ML18214A480

Final SER (Non-Prop) - TR-0116-21012, Revision 1, NuScale Power Critical Heat Flux Correlations.

ML18214A480
Person / Time
Site:	PROJ0769
Issue date:	08/08/2018
From:	Renee Taylor NRC/NRO/DLSE/LB1
To:
BAVOL B/415-6715
Shared Package
ML18214A478	List: ML18214A479 ML18214A480
References
Download: ML18214A480 (44)

Text

U.S. NUCLEAR REGULATORY COMMISSION SAFETY EVALUATION BY THE OFFICE OF NEW REACTORS TOPICAL REPORT - 0116-21012, REVISION 1 NUSCALE POWER CRITICAL HEAT FLUX CORRELATIONS NUSCALE POWER, LLC PROJECT NO. PROJ0769 Enclosure 1

TABLE OF CONTENTS 1.0 Introduction ....................................................................................................................... 1 2.0 Regulatory Evaluation ....................................................................................................... 2 3.0 Technical Evaluation ......................................................................................................... 3 3.1 Experiential Data ............................................................................................................... 6 3.1.1 Credible Test Facility ......................................................................................................... 6 3.1.1.1 Test Facility Description .............................................................................................. 6 3.1.1.2 Test Facility Comparison ............................................................................................. 6 3.1.2 Accurate Data ................................................................................................................... 7 3.1.2.1 Test Procedures .......................................................................................................... 7 3.1.2.2 Statistical Design of Experiment.................................................................................. 8 3.1.2.3 Accurate Method ......................................................................................................... 8 3.1.2.4 Instrumentation Uncertainties...................................................................................... 9 3.1.2.5 Repeated Test Points .................................................................................................. 9 3.1.2.6 Quantified Heat Losses ............................................................................................. 10 3.1.3 Reproduced Local Conditions ......................................................................................... 11 3.1.3.1 Equivalent Geometries .............................................................................................. 11 3.1.3.2 Equivalent Grid Spacers............................................................................................ 12 3.1.3.3 Axial Power Shapes .................................................................................................. 12 3.1.3.4 Radial Power Shapes ................................................................................................ 13 3.1.3.5 Differences between Test and Reactor ..................................................................... 13 3.2 Model Generation ............................................................................................................ 14 3.2.1 Appropriate Mathematical Model .................................................................................... 14 3.2.1.1 Necessary Parameters .............................................................................................. 14 3.2.1.2 Model Form ............................................................................................................... 14 3.2.2 Model Coefficient Generation .......................................................................................... 16 3.2.2.1 Training Data ............................................................................................................. 16 3.2.2.2 Coefficient Generation............................................................................................... 16 3.3 Model Validation .............................................................................................................. 17 3.3.1 Validation Error ............................................................................................................... 17 3.3.2 Data Distribution .............................................................................................................. 18 3.3.2.1 Validation Data .......................................................................................................... 18 3.3.2.2 Application Domain ................................................................................................... 18 ii

3.3.2.3 Expected Domain ...................................................................................................... 19 3.3.2.4 Data Density .............................................................................................................. 19 3.3.2.5 Sparse Regions ......................................................................................................... 20 3.3.2.6 Restricted Domain ..................................................................................................... 20 3.3.3 Consistent Model Error ................................................................................................... 20 3.3.3.1 Poolability .................................................................................................................. 20 3.3.3.2 Nonconservative Subregions .................................................................................... 21 3.3.3.3 Model Trends ............................................................................................................ 22 3.3.4 Quantified Model Error .................................................................................................... 23 3.3.4.1 Error Data Base......................................................................................................... 23 3.3.4.2 Statistical Method ...................................................................................................... 23 3.3.4.3 Appropriate Bias for Model Uncertainty..................................................................... 24 3.3.5 Model Implementation ..................................................................................................... 24 3.3.5.1 Same Computer Code............................................................................................... 24 3.3.5.2 Same Methodology ................................................................................................... 25 3.3.5.3 Transient Behavior .................................................................................................... 25 4.0 Limitations ....................................................................................................................... 25 5.0 Conclusions ..................................................................................................................... 25 6.0 References ...................................................................................................................... 26 A. Critical boiling Transition Model Assessment Framework ............................................... 29 A.1 Introduction ..................................................................................................................... 29 A.2 Existing CHF Correlations ............................................................................................... 29 A.3 CBT Assessment Framework ......................................................................................... 31 A.3.1 G1 Experimental Data ..................................................................................................... 31 A.3.1.1 G1.1 Credible Test Facility ........................................................................................ 32 A.3.1.2 G1.2 Accurate Measurements................................................................................... 32 A.3.1.3 G1.3 Reproduction of Local Conditions..................................................................... 33 A.3.2 G2 Model Generation ...................................................................................................... 34 A.3.2.1 G2.1 The Mathematical Form.................................................................................... 35 A.3.2.2 G2.2 Method for Determining Coefficients ................................................................ 35 A.3.3 G3 Validation through Error Quantification ..................................................................... 36 A.3.3.1 G3.1 Calculating Validation Error .............................................................................. 37 A.3.3.2 G3.2 Data Distribution in the Application Domain ..................................................... 37 A.3.3.3 G3.3 Inconsistency in the Validation Error ................................................................ 38 A.3.3.4 G3.4 Calculating Model Uncertainty .......................................................................... 39 iii

A.3.3.5 G3.5 Model Implementation ...................................................................................... 39 iv

1.0 INTRODUCTION

By letter dated November 30, 2017 (Agencywide Documents Access and Management System (ADAMS) Accession No. ML17335A089), NuScale Power, LLC (NuScale), submitted Topical Report (TR)-0116-21012, Revision 1, Critical Heat Flux Correlations, to the U.S. Nuclear Regulatory Commission (NRC) staff for review. This revision replaced the submittal dated October 5, 2016 (ADAMS Accession No. ML16279A363), NuScale Power Critical Heat Flux Correlation NSP2. The revised submittal implements an additional NSP4 critical heat flux (CHF) correlation and incorporates changes associated with NRC requests for additional information (RAIs). The purpose of TR-0116-21012 is to provide the bases for NRC approval to use the NSP2 and NSP4 CHF correlations in VIPRE-01, within their range of applicability, along with their associated correlation limits for the NuScale design certification application and safety analysis of the NuScale Power Module (NPM) with NuFuel-HTP2TM fuel.

Table 1, List of Key Correspondence, contains the key correspondence between the NRC and NuScale. This includes preapplication correspondence, RAIs, responses to RAIs, audit documentation, and other correspondence relevant to this review.

Table 1. List of Key Correspondence Sender Document Document Date Reference Critical Heat Flux Test NuScale January 24, 2014 1 Program Technical Report NuScale Critical Heat Flux Correlation Topical Report NRC December 10, 2015 2 Preapplication Engagement with NRC Audit Plan for NuScale NRC Critical Heat Flux Testing at May 2, 2016 3 KATHY Audit Report for NuScale NRC Critical Heat Flux Testing at August 12, 2016 4 KATHY NuScale Topical Report October 5, 2016 5 Request for Supplemental NRC December 5, 2016 6 Information NuScale Supplemental Information December 29, 2016 7 NRC Acceptance Letter February 2, 2017 8 Request for Additional NRC May 8, 2017 9 Information (eRAI 8795)

Audit Plan for Regulatory Audit of NuScale Topical NRC May 30, 2017 10 Report TR-0116-21012, Revision 0 Response to Request for NuScale Additional Information July 7, 2017 11 (eRAI 8795) 1

Sender Document Document Date Reference Request for Additional NRC August 21, 2017 12 Information (eRAI 8931)

Response to Request for NuScale Additional Information September 25, 2017 13 (eRAI 8931)

Audit Summary for NuScale NRC November 27, 2017 14 CHF NuScale Topical Report Revision November 30, 2017 15

2.0 REGULATORY EVALUATION

Title 10 of the Code of Federal Regulations (10 CFR), Sections 52.47, Contents of Applications; Technical Information, and 10 CFR 52.79, Contents of Applications; Technical Information in Final Safety Analysis Report, require a final safety analysis report (FSAR) to describe and analyze the design and performance of the structures, systems, and components.

Safety evaluations (SEs) performed to support the FSAR include accident analyses to (1) demonstrate that specified acceptable fuel design limits (SAFDLs) are not exceeded during normal operation, including the effects of anticipated operational occurrences (AOOs), and (2) determine the number of fuel failures associated with CHF that need to be included in the radiological consequences for postulated accidents. An approved CHF correlation is used in establishing an SAFDL for use in such analyses. Thus, an approved CHF correlation is used to establish a partial basis for demonstrating compliance with the following applicable regulations from 10 CFR Part 50, Domestic Licensing of Production and Utilization Facilities, which refer to the general design criteria (GDC) of Appendix A, General Design Criteria for Nuclear Power Plants.

• GDC 10, Reactor Design, which requires that the reactor core and associated coolant, control, and protection systems be designed with appropriate margin to assure that SAFDLs are not exceeded during any condition of normal operation, including the effects of AOOs.

• GDC 12, Suppression of reactor power oscillations, which requires that the reactor core and associated coolant, control, and protection systems be designed to assure that power oscillation which can result in conditions exceeding SAFDLs are not possible or can be reliably detected and suppressed.

