its approach.

Furthermore, in RAI 126.d, [''' ''''''' ''''' '''''''''''''''''''''' '' '''''''''''''''''''''''' ''' '''''''''''''

''''''''''''''''''' '''''''''']

Framatome responded to the NRC staffs questions [' ''''''''''''''''''''''' '''' ''''''''''''''''' ''''''''''

''''''''''''''''''''''''' ''''''''''''''''' ''' ''''''''''''''''''''' ' '''''''''''''' ' ''''''''''''''''''''''''''' ''''''''''''''''''''''''' ''''''''''''''''''].

As documented in response to RAI 126.d, the updated sensitivity studies show significantly increased robustness against perturbation. Framatomes response states that the results support use of timesteps 0.005 seconds and smaller. The NRC staff found that these sensitivity studies addressed the issues identified in RAIs 112.a and 126.d and considered the results consistent with expectations. In particular, for timesteps within the qualified range ['''''''''''''

''''''' ''''' ''''''''' '''''''''''''''], the largest peak cladding temperature variation observed in the revised sensitivity analysis was [' '''' '''''''' '''''''' ].

3.5.2 Hot Channel Axial Nodalization A sensitivity study of the axial nodalization for the hot channel is described in Section 7.9.4 of ANP-10332P. In particular, Framatomes study varied the number of hydraulic nodes [''''' '

'''' ''''''''' '''''''''''''''''' ' ''''''''''''''''''' ' ''''''''''''''''''''''''' ''''' ']. The NRC staff found this result unacceptable and, in RAI 112.b, requested that Framatome justify such a large variance.

Framatome responded by providing a revised sensitivity study of hot channel axial nodalization that was unaffected by the practice of ['''''''''''''' '''''''' '' ''''''''''''''''''''''''' '''''''''''''''''' '''''''''''''''

'''''''''''], as the study included in ANP-10332P that displayed heightened sensitivity apparently

- 129 -

had been. The revised comparison showed a sensitivity of approximately 40-45 °F, which the NRC staff found reasonable.

Framatomes sensitivity study only considered the hydraulic nodalization of the hot channel.

Framatome stated that [''' ''''''' '''''''''''''' '''''''''''''''''''''' '' '' '''' '''''' ' ''''''''''' ''''''''''''''''

''''''''''''''' '''''''''' '''''''''' '''' '''''''''' '''''''''''''''''' ''''''' '''''' ''''''''''''''''' ''''']. According to the draft modeling guidelines and demonstration cases audited by the NRC staff, Framatome specifies

['' '''''''''' ' ''''''' '''''' '''''''' ''''' '''''''''''''''' ''''''' '' ''''''''''''''''''' ''' ' ' ''''''''''''']. In light of the fairly detailed heat structure nodalization scheme employed, Framatome assumed the nodalization to be converged.

While agreeing that the [''''' '''''''''''''' ''''''''''''''''''''''] used by Framatome appears reasonable, the NRC staff did not agree with the logic expressed in ANP-10332P [''''''''''''''''''' '''''''

'''''''''''''''''' '''''''''''''''''''''''''' ']. However, the NRC staff ultimately found Framatomes focus on the

['''''''''''''''''' ''''''''''''''''''''''' '''''''''''''' '' ''''' ''''''''''''''''' ''''''''''''''''''''''''''''] reasonable based upon the level of detail employed and evidence observed during an audit of the demonstration analysis cases, wherein [''''' '''''''''''''' ''''''''''' '''''''''''' ' '' ''''''' '''''''''''''' '''''''' ''''''''''''''''' '''''''''''''''''

'''''''''''''''''''''''''''' ''''''''''''''''''''' '''''''''''''''''''''''''''''' '''''''''''''''''''''''' ''''''''''].

Furthermore, the hot channel nodalization sensitivity studies performed by Framatome ['' ''

''''''''' '''''''''''''''''' '''''' ' ''''''''''''''''' '''''''' ''''''''''''''' '''''''''''''' ''''''''']. However, this issue was ultimately addressed in Framatomes updated response to RAI 13.c, as discussed above in Section 3.3.4.1.2.

3.5.3 Core Axial Power Shape Section 7.9.5 of ANP-10332P provides the results of Framatomes sensitivity study of core axial power shape, which considered two cases:

• varying only the hot channel between mid- and top-peaked axial power shapes

• varying the remainder of the core between mid- and top-peaked axial power shapes 3.5.3.1 Hot Channel Axial Power Shape Varying the axial power shape for the hot channel was observed to ['''''''''''''''' ' ''''''''''''''''''''

''''''''] On this basis, Framatome concluded ['''' '''''''''''''''''''' ''''' ''''''''' ''''''''''' '' ''

'''''''''''''''' ''''''''''''''''''''''''' ' ''''''''''''''''''' '''''' ''''''''' ''''' ''''''''''''''''''' ''''''''''''''].

While it appears reasonable ['''' '' ''''''''''''''''''''''''''' ''''''''''''''''' '''''''''''''''''''' '''''''''''''''' ' ''''

''''''''''''''''' '''''''' '''''' '''''''''''''''''''''''' '''''''''' ''''''' ''''''''''' '''''''''''''''''], the NRC staff observed in RAI 111.c that the demonstration case results presented in Section 7.7.4 of ANP-10332P show that [''' ''''''''''''' '''''''''''''''''''' ''''' ''''' '''''''''' ''''''' '''''''''''''' ''''''''''''''''''''''' ' '' '''''''''''''''''''''''

'''''''''' ''''' '''' ''''''''''''''''''''''''' '''''''''''''''''''' '''''']. ANP-10332 did not explain or even acknowledge this apparent contradiction.

- 130 -

However, Framatome clarified the situation in response to RAI 111.c, stating [''''' ''

'''''''''''''''''''''' ''''' '' '''''''''''''''''''' '''''''' '' '''''''''''''''' ''''''' ''' ''''''''''''''''''''''''' '''''''''''''''' '''''''''''''''' ]

From the NRC staffs perspective, [''''''''''''''''' '' '''''''''''''''''''''' ''''''''''''''''''' ''''''''''''''''''' '''''''''''''''

''' '''''' ' '''''''' '''''''''''''''' '''''''''''''''''''''''], according to Limitation and Condition 8 in Section 5.0,

['''''''''''''''' '' ''''''''''''''' '''' ''' '''''''''''''''''''''' '''''''' ''''''''''''' '''''''''''''''' '' '''''''''''''' '''''''''''''''''' ''

'' ''''''''''''''''''''''' '''''' ''''''''' '''''''''' ''''' '' '''''''''''''''''' '' ''''''''''''''' ' ''' '' '''''''''''''']:

• The BWR/6 SF-LPCS scenario discussed in Framatomes response to RAI 15 ['''''''''''''

''''''''''''']

• ANP-10332P and Framatomes responses to RAIs 15 and 111 confirm [''''' ''

''''''''''''''''''''''''' ' ''''''' '''''''''''''''''''''''''''' ''''''''' '''''''''''''''']

Therefore, the NRC staff found the results of the hot channel axial power shape study inconclusive and reasoned that ['''''' '''' ''''' ''''''''''''''''''''' '''''' '''''''''' ''''''''''' ''''''''' '

'''''''''''''''''''' '' '' '' ''''''''''''''], as specified in Section 6.2.1 of ANP-10332P.

3.5.3.2 Average Channel Axial Power Shape Section 7.9.5 of ANP-10332P describes a second sensitivity study, wherein the hot channel was set to a top-peaked axial power shape, and the remainder of the core was varied between the mid- and top-peaked profiles. Both small- and large-break LOCA scenarios were considered.

['''''' '''''''''''''''''' '''''''''''''''' ''''''''''''''''''''' '''''' ''''' ' ']. The NRC staff ultimately found these results reasonable and supportive of Framatomes approach [ '''''''' ' ''''''''''''''''''''''''' ''''''''''

' '''''''''''''' ''''''''''''''''']

3.5.4 Core Radial Power Shape As described in Section 7.9.6.1 of ANP-10332P, Framatome performed analysis for both a large- and a small-break scenario to illustrate the sensitivity of the core to radial power shape.

In the reference case, [''''''''''''' ''''''' ''''''''''''''''' ' '''''''' ' '''' ' '''''''' ''''''''''''''''' '''' '''''''''''''

''''''''''''''''''''''' '''' ''''''''''''''''' '''''''''''' ' ''''''' '''''''''''''''' '''''''''''''''''] The NRC staff disagreed with Framatomes conclusion and requested further explanation in RAI 112.c.

- 131 -

Framatomes response to RAI 112.c provided revised results that demonstrated little sensitivity for the large-break scenario once the issue with ['''''''''''''' '''''''' ''' ''''''''''''''''''''''' ''''''''''''''''

''''''''''''''' ''''''''''] was resolved.

The NRC staff found the results of the revised sensitivity study reasonable but, based on other considerations discussed below, ultimately was unable to conclude that [''''' ''''''' ''''''''

'''''''''] would not significantly impact the figures of merit calculated by the evaluation model. ['

''''''''''''''''''''' ''''''''' ' '''''''''''''''''''''' '''''''''''''' ' '''''' ''' '''''''''''' ''''''' ] Therefore, the NRC staff concluded that (1) as per Limitation and Condition 14, [''''''''''''''' ''' ''''''''''''''' ''''''' '

''''''' ''''''' '''''''''' '''''''''' '''''' '' '''''''''''''' '''''''''''''''''''''''''''''' ' ''''''''''''''''''''''' '''''''''''''''']

3.5.5 Fuel Channel Grouping / Parallel Channel Flow As described in Section 7.9.6.2 of ANP-10332P, Framatome performed a ['''''''''''' '''''''''''''''''

'''''''' ' ''' ''''''''''''' '''''''''''''''' ''''' ''''''''''' ''''''''''' ' ''''''''''' ''''''''''''''' '''' '''''''''''''''''].58 Framatomes sensitivity study built off of one of the SSTF cases used as an integral effects assessment of the AURORA-B LOCA evaluation model [''''' ''' '''''''''' '''' ''''''''''''''''''' '''''''''

'''' '' ' '''''''''''''''''''' ''''''''''' '''''''''''''' ''''''''''''''].

Framatome performed four sensitivity cases, in addition to the base case. [''''' ''''''' ''''''

'''''''''''''''''' '' ''''''''''''''' '''''''''''''''''' '''' ''' ''''' '''''''''''''''''''' ''''''' ''''''' ]

Unlike the SSTF assessment case described in Section 3.4.4.3, ['' '''''''''''''' '''''''''''''''''

'''''''''''''' '''''''''''''''''' '''''''' '''''''''''''''''''' ''' ''''''''''''''''''' '''''''''''''''''''' '''''''''']

Framatome characterized the flow regime [ ' ''''''' '''''''''''''' '''''''' '''''''''''''''''''' '' '''''''''''''''

'''' '''''''''''''' ' '' '''''' ' '''' ''' '''''''''''''''' '''''''''''''' ' ''''''''' ], Framatome recognized this method as an approximation capable of providing an overall sense of channel flow regime.

Based upon the results presented in ANP-10332P, Framatome observed ['''' ' ''' ''''''

'''''''''''''''''' ''''''''''''''' ''''''''''''''' ''''''''''' ' ''''''' ''''''' '].

58 Parallel channel flow and its influence on the LOCA event are described above in Section 3.3.4.3.3.

- 132 -

The NRC staff performed its own assessment of the results presented in ANP-10332P and questioned several of Framatomes conclusions. [ '''''''''''''''''' ''' ''''''' ''''''''' ''''''''''''''''''

''''''''''''] The situation is illustrated below in Figure 4.

The results shown in Figure 4 drew the NRC staffs attention, since this presentation of the results calculated by Framatome appears to contradict the conclusion stated in ANP-10332P that ['''''''''''''''''''''' ''''''''''' ''''''''''' ''''''''''''' ' '' ''''' ''''''''''''''' '''''''''''''''''''] Furthermore, as observed previously, it is not obvious a priori whether the co-current or countercurrent flow regime will provide more limiting results.

Figure 4: Average Fraction of Time ['''''''''' ''''''''''''''''''''''''''''''''' ''''''

'' ''''''' ''''''''''''''' '''''''''''''''' '''''''''''''''''' ''''''''] in ANP-10332P The NRC staffs review further observed [''' ''''''''''''''''''''''''' ' '' '''' '''''''''''''' ''''''''''''''''''' '

''''''''' ''''''''''''''''''' ' ''''''''''' ''''''''' ]. While the NRC staff found Framatomes primary scaling rationale for the SSTF reasonable ['''''' ''''''''''''''''''' '' ''''''''' ''''''''''''''''''' ''''''''''' '

''''''''' ''''''' ''' '''''''''''''''' ''''''''''''''' ''''''''''''''''' ''''''''''''''''' '''' ''''''' ''''''''''''' ''''''''''] remains a relevant similarity parameter that was not specifically accounted for in ANP-10332P. In particular, [''' '''' ' ''''''''''''''' '''''''''] may influence which parallel flow regime it is predicted to enter as well as the duration over which it is predicted to maintain that flow regime.

- 133 -

Table 11: Fuel Channel Grouping Resolution for AURORA-B LOCA Evaluation Model Operating BWRs SSTF Small Large Fuel Channels ' ''''' '''''

S-RELAP5 Channel Groups ' ' '

Fuel Channels per Group '''' ''''' '''''

Beyond this, the NRC staff noted that a number of additional factors play a role in the complex and sensitive determination of which flow regime a given channel will enter during the blowdown phase of a LOCA event:

• the radial power distribution in the core (i.e., the flatter the power distribution, the less predictability expected)

• the core spray distribution

• upper plenum phenomena, including o spray droplet condensation o spray nozzle submergence59 o liquid accumulation o mixing/subcooling distribution of accumulated liquid The NRC staff questioned the degree of accuracy in the modeling of these phenomena in the SSTF sensitivity study and in the AURORA-B LOCA evaluation model in general. These concerns led the NRC staff to issue multi-part RAI 32, which addressed concerns with the capability of the AURORA-B LOCA evaluation model to predict parallel channel flow behavior.

Recognizing the challenge with addressing this set of complex issues, the NRC staff further indicated in RAI 32 that Framatome could alternatively satisfy the NRC staffs concerns by demonstrating that significant differences in the calculated figures of merit do not result when the flow regime for the hot channel is varied between co-current upflow and countercurrent flow, such that the impacts of flow regime transition may be considered bounded by the impact of the hot-channel downflow restriction and other evaluation model conservatisms.

In response to RAI 32, Framatome ['''''''''''''''' ' ''' '''''''''''''''''''''' '''''''''''''''''' ''''''''''''''''''' '''''''''

'''''''''''''''''' ''' '''''''''''''''''''''''''''''''' '''' ''''''''''''']:

• [''''' '''' '''''''''''''''' '''''''' ''''''''' ''' '''''''''''''''''''''''''''''' ''''' ''''''''''''''''' '''''''''''''''''''''''' ''

''''''''''''''' ' ''''''''''''''''' ' '' ''''''''''''''''''''' ''''''''''' '''''''''']

• [''''' ''''''''''' '''''''''''''''''' ''''''' ''''''''''''' ''' ''''' '' ''''''' ' '' ''''''''''''' ''''''''''' ''''

59 Note that the AURORA-B LOCA evaluation model [assumes the same core spray distribution regardless of whether the spray nozzles are submerged or not].

- 134 -

'''''''''''' ''''''' '''''' ''' '''''''''''''''' '''''''' '''''''''''''''''' '' '''''''''''''''' '''''''''''''''''' '' ']

• [''''' ''''''' '''''''''''''''''''' ''''''''' '''' '''''''''''''''' '''''''''''''' '' ''' '''''''''' ''''''''''' ''' '''''''

''''''''''' ' ''''''''' ]

The NRC staff found the results of Framatomes sensitivity studies relevant and informative.

The studies generally increased the NRC staffs assurance [''''' '' '''''''''''''''''' '' ''''''''''''']

in S-RELAP5 is conservative. However, the NRC staff ultimately concluded that the sensitivity studies were not fully responsive to the request made in RAI 32 because they did not satisfactorily characterize the impact on the calculated figures of merit that would result from a transition between the co-current and countercurrent flow regimes. In particular, the behavior of one of the parameters used to infer the hot channel flow regime ['''''' '' ''''''''''''''' ''''''''] was not adequately explained in the RAI response, which ultimately reduced the NRC staffs confidence in the conclusions drawn from the sensitivity study. [''' '''''''''''''''''' ' ''''''''''''''''''''

'''' '''''''''''''''' ' '' '''''''''''''''''''' '''''''''''''''' '''''''' '' '''''''''''''''' '''''''''''''''' '''''''''''''''''''' ''''''''''''''''];

the underlying rationale for the observed behavior and its impact on the NRC staffs concerns was not adequately explained. ['''''''''''''''''''''''''' '''''' ' ''''''' '''''''''''''''''''' ''''''''' ''''''''''' ''' ''''

''''' ''''' '''' '''''''''''''' '''''' '''' ''''''''''''' '' ''''''''''''''''''''''''' ''''''''''''''''''' ''''''''''''''''']

In its updated response to RAI 32 dated October 31, 2018 (Reference 66), Framatome provided

[''''''''''''''''' ''''''''''''''''''' '' '''''' ' ''''''''''''''''''' '''''''''''''''''''' ''''''''' '''''''' '''''''''''''''''''''''''''''''

'''''''''''''' ' ''''''''''' ''''] Although limited description was provided in the updated RAI response, the NRC staff found the sensitivity study valuable because it increases confidence that variation in fuel channel nodalization (e.g., due to plant design differences) would not have a significant effect on the calculated figures of merit with the conservative hot-channel downflow restriction in place.

Despite the somewhat compelling evidence shown in the sensitivity studies performed by Framatome in response to RAI 32, the NRC staff concluded that the fundamental issue in RAI 32 had not been adequately addressed. In particular, the NRC staffs review did not find sufficient evidence to conclude (1) that S-RELAP5 can accurately predict parallel channel flow behavior, (2) that one parallel channel flow regime will generally lead to more limiting heat transfer results than another, or (3) that a transition between parallel channel flow regimes would not significantly affect calculated figures of merit. As such, the NRC staffs review of [''

''' ''''''''''''''' ''''''''''''''''''''''''''''''''' '''''''''''' ''''' ''''''''''''''''' ''''''''] contributed to the NRC staffs decision to impose Limitations and Conditions 8 and 14, ['''''''' ''''''' '''''''' ' '''' ''''''''''''''''''''

- 135 -

''''''''''''].

3.5.6 Break and ECCS Injection Nodalization In RAI 41, the NRC staff observed that ANP-10332P does not appear to provide a basis for the satisfaction of Section I.C.1.d of Appendix K to 10 CFR 50, which states that the nodalization in the vicinity of and including the broken or split sections of piping and the points of ECCS injection shall be chosen to permit a reliable analysis of the thermodynamic history during blowdown. In response, Framatome performed a nodalization sensitivity study to illustrate the consistency of its predictions. As shown in Table R41-1 of its response, Framatome

[''''''''''''''''''''' ''''' '''''''''''''''' ''''''''''' '''' '''''''' ''''''''''''' '''''''''''''''''''''' ''''''''''''''''' ''''''' '''' ''''''''''

' '''''' ''''''''''''''' ''''''''''''''''''''' ''''''''' ' '' '''''''''''''''' '''''''] Based upon this relatively mild variation, the NRC staff found the response to RAI 41 acceptable for the conservative AURORA-B LOCA evaluation model.

