ML18157A138

Technical Evaluation of the Preliminary Safety Evaluation Report Supporting the Conversion to Low-Enriched Uranium Fuel for the National Bureau of Standards Reactors

ML18157A138
Person / Time
Site:	National Bureau of Standards Reactor
Issue date:	05/31/2018
From:	Choi S, Karimi R, Khatib-Rahbar M, Libby M Energy Research
To:	Office of Nuclear Reactor Regulation
References
ERI/NRC 18-204
Download: ML18157A138 (88)

Text

ERI/NRC 18-204 TECHNICAL EVALUATION OF THE PRELIMINARY SAFETY EVALUATION REPORT SUPPORTING THE CONVERSION TO LOW-ENRICHED URANIUM FUEL FOR THE NATIONAL BUREAU OF STANDARDS REACTOR Work Performed under the Auspices of the United States Nuclear Regulatory Commission Office of Nuclear Reactor Regulations Washington, D.C. 20555 P. O. Box 2034 Rockville, Maryland 20847 May 2018

Intentionally left blank ERI/NRC 18-204 TECHNICAL EVALUATION OF THE PRELIMINARY SAFETY EVALUATION REPORT SUPPORTING THE CONVERSION TO LOW-ENRICHED URANIUM FUEL FOR THE NATIONAL BUREAU OF STANDARDS REACTOR FACILITY OPERATING LICENSE NO. TR-5 DOCKET NO. 50-184 May 2018 M. Libby, R. Karimi, S. Choi and M. Khatib-Rahbar Energy Research, Inc.

P.O. Box 2034 Rockville, Maryland 20847-2034 Work Performed under the Auspices of the United States Nuclear Regulatory Commission Office of Nuclear Reactor Regulations Washington, D.C. 20555 Under Contract Number NRC-HQ-25-14-E-0005 Task Order Number 9

Intentionally left blank ERI/NRC 18-204 TABLE OF CONTENTS TABLE OF CONTENTS .................................................................................................................. i LIST OF TABLES .......................................................................................................................... iii LIST OF FIGURES ........................................................................................................................ v LIST OF ACRONYMS ................................................................................................................... vii LIST OF ABBREVIATIONS............................................................................................................ix LIST OF UNITS .............................................................................................................................ix

1. INTRODUCTION .................................................................................................................... 1

2. TECHNICAL EVALUATION .................................................................................................... 3 2.1 Summary of Reactor Facility Changes ........................................................................... 3 2.2 NBSR General Description ............................................................................................ 3 2.3 Comparison with Similar Facilities Already Converted ................................................... 7 2.4 Fuel and Core Design .................................................................................................... 8 2.5 Nuclear Design ............................................................................................................ 10 2.5.1 Calculational Methodology ............................................................................... 10 2.5.2 Core Reactivity ................................................................................................. 13 2.5.3 Point Kinetics Parameters ................................................................................ 15 2.5.4 Reactivity Coefficients ...................................................................................... 16 2.5.5 Burnup Effects on Analyses ............................................................................. 17 2.5.6 Power Distribution ............................................................................................ 18 2.5.7 Conclusions ..................................................................................................... 18 2.6 Thermal-Hydraulic (T&H) Design ................................................................................. 19 2.6.1 Calculation Methodology .................................................................................. 19 2.6.2 Thermal-Hydraulic Results ............................................................................... 20 2.6.3 Conclusions ..................................................................................................... 20 2.7 Accident Analysis......................................................................................................... 21 2.7.1 Introduction ...................................................................................................... 21 2.7.2 Methodology for Accident and Transient Events............................................... 22 2.7.3 Reactivity Insertion Events ............................................................................... 24 2.7.4 Loss-of-Flow Events ......................................................................................... 34 2.7.5 Loss-of-Coolant Accident (LOCA) .................................................................... 41 2.7.6 Natural Circulation at Low Power Event ........................................................... 54 2.7.7 Flow Blockage Accident - the Maximum Hypothetical Accident ........................ 54 2.7.8 Mishandling, Malfunction, or Misloading of Fuel ............................................... 62 2.7.9 Experiment Malfunction .................................................................................... 63 2.7.10 Loss of Normal Power ...................................................................................... 64 2.7.11 Equipment Malfunction ..................................................................................... 64 2.7.12 External Events ................................................................................................ 64 2.7.13 Accident Analysis Conclusions ......................................................................... 64 2.8 Fuel Storage ................................................................................................................ 64 i

ERI/NRC 18-204 2.9 LEU Startup Plan ......................................................................................................... 66 2.10 Proposed Changes to License Conditions and Technical Specifications...................... 66 2.10.1 Proposed Changes to License Conditions ........................................................ 66 2.10.2 Proposed Changes to Technical Specifications ................................................ 66 2.10.3 Conclusions ..................................................................................................... 67

3. CONCLUSIONS.................................................................................................................... 69

4. REFERENCES ..................................................................................................................... 71 ii

ERI/NRC 18-204 LIST OF TABLES Table 2-1 HEU and the Expected LEU Fuel Characteristics ................................................... 5 Table 2-2 Key Neutronics Parameters .................................................................................. 14 Table 2-3 Shutdown Margin and Excess Reactivity (%k/k) ................................................. 15 Table 2-4 NBSR Point Kinetics Parameters ......................................................................... 16 Table 2-5 TRACE and RELAP5 Uncontrolled Shim Safety Arm Withdrawal Results............. 31 Table 2-6 Maximum Reactivity Event Results ....................................................................... 33 Table 2-7 Break Characteristics from the PSAR ................................................................... 43 Table 2-8 18-inch GBLOCA - Case 1 .................................................................................... 44 Table 2-9 14-inch GBLOCA - Case 2 .................................................................................... 45 Table 2-10 10-inch GBLOCA - Case 3 .................................................................................... 46 Table 2-11 TRACE and RELAP Comparison of Draindown Analysis for Case 2 ..................... 53 Table 2-12 Estimates of MHA Source Term Nuclide Inventory ............................................... 56 Table 2-13 Leak Rates to Confinement and Release Rates to Stack ...................................... 57 Table 2-14 Ten-Minute Dose (TEDE) to NBSR Staff after MHA .............................................. 58 Table 2-15 MHA Occupational Worker Dose Estimates in Restricted Locations ..................... 58 Table 2-16 Dose (TEDE) to an Individual after MHA ............................................................... 60 Table 2-17 Offsite Dose Consequences ................................................................................. 61 Table 2-18 Accident Analysis Summary.................................................................................. 65 Table 3-1 Conversion Open Items ........................................................................................ 69 Table 3-2 General Open Items ............................................................................................. 70 iii

ERI/NRC 18-204 Intentionally left blank iv

ERI/NRC 18-204 LIST OF FIGURES Figure 2-1 NBSR Core and Irradiation Facilities ....................................................................... 6 Figure 2-2 NBSR Confinement Building, Elevation View .......................................................... 7 Figure 2-3 MTR Fuel ................................................................................................................ 8 Figure 2-4 Core Plate............................................................................................................... 9 Figure 2-5 RELAP5 Model ..................................................................................................... 23 Figure 2-6 TRACE Model ....................................................................................................... 23 Figure 2-7 Startup Event Cladding Temperature .................................................................... 26 Figure 2-8 Startup Event CHFR ............................................................................................. 26 Figure 2-9 Maximum Reactivity Insertion Event Cladding Temperature ................................. 27 Figure 2-10 Maximum Reactivity Insertion Event CHFR .......................................................... 28 Figure 2-11 Maximum Reactivity Insertion Event OFIR ............................................................ 29 Figure 2-12 TRACE Model for Confirmatory Analysis .............................................................. 30 Figure 2-13 Power Response to Uncontrolled Shim Safety Arm Withdrawal ............................ 31 Figure 2-14 Fuel Temperature Response to Uncontrolled Shim Safety Arm Withdrawal .......... 32 Figure 2-15 Power Response to Maximum Reactivity Event .................................................... 33 Figure 2-16 Temperature Response to Maximum Reactivity Event .......................................... 34 Figure 2-17 Cladding Temperature after LOSP ........................................................................ 35 Figure 2-18 Cladding Temperature After Pump Seizure........................................................... 36 Figure 2-19 Cladding Temperature after Valve Closure ........................................................... 37 Figure 2-20 Cladding Temperatures after Flow Control Valve Closure ..................................... 38 Figure 2-21 Cladding Temperature after DMW-19 Closure ...................................................... 40 Figure 2-22 Confirmatory Case-2 LOCA Graphic - Post Blowdown.......................................... 51 Figure 2-23 Core Water Level in the Hot Plate ......................................................................... 52 v

ERI/NRC 18-204 Intentionally left blank vi

ERI/NRC 18-204 LIST OF ACRONYMS Acronym Definition ADAMS Agency-wide Documents Access and Management System ASAI all shim safety arms inserted BOC beginning of cycle BWI Baltimore-Washington International CFR Code of Federal Regulations CHF critical heat flux CHFR critical heat flux ratio CNS cold neutron source DAC derived air concentration DNB departure from nucleate boiling DNBR departure from nucleate boiling ratio DOE Department of Energy ECP estimated critical position ECT emergency cooling tank EOC end of cycle ERI Energy Research Incorporated GBLOCA guillotine break LOCA HEU highly enriched uranium IRT inner reactor tank ITC isothermal temperature coefficient LCC limiting core configuration LCO limiting condition for operation LEU low enrichment uranium LOCA loss of coolant accident LOFA loss of flow accident LOSP loss of offsite power LSSS limiting safety system setting MHA maximum hypothetical accident MTC moderator temperature coefficient MTR materials test reactor NBSR National Bureau of Standards Reactor NBSR-14 2004 SAR vii

ERI/NRC 18-204 Acronym Definition NCNR National Center for Neutron Research NIST National Institute of Standards and Technology NRC Nuclear Regulatory Commission OCC operating core configuration OFI onset of flow instability OFIR onset of flow instability ratio PSAR Preliminary Safety Analysis Report RAI request for additional information RG Regulatory Guide SAR Safety Analysis Report SBLOCA small break LOCA SDM shutdown margin SDP shutdown pumps SER Safety Evaluation Report SFA single failure analysis SL safety limit SOE sequence of events SOE sequence of events SR surveillance requirement TEDE total effective dose equivalent TER Technical Evaluation Report TS technical specification viii

ERI/NRC 18-204 LIST OF ABBREVIATIONS Abbreviation Definition D2O heavy water H2O light water SU startup T&H thermal and hydraulic LIST OF UNITS Units Definition k/k / %k/k reactivity / reactivity x 1000.0

°C degrees centigrade cfm cubic-feet-per-minute Ci Curie g grams gpm gallons-per-minute l liter m / cm meter / centimeter MW MegaWatt pcm percent millirho rem / mrem roentgen-equivalent-man / millirem s / s second / microsecond ix

ERI/NRC 18-204 Intentionally left blank x

ERI/NRC 18-204

1. INTRODUCTION Title 10 of the Code of Federal Regulations (10 CFR) Section 50.64 requires licensees of research and test reactors to convert from the use of high-enriched uranium (HEU) fuel to low-enriched uranium (LEU) fuel, unless specifically exempted. The National Bureau of Standards Reactor (NBSR) is operated by the National Institute of Standards and Technology (NIST) National Center for Neutron Research (NCNR). NIST is a division of the U.S. Department of Commerce. NCNR (the licensee) has been involved in an ongoing process led by the Department of Energy (DOE) to develop a fuel design that can be used to convert the NBSR from the use of HEU to LEU fuel.