• 10 CFR 52.47(a)(2)(iv) and GDC 19, Control Room, as they relate to the evaluation and analysis of the radiological consequences of postulated accidents.

NUREG-0800, Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants, Section 4.4, Thermal and Hydraulic Design, describes an acceptable approach for a CHF correlation to meet GDC 10 and GDC 12 as establishing a 95-percent probability at the 95-percent confidence level that the hot rod in the core does not experience a boiling transition condition during normal operation and AOOs.

The scope of the NRC staffs review addresses the applicability of the NSP2 and NSP4 CHF correlations, and their associated CHF ratio (CHFR) limits, for use in performing safety analyses of the NPM with NuFuel-HTP2TM fuel.

2

3.0 TECHNICAL EVALUATION

The purpose of TR-0116-21012 is to provide the bases for NRC approval to use the NSP2 and NSP4 CHF correlations in VIPRE-01, within their range of applicability, along with their associated correlation limits for the NuScale design certification application and safety analysis of the NPM with NuFuel-HTP2TM fuel. The NRC staff used the critical boiling transition model assessment framework summarized in Table 2, List of All Goals, which considers the framework presented in Appendix A of this SE. For brevity, Table 2 only contains the goals (G).

The sections below provide the evidence.

Table 2. List of All Goals The critical boiling transition model can be trusted in reactor GOAL safety analyses.

The experimental data supporting the critical boiling G1 transition model are appropriate.

The experimental data have been collected at a credible test G1.1 facility.

G1.1.1 The test facility is well understood.

The test facility has been verified by comparison to an G1.1.2 outside source.

G1.2 The experimental data have been accurately measured.

The test facility has an appropriate quality assurance G1.2.1 program.

The experiment has been appropriately statistically designed (i.e., the value of a system parameter from any test was G1.2.2 completely independent from its value in the test before and after the test).

The method used to obtain critical boiling transition data G1.2.3 results in an accurate measurement.

The instrumentation uncertainties have been demonstrated G1.2.4 to have a minimal impact on the measured critical heat flux or critical power.

The uncertainty in the critical heat flux or critical power is G1.2.5 quantified through repeated tests at the same state points.

The heat losses from the test section are quantified, G1.2.6 appropriately low, and duly accounted for in the measured data.

The test bundle reproduced the local conditions in the G1.3 reactor fuel bundle.

The test bundle used in the experiment should have G1.3.1 geometric dimensions equivalent to those of the fuel bundle used in the reactor for all major components.

The grid spacers used in the test bundle should be G1.3.2 prototypical of the grid spacers used in the reactor assembly.

3

The critical boiling transition model can be trusted in reactor GOAL safety analyses.

The axial power shapes in the test bundle should reflect the G1.3.3 expected or limiting axial power shapes in the reactor bundle.

The radial power peaking in the test bundle should reflect the G1.3.4 expected or limiting radial powers in the reactor bundle.

Any differences between the test bundle and the reactor bundle should have a minimal impact on the flow field. This G1.3.5 includes components that are not in the reactor bundle but that are needed for testing purposes.

G2 The model was generated in a logical fashion.

G2.1 The mathematical form of the model is appropriate.

The mathematical form of the model contains all the G2.1.1 necessary parameters.

The reasoning for choosing the mathematical form of the G2.1.2 model should be discussed and should be logical.

The process for determining the models coefficients was G2.2 appropriate.

The training data (i.e., the data used to generate the G2.2.1 coefficients of the model) should be identified.

The method for calculating the models coefficients should be G2.2.2 described.

The method for calculating the R- or K-factor and the additive constants (for both full-length and part-length rods) should be described. Further, a description of how such values are G2.2.3 calculated if dryout is not measured on the rod under consideration should be provided (boiling-water reactors only).

The model has sufficient validation as demonstrated through G3 appropriate quantification of its error.

G3.1 The correct validation error has been calculated.

The validation error is appropriately distributed throughout G3.2 the application domain.

The validation data (i.e., the data used to quantify the G3.2.1 models error) should be identified.

The application domain of the model should be G3.2.2 mathematically defined.

G3.2.3 The expected domain of the model should be understood.

There should be adequate validation error data density G3.2.4 throughout the expected and application domains.

Sparse regions (i.e., regions of low data density) in the G3.2.5 expected and application domains should be identified and justified to be appropriate.

4

The critical boiling transition model can be trusted in reactor GOAL safety analyses.

G3.2.6 The model should be restricted to its application domain.

Any inconsistencies in the validation error have been G3.3 accounted for appropriately.

The validation error should be investigated to ensure that it G3.3.1 does not contain any subgroups that are obviously not from the same population (i.e., non-poolable).

The expected domain should be investigated to determine if G3.3.2 it contains any non-conservative subregions which would impact the predictive capability of the model.

The model is trending as expected in each of the various G3.3.3 model parameters.

The models uncertainty has been appropriately calculated G3.4 from the validation error.

The validation error statistics should be calculated from an G3.4.1 appropriate database.

The validation error statistics should be calculated using an G3.4.2 appropriate method.

G3.4.3 The models uncertainty should be appropriately biased.

G3.5 The model has been correctly implemented.

The model has been implemented in the same computer G3.5.1 code that was used to generate the validation error.

The models prediction of the critical boiling transition (CBT) is being applied using the same evaluation methodology as it G3.5.2 was when predicting the validation data set for determining the validation error.

The model results in an accurate or conservative prediction G3.5.3 when it is used to predict transient behavior.

In addition to the NSP2 and NSP4 CHF correlations, TR-0116-21012 discusses a third CHF correlation, NSP1. Section 5.3 of the TR explains that the NSP2 CHF correlation is simply the NSP1 CHF correlation with an additional NSPX factor. Therefore, anything that affects the NSP1 CHF correlation also affects the NSP2 CHF correlation.

5

3.1 Experiential Data 3.1.1 Credible Test Facility 3.1.1.1 Test Facility Description Test Facility Description The test facility is well understood.

G1.1.1, Review Framework for Critical Boiling Transition Models Section 3.1 of the TR describes the facilities used to obtain the data in the development and validation of the NSP2 and NSP4 CHF correlations, which includes Stern Laboratories in Hamilton, Ontario, Canada, and the KArlstein Thermal HYdraulic test loop (KATHY) in Karlstein, Germany. In addition to using these facility descriptions, the NRC staff conducted an inspection at Stern Laboratories (References 16-19) and an audit at the KATHY test loop (Reference 4).

Based on the descriptions provided by NuScale and the information obtained by the staff during the inspection and audit activities, the NRC staff finds that the test facilities NuScale used to develop the NSP2 and NSP4 CHF correlations are well understood.

3.1.1.2 Test Facility Comparison Test Facility Comparison The test facility has been verified by comparison to an outside source.

G1.1.2, Review Framework for Critical Boiling Transition Models Benchmarking tests were conducted at the CHF testing facilities of Stern Laboratories, the KATHY test loop, and the Heat Transfer Research Facility at Columbia University (Reference 20). The results of these benchmarking activities demonstrated that the variation of test data obtained from these facilities was within a reasonable range, and that a comparison of CHF data among the three facilities showed consistent results from all three facilities (Reference 20). Based on the existence of a benchmarking study involving Stern Laboratories and the KATHY test loop in the literature and the results presented in that study, the NRC staff finds that the test facilities NuScale used to develop the NSP2 and NSP4 CHF correlations have been adequately compared to outside sources.

6

3.1.2 Accurate Data 3.1.2.1 Test Procedures Facility Quality Assurance The test facility has an appropriate quality assurance program.

G1.2.1, Review Framework for Critical Boiling Transition Models Section 3.1.1.3 and Section 3.1.2.7 of the TR provide a high-level description of the data collection process at Stern Laboratories and the KATHY test loop, respectively. These sections describe the instrumentation used to measure heater power, coolant flow, pressure, and temperature. The TR describes the heater power, coolant temperature, and pressure measurements as redundant and diverse at both facilities. The TR describes the flow measurement at Stern Laboratories as redundant and diverse but did not provide similar information for flow measurement at the KATHY facility. Accordingly, the NRC staff issued RAI 8795, Question 29725 (Reference 9), asking the applicant to describe how flow measurement redundancy and diversity were treated for the KATHY test loop. The applicants response, provided in a letter dated July 7, 2017 (Reference 11), stated that the flow measurement [

]. The NRC staff finds this response acceptable because it describes [ ] instrumentation capable of providing a check against instrument malfunction. The NRC summarized the redundancy and diversity capabilities for the power, flow, pressure, and temperature measurements at Stern Laboratories and the KATHY test loop in Table 3, Instrumentation Redundancy and Diversity.

Table 3. Instrumentation Redundancy and Diversity Stern KATHY Power Current and voltage at power Two independent and diverse supply methods for measuring electric current through the

[ test bundle: a Faraday effect meter and shunts

]

Flow [ Orifice with [

]

]

Pressure Pressure taps installed [ Two pressure transducers at the outlet of the test section

]

Temperature Resistance temperature Two independent, calibrated detectors for [ resistance temperature detectors

]

7

In addition to reviewing the information in the TR, the NRC staff conducted an inspection at Stern Laboratories (Reference 17) and an audit at the KATHY test loop (Reference 4). During the inspection and audit activities, the NRC staff determined that the test procedures appropriately incorporated the testing requirements, that all of the instrumentation is routinely calibrated, and that the calibration records were up to date at the time of testing. Based on the instrumentation redundancy and diversity, the NRC staffs inspection findings, and the staffs audit observations, the staff finds that the test facilities used to collect CHF data maintained appropriate quality assurance programs.