3.6 Demonstration Analyses Section 7.7 of ANP-10332P and various RAI responses describe the demonstration analyses Framatome completed to illustrate application of the full AURORA-B LOCA evaluation model to two sample plants (i.e., a BWR/4 and a BWR/6).

The NRC staff paid particular attention to the demonstration analyses in ANP-10332P, auditing input and output files for four demonstration cases. The NRC staff has previously related

['''''''''''' ''' ''''''''''''''''' '''''''''''''''''''' ' '''''''''''''''''''''''''''' ' '''''''''''''''''''''''' '''' ''''''''''''''''''' '''''''''''' ''''

''''' '''''''''''''''' '''' ''''''''''' ' ''''''''''''''''''''''' ' '''''''''''''' '''' ''''''''''''''''''''''' '''''''''' ''''''''''''''''''''' '].

Consequently, the NRC staff found the demonstration cases valuable as both reference material for interpreting and understanding the evaluation model and as a source of insight into code stability and predictability with all Appendix K models and other required features activated.

The NRC staffs review of the demonstration analyses from ANP-10332P, including the results of its audit of demonstration case input decks, is presented below. Since the majority of the NRC staffs review ['''''''''''''' ' '''' ''''''''''''''' '''''''''''' '''''''''''''''''], discussion primarily focuses on that material. However, in RAI 15, the NRC staff requested that Framatome ['''''''''''''''' ''''

''''''''''''''' '' ''''''''''''''' '''''''' ''''''''''' ''''' '''''''''''''''''''''' ''''''''''''''''''''' '''''''''' '''''''''''''''] on the demonstration analysis cases. This information will be discussed specifically in Section 3.6.3.

It should be emphasized that the NRC staffs audit did not encompass specific plant parameters used in the demonstration calculations. Hence, no judgment is passed upon their suitability for plant-specific application. Rather, [''''''' '''''''''''''''''''' '''''''''] must be selected conservatively on an individual basis for each analysis, and will be subject to review, as appropriate, by the NRC staff in connection with license amendment requests or other regulatory processes. The NRC staff has designated this position as Limitation and Condition 22 in Section 5.0 of this SE.

- 136 -

3.6.1 BWR/4 and BWR/6 Demonstration Cases Sections 7.7.4, 7.7.5, and 7.7.6 of ANP-10332P contain analyses demonstrating application of the AURORA-B LOCA evaluation model to two sample plants (i.e., a BWR/4 and a BWR/6).

These demonstration analyses, as supplemented by responses to RAIs 15 and 56, among others, focus primarily on the break spectrum analysis step in the plant safety analysis process described in Section 3.3.5.1 of this SE. In particular, for each sample plant, the demonstration analyses considered the spectrum of postulated recirculation line breaks (both suction and discharge), two axial power profiles (mid- and top-peaked), and a sampling of potentially limiting single-failure scenarios.60 Description is provided in ANP-10332P for each analyzed scenario concerning the assumed plant parameters, initial conditions, failure scenarios, event sequences, and calculated figures of merit. Plots of key calculated parameters as a function of time are included. For the demonstration analyses presented in ANP-10332P ['''' ''''''''' '''''''' ''' '''''''''''''''' '''''''''''''

'''''''''''''''''], Framatome reported limiting conditions and results for each analyzed failure scenario61 as tabulated below. Note in particular that the calculated results in Table 14 are affected by the error in S-RELAP5s calculation of cladding oxidation that was described above in Section 3.3.1.2.4. The impact of correcting this error will be discussed subsequently.

Table 12: Limiting Results for BWR/4 Demonstration Case ([''''''''''''''''] Code Version)

Limiting Condition Local Hot Rod Failure PCT Oxidation Oxidation Scenario Location Size (ft2) Peaking (°F)

(percent) (percent)

SF-BATT ''''''''''''''''''' '' '''' ''''''' 1.00 '''''''

SF-LPCI ''''''''''''''''''''' '' '''' ''''''' '''' '''''''''''''''

SF-LOCA ''''''''''''''' '''''''' '''' '''''' ''''' ''''''

Table 13: Limiting Results for BWR/6 Demonstration Case ([''''''''''''''''] Code Version)

Limiting Condition Local Hot Rod Failure PCT Oxidation Oxidation Scenario Location Size (ft2) Peaking (°F)

(percent) (percent)

SF-LPCI ''''''''''''' ''''''''' ''' ''''''' ''' ''''''''''''''''''

SF-LPCS '''''''''''''' '''''''''' '''' ''''''' '''' '''''''''''''''''''

SF-HPCS '''''''''''''' '''''''''' '''' ''''''' '''' '''''''''''''''''''

60 It should, however, be emphasized that the break spectrum analysis for an actual plant safety analysis

['''' ' '''''''' '''''''''''''''' '''''' ''''''''''''''''''' ''''''''''''''''''''''''''' ' '''''''''''''''''' ''''''''''''''''''''''''''''''''''' '''' '''''''''''

''''' '''''''''''''''''''' '''''''''''''''''''' ''''''''' '''''''''''''''' '''''' '''''''''''' '''''''''''''''''''''''''''].

61 Note that, while a specific description of each single failure scenario is not relevant to this SE (reference instead Tables 7-36 and 7-52 of ANP-10332P), it is worth noting that the scenarios considered in Table 12 and Table 13 involve postulated single failures; whereas, Table 14 involves a beyond-design-basis scenario with multiple failures that leaves available only a single train of the low-pressure core spray system.

- 137 -

Table 14: Limiting Results for BWR/4 Reduced ECCS Case (['''''''''''''''''] Code Version)

Limiting Condition Local Hot Rod PCT (°F) Oxidation Oxidation Location Size (ft2) Peaking (percent) (percent)

'''''''''''''''' 0.4 '''' ''''''''62 ''''' '''''63 The NRC staffs review of the demonstration analyses in ANP-10332P resulted in a number of questions, which will be summarized briefly below. Issues relating particularly to the NRC staffs audit of the input and output files for four demonstration case input decks will be presented below in Section 3.6.2.

In RAI 27, the NRC staff requested that Framatome ['''''''''''' ''''''''''''' '' ''' ''' '''''''''''''''

'' ''''''''''''''''''''''''''' ''''''''''''''''' ''''''''''''''''''' ' '''''''''''''''''']. Framatomes response affirmed that the ANP-10332P demonstration analyses ['''''''''''''' ' '''''''''''''''' '''''''''''''''' ''''''''''' ''''' '''''''''''''' '''''

''''''''''''''''''''' ''''''''''' ' '' ' '''''''''''''''''''''''''' ' '''''''''' '''''' ''''''''' ''''''''''''''''']. The NRC staff found Framatomes response consistent with the intent of the EMDAP paradigm.

In RAI 56, the NRC staff requested that Framatome complete additional simulations for the BWR/6 case ['''' ' ''''''''''''''''''' ''''''''' ''''''''''''' ''''''''' ''''' ''' '''''''''''''' '''''' ''''''''''''''' '

'''''''''''''''''''''''' ''''''''''''''''''''''''']. The NRC staff considered these simulations of interest due to the differences in the ECCS configuration between the BWR/4 and BWR/6 designs and [''

''''''''''''''']. Framatomes response provided the requested results, which, as may be inferred from Table 13, [''''''''''' ''''''''''''' '''''''' '''''''''' '''''''' '''''''''''''' '''''''''''''''''''''''''''' ''''' ''

''''''''''''''''''' '''''''''] presented in ANP-10332P.

In RAI 54, the NRC staff requested that Framatome explain deviations between [''

'''''''''''''''']. In Section 7.7 of ANP-10332P, Framatome presented peak cladding temperature results for the BWR/4 and BWR/6 demonstration cases in ['''''''''''' '''''' '''''''''' ''''' ''''''''''

'''''''' ']. As shown in the example spectrum plot included below as Figure 5, the NRC staffs review in many cases found a disconnect between the calculated results for [''

'''''''''' ']. Although perfect agreement is not expected, the large magnitude of some deviations

['''''' ''''' '''] caused the NRC staff to request additional explanation.

62 Note that Section 7.7.5 of ANP-10332P cites ['''' '''''''''''''''' '''''''' ''''''''''''''' '''''''''''''''''''''''''' '''''''''' ''''''''

'''' '''''' ''' '''''''''''''''' '''''''''''''''''''''''' '''''''''''''''''''' ''''''''''''''''' '''' '''' ''''' ''''''' ' '' '''''''''''' '']

63 ANP-10332P notes that the [''' ''''''''''''''''' '''''''''''''' '''' ''''''''''''''''''''' ' '''''' ''''''''''''''], which conservatively satisfies the 1 percent criterion in 10 CFR 50.46.

- 138 -

Figure 5: Break Spectra for BWR/6 Demonstration Case, Recirculation Line Break, Low-Pressure Core Spray Single-Failure64 In response to RAI 54, Framatome explained [''''' '' '''''''''''''''''' ''''''''''''''''''''' ' ''''''''''''''''' '''''''

'''''''''''''''' ''''''''''''''''''''' ' '' '''''''''''' ''''''''''' '''' ''''''''''''''''' ' ''''''''''''''''''' ' ''''''''' ''''''''''''''' ].

In particular, Framatomes [''''''''''''''''' ' ''''''''''''''''''''''''''''''''''''''''''''' ''''''''' ''''''''''''' ''''''''''''''' ''''

''''''''''' '''''''''' '''' ] The NRC staff agreed with Framatomes response to RAI 54, finding it physically well-reasoned and supported by sufficient evidence.

In RAI 58, the NRC staff requested that Framatome illustrate the margin associated with the conservative positions specified in Appendix K by [''''''''''''''''''''''' '''''''''''''''''''' '''''''''''''''''''' '''''

64 Note that in the figure legend, [''''''''''' '''''''''''''''' ''''''''''''''''''''''''''' ''''''''''''''''' ''''''''''']

- 139 -

''''''''''''''''''''''''''' '''' '''''''''''''''''''''''' ''''''''''''''' '''''''''''''''''''''''' ''''''''''''' '''''''''''''''''' '''''''''''''''''''' '''''''''''']

• ['''''' ''''''' ''''''' '''''''''''''''' ''''''''''' ''''''''' '''''' ''''''''''''''' '''''''''''''''''''''''' ''''' '''''' '''''''' ],

and

• [''''' '''''' '''''''''''''''' '''''''''''''''''''''''''''' ''' ''''''''''''''''' '''''''' '''''''''''''''''''''' '''''''''''''' ''' ''''''''''

'' '''''''''''''''''''''''' ''''''''''''''''' '''''''''''''''''''']

As a result, the NRC staff could not make an informed assessment of the conservatism of the AURORA-B LOCA evaluation model.

Framatome responded to RAI 58 [ '''''''''''''''' '''''''''''' '''''''''' ''''' '''''''''''''''''''''' ''''''''' '''''''''' '

''''''''''''''''''''''' ''''''''''' ''''''''''''''''''' ''''''''''' ''''''' ''''''''''''''''''' ''''''''''''''''' '] with the following changes:

• [''''''''''''''' '' ''''''''''''''''''''''''' '''''''''''''''' ''''''''''''''''']65

• [''''''''''''' '''''''''''''''' ' '''''''' ' ''''''''''''''' ''''' '''''''''''''''' ''''''''''']

• [''''''''''''' ' '' ''''''''''''''''' '''' ''''''''''' '''''''''' ''''']

• ['''''''''''''''''''''''''' ''' ''''''''''' ''''''''''' ''''' '''''''''' ''''' '' '''''''''''''''''''''''''' '''''''''''''''''''' ''''''']

For both the [''''''''''''' ''''''''''' ''''' '''''''''''''''''''' '''''''''' '''''''''''''''''''''''''''''] case scenarios included in the sensitivity study, a significant conservative margin of ['''''' '''' ' '''' '''''''''''''''' '

'''''''''''''''''''''' '''''' '' '''''] sensitized conservatisms.

3.6.2 NRC Staff Audit of Demonstration Case Files As noted above, to better understand and assess the AURORA-B LOCA evaluation model and its application, the NRC staff requested that Framatome provide files containing input decks, output decks, and graphics files for four demonstration cases included in ANP-10332P:

• BWR/4, mid-peaked axial profile, double-ended guillotine break on the recirculation pump suction line, with an opposite-unit-false-LOCA-signal single failure,

• BWR/6, mid-peaked axial profile, double-ended guillotine break on the recirculation pump suction line, with a low-pressure core spray single failure,

• BWR/4, top-peaked axial profile, 0.1-ft2 split break on the recirculation pump suction line, with a battery single failure, and

• BWR/4, top-peaked axial profile, 0.4-ft2 split break on the recirculation pump discharge line, with multiple failures (i.e., crediting only a single train of low-pressure core spray).

Regarding the fourth case, the NRC staff specifically requested an audit of this case in RAI 115 to illustrate the behavior of S-RELAP5 at conditions near regulatory limits, with particular focus 65 Note in particular that the ['''''''''''''''''''' ''''''''''''''''' '''''''''''''''''''' '''''''' ' ''''''''''''''''''''''''' '''''''''''''''' ''

'''''''''''''''''''''' ''''''''' ''''''''''''''''''' '''''''''], is not explicitly required by Appendix K to 10 CFR 50.

- 140 -

on the prediction of cladding swelling and rupture. Note that the highest peak cladding temperature in the three demonstration cases initially audited by the NRC staff [''''' '''''''' ''

''''''''' '''''''''''''''] neither in rupture nor significant swelling.

As described further below, the NRC staffs audit of these input decks resulted in a number of valuable observations. In particular, several unanticipated issues were identified that required additional effort from Framatome to address and influenced several limitations and conditions in Section 5.0 of this SE.

3.6.2.1 Impact of Spacer Grids In auditing certain demonstration cases provided by Framatome, the NRC staff observed that

[''' '''' '''''' '''''''''''''''' '''''''''' ''' '' '''' '''''''''''''''''' ''''' '' '' '''''''''''' ''''''''''''''''''''''''''

'''''''''''''']. The cause of this behavior was not discussed in ANP-10332P, ['''''''''' '''''''''' '''''''''

'''''''''''''''''''''''' ''''''''''' '''''''''] Therefore, the NRC staff requested in RAI 60 that Framatome explain this behavior, ['''''''''' '''''''''''''''''' '''''''' '' ''''''''''''''''' ''''''' ''''''''' '' ''''' '

'' ''''''' '' ''' '''''' ''''''' '''' '''''''' '''''''''''] in Figure 6.

In response to RAI 60, Framatome explained that the ['''''''''''''''''''''''''' ''''''''''''''''''' ' '''' '''''

'''' '' ''''''''''''''''' ' ''''''''' ''''' ''' '''' '''''''''''' '' ''''''''''''''' ''''''''''''''''''''''''].

Framatome stated that ['''' ''''''''''''''''' '''''''''''''''''''''''''' ''''''' ''''' '''''''''''''''''''''' ' ''''''''' '''''''''''''''''

'''''''''''''''''''''''''''''''] similar behavior to that shown in Figure 6.

- 141 -

Figure 6: [''''''''''''''''' '''''''''''''''''] Temperature [''''''''''''''''] Observed by NRC Staff During Audit The NRC staffs review of Framatomes response to RAI 60 noted [''''' ''''' ''' '''''''''''' '''''''''''

''''' ' '''''66 ''''' ''' '''''''' ''''''''''''''' '''''''''''''''''''''' '''''''''''''''' '''''''' '' '''''' ''''''''''''''''''

''''''''' '' '''' '''''''''' ''''' ''''''''''''''''' '''''''''' '''''''' ' ''''''''''''], the NRC staff did not obtain confidence that plant-specific analysis should allow crediting a significant temperature reduction based upon this approach.

66 Note that this value is consistent with the NRC staffs earlier estimate, based upon audit review of demonstration case results, [ ' ''''''''''''' ' ''''''''''''''''''''' '''''''''''''''''''] (e.g., see Figure 6).

- 142 -

Furthermore, the response to RAI 60 did not address related questions from the NRC staff,

[''''''''''''''''' ''''' '''''''''' '''''''''''''''''' '''''''''''''''''''''''''' ''''''' ''''''''''' ''''''' ''''' ''''''' '''''

'''''''''''''''''''''''''' ''''''''''''' '' '''''''''''' '''''''''''''''''''' ''''' '''''''''' '''''''''''''' ''''' ''''''67' ''

'''''''''''''''''' ''''''''''''''''']. A further audit of the demonstration case output by the NRC staff found

['''' '''''''''' ''' '''''''''''''''''''''''''' ''''''''''''''' ''''''' ' '''''' ''''''''''''''''''' '' ''' '''''''''''''''''''''''''

'''' ''''''''' ''''''''' ''''''']. The NRC staff hypothesized ['' ''''' ''''''' ''''''''' '''''''''''''''''' ''''

''''''''''''''''''']. However, the audit did not provide conclusive demonstration of this hypothesis, and neither did Framatomes response adequately clarify the issue.

Framatome provided additional information relevant to the NRC staffs concerns in its updated response to RAI 60 dated October 31, 2018 (Reference 66). In particular, Framatome

['''''''''''''''''' '''''''''''''''''' '''''''' ''''' '''''''''' ''' ''''' '''''''''''''' '''''''''''''''''' ''''''''''' ''' ''''''''''''

'''''''''''''''''''''''' '''''''''''''''''' '' '''''''''''' ''' ''''''''''''''''' '''''' '''''''''''''''''''' ]

Table 15: Sensitivity of Peak Cladding Temperature to Interphase Friction Enhancement Peaking Peak Cladding Temperature (°F) Peak Cladding Elevation Unenhanced Enhanced Temperature (fraction of Friction Factor Friction Factor Difference (°F) core height)

[''''' '''''''''''' ''' '''''''''''''''''' ''''''' '''''''''''' ''''' ''''''''''''''''''' '' '''''''''''''''''''' '''''''''

''''''''''''''' ''''''''''''''''''''''''''']. However, the NRC staff observed another relevant trend in the results calculated by Framatome, which is emphasized below in Table 16.

67 THTF mixture level tests are discussed in Section 3.4.3.5 of this SE.

- 143 -

Table 16: Sensitivity of Axial Peaking Elevation Relative to Baseline Case Peaking Unenhanced Friction Factor Enhanced Friction Factor Elevation Peak Cladding Temperature Peak Cladding Temperature (fraction of Temperature Difference vs. Temperature Difference vs. Base core height) (°F) Base Case (°F) (°F) Case (°F)

[''''' ''''''''''''' ''''''''''''''''''' ' Table 16 ''''''''''' ''''''''' ' ''''''''''''''''''''''''''' ' '' '''''' ''''''''''''''

''''''''''''''''''''''''' ' '' '''''''''''''''''' '' ''''''' '''''''''''' '''''''''''''''''' ' '''''''''''' ''''''''''''' '''''''''' '].