In a letter dated December 30, 2014, the licensee submitted a Preliminary Safety Analysis Report (PSAR) (Ref. 1) wherein it provided technical information describing the analysis methods which will be utilized when NIST submits the request for licensing action for Nuclear Regulatory Commission (NRC) approval of the planned conversion. This PSAR proposes certain changes to the safety analysis report (NBSR-14) (Ref. 2) as approved by the NRC in the Safety Evaluation Report (SER) [Ref. 3]) in 2004. To support this review, the licensee has also submitted an updated version of the safety analysis report (SAR) (Ref. 4) which has not yet been reviewed by the NRC staff.

This review is accomplished by the Energy Research Inc. (ERI) staff and it also considers the previous NRC staff assessment performed as part of the consideration of Fukushima-related issues (Ref. 5 and Ref. 6), the previously approved SER, the PSAR, the revised SAR and the supporting documents provided by the licensee. This Technical Evaluation Report (TER) establishes the level of agreement between the PSAR and the guidance in NUREG-1537 (Ref. 7).

During this process certain aspects of the PSAR review required additional information that the ERI staff requested by preparing the input to requests for additional information (RAI) (Ref. 8 and Ref. 9). The licensee responses to these RAIs (Ref. 10 and Ref. 11) provides significant additional information. The present TER includes an evaluation of the responses to these RAIs by the licensee.

This report refers to the two RAI documents (Set 1 and Set 2). Hence, references to RAIs will include the RAI number and the set number (e.g., RAI-1S1 is RAI-1 from the Set 1 document

[Ref. 10]) and Set 1 provides the responses to RAI-1S1 through RAI-15S1. Set 2 (Ref. 11) provides the responses to RAI-1S2 through RAI-11S2.

The responses have been categorized in the following manner:

• The response is relevant to conversion and is considered satisfactory. The acceptability of the response is documented in the TER and there is no outstanding technical issues associated with such a response.

• The response is determined to be unsatisfactory. Either additional information would be required (e.g., fuel irradiation data) or the response needs to be revised to satisfy the guidance in NUREG-1537 in order to form the conclusions required to support conversion.

A conversion open item (COI) statement is provided herein indicating the shortcoming in the response by the licensee. An ERI letter report is issued that parses this information from the TER (Ref. 12).

1

ERI/NRC 18-204

• The response is determined to be unsatisfactory. The information in the response indicates that there is an issue that has arisen because of the review of the conversion PSAR, but which is of broader importance than for aspects of the fuel conversion. A general open item (GOI) statement is provided herein identifying the shortcoming in the response. An ERI letter report is issued that parses this information from the TER (Ref. 13).

2

ERI/NRC 18-204

2. TECHNICAL EVALUATION 2.1 Summary of Reactor Facility Changes According to PSAR Section 4.2.1.1, the NBSR is presently fueled with HEU at a nominal uranium-235 (U-235) enrichment of ~93%. The fuel is in the form of U308 in an aluminum powder dispersion that is clad in aluminum alloy. Each fuel element is constructed of 17 plates in each upper and lower half of the core (34 plates per fuel element) and uses what is called the Materials Test Reactor (MTR) plate geometry. In the case of NBSR, the plates are curved.

As stated in PSAR Section 1.1.1, the licensee anticipates that the only change to the facility from the anticipated conversion is the fuel plates used within the assembly. At this time, the anticipated fuel meat for the LEU conversion uses U10Mo metal alloy foils with aluminum alloy cladding and a zirconium interlayer between the foil and cladding. The external geometry for the LEU fuel element is expected to be identical to that of the present HEU fuel. The thickness of the fuel meat and the cladding are expected to change. The U-235 content of each fuel element is expected to increase from 350 grams for HEU to 383 grams for LEU using the uranium alloy that is less-than 20% enriched in U-235.

As stated in PSAR Chapter 14, the licensee anticipates that certain changes to Technical Specifications (TS) will be required to support the planned fuel conversion at the facility. They involve changes to the safety limits and design features that are directly connected to the fuel conversion. They are described in PSAR Section 14.1 and 14.2.

In PSAR Table 1.1 the licensee provides a general description of fuel element changes and they are then reproduced in Table 2-1.

It is noted from Table 2-1 that the fuel meat mass is increased as a consequence of the design change from HEU to LEU. In RAI-9S2 the licensee confirms that the weight of the fuel meat in one HEU fuel plate is ~1.069 kg and the corresponding LEU mass is 2.164 kg. However, since the fuel design is preliminary, a complete evaluation is not available at this time. At the time that the fuel design is finalized it will be necessary for the licensee to obtain and provide assurance that the mass difference is acceptable for the dynamic response of the fuel to events such as seismic events and fuel handling. For this reason, the ERI staff finds that the response to RAI-9S2 does not resolve the information requirement; therefore, it needs to be reconsidered as part of the further submissions by the licensee. (COI-1) 2.2 NBSR General Description The NBSR, as described in PSAR Section 1.1, is a heavy water (D2O) moderated and cooled, tank type reactor designed to operate at 20 MegaWatts (MW). It is a custom designed variation of the Argonne CP-5 design reactor. It differs from CP-5 in its power rating, core configuration and changes to accommodate production of cold neutrons, but otherwise, it retains many other aspects of the CP-5 design. The three most notable modifications to the basic design are:

• a gap between the upper and lower fuel regions in each fuel element to reduce the fast neutron background near the beam tube axis; 3

ERI/NRC 18-204

• a double plenum at the bottom of the reactor vessel to provide optimized cooling to the core; and

• a method for the remote handling of fuel elements during refueling.

The NBSR facility is described in NBSR-14 Section 6.1.2. The NBSR is an above-ground concrete structure housed within a confinement building. The confinement building is specifically designed to support reactor operation and consists of a single building which houses the reactor and other support areas. The NBSR confinement is a steel-reinforced concrete structure and during normal operation, it is maintained under a slight negative pressure with respect to the ambient conditions.

NBSR includes a large number of experimental facilities. The principle experimental facilities that influence reactor behavior are the substantial number of beam tubes illustrated in Figure 2-1 (Ref. 1).

According to the NBSR-14 Chapter 3, the reactor confinement building is designed and constructed to have a maximum building leakage rate of less than 24 cubic-feet-per-minute (cfm) per inch of differential pressure. The lower-level of the confinement building includes the equipment room and the coffer dam that are discussed further in consideration of the loss of coolant accident (LOCA) analysis.

According to NBSR-14 Section 5.1, the reactor coolant system includes the following:

• The primary coolant system,

• The secondary coolant system,

• The primary coolant purification system,

• The primary coolant makeup system,

• The Helium Sweep System,

• The Nitrogen-16 control system, and

• The D2O experimental cooling system.

4

ERI/NRC 18-204 Table 2-1 HEU and the Expected LEU Fuel Characteristics Property Units HEU LEU as expected Fuel Meat Composition U3O8-Al 1 U10Mo 2 U-235 g 350 383 U-238 g 26 1556 O g 68 0 Al g 625 0 Mo g 0 215 Total grams 1069 2154 Fuel meat density g/cm3 3.61 17.2 Fuel thickness in 0.020 0.0085 Fuel width in 2.415 2.415 Fuel length in 11 11 Total fuel volume cm3 296.0 125.9 Fuel plate length in 13 13 Fuel plate width in 2.68 2.68 Fuel plate thickness in 0.0501 0.0501 Fuel plate radius of curvature in 5.5 5.5 Average U-235 burnup  % 70 60 1 The current HEU fuel design is a dispersion powder.

2 The proposed LEU fuel design is a metal alloy.

5

ERI/NRC 18-204 Figure 2-1 NBSR Core and Irradiation Facilities 6

ERI/NRC 18-204 Figure 2-2 NBSR Confinement Building, Elevation View The primary purposes of the reactor coolant systems are: to remove the core-generated fission and decay heat; to dissipate the heat to the environment; and to serve as barriers to prevent fission product release to the environment. The primary coolant is heavy water (D2O) and the secondary coolant is light water (H2O). According to NBSR-14 Section 5.3.2.2, the ultimate heat sink for the NBSR is the hybrid wet/dry, plume abatement cooling tower.

The primary coolant system normally operates under conditions of forced flow. Primary coolant enters the bottom of the reactor vessel through the inner and outer plenums. The inner plenum feeds primary coolant to the center six fuel assemblies via a 10-inch pipe, while the outer plenum feeds the remaining twenty-four fuel assemblies via a 14-inch pipe that is concentric to the 10-inch pipe. The coolant flows up through the fuel before exiting from the bottom of the vessel through two 18-inch outlet pipes.

2.3 Comparison with Similar Facilities Already Converted There have been changes to other plate-type fueled domestic reactors in the past. The changes to those facilities were successful in that the expected performance was validated by the operation of the revised fuel design. The same approach is now being undertaken under the leadership of DOE. However, those previous changes to other reactors employed uranium-silicide dispersion fuel and not the alloy design currently under consideration for NBSR.

7

ERI/NRC 18-204 2.4 Fuel and Core Design According to discussions with the licensee, NBSR currently expects to gradually introduce the use of the LEU fuel as refueling requires. During the transition, both fuel types will be present, and their interactions need to be considered. The HEU fuel elements are described in NBSR-14.

The current LEU fuel is illustrated in Figure 2-3.

Figure 2-3 MTR Fuel The behavior of the LEU metal alloy fuel has not yet been fully characterized and the irradiation tests are currently ongoing. Thus, this LEU fuel is not yet approved for use in research and test reactors by the NRC. The licensee submitted this application because it offers the opportunity to capture the current state-of-knowledge and illustrates how this fuel change will be analyzed when final characterization is completed, and it is released for use.

In the first response to RAIs the licensee provided results of neutronics analysis (Ref. 10) with the fuel arrangement shown in Figure 2-4. Since the fuel elements in each row are offset from the previous row, the core arrangement takes on an unusual configuration in the actual core. The curved plates in the rectangular fuel elements are located in a hexagonal grid.

8

ERI/NRC 18-204 Figure 2-4 Core Plate RAI-1aS2 requests that the licensee describe how the core loading will transition from HEU to LEU, and further, to describe the limiting core configuration (LCC) that envelops the transition and LEU cores. The licensee response states that PSAR Section 4.5.1.2 describes the model used to calculate the performance parameters for both the HEU and LEU equilibrium cores. The licensee has considered several transition schemes to get from the current HEU core to an equilibrium LEU core, but at this time indications are that the fuel supply may be the determining factor and it is premature to speculate about this factor at this time. The ERI staff understands that this response represents the state of knowledge available at this time. However, the submittal requesting formal acceptance is required to define the operational core configuration (OCC) to be loaded into NBSR and an LCC that is shown to be bounding for that OCC. For this reason, the ERI staff finds that the response to RAI-1aS2 does not resolve the information requirement.

Therefore, additional information the RAI needs to be obtained as part of the future submittals by the licensee. (COI-2)

In reviewing Figure 2-4, it is noted that the NBSR fuel consists of a series of curved plates.

However, the PSAR does not indicate the orientation of the curved fuel plates with respect to each other. RAI-1bS2 requests the licensee to identify whether the core performance is sensitive to the fuel orientation. In the response the licensee states that the location of the edge of the curved portion of the fuel plate changes by ~0.4 cm from the middle to the edge whereas the neutron diffusion length in heavy water is ~120 cm with 0.16% H2O impurity. The licensee states that this difference in location is insignificant neutronically. Furthermore, the concave-convex orientation of any bundle is not known and has never been tracked. The ERI staff finds that this response is acceptable because of the difference between the curvature and the diffusion length cited. Based upon the information provided by the licensee in the response, the ERI staff finds that RAI-1bS2 is acceptably resolved.

9

ERI/NRC 18-204 RAI-1cS2 reports differences between certain parameters in two PSAR tables. In the response the licensee states that the incorrect parameter in both tables is the total moderator coefficient for HEU fuel at beginning of cycle (BOC). The components of the coefficient for scattering kernel and density are correct and the total reactivity coefficient should be -0.0312 %k/k/ºC. Based upon the information provided by the licensee in the response, the ERI staff finds that RAI-1cS2 is acceptably resolved.