3.1.2.2 Statistical Design of Experiment Statistical Design of Experiment The experiment has been appropriately statistically designed (i.e., the value of a system parameter from any test was completely independent from its value in the test before and after it).

G1.2.2, Review Framework for Critical Boiling Transition Models Section 3.1 of TR-0116-21012 describes the data sources used to develop the CHF correlation and provides the test matrices used for testing at Stern Laboratories and the KATHY test loop.

The NRC staff was not able to identify any discussion regarding the statistical design of the experiments. Accordingly, it issued RAI 8795, Question 29724 (Reference 9) asking the applicant to describe how the statistical design of experiments was treated during testing at Stern Laboratories and the KATHY test loop. The applicants response, provided in its letter dated July 7, 2017 (Reference 11), stated that data were taken for each combination of pressure, mass-flux, and subcooling with no particular state point emphasized. The applicant stated that taking data at each state point combination allows for a systematic evaluation of trends with regard to pressure, mass flux, and subcooling. The NRC staff finds that this method of collecting data is consistent with existing precedent (Reference 21 and Reference 22).

Based on the consistency with precedent and the repeatability tests, discussed in Section 3.1.2.5 of this SE, the NRC staff finds the experiment has been appropriately statistically designed.

3.1.2.3 Accurate Method Accurate Method The method used to obtain critical boiling transition data results in an accurate measurement.

G1.2.3, Review Framework for Critical Boiling Transition Models Section 3.1.1.3 and Section 3.1.2.7 of the TR provide a high-level description of the data collection process at Stern Laboratories and the KATHY test loop, respectively. These sections describe the method used to obtain steady-state CHF data points by first establishing the desired conditions (i.e., pressure, temperature, and flow) and then increasing heater power until a CHF event is detected. The NRC staffs audit summary (Reference 4) describes additional details about this process. The NRC staff finds this process produces an accurate 8

measurement of CHF because it results in the direct observation of critical boiling transition at a known heat flux, and is therefore acceptable.

3.1.2.4 Instrumentation Uncertainties Instrumentation Uncertainties The instrumentation uncertainties have been demonstrated to have a minimal impact on the measured heat flux or critical power.

G1.2.4, Review Framework for Critical Boiling Transition Models The TR provides the instrument uncertainties for measuring system pressure, mass flux, inlet coolant temperature, and bundle power in Table 3-3 and Table 3-12 of the TR for the tests at Stern Laboratories and the KATHY test loop, respectively. The NRC staff inquired about the impact of instrument uncertainty during the audit of the KATHY test loop (Reference 4) and was informed that sufficient test data are taken such that the standard deviation of the measured-to-predicted values for the CHF correlation encompass uncertainties associated with instrumentation measurement, test repeatability, and test reproducibility. Section 3.2.2 of Revision 1 of the TR incorporates this information. The NRC staff performed calculations to further investigate the impact of instrument uncertainty on the NSP2 and NSP4 CHF correlations. These calculations determined sensitivity derivatives of the CHF correlations with respect to pressure, temperature, mass flux, and power, and then used these sensitivity derivatives and the instrument uncertainties provided in the TR to estimate the resulting uncertainty in the calculated CHFR. The results of the calculations showed that the CHFR limits account for much more uncertainty (approximately a factor of 5) than what can be attributed to instrument error. Therefore, the NRC staff concluded that instrument uncertainty is not the major contributor to the uncertainty in the NSP2 and NSP4 CHF correlations.

3.1.2.5 Repeated Test Points Repeated Test Points The uncertainty in the critical heat flux or critical power is quantified through repeated tests at the same state points.

G1.2.5, Review Framework for Critical Boiling Transition Models Section 3.1.2.7 of the TR discusses the use of reference points at the KATHY test loop that are used to verify test repeatability. Additionally, the NRC staff conducted an audit of the KATHY test loop and noted that, Test repeatability at the KATHY test loop is verified by [

]. This is in contrast to CHF testing at Stern, where random state points within the testing domain are rerun (Reference 4). The NRC staff issued RAI 8795, Question 29726, asking NuScale to provide evidence to show that the variability in the repeatability tests was sufficiently low (Reference 9).

The applicants response, provided in its letter dated July 7, 2017 (Reference 11), stated the following:

9

• CHF testing, at both Stern Laboratories and the KATHY test loop, had established acceptance criteria on test repeatability of [

].

• [

].

• [

].

The NRC staff notes that the repeat test points that [ ] occur for state points that are outside the requested range of applicability for the CHF correlations. The NRC staff finds the applicants response acceptable because the stated acceptance criteria are sufficient to demonstrate low variance in test repeatability (i.e., uncertainty associated with test repeatability is much less than the uncertainty accounted for in the CHFR limits), and the test repeatability criteria are satisfied in the ranges of applicability for the CHF correlations. Based on the number of repeat test points collected, and the acceptance criteria used for these tests, the NRC staff finds that the applicant has appropriately quantified the uncertainty in CHF through repeated tests.

3.1.2.6 Quantified Heat Losses Quantified Heat Losses The heat losses from the test section are quantified, appropriately low, and duly accounted for in the measured data.

G1.2.6, Review Framework for Critical Boiling Transition Models Section 3.1.1.3 of the TR states that the heat losses at Stern [

] and are accounted for in the CHF correlation validation. Based on the heat losses being accounted for in the KATHY test loop data, which is used to establish the 95/95 CHFR limits for both the NSP2 and NPS4 CHF correlations, the NRC staff finds that the treatment of heat losses is acceptable.

10

3.1.3 Reproduced Local Conditions 3.1.3.1 Equivalent Geometries Equivalent Geometries The test bundle used in the experiment should have equivalent geometric dimensions to that of the fuel bundle used in the reactor for all major components.

G1.3.1, Review Framework for Critical Boiling Transition Models Sections 3.1.2.2 to 3.1.2.5 of the TR describe test sections used to obtain data from the KATHY test loop for prototypical NuFuel-HTP2TM fuel (K9000, K9100, K9200, and K9300). The TR describes the K9000, K9100, K9200, and K9300 test bundles as representing the NuFuel-HTP2TM design. NRC staff compared the geometric dimensions of the these test sections (provided in TR Table 3-6, 3-7, 3-8, and 3-10 for the K9000, K9100, K9200, and K9300, respectively) against the geometric dimensions for NuFuel HTP2TM fuel assemblies (provided in TR Table 2-1). This comparison showed that the simulated fuel used in the test sections used equivalent rod diameter, gap pitch, guide tube diameter, grid spacer heights, and grid spacer locations (grid spacer locations were within [

]). Based on this comparison between the NuFuel-HTP2TM design and the simulated fuel used to develop validation data, NRC staff finds that these test bundles have equivalent dimensions to those of the fuel bundle in the reactor for all major components.

In addition to the data collected from the K9000 - K9300 test sections, the TR describes two other categories of experimental data. Section 3.1.1.1 of the TR describes the Stern test sections used to obtain data for a preliminary fuel design (U1, U2, and C1); this data was used to develop the NSP1 CHF correlation (note that NSP1 is included in the NSP2 CHF correlation, however, separate approval is not being sought for the application of NSP1 to NuFuel-HTP2TM fuel). Section 3.1.2.1 of the TR describes a test section used to obtain legacy data from the KATHY test loop for a design that used HMPTM grids (K8500 HMPTM); this data was used to develop the NSPX factor and NSP2 CHF correlation. The geometric dimensions associated with these test sections are provided in TR Table 3-2 and TR Table 3-5 for the Stern tests and K8500 tests, respectively. NRC staff notes that [

]. However, NRC staff finds the use of this data to develop the NSP1 and NSP2 CHF correlations acceptable because this data is not used for validation (note that the K9000 - K9300 test data is used to validate the NSP2 and NSP4 CHF correlations).

11

3.1.3.2 Equivalent Grid Spacers Equivalent Grid Spacers The grid spacers used in the test bundle should be prototypical of the grid spacers used in the reactor assembly.

G1.3.2, Review Framework for Critical Boiling Transition Models Section 3.1.2 of the TR describes the grid spacers used in the testing to obtain validation data as the HMPTM and HTPTM that are prototypical of the NuFuel-HTP2TM fuel design. In addition, NRC staff performed an audit of the CHF testing at the KATHY test loop and observed that HMPTM and HTPTM mixing grids used in the testing are consistent with the prototypical design and are located at the same axial locations as the prototypical design (Reference 4). Based on the discussion in this paragraph, the NRC staff finds that the grid spacers used in the test bundles for obtaining validation data are prototypical of the grid spacers used in the NuFuel-HTP2TM fuel design.

3.1.3.3 Axial Power Shapes Axial Power Shapes The axial power shapes in the test bundle should reflect the expected or limiting axial power shapes in the reactor bundle.