Table 16 further suggests this dependency is not associated with ['' ''''''''''''''''''''' ''''''''''

''''''''''''''''''''''''''''''], contrary to the premise on which Framatome based its response to RAI 60, as described above. Framatome further reiterated this premise in its updated response to RAI 60,

['''''''''''''''''''' '' '''''''''''''''' ''''' ' ''''''''''' '''' ''''''''''' ''''''' ' '' ''''''''''''''''' '''''''''''''''''''''''' ' '''

'''''''''''''''''''''''''']

Ultimately, the NRC staffs review concluded that (1) [''' '''''''''''''''' '''''''''''' '''''''' ' '''

'''''''''''''''''''''''' '''''''' '''''''''''''''''''''' '''''''''''] has the potential to affect the calculation of relevant figures of merit significantly and (2) the evaluation model proposed by Framtome does not adequately account for this significant effect. Therefore, the NRC staff imposed Limitation and Condition 23 [ '''''''''''''' '''' ''''''''''' '''''''''''''''' '''''''''''''''''''''' ''''' '''' '''''''''''''''''''''''' '''''''''''

''''''''''''''''''''' '''''''''' ''''' ''''''''''''''''' '''''''' '''''''''''' '''''''' ''' ''''''']

3.6.2.2 Hot-Channel [''''''''''''''''''' '''''''''''''''''''''']

The NRC staff audited Framatomes implementation [ '''' '''''''''''''''''' '''''''''''''''' ' '' ''

'''''''''''''''' ' '''''''''' '''' ] for several of the demonstration cases. The NRC staff confirmed that, in all audited cases, [''''''''' '''''''''''''''' '''' ''''''''''''''' ' '''' '' ''''''''''''''' '''''''''''''' '' ''''

'' ''''''' '''''''''''''' '''''''''''''''''''''''''' '' '' ''']. In fact, similar behavior was observed between the BWR/4 and BWR/6 large-break demonstration cases.

However, the NRC staffs audit found that Framatomes stated criterion ['''' ''' '''''''''''''''' '

''''' '''''''''''] was not an accurate reflection of the modeling practice actually used in the three BWR/4 demonstration cases. [' '''''''''''''''''' '' ''''' '' ''''''''''''''''' '''''''''''''''' '''' ''''''' ''

'''''''' '''''''''' '' ''''''''''' '''''''''' ''''' ''''''''''' ' ''''' ''''''''''''''''''''''''''' ' ''''''''''''''''] During the review, Framatome clarified that, due to an input error, [''' ''''''''''''''' ''''''''''''''''''' ''''' ''''''''

'''''''''''''' ''''''''' ''''''' '''''''''''''''' ' '' '''''''''' '''''''''''''']. However, Framatome observed that the correct setpoint ['' ''''''''' ''' ''''''''''''''''''' ''''''''''''''''''] was used in the BWR/6 demonstration case. [''''''''''''''' '''' ''''''''''''''' ''''''''''''''''' ''' ''''''' '''''''''''''''' '' ''''' ' '''''' ''''''''''''''

- 144 -

'''''' '' ''''''''''''' '''''' '''''''''''''' ''' ''''''''' '] While a mild reduction in calculated cladding oxidation could have resulted, this possibility does not detract from the purpose of the demonstration analyses.

In response to RAI 111, Framatome concluded that the ['''''''''''''''''''''''' '''''''''''''''' '''''''''''''''']

should provide a conservative impact for limiting small- and large-break LOCAs. The NRC staffs audit of the demonstration cases indicated that, although in some cases ['''' ''''''''''''''''

''''''''''''''''''] appeared to have a significant, conservative impact on the calculated figures of merit, in other cases its effect did not appear significant. For example, [' '''' '''''''''''''' ''''''

'''''''''''''' ' ' ''''''''''''''''''''' '''''''''''''' '''''''''''''''''''' ''''''''''''''''''''''''''' ' ] Evidence supporting this viewpoint was observed in the audit of both the BWR/4 and BWR/6 large-break demonstration cases, [' '''' ''''''''''''''''''''' '''''''''''''''''' '''' '' '''''''''''''' '' '''''''''''''''' ''' '''' '''''''''''''''''' '''''''''

'' ''' ''''''''''''' ''' '''''''''''''''''''''''' ''''''' ''''''' ''' '''''''''''''''' '''''''''''''''''''' '''' '''''].

Framatomes response to RAI 111 ['''''''''''''''''''''''''''' '''' '' '''''''''''''''''''' ''''''''''''''' ''''''''''''''''''' '''''

'' '''''''''''] Because it relied solely upon qualitative reasoning and did not address apparently contradictory quantitative evidence from the audited demonstration analysis cases, the NRC staff did not find Framatomes response to RAI 111 compelling [''''' '''''''''''''' '' '''''''''''''''''

''''''''''' ''''''''''''''''''']

On the other hand, the available evidence indicates ['''' ''''''''''''''''' ' ''' '''''''''''''''''' ''''''''''''''''''''

'''''''''''''''' ''''' '''''''''' '''''' ''''''''''''''''''''''''' '''''''''''] While ultimately agreeing with Framatomes conclusion that ['''''''''''''''''''''''''''''''''' '' '' ''''''''''''''''''' ''''''''''''''''''] is appropriate and conservative, the NRC staff found that the magnitude of the conservatism may vary significantly for large break sizes.

- 145 -

3.6.2.3 Heat Transfer Modes and Appendix K Lockouts The NRC staff examined the output of Framatomes demonstration cases to assess the heat transfer modes reported by S-RELAP5, particularly for the hot node of the hot rod.

During the period of core heatup, the NRC staff confirmed that the heat transfer mode for the hot node of the audited small- and large-break LOCA cases [''''''''''''''''''''' ''''''''''''''' '''''''''''''''''''''''''

''''''''''' ''''''''''''] than was seen in similar LOCA scenarios in NRC staff confirmatory calculations using the TRACE code.

While the S-RELAP5 and TRACE results are not directly comparable, due in part to modeling differences, the NRC staff noted that, for the AURORA-B LOCA evaluation model, Framatome

['''' ''''''' '''' ''''' '''''''''''''' '' '''''''''''''''''''''''''' '''''''''''''''' '''' '''''''''''' ''' '''''''''''''''''''''''' '''''''''

''''' '''''''''''''' ''''' '''''''''' '''''''''''''''']. The NRC staff found the revised S-RELAP5 transition criterion [''''''''''' '''''' '' ''''''''''] used in comparable thermal-hydraulic codes including RELAP5-3D ['''''''' '''''''''''''''''''''' ''''] and TRACE ['''''''''''''' '''''''''''''''''''''' ''']. From the NRC staffs perspective, S-RELAP5s usage of such a precise criterion is questionable because existing thermal-hydraulic codes are not capable of accurately predicting ['''''''''''''''' ' '''''

''''''''''''] on the order of [''''' '''''''''''''']. Furthermore, at such miniscule quantities, it is doubtful that ['' '''''' ''''''''''' '''''''' '''' ''''''''''''''''''''''''' '''' '' '''''''''''''''''''' ''''' ''''''''''''''' ''''''''''''''].

Hence, it appears that noise-level variation in the S-RELAP5-predicted [''''' ''''''''''''''] could be responsible for arbitrary transitions between [''' ''''' '''''''''' '''' '''''''''''''''''''''''''' ''''''''' '''''''

'''''''''''''' '''''''''''] A sample plot of the heat transfer mode for the peak node of the hot rod from the S-RELAP5 prediction of the BWR/4 large-break LOCA case is provided below as Figure 7.

- 146 -

Figure 7: S-RELAP5-Predicted Heat Transfer Mode at Peak Node for BWR/4 Large-Break Demonstration Case Ultimately, however, the NRC staff found no evidence that ['''

''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' ''''] resulted in a nonconservative impact on the assessment and validation results or demonstration cases. In particular, for the demonstration cases,

[''''''''''''''''' ''''''''''''' ' ''''' ''''''''''''' ''''''' ''''' ''''' '''''''''''''''' ' ''''''''''''''''''' ''''' ''''''''''''''

''''''''''''''''''''''''']. Therefore, despite reservations concerning the physical and computational bases for Framatomes proposed transition criterion, considering the evidence from the NRC staffs audit of the demonstration cases, as well as the generally conservative assessment comparisons discussed in Section 3.4, the NRC staff ultimately found the current approach acceptable for the conservative AURORA-B LOCA evaluation model.

In addition, the NRC staff considered it worthwhile to assess the impacts of the Appendix K heat transfer lockouts in the demonstration cases. The times of imposition and release of both required heat transfer lockouts are shown below in Table 17.

- 147 -

Table 17: Appendix K Heat Transfer Lockout Imposition in Demonstration Cases Nucleate Boiling Transition Boiling Lockout Case Lockout Imposition Lockout Imposition Release Time (s) Time (s) Time (s)

It is apparent that, as the break size is reduced, imposition of the Appendix K heat transfer lockouts [ '''''''''''' ''''' ''''' ''''''''''''' ' '''''''''''''''''] Limited evidence available from the demonstration cases suggests that the lockouts have [' '''''''''''''''''''''''' ''' ''' '''''''''''''''''''''''''''''''

'' '''''' ' ] Framatomes response to RAI 17 further suggests that, [' '' ''''''''''''''''''''''

'' ''''''''' '''''''''''''''''']68 For all audited scenarios, the cladding temperature rise was observed to be turned around and effectively mitigated [' ''' ''' '''''''''''' '''''''''''''' ''''' ''''''''''''' '''''''''''']. Around the time of cladding temperature turnaround, the flow regime was observed [ ''''''''''''''' '''''''''''''' '''' ''''

'''''''''''' '''''' ''''''].

3.6.2.4 Swelling and Rupture The NRC staffs audit of Framatomes demonstration input decks discovered that swelling and rupture modeling ['''' '''''' ''''''''''''''' ''''' ' '''' '' '''' '''''''''''''' '' ''''''''''''''''

'''' '''''''''''''''' '''''''' ''''], which resulted in the issuance of RAI 108.

Framatomes response to RAI 108 justified its approach by noting that the modeling of swelling and rupture ['' ''''''''''''' '''' '''''''''''''''''''''''' ''''' ''''' ''''''''' '''''''''' '' ''''''''' '''''''''''''''''''' ''''''

''''''''''''''''' ''''''''''''''''''''''''' '''''''''' '' ''''''''''''''''''' ''''''''''''''''' '''''''''' ] As discussed further below in Section 3.6.2.7, the NRC staff found Framatomes approach ['' ''''''''''''''''' '''''''''''''' ''''' ''''''''''

' '''''''''''' '''' '''''''''''''''''''''''' ''''''] to be conservative and appropriate. However, the NRC staff disagreed with Framatomes characterization that Appendix K may be violated by [''''''''''''''''

''''''''''''''''''''']

68 Positing an average heatup rate [ '' ''''' '''''''' '''''''' ''''''''''''''''' ''''''''''''''''''''' ''' '''''''''''' ' '].

- 148 -

3.6.2.5 Calculation of Oxidation As noted above, an error in S-RELAP5 resulted in an underestimation of the calculated cladding oxidation that particularly affected one of the demonstration cases in ANP-10332P (see Table 14, above). In conjunction with reviewing this error, the NRC staff further requested in RAIs 115 and 126 that Framatome provide updated results for the affected demonstration case from Section 7.7.5 of ANP-10332P and make the analysis files available for audit. As noted above, this case considered a bounding, beyond-design-basis scenario with reduced ECCS availability.

Because the three other demonstration cases initially audited by the NRC staff [''' '''' '''''''''''''

''''''''''''''''''''''''''' ''''''''''''''' '''''''''''''''''' ''''''''''].

Framatomes responses to RAIs 115 and 126 provided updated demonstration case results;

['''''''''''''''' ''''''''''' '' '''''''''''''' '''''''''''''''''' ''''''' '''''''''''''' ''''' ' '''''' ''''''' ' ''''''''''''''''''

''''''''''''''''' ''''''''''''' '' ''''''''''''''''''''''' ''''''''' '''' '' ' '''' '''''''''''''''''' '''].

In light of unresolved issues with Framatomes corrected methodology for calculating the percentage of the cladding thickness affected by oxidation (see Section 3.3.1.2.4, above), the NRC staff performed an independent confirmatory calculation using the TRACE code with a simplified single-channel model. Appropriate, time-dependent boundary conditions were devised to reproduce temperature profiles similar to that calculated by Framatome for the Ring 1 hot channel in the BWR/4 reduced ECCS case described in response to RAI 115. While exact reproduction of the S-RELAP5 temperature profile (and hence calculated local oxidation) was not feasible, the NRC staff generated representative curves that approximately envelop Framatomes calculated temperature profile, as shown below in Figure 8.

- 149 -

Figure 1: Cladding Temperature Profiles Used in NRC Staff Confirmatory Calculation for Cladding Oxidation Table 18 shows the results of the confirmatory calculation, which indicate that the maximum local oxidation values calculated by TRACE using the Baker-Just correlation similarly envelop the corresponding value calculated by S-RELAP5. While such a code-to-code comparison does not constitute validation, such a demonstration of expected behavior increases the NRC staffs confidence that the calculation of oxidation in S-RELAP5 is being performed appropriately.

Table 18: Results of TRACE Confirmatory Calculation of Cladding Oxidation Curve Calculated Oxidation S-RELAP5 '''''' ''''''''''''''

TRACE - Underestimate ''''' '''''''''''''

TRACE - Overestimate ''''' ''''''''''''''

Comparing these results to those presented earlier in Table 14 shows that [''''''''''''''''' '''

''''''''''''''''''''''''''''''''''] In particular, when cladding ['''''''''''''''' '''''''''''''' '''''''''''''''''' '' '''

- 150 -

'''''''''''''''''''' '' ''''''''''''''''' ''''''''''''''''''' '' ''''''''''''''' '''''''' ''''''''' '''' ''''''' ' '''''''''''''''''''' '']

on the calculation of maximum local oxidation.

3.6.2.6 BWR/4 Reduced ECCS Case The NRC staffs audit of the BWR/4 reduced ECCS demonstration case, discussed in response to RAI 115, discovered a significant difference in [''' ''''''''''''''''''' '''''''''' ''''''' '' ''' ''

''''' '''''''''''' ' '''''''''' ''''''' '''''''' ' '']. As shown below in Table 19, while the peak cladding temperature ['''''''''' '''' '' '' '''''''''''''''''' ''''' ''''''''''''''''' ''''' ''''''''''' ''''''' ''''''''

'''''''''''''''''''''' ''].

Table 19: Figures of Merit for BWR/4 Reduced ECCS Case69 Figure of Merit '''''' ' ''''''' '

Peak Cladding Temperature (°F) '''''''' ''''''

Maximum Local Oxidation '''''' ''

(percent)

Figure 9, below, plots S-RELAP5-predicted cladding temperatures [ '' ''''''''''' ''''''''' ' ''

''''''''' ' ''''' '''''''''' ''''''''' '''''''''' ''' ''''''''''' '''''''''''''''' ''''''''''''''''' '''''''''''' '''''''''''] While results are shown in this SE only for ['' '''''''''''''''''''''' '''' '''''' ' ''' ''''' ''' '' '''''''''''''''],

similar systematic differences in behavior were also seen between ['''''''''''''''''''''''''''''' '''''''''''''''''''''''

'''''''''''''' '''''''''''''''''' ''''''''' '''' ]

In investigating the cause of the significant discrepancy ['''''''''''''''' ''' '''''' '' ''''' ''

'''''''''''''''' '' '''' '''''''' ''''''''''' ''''''''' ], the NRC staff observed ['' '''''''''''''''''''''''''''' '

''''''''''''''' '' '''''''''''''''''''''''' ] 70,71 69 The analysis described in response to RAI 115 ['''''''' '''''''''''''' ''''''''''' ''' '''''''''''''' '''''''''''''

''''''''''' '' ''' '''''''''' '''' ] Therefore, the magnitude of the calculated figures of merit differs from the values shown above in Table 14.

70 In its review of this issue, the NRC staff noticed an inconsistency between the nodalization Framatome used in the demonstration cases and the nodalization diagram shown in Figure 6-6 of ANP-10332P. In particular, [''' '''''''''''''''''''''''''' '''''''' '''''' '''' '''''''''''''''' ''' '''''''''''''''''''''''' '''' ''''''''''''''''' ''' ' '''''''' ''''

''''''''''''' ''''' ''' ''''''' '''''''''''''''' ''''''''''''''''' '''''''''' ''' '''''''''' ''''''' ''''''''''''''''' ''' '''''].

71

['''''''''' '''''''''''''' ''''''' '''''''''''''''''''''''' '' '''''''''''''''''''''''''''' ''''''' ''''''' '''' ' ''' '''''''' '''''''''''''

'''' '''''' ''''''''''''''''' ''' ''''''''''''''''''''' '''''''''''' '''''''''''''' '' '''''''''''' ''''''']

- 151 -

Figure 2: S-RELAP5-Predicted Peak Cladding Temperature at Limiting Nodes for BWR/4 Reduced ECCS Case Figure 3: S-RELAP5-Predicted Lower Plenum ['''''''''' ''''''''' ''''''''''''''' ''''''''' '' '''' ]

for BWR/4 Reduced ECCS Case

- 152 -

The NRC staff is not aware of direct evidence sufficient to confirm or refute asymmetric behavior such as that predicted by S-RELAP5 for the BWR/4 reduced ECCS case. In particular, [''''''

''''''''''''''''''''''''''' '''''''''''''''''' '''' ''''''''''''''''''' '''''''''''''''''''''''''''' '''''''' '''''''''' '''''''''''''] While code comparisons are not definitive, as a source of insight into the S-RELAP5 predictions, the NRC staff performed a set of confirmatory calculations using the TRACE code. The NRC staffs calculations corroborated the potential for divergence [''''''''''''''''' ''' ''''''''''''''''' ''''''''''''''''''''''

'''''''' ' '' ''''''' '''' '' '''''''''''''''' ''''''''''''''''''''' '' '''''''''''''''' ']. While [''''''''' '''''''''''

''''''''] observed in the NRC staffs calculations were significantly reduced [''''''''''''''''''''''''' ''''

''''' '''''''''''''] and more stable relative to those calculated by S-RELAP5, ['''''''''''''''''' ''''

''''''''''''''' '''''' ''''''''''''''''''' ''''''''' ]

Considering the evidence discussed above, the NRC staff determined the following:

• [''''''''''' ''''''''''''''''''''''' '''''''''''''''' ''' '''''' '' ''''' ' '' '''''''''''''''' ''''' ''''''''' '''''''''''''''''''''

'''''''' ' '' ''''''' '''''''''''] were observed only in the BWR/4 reduced ECCS case, which involves challenging, beyond-design-basis conditions. In particular, with the low-pressure coolant injection system assumed inoperative, coolant injected by the single available train of low-pressure core spray essentially saturates prior to draining into the lower plenum. [''''' '''''''''''''''''' '''''''''''''' '''' '''''''''''''''''' '''' ''''''''''''''''' ''

''''''' '''''''''''''' '''''''''''''''' '''''' ''''''''''' '''' '' ''' ''''''''''''''''' '''''''''''''''''''''''' '''''''''''''''].

• Provided that ['''''''''''''' '' '''''''''''''''''' ''' '''''''''''' ' '''''''''' ' '''' ''' ''''''''''' '''''

'''''''''''''' ''''''''''''''''''' ''''' ''''''''''''''' '''' ''''''''''''''''' ''' ''''''''''''' ''''''''' '''''''''''''''''''' ''''''']

result in a more conservative overall prediction of figures of merit.