The PSAR does not describe the control elements in the NBSR. Instead, it states that they are unchanged as a consequence of any HEU to LEU change. NBSR-14 describes the control elements as consisting of 1 regulating rod and 4 shim safety arms. The regulating rod is located in a thimble within the periphery of the active fuel assemblies and the shim safety arms swing between fuel assemblies in an arc that is described in NBSR-14.

2.5 Nuclear Design 2.5.1 Calculational Methodology PSAR Section 4.5.1 describes the neutronics methodologies used to describe the NBSR behavior. Therein, the licensee states that MCNP is used to develop a basic model of the NBSR and MCNPX v2.7.0 which includes a burnup capability used to determine the core inventory and to analyze the core neutronics of the NBSR for the current report. The text of the PSAR uses the terms MCNP and MCNPX and it is sometimes unclear which code is actually used. However, the ERI staff are sufficiently clear on the capabilities of both codes and this lack of clarity does not affect the conclusions of this review. The information discussed in this section establishes the design basis for the content of other chapters, specifically the thermal-hydraulic (T&H) analysis, safety analysis, and portions of the TSs. To demonstrate that the neutronics models are suitably predictive, calculations are presented demonstrating the ability to reproduce measurements from an operational core configuration (OCC).

There are three aspects of this topic that need to be individually discussed. They are the ability of the model to suitably predict (1) reactivity, (2) power distributions, and (3) temperature effects.

According to PSAR Section 4.5.1.2 (Ref. 1) it is not unusual for the calculated value of keff to exhibit a bias. The ERI staff conceptually agrees and notes that a bias is a consistent deviation from the actual value given a sufficiently robust set of comparisons. The individual deviations are typically determined by calculating the value of keff for a known critical condition and comparing that value to the known value, which for a critical configuration is unity. If the deviation is reasonable in magnitude and consistent then it may be appropriate to use this bias to adjust the calculated predictions. In the PSAR the licensee referenced the Bess TRIGA results (Ref. 14) as a basis for their bias. RAI-2S2 states that the reactor described in the Bess paper uses TRIGA fuel (Fuel Life Improvement Program fuel and 30/20 fuel with erbium). This fuel has low burnup and a very significant calculated bias of 1000 percent millirho (pcm) is claimed. The eigenvalue spread appears to be too large to be a defendable bias and the paper focuses on library evaluations that are not true ECPs. This paper does not appear to be applicable to NBSR calculations. In response to RAI-2S2 the applicant withdrew the reference to these results. Based upon the information provided by the licensee in the response, the ERI staff finds that RAI-2S2 is acceptably resolved.

10

ERI/NRC 18-204 The methods used by the licensee are detailed in the response to RAI-1S1. In response to that RAI the licensee states that the neutronics calculations include 3-D models using the MCNPX code. MCNPX is used to provide the power distribution and reactivity information used by NBSR and for most routine neutronics calculations. The licensee then states that to provide full qualification of the use of the code, it is then used to reproduce measurements made at the NBSR.

The licensee uses measurements of critical shim safety arm position (hence worth) and the isothermal temperature coefficient (ITC) for comparisons. PSAR Section 4.5.1.1 states that the MCNPX model calculates the reactivity for the shim safety arm position (the estimated critical position [ECP]) of ~1.006 consistently using an idealized fuel cycle of 38.5 days implying a deviation for the present methodology of 0.6%k/k. Furthermore, the response includes Figure 1-1 that illustrates several cycles of consistent ECP predictions. The ERI staff finds that the NBSR bias of 0.6 %k/k is acceptable because the MCNPX model consistently replicates the measured core reactivity thus demonstrating that the model is suitably predictive for core reactivity. Based upon the information provided by the licensee in the response, the ERI staff finds that RAI-1S1 is acceptably resolved.

As stated in PSAR Section 4.5.2, the safety analyses require a series of calculations that model the behavior of the NBSR under a set of off-normal conditions to show that fuel integrity will not be compromised under any credible scenario. RAI-2S1questions the use of 60 homogenized compositions to represent the spatial burnup of 1020 fuel plates in the active core. In the response the licensee describes the details of how the fuel is modeled. The response referred to the supplied Brown report (Ref. 15) which comprehensively discusses this topic. However, the ERI staff can find no indication that the information and factors determined in the Brown report are utilized in the PSAR analysis. (COI-5)

Since the response to RAI-2S1 was not fully acceptable, the issue was raised again in RAI-3S2 requesting justification for the small number. In the response to that RAI, the licensee states that although the number of compositions in the model is small when compared to the number of fuel plates, the model has been validated (see RAI-1S1 and 2S1) and the model used successfully to renew the license of the NBSR in 2010, only had 30 unique compositions. The ERI staff finds that 60 is an improvement over 30, but, it does not provide the justification requested. (COI-5)

In the response to RAI-15S1 the licensee supplied the Brown report which provides a comprehensive analysis regarding the effect of the MCNPX model on the accuracy of burnup and power distributions. In the summary of that report the following statements are made:

Detailed analysis of the burn-up results justifies lumping the majority of the burn-up nodes. However, nodes near thermal flux peaks cannot be realistically lumped.

Lumping nodes near thermal flux peaks with dissimilar nodes will result in unrealistic burn-up and power distributions throughout the fuel element. Non-physical lumping of the nodes near thermal flux peaks will also bias the reactivity worth of the fuel element, although the impact of this bias on the full-core reactivity and the fuel cycle length has not been assessed.

11

ERI/NRC 18-204 The ERI staff finds conflicts between these statements and how the NBSR core is modeled by the licensee. Since the conclusion is that lumping results in unrealistic results, there seems to be no justification for the lumping used in the licensee models. (COI-6)

There are several potential applications of this work. The single-element results could be utilized to generate burn-up correction factors for the full-core power distributions.

These correction factors would be applied to the whole core power distribution as a function of irradiation history, axial location, and plate. Another approach would be to utilize the single-element results to inform the calculation of a new equilibrium core utilizing an improved scheme for lumping the fuel element inventories if more fuel compositions were to be modeled. In this new scheme, the location of the thermal flux peaks inherent in the NBSR fuel element design would be incorporated into the choice of nodes for lumping the fuel element burn-up.

In concept, the ERI staff agrees with this analysis. However, it is unclear whether such correction factors are used in the submitted PSAR analysis or if some other technique (e.g., uncertainty factors) is used to account for the observed effects. (COI-6)

This study ignores the effect of the shim arms, which have significant impact on the local burn-up distribution. In this study the synergistic burn-up of the mid-plane gap and plate-to-plate self-shielding were ignored, and the two effects were isolated by selection of distinct burn-up meshing for each effect. However, in reality the axial and plate-wise burn-up will occur synergistically, further altering the burnup and power distributions. Quantification of these effects could be the focus of future studies.

The response does not help in developing an understanding whether the shim safety arm effects are factored into the PSAR analysis. (COI-6)

Regarding temperature effects, these are typically assessed using comparisons of isothermal temperature coefficient measurements and calculations. Although calculations of such coefficients are provided, no comparisons to measurements are included. (COI-4)

These MCNPX analyses requires data such as the limiting planar and axial power distributions, neutron kinetics data, reactivity feedback coefficients, control blade worths, control blade drop times, fluid flow rates, pump-on and pump coast-down characteristics, as well as the operating conditions at the beginning of specific transient scenarios. These include: power level, inlet temperature, system coolant pressure, and operating history for the consideration of decay heat.

The Brown report makes an important contribution towards documenting the effects that need to be modeled in order to arrive at a sufficiently conservative LCC. However, there is no evidence that the information and factors determined in this report is incorporated into the PSAR. An alternative to the Brown report is to use a large enough number of materials to model the burnup distributions. However, since neither is in evidence in the PSAR, the ERI staff finds that although the information supplied in the response to RAI15S1 is useful, the response to RAI-2S1 and RAI-3S2 do not resolve the information requirement allowing the formation of a basis of acceptability for the MCNPX model to describe the power distribution for the analysis of the LCC.

(COI-5) 12

ERI/NRC 18-204 The parameters determined from HEU and LEU analyses supplied by the licensee are listed in Table 2-2. However, since the discussion in this section indicates that certain aspects of the analysis requires further explanation, the acceptability of this information cannot be determined at this time. These issues need to be reconsidered as part of the future submittals by the licensee.

(COI-3) 2.5.2 Core Reactivity It is recognized that the core reactivity is affected by factors that include the combination of excess reactivity and control element position and worth. In addition, the specific issue of shutdown margin is important to safety. One way to interpret the acceptability of the neutronics model - in this case MCNPX - is to examine the ability of that model to estimate the estimated critical position (ECP) of the control elements. This is done by using historical data of known critical conditions and modeling those configurations in MCNPX.

RAI-6S1 raises the question of whether the coarse burnup distribution afforded by use of only 60 materials is suitable for obtaining accurate ECPs. In response, the licensee states that the responses to RAI-2 and -5 explain this issue. However, the review of those responses does not support such a conclusion. The licensee demonstrates in the response to RAI-1S1 the ability to calculate shim safety arm reactivities that match critical measurements to within 0.3 %k/k. This is somewhat complicated by the statement in the response that shim safety arms are periodically changed, but without indicating in the graphic provided when these changes take place. It is noted that in more recent cycles the discrepancy between calculations and measurements is significantly increasing. Although the average ECP is acceptable, the ERI staff finds that the model accuracy is diminishing with time. This and other aspects of the model indicate that a more complete model is required to support future safety analyses. For this reason, the ERI staff finds that the response to RAI-6S1 does not resolve the information requirement allowing the formation of a basis of acceptability. (COI-4)

NUREG-1537, Section 4.5.1, recommends that the licensee establish an LCC. LCC is defined as the core configuration that would yield the highest power density using the fuel authorized for use in the reactor. The PSAR compares the characteristics of the HEU and LEU fuel to determine the LEU characteristics that should be expected to be present an LCC in subsequent submittals.

It is for this reason that this review considers the applicability of the methodologies employed and whether the TS presented address the necessary topics suitably.

The PSAR addresses the issue of shim safety arm worth comparisons in a unique manner.

Typically, licensees present calculations and measurements of individual shim safety arm worths and show acceptable comparisons. In the PSAR the licensee shows the acceptability in concert with the ECP results. The ERI staff finds that this is an acceptable approach and since the ECP results are acceptable, the ERI staff further finds that the control element worths are also acceptable.

13

ERI/NRC 18-204 Table 2-2 Key Neutronics Parameters Parameter HEU Core LEU Core Excess reactivity (%k/k) 6.7 6.3 Shutdown margin with highest worth shim safety arm out (%k/k) -10.1 -10.8 keff with moderator at dump level, BOC 0.9857 0.9849 keff with moderator at dump level, EOC 0.9124 0.9215 Shim arm worth, BOC (%k/k) 24.9 24.2 Shim arm worth, EOC (%k/k) 27.2 26.0 Regulating rod worth, BOC (%k/k) 0.50 0.53 Regulating rod worth, EOC (%k/k) 0.45 0.43 Moderator temperature coefficient, BOC (%k/k/°C) -0.0397 -0.0280 Moderator temperature coefficient, EOC (%k/k/°C) -0.0275 -0.0228 Void coefficient, all thimbles voided, BOC (%k/k/liter) -0.038 -0.039 Void coefficient, all thimbles voided, EOC (%k/k/liter) -0.031 -0.032 Void coefficient, all FEs voided, BOC (%k/k/liter) -0.019 -0.018 Void coefficient, all FEs voided, EOC (%k/k/liter) -0.022 -0.022 Reactivity insertion for CNS flooded, BOC (%k/k) 0.24 0.15 Reactivity insertion for CNS flooded, EOC (%k/k) 0.25 0.15 Reactivity insertion flooding of one beam tube, BOC (%k/k) 0.27 0.26 Reactivity insertion flooding of one beam tube, EOC (%k/k) 0.20 0.26 Peak half-element relative power, BOC 1.28 1.35 Peak half-element relative power, EOC 1.18 1.15 Peak half-element relative power with misloaded fuel 1.93 1.83 Delayed neutron fraction, BOC 0.00665 0.00649 Delayed neutron fraction, EOC 0.00662 0.00649 Recommended prompt neutron lifetime, BOC (s) 650. 600.