G1.3.3, Review Framework for Critical Boiling Transition Models Section 3.1.1 and Section 3.1.2 of the TR describe the experiments at Stern Laboratories and the KATHY test loop, respectively. The description of the experiments at Stern Laboratories show that data were obtained using uniform and symmetric cosine axial power profiles.

Additionally, as described in Section 4.3 of the TR, the applicant uses a nonuniform flux factor to adjust CHF predictions for the effect of variations of axial power profiles and demonstrated that this factor adequately adjusts data obtained for a uniform power profile to nonuniform shapes.

Based on the testing performed with uniform and nonuniform power shapes and the development of a nonuniform flux factor, the NRC staff finds that the axial power shapes used in the test bundles at Stern Laboratories are suitably representative of the expected or limiting axial power shapes used in the reactor bundle.

The description of the experiments at the KATHY test loop show that data [

]. Additionally, as described in Section 7.5 of the TR, the applicant uses a nonuniform flux factor to adjust CHF predictions for the effect of variations of axial power profiles and demonstrated that this factor adequately adjusts [

]. Based on the testing performed with [

] and the development of a nonuniform flux factor, the NRC staff finds that the axial power shapes used in the test bundles at the KATHY test loop are suitably representative of the expected or limiting axial power shapes used in the reactor bundle.

12

3.1.3.4 Radial Power Shapes Radial Power Shapes The radial power peaking in the test bundle should reflect the expected or limiting radial power in the reactor bundle.

G1.3.4, Review Framework for Critical Boiling Transition Models The TR Figures 3-4, 3-5, and 3-11 to 3-15 show the radial power distribution used in the tests.

As described in the NRC staffs audit report of the KATHY test facility, the radial power distribution is selected to produce CHF in the interior rods to negate the impact of the wall effect. The radial power distributions in the TR show power distributions that (1) produce peak heat flux values within the interior rods and (2) produce a subchannel surrounded by peaked rods such that it is expected to result in a hot or limiting subchannel. Based on these observations, the NRC staff finds the radial power distributions used in the TR reflect a limiting radial power distribution for developing and validating the NSP2 and NSP4 CHF correlations.

3.1.3.5 Differences between Test and Reactor Differences between Test and Reactor Any differences between the test bundle and the reactor bundle should be addressed. This includes components that are not in the reactor bundle but are needed for testing purposes.

G1.3.5, Review Framework for Critical Boiling Transition Models The NRC staff found the following differences between the K9000, K9100, K9200, and K9300 test bundles, which are used to develop validation data, and the NuFuel-HTP2TM: (1) the test bundles used a 5x5 grid as compared to the 17x17 grid used in the prototype, and (2) the test bundles used intermediate support grids to counteract the electromagnetic forces on the heater rods. During an audit of the KATHY test loop, the NRC staff inquired about these differences and noted that (1) early CHF testing at the Columbia University Heat Transfer Research Facility loop demonstrated that a 5x5 grid layout is sufficiently large to negate the impact of the wall effect, and (2) based on the results of previous tests, Areva, Inc. has determined that the intermediate support grids produce a small flow resistance and have a negligible contribution to mixing (Reference 4). The NRC finds these differences acceptable because they have a negligible impact on the test results and are consistent with established practice.

13

3.2 Model Generation 3.2.1 Appropriate Mathematical Model 3.2.1.1 Necessary Parameters Necessary Parameters The mathematical form of the model contains all necessary parameters.

G2.1.1, Review Framework for Critical Boiling Transition Models Sections 4.2 and 5.2 of the TR discuss the form of the NSP2 CHF, while Section 7.4 discusses the form of the NSP4 CHF correlation. The NSP4 correlation includes pressure, local mass flux, local equilibrium quality, boiling length, [ ], and hydraulic diameter as correlation parameters. The NSP2 CHF correlation includes these same parameters plus the [

]. The NRC staff finds the parameters chosen for the NSP2 and NSP4 correlations are consistent with parameters used in previously approved CHF correlations (Reference 21 and Reference 22). Additionally, the coefficient generation process, described in Section 3.2.2 of this SE, checks that statistically significant parameters are captured by the correlation; and the correlation validation process, described in Section 3.3 of this SE, verifies adequate performance of the correlation (i.e., the final correlation includes the necessary parameters). Based on the use of correlation parameters that are consistent with previously approved CHF correlations, the coefficient generation process, and the validation process, the NRC staff finds that the NSP2 and NSP4 CHF correlations contain the necessary parameters.

3.2.1.2 Model Form Model Form The reasoning for choosing the mathematical form of the model should be discussed and be logical.

G2.1.2, Review Framework for Critical Boiling Transition Models Section 4.2 and Section 7.4 of the TR discuss the form of the base CHF correlation for NSP2 and the form of the NSP4 CHF correlation, respectively. These correlations are quadratic equations consisting of the parameters discussed in Section 3.2.1.1 of this SE. During an audit, the NRC staff reviewed engineering calculations associated with correlation development and noted that (1) NuScale considered several forms of the CHF correlation, (2) global sensitivity analyses were performed for each form of the CHF correlation to confirm consistency with known physical trends, and (3) the quadratic formulation was selected as the final form based on the results of the global sensitivity analyses (Reference 4). Additionally, the correlation validation process, described in Section 3.3 of this SE, verifies adequate performance of the correlation. Based on the description of the CHF correlation forms in Sections 4.2 and 7.4 of the TR, the information obtained by the NRC staff during an audit, and the correlation validation process, the staff finds that the mathematical form of the NSP2 CHF correlation and the NSP4 correlations are adequately described and logical.

14

The final NSP2 CHF correlation, Equation 5-2 of the TR, includes an NSPX factor. Section 5.2 of the TR describes the NSPX factor as an adjustment to the base CHF correlation for analyzing the thermal performance of NuFuel-HTP2TM fuel. The NSPX factor is applied because the base CHF correlation is developed using data from the Stern tests, which did not use prototypical NuFuel-HTP2TM geometry (see Section 3.1.3.1 of this SE). Section 5.2 of the TR explains that the NSPX factor imposes a mean predicted-to-measured ratio of [ ] when the NSP2 CHF correlation is applied to the K8500 data. Additionally, during an audit, the NRC staff observed that the K8500 data, which use a heater rod longer than the NuFuel-HTP2TM fuel, showed reduced thermal performance than the data collected for NuFuel-HTP2TM fuel (Reference 4).

Therefore, the NRC staff recognizes that the NSPX factor, which is derived from the K8500 data, introduces conservative bias into the NSP2 correlation (Reference 4). Furthermore, the NRC staff recognizes that the correlation validation process, described in Section 3.3 of this SE, uses data that are independent of the K8500 data and addresses uncertainties associated with any lack of fit that may be introduced by the NSPX factor. Based on the conservative development of the NSPX factor and the use of a validation process over the correlations range of applicability, the NRC staff finds the NSPX factor acceptable.

A nonuniform flux factor, described in Section 4.3 of the TR, is applied to adjust NSP2 CHF predictions for the effect of variations of axial power profiles. As described in Section 4.3 of the TR, the nonuniform flux factor is based on the original Tong factor (Reference 23). The parameters in the original Tong factor, which are carried over into the nonuniform flux factor for the NSP2 CHF correlation, are based on test parameters that do not bound the range of applicability for the NSP2 CHF correlation. However, as described in Section 4.3 of the TR, the applicant [

]. Additionally, the correlation validation process, described in Section 3.3 of this SE, verifies adequate performance of the correlation over its range of applicability. Appendix A of the TR contains the local conditions used to develop the CHF correlations and shows that the Tong factors are always applied as a penalty. Accordingly, the NRC staff established Limitation 1 on the application of the NSP2 and NSP4 correlations.

Based on the use of Stern test data to modify the base Tong factor, the validation process that covers the correlation range of applicability, and pursuant to Limitation 1, the NRC staff finds the applicants use of the Tong factor in the NSP2 CHF correlation acceptable.

The final NSP4 CHF correlation is provided as Equation 7-1 of the TR and is developed using data obtained from experiments that are consistent with NuFuel-HTP2TM fuel. A nonuniform flux factor, described in Section 7.5 of the TR, is applied to adjust NSP4 CHF predictions for the effect of variations of axial power profiles. Unlike the approach taken for the NSP2 nonuniform flux factor, the applicant refit the parameters used in the original Tong factor using data obtained from experiments that are consistent with NuFuel-HTP2TM fuel. Additionally, the correlation validation process, described in Section 3.3 of this SE, verifies adequate performance of the correlation over its range of applicability. Based on the use of prototypical test data to develop a nonuniform flux factor, as well as the validation process that covers the correlation range of applicability, and subject to Limitation 1, the NRC staff finds the applicants nonuniform flux factor for the NSP4 CHF correlation acceptable.

15

3.2.2 Model Coefficient Generation 3.2.2.1 Training Data Training Data The training data (i.e., the data used to generate the coefficients of the model) should be identified.