The NRC staff recommends that future plant-specific implementation reviews continue to pay attention to [''' ''''''''''''' ' '''''''''''''''''''' '''''''''''''''' '' '''''''''' '''''''' '''''''''''''''''''' '' '' ''

''''''''''''''''' ' '''''''' ']. To this end, when reporting results in submittals to the NRC, licensees are expected to provide figures of merit ['' ''''''' ''' '''''''''''''''] in accordance with Limitation and Condition 14.

3.6.2.7 Figure-of-Merit Determination In auditing the demonstration case output decks provided by Framatome, the NRC staff unexpectedly discovered that S-RELAP5 predicted figures of merit for some fuel rod heat structures that were more limiting than the results Framatome reported for these cases in ANP-10332P. In particular, the NRC staff observed that

• the ['''''' ' ''' '''''''''''''''] in one audited demonstration case had a peak cladding temperature ['''''''''''''''''''''''''''''' ' ' ''''''''''''' '' ''''''''' ''''''''''''' '''''''''''''

'''''''''''''''''''''''''''],

• an ['''''''''''''' ''''''''''''''' ' ''''' ] for one audited demonstration case ['''' '''''''''' ''''''''

'''''''''''' ''''' ], and

- 153 -

• for ['''''''''''''''''''''''''''' ''''' ' ''' '' ''''''''''''''''''] in the cases audited by the NRC staff, the cladding temperature and local oxidation [ ''' '' '' '''''''' '''''''''''''''''''' ''''''''''''''''''' '

''''' ''' '''''''''''''''''''''''''' '''].

[''''''''''' '''''''''''''''' ''''' ''''''''''''''''' ''''''''''' '' '''''''''''''' ''''''''''' '''''''''' ''''''''''''' ''''''''''''''''''].

The NRC staff addressed these observations in RAIs 61 and 107, requesting that Framatome explain and justify the criteria used for determining figures of merit for reporting purposes.

Framatome responded to RAIs 61 and 107 [' ''''' ''''''''' ' ''''''''''''''''' '''''''''' '''

'''''' '''''' '' '''''''''''''' '' '''].

With regard to the NRC staffs first point, above, Framatome stated [''''' ''' ''''''''''''''''''''''''''''

''''''''''''''''' ''''''''''''''''''''''''' ''''''''''''''''''''''' ' ''' ''''''' '' ' ' ''''''''''''''' ''''''''''''' '''''' ''' ]. As noted above in the discussion regarding Limitation and Condition 14, the NRC staff [''' '''

'''''''''''''''''' '''''''''''''''' ''''''''''''''' ''''''' ''''' ' '' ''''''''''''''''''' ''''''''''''''''''''''']

While the [''''' ' '''' '''''''''''''' ''''''''''''''''' '''''''''''' '' '''''''''''' ''''''' '' ''''''''''''' ''''''''''''''''

'''''''' ''''''''''''''''''''''''''''' ''''''], a set of confirmatory analyses performed by the NRC staff using the TRACE code to assist in understanding the impacts of hot channel radial placement found that the more limiting core ring could not be deterministically predicted a priori. While the NRC staff observed that biasing the core spray distribution in TRACE to favor Ring 2 ([' '''''' '

'''''''''''''''''''''''' '' '''''''''''''''''''' ''''''''''' ''''''''''''''''''''' '''''''''']) generally tended to reduce cladding temperatures relative to Ring 1, in some cases Ring 2 cladding temperatures still set the limiting condition. The apparent randomness associated with the predicted limiting core ring in the NRC staffs confirmatory analyses appears [ '''''''''''''' ''' '''''''''''''''''''' ''''''''' ''''''''''''''''' ''''''''''

'''''''''''''''''''''''' ''''' '''''''''''''''''''''' '''''''''' ''''''''''']

Ultimately, while the AURORA-B LOCA evaluation model has substantial conservative margin,

['''' ''''''''''''' ' '''''''''''''''' '''''''''''''''' '''''''''''' '' '''''''''''''''''''''''''''''''''''''' ''''''''''''''''''''''''' '''''''' '''''

''''''''''''''''''''''' '''''''''' '''''''''''''''''' '''''''''''''''''' '''' '''' ' ''''''''' '''' '''''''''''' ''''''''''''''''']. Hence, the NRC staff designated Limitation and Condition 14 ['' ''' '''''''''''''''''''''''' ''''''''' '''''''''''''''''''

''''''''' ''''''''''''''' ''''''''''''' '' '''''''''''''''''' ' '''''''' ' '''' '' '''''''''' '''''].

With regard to the NRC staffs [''''''''''' '''''''' '''''''''''' ''''''''''''''''''' '''' '''''''''''''' '''''''''''''''''''''' '

''''' '''''''''''''''' '''' '''''''''''' ''''''''''''''''''' '''' ' '' ''''''''''''' '' '' ''' '''''''''''''''''''''''''' '''''''],

Framatome stated that the difference is attributable [' '' ''''''''''''''''''''''' '''''''''' ''''''''''''''''''''

- 154 -

''''''''''''''' '''' ''''''''''''''72 '''''''''''' ' '' ''''''''''''''''''''' '''''''''' ''''''' ''''''''''''''''''''''' '' '' '''' ''''''

'''''''''''''''''''' '''''''''''''''' ' '' '' '''' ''''''''''''''' ''''' '], the NRC staff considered it reasonable to infer [''''' '''''''''''''''' ' '''''''''''' '''' ''''''''''''''' '''''' ''''''''''''''' '''''''''''''' '' '''''''''''''' ''''' ''''''''

'''''''''' ' '''''''''''''' '''''''''''''''''' '''''''''''' '' ''' ''''''''''''''' ''''' '''], including the following:

• ['''''''''''''''' '''''''''''''' '''''' '''''''''''''] due to an expanded rod surface area,

• [''''''''''''''''' '''''''''''''''''' ' ''''''' ''''''''''' ''''' ''''' ''''''''''' '' '''''''''''''''] due to an increase in the fuel-to-cladding gap radius, and

• [''''''''''''''''''' '''''' '''''''''''''''''''' ''''' ''' ''''''''''''''''''''''''' ''''''''''''''''] due to an expanded rod surface area.

At high cladding temperatures approaching regulatory limits, the effect of increased heat generation from the metal-water reaction becomes quite significant. However, at lower temperatures ['''''''' ''''''''''''''''''''''''''' '' ''''''''''''''' ''''' '''' '''' ''' ''''''''''' ''''' ''''''

''''''''''' '''''''''''''''' '''''''''''''''''''''''''' '''' '''''''''''''''' '' '''''''''''']. In particular, the evidence provided by Framatome indicates [''''' ''''''''''' '''''''''''''' ''''' '''''''''''''' ''''' ''''''''''''''' ' '''''''''''''' '''''] will in general lead to conservative predictions [' ''' '''' '''' ''''''''''''''''''''' ''''''' '''''''' ' ''''''''''''''''

''''''' ''''''''''''''' '''''' '''''''' ' ' ''''''''''''''' '''''''''''''' '''''' '''''' ' ''' ''' '''' ''''''''''''''''''''''''' ''].

Therefore, the NRC staff considers Framatomes [''''''''''''''''' '''''''''''''''' '' ''''''''''''''' ''''''] to be conservative and acceptable. Furthermore, since the evaluation model [''''''''''''''''' '''''''''''''''''

''''''''''''''''''''''' '''''''' ''''''''''']

With regard to the NRC staffs third point, above, regarding [''' ''''''''''''''''''''''' '' ''''''''''''''''''''''''''''

'''''''''''''''''' '''' ''''''''''' ''' '' ''''''''''''''''''''''''''''''' '''' '' ], Framatome stated that, in the four audited demonstration cases, [''' '''''''' ''''''''''''''' ''''''''''''''''''''''''' ''' ''''''''''''''''''''''''' ''''''

''''''''''''''''''''''''' '''''''' '''''''''''''''''''''' '''''''' ''''''''''''''''' ' '''''''''''''''' '''''' '''''''''''''''' ''''''' '' '''''''']

Framatome attributed the tendency [ '' '''''''' '''''''''''''' '''''''''''''''''''''''''' '' '''''''''''''''''''''' '''''

''''''''''' ''''''''''''''' '''' ''''''''''''' ' '' ''''''''''''']. To illustrate the effect, Framatome performed a sensitivity analysis that was described in its response to RAI 107. The scenario used for the sensitivity case was the BWR/4 reduced ECCS case audited by the NRC staff, [''''''''' '''''''''''''

'''''''''''''''''''''' '''' '''''''''''''''''' ' '''''''''''''''''' ''''''''''''''' ''' '''''''''' ' '''' '''''''''''''''''''''''' '''''''']

The NRC staffs review of Framatomes response and related information found that

['''''''''''''''''''''''''''' '''''''''''''''' '''' ''''''''''' ' ''''''''''''''''''''''''' '''''' '''''''''''''''''''' ' '' '''''''''

72 Note that [''''''''''''''' '''''''''' '' ''' ''''''''''''''''''' '''''''''''''''' ''' ''' ''''''' '''' ''''''].

- 155 -

'''''''' ''''''''''''''''''''']. Therefore, the NRC staff concluded that, in general, [ ''''''''' '

'''''''''''''''''] Therefore, the NRC staff ['' '''''''''''''' '''''''''''''''''''''' ' ''' ''''''''''''''''' ''

'''''''''''''' '''' ''''''''''''' '''''''''' ''''''''''''''''''' '''''' '''''''''''''''' '''''''''' ''''''''''''''''] Therefore, the NRC staff concluded that ['''''''''''''''''''''' ''''''''''' '''''''''''''''''''''''''''''''' '''''' ' '''''''''''''''''''' ''''''''''''''''''''''' ''

''''''''''''''''' ''''''''''''''''''''''''' '''''' '''''' '' '''''''''''''''''''''''''''' ' '''''''''' ''']. This position is captured as Limitation and Condition 24 in Section 5.0 below.

Framatome further added ['''' ' '''''''''''''''''''''' '''''''''''' '''' ' ''''''' '''''''''' '' '''''''

''''''''''''''' ''''''''''''''''''''''' ''''' ''''''''''''''''' ''''' '''''''''''''''' ' '''''' ''''''''''']. The NRC staff considered

[''''' ' ''''''''''''' '''''''''''''''''''''' ''''''''''''' ''''' ''' ' '''''''''''''''' ''''''''''' ''''''' ''' ''''''''''''''''''''''

'''' '''''''''''''''' ''''''''''' ''''' '' '''''''''' ''''''''''''''' '''''''''''''''' '''''''''' '' ''''''''' '''' '''''''' ''''''''' ]

3.6.3 Impact of ['''''''''''''''] Code Version [''''' ''''''' ''''''''''''''''''''''''']

The NRC staffs review of the demonstration analysis results described above generally focuses upon the results provided in ANP-10332P ['''' ''''''' '''''''''''''''''''' '''' ''' ''''''''''''''' '''''''''''''

'''''''''''''''''''']. However, in RAI 15, the NRC staff requested that Framatome address any impacts on the demonstration analyses resulting from its decision to update the [''''''''''''''''''''

''''''''''''''''''''''' ''''''''''' ''''''''''''''''' ''''''''' '''''''''''''' ''' ''''''''''''''''''' ''''''']. In response, Framatome provided tables of peak cladding temperature results for several demonstration analysis scenarios [''''''''''''''''''''' '''''''' '' ''''''''''''''''''' '''''''''''' ''''''''''''''''''']. The limiting peak cladding temperature for each single-failure scenario, as reported by Framatome, is summarized in the following tables, along with the difference in the peak cladding temperature ['''''''''''''''''''' ' '''

'''''''''''''''''' ''' '''''''''''''''' ''''''''''''' ' '''''''''''''''''' ''''''''''''''''''''''' ' ''''''''''''''''''''''''''' ' '''''''''''''''''''''''''].

Table 20: Limiting Results for BWR/4 Demonstration Case (['''''''''''''''] Code Version)

Failure Limiting Condition PCT (°F) PCT (°F)

Scenario Location Size (ft2) Peaking SF-BATT '''''''''''''''''' '''''' '''' ''''''' '

SF-LPCI '''''''''''''' '''''''' ''''' '''''''' '''

SF-LOCA '''''''''''''' ''''''''' '''' ''''''' ''''

- 156 -

Table 21: Limiting Results for BWR/6 Demonstration Case ([''''''''''''''''''] Code Version)

Failure Limiting Condition PCT (°F) PCT (°F)

Scenario Location Size (ft2) Peaking SF-LPCI ''''''''''''''' '''''''' ''''' '''''' '''

SF-LPCS ''''''''''''' ''''''''' '''' ''''''' ''

SF-HPCS ''''''''''''' '''''''''''' ''''' ''''''' '''

Table 22: Limiting Results for BWR/4 Reduced ECCS Case ([''''''''''''''''''] Code Version)73 Limiting Condition PCT (°F) PCT (°F)

Location Size (ft2) Peaking

['''''''''''''''''''' '' ''''' 2119 '']

The above results are largely comparable to the ['''''''''''''' ''''''''''''' ''''''''''''''''''' ''''''''''''''''''''' '

''''''''' '' '''''''''''' ].74 However, the [''''''''''''''''''' '''''''''' '''''''''''''''''' ''''''''''''' ' ''''' ''''''''''''''''''' '

''''''' '''''''''''''''' ''''''''''''''''''''''''' ''''''''''' '' '''''''''''''''' '''''''']. Note that this trend is the opposite of the integral effects comparison results in Table 10. Framatome stated that unspecified error corrections were made in the demonstration case re-runs [''''' '' '''''''''''''''''' '''''''' '''''''''] that affected these comparisons.

In its updated response to RAI 15 dated October 31, 2018 (Reference 66), Framatome

['''''''''''''''''' '''''''''''''''''' '''''''''''''''' '''''''''''''' '''' ''''''''''''''''' '' '''''''''' ''' ''''''''''''''''''' '''''' '''''''

'''''''''' ''''''''''''''''''' '''''''''''''''' '''''''''''''''''''''''''''], including the following:

• ['''''''''''''' '''''''' '''''''''' '''''''' '''''''''' ''''' '''''']

• ['''''''''''''''''' ''''''''''''''' ' '''''''' ''''''''''''''' '''' ''''''''''''' ''''' ''''' '''''''''''''''''' ''''''''''''''

'' ''''''''''''''''']

• [''' ''''''''' ''''''''''' ''''''''' ''''']

• ['''''''''''''''''''''' ' '''''''''''''''''' ' ''''''''''' ''''' ''''''']

However, for some break sizes that were ultimately not limiting for the demonstration plant, unexpectedly large differences in peak cladding temperature were observed, ['''' ''''''''''

''''''''' '''''''' ' ''''''''''' '''''''''']. The origin of the sensitivity could not be discerned from Framatomes response; nor could the NRC staff conclude that the limiting breaks from the demonstration plant break spectrum would necessarily be limiting for all plants to which the AURORA-B LOCA evaluation model would be applied in the future. Hence, the NRC staff could not conclude these large sensitivities are insignificant and the NRC staff imposed Limitation and 73 Note that the ['''''''''''''''''' '''''' ''''''''''''' '''''''' ''''''''''''''' ''' '''''''''''' '' '' '''''''''' ''''''''''''''' '''''''''''

''''' ''' '' ''''''''''''''''''''''' ''''''''''''''''' ''''' '''''''''''''''''].

74 Note in particular that the calculated results in Table 14 are ['''''''''''' ' '' ''''''' '''''''''''''''''''''

'''''''''''''''''''' ''' '''''' '' ' '''''''''''''''''' ''''''''''''''''''''''''].

- 157 -

Condition 25 [' '''''' ''''''''''''''' '''''''''''''''''''''' '''''''''''''' ''''''''''''''''' '''''''''''''''''''' '' ''

'''''''''' '''''''''''''''''''''''''].

For the BWR/4 SF-LOCA scenario shown above in Table 20, the NRC staff observed that [''''

''''''''''''''''''''' ''''''' ' ''' ''''''''''']. As made clear in response to RAI 111, the non-representative practice [' ''''''''''' ''''''''''''''''''''' ''''''''' ''''''''''''''''' '''''''''''''' ''''''''''' ''''''''''''''' '''''''' '''' '''''''''''''''''''

''''''''''''''''''''''''' ''' '''''''''''''' '''''''''].

3.6.4 AURORA-B Comparison Cases Versus EXEM BWR-2000 Model In addition to reviewing the demonstration cases provided in ANP-10332P, the NRC staff requested in RAI 3 that Framatome provide a comparison of the predictions of the AURORA-B LOCA evaluation model against Framatomes current LOCA evaluation model, EXEM BWR-2000. In response, Framatome provided comparative analyses for one small-break and one large-break case that were intended to model the same plant experiencing an identical LOCA scenario from equivalent initial conditions.

Plots of key parameters for the comparison cases for the AURORA-B LOCA and EXEM BWR-2000 evaluation models are included in response to RAI 3 and generally show reasonable agreement. In particular, the plots ['''''''' '''''' ' '''''' '''''''''''''''''''''''' ''''''''' ''' ''''''''''''''''''''''''

''''' '' '''''''''' ''''' '''''''''''''''''''''' ''''''''''' '''' '''''''''''''''''''''' ''''''' ''''''''''''''''''' '' '' 75 '''''''''

''''''''''''''''''' '''''''''''''''''' ''''''''''''''''' ' ''''''''''''''''' ' '''''''' '']

Table 23: Approximate Peak Cladding Temperature Difference [''''''''''''''''''''' '''''

''''''''''''''''''''''''''''''' ' ''' ''''''''''''''''''''''''' ''''''''''''''''''' '''''''''''''''''''''] in EXEM BWR-2000 Comparison Cases Small-Break LOCA Large-Break LOCA

[''''''' ' '''''' ]

Framatome observed that perfect agreement between the AURORA-B LOCA and EXEM BWR-2000 evaluation models is not expected due to fundamental differences in code modeling capabilities. For example, the EXEM BWR-2000 model uses a three-equation, drift-flux homogenous-equilibrium formulation (as opposed to the two-fluid model employed in S-RELAP5). Framatome further stated that the RELAX model used in the EXEM BWR-2000 evaluation model contains ['''''''' ' '''''''''''''''''''''''''''] volumes, as compared to the S-RELAP5 input deck for the AURORA-B LOCA evaluation model ['''' ''''' ''''' '''''''] volumes.

75 While not included in this comparison, [' ''''''''' ''''''''' ''''''''''' '''' ' ''''''''' '''''''''''''''''''' ' ''''''''''''''

''''''''''''''''''''''''''''''''''''] of the EXEM BWR-2000 evaluation model (e.g., Reference 60).

- 158 -

Although some differences are apparent in the comparison plots, the NRC staff found the level of agreement reasonable overall and concluded that the information Framatome presented in response to RAI 3 increases confidence in the conservatism of the AURORA-B LOCA evaluation model. In particular, Framatomes response to RAI 3 ['''''''''''''' ''''''''''''''''''''''' '''''

'''''''''' '''''''''''' '' '''' ''''''''' ''''''''''''''''' ' ' '''''''' ''''' ''' '''''''''' '''''''''''''' ''''''''''].

4.0 ADMINISTRATIVE REQUIREMENTS 4.1 Documentation Acceptance criteria for the documentation of evaluation models for reactor safety analysis are provided in Chapter 15.0.2 of the SRP. The majority of the documentation necessary to support the AURORA-B LOCA evaluation model is contained in ANP-10332P itself. This includes an overview of the evaluation model, a description of the BWR LOCA event and relevant phenomena, the code assessment for S-RELAP5, justification that the requirements of Appendix K to 10 CFR 50 have been addressed, and description of a quality assurance plan.