Recommended prompt neutron lifetime, EOC (s) 750. 700.

14

ERI/NRC 18-204 TS 3.1.2, Reactivity Limitations, states that the core cannot be loaded such that the excess reactivity can exceed 15%k/k and it also states that the NBSR shall not be operated if it cannot be kept shutdown with the most reactive shim safety arm fully retracted. According to PSAR Section 4.5.2.2 (Ref. 1) to determine if these conditions are met, keff was calculated under the following conditions: all shims inserted (shutdown reactivity), all shim safety arms withdrawn (excess reactivity), and three of the four shim safety arms inserted with the other withdrawn (shutdown margin, SDM). The calculations were performed at the most reactive state point in the cycle, which is BOC with four fresh fuel elements and no Xe-135 poison. The results are listed in Error! Not a valid bookmark self-reference.. These calculations are made at the most reactive time for each core which is cycle BOC with 4 fresh fuel assemblies. The ERI staff finds that the excess reactivity of the HEU and LEU cores described satisfy the excess reactivity and single-shim removed shutdown requirements.

Table 2-3 Shutdown Margin and Excess Reactivity (%k/k)

Condition HEU LEU Shutdown reactivity (all shim safety arms inserted [ASAI]) -18.2 -18.3 SDM ASAI except that shim 1 is out -12.1 -12.2 SDM ASAI except that shim 2 is out -11.1 -11.2 SDM ASAI except that shim 3 is out -10.1 -10.8 SDM ASAI except that shim 4 is out -11.6 -11.9 Excess reactivity (all shim safety arms out) 6.7 6.3 The review of the licensees analysis indicates that HEU and LEU fuel designs satisfy both excess reactivity and shutdown margin under the current TS. Therefore, the ERI staff concludes that the ability to shut down the reactor will be maintained after conversion.

2.5.3 Point Kinetics Parameters PSAR Section 4.5.4.1 provides the methodology used to calculate the point kinetics parameters.

MCNP5-1.60 was used to perform the calculations. These parameters are then used to support the point kinetics analyses. The calculated values of eff, i, and i are shown in Table 2-4. The results show that the LEU core has a slightly smaller delayed neutron fraction relative to the HEU core. This is a result of the increase in the contribution from Pu-239 fission which has a smaller delayed neutron fraction than U-235.

15

ERI/NRC 18-204 Table 2-4 NBSR Point Kinetics Parameters 3 BOC HEU LEU Group i (1/s) i i (1/s) i 1 0.01249 0.00022 0.01249 0.00020 2 0.03182 0.00111 0.03177 0.00108 3 0.10938 0.00107 0.10942 0.00105 4 0.31700 0.00301 0.31731 0.00301 5 1.35386 0.00092 1.35205 0.00085 6 8.63611 0.00032 8.65543 0.00030 eff = i 0.00665 0.00649 EOC HEU LEU Group i (l/s) i i (1/s) i 1 0.01249 0.00021 0.01249 0.00020 2 0.03182 0.00112 0.03176 0.00109 3 0.10938 0.00110 0.10942 0.00102 4 0.31700 0.00302 0.31730 0.00301 5 1.35374 0.00087 1.35118 0.00087 6 8.63558 0.00030 8.65038 0.00030 eff = i 0.00662 0.00649 The ERI staff has reviewed the kinetics parameters provided and finds that they are typical of MTR designs, consistent with previously supplied parameters for this reactor, and the methodology employed is acceptable as are the parameters themselves.

2.5.4 Reactivity Coefficients According to PSAR Section 4.5.2.4 the moderator temperature coefficient (MTC) was calculated using the two ways of representing the temperature change. First the scattering kernel was changed from 20°C to 77°C. The value of keff (and k/k) was calculated and then divided by the temperature change. Second, the density was changed from that at 46°C to 96°C in 100°C increments maintaining the 20°C scattering kernel. For each temperature step the value of keff was calculated and k/k was divided by the temperature change. The values of k/k/°C were 3 Three of the four values of eff in PSAR Table 4.16 are incorrect in the last decimal place; corrected here.

16

ERI/NRC 18-204 then averaged. The values of reactivity change per degree from the scattering kernel change are added to the values calculated with the density change. The MTC results are similar for the HEU and LEU cores.

According to PSAR Section 4.5.2.5 the methodology for the void reactivity coefficient analysis is similar to the methodology for calculating the MTC. The region was first voided, the reactivity change (k/k) calculated, and the reactivity change is then divided by the volume of the void. The results demonstrate that a void forming anywhere within the NBSR active core will result in negative reactivity feedback. The magnitude of the feedback is similar for the HEU and LEU cores.

RAI-1S1 requests information to establish a basis for the acceptability of temperature coefficient calculations. The response by the licensee demonstrates their acceptability by comparing measurements and calculations of the isothermal temperature coefficient. The measured coefficient was determined in 1969 to be -(0.023-0.035) %k/k/°C and the equivalent calculated value is -0.031 %k/k/°C for the current HEU core. The ERI staff finds that this level of agreement indicates that the MCNPX model employed is suitable for the calculation of temperature-based coefficients. Based upon the information provided by the licensee in the response, the ERI staff finds that RAI-1S1 is acceptably resolved regarding temperature-based coefficients.

The calculated coefficient values for the HEU and LEU cores are reported in the PSAR and reproduced in Table 2-2.

2.5.5 Burnup Effects on Analyses Fuel depletion, or burnup, typically has four primary effects on reactor core operation. It:

(1) introduces spatial differences in fissile material distribution, (2) creates a distribution of fission products that reflect the power distribution, (3) alters fuel thermal conductivity, and (4) changes core reactivity.

According to PSAR Section 4.5.1.2 the depletion of the NBSR fuel is modeled by using the MCNPX code which utilizes an explicit treatment for burnup. However, the model as described uses very few (60) unique materials to track the depletion of 1020 plates in the core. In the Brown document provided in the response to RAI-15S1, the authors cite the significant improvement in the results accomplished by increasing the number of materials from 30 to 60. However, it is unclear whether the use of 60 materials is suitable. The responses to RAI-2S1 and 3S2 confirm the use of 60 materials, but do not provide a defendable basis for the adequacy of this number.

This number of materials is a very small percentage of the number used in that single assembly sensitivity study in the Brown report. The PSAR and RAI responses do not provide a justification for the number of materials used, nor is any methodology indicated for accounting for the cited effects via biases or uncertainties. The ERI staff finds that the burnup effects and their implications for the determination of the conditions for the conservative determination of thermal power may not be suitably conservative. (COI-5) 17

ERI/NRC 18-204 2.5.6 Power Distribution PSAR Section 4.5.3 describes the power distributions that the MCNPX model determines are applicable to both HEU and LEU cores. The radial power distributions demonstrate that some modest differences exist between them. As shown in PSAR Table 4.13, the maximum half-element power at BOC increases from 427 kW to 449 kW when going from HEU to LEU fuel. At EOC there is a decrease in the maximum half-element power.

There are two plena in the NBSR dividing the coolant flow between the six innermost fuel elements and the other 24 fuel elements. The analysis shows that there is an 8.4% increase in the power in the innermost six fuel elements at BOC when going from HEU to LEU fuel and at EOC there is an 11% increase, though the total power generated by the inner six FEs is smaller at EOC than at BOC. This indicates that when converting from HEU to LEU fuel, there is a net power shift from the peripheral regions of the core towards the inner portion of the core.

The licensee states that the power shift is due primarily to the larger loading of U-238 in the LEU core relative to the HEU core. It acts as an absorber, reducing the leakage out of the core. The consequence of the reduced leakage is to raise power in the central region.

The licensee also states that at BOC, there is more power generated in the lower half of the core than there is in the upper half of the core. This is due to the shim safety arms suppressing the power in the upper half of the core. Because of this the burnup is initially reduced in the upper half of the core. When the shim safety arms are swung out of the core as EOC is approached, the power is shifted to the upper half of the core. The responses to RAI-5S1 and RAI-5S2 provide statements in response to this review. However, the analysis presented earlier in Sections 2.5.1, 2.5.5, and 2.5.6 raise some questions regarding whether the determination of peaking factors and the associated temperatures are sufficiently conservative.

As discussed previously, the PSAR and RAI responses do not provide a justification for the determination that the limiting power distribution for NBSR HEU or LEU cores is established by the descriptions in the PSAR. The ERI staff finds that the power distribution determined for the HEU and LEU cores is not clearly demonstrated to be limiting and thus the power distribution determined in the PSAR may not be suitably conservative. For this reason, the ERI staff finds that the responses to RAI-5S1 and RAI-5S2 do not resolve the information requirement allowing the formation of a basis of acceptability. (COI-5) 2.5.7 Conclusions The analysis submitted by the licensee in the PSAR indicates that the key neutronic characteristics of the NBSR core (i.e., excess reactivity, control element worths, shutdown margin, reactivity coefficients, and other dynamic parameters) will not change significantly due to the conversion from HEU to LEU fuel. The operational characteristics will be analyzed using computer codes that have been previously considered and found to be acceptable. However, the PSAR does not provide sufficient assurance that the manner in which these codes are used to model the NBSR are sufficiently conservative so as to bound the expected operation of the facility.

Even including consideration of RAI responses, the ERI staff is unable to form the basis of a 18

ERI/NRC 18-204 finding that the power distributions calculated are acceptable to support the thermal-hydraulic analysis objectives of this review. (COI-5) 2.6 Thermal-Hydraulic (T&H) Design According to PSAR 4.6.1 the design basis of the thermal-hydraulic design of the NBSR is that there shall be no fuel damage during normal operation and no fuel damage resulting in release of fission products from any credible accident. For normal operating conditions, the criterion chosen is that the heat transfer to the primary coolant shall not exceed critical heat flux (CHF), including any instability; the latter being defined by onset of flow instability (OFI). This will preclude blistering and the potential for fuel damage. The temperature at which blistering might occur is the Safety Limit in the Technical Specifications and hence, also, a criterion for fuel damage.

The cooling system of NBSR is unchanged by the fuel conversion. According to the description in NBSR-14 there are four primary circuit cooling loops, each having a single centrifugal pump (DP-1, DP-2, DP-3, and DP-4) that are connected to common input and output headers. Typically, only three of these loops are operational while at full power.

2.6.1 Calculation Methodology In PSAR Section 4.6.2.4 the licensee states that the thermal-limit analysis is performed using RELAP5 (Ref. 16) with the power distribution obtained from the MCNPX analysis discussed previously. This code has been used extensively for reactor analysis, including previous analysis of NBSR, and its use is appropriate to support the HEU to LEU conversion.

RAI-3S1 requests information justifying the use of the Sudo-Kaminaga correlation as discussed in PSAR 4.6.2.1 for the calculation of the critical heat flux (CHF) and the Saha-Zuber correlation for the calculation of the onset of vapor generation, which according to PSAR Section 4.6.2.2 is used by the licensee as an indicator for the onset of flow instability. In the response (Ref. 10) the licensee provides a discussion of the basis for using the Sudo-Kaminaga correlation for the determination of critical heat flux and Saha-Zuber correlation for the determination of net vapor generation. The response also provides additional details regarding the use of these correlations to calculate the critical heat flux ratio (CHFR) and the experimental basis for their acceptability.