G2.2.1, Review Framework for Critical Boiling Transition Models Section 4.4 and Section 7.6 of the TR describe the statistical technique used in the CHF correlation coefficient generation process for the NSP1 (which is incorporated into NSP2) and NSP4 CHF correlations, respectively. Section 4.4 of the TR states that a five-fold cross-validation process is used in generating the NSP1 coefficients, and Section 7.6 states that a three-fold cross-validation process is used in generating the NSP4 coefficients. Section 4.1 of the TR describes the k-fold process as randomly portioning the data into k subsets, then holding one subset for validation testing and using the remaining data to train the correlation (i.e., obtain correlation coefficients). This process is repeated k times (i.e., until all of the data have been used in validation). The NRC staff finds the k-folds process for selecting training data acceptable because it is an established statistical technique for training correlations.

Additionally, Section 7.1 of the TR states that the difference between a five-fold and three-fold cross validation is small. During an audit, the NRC staff noted that NuScale performed sensitivity studies on the number of groups used in the k-folds cross-validation process and that the results of the sensitivity study showed negligible differences between three and five groups (Reference 4). Based on the information provided in Section 7.1 of the TR, and the NRC staffs audit observations, the staff finds the number of subsets used in the cross-validation process for the NSP1 and NSP4 CHF correlations acceptable.

In addition to the Stern data used in the development of the NSP1 CHF correlation and the prototypical NuFuel-HTP2TM data used in the development of the NSP4 CHF correlation, NuScale uses legacy data obtain from the KATHY test loop, K8500 data, to develop the NSPX factor. The NRC staff recognizes that the correlation validation process, described in Section 3.3 of this SE, uses data that are independent of the K8500 data. Based on the independence of the NSP2 validation data, the NRC staff finds the use of K8500 data to develop the NSPX factor acceptable.

3.2.2.2 Coefficient Generation Coefficient Generation The method for calculating the models coefficients should be described.

G2.2.2, Review Framework for Critical Boiling Transition Models Sections 4.2 and 7.4 of the TR describe the process used to develop the NSP1 and NSP4 CHF correlations. The TR describes the iterative process that removes noncorrelating parameters from the quadratic form discussed in Section 3.2.1.2 of this SE. Once the final form of the 16

correlation is obtained, NuScale performs the k-fold cross-validation process described in Section 3.2.2.1 of this SE. NuScale performs a linear least squares fit, which is an established technique for performing linear regression, to obtain correlation coefficients k times during the k-folds process. NuScale averaged the coefficients obtained from the cross-validation process to obtain the final coefficients. During an audit, the NRC staff independently replicated the NSP1 CHF correlation using the group structure and data provided by NuScale (Reference 14).

Additionally, the correlation validation process, described in Section 3.3 of this SE, verifies adequate performance of the correlation over its range of applicability. Based on the use of an established technique to fit the CHF correlations and a validation process that covers the correlation range of applicability, the NRC staff finds the coefficient generation process for the NSP1 (which is included in NSP2) and NSP4 CHF correlations acceptable.

3.3 Model Validation 3.3.1 Validation Error Validation Error The validation error has been correctly calculated.

G3.1, Review Framework for Critical Boiling Transition Models NuScale uses a database to develop and validate the NSP2 and NSP4 CHF correlations that depends upon reduced data (i.e., the raw data from CHF testing are further processed by calculation). Section 3.3.2 of the TR describes this data reduction process as using VIPRE-01 to [ ]. The TR explains that the Stern data [

], and the data collected from the KATHY test loop [

]. The validation data for the NSP2 and NSP4 CHF correlations are obtained from testing the NuFuel-HTP2TM simulated fuel.

Section 3.3.2 of the TR explains that, to obtain local conditions from the NuFuel-HTP2TM testing data for validation of the NSP2 CHF correlation, NuScale took the local conditions [

]. The NRC staff finds this approach acceptable because it results in a conservative CHFR limit.

Section 3.3.2 of the TR explains that, to obtain local conditions from the NuFuel-HTP2TM testing data to develop and validate the NSP4 CHF correlation, NuScale took the local conditions [

]. The local conditions, provided in Appendix A of the TR, show that CHF [

]. Therefore, the NRC staff finds that it is reasonable to select local conditions [ ] for developing and validating the NSP4 CHF correlation.

Based on the acceptable identification of the CHF location for obtaining validation data, the NRC staff finds that the validation error is calculated using an acceptable measured value.

17

3.3.2 Data Distribution 3.3.2.1 Validation Data Validation Data The validation data (i.e., the data used to quantify the models error) should be identified.

G3.2.1, Review Framework for Critical Boiling Transition Models Section 4.5, Section 6.3, and Section 7.7 describe the validation process used to establish the CHF correlation limits for the NSP1, NSP2, and NSP4 CHF correlations, respectively. These sections of the TR explain that [

]. Based on the use of the k-folds method and the [ ], the NRC staff finds the use of all the data to perform validation and to determine the CHFR limit acceptable.

3.3.2.2 Application Domain Application Domain The application domain of the model should be mathematically defined.

G3.2.2, Review Framework for Critical Boiling Transition Models Table 8-2 and Table 8-4 of the TR provide the application domains for the NSP2 and NSP4 CHF correlations, respectively. The NRC staff identifies the application domains provided in the TR as hyper-rectangles and finds that defining the application domain as a hyper-rectangle is consistent with current practice. NRC staff finds the definition of the application domain acceptable because it is (1) consistent with current practice, and (2) regions within the application that are not supported by data (e.g., low flow with low quality, and high flow with high quality) are not observed in testing and are not realistically expected to occur during normal operation, including AOOs.

18

3.3.2.3 Expected Domain Expected Domain The expected domain of the model should be understood.

G3.2.3, Review Framework for Critical Boiling Transition Models As described in Section 2.5 of Reference 1, the applicant performed preliminary analyses of the NPM to develop conditions covering normal operation and AOOs. The NRC staff compared the range of test conditions, identified in Table 2-5 of Reference 1, with the range of applicability of the NSP2 and NSP4 CHF correlations, provided in Table 8-2 and Table 8-4 of the TR, and determined that the range of test conditions encompasses the range of applicability for both the NSP2 and NSP4 CHF correlations. Based on the performance studies of the NPM presented in Reference 1, the NRC staff finds the expected range acceptable.

3.3.2.4 Data Density Data Density There should be an appropriate data density throughout the expected domain.

G3.2.4, Review Framework for Critical Boiling Transition Models Table 8-2 and Table 8-4 of the TR mathematically define the application domains for the NSP2 and NSP4 CHF correlations, respectively. Based on prior experience with CHF correlation reviews, the NRC staff recognizes that the defined application domain of a CHF correlation contains regions where there are no underlying experimental data and where the correlation will not be used. Accordingly, on August 21, 2017, the NRC staff issued RAI 8931, Question 30134 (Reference 12), requesting that NuScale demonstrate adequate data density throughout the expected domain of application. NuScales response, provided in a letter dated September 25, 2017 (Reference 13), included several plots that show the data collection within the expected domain. These plots show that the expected mass flux and pressure regions are densely populated with data. These plots also show that the local equilibrium quality data fall mostly outside the expected domain. The NRC staff finds these results acceptable because CHF data are collected by driving test rods into CHF, which results in much higher thermodynamic qualities than are expected to occur. Based on the information provided by NuScale in response to RAI 8931, Question 30134, the NRC finds that there is appropriate data density throughout the expected domain.

19

3.3.2.5 Sparse Regions Sparse Regions Sparse regions (i.e., regions of low data density) in the expected domain should be identified and justified to be appropriate.

G3.2.5, Review Framework for Critical Boiling Transition Models As described in Section 3.3.2.4 of this SE, the expected mass flux and pressure regions are densely populated with data, but the local equilibrium quality observed in the data mostly falls outside the expected domain. The NRC staff recognizes that this occurs because CHF data are collected by driving test rods into CHF, which results in much higher equilibrium qualities than are expected to occur in the actual plant (i.e., CHF testing forces CHF to occur, but CHF is not expected to actually occur in the plant during normal operation and AOOs). However, NRC staff recognizes that if reactor power was increased substantially, due to an unexpected and unforeseen event, such that thermal margin to CHF was challenged, the increased energy added to the coolant would result in equilibrium qualities in the region where CHF data was collected. Accordingly, the NRC staff finds that the sparse regions in the expected domain are identified and appropriately justified.

3.3.2.6 Restricted Domain Restricted Domain The model should be restricted to its application domain.

G3.2.6, Review Framework for Critical Boiling Transition Models Table 8-2 and Table 8-4 of the TR mathematically define the application domains for the NSP2 and NSP4 CHF correlations, respectively. Additionally, Section 1.1 of the TR states that approval of the NSP2 and NSP4 CHF correlations is requested within the domain of applicability defined in these tables. The NRC staff finds that this is consistent with established precedent and is sufficient for restricting the domain of applicability.

3.3.3 Consistent Model Error 3.3.3.1 Poolability Poolability The validation error should be investigated to ensure that it does not contain any subgroups that are obviously not from the same population (i.e., non-poolable).

G3.3.1, Review Framework for Critical Boiling Transition Models Section 3.3.2.1 of this SE explains the process of splitting up the validation data. The NRC staff further investigated this process during an audit, where it observed that the NuScale process for 20

developing the CHFR limit involved several statistical tests to determine (1) whether subregions can be combined (i.e., pooled), and (2) whether the data can be treated as normally distributed.