With the exception of the quality assurance plan (discussed in the following section), this information has been described and evaluated in the foregoing sections of this SE.

In addition, Chapter 15.0.2 of the SRP calls for theory and users manuals for any codes supporting the evaluation model. As noted above, the AURORA-B LOCA evaluation model is based on the S-RELAP5 code, which incorporates a kernel of routines from RODEX4.

Therefore, in support of the present review of ANP-10332P, the NRC staff requested and Framatome supplied for information current versions of its S-RELAP5 code theory manual (Reference 11) and users manual (Reference 12).

The NRC staff frequently referenced these manuals in support of its review of ANP-10332P, particularly during the audit of the demonstration analyses described above in Section 3.6.2.

The NRC staff further sampled relevant sections of these manuals to assess whether they contain adequate descriptions of the S-RELAP5 code, with particular focus upon the recent changes made to support implementation of the AURORA-B LOCA evaluation model.

While the NRC staff concluded that these manuals generally provide the information necessary for analysts to apply the S-RELAP5 code to plant-specific analyses under the AURORA-B LOCA evaluation model, the NRC staffs sample review identified several issues with the code manuals:

• As detailed above in Section 1.2, Framatome [''' '''' ''''''''''''' ''''''''''''''''' ' ''''

'' '' '''''''''' ''''' '''''''''''''''''''''''''''' '''''''''''''''''''' '''''' ] While finding the theory manual generally understandable, the NRC staff believes that improved clarity regarding application-specific modeling practices would be useful. Furthermore, simplification of code manuals by [''''''''' ''''''''''''''''''''''' '''''''''''''''''''''''''''''''' '''''''''''''''''''''' ''''''''''''''''''''''' ''''

'''''''''''''''' ''''''''''' ' ''''''''''' ' '''''' '''''''''''''''''''''''''''''' ''''''''''''''''''''''''''] could be helpful.

• The NRC staff noted that general updating of the code theory manual may be appropriate. For example, [''' ''''''' '''''''''' ''''''''' '''''''''''''''''''' ' ''' '''''' '''''''''''

- 159 -

''''''' ''''''''''''''''''''''''''''''' ' ''''''''''''''' '''''''']

• As discussed above in Section 3.3.1.1, the NRC staff found that revisions to the derivation of the momentum equations in the S-RELAP5 code theory manual had not been made in response to comments from the Advisory Committee on Reactor Safeguards.

• As discussed above in Section 3.3.1.2.4, the NRC staff found that insufficient clarity in the S-RELAP5 code theory manual likely contributed to the persistence of the error in calculating the cladding thickness affected by oxidation. In particular, the equation for the reacted mass of cladding was cast in a non-intuitive form (e.g., compared to theory manuals for TRACE (Reference 44) and RELAP5-3D (Reference 49)), likely making the error difficult to discern. While Framatomes response to RAI 114 was intended to address the NRC staffs concern with this code error, further staff review found that the proposed theory manual revision discussed therein contained additional errors and unclear parameter definitions. Finally, although Framatomes updated response to RAI 114 (Reference 66) addressed the most significant issues identified by the NRC staff, as discussed in Section 3.3.1.2.4, room remains for further improvement.

• Criteria for reporting the heat transfer modes [' '''''' '''''''''''''''' ''''''''''' ''''''''''' ''''''''''''''''

''''''''''''''''''''' '''''' ' '''' '].

• As noted above in Section 3.3.4.1.2, the NRC staff found that the S-RELAP5 code theory manual [''' '' '''''''''''''' '''''''''''''''' ''''''' '''''''''''''' '' ''''''''''''''''' ''''''''''''''''''''''''''''''

'''''''''''''''' '''''''''''''' ''''' '''''''''''].

While the NRC staffs audit of the S-RELAP5 code theory and users manuals generally found them sufficient to support application of the AURORA-B LOCA evaluation model, the NRC staff recommends that Framatome continue to focus on documentation improvements in these areas. The progress of this effort should be monitored during future NRC staff reviews involving the AURORA-B code system.

At a further level of detail, Framatome is developing a modeling guidelines document (Reference 26) and procedures for performing plant-specific analysis with the AURORA-B LOCA evaluation model. The NRC staff audited a draft of the modeling guidelines document for the AURORA-B LOCA evaluation model and, in response to RAI 62, ['''''''''''''''''''''' '''''''''''' '

'''''''''' ''''''''''''''''''' ''' ''''''''''''''' '''''''''''' ' ''''''''''''''''''''''''''].

Particularly in light of the [''''''''''''' '''''''''''''''''''' '''''''''''''' '''''''''' ' ''''''''''] S-RELAP5 code theory manual, the NRC staff considered the modeling guidelines document essential for ensuring consistent implementation of the AURORA-B LOCA evaluation model. As such, the NRC staff ['''''''''''''' ''''''''''''''''''''''' ''''''''''''''''' ''''''' ''''''''''''''''''''' ''''' '''''''''''''''''''' '''''''''''''''''' '''''

- 160 -

''''''''''''''''''''''' ' ''''''''''''''''''''' '''' '''' ''''''''''''] This expectation is formalized as Limitation and Condition 26.

Although it is possible that supporting documentation such as code manuals and modeling guidelines may be subject to audit reviews in the future, these documents are not formally a part of the approved evaluation model; hence, such documents are beyond the scope of the present review and are not explicitly approved. While Framatome may in general make changes to these documents, consistent with discussion below in Section 4.3, any changes implemented must not alter the approved AURORA-B LOCA evaluation model, as described and documented in ANP-10332P and associated RAI responses, or the basis for any conclusions made in this SE.

4.2 Quality Assurance Plan As noted by Framatome in Sections 3.1 and 10.0 of ANP-10332P, because analyses generated using the AURORA-B LOCA evaluation model are important to the safety of nuclear power plants, the evaluation model must be maintained under a quality assurance program that meets the criteria set forth in Appendix B to 10 CFR 50. According to Chapter 15.0.2 of the SRP, a vendors quality assurance plan for accident and transient analysis methods should address, at a minimum, design control, document control, software configuration control and testing, and corrective actions.

Section 10.0 of ANP-10332P summarizes Framatomes quality assurance program, as applicable to the AURORA-B LOCA evaluation model. Framatome stated that its quality assurance program applies throughout all stages of the analysis process, including verification and validation. Framatome further stated that its quality assurance program covers procedures for design control, document control, software configuration control and testing, error identification, and corrective actions. In addition the program covers training for personnel involved with performing analyses, code development, and code maintenance.

The NRC staff generally found the quality assurance program described in ANP-10332P to be appropriate and consistent with regulatory guidance in the SRP.

However, in RAI 64, the NRC staff questioned whether S-RELAP5 contains input checking features consistent with Chapter 15.0.2 of the SRP. In particular, if required inputs are missing or outside an acceptable range, the code should generate error messages and cease executing the calculation. The NRC staff further questioned how the code would respond if the parameters passed to a given model or correlation are outside its validated range, since this scenario also could lead to the generation of nonphysical or otherwise unexpected results.

Framatome responded that ['''''''''''''''''''''' '''''''''''''' '' '''''''' '''''''''''' ''''' '' '''''''''''''''''' '''''' '

'' ''''''''''''' '''''''''''''' '''''' '''' '''''''' '' '''''''''''''''''''''' '''''''''''''''''' ] The NRC staff found Framatomes response to RAI 64 consistent with regulatory guidance and, hence, acceptable.

The NRC staff further noted that, in accordance with Appendix B to 10 CFR 50, reviews performed by vendors and licensees would be responsible for assessing and, as necessary correcting, unexpected behaviors in calculations supporting plant safety analyses.

- 161 -

Chapter 15.0.2 of the SRP and RG 1.203 call for independent reviews of evaluation models at key steps in the development process. The NRC staff recognizes that truly independent review of proprietary vendor evaluation models is typically not achieved prior to submission for NRC staff review. In the case of S-RELAP5, due to the similarity of many features with other code versions in the RELAP5 family that have been subject to extensive peer review over the years, this guidance position may be considered partially satisfied. However, [''''' ' ''''''''''''''''''''

'' '''''''''''''''''''''' '''''''''''''']. In this regard, past and present NRC staff reviews of evaluation models reliant upon the S-RELAP5 code may be considered as independent reviews of code development and assessment; as such, the NRC staff considers the intent of the guidance satisfied.

In RAI 27, the NRC staff requested that Framatome explain how the required set of modeling options for the AURORA-B LOCA evaluation model will be implemented and verified in a manner that minimizes the potential for error and variation. Framatomes response identified that, [''''''' ''''''''''''''''''' ''' ''''''''' '''''''''' ''''''''''''''''''' ''''''''' '''''''''' '''''''''''''''''''' ''''''' ''

''''''''''''''' '''''''''''' '''''''''''''''''''' ''''''''''''''] As a result, the NRC staff concludes there is confidence that the potential for [''''''''''''' ''''''' '''''''''''''''''''] with the inadequate specification of all required modeling options is minimized.

As discussed above, the NRC staffs review found the quality assurance program Framatome proposed for the AURORA-B LOCA evaluation model to be adequate.

4.3 Update Process As submitted on March 25, 2014, in Section 12 of ANP-10332P (Reference 1), Framatome proposed a process for making certain types of modifications to the AURORA-B LOCA evaluation model autonomously in the future. However, based upon its review of Section 12 and associated RAI responses (e.g., RAIs 63, 65, and 127), the NRC staff did not agree with certain aspects of the proposed update process (Reference 65). On March 8, 2019, prior to the issuance of the NRC staffs final SE, Framatome informed the NRC staff of its intention to delete Section 12, in its entirety, from the approved version of ANP-10332P. In response, the NRC staff has abbreviated this section of the SE relative to the draft dated August 31, 2018 (Reference 65).

Based upon its review of ANP-10332P, the NRC staff does not object to autonomously implemented changes that do not substantially alter the approved AURORA-B LOCA evaluation model. For instance, as discussed previously in the NRC staffs SE for ANP-10300P (Reference 8), the NRC staff generally considers certain code changes as having low potential to impact the approved evaluation model, for example:

• enhancing source coding and its structure

• porting a code to a different computational platform

- 162 -

• parallelizing the existing numerical solution method across multiple processors

• updating the physical properties of water to current standards In each of these examples, the effect of such code changes on an approved evaluation model is expected to be minimal, which would be confirmed by the vendors continuity of assessment and quality assurance processes. Furthermore, since the above examples also reside at a level of detail that is typically below what the NRC staff reviews and approves, such changes appear unlikely to affect descriptions or information provided in ANP-10332P and associated RAI responses, or the basis for any conclusions made in the NRC staffs SE.

As a counterexample, the NRC staff does not agree with Framatomes proposal in Section 6.3.7.1 of ANP-10332P that ['''''''''''''''''''''''''' ' '''' '''''''''''''''''''' ''' ''''' ''''''''''''''''''''''

'''' ''''''''''''''' ''''''''' '''''''' '''''''''''''''''''''] may be changed in the future without prior NRC review. [''''' ''''''''''''''''''''' '''''''''''''''''' ' '''''''''''''''''''''''''''' ' ''''''''''''''''''''''' ' ''' '''''' ''''''''' ''''''''''

'''''''''' ' ''''''''''''''''''''''' '''''''''''''''''''' ''''' ''' '''''''''''''''' ''''''''''''''''''''''''''''''] Significant deviations

[''''' ''' ''''''''''''''''''''''' ''''''''''''''''' ''''''''''''''''''''''' ''''''''''''''' '''''''''' '''''''' '''''''''''''''''''''''] are not specifically approved by this SE and would need to be addressed in plant-specific licensing reviews.

Future applications of the AURORA-B LOCA evaluation model may involve analysis of fuel designs that were not explicitly considered during the NRC staffs review of ANP-10332P. While the NRC staffs review of the AURORA-B LOCA evaluation model is not explicitly tied to any particular fuel design(s), the NRC staffs evaluation of [''''''''''''' ''''''''''''' ''''''''''''''''''''''' ''''''''

'''''''''''''''''''''''''' ''''''' ''''''''' '' '''''''''''''' ' '''' ' '''' '''''''''''''']. In light of this dependency, for new or modified fuel designs, the NRC staff finds necessary the demonstration of continued (1) compatibility of the AURORA-B LOCA evaluation model field equations and closure relations, and (2) applicability of the AURORA-B LOCA evaluation model assessment and validation effort. Hence, in practice, Framatome must ensure the continued applicability of existing calculational methods and assessments to all new or modified fuel designs. This stipulation applies, not only to the calculational methods and assessments of the AURORA-B LOCA evaluation model, but also extends to those of all supporting methodologies. In particular, NRC staff review and approval is necessary prior to application of the AURORA-B LOCA evaluation model to any new or modified fuel designs exhibiting deviations from existing fuel design properties and behaviors, such that existing calculational methods or evaluation model assessments for the AURORA-B LOCA evaluation model or supporting methodologies are no longer applicable. This expectation is captured in Limitation and Condition 27. Furthermore, implementation of new, previously approved critical power correlations in steady-state S-RELAP5 calculations as a means of validating (or recalibrating) ['' ''''''''''''''''''''''''' '''

'''''''''''''''''''''''' ' ''' '''''''''' ' '''' ''''''''''''''''''''' '''''' '''''''''' ''''''' ] is acceptable to the NRC staff.

Regarding potential future evaluation model modifications to support the analysis of lead test assemblies (e.g., incorporation of unapproved models and correlations), the NRC staff considers this topic beyond the scope of the present SE. Licensees are responsible for assessing the need for NRC staff review prior to implementing methodology changes, including those they may deem necessary for the analysis of lead test assemblies. The acceptability of

- 163 -

future changes implemented to address lead test assemblies will be evaluated by the NRC staff, as necessary, in subsequent reviews.

In response to RAI 65, Framatome discussed its perspective regarding the latitude 10 CFR 50.46 allows for implementing evaluation model changes. While the NRC staff found the practical result of Framatomes response to RAI 65 reasonable [''''' '''''''''''''''' ''''''''''''''''''

' '''''''''''''''''' ''''''''''''], the NRC staff does not endorse the interpretation of 10 CFR 50.46 provided in Framatomes RAI response. However, in light of Framatomes commitment to delete Section 12 from the approved version of ANP-10332P, the NRC staff no longer considers Framatomes response to RAI 65 relevant to the AURORA-B LOCA evaluation model.

Verification that code changes ['''''''''''''' ''''''''''''''' ''''''''''''''''''' ''''''''' ' ''''''''''''''''' ''''''''''''''''''''

'''''''' '''''' ' '''''''''''''''''' ' '''''''''''''''''''' ''''''' '''''''''''''''''''''''''' ''''''''''''''' ''''''''''''''''''''''' '']. In RAIs 63 and 127, the NRC staff requested additional information concerning [''' ''''''''''''''''' '''

''''''''''''''' '''''''''' '''''''''''' '''''' '''''' ''''''' ''''''''''] While the NRC staffs review of Framatomes responses to these RAIs found that the continuity of assessment process for the AURORA-B LOCA evaluation model incorporates many essential features and good practices, the NRC staff ultimately did not obtain sufficient confidence that [''' ''''''''''''''''' '''''''''''' ''''''''' ' ''''''''''''''' '

''''''''''' ''''''''''''''''''''' '''''''''''''''''''''' '' '''''''''''''''''''' ' ''''''''''''''''''''''' ] associated with the proposed update process that Framatome will delete from Section 12 of the approved version of ANP-10332P, the NRC staff [''' '''' '''''''''''''''' ' '''''''''''''''''' ''''' '''''''''''''' '''''''''''''''''

'''''''''''''''''''''''''' '' ''''''''''''''' ''''''''''''].

In conjunction with its acceptance of the methodology in ANP-10332P, the NRC staff does not specifically approve any new processes for implementing updates or changes that would impact the AURORA-B LOCA evaluation model, as documented in ANP-10332P, Framatomes RAI responses, or the basis for any conclusions made in the NRC staffs SE (e.g., changes to code models, required modeling options and assumptions, nodalization, required inputs, etc.).

As discussed above, Framatome has committed to the removal of Section 12, in its entirety, from the approved version of ANP-10332P. Consequently, the NRC staff finds that the information concerning the proposed update process that Framatome submitted in Section 12 of ANP-10332P in March 2014, along with that contained in associated RAI responses (e.g., 63, 65, 127), is no longer relevant to the AURORA-B LOCA evaluation model. The NRC staff has retained sufficient discussion of this topic in its final SE to promote the general consistency of (1) plant-specific applications of the AURORA-B LOCA evaluation model and (2) update practices among various AURORA-B evaluation models (i.e., References 1, 7, 9).

5.0 LIMITATIONS AND CONDITIONS For reasons explained in the foregoing evaluation, the NRC staff finds the following limitations and conditions necessary, in general, to support application of the AURORA-B LOCA evaluation model for performing safety analysis for BWR/3-6 plants:

1. The AURORA-B LOCA evaluation model shall be supported by an approved nodal core simulator and lattice physics methodology. Plant-specific licensing applications referencing the AURORA-B LOCA evaluation model shall identify the nodal core simulator and lattice

- 164 -

physics methods supporting the AURORA-B LOCA analysis and reference an NRC-approved TR confirming their acceptability for the intended application. (Section 1.1)

2. The full, stand-alone version of the RODEX4 code shall be used in accordance with an approved methodology to supply steady-state fuel thermal-mechanical inputs to the AURORA-B LOCA evaluation model. (Section 1.1)

3. The AURORA-B LOCA evaluation model may not be used to perform analyses that result in any of its constituent components or supporting codes (i.e., S-RELAP5, RODEX4 kernel, RODEX4, core simulator and lattice physics methods) being operated outside approved limits documented in their respective TRs, SEs, code manuals, and plant-specific licensing applications. (Section 1.2)

4. TR ANP-10332P does not provide a technical basis to support satisfaction of the requirement in 10 CFR 50.46(b)(5) for long-term core cooling, and, as such, has not been approved for this purpose. (Section 2.1)

5. As discussed above in Section 2.1, the conclusions of this SE apply only to the use of the AURORA-B LOCA evaluation model for the purpose of demonstrating compliance with relevant regulatory requirements in effect at the time the NRC staffs technical review of ANP-10332P was completed (i.e., as of December 31, 2018). (Section 2.1)

6. This SE does not constitute approval of the PIRT rankings in Table 4-1 of ANP-10332P.

Framatomes PIRT rankings represent an informed opinion of phenomenon importance that the NRC staff referred to as supporting information in its review of the AURORA-B LOCA evaluation model; beyond this, they remain subjective judgments that are not integral to the acceptability of the evaluation model. (Section 3.2)

7. ['''' '''''''''''''''''''''''''' '''''' ''''' '''' '''''''''''''''''' '''''''' '''''' ''''' ''''''' '' ' ''''''''''' ' ''' '''''

'''''''''' '''''''''''''''''''''' '''''''''''''''' ' '' '''''''''''''''''''''' '''''' '''''''''' ''''''''] (Section 3.3.1.2.2)

8. [''''' ''''''''''''''''' '''''''''''''''''''' ''''''' '' ' '''''' '' '''' ''' '''''''''''''' ''''' '' '''''' ''' ''

''''''''''''''' '''''''''''''''''''] (Section 3.3.4.1.1)

9. Safety analyses performed with the AURORA-B LOCA evaluation model may not credit a limit on [''''''' ''''''''''''''' ''''''''''], absent a plant-specific determination from the NRC staff that such credit is consistent with the requirements of 10 CFR 50.36.76 Absent such a determination, ['' ''''''''' ''''''''''''' ''''''' ' ''' ''''''''''''''' '''''' ' '' ' '''''''''''''''' '''''''

''''''''''''''''''''''''''] (Section 3.3.4.1.1)

10. To ensure adequate conservatism in future plant-specific safety analyses, absent specific NRC staff approval for higher values, this SE limits credit for gamma energy deposition 76 Such a determination could be made during the NRC staffs review of a license amendment request implementing the AURORA-B LOCA evaluation model or via a separate licensing action.