Based upon the information provided by the licensee in the response, the ERI staff finds that RAI-3S1 is acceptably resolved.

RAI-4S2 asks for clarification of the applicable departure from nucleate boiling ration (DNBR) limit and values used by the licensee to characterize NBSR performance. In their response, the licensee explains that they began using a statistical approach to the determination of CHFR with the submission of its request for license renewal in 2004. In addition, they chose to include the analysis of the minimum onset of flow instability ratio as added assurance of thermal margin.

However, the licensee explains in the response that the resulting CHFR remains greater than 2.00 for all reactivity insertion and loss of flow events.

The ERI staff finds that the calculational framework and correlations used to perform the thermal-hydraulic analyses are appropriate for the application to NBSR. Based upon the information provided by the licensee in the response, the ERI staff finds that RAI-4S2 is acceptably resolved.

19

ERI/NRC 18-204 2.6.2 Thermal-Hydraulic Results The following statements are extracted from PSAR Section 4.6:

The power distribution in the fission plates is assumed to be given by the fission density as calculated by the computer code MCNPX. This is a conservative assumption, as 14% of the energy is in the form of y-rays and neutrons, and will be deposited much more uniformly throughout the core.

The ERI staff agrees with this statement.

Another conservatism is the fact that burnup is assumed to be uniform over each half-element. In reality the distribution of burnup in a half-element is roughly proportional to power density and this tends to lower high power densities. This does not apply to fresh fuel but, as will be discussed below, the highest powers are in burned fuel elements.

Although the text states that fresh fuel will be discussed below the reviewers find no additional discussion material regarding fresh fuel. In addition, the position of the licensee that uniform burnup is a reasonable assumption is not demonstrated.

With four new fuel elements, criticality occurs when the shim safety arms are inserted furthest into the core. This insertion results in flux compression into the bottom half of the fuel elements.

For the purposes of thermal limit analysis, the SU hot-spot cases for the inner and outer plenums are listed in Table 4.23 and the SU hot-stripe cases for the inner and outer plenums are listed in Table 4.24. The values reported here are derived from the detailed power distribution calculation in MCNPX, with the exception of the peak heat fluxes, which are calculated using RELAP5.

The ERI staff finds that it is unclear whether the limiting power distribution is used for the RELAP5 model; thus, it is not possible to determine the acceptability of the RELAP5 model under the stated conditions. (COI-6) 2.6.3 Conclusions The ERI staff concludes that the thermal-hydraulic analyses for the NBSR conversion were performed using qualified methods. The applicability of the analytic methodology is confirmed by the extensive validation effort that has been undertaken over the course of the RELAP5 computer code development effort. However, questions remain about the suitability of the power distribution used to characterize the thermal limits for NBSR under normal conditions. The ERI staff is unable to form the basis of a finding that the thermal-hydraulic conditions calculated are acceptable to support the objectives of this review.

20

ERI/NRC 18-204 2.7 Accident Analysis 2.7.1 Introduction PSAR Chapter 13 presents the results of analyses describing a comprehensive series of accidents and transients (events). To support these analyses neutronics models are used by the licensee to calculate physics parameters for both the HEU- and LEU-fueled cores as discussed in PSAR Chapter 4. These models provide the BOC and EOC power distributions, neutron kinetics parameters, and the reactivity worths of the regulating rod and the shim safety arms.

As discussed in PSAR Section 13.1, the event analyses are described by the licensee as generally employing conservative, but credible assumptions that are expected to result in the most severe consequences that will bound all events allowed by technical specifications. The progression of each event is analyzed by the licensee to the extent necessary to determine the type of hazard the event presents. Generally, this is either a challenge to the fuel safety limit, a radiological limit, or both. The flow blockage in one fuel element event is not considered credible by the licensee and is selected by the licensee to be the maximum hypothetical accident. The results of the events with the LEU fuel are compared to those of the HEU fuel in the PSAR to illustrate the suitability of the LEU fuel design to support NBSR fuel conversion.

Furthermore, the licensee states that events are analyzed at two points in the fuel cycle: the beginning of a cycle (BOC) before equilibrium xenon has built into the core and end-of-cycle (EOC) when the fission product inventory is maximized, and the fuel is heavily depleted. The licensee states that at BOC, there are typically four fresh fuel elements inserted into the core.

Some short-lived fission product poisons such as Xe-135 have decayed away during the refueling period and others, such as Sm-149 have peaked. The licensee states that power peaking is highest at BOC making it the limiting condition for some events. Other events are most limiting at EOC because the available negative reactivity for shutdown is lowest when the shim safety arms are deeply inserted. The events are then analyzed for both all HEU and all LEU cores providing 4 Scenarios for each event analyzed:

• HEU at BOC

• HEU at EOC

• LEU at BOC

• LEU at EOC As stated in PSAR Section 13.1, the acceptance criterion used by the licensee for all analyzed events is that there may be no loss of fuel integrity and the dose to occupational workers and the public are limited by 10 CFR Part 20. The fuel blister temperature currently used as the HEU safety limit (SL), 450°C, is used as a conservative surrogate to preclude the release of fission products and to act as the acceptance criterion for HEU events.

As discussed in PSAR Section 13.1, the information available regarding the blister temperature of LEU fuel is still being interpreted. The blister threshold has been determined experimentally as a function of fission density, but many more tests are yet to be completed. For the LEU fuel, the maximum fission density is conservatively estimated to be 7.2x1021 fissions/cm3. The licensee 21

ERI/NRC 18-204 states that the peak fission density typically occurs at the bottom of upper section fuel plates near the midplane gap at EOC. The isothermal blister threshold based on the experimental data currently available is 435°C at this fission density.

2.7.2 Methodology for Accident and Transient Events The analysis methodology to support accident and transient analysis is described in PSAR Sections 13.2 and 13.3. Significant portions of the event analyses are performed using three codes:

• RELAP-5 MOD3.3 simulates a wide variety of thermal-hydraulic transients in nuclear systems involving mixtures of steam, water, non-condensable gases, and solute in a fueled vessel. RELAP5 is one of the most widely used system codes for analyzing power and research reactor accidents and transients. The Advanced Test Reactor, the High Flux Isotope Reactor, the High Flux Beam Reactor have all been analyzed successfully with RELAP5 all of which used fuel similar in configuration to the HEU fuel in NBSR. The RELAP5/MOD3.3 model of the NBSR simulates the transport of heat and coolant in the primary system. A schematic diagram showing the main components of the NBSR primary system is shown in PSAR Figure 13.1 (see also PSAR Figure 4.5). PSAR Figure 13.2 shows the corresponding nodal diagram for the RELAP5 model. The reactor vessel is divided into several interconnected hydrodynamic volumes and heat structures with internal heat generation used to model the fuel plates.

• The TRACE computer code is used to calculate the water level inside the reactor vessel after a LOCA. The model is based on the RELAP5 model developed for other safety analyses as described above. It is converted into a TRACE model because of the ability of TRACE to predict water level accurately at atmospheric pressure with non-condensable gases present. The TRACE model is used to determine the draining of the pool for LOCA which is an aspect of the event that is independent of whether the fuel is HEU or LEU.

• The fuel plate temperature following a LOCA is calculated using the HEATING-7.3 code.

It provides peak fuel temperature spatially and as well as over time. HEATING-7.3 solves steady-state and/or transient heat conduction problems in one-, two-, or three-dimensional Cartesian, cylindrical, or spherical coordinates. The modeling for HEU and LEU fuel is almost identical except for the different dimensions of the fuel meat and cladding.

Further details of the methodology are provided in PSAR Sections 13.2 and 13.3. Figure 2-5 and Figure 2-6 illustrate the major features of the respective RELAP and TRACE models.

The ERI staff find the use of RELAP5 and TRACE for the purposes stated in the PSAR appropriate and acceptable. Similarly, the use of HEATING7.3 computer code is appropriate and acceptable for the purposes stated in the PSAR.

22

ERI/NRC 18-204 Figure 2-5 RELAP5 Model Figure 2-6 TRACE Model 23

ERI/NRC 18-204 RAI-7S1 questions combining the coolant loops into a single flow path in the NBSR RELAP model in representing multi-loop effects conservatively.

In response (Ref. 10) the licensee explains how they modeled multiple cooling loops using a single flow path and states that analysis is still able to simulate single component failures.

However, it is unclear that this modeling approach would allow the licensee to determine the behavior of the system in a suitably conservative manner when single failures are simulated. It is more appropriate to model the actual piping arrangement, valve alignment, and instrument response rather than using the cited approximations in the response. The ERI staff suggests that the analysis models be modified to represent the NBSR system with greater fidelity.

RAI-6cS2 raises the issue of how single failure analysis (SFA) is applied to the analysis of NBSR, how the SFAs are determined, and what effect they have on the consequences. In response (Ref. 10) the licensee states:

NUREG 1537 does not provide any guidance on assuming a limiting single failure.

However, Chapter 13 in the PSAR does discuss the many conservative assumptions used in the analysis.

The ERI staff finds that this is not an accurate statement. The guidance to consider single failures for normal operation, transient events, and accidents is discussed throughout NUREG-1537 Part

1. It is discussed twice in Section 3.1, once in Section 4.2.2, once in Chapter 6, twice in Section 7.2.1, once in Section 7.2.4, once in Section 7.4, once in Section 7.5, and once in Section 13.2(3).

Furthermore, it is inappropriate to substitute many conservative assumptions used in the analysis for the clear guidance provided to perform actual SFAs. The ERI staff finds that the analyses provided in the PSAR and supplemented by the RAI responses, specifically RAI-6cS2, does not suitably demonstrate the ability to accommodate single failures and the licensees position on this requirement in the review process prevents the ERI staff from forming the basis of a finding of acceptability for the supplied safety analysis. (GOI-1) 2.7.3 Reactivity Insertion Events Reactivity Insertion events are described in PSAR Section 13.4.

Startup Event (Uncontrolled Shim Safety Arm Withdrawal) - Licensee The analysis of a startup event is described by the licensee in PSAR Section 13.4.2 where it assumes that the reactor is initially critical at a power level of 100 W. The licensee postulates that the initiating event is an operator error whereby the shim safety arms are withdrawn until the reactor is scrammed by one of the high-power level trips.

RAI-8S1 requests clarification of the conditions of this analysis to understand whether the transient analyzed is sufficiently conservative to bound an actual event of this type. The RAI also requests information pertaining to the maximum blade withdrawal rate utilized, the worth of the maximum worth blade, whether any uncertainties are utilized, and assurance that the supplied analysis models this event in a suitably conservative manner.

24

ERI/NRC 18-204 In response (Ref. 10) the licensee explains that the event is modeled using a reactivity insertion rate for the shim safety arm withdrawal equal to 5x10-4 k/k/s. This rate is greater than the maximum measured and calculated rate at any shim safety arm initial position and greater than the rate that moving the regulating rod would produce. The scram setpoint employed (130%) is greater than the limiting safety system setting (LSSS) (125%) to be conservative. The analysis does not consider any fuel or moderator feedback, which is also conservative. After the reactor scram, all shim safety arms are assumed to insert fully using the TS value for the blade drop time.

Based upon the information provided by the licensee in the response, the ERI staff finds that RAI-8S1 is acceptably resolved.

The results shown in Figure 2-7 demonstrate the sensitivity of the fuel cladding temperature to fuel type. The power increase starts earlier with the LEU fuel than with the HEU fuel because of the smaller delayed neutron fraction and shorter neutron generation time with the LEU fuel. The ERI staff also observe that the power rises faster at BOC than at EOC which is the result of the difference in neutron generation time and the shim safety arm position at BOC. Similarly, Figure 2-8 shows the effect that the increase in reactor power has on the reduction of the CHFR.