The NRC staff developed a flow chart, provided in Figure 1, to clarify the NuScale process for developing the CHFR limit. The statistical tests used by NuScale, and shown in Figure 1, are consistent with those described in NUREG-1475, Revision 1, Applying Statistics (Reference 24). Based on the description of these methods in NUREG-1475, the NRC staff finds that these statistical tests are consistent with established practice and are therefore acceptable. Additionally, based on NuScale [

], the NRC staff finds the process used to select the CHFR limit acceptable.

Figure 1. [ ]

3.3.3.2 Nonconservative Subregions Nonconservative Subregions The expected domain should be investigated to determine if it contains any non-conservative subregions which would impact the predictive capability of the model.

G3.3.2, Review Framework for Critical Boiling Transition Models The NRC staff analyzed the measured-to-predicted performance of the NSP2 CHF correlation.

This analysis showed that a subregion of reduced margin existed within the application domain 21

of the NSP2 CHF correlation (at low mass flux and high quality), which caused the NRC staff to question whether the NSP2 correlation limit, proposed in Revision 0 of TR-0116-21012, is suitable for application within this subregion. Accordingly, on August 21, 2017, the NRC staff issued RAI 8931, Question 4.4-6 (Reference 12), asking NuScale to provide a means to address the low-margin subregion to ensure that CHF will not be experienced at the CHFR limit at the 95/95 level. NuScales response, provided in its letter dated September 25, 2017 (Reference 13), proposed to reduce the application domain to remove the subregion in question. The NuScale response further demonstrated the efficacy of this approach by plotting the most limiting points and showing that, by restricting the application domain, the clustered region of low margin is eliminated. The NRC staff finds this response acceptable because it eliminates the region of reduced margin. Based on the NuScale response to RAI 8931, Question 4.4-6, the NRC staff finds that the NSP2 CHFR limit is suitably conservative over the revised application domain. The NRC staff has confirmed that NuScale has incorporated the changes associated with this RAI response into Revision 1 of TR-0116-21012.

Figures 7-4 and 7-5 of the TR show that the predicted-to-measured data points exceeding the NSP4 CHFR limit occur at low mass fluxes and high qualities. Further analysis by the NRC staff showed that three data points, exceeding the NSP4 CHFR limit, occur in close proximity to each other in the low mass flux, high-quality region. However, Table 7-3 of the TR shows that the low mass flux, high-quality region is captured by the CHFR limit calculation process by partitioning the data by mass flux and local quality. Additionally, Table 3-1 of TR-0915-17564, Subchannel Analysis Methodology, dated February 15, 2017 (Reference 25), shows that normal and off-normal conditions, encountered for the NPM, are not expected to occur in the low mass flux, high-quality region. Based on the analyses presented in Section 7.7 of the TR, the NRC staff finds that the NSP4 CHFR limit is suitably conservative over the application domain established by Table 8-4 of the TR.

3.3.3.3 Model Trends Model Trends The model is trending as expected in each of the various model parameters.

G3.3.3, Review Framework for Critical Boiling Transition Models Sections 4.7 and 7.9 of the TR show the results of global sensitivity analyses for the NSP1 (which is included in the NSP2 CHF correlation) and the NSP4 CHF correlations. These studies show that the [

]. The NRC staff recognizes that these trends are consistent with previously reviewed CHF correlations. NuScales sensitivity studies also show that the CHF [

]. Data collected at Stern Laboratories (shown in Figure 4-1 of the TR) and the KATHY test loop (shown in Figure 7-10 of the TR) support this trend in the NSP2 and NSP4 CHF correlations. The NRC staff recognizes that this trend differs from the trend observed in previously reviewed CHF correlations. The NRC staff further investigated this trend through the comparison with historical CHF data provided in the 2006 CHF lookup table (Reference 26).

The NRC staff found that the trend of increasing pressure resulting in lower CHF is observed in the Groeneveld CHF lookup table. Based on the consistency in physical trends observed between the physical data and the CHF correlation, and [ ], the NRC staff finds that the NSP2 and NSP4 CHF correlations trend as expected in each of the various model parameters.

22

Section 4.4 and Section 7.6 of the TR show the measured-to-predicted performance of the NSP1 and NSP4 CHF correlations, respectively. The plots provided show that the measured-to-predicted performance of the NSP1 and NSP4 CHF correlations is consistent over the domain of applicability, [

] Section 3.3.3.2 of this SE presents the NRC staffs evaluation of this region.

Section 6.3 of the TR shows the measured-to-predicted performance of the NSP2 CHF correlation. The plots provided show that the accuracy of the NSP2 CHF correlation is dependent upon the location within the application domain. Because this trend was not observed in the NSP1 CHF correlation, which is included in the NSP2 correlation, the NRC staff recognizes that this trend is attributed to the NSPX factor (discussed in Section 3.2.1.2 of this SE). Section 5.2 of the TR describes the NSPX factor as a conservative multiplier that assures that the mean predicted-to-measured CHF value is conservatively biased [ ]. The NRC staff finds the measured-to-predicted trend for the NSP2 CHF correlation acceptable because, as described in Section 3.3.2.1 of this SE, the limiting region sets the CHFR limit.

3.3.4 Quantified Model Error 3.3.4.1 Error Data Base Error Data Base The models error should be calculated from an appropriate data base.

G3.4.1, Review Framework for Critical Boiling Transition Models As described in Section 3.3.2.1 of this SE, NuScale [

] to establish the CHFR limit for the NSP2 and NSP4 CHF correlations. Based on the use of the [ ] to set the CHFR limits, the NRC staff finds that the error data base used to determine the CHFR limits for the NSP2 and NSP4 CHF correlations acceptable.

3.3.4.2 Statistical Method Statistical Method The models error should be calculated using an appropriate statistical method.

G3.4.2, Review Framework for Critical Boiling Transition Models Section 3.2.5 of the TR describes the statistics used to develop the correlation limit. As evaluated in Section 3.3.3.1 of this SE, the statistical tests for poolability and normality are consistent with established practice and are therefore acceptable. If normality testing determines that the data are normally distributed, NuScale determines the CHFR limit by adding the standard deviation times an appropriate tolerance factor (i.e., sufficient to establish a 95/95 CHFR limit) to the predicted-to-mean distribution average. If normality testing determines that 23

the data are not normally distributed, then NuScale uses nonparametric statistics to establish a 95/95 CHFR limit. The NRC staff finds NuScales statistical methodology acceptable because it is consistent with established practice.

3.3.4.3 Appropriate Bias for Model Uncertainty Appropriate Bias The models error should be appropriately biased in generating the model uncertainty.

G3.4.3, Review Framework for Critical Boiling Transition Models Section 4.5, Section 6.3, and Section 7.7 of the TR develop the CHFR limits for the NSP1, NSP2, and NSP4 CHF correlations, respectively. The CHFR limit of 1.17 for the NSP1 correlation, presented in Section 4.5 of the TR, is based on Stern data and is selected from the value obtained for the limiting subset of data. NuScale uses the NSP1 limit of 1.17 for the NSP2 correlation and shows that this limit is bounding in Table 6-2 of the TR. NuScale selected a CHFR limit of 1.21 for the NSP4 correlation, which conservatively rounds up and bounds the value obtained for the most limiting subset of data presented in Table 7-4 of the TR. Based on the use of bounding values the CHFR limit, the NRC staff finds that the CHFR limits for the NSP2 and NSP4 CHF correlations are appropriately biased.

3.3.5 Model Implementation 3.3.5.1 Same Computer Code Same Computer Code The model has been implemented in the same computer code that was used to generate the validation data.

G3.5.1, Review Framework for Critical Boiling Transition Models Section 3.3.1 and Section 3.3.2 of the TR show that the VIPRE-01 models are used to perform the data reduction calculations in accordance with TR-0915-17594, Subchannel Analysis Methodology (Reference 25). To ensure that the NSP2 and NSP4 CHF correlations are used in a manner consistent with their validation, the NRC staff established Limitation 2 on the use of VIPRE-01 calculations using the NSP2 and NSP4 CHF correlations. Based on the description in Section 3.3.1 and Section 3.3.2 of the TR, and pursuant to Limitation 2, the NRC staff finds that the NSP2 and NSP4 CHF correlations are implemented using the same computer code used to generate validation data.

24

3.3.5.2 Same Methodology Same Methodology The models prediction of critical boiling transition is being applied in the same manner as it was when predicting the validation data set.

G3.5.2, Review Framework for Critical Boiling Transition Models As described in Section 3.3.5.1 of this SE, the NRC staff established Limitation 2 to ensure that the NSP2 and NSP4 CHF correlations are used in a manner consistent with their validation.

Based on the description in Section 3.3.1 and Section 3.3.2 of the TR, and pursuant to Limitation 2, the NRC staff finds that the NSP2 and NSP4 CHF correlations are being applied in the same manner as when predicting the validation data set.

3.3.5.3 Transient Behavior Prediction of Transient Behavior The model results in an accurate or conservative prediction when it is used to predict transient behavior.

G3.5.3, Review Framework for Critical Boiling Transition Models During an audit, the NRC staff observed that NuScale performed [

] Based on the results of these tests, the NRC staff finds that the NSP2 and NSP4 CHF correlations provide suitably conservative predictions for CHF when used to predict transient behavior.