- 165 -

outside of a fuel rod to no more than ['''''''''''''''''''' ' '' '''''''''''''''''' ''''''''''' '''''].

(Section 3.3.4.1.1)

11. Plant-specific licensing applications referencing the AURORA-B LOCA evaluation model shall adequately justify the averaging method for determining the temperature ramp rate used in the calculation of cladding swelling and rupture. (Section 3.3.4.1.2)

12. The Appendix K lockout preventing the return to nucleate boiling shall be ['''''''''''''' ' '

'''''''''''''''''''' '''''' '''''']. (Section 3.3.4.1.3)

13. ['''' ''''' ''''''''' ' '''''''''''''''''''''''' ''''''''''' '''''''''''' ''''''' ''''''''''''''' ''''''''' '''' ''''' ''''''''''']

shall be taken into account when determining the start of the refill and reflood phases and the release of Appendix K heat transfer lockouts. (Section 3.3.4.1.4)

14. Plant-specific licensing applications referencing the AURORA-B LOCA evaluation model

['''''' '' '''''''''' '''' ''' '''''''''''''' ''''''' '''''''''' '' ''''' ' ' '''''' '''''' '''''''''''''''''

' ''''''''''''''''''''''''' '''''''''''']. When figures of merit are reported in licensing submittals to the NRC, they shall show results for [''' '' ''''''''''''''''' ''''''' '''''''' '' ''''' '''''''''''' '

'''''''''''''''' '''' ''''''''''''' ' ''''''''' ''''''''''''''''''''] (Sections 3.3.4.2.1, 3.5.4, and 3.6.2.6)

15. ['''' '''''' ''''''' ''''''''''''''''''''' ''''''''''''''''''''' '' ''' ''' ''''''' ' '''''''''''''' ' '''''''''''''

''''''''''''''' '''''''' ''''''''''''''''''''''''''' ''''' ''' '' '']. (Section 3.3.4.2.1)

16. Plant-specific licensing applications referencing the AURORA-B LOCA evaluation model shall justify that the input conditions assumed in the analysis are bounding across the entire approved operating domain, which may include, for example, extended power uprates, extended flow windows, equipment out of service (e.g., automatic depressurization system valves, feedwater heaters, single-loop operation), and feedwater temperature reduction. If necessary, analysis of multiple initial operating states shall be performed to ensure that the most limiting conditions with respect to the acceptance criteria of 10 CFR 50.46 have been calculated. (Section 3.3.5.4)

17. To assure satisfaction of GDC 35 (or similar plant-specific design criterion), [ '

'''''''''''''' ''''''''' ''' '''''''''''' '' ''''' ''''''''''''''''' '''''''']. (Section 3.3.5.4)

18. Safety analyses performed with the AURORA-B LOCA evaluation model shall include justification for any credit taken for the drywell high pressure trip signal. (Section 3.3.5.5)

19. Safety analyses for mixed-core configurations shall appropriately justify application of the AURORA-B LOCA evaluation model and any supporting methodologies (e.g., nodal core simulator and lattice physics methods, fuel thermal-mechanical performance methods) to legacy fuel assemblies designed by other vendors. Furthermore, ['' '''''''''''''''''''''''' '''''''''

- 166 -

'''''''''' '''''''''''''' ''''''''''''''''' '''' '''''''''' ''''' ''''''''''''] (Section 3.3.5.5)

20. Simulations supporting plant safety analyses should be run to completion of quenching on all potentially limiting fuel rods.77 If premature termination occurs, ['''''''''''''' ''''''''''''''''''''''

''''''''''''' '''''''''''''''''' '''''''''''''''''''''''''' ' ' '''''' '''''''''''''''''''''''] (Sections 3.3.5.7 and 3.6.2.7)

21. As discussed in Section 3.3.5.7, Framatome used a non-representative modeling practice of

['''''''''''' '''''''''''''''''''''''' '''''''''''''''' '''''''''''''''' '''''''''' ''''''''''''''' ''''''' ''''''''''' ''''''''''''''''''''''''''

'''''' '''''''''''' '''''''''''''''''''''' ''''']. Prior to implementing this practice in future plant safety analyses, the practice must be adequately defined in the AURORA-B LOCA modeling guidelines. Furthermore, this practice may not be implemented in the safety analysis for any given plant without explicit plant-specific approval by the NRC staff (e.g., in conjunction with a license amendment request to implement the AURORA-B LOCA evaluation model).

Licensees requesting credit for this non-representative modeling practice must adequately describe the extent of its intended use and justify its conservatism. The justification must address the potential for [''''''''''''''''''''''' '''''''''' ''''''' ''''''''''''''''''''''''' ''''''''''''''''''''''' ''''''''''''''''''

'''''''''''''''' ''''''''''' ' '''''''''''''''''''''''''] excessive sensitivity to timestep and nodalization variations, as discussed further in Section 3.5. (Sections 3.3.5.7 and 3.5)

22. The NRC staff has not specifically reviewed any plant parameters in ANP-10332P or deemed them acceptable for use in plant safety analyses. Therefore, each licensee using the AURORA-B LOCA evaluation model is responsible for confirming that all plant-specific design parameters are consistent with the assumptions made in the analysis. This includes, for example, [''''''''''''''''''''''' ''''' '' ''''''' ''''''''''''''' '''''''''' ''''''''''''''''' '' '' ''''''''''''' '

'''''''''''' '''''''''''''' ''' ''''''''''''''''''''''''''''' ' '' '''''''''' '''''''' ''''''''''] (Section 3.6)

23. Safety analyses performed with the AURORA-B LOCA evaluation model shall include justification that ['' '''''''''''''''''' ''''''''''''' ''''''' ''''''''' ''''''''''''''''''''''' ''''''''''' ' '''

' ' '''''' '''''''''''''''''' '''''''' '''''''''''''' '''''''' ''' ''''''''''''''''' '''''''''' ' '''''''''' ''''']

(Sections 3.4.3.5 and 3.6.2.1)

24. ['''''''''''''''''''''''''' ''''' ''''' '''''' '''''''''''''' '''' ''' '''' ''''' '''''''''''''''' '' ''''''''''''''''''''' ''''''''

77 ['' ''''''''''' '''''''''''' '''''''''''' '' ''' '''''''''''''''''' '''''''''''' ''''' ''''''''''''''' '' ''' '''' '''''''''''''''''''' ''''''

''' ''' '''''''''''''''''' ''''''' '''' ]

- 167 -

'''''''' ''''''''''''''''' ' '''''''''''''''''''''''''' '''''''''''''''''''] (Sections 3.3.4.2.1 and 3.6.2.7)

25. Plant-specific licensing applications referencing the AURORA-B LOCA evaluation model shall justify the acceptability of the following evaluation model changes Framatome implemented during the NRC staffs review of ANP-10332P:

• [''''''''''''''''' ''''''' ''''''''''''' '''''''''' '''''''' ' ''''' ''''']

• ['''''''''''''''''''' '''''''''''''''' '''''''' '''''''''''' ''''' '''''''''''' '''' '''''' ''''''''''''''''' ''''''''''''

'' ''''''''''''''']

• ['' ''''''''' ''''''''''''' ''''''''' '''''']

• ['''''''''''''''''''''' ''''''''''''''''''' ' ''''''''''''''' '''''' '''''''']

(Section 3.6.3)

26. Plant-specific licensing applications referencing the AURORA-B LOCA evaluation model shall confirm that ['' '''''' '''''''''''''''' '''''''''''''' '''''''''''''' ' '' '''''''' '''''''''' ''''''''''''''' ''

''''''''''''''' ''''''''''''''''' '''''''''''''''''78 ''''''' '''''''' ''''''''''''''' ''''''''''''' ''' ''''''''''''''''''' '''''''''''''' ''''

' ''''''''''' ' '' ''''''''' ' ''''''''''''''''''''''' '''''''''''''''''''' ''''''' ''''''' ''''' ''''''''''''''''] (Section 4.1)

27. As discussed in Section 4.3 of this SE, new or modified Framatome [''''' ''''''''''''' '''''' '

'' ''''''''' ''''''''''''''''''''] (Section 4.3) 6.0 COMPLIANCE

SUMMARY

The foregoing evaluation has assessed the AURORA-B LOCA evaluation model, as described in ANP-10332P, that Framatome has proposed for evaluating the loss-of-coolant accident for BWRs with internal jet pumps (i.e., BWR/3-6 designs). Based upon the NRC staffs review of this TR (Reference 1), Framatomes RAI responses (Reference 5), and other supporting materials, the NRC staff proposed in Section 5.0 a series of limitations and conditions that are deemed necessary, in general, to ensure that safety analyses performed with the AURORA-B LOCA evaluation model will comply with applicable regulatory requirements. This section of the NRC staffs SE provides a summary assessment of the consistency of the AURORA-B LOCA evaluation model, as modified by limitations and conditions in Section 5.0, with applicable regulatory guidance and requirements.

78 Note that the draft version of the modeling guidelines document audited by the NRC staff during this review is listed as Reference 26 to this SE.

- 168 -

6.1 Conformance with Relevant Regulatory Guidance As noted previously in Section 2.2 of this SE, the primary sources of regulatory guidance applicable to the AURORA-B LOCA evaluation model include Chapter 15 of the SRP and RG 1.203.

With regard to Chapter 15 of the SRP, this SE has documented the basis for conformance in all major review areas including

• documentation (Section 4.1)

• evaluation model (Section 3.3)

• identification of accident sequence and phenomenon ranking (Section 3.2)

• assessment and validation (Section 3.4)

• addressing uncertainty via conformance to Appendix K to 10 CFR 50 (Section 6.2.1)

• quality assurance plan (Section 4.2)

With regard to RG 1.203, Framatome used the EMDAP structure to organize ANP-10332P, and further made a reasonable effort to show compliance against the EMDAP criteria. While the NRC staff found ANP-10332P generally responsive to the EMDAP criteria, a reduction in implementational rigor was evident in certain areas, particularly the identification and ranking of relevant phenomena and the assessment and validation of the evaluation model. However, as noted above, RG 1.203 allows for a reduction in the rigor of EMDAP implementation where justified by sufficient evaluation model conservatism. As noted above, the NRC staff found sufficient evidence of conservatism in the Appendix K-based AURORA-B LOCA evaluation model to justify such a reduction in rigor. Thus, RG 1.203 may be considered satisfied.

6.2 Compliance with Applicable Regulatory Requirements As discussed above in Section 2.1, LOCA evaluation models may demonstrate compliance with 10 CFR 50.46 by either (1) realistically analyzing the LOCA event and explicitly accounting for uncertainty or (2) conservatively analyzing the LOCA event in accordance with the prescriptive requirements of Appendix K to 10 CFR 50. The AURORA-B LOCA evaluation model follows the second option.

6.2.1 Conformance to Appendix K to 10 CFR 50 As a practical convenience, the foregoing SE has presented its assessment of the AURORA-B LOCA evaluation model according to a format developed for best-estimate models that explicitly account for uncertainty. In particular, this organizational structure was selected for compatibility with ANP-10332P (the structure of which mirrors the EMDAP described in RG 1.203), as well as the review guidance in Chapter 15.0.2 of the SRP.

While the modern perspective of the EMDAP has contributed additional depth and rigor to the review process, ultimately, the NRC staff still deemed it necessary to document specifically the basis for its conclusion that the requirements of Appendix K have in fact been satisfied by the AURORA-B LOCA evaluation model. Appendix K to 10 CFR 50 specifies both (1) required and acceptable features for evaluation models used to analyze the LOCA event and (2) documentation requirements. The NRC staffs rationale for concluding that the AURORA-B LOCA evaluation model described in ANP-10332P, as modified by this SE, conforms to these Appendix K requirements is summarized below in Table 24.

- 169 -

Table 24: Conformance of AURORA-B LOCA Evaluation Model with Appendix K to 10 CFR 50 Paraphrased Appendix K Requirement Framatomes Approach NRC Staff Evaluation I. Required and Acceptable Features of the Evaluation Models I.A Sources of Heat During the LOCA Assume continuous operation with a reactor power at least 1.02 Assume infinite operation at 102 percent of licensed Acceptable, since Framatome times licensed limit unless a lower alternative value is power unless a lower alternative has been previously will assume infinite operation demonstrated to be acceptable. deemed acceptable by NRC staff. Power level at a power level that is greater uncertainties larger than 2 percent are considered than or equal to the conservative relative to Appendix K requirements, requirement.

and hence acceptable.

Assume maximum peaking factors allowed by technical [''' '''''''''''' ''''''''''''''''''' ' ''' ''''''''''''''''''' ['''''''''''''''''''''' ''''''

specifications. '''''''''''''''' ''''''''''' ''''''''' '''' '''''''''''''''''''' ''''' '''''''''''''''''''''''' ''' ''''''''''''''

''''''''''''''' '''''''' ''''' ''''''''''' ''''''''' ''''''''''' ''''' ''''''''']

''''''''''' ''''''''' ''''''''''''''''''' ''''''''''''''''' '''''']

Select combination of power shape and peaking factor from [''''''''''''''' ''' ''''''''''' ''''''''''''' '''''''''''''''''''''' '''' Acceptable, as modified by across core lifetime that results in the most severe consequences ''''''''''''''''''' ' '''' ''''''' ''' ''''''''''' ''''' '''''''''''''' Limitation and Condition 22.

for the spectrum of postulated breaks and single failures. '''''''''' ''' ''''''' '''''''''''] See Section 3.3.4.1.1 above.

Also, Limitation and Condition 16 relevant to assuring a conservative axial power distribution.

I.A.1 Initial Stored Energy in the Fuel Steady-state temperature distribution and stored energy shall be Stored energy is maximized [' ''''''''''''''''' ''' '' Per Section 3.3.4.1.1 above, calculated for the burnup that yields the highest cladding '''''''''''''' ''''' ''''' ''''''''''''' ''''''''' '''''''''''''' '''''''' general analytical approach is temperature (or stored energy). ''''''''''''''''''' ''''''''''''''''''''''' '''''''''''''''''' ''''''''' ''''''''''''' acceptable.

' ''''''''''''''''''' '''''''' '' ''''''''''''''''''' '''''''''''' '''''''''''''''''''']

- 170 -

Paraphrased Appendix K Requirement Framatomes Approach NRC Staff Evaluation

['''''''''''''''' '''''' ''' '''''''''''''''' '' ''''''' ' ''''''''' ''''''''''

'''''''''''''''''''']

Evaluate thermal conductivity of fuel as a function of burnup and The RODEX4 code will be used to determine fuel Acceptable, as evaluated temperature, taking into consideration differences in initial density. initial conditions. RODEX4 contains the required above in Sections 3.3.2 and Evaluate gap conductance as a function of burnup, taking into modeling capabilities. 3.3.4.1.1. The RODEX4 code consideration fuel densification and expansion, composition and has been previously approved pressure of gases, the initial cold gap dimension with tolerances, by NRC staff for use in and cladding creep. accordance with BAW-10247P. Note Limitation and Condition 2.

I.A.2 Fission Heat Fission heat shall be calculated using reactivity and reactor Use S-RELAP5 point kinetics model, which Acceptable, as discussed in kinetics. Framatome states in Section 6.2.3 of ANP-10332P is Section 3.3.4.1.1.

identical to that in RELAP5.

Shutdown reactivities resulting from temperatures and voids shall Conservative neutronic parameters from an approved Acceptable. See Sections 1.1 be given their minimum plausible values, including allowance for lattice physics code will be used. and 3.3.1.2.4, and Limitations uncertainties. and Conditions 1 and 22.

Control rod insertion may be assumed if calculated to occur. A bounding scram reactivity curve will be used based Acceptable, as discussed in on values that underestimate the true rod worth. Section 3.3.5.5.

I.A.3 Decay of Actinides Heat from radioactive decay of actinides, including isotopes of An actinide model consistent with 1979 version of the Acceptable, as discussed in neptunium, plutonium and uranium, shall be calculated in American Nuclear Society Standard 5.1 is used. [' Sections 3.3.4.1.1 and accordance with fuel cycle calculations and known radioactive ''''''''''''''''' ''''''''''''''''''''''''''''''''' ''''' '''''''' '''' ' 3.4.1.2, since the 1979 properties. The actinide decay heat chosen shall be that '''''''''''''''' '' '''''''''''''''''''''''''''' '''''' ''' ''''''']. The American Nuclear Society appropriate for the time in the fuel cycle that yields the highest default values from the 1979 American Nuclear Standard 5.1 is being followed calculated fuel temperature during the LOCA. Society Standard are used for the energy yield and [''' '''''''''''''''''''''''''''

decay constant for uranium-239 and neptunium-239. '''''''''''''''''''''''''''''''' ''''''''' ''' '

''''''''''''' '' '''''''''''''''''''''']

- 171 -

Paraphrased Appendix K Requirement Framatomes Approach NRC Staff Evaluation I.A.4 Fission Pro duct Decay Heat generation from fission product decay shall be assumed to Fission product decay heat is determined according Acceptable, since be 1.2 times the proposed 1971 American Nuclear Society to the proposed 1971 American Nuclear Society requirement is being followed.

Standard 5.1 value for an infinite operating period. Standard 5.1 for an infinite period of operation using required 1.2 multiplier.

The fraction of locally generated gamma energy deposited in the Gamma energy deposition is determined using a Method generally acceptable, fuel (including the cladding) may be different from 1.0 if justified lattice physics code licensed for reload analysis. as discussed in Section by a suitable calculation. Based on such a calculation for ATRIUM 10 and 11 3.3.4.1.1. However, see fuel designs, [''' ''''''''''''''' '' '''''''''' '''' ' '''''''''''''''''' Limitation and Condition 10.

''''''''''' ''''''''''''' '''' '''''' ''' ''''''''''''' '''''''']

I.A.5 Metal-Water Reaction Rate The rates of energy release, hydrogen generation, and cladding The Baker-Just equation is used for energy release, Acceptable, since required oxidation from the metal-water reaction shall be calculated using hydrogen generation, and cladding oxidation from the model is used.

the Baker-Just equation metal-water reaction.

The reaction shall be assumed not to be steam limited. Reaction is conservatively assumed not to be steam Acceptable, since required limited. assumption is used.