The previously stated position of the ERI staff concerning the lack of an appropriate SFA prevents the formation of a basis for the endorsement of this otherwise acceptable analysis. (GOI-1)

Maximum Reactivity Insertion Event - Licensee The maximum reactivity insertion event is analyzed by the licensee in PSAR Section 13.4.3 using the RELAP5 computer code employing NBSR developed point kinetics parameters discussed in PSAR Section 13.2.4. For conservatism, the calculation does not consider any fuel or moderator reactivity feedback. For this event, a ramp reactivity insertion of 0.500 %k/k is assumed to occur in 0.5 s. This amount of reactivity is the Technical Specification limit for the reactivity of any experiment. The power increases until reaching the high-power scram setting after which the shim safety arms are inserted.

25

ERI/NRC 18-204 Figure 2-7 Startup Event Cladding Temperature Figure 2-8 Startup Event CHFR Figure 2-9 displays the calculated clad temperatures for the event analyzed for the 4 Scenarios.

The temperature reported is for the fuel element nodes where the highest value is predicted. The clad temperatures increase rapidly after the reactivity insertion. After reaching peak values between 0.42 s and 0.51 s, they decrease rapidly due to the insertion of the shim safety arms after the reactor scram signal. The peak clad temperatures range from 124°C to 128°C, which is 26

ERI/NRC 18-204 well below the blister temperature, and corresponds to temperature increases of approximately 26°C to 29°C.

Figure 2-9 Maximum Reactivity Insertion Event Cladding Temperature The CHFR is evaluated using the Sudo-Kaminaga correlation and the value as a function of time is shown in Figure 2-10. The hydraulic nodes used for the figures are where minimum CHFR takes place at each timestep. The CHFRs reach minimum values between 0.40 s and 0.48 s.

Then they increase very rapidly and become larger than 37 from 1.0 s in all cases. The calculated minimum CHFR is greater than 2.00 for any of the 4 scenarios.

27

ERI/NRC 18-204 Figure 2-10 Maximum Reactivity Insertion Event CHFR The onset of flow instability ratio (OFIR) is evaluated using the Saha-Zuber criteria and the behaviors for the 4 scenarios are shown in Figure 2-11. The nodes in the figures are the ones where minimum OFIR occurs among all hydraulic nodes in the core region. The OFIRs reach minimum values between 0.40 s and 0.49 s. Then they increase very rapidly and become larger than 50 from 1.0 s in all cases. The minimum OFIRs are shown in PSAR Table 13.8 along with the corresponding times and hydraulic node number. The minimum OFIR values are similar in all cases. It is observed from PSAR Figure 13.22 and Table 13.8 that the minimum OFIRs are all much larger than 3.19.

28

ERI/NRC 18-204 Figure 2-11 Maximum Reactivity Insertion Event OFIR The response to RAI-6bS2 illustrates the use of the statistical methodology for the determination of CHFR. However, the results from PSAR Table 13.7 show that the calculated CHFR is greater than 2.00 for HEU or LEU at BOC and EOC conditions at all times during the event. Furthermore, as stated by the licensee in the response to RAI-8S1, the conditions used for this event include the assumption of a reactor scram setpoint used is 130% of the licensed NBSR power instead of the TS 3.1.1 value of 125% making the results even more conservative than stated.

Based upon the information provided by the licensee in the response, the ERI staff finds that the CHFR values cited by the licensee satisfy the guidance without using the statistical technique discussed. The ERI staff takes no position on the acceptability of using the statistical approach since it has not been discussed in sufficient detail in the application. RAI-6bS2 is acceptably resolved.

The previously stated position of the ERI staff concerning the lack of an appropriate SFA prevents the formation of a basis for the endorsement of this otherwise acceptable analysis. (GOI-1)

Confirmatory Analysis of Reactivity Events The ERI staff analysis of reactivity-based events uses a TRACE model (Ref. 17). This model was developed by ERI using the RELAP5 input that was obtained from the licensee (Ref. 10). The NIST RELAP model was then reviewed, adjusted, and validated by the ERI staff. That model was then converted to a TRACE model for subsequent analysis. The TRACE model is illustrated in Figure 2-12. This model uses 7 groups of assemblies to model the 30-assembly core. The ERI staff finds that the upper and lower fuel regions are properly represented.

29

ERI/NRC 18-204 Figure 2-12 TRACE Model for Confirmatory Analysis Startup Event (Uncontrolled Shim Safety Arm Withdrawal) - Confirmatory Analysis The ERI staff analysis of the uncontrolled shim safety arm removal is compared with the licensees equivalent results. The event occurs when the shim blades are all withdrawn simultaneously at 0.50 %k/k/s until the high-power-level scram setpoint is reached at 26 MW (130% of full power) per TS 2.2(2). After passage of the shim blade insertion delay time of 0.0983 s the shim safety arms then fully inserted. The maximum time step of 0.0005 s is used for the TRACE calculation.

The ERI staff analysis assumes that the limiting single failure for this event is the failure of one of the two high power scrams and the startup rate scrams are also not credited.

The results of these calculations are compared to those presented in the PSAR and both are shown in Table 2-5, Figure 2-13, and Figure 2-14. The temperatures reported are for the fuel element nodes where the highest value is predicted. The temperatures increase rapidly after the reactivity insertion. After reaching peak values, they decrease rapidly due to the insertion of the 30

ERI/NRC 18-204 shim safety arms after the reactor scram signal. The resulting temperatures are well below the fuel blister temperature. In both models no reactivity feedback is credited, thus making the results additionally conservative with respect to the expected behavior.

Table 2-5 TRACE and RELAP5 Uncontrolled Shim Safety Arm Withdrawal Results Parameters TRACE LEU from RELAP in PSAR Reactor scram at 26 MW 15.05 s 15.7 s Peak power 36.7 MW 38.5 MW Tmax 124.°C 129.°C Minimum DNBR 2.34 4 / 2.38 5 2.19 6 Figure 2-13 Power Response to Uncontrolled Shim Safety Arm Withdrawal 4 Critical heat flux determined using the AECL-IPPE correlation within TRACE.

5 Critical heat flux determined using the Sudo-Kaminaga correlation external to TRACE.

6 Critical heat flux determined using the Sudo-Kaminaga correlation.

31

ERI/NRC 18-204 Figure 2-14 Fuel Temperature Response to Uncontrolled Shim Safety Arm Withdrawal The ERI staff finds that the results for the uncontrolled shim safety arm withdrawal event are confirmed.

Maximum Reactivity Insertion Event - Confirmatory Analysis The ERI staff analysis of the withdrawal of a 0.500 %k/k maximum reactivity event results in a power increase from that reactivity change. In this scenario, the event occurs as the worth of the experiment is inserted instantaneously. The shim safety arms undergo an insertion delay time of 0.0983 sec as in the NBSR RELAP5 model. The ERI staff analysis assumes that the limiting single failure for this event is the failure of one of the two high power scrams. The event terminates on a high-power-level scram at 26 MW (130% of full power) per TS 2.2(2), and the fuel temperature remains substantially below the SL, TS 2.1, fuel temperature of 450°C. In addition, the DNBR, remains above 2.0 at all times. As illustrated in Table 2-6, Figure 2-15, and Figure 2-16 the TRACE and RELAP results are reasonably similar in terms of timing and consequences.

In both models no reactivity feedback is credited, thus making the results additionally conservative with respect to operational expected-behavior.

32

ERI/NRC 18-204 Table 2-6 Maximum Reactivity Event Results Parameters TRACE LEU from RELAP in PSAR Reactor scram at 26 MW 0.25 s 0.26 s Peak power occurs @ 0.34 s 0.47 s Peak power 31.4 MW 32.3 MW Tmax 131.°C 127.°C Figure 2-15 Power Response to Maximum Reactivity Event 33

ERI/NRC 18-204 Figure 2-16 Temperature Response to Maximum Reactivity Event The ERI staff finds that the analysis results for the Maximum Reactivity Event are confirmed.

2.7.4 Loss-of-Flow Events The loss-of-flow events are discussed in PSAR Section 13.5.1. RAI-9S1 and RAI-10S1 request clarification on whether the LOFA analysis is suitably conservative given the concern over the ability to model backflow after a pump trip. In response (Ref. 10) the licensee explains that in the response to RAI-7 the licensee clarified how backflow from a damaged pump is prevented using the attending control valve, preventing backflow through that pump. However, this explanation does not fully address the ERI staff concern, which is that the remaining pumps are expected to continue to operate. Using SFA, failure of the control valve to actuate allows backflow through that pump, thus reducing flow to the inner and outer plenums. Having failed to consider this failure mode means that the analysis of this event is not sufficiently conservative. For this reason, the ERI staff finds that the responses to RAI-9S1 and RAI-10S1 do not resolve the information requirement allowing the formation of a basis of acceptability. (GOI-1) 34

ERI/NRC 18-204 Loss-of-Flow due to Loss of Offsite Power In this event the licensee assumes that all three primary pumps trip upon loss of offsite power (LOSP). The licensee states that this bounds any loss-of-normal-power event. As the three primary coolant pumps coast down the primary coolant flow drops to a value where one or more of the primary coolant flow monitors generates a scram signal. The scram is assumed to occur 0.4 s after the coolant flow has reached the trip value considering instrumentation delays. The primary pump discharge valves start closing at 1.0 s on the primary pump trip signal. The stroke time of the valves is assumed to be 2.0 s. The shutdown pumps (SDP) operate on both alternating and direct current power and they continue to operate normally during the LOSP. In the present analysis only one of the two SDPs are assumed to operate to remove the decay power. The valves at the outlets of the SDPs begin to open at 0.7 s with the stroke time of 1.5 s. The normal flowrate of one SDP is 800 gpm (~55 kg/s).

Figure 2-17 Cladding Temperature after LOSP Figure 2-17 shows the cladding temperature behavior for this event in the fuel element nodes corresponding to the RELAP5 prediction of the highest cladding temperature. The temperatures increase due to reduced heat transfer as fluid velocity decreases after the primary pumps trip.

The temperature reaches a maximum value at 1.41 s with both HEU and LEU fuels at BOC and at 1.45 s with both HEU and LEU fuels at EOC and then begins decreasing rapidly because of decreasing reactor power. The peak cladding temperatures range from 118°C to 124°C which is significantly below the fuel blister temperature.

Loss-of-Flow due to Seizure of One Primary Pump This event is described in PSAR Section 13.5.2 where it is assumed that through some failure, such as a faulty bearing, the rotor of one pump suddenly locks. Because of fluid momentum, 35

ERI/NRC 18-204 coolant flow through the primary loop will decrease over a finite time interval until a one-third flow reduction is achieved. Since the RELAP5 model lumps all three pumps into one effective pump, the seizure of one of the pumps is modeled by an instantaneous step reduction in the pump speed to two-thirds of full speed. This licensee states that is conservative since the flow with only two pumps operating would be more than two-thirds of full flow. The responses to RAI-7S1 and RAI-9S1 (Ref. 10) discuss this in some detail.

Figure 2-18 Cladding Temperature After Pump Seizure Figure 2-18 illustrates the change in cladding temperatures in response to the pump seizure (locked rotor) event displaying the results for the fuel element nodes corresponding to the highest cladding temperature. The cladding temperatures initially increase due to heat transfer becoming less efficient as the mass flow rate decreases after pump seizure. The temperature reaches a maximum shortly after reactor scram and then decreases rapidly because of the reduction in reactor power. The peak cladding temperatures range from 114°C to 121°C which are much lower than the fuel cladding blister temperature.