4.0 LIMITATIONS The NRC staffs conclusions about TR-0116-21012 are subject to the following limitations:

Limitation 1 The nonuniform flux factors used in the NSP2 and NSP4 CHF correlations must always be greater than or equal to one. Section 3.2.1.2 of this SE describes the basis for this limitation.

Limitation 2 Analyses using the NSP2 and NSP4 CHF correlations must be performed in accordance with TR-0915-17564. Section 3.3.5.1 of this SE provides the basis for this limitation.

5.0 CONCLUSION

S The NRC staff approves the use of NuScale TR-0116-21012, Revision 1, subject to the limitations identified in Section 4.0 of this SE. In particular, the NRC staff finds that (1) the 25

NSP2 CHF correlation is acceptable for use in performing safety analyses of the NPM with NuFuel-HTP2TM fuel, with a CHFR limit of 1.17, over the range of applicability provided in Table 8-2 of the TR, and (2) the NSP4 CHF correlation is acceptable for use in performing safety analyses of the NPM with NuFuel-HTP2TM fuel, with a CHFR limit of 1.21, over the range of applicability provided in Table 8-4 of the TR. These findings are based on the following three observations:

1. The experimental data supporting the NSP2 and NSP4 CHF correlations are appropriate as evidenced by meeting all the supporting goals discussed in Section 3.1 of this SE.

2. The NSP2 and NSP4 CHF correlations were generated in a logical fashion as evidenced by meeting all the supporting goals discussed in Section 3.2 of this SE.

3. The NSP2 and NSP4 CHF correlations have sufficient validation, demonstrated through appropriate quantification of their error, as evidenced by meeting all the supporting goals discussed in Section 3.3 of this SE.

6.0 REFERENCES

1. TR-1113-5374-NP, Critical Heat Flux Test Program Technical Report, January 2014 (ADAMS Accession No. ML14024A455)

2. PM-1115-19264-NP, Rev. 0, NuScale Critical Heat Flux Correlation Topical Report Pre-Application Engagement with NRC, December 10, 2015 (ADAMS Accession No. ML15335A010)

3. Audit Plan for NuScale Critical Heat Flux Testing at KATHY, May 2, 2016 (ADAMS Accession No. ML16119A154)

4. Non-Proprietary Audit Report Summary for NuScale Critical Heat Flux Testing at KATHY (PROJ0769), August 12, 2016 (ADAMS Accession No. ML16204A217)

5. NuScale Power, LLC, Topical Report TR-0116-21012, Rev. 0, NuScale Power Critical Heat Flux Correlation, NSP2, October 5, 2016 (ADAMS Accession No. ML16279A363)

6. Acceptance Review of NuScale Topical Report TR-0116-21012, Rev. 0, NuScale Power Critical Heat Flux Correlation, NSP2, December 5, 2016 (ADAMS Accession No. ML16334A234)

7. NuScale Power, LLC, Submittal of Response to NRC Request for Supplemental Information to TR-01160-21012 (NRC PROJ0769), December 29, 2016 (ADAMS Accession No. ML17003A004)

8. Acceptance Letter for the Review of NuScale Topical Report TR-01160-21012, Rev. 0, NuScale Power Critical heat Flux Correlation, NSP2, (PROJ0769), February 2, 2017 (ADAMS Accession No. ML17030A049)

9. NuScale SMR TR RAIs, Topical Report Request for Additional Information Letter No. 13 (eRAI No. 8795), Section 04.04, Thermal and Hydraulic Design, May 8, 2017 (ADAMS Accession No. ML17128A468) 26

10. Audit Plan for the Regulatory Audit of NuScale Topical Report TR-0116-21012, NuScale Power Critical Heat Flux Correlation NSP2, Revision 0, May 30, 2017 (ADAMS Accession No. ML17138A113)

11. NuScale Power, LLC, Response to NRC Request for Additional Information No. 13 (eRAI No. 8795) on NuScale Topical Report TR-0116-21012, NuScale Power Critical Heat Flux Correlation NSP2, Revision 0, July 7, 2017 (ADAMS Accession No. ML17188A461)

12. NuScale SMR TR RAIsTopical Report Thermal Hydraulic StabilityRequest for Additional Information Letter No. 8922 (eRAI No. 8931), August 21, 2017 (ADAMS Accession No. ML17233A127)

13. NuScale Power, LLC, Response to NRC Request for Additional Information No. 8931 (eRAI No. 8931) on the NuScale Topical Report, NuScale Power Critical Heat Flux Correlation NSP2, TR-0116-21012, Revision 0, September 25, 2017 (ADAMS Accession No. ML17268A385)

14. Audit Summary for NuScale CHF, Non-Proprietary Version, November 27, 2017 (ADAMS Accession No. ML17278A168)

15. NuScale Power, LLC, Topical Report, TR-0116-21012-NP, Revision 1, Critical Heat Flux Correlations, November 30, 2017 (ADAMS Accession No. ML17335A089)

16. U.S. Nuclear Regulatory Commission Inspection Report No. 99901418/2013-201 and Notice of Violation, April 18, 2013 (ADAMS Accession No. ML13098A338)

17. NP-LO-0513-3717, NuScale Power, LLC, Response to Notice of Violation - NRC Inspection Report No. 99901418/2013-201 and Notice of Violation, May 21, 2013 (ADAMS Accession No. ML13141A593)

18. NRC Acknowledgement Letter, NuScales Response to the U.S. Nuclear Regulatory Commission Inspection Report No. 99901418/2013-201-01 Notice of Violation, June 6, 2013 (ADAMS Accession No. ML13154A456)

19. Notification of Re-Characterization of Findings in NRC Inspection Reports 9901351/2015-201 and 9991418/2013-201, February 6, 2017 (ADAMS Accession No. ML17027A267)

20. Y. Sato, J. Takeuchi, D. Komiyama, Y. Uno, R.A. Fortman, and G.I. Hadaller, Qualification of CHF Test Facility at Stern Laboratories using Mitsubishi PWR Fuel, NURETH-16, Chicago, IL, August 30September 4, 2015

21. APR1400-F-C-TR-12002-NP, Revision 0, KCE-1 Critical Heat Flux Correlation for PLUS7 Thermal Design, November 30, 2012 (ADAMS Accession No. ML130180119)

22. ANP-10341NP, Revision 0, The ORFEO-GAIA and ORFEO-NMGRID Critical Heat Flux Correlation, August 31, 2016 (ADAMS Accession No. ML16238A082)

23. L.S. Tong, H.B. Currin, P.S. Larsen, and O.G. Smith, Influence of Axially Nonuniform Flux on DNB, AlChE Chemical Engineering Progress Symposium, Series 62, (1966),

64:35-40 27

24. NUREG-1475, Revision 1, Applying Statistics, March 2011 (ADAMS Accession No. ML11102A076)

25. NuScale Power, LLC, Proprietary Marking Changes to Subchannel Analysis Methodology, Topical Report, TR-0915-17564 Revision 1 (NRC Project No. 0769),

February 15, 2017 (ADAMS Accession No. ML17046A333)

26. D.C. Groeneveld, J.Q. Shan, A.Z. Vasic, L.K.H. Leung, A. Durmayaz, J. Yang, S.C.

Cheng, A. Tanase, The 2006 CHF look-up table, Nuclear Engineering and Design, Vol. 237, p. 1909-1922, September 2007 28

A. CRITICAL BOILING TRANSITION MODEL ASSESSMENT FRAMEWORK A.1 Introduction Critical boiling transition (CBT) is defined as a transition from a flow regime that has a higher heat transfer rate to a flow regime that has a significantly lower heat transfer rate. In scenarios where the heat transfer is controlled by the heat flux (such as in nuclear fuel bundles), the reduction in heat transfer rate results in an increase in the surface temperature such that the heat flux can be maintained. If the reduction in the heat transfer rate and resulting increase in surface temperature is large enough, the surface may weaken or melt. In a nuclear power plant, this condition could result in fuel damage. The term CBT is meant to encompass many other terms that have been used to describe this phenomena including critical heat flux, departure from nuclear boiling, and dryout.

To ensure that the fuel is not damaged during normal operation or AOOs, computer simulations of the fuel are conducted to predict the thermal-hydraulic conditions that would occur in the fuel during various scenarios. The resulting thermal-hydraulic conditions are then input to a CBT model that determines whether a CBT has occurred and, if not, the margin to a CBT occurrence (i.e., the additional power that would be required to initiate a CBT). The NRC has historically accepted that one way to demonstrate the avoidance of fuel damage during all normal operation and AOOs is to demonstrate that there is margin to a CBT.

Because of the importance of CBT models, a major focus in reactor safety analysis is to determine that they are trustworthy. The NRC has reviewed many CBT models over the years and has documented why each model was found acceptable in the corresponding safety evaluation.

A.2 Existing CHF Correlations The NRC staff has reviewed several CBT models throughout its history. Table A-1 provides a list of several correlations that have been used in the safety analyses of light water reactors, both boiling water reactors and pressurized water reactors (note that Table A-1 is not comprehensive).