For rods calculated to rupture, the inside of the cladding shall be When cladding rupture is predicted, ['' ''''''''''''''''''' Acceptable, since required assumed to react after the rupture according to the Baker-Just ''' '''''''''''''''''''''' '''''''''''''' '' '' '''''''' '''''''''''' ''' models and assumptions are equation and shall be assumed not to be steam limited. The '''''''''''''''' ' '''''''''''''''' '''''''''''''' ' ''''''''''''''''' ''''''' used.

affected area shall be assumed to extend around the cladding '''' ''''''''''' '''' ''' '''''''''' '''''''' ''''''' '''

inner circumference and axially no less than 1.5 inches each way '''''''' '' ' '''''''''''' '''''''''''' '''' ''' ''''''''''''' ' '''

from the location of the rupture. ''''''''''' ''' '''''''''''''' ''''''''''''''' ''' ''''''''' ''']

The Baker-Just equation is used, and the reaction is assumed not to be steam-limited.

I.A.6 Reactor Internals Heat Transfer Heat transfer from piping, vessel walls, and non-fuel internal Heat transfer from significant structures is accounted Acceptable; evaluation model hardware shall be taken into account. for, including the reactor vessel, recirculation loop includes required capability piping, lower plenum internals, steam dryers, core and will account for significant shroud, core shroud dome and steam separators. sources of metal mass.

I.A.7 Pressurized Water Reactor Primary-to-Secondary Heat Transfer Not applicable to BWRs.

- 172 -

Paraphrased Appendix K Requirement Framatomes Approach NRC Staff Evaluation I.B Swelling and Rupture of the Cladding and Fuel Rod Thermal Parameters Include time-dependent modeling of cladding swelling and rupture Cladding swelling and rupture is performed using a Generally, acceptable, as according to the cladding axial temperature distribution and the model derived from NUREG-0630 (Reference 29) discussed further in pressure difference across the cladding. Swelling and rupture that was incorporated into prior approved evaluation Section 3.3.4.1.2. However, calculations shall be based on applicable data such that the models (e.g., XN-NF-82-07(P)(A), Revision 1 note Limitation and degree of swelling and incidence of rupture are not (Reference 30), and EMF-2361(P)(A) Condition 11.

underestimated. (Reference 16)).

The degree of swelling and rupture shall be accounted for in The RODEX4 kernel determines thermal and elastic Acceptable, as discussed in calculations of gap conductance, cladding oxidation and cladding strains. S-RELAP5 determines plastic Section 3.3.4.1.2.

embrittlement, and hydrogen generation. strains, which are added to the gap widths computed by the RODEX4 kernel. The RODEX4 kernel calculates transient gap conductance based on the dynamic state of the fuel during LOCA event.

['''''''''''''''''''''''' ''''''''''''''' '''''''''''''''''' '''''''''''' '

''''''''''''''''''].

Calculations of fuel and cladding temperatures as a function of ['''''' ''''''''''''''''''''''''' '''' '''''' ''''''''''''' ''''''''''''''''''''''' Acceptable, since evaluation time shall use values for gap conductance and other thermal ''' ''''''''''''''''''' '''''''''''' ''''''''''''''''''''''' '''' '''''' model includes required parameters as functions of temperature and other applicable ''''''''''''''''''' ''''''''''''''''''''''''''''' '''''''''''''''' '''''''' ''''''''''''''''''' modeling capability. The time-dependent variables. The gap conductance shall be varied '''' ''' '''''''''''''''' ''''''''' '''''''''''''''' '''' ''''''''''''''''' RODEX4 kernel is based on in accordance with changes in gap dimensions and any other ''''' ''''''' '''''''''''''''''''' ''''''''''''''''' ''' '''''''''''''''''''''' ''] approved stand-alone applicable variables. methodology.

I.C Blowdown Phenomena I.C.1.a Break Characteristics and Flow A spectrum of possible pipe breaks shall be considered, including The methodology will be used for the entire spectrum Acceptable, as described in instantaneous double-ended breaks ranging in cross-sectional of postulated breaks and will follow this requirement. Section 3.3.5.3.

area up to the largest pipe in the primary coolant system. The analysis shall also include the effects of longitudinal splits, with the split area equal to the cross-sectional area of the pipe.

I.C.1.b Discharge Model Once the discharging fluid reaches two-phase conditions, the Moody model is used to calculate two-phase critical Acceptable, as described in discharge rate shall be calculated with the Moody critical flow flow. [''''''''''''''''' ''''''''''''''''''''''' ''''' ''' ''' '''' ''' Section 3.3.4.1.3.

model. At least three values of discharge coefficient shall be ''' ' ''''''].

used, spanning the range from 0.6 to 1.0. If the results indicate

- 173 -

Paraphrased Appendix K Requirement Framatomes Approach NRC Staff Evaluation that even lower values of discharge coefficient are limiting, the range of discharge coefficient shall be extended until the maximum cladding temperature is achieved.

I.C.1.c End of Blowdown Not applicable to BWRs.

I.C.1.d Noding Near the Break and the ECCS Injection Points Noding in the vicinity of and including the broken or split sections A sensitivity study considering nodalization near Generally acceptable, as of pipe and the points of ECCS injection shall be chosen to permit break and injection points is discussed in response to discussed in Sections 3.3.4.3 a reliable analysis of the thermodynamic history in these regions RAI 41. [''''''''' '''''''''''''''' ' '''''''''''''' ''''''''''''''' ' and 3.5.6.

during blowdown. '''''''''''''''''''''''' '''''''''''' ''' '' ' ''''''''''' '''''''''''''''''''''

'''''''''''''''' '''''''' '''''''''''''']

I.C.2 Frictional Pressure Drops Calculate frictional losses in pipes and other components Similar correlations to those listed in Appendix K are Acceptable. The correlations including the reactor core with models that include realistic used. The S-RELAP5 friction model is essentially the used are well established, variation of friction factor with Reynolds number, and realistic two- same as used in RELAP/MOD2, ['''''''''' '''' '' they have been previously phase friction multipliers that have been adequately verified by ''''''''''''''''''''''''''''''' ''' '''''''''''''''''''' '''''''''''''''' ' reviewed by the NRC staff in comparison with experimental data, or models that prove at least ''''''''''''''' ' ' ''''''''''' '''''''''''' '''''''''''''''''''' ' '''''''' past reviews involving equally conservative with respect to maximum clad temperature ''''' ''''''''' ''''''''''''''''' ''''''''''''' ''' ''''''' ''''''''' '' S-RELAP5, and pressure calculated during the hypothetical accident. A modified Baroczy ''''''''''''''''''''' '''''''''''' '''' ''''''''' '''''''''''''''' '''''' ''''' drop comparisons for the correlation or combination of the Thom correlation for pressures '' ''''''''''''''''''''''''' ''' '''' ''''''''''' '''' ''''''' '''' assessments presented in above or equal to 250 psia and the Martinelli-Nelson correlation ''''''''' '''''''''''''''''''''' '''''''''''''''''''''''' ''''''''' '''''''''''''''''' Section 3.4 of this SE for pressures lower than 250 psia is acceptable for calculating '''''' ''' ''''''''''''' '''''''''''''''''' ''' '''''' '''''''' '''' generally appear reasonable.

two-phase friction multipliers. ''''''''''''''''''''''' '''''''''''''''''' '''''' ''''' ''''''''''' ''''''''' Note Limitation and

''''''''' '''''''''' '''''' '''''''''' '''''''''''''''''' ''''''''''''''''' Condition 22 ['''''''''''''''

'''''''''''''''''''''''' '''''''' '''''''''''''' '] ''''''''''''''''''''''''' '''''''''''''''' '''

''''''''''''' '''''''''''' ''''''''''''''''''].

I.C.3 Momentum Equation Conservation of momentum equation shall take into account S-RELAP5 includes all required effects in a manner Acceptable, since (1) temporal change of momentum, (2) momentum convection, similar to current state-of-the-art codes. requirement is satisfied.

(3) area change momentum flux, (4) momentum change due to compressibility, (5) pressure loss resulting from wall friction, (6) pressure loss resulting from area change, and (7) gravitational

- 174 -

Paraphrased Appendix K Requirement Framatomes Approach NRC Staff Evaluation acceleration. Omissions shall be justified by comparative analyses or experimental data.

I.C.4 Critical Heat Flux Correlations developed from appropriate steady-state and Framatome discussed in response to RAI 64 how Acceptable, as reviewed transient-state experimental data are acceptable for use in checks are implemented to ensure that the range of above in Section 4.2.

predicting the critical heat flux (CHF) during LOCA transients. parameters is maintained within specified limits.

The computer programs in which these correlations are used shall Framatome expects the range of the Groeneveld contain suitable checks to assure that the physical parameters CHF lookup table to encompass the LOCA event.

are within the range of parameters specified for use of the correlations by their respective authors.

Steady-state CHF correlations acceptable for use in LOCA ['''''''''''''''''''' ''''''''' ' ''''''''''''''''''' ''' Acceptable, as reviewed transients include, but are not limited to, the following: W-3, '''''''''''''''''''''''''''''' '''''''''''''''''''''''' ''''''''''''''''''' ' above in Section 3.3.1.2.2.

B&W-2, Hench-Levy, Macbeth, Barnett, Hughes. ''''''''''''''''' '''' '''''' '''''''''''' ''' ] However, note Limitation and Condition 7.

Correlations of transient CHF data may be accepted if A transient CHF correlation is not used. Use of a transient CHF comparisons between data and the correlations demonstrate that correlation is optional, not correlations predict values of CHF which allow for uncertainty in required. Framatome chose the experimental data throughout the applicable range of to use an acceptable parameters. Where appropriate, the comparisons shall use steady-state CHF correlation statistical uncertainty analysis to demonstrate conservatism of the instead.

transient correlation. Transient CHF correlations acceptable for use in LOCA transients include, but are not limited to, the GE transient CHF correlation.

After CHF is first predicted at an axial fuel rod location during Logic has been implemented in S-RELAP5 to ensure Acceptable, as reviewed blowdown, the calculation shall not use nucleate boiling heat that, [''''''''''''''' ''' '''''''' ''''''''''''''''' ''''''''' ''''''''''''''' above in Section 3.3.4.1.3.

transfer correlations at that location subsequently during the '' ''''''' '''''''' ''''' '''' '''''''''''''' '''''''''''' However, note Limitations blowdown even if the calculated local fluid and surface conditions ''' '''''''''''''''''''''''''''' '''' ''''''' '''''''''' ''' '''''''''''''''''' and Conditions 12 and 13.

would apparently justify the reestablishment of nucleate boiling. ''' ''''''''''''''''''' ''''''''' ' ''' ''''''''''' ''''''''

Heat transfer assumptions characteristic of return to nucleate ].

boiling (rewetting) shall be permitted when justified by the calculated local fluid and surface conditions during the reflood portion of a LOCA.

- 175 -

Paraphrased Appendix K Requirement Framatomes Approach NRC Staff Evaluation I.C.5 Post-CHF Heat Transfer Correlations Correlations of heat transfer from the fuel cladding in the post- The McDonough, Milich, King correlation is used for Acceptable. McDonough, CHF regimes of transition and film boiling shall be compared to transition boiling. Milich, King correlation is applicable steady-state and transient-state data using statistical approved in Appendix K.

correlation and uncertainty analyses. Such comparison shall [''''' '''''''''''''''' '''' '''''''''''''''' '''''''''''''''''''''' ''' '''''' ' Framatome validated other demonstrate that the correlations predict conservative values of ''' ''' ''''''''' ''''''''''''' '' '''' '''''''] correlations including heat transfer coefficient relative to data throughout the intended [''''''''''''' '''' ''''''''''''''']

range of application. The comparisons shall quantify the relation Logic has been implemented in S-RELAP5 to ensure against test data, as of the correlations to the statistical uncertainty of applicable data. that, [''''''''''''''' '' '''''''' ''''''''''''''''''' ''''''''''''''''''''''' described above in Acceptable correlations for post-CHF boiling regimes include the ''''''''''''''''' '''''''''''''''''' ''''' ' '''''''' '''''' ' ''' Section 3.4.

Groeneveld flow film boiling correlation and the Westinghouse '''''' ''''''''''''''''''' ''''''''' ''''''' '''''''''''''' '''''' '''''

correlation of steady-state transition boiling. The transition boiling ''' ''''''''''''''''''''''''' '''''''''' '''' '''''''''''''''''''''''' ''' Logic to lock out a return to correlation of McDonough, Milich, and King is suitable for use ''''''''''''''''' ''''' '''''''' '''''''''' '' '''''''''''''''''''''' ''' transition boiling under the between nucleate and film boiling. All these correlations are ''''''''''''''''''''' ''''''''''' ''' ] conditions specified in restricted as follows: Appendix K has been

• Groeneveld correlation shall not be used near its low- implemented.

pressure singularity

• The nucleate boiling term of the Westinghouse correlation and the entire McDonough, Milich, King correlation shall not be used during the blowdown phase after the temperature difference between cladding and saturated fluid exceeds 300 °F.

• Transition boiling heat transfer shall not be reapplied for the remainder of the LOCA blowdown, even if the cladding superheat returns below 300 °F, except for the reflood portion of the LOCA when justified by the calculated local fluid and surface conditions.

Evaluation models approved after October 17, 1988, shall not use The Dougall-Rohsenow correlation is not used. Acceptable.

the Dougall-Rohsenow flow film boiling correlation under conditions where nonconservative predictions of heat transfer result.

- 176 -

Paraphrased Appendix K Requirement Framatomes Approach NRC Staff Evaluation I.C.6 Pump Modeling Characteristics of rotating primary system pumps shall be derived Model is the same as that from RELAP5/MOD3, ['''' Acceptable. As noted above from a dynamic model that includes momentum transfer between '' ''''''''''''''''' ''' ''''''''''''''''''' ''''''''' ''''''''''''''''''''''''''' in Section 3.3.1.2.3, the fluid and the rotating member, with variable pump speed as a '''''''''''''''''''''' '''''' '' '''''''' '' ''''''''' '''''''''''''''''' centrifugal pump model used function of time. The pump model resistance used for analysis '' ''''''''''''' ''''''''' ''''''''''''''''' '''''''''''''''' ''''' '''''''' for AURORA-B LOCA should be justified. The pump model for the two-phase region ''''''''' '''''''''''''''' ''''''''''''''''''''''''''' '''''''''''''''''''''''''''' evaluation model is similar to shall be verified by applicable two-phase pump performance data. '''''''''''''''''''''' ''''''''''' ''' ''' '''''' ''''''''''''''''''''''''' ''' those previously reviewed for For BWRs, after saturation is calculated at the pump suction, ''''''' '''''''' '''' ''''''''' ''''''''' '' '''''''''''''''''' other S-RELAP5-based pump head may be assumed to vary linearly with quality, going to ''''''''''''''''' '''' ''''' ''''''''' ''''''''''] methods.

zero for one percent quality at the pump suction, so long as the analysis shows that core flow stops before the quality at pump suction reaches one percent.

I.C.7 Core Flow Distribution During Blowdown Not applicable to BWRs.

I.D Post-Blowdown Phenomena; Heat Removal by the ECCS I.D.1 Single Failure Criterion Analyze possible failure modes of ECCS equipment and effects Single failure scenarios applicable to each plant are Acceptable, since calculated on ECCS performance. Analysis must assume occurrence of the evaluated on an individual basis to determine the results will encompass limiting most limiting single failure. limiting case. single failure for each plant.

I.D.2 Containment Pressure The containment pressure for evaluating cooling effectiveness Assume containment is at atmospheric pressure. Acceptable; conservatively during reflood and spray cooling shall not exceed a conservatively satisfies requirement and calculated minimum value. The calculation shall include the consistent with past Appendix effects of operation of all installed pressure-reducing systems and K evaluation models.

processes.

I.D.3 Calculation of Reflood Rate for Pressurized Water Reactors Not applicable to BWRs.

I.D.4 Steam Interaction with Emergency Core Cooling Water in Pressurized Water Reactors Not applicable to BWRs.

I.D.5 Refill and Reflood Heat Transfer for Pressurized Water Reactors Not applicable to BWRs.

- 177 -

Paraphrased Appendix K Requirement Framatomes Approach NRC Staff Evaluation I.D.6 Convective Heat Transfer Coefficients for Boiling Water Reactor Fuel Rods Under Spray Cooling Following the blowdown period, convective heat transfer shall be Convective heat transfer is calculated according to Acceptable, as discussed calculated using coefficients based on appropriate experimental heat transfer correlations that have been validated above in Sections 3.3.1.2.2, data. according to test data. [' ''''''''''''''''''''''''' ''''''''''''''''''' ' 3.3.4.1.1, 3.4, et al.

'''''''' '''''''''''''''' ' ''''''''''''''''' '' ''''''''''' ''' '''''''''''''']

For reactors with jet pumps and having fuel rods in a 7x7 fuel The convective heat transfer coefficients provided in Acceptable; use of assembly array, certain convective heat transfer coefficients Appendix K are not used. Appendix K values (that provided in Appendix K are stated to be acceptable. strictly apply to antiquated 7x7 fuel) is not required.

I.D.7 The Boiling Water Reactor Channel Box Under Spray Cooling Following the blowdown period, heat transfer from, and wetting of, Heat transfer from channel box is calculated Method described above in the channel box shall be based on appropriate experimental data. according to heat transfer correlations that have been Section 3.3.1.2.2 is validated according to test data. acceptable per assessments reviewed in Section 3.4.

For reactors with jet pumps and fuel rods in a 7x7 fuel assembly The convective heat transfer coefficients provided in Acceptable; use of array, Appendix K identifies specific heat transfer coefficients and Appendix K are not used. Appendix K values (that a wetting time correlation as acceptable. strictly apply to antiquated 7x7 fuel) is not required.

II. Required and Acceptable Features of the Evaluation Models A description shall be furnished that is sufficiently complete to The required descriptions are generally provided in Acceptable. Summary level permit technical review of the analytical approach, including the ANP-10332P. Supplementary information is provided description is provided in equations used, their approximations in difference form, the in other sources made available for NRC staff review, ANP-10332P, as assumptions made, and the values of all parameters or the particularly the S-RELAP code theory manual supplemented by code theory procedure for their selection, as for example, in accordance with a (Reference 11). manual, modeling guidelines, specified physical law or empirical correlation. and other references. Note Limitations and Conditions 25 and 26.

A complete listing of each computer program, in the same form as Framatome stated the code would be furnished upon Code documentation and used in the evaluation model, must be furnished to the Nuclear request. demonstration case files were Regulatory Commission upon request. audited in lieu of requesting code listing.

- 178 -

Paraphrased Appendix K Requirement Framatomes Approach NRC Staff Evaluation Solution convergence shall be demonstrated by studies of system Required sensitivity studies were performed in Acceptable; as discussed in modeling or noding and calculational time steps Sections 7.9.3 and 7.9.4 of ANP-10332P. These Sections 3.5 and 3.3.4.1.3, studies were supplemented in response to RAIs 112 many sensitivity studies in Appropriate sensitivity studies shall be performed to evaluate the and 126. ANP-10332P were not effect on the calculated results of variations in noding, accepted by NRC staff.

phenomena assumed in the calculation to predominate, including Revised sensitivity studies in pump operation or locking, and values of parameters over their responses to RAIs 112 and applicable ranges. For items to which results are shown to be 126 addressed staffs sensitive, the choices made shall be justified. concerns. Note Limitations and Conditions 17, 21, 23.