The previously stated position of the ERI staff concerning the lack of an appropriate SFA prevents the formation of a basis for the endorsement of this otherwise acceptable analysis. (GOI-1)

Loss-of-Flow due to Throttling of Coolant Flow to Outer Plenum This event is described in PSAR Section 13.5.3 where the licensee assumes that the flow control valve DWV-1 fails by closing in 60 s, thus reducing the flow through the outer plenum and generating a reactor scram signal 0.4 s after the flow reaches the low flow trip point of 4,700 gpm.

The complete closure of the flow control valve isolates the lower plenum of the outer core and cuts off the supply of forced coolant. The results of RELAP5 simulation shows that since all 36

ERI/NRC 18-204 coolant channels in the fuel elements in the outer core share the same inlet and outlet plena, closed loop recirculation flow paths are established between the hotter and cooler coolant channels in the outer core. Buoyancy induces upflow through the hotter coolant channels, while downflow through the cooler channels completes the closed flow loop. The recirculation flow removes heat from the fuel elements by natural convection. The response to RAI-7S1 and RAI-10S1 (Ref. 10) provide an acceptable justification for using a single loop to model multi-loop effects when not subject to SFA (e.g., normal operation).

Figure 2-19 Cladding Temperature after Valve Closure Figure 2-19 shows that the cladding temperature at BOC with LEU fuel is much lower than in the other cases after ~110 s. This is the result of relatively large negative flow through the limiting CHFR channel compared to the other three cases. The mass flow rate in the limiting CHFR channel at 110 s is about -0.03 kg/s at BOC with LEU fuel while it is around 0.02 kg/s in the other cases, which enhances heat transfer from the fuel to the coolant in the former case. It also shows that the cladding temperature at BOC with HEU fuel is slightly higher than at EOC with both HEU and LEU fuels. This is because the power fraction is higher at BOC than at EOC in the hottest cell. The peak cladding temperatures vary from 116°C to 118°C which are much lower than the fuel cladding blister temperature.

The previously stated position of the ERI staff concerning the lack of an appropriate SFA prevents the formation of a basis for the endorsement of this otherwise acceptable analysis. (GOI-1)

Loss-of-Flow due to Throttling of Coolant Flow to Inner Plenum This event is described in PSAR Section 13.5.4 where the licensee assumes that the flow control valve DWV-2 closes, decreasing the flow through the inner plenum and generating a reactor 37

ERI/NRC 18-204 scram signal 0.4 s after the flow reaches the low flow trip point of 1,200 gpm. The 8-inch flow control valve has a stroke time of 30 s. The complete closure of the flow control valve isolates the lower plenum of the inner core and at the same time cuts off the supply of forced coolant flow.

The results of RELAP5 simulation shows that since all coolant channels in the fuel elements in the inner core share the same inlet and outlet plena, closed loop recirculation flow paths are established between hotter and cooler coolant channels in the inner core. Buoyancy induces upflow through the hotter coolant channels, while downflow through the cooler channels completes the closed flow circulation loop. The recirculation flow removes heat from the fuel elements by natural convection.

Figure 2-20 Cladding Temperatures after Flow Control Valve Closure As illustrated in Figure 2-20, the general behavior of the cladding temperature for LEU fuel is very similar to that of the HEU fuel. The cladding temperature starts increasing due to reduced heat transfer as the mass flow rate to the inner core decreases. The temperature reaches a first peak shortly after reactor scram and then begins decreasing rapidly because of reduction of the reactor power. The temperature starts increasing again from around 17 s as the mass flow rate decreases further and heat transfer from the fuel elements to the coolant becomes less efficient.

The cladding temperatures show some oscillatory behavior from around 30 s to 100 s because of fluctuations of the mass flow rates in these channels. As the valve at the inlet to the inner plenum is closed, the coolant flow velocity decreases very rapidly, fluctuates around zero, and then stabilizes to natural circulation inside the inner core. During these flow fluctuations, RELAP5 predicts almost zero flow velocity a few times when flow direction changes. This results in poor heat transfer from the fuel to the coolant and increased cladding temperatures. As flow velocity increases after changes in flow direction, the heat transfer improves again and the cladding temperature decreases. This behavior occurs a few times in all cases causing oscillations in the 38

ERI/NRC 18-204 cladding temperature. The figure also shows that the cladding temperatures at BOC with both HEU and LEU fuel are much lower than those in the other cases, after ~60 s. This is caused by relatively large negative flow through those channels while positive flow is predicted in the other two channels at EOC with the HEU and LEU fuels. The peak clad temperatures vary from 115°C to 118°C which is much lower than the blister temperature.

The previously stated position of the ERI staff concerning the lack of an appropriate SFA prevents the formation of a basis for the endorsement of this otherwise acceptable analysis. (GOI-1)

Loss-of-Flow due to Closure of Valve DWV-19 This event is described by the licensee in PSAR Section 13.5.5. Therein, they state that valve DWV-19 is a motorized 18-inch butterfly valve, mounted in the primary coolant outlet line. Even though this valve is only used during maintenance when the reactor is shut down, they assume that it could receive a spurious signal while the reactor is operating at full power and close resulting in a loss of primary flow. According to the PSAR this valve is not shown explicitly in the RELAP5 nodal diagram in Figure 13.2 where it would be within node 10. However, it was added to the RELAP5 model by the licensee in order to simulate this event and it is shown in the corresponding TRACE nodalization diagram in Figure 2-6 as valve 102. The valve has a measured stroke time, fully open to fully closed, of 21 s. In the future, this valve may be replaced with one having a closing time of 120 s. Since the shorter closing time results in more significant consequences, it is the only one modeled.

According to the PSAR, the analysis is done only for the HEU core. Because of the low probability of this event, the large safety margin, and the similarity in the response for both the HEU and LEU fueled cores in all of the above analyzed loss-of-flow events, it is not necessary to separately analyze the LEU core. The ERI staff agree with the position that the fuel similarity allows the HEU core results to justify the LEU behavior also.

39

ERI/NRC 18-204 Figure 2-21 Cladding Temperature after DMW-19 Closure Figure 2-21 shows cladding temperatures from this event in the fuel element nodes corresponding to the channel cell where the highest peak cladding temperature. The temperature increases slightly and reaches 102.1°C at 9.3 s as the coolant mass flowrate decreases. It decreases significantly after reactor scram. The temperature starts increasing slowly again from around 11 s as the coolant flowrate to the core becomes smaller.

Several postulated loss-of-flow events are simulated by the licensee. The simulations involve both forced flow and natural circulation flow cooling in the core. Two limiting state-points in a fuel cycle have been considered, namely, BOC and EOC. Primary system flow, reactor power and peak clad temperature have been examined in detail along with minimum CHFR and OFIR. In all cases, the system behavior is generally similar for either HEU or LEU fuel. The RELAP5 predictions follow expected behavior given the scenarios analyzed. The calculated total system flowrate is smaller than the measured data in the case with loss of offsite power leading to more conservative results. The reactor power starts decreasing very rapidly from full power after reactor scram due to low flow. For the events with loss-of-offsite-power and throttling of coolant flow to the outer plenum, the scram signal is a result of low flow to the outer plenum. For the event with throttling of coolant flow to the inner plenum, the scram signal is a result of low flow to the inner plenum. Cladding temperature rises from time zero because the mass flow rate decreases in the core. All events are terminated by a reactor scram and there is no indication of any challenge to the SL. The reactor remains flooded with water for all events with a loss of forced circulation. The minimal natural circulation that occurs is sufficient to prevent overheating of the fuel.

40

ERI/NRC 18-204 The previously stated position of the ERI staff concerning the lack of an appropriate SFA prevents the formation of a basis for the endorsement of this otherwise acceptable analysis. (GOI-1) 2.7.5 Loss-of-Coolant Accident (LOCA)

The ERI staff has reviewed the supplied LOCA analysis. Although one may not expect this given the stated purpose of the PSAR - to consider the changes required to support the change in the fuel design - this extensive review is undertaken because of the fundamental differences between the previously supplied analysis from 2004 in NBSR-14 and the analysis submitted in 2017 in the PSAR. The analysis presented in 2004 is 11/2 pages in length and largely deals with radiological consequences, whereas the PSAR version is 21 pages, of which 1 page is radiological in nature.

The basis of the LOCA analysis in the PSAR is provided in the Baek 2014 report (Ref. 18). In that report other events and accidents are readily identifiable by common names (e.g., Loss of Offsite Power, Reactivity Insertion Accident, etc.), the LOCA Accident is euphemistically identified as Heat Removal by Flow from the Inner Emergency Cooling Tank. The text of that document states in Section 5.6 that coolant will drain from the interior of the fuel elements.

However, in 2017 a paper was published that in part discussed the history of the NBSR LOCA analysis (Ref. 19). That paper states:

The SAR also refers to an analysis 7 of why there is sufficient water without forced cooling to cool the fuel elements (FEs) in a LOCA if the flow channels between two fuel plates (or a fuel plate and an outside plate) remain filled with water.

A more recent study 8 showed that some scenarios may lead to the draining of water from the coolant channels The ERI staff review of NBSR-14 finds that it does not clearly state what the condition of the fuel is post LOCA blowdown - covered or uncovered with D2O.

Discussions with the NBSR staff (Ref. 20) clarified this issue by informing the ERI and NRC staff that the NBSR-14 analysis of the limiting LOCA concluded that in that previous analysis, subsequent to a break in the inlet pipe the fuel channels remained flooded and were provided supplemental cooling by the inner reactor tank (IRT) flow to compensate for losses due to decay heat. The emergency cooling tank (ECT) provided makeup to the IRT, and sump recirculation 7 L.-Y. CHENG et al., Physics and Safety Analysis for the NIST Research Reactor, BNL-NIST-0803, Rev.

1, Brookhaven National Laboratory (2004).

8 J. S. BAEK, L.-Y. CHENG, and D. J. DIAMOND, Analysis of LOCA Scenarios in the NIST Research Reactor Before and After Fuel Conversion, Proceedings of Conference on Nuclear Reactor Thermal-Hydraulics (NURETH 16), Chicago, Il (2015).

41

ERI/NRC 18-204 replenished the ECT. The IRT, ECT, and sump recirculation thus provide the primary success path for mitigating LOCA events as discussed in 10 CFR 50.36.

Now, in the 2014 published PSAR Section 13.6.1 the licensee states:

A recent study (Baek, 2014a) indicates that some scenarios may lead to the draining of water from the coolant channels. In these cases, the cooling on the inside of the fuel element (see Figure 13.9 and Figure 13.10) is from water that falls only along the inner surface of one of the side plates in the fuel elements; the remaining surfaces within the fuel element will be exposed to gas. The analysis herein discusses what would happen when this was the cooling available in scenarios where the coolant channels within the fuel elements are drained.

This statement confirms that the previously submitted analysis that provided the basis for the NRC issuance of the license renewal SER in 2009 erroneously concluded that the fuel channels remained flooded. In a meeting with the NIST staff (Ref. 20) the licensee informed the NRC and ERI staff that they have provided no notification of this information other than the PSAR submittal made on January 7, 2015. (GOI-2)

In the revised analysis documented in the PSAR the fuel does not remain flooded - it is voided fully instead within seconds of the break. The fuel is then cooled by the flow of water from the IRT through the distribution pan where a small stream of water then enters the top of the fuel channel and cools the fuel plates indirectly via conduction through the side plate. This revised analysis uses new methods and assumptions to demonstrate cooling of the fuel and these methods have not been previously reviewed by the NRC staff. The several new features of this revised LOCA analysis resulted in a series of RAIs to obtain the information required for this review.

In PSAR Section 13.3 the LOCA analysis is described as being significantly influenced by the performance of valve DWV-19. The ERI staff can find no such valve described in any of the PSAR drawings. RAI-4S1 requests that the licensee clarify this observation. In the response (Ref. 10) the licensee provides a revised drawing that clearly indicates the position of this important valve.

Based upon the information provided by the licensee in the response, the ERI staff finds that RAI-4S1 is acceptably resolved.

The LOCA evaluation is presented in PSAR Section 13.6. The licensee identified three large breaks and three small breaks for consideration and their features are described in PSAR Table 13.17 and reproduced below.

In PSAR Section 13.6.2.2 the licensee explains that:

SBLOCAs were considered for the same three locations addressed for GBLOCAs as shown in Table 13.17. The difference between an SBLOCA and GBLOCA is that the operator has time to take action. For the case with the break at the vessel outlet this makes no difference and the sequence of events proceeds as in the case with the GBLOCA (Case 1) except at a much slower rate.

42

ERI/NRC 18-204 The characteristics of the cases 1-3 are generally understandable.

Table 2-7 Break Characteristics from the PSAR Case No. Type of Break / Location Size / Remark Guillotine Break 1 18-inch pipe between the reactor vessel outlet and DWV-19 2x0.1508 m2 2 14-inch pipe between DWV-1 and the outer plenum 2x0.089 m2 3 10-inch pipe between DWV-2 and the inner plenum 2x0.0509 m2 Small Break 4 18-inch pipe between the reactor vessel outlet and DWV-19 Not simulated 5 14-inch pipe between DWV-1 and the outer plenum TRACE simulation 6 10-inch pipe between DWV-2 and the inner plenum TRACE simulation From PSAR Section 13.6.1:

This analysis is independent of whether the fuel is HEU or LEU. The second (see also Section 13.3.2) investigates the fuel and clad temperature behavior for those cases where the water has drained from some of the flow channels. The latter analysis is done for both HEU and LEU fuel elements. Note that this analysis differs somewhat from that reported in (Baek, 2014a) and (Baek, 2014b). For those previous studies the shutdown pumps came on automatically when the primary pumps trip due to low water level. However, new instrumentation is being introduced so that upon receiving a LOCA signal (low water level), the shutdown pumps will not start and the outlet valves (VALVE-83 in Figure 13.7) will not automatically open. The analysis reported on herein assumes this new mode of operation.

It is clear from this statement that:

• the licensee is currently changing the instrumentation related to LOCA events,

• the sequence of events (SOE) supporting the previously approved SER will be altered by these instrumentation changes being made by the licensee, and

• the analysis submitted in the PSAR assumes completion of instrumentation changes which is as of yet not complete.

According to statements made by the licensee during the meeting (Ref. 20) they are making these changes under the authority granted them under 10 CFR 50.59. (GOI-3)

The treatment of the LOCA events by the licensee has focused on the breaks of only three lines.

The small break LOCAs use reduced break sizes for the large break cases. For this reason, the large break cases are more limiting for figures of merit and only the large break cases are considered in this document. The following discussions refer to PSAR Figure 13.7 which is reproduced as Figure 2-6 above.

43

ERI/NRC 18-204 Licensee Analysis of Water Drainage - Case 1 A guillotine break is assumed to occur on the 18-inch pipe between the vessel outlet and DWV-

19. The event is simulated by closing VALVE-102 between the vessel outlet and the main coolant pumps and opening VALVE-3 and VALVE-22 near VALVE-102 with flow areas the same as the 18-inch pipe flow area in the TRACE model. The water level in the upper plenum drops rapidly from its normal operating level as the coolant is discharged through the simulated break. The break flow is only from the vessel side. There is no discharge of coolant from the pump side because the shutdown pump outlet VALVE-83 remains closed and PUMP-20 keeps the coolant flowing away from the break until they trip at 2.6 s. At that point, the licensee assumes that VALVE-26 closes in three seconds as designed. For this reason, there is no path for the coolant to drain from the fuel elements.

The water level inside the vessel, but outside the fuel elements, reaches the setpoint at 2.6 s and a LOCA signal is generated along with reactor scram. The water level reaches the elevation of the top of the upper fuel plates at 9.4 s, the bottom of the upper fuel plates at 10.5 s, and the top of the hold-up pan at 10.9 s. The water level inside the fuel elements is not affected by the break because there is no pathway from the bottom of the elements through the inner or outer plenums to the break (i.e., the interior of the fuel channels remained flooded).

Table 2-8 18-inch GBLOCA - Case 1 Time (s) Event Description 0.0 Guillotine break on the 18-inch pipe between the vessel outlet and DWV-19.

Water level drops in the upper plenum.

Water flows into the vessel from the IRT via the distribution pan.

2.5 Flowrate at the vessel outlet pipe decreases to 54,700 gpm.

2.6 LOCA signal is generated on low level.

First reactor scram signal is generated due to low level.

Main coolant pumps are tripped.

5.6 Valves at the main coolant pumps outlets are completely closed.

9.4 Water level outside the fuel elements reaches the top of the upper fuel.

10.5 Water level outside the fuel elements reaches the bottom of the upper fuel.

10.9 Water level outside the fuel elements reaches the top of the hold-up pan.

30.0 Simulation ends.

Licensee Analysis of Water Drainage - Case 2 A guillotine break is assumed to take place on the 14-inch pipe between the control valve DWV-1 and the outer plenum. The event simulated by closing VALVE-51 at the outer plenum inlet pipe and opening VALVE-1 and VALVE-2 in the TRACE model. The flow areas of those valves are the same as the 14-inch pipe flow area. The coolant is initially discharged through both sides of 44

ERI/NRC 18-204 the break in this case. The break flow from the pump side at VALVE-1; however, is limited due to the closure of the valves at the outlets of the primary pumps when the main pumps trip at 1.5 s.

The main pumps trip on a low-level signal at 1.5 s and the pump discharge valves are closed at 4.5 s. The coolant is drained only through the flow channels after the closure of the valves and the vessel water level stops decreasing at after 10 s as it reaches the top of the fuel elements.

The interior of the fuel channels begins draining from around 8.6 s and are completely drained in 12.4 s. At this time the fuel plates are no longer in constant contact with coolant and they never become reflooded.

Table 2-9 14-inch GBLOCA - Case 2 Time (s) Event Description 0.0 Guillotine break on the 14-inch pipe between DWV-1 and the outer plenum.

Water level drops in the upper plenum.

Water flows into the vessel from the IRT via the distribution pan.

0.4 Flowrate at the inner plenum inlet pipe decreases to 51,200 gpm.

0.8 First reactor scram signal is generated due to low inner plenum flow.

1.5 LOCA signal is generated on low level.

Main coolant pumps are tripped.

4.5 Valves at the main coolant pump outlets are completely closed.

8.6 The fuel plates start to be uncovered in the upper section.

12.4 The fuel plates are completely uncovered in the lower section.

30.0 Simulation ends.

Licensee Analysis of Water Drainage - Case 3 In this scenario a guillotine break is assumed to occur on the 10-inch pipe between the control valve DWV-2 and the inner plenum. This event is simulated by closing VALVE-70 at the inner plenum inlet pipe and opening VALVE-23 and VALVE-32 in the TRACE model. The response to this event is similar to that in Case 2, but the events occur with different timing because of the smaller break flow area in Case-3. A reactor scram occurs at 0.8 s on the lower outer plenum flow. The primary pumps trip at 2.2 s on the LOCA signal and the valves at the outlets of the pumps close in three seconds. After closure of the valves coolant drains only from the coolant channels. The uncovery of the interior of the fuel channels begins at 12.7 s and the fuel channels are completely drained at 15.6 s. At this time the fuel plates are no longer in constant contact with coolant and they never become reflooded.

45

ERI/NRC 18-204 Table 2-10 10-inch GBLOCA - Case 3 Time (s) Event Description 0.0 Guillotine break on the 10-inch pipe between DWV-2 and the inner plenum.

Water level drops in the upper plenum.

Water flows into the vessel from the IRT via the distribution pan.

0.4 Flowrate at the outer plenum inlet pipe decreases to 4,700 gpm.

0.8 First reactor scram signal is generated due to low outer plenum flow.

2.2 LOCA signal is generated on low level.

Main coolant pumps are tripped.

5.2 Valves at the main coolant pump outlets are completely closed.

12.7 The fuel plates start to be uncovered in the upper section.

15.6 The fuel plates are completely uncovered in the lower section.

30.0 Simulation ends.

ERI Staff Consideration of Water Drainage The licensee analysis of water draining credits the various instrumentation signals including, the LOCA signals, scram signals, pump trips, valve closures, and measurement signals. In Case 1 this is particularly of concern because the conclusion that no water drains from the fuel elements hinges upon the performance of VALVE-83 remaining closed, PUMP-20 tripping when commanded, and VALVE-26 closing when commanded.

The reviewers identified several concerns with this analysis and posed several questions in the multi-part RAI. In the response to the multi-part RAI (Ref. 10) the licensee provides the following information.

RAI-7aS2

a. For each case revise the sequence of events (SOE) to include the time when IRT flow is the only coolant supplied to the fuel, the time for ECT actuation, the time for recirculation flow from the pump to initiate and the point of discharge for recirculation flow. This should consider the manual actuation of the ECT and the time required for operators to diagnose the accident, decide whether to initiate bottom fill or top fill, and the time to accomplish this activity. In addition, consider the manual actuation of recirculation flow in the same manner.

The response does not modify the SOE as requested. It does not identify the time for ECT actuation. It does not address the question of whether operator action is, or is not, required. Nor does it address the timing of when such operator actions can be achieved. For this reason, the ERI staff finds that the response to RAI-7aS2 does not resolve the information requirement allowing the formation of a basis of acceptability. (GOI-4) 46

ERI/NRC 18-204 RAI-7bS2

b. For LOCA case 2, indicate when the vessel level reaches the top-of-fuel and what the level of the IRT is at that time (i.e., showing graphically, the IRT water level and discharge rate vs. time). Indicate at what time the IRT is fully drained.

The response identified the requested timing information and that information is confirmed by independent calculations by the ERI staff. Based upon the information provided by the licensee in the response, the ERI staff finds that RAI-7bS2 is acceptably resolved.

RAI-7cS2

c. For each LOCA case indicate what the limiting single failure is and ensure that the consequences of this failure are reflected in the provided results for that case.

In response to RAI-7cS2 the licensee responded that no single failure is assumed in the LOCA analysis. In addition, they referred to the response to RAI-6cS2 where they stated:

NUREG-1537 does not provide any guidance on assuming a limiting single failure.

However, Chapter 13 in the PSAR does discuss the many conservative assumptions used in the analysis.

The ERI staff finds that NUREG-1537 guidance on single-failure is discussed twice in Section 3.1, once in Section 4.2.2, once in Chapter 6, twice in Section 7.2.1, once in Section 7.2.4, once in Section 7.4, once in Section 7.5, and once in Section 13.2(3). The PSAR takes no exception to this guidance and does not propose an alternative. The statement many conservative assumptions used in the analysis is not an acceptable substitute. (GOI-1)

RAI-7dS2

d. For all LOCA cases, provide the water level in the fuel and the maximum cladding temperature as a function of time using a time scale that fully covers the participation of all elements of the core cooling system used in the SOE.

The response fails to recognize the need to modify the SOE and to fully describe the operational characteristics of the IRT, ECT, and sump recirculation. The time scale of 30 seconds fails to realistically show when the fuel no longer needs to be cooled with water (i.e., at what point can the addition of cooling water be suspended), it does not show any manual evolutions that are required (ECT valve alignment or recirculation), and it does not provide an acceptable SFA. For this reason, the ERI staff finds that the response to RAI-7dS2 does not resolve the information requirement allowing the formation of a basis of acceptability. (GOI-5)

RAI-7eS2

e. The supplied RELAP5 model input shows that the volume of the IRT is ~738 gallons and the volume of