29

Table A-1. List of CBT Models CBT Model Date ML 1 B&W-2 1970 ML082490748 2 CE-1 1976 ML083010357 3 XNB DNB 1983 -

4 WRB-1 1984 ML080630433 5 WRB-2 1985 ML092430561 6 CE-1 modified) 1985 -

7 ANFP DNB 1990 -

8 WRB-2M 1999 ML081610106 ABB-NV and 9 2004 ML042610371 ABB-TV 10 BHTP 2004 ML052500092 11 HTP 2005 ML051020019 12 WSSV and WSSV-T 2007 ML072570633 ABB-NV (extended) 13 2008 ML081280713 and WLOP 14 WNG-1 2010 ML100850532 15 WRB-1 and WRB-2 2013 ML13284A071 16 KCE-1 2012 ML130180119 17 ORFEO 2016 ML16238A082 18 GE transient CHF 1971 -

19 GEXL 1977 ML092820214 20 ANFB 1990 ML081820434 21 D2 1999 ML993470286 22 D1 1999 ML003767392 23 GEXL96 2001 ML003755947 24 GEXL10 2001 ML012760512 25 GEXL80 2004 ML043210062 26 D4 2005 ML051260213 27 GEXL97 2008 ML082070090 28 GEXL17 2009 ML091830641 29 SPCB 2009 ML093650230 30 GEXL14 2011 ML111290535 31 ACE/ATRIUM-10 2014 ML14175A228 ACE/ATRIUM-10 32 2014 ML14183A748 XM 33 D5 2013 ML13333A276 30

A.3 CBT Assessment Framework The CBT assessment framework was developed by NRC staff based on engineering judgement. This engineering judgement is informed by the NRC staffs experience conducting CHF reviews as described in Section A.2 of this SE, and uses a top-down approach. This top-down approach starts with a high level goal (G) that the CBT model can be trusted in reactor safety analyses. This high level goal is decomposed into lower level sub-goals, which are further decomposed until each sub-goal is sufficiently precise as to be satisfied directly by evidence. The top level of this framework is provided in Figure A-1 and is further decomposed in Sections A.3.1, A.3.2, and A.3.3 of this SE.

Figure A-1. Decomposition of G Main Goal A.3.1 G1 Experimental Data Experimental data are the cornerstone of a CBT model. The data are used to generate the coefficients of the model and validate the model. Additionally previous data are often used to generate the form of the model. Therefore, it is essential that the experimental data are appropriate. The three subgoals in Figure A-2 are used to demonstrate that the experimental data are appropriate.

Figure A-2. Decomposition of G1Experimental Data 31

A.3.1.1 G1.1 Credible Test Facility Test facilities that are used to measure the CBT primarily focus on measuring key flow parameters which occur during the CBT event. Although such experimental data could be collected at many facilities, the time, effort, and resources needed to set up a credible facility are quite significant; therefore, most CBT data come from one of the following facilities:

• Columbia Universitys Heat Transfer Research Facility (closed in 2003).

• General Electric Companys ATLAS test loop facility in San Jose, CA (closed).

• Stern Laboratories in Hamilton, Ontario (still in use).

• AREVAs KATHY loop in Karlstein, Germany (still in use).

• Westinghouse Electric Corporations FRIGG and ODEN loops in Vsters, Sweden, for BWRs and PWRs (still in use).

The two subgoals in Figure A-3 are used to demonstrate the credibility of the test facility. NRC staff determined, based on prior experience, that this level of decomposition is sufficient, such that each subgoal could be supported directly by evidence.

Figure A-3. Decomposition of G1.1Credible Test Facility A.3.1.2 G1.2 Accurate Measurements The test facility should provide accurate measurements of all important experimental parameters, including the measurement of critical heat flux or critical power. It is important to note that the critical heat flux or critical power is not a directly measured parameter (like flow rate or pressure); instead, it is inferred from the power, axial and radial power peaking, and a determination that a CBT has occurred. The six subgoals provided in Figure A-4 are used to demonstrate the accuracy of the measurements. NRC staff determined, based on prior experience, that this level of decomposition is sufficient, such that each subgoal could be supported directly by evidence.

32

Figure A-4. Decomposition of G1.2Accurate Measurements A.3.1.3 G1.3 Reproduction of Local Conditions The local conditions in the reactor fuel bundle should be reproduced in the test bundle to ensure that experimental data taken in the laboratory apply to the reactor fuel bundle placed in the reactor. The five subgoals in Figure A-5 are used to demonstrate the reproduction of local conditions. NRC staff determined, based on prior experience, that this level of decomposition is sufficient, such that each subgoal could be supported directly by evidence.

33

Figure A-5. Decomposition of G1.3Reproduction of Local Conditions A.3.2 G2 Model Generation A CBT model should be generated in a logical fashion. This statement is intentionally broad because the decision to trust the model rests mostly on the validation data.

Although any number of methods could be used to generate a CBT model, understanding what method was used and the reasoning behind that method is helpful. The two subgoals in Figure A-6 are used to demonstrate that the model was generated in a logical fashion.

34

Figure A-6. Decomposition of G2Model Generation A.3.2.1 G2.1 The Mathematical Form The mathematical form of the model should be appropriate in that all relevant parameters appear as variables in the model and the general mathematical behavior of each variable is consistent with known physical behavior. Typically, the mathematical form of the model is chosen based on an organizations past experience. The two subgoals in Figure A-7 are used to demonstrate that the mathematical form of the model is appropriate. NRC staff determined, based on prior experience, that this level of decomposition is sufficient, such that each subgoal could be supported directly by evidence.

Figure A-7. Decomposition of G2.1The Mathematical Form A.3.2.2 G2.2 Method for Determining Coefficients The process for determining the values of the models coefficients should be appropriate. Again, as with the choice of the models mathematical form, the values of the coefficients can be chosen in a number of ways. Although only a single set of the coefficients would result in the lowest error, as judged by some norm (e.g., the Euclidian norm), minimizing this error is often not the most important criteria when determining the coefficient values. Instead, great care is usually taken to ensure that the model reflects appropriate physical behavior rather than simply minimizing the error. Thus, many of the coefficients for a model are chosen to ensure that the model has certain desired trends. The three subgoals in Figure A-8 are used demonstrate that the method for determining the coefficients is appropriate. NRC staff determined, based on prior experience, that this level of decomposition is sufficient, such that each subgoal could be supported directly by evidence.

35

Figure A-8. Decomposition of G2.2Method for Determining Coefficients A.3.3 G3 Validation through Error Quantification Because of the desire to ensure that the models prediction is conservative, any bias or uncertainty, or both, in the models prediction of CBT should be adequately quantified such that safety analyses can account for it. This process is uncertainty quantification. The first step in this process is to use the experimental data (i.e., the validation data) along with the models prediction of that experimental data to calculate the validation error. If the validation error is appropriately distributed through the models application domain and if any inconsistencies in the validation error are accounted for, statistics from the validation error can be used to determine the models uncertainty. The five subgoals in Figure A-9 are used to demonstrate that the model has sufficient validation through the quantification of its error.

36

Figure A-9. Decomposition of G3Validation through Error Quantification A.3.3.1 G3.1 Calculating Validation Error Base on prior experience, NRC staff determined that G3.1 can be supported directly by evidence and no further decomposition is necessary.

A.3.3.2 G3.2 Data Distribution in the Application Domain The validation error data should be appropriately distributed throughout the application domain.

The six subgoals in Figure A-10 are used to demonstrate that the validation error is appropriately distributed throughout the application domain. NRC staff determined, based on prior experience, that this level of decomposition is sufficient, such that each subgoal could be supported directly by evidence.

37

Figure A-10. Decomposition of G3.2Data Distribution in the Application Domain A.3.3.3 G3.3 Inconsistency in the Validation Error Statistics from the validation error is used as estimates of parameters from the population of the model application error in order to quantify the uncertainty of the CBT model. This assumes that the model application error can be described as a single population with the same distribution and parameters (e.g., mean, variance) over the entire application domain and that the validation error is a representative sample of this distribution. The three subgoals in Figure A-11 are used to demonstrate that any inconsistencies in the validation error have been appropriately addressed. NRC staff determined, based on prior experience, that this level of decomposition is sufficient, such that each subgoal could be supported directly by evidence.

Figure A-11. Decomposition of G3.3Inconsistencies in the Validation Error 38

A.3.3.4 G3.4 Calculating Model Uncertainty The CBT models uncertainty is quantified using statistics from the validation error as estimates of the parameters of the population of the model application error. Thus, the calculation of those statistics is a major focus and should be appropriate and conservative such that an uncertainty that is greater than or equal to the models actual application uncertainty is calculated. The three subgoals in Figure A-12 are used to demonstrate that the validation error has been appropriately quantified. NRC staff determined, based on prior experience, that this level of decomposition is sufficient, such that each subgoal could be supported directly by evidence.

Figure A-12. Decomposition of G3.4Quantification of the Models Error A.3.3.5 G3.5 Model Implementation Once the models uncertainty has been quantified by experimental data, the model can be applied in a reactor safety analysis. However, the implementation of the model in the analysis should be consistent with its use during validation. The three subgoals in Figure A-13 are used to demonstrate that the model has been correctly implemented. NRC staff determined, based on prior experience, that this level of decomposition is sufficient, such that each subgoal could be supported directly by evidence.

39

Figure A-13. Decomposition of G3.5Model Implementation 40