To the extent practicable, predictions of the evaluation model, or Comparisons of the AURORA-B LOCA evaluation Acceptable; assessment portions thereof, shall be compared with applicable experimental model against experimental data are provided in comparisons are evaluated in information. ANP-10332P in Sections 5.4 and 7. Section 3.4, above.

General Standards for AcceptabilityElements of evaluation Demonstrate compliance with the required features of This table summarizes the models reviewed will include technical adequacy of the Section I of Appendix K. NRC staffs assessment that calculational methods, including: For models covered by the AURORA-B LOCA 10 CFR 50.46(a)(1)(ii), compliance with required features of evaluation model conforms to section I of this Appendix K; and, for models covered by requirements of Appendix K.

10 CFR 50.46(a)(1)(i), assurance of a high level of probability that the performance criteria of 10 CFR 50.46(b) would not be exceeded.

- 179 -

6.2.2 Compliance with Relevant Criteria from 10 CFR 50.46 The AURORA-B LOCA evaluation model is intended to provide a methodology for BWR/3-6 licensees to perform safety analyses capable of demonstrating compliance with certain criteria specified in 10 CFR 50.46, namely

• the peak cladding temperature does not exceed 2200 °F,

• the calculated local cladding oxidation does not exceed 17 percent of the cladding thickness prior to oxidation,

• the total amount of hydrogen generated from the metal-water reaction is limited to 1 percent of the hypothesized theoretical maximum, and

• the reactor core is maintained in a coolable geometry.79 As an alternative to best-estimate models that explicitly account for uncertainty, 10 CFR 50.46 allows the development of LOCA evaluation models in conformance with the required and acceptable features of Appendix K to 10 CFR 50. The NRC staff has concluded that the AURORA-B LOCA evaluation model, as modified by this SE, conforms to the required and acceptable features of Appendix K. As discussed above in Section 3.3.5, the AURORA-B LOCA evaluation model would consider a sufficient number of postulated LOCA scenarios with break sizes, locations, and characteristics to ensure that the limiting condition is calculated.

Therefore, it may be used to perform safety analyses for BWR/3-6 licensees for the purpose of demonstrating compliance with 10 CFR 50.46.

6.2.3 Compliance with General Design Criterion 35 Inasmuch as 10 CFR 50.46 and Appendix K specify detailed requirements for demonstrating adequate ECCS performance for each boiling or pressurized light-water reactor fueled with uranium oxide pellets within cylindrical zircaloy or ZIRLO cladding, the NRC staffs conclusions above that analyses performed using the AURORA-B LOCA evaluation model would satisfy these requirements leads directly to an identical conclusion regarding GDC 35.

However, it is worth reemphasizing that GDC 35 specifies that adequate ECCS performance shall be assured both with and without offsite power available. As noted above in Section 3.3.5.4, Limitation and Condition 17 [''''''''''''' ''''''''''''''''''''''''' '''''''''''''''' ''''''''''''''''''''''''

'''''''''''''''' ' '' ''''''''''''' ''''''''''' ''' '''''''''''' ''''''' ' '' '''''''''' ''''''''].

6.2.4 Proposed Rule 10 CFR 50.46c Sections 2.5.1 and 4.4 of ANP-10332P discuss how the AURORA-B LOCA evaluation model might be applied under the hypothetical scenario that the Commission approves proposed rule 10 CFR 50.46c without substantial modification. However, as documented above in Limitation and Condition 5, at the present time, the NRC staff is unable to conclude that the AURORA-B 79 Note that, although BWR/3-6 licensees must also demonstrate that sufficient long-term core cooling is provided to maintain an acceptable core temperature, as discussed above in Section 2.1, this demonstration is beyond the scope of the AURORA-B LOCA evaluation model.

- 180 -

LOCA evaluation model may be applied directly under proposed rule 10 CFR 50.46c. As described above in Section 2.1, the NRC staffs review of ANP-10332P was performed based solely upon existing regulations and guidance; therefore, it passes no judgment on the acceptability of the AURORA-B LOCA evaluation model under a hypothetical implementation of 10 CFR 50.46c. Should the Commission approve the 10 CFR 50.46c rulemaking, the NRC staff anticipates that further guidance will be provided regarding the demonstration of compliance against its requirements for approved LOCA evaluation models.

7.0 CONCLUSION

Based upon the foregoing evaluation, the NRC staff has concluded that, contingent upon the satisfaction of the limitations and conditions defined in Section 5.0 of this SE, the AURORA-B LOCA evaluation model described in ANP-10332P provides an acceptable methodology for performing safety analyses for BWR/3-6 plants for the purpose of demonstrating compliance with the criteria in 10 CFR 50.46 (b)(1)-(4) in accordance with the required and acceptable features prescribed in Appendix K to 10 CFR 50.

8.0 REFERENCES

80

1. ANP-10332P, Revision 0, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Loss of Coolant Accident Scenarios, AREVA Inc.,

February 2014 (ADAMS Package Accession No. ML14091A220).

2. Audit Report for the May 16 to May 18, 2017, Audit in Support of the Review of ANP-10332P, Revision 0, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Loss of Coolant Accident Scenarios, NRC, December 5, 2017 (ADAMS Package Accession No. ML17262B138 (Non-Public)/ML17262B196 (Public)).

3. Request for Additional Information Regarding AREVA Inc., Topical Report ANP-10332P, Revision 0, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Loss of Coolant Accident Scenarios, NRC, October 24, 2017 (ADAMS Package Accession No. ML17257A302 (Non-Public)/ML17257A309 (Public)).

80 Note that certain references may be proprietary and, hence, unavailable to the general public. These references may be denoted by a P in the suffix of the report number or otherwise labeled as proprietary in the list below; conversely, a report number suffix including NP is typically indicative of a redacted, non-proprietary version of a proprietary document. Where applicable, accession numbers for the NRCs Agencywide Document Access and Management System (ADAMS) are provided for the readers convenience.

- 181 -

4. Request for Additional Information Regarding AREVA Inc. Topical Report ANP-10332P, Revision 0, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Loss of Coolant Accident Scenarios, NRC, January 29, 2018 (ADAMS Package Accession No. ML17360A202 (Non-Public)/ML17360A207 (Public)).

5. Response to Request for Additional Information Regarding Topical Report ANP-10332P, Revision 0, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Loss of Coolant Accident Scenarios, Framatome Inc.,

April 30, 2018 (ADAMS Package Accession No. ML18122A227).

6. NUREG-0800, Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants: LWR Edition, NRC, June 1987 (ADAMS Package Accession No. ML052340514).

7. ANP-10300P, Revision 0, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Transient and Accident Scenarios, AREVA Inc.,

December 2009 (ADAMS Package Accession No. ML100040163 (Non-Public)/ML100040157 and ML100040158 (Public)).

8. Final Safety Evaluation for AREVA Inc. Topical Report ANP-10300P, Revision 0, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Transient and Accident Scenarios, NRC, January 5, 2018 (ADAMS Package Accession Nos. ML17346B039 (Non-Public)/ML18124A214 (Public)).

9. ANP-10333P, Revision 0, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Control Rod Drop Accident (CRDA), AREVA Inc.,

March 2014 (ADAMS Accession No. ML14098A332 (Public)/ML14098A333 (Non-Public))

10. Final Safety Evaluation for Topical Report ANP-10333P, Revision 0, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Control Rod Drop Accident (CRDA), NRC, March 15, 2018 (ADAMS Accession No. ML18038B277 (Non-Public)/ML18128A002 (Public)).

11. FS1-0009406, Revision 7.0, FSQA-08-S-RELAP5 Models and Correlations Code Manual (Theory), AREVA Inc., August 2, 2016.

12. FS1-0009130, Revision 6.0, FSQA-07-S-RELAP5 Input Data Requirements (Users Manual), AREVA Inc., March 31, 2017.

13. NUREG-1230, Compendium of ECCS Research for Realistic LOCA Analysis, Final Report, NRC, December 1988 (ADAMS Accession Nos. ML053490333 (Part 1),

ML053620415 (Part 2)).

14. Opinion of the Commission, In the Matter of Rulemaking Hearing on Acceptance Criteria for Emergency Core Cooling Systems for Light-Water-Cooled Nuclear Reactors, Atomic Energy Commission, December 28, 1973.

- 182 -

15. NEDO-20566A, General Electric Company Analytical Model for Loss-of-Coolant Analysis in Accordance with 10 CFR 50, Appendix K, General Electric Company September 1986.

16. EMF-2361(P)(A), Revision 0, EXEM BWR-2000 ECCS Evaluation Model, Framatome ANP, May 2001 (ADAMS Accession No. ML012050409).

17. Groeneveld, D. C., et al., The 2006 CHF Look-Up Table, Nuclear Engineering and Design, 237, Pages 1909-1922, 2007.

18. Regulatory Guide 1.157, Best-Estimate Calculations of Emergency Core Cooling System Performance, NRC, May 1989.

19. Regulatory Guide 1.203, Transient and Accident Analysis Methods, NRC, December 2005.

20. XN-NF-82-49(P)(A), Revision 1, Exxon Nuclear Company Evaluation Model EXEM PWR Small Break Model, Exxon Nuclear Company Inc., June 1986.

21. Request for Review of EMF-2310(P), Revision 0, SRP Chapter 15 Non-LOCA Methodology for Pressurized Water Reactors, Siemens Power Corporation, November 1999 (ADAMS Package Accession No. ML993470050).

22. Request for Review of EMF-2328(P), Revision 0, PWR Small Break LOCA Evaluation Model, S-RELAP5 Based, Siemens Power Corporation, January 10, 2000 (ADAMS Package Accession No. ML003675792).

23. Request for Review of EMF-2103(P), Revision 0, Realistic Large Break LOCA Methodology for Pressurized Water Reactors, Framatome ANP, August 2001 (ADAMS Package Accession No. ML012400043).

24. EMF-2100(P), Revision 4, S-RELAP5 Models and Correlations Code Manual, Framatome ANP, May 2001 (ADAMS Accession No. ML012880298 (Non-Public)).

25. EMF-2103(P)(A) Revision 0, Realistic Large Break LOCA Methodology for Pressurized Water Reactors, Framatome ANP, April 2003 (ADAMS Package Accession No. ML032691410).

26. FS1-0034030, Revision 1.0, Guidelines for Input Development and Problem Execution for AURORA-B LOCA, AREVA Inc., Undated Draft Version, Audited October 2017.

27. BAW-10247PA, Revision 0, Realistic Thermal-Mechanical Fuel Rod Methodology for Boiling Water Reactors, AREVA Inc., April 2008 (ADAMS Package Accession No. ML081340220 (Non-Public)/ ML081340208 (Public)).

28. Staff Assessment of AREVA Codes and Methods with Regards to Thermal Conductivity Degradation - Experimental Data, NRC, March 23, 2012.

- 183 -

29. NUREG-0630, Cladding Swelling and Rupture Models for LOCA Analysis, NRC, April 1980 (ADAMS Accession No. ML053490337).

30. XN-NF-82-07(P)(A), Revision 1, Exxon Nuclear Company ECCS Cladding Swelling and Rupture Model, Exxon Nuclear Company, November 1982 (ADAMS Accession No. ML081710143 (Non-Public)).

31. NUREG-2160, Post-Test Examination Results from Integral, High-Burnup, Fueled LOCA Tests at Studsvik Nuclear Laboratory, NRC, August 2013 (ADAMS Accession No. ML13240A256).

32. Review of the Siemens Power Corporation S-RELAP5 Code to Appendix K Small--Break Loss-of-Coolant Accident Analysis, Advisory Committee on Reactor Safeguards, February 13, 2001 (ADAMS Accession No. ML010460306).

33. Framatome ANP S-RELAP5 Realistic Large-Break Loss-of-Coolant Accident Code, Advisory Committee on Reactor Safeguards, December 20, 2002 (ADAMS Accession No. ML023570289).

34. Acceptance for Referencing of Licensing Topical Report EMF-2328(P), Revision 0, PWR Small Break LOCA Evaluation Model, S-RELAP5 Based, NRC, March 15, 2001 (ADAMS Accession No. ML010800365).

35. Technical Evaluation of AREVA Topical Report ANP-10332P, Revision 0, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Loss of Coolant Accident Scenarios, Brookhaven National Laboratory, October 15, 2018.

36. NUREG/CR-2566, BWR Refill-Reflood Program Task 4.4 - CCFL/Refill System Effects Tests (30° Sector) Evaluation of Parallel Channel Phenomena, NRC, November 1982 (ADAMS Accession No. ML092940124 (Non-Public)).

37. Acceptance for Referencing of Licensing Topical Report EMF-2310(P), Revision 0, SRP Chapter 15 Non-LOCA Methodology for Pressurized Water Reactors, NRC, May 11, 2001 (ADAMS Accession No. ML011310533).

38. Safety Evaluation on Framatome ANP Topical Report EMF-2103(P), Revision 0, Realistic Large Break Loss-of-Coolant Accident Methodology for Pressurized Water Reactors, NRC, April 9, 2003 (ADAMS Accession No. ML030760312).

39. Peterson, L.J., and S.M. Bajorek, Experimental Investigation of Minimum Film Boiling Temperature for Vertical Cylinders at Elevated Pressure, ICONE10-22520, April14-18, 2002.

40. EMF-2158(P)(A), Revision 0, Siemens Power Corporation Methodology for Boiling Water Reactors: Evaluation and Validation of CASMO-4/MICROBURN-B2, Siemens, October 1999 (ADAMS Package Accession No. ML003698556).

- 184 -

41. XN-NF-80-19(P)(A), Volume 3, Revision 2, Exxon Nuclear Methodology for Boiling Water Reactors, THERMEX: Thermal Limits Methodology Summary Description, Exxon Nuclear Company, January 1987 (ADAMS Accession No. ML081340305).

42. NUREG/CR-2230, BWR Small Break Simulation Tests With and Without Degraded ECC Systems: BWR Blowdown/Emergency Core Cooling Program, NRC, January 1982 (ADAMS Accession No. ML083440268 (Non-Public)).

43. NUREG/CR-4128, BWR Full Integral Simulation Test (FIST) Phase II Test Results and TRAC-BWR Model Qualification, NRC, October 1985 (ADAMS Accession No. ML092940131 (Non-Public)).

44. TRACE V5.840 Theory Manual, Volume 1: Field Equations, Solution Methods, and Physical Models, NRC, March 6, 2013.

45. NUREG/CR-2568, BWR Refill-Reflood Program Task 4.4 - CCFL/Refill System Effects Tests (30° Sector) SSTF System Response Test Results, NRC, April 1983 (ADAMS Accession No. ML092940130 (Non-Public)).

46. Groeneveld, D. C., et al., Lookup Tables for Predicting CHF and Film Boiling Heat Transfer: Past, Present and Future, Nuclear Technology, 152, Pages87-104, 2005.

47. Lee, M., A Critical Heat Flux Approach for Square Rod Bundles Using the 1995 Groeneveld CHF Table and Bundle Data of Heat Transfer Research Facility, Nuclear Engineering and Design, 197, Pages 357-374, 2000.

48. RELAP5/MOD3.3 Code Manual, Volume IV: Models and Correlations, Information Systems Laboratories, Inc., December 2001.

49. INEEL-EXT-98-00834, Revision 2.3, RELAP5-3D Code Manual, Volume IV: Models and Correlations, Idaho National Laboratory, April 2005.

50. FS1-0036275, Revision 1.0, COA Test Suite Cases for RIA Response: USEP12 vs UMAR16, Framatome Inc., March 21, 2018.

51. Sun, K. H., Flooding Correlations for BWR Bundle Upper Tieplates and Bottom Side-Entry Orifices, as taken from Multiphase Transport, Fundamentals, Reactor Safety, Applications, Volume 3, edited by T. Nejat Vezirolu, 1980.

52. NUREG/CR-2435, Dispersed Flow Film Boiling in Rod Bundle Geometry - Steady State Heat Transfer Data and Correlation Comparisons, Oak Ridge National Laboratory, March 1982 (ADAMS Accession No. ML063190260).

53. TRACE V5.0 Assessment Manual, Appendix B: Separate Effects Tests, NRC, Undated.

54. EMF-2103P-A, Revision 3, Realistic Large Break LOCA Methodology for Pressurized Water Reactors, AREVA Inc., June 2016 (ADAMS Package Accession No. ML16286A579).

- 185 -

55. NEDO-10183, Core Standby Cooling Systems for the General Electric 1969 BWR Standard Plants, General Electric, May 1970 (ADAMS Accession No. ML070670272 (Non-Public)).

56. NUREG-2121, Fuel Fragmentation, Relocation, and Dispersal During the Loss-of-Coolant Accident, NRC, March 2012 (ADAMS Accession No. ML12090A018).

57. NUREG/CR-2786, BWR Refill-Reflood Program Task 4.4 - CCFL/Refill System Effects Tests (30° Sector) Evaluation of ECCS Mixing Phenomena, NRC, May 1983 (ADAMS Accession No. ML070610551 (Non-Public)).

58. TRACE V5.0 Assessment Manual, Appendix C: Integral Effects Tests, NRC, Undated.

59. Incropera, F.P. and D.P. DeWitt, Fundamentals of Heat and Mass Transfer, Fifth Edition, 2002.

60. Browns Ferry Nuclear Plant, Unit 1 - Issuance of Amendments Regarding the Transition to AREVA Fuel (TAC NO. ME3775) (TS-473), NRC, April 27, 2012 (ADAMS Accession No. ML12086A285 (Non-Public)).

61. SECY-16-0033, Draft Final Rule - Performance-Based Emergency Core Cooling System Requirements and Related Fuel Cladding Acceptance Criteria (RIN 3150-AH42), March 16, 2016, (ADAMS Accession No. ML15238A947).

62. NRC Regulatory Issue Summary 2016-04, Clarification of 10 CFR 50.46 Reporting Requirements and Recent Issues with Related Guidance Not Approved for Use, April 19, 2016.

63. Hejzlar, P. and N. E. Todreas, Consideration of Critical Heat Flux Margin Prediction by Subcooled or Low Quality Critical Heat Flux Correlations, Nuclear Engineering and Design 163 (1996) 215-223.

64. Todreas, N. E. and P. Hejzlar, Letter to the Editor on the Paper by D. C. Groeneveld and L. K. H. Leung, The 1995 Look-Up Table for Critical Heat flux in Tubes, Nuclear Engineering and Design 163 (1996) 25-26.

65. Draft Safety Evaluation by the Office of Nuclear Reactor Regulation, Topical Report ANP-10332P, Revision 0, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Loss of Coolant Accident Scenarios, NRC, August 31, 2018.

66. Response to NRC Request for Additional Information, AURORA-B: An Evaluation Model for Boiling Water Reactors; Application to Loss of Coolant Accident Scenarios, ANP-10332Q1P, Revision 1, October 31, 2018.

67. Groenveld, D. C., Reply to the Letter to the Editor from Todreas and Hejzlar on the Paper Entitled The 1995 CHF Look-Up Table, Nuclear Engineering and Design 163 (1996) 27-28.

- 186 -

68. Nukiyama S., The Maximum and Minimum Values of the Heat Q Transmitted from Metal to Boiling Water under Atmospheric Pressure, International Journal of Heat and Mass Transfer, Volume 9, Issue 12, Pages 1419-1433, December 1966.

Principal Contributors: J. Lehning, NRC L.-Y. Cheng, Brookhaven National Laboratory U. S. Rohatgi, Brookhaven National Laboratory Dated: March 26, 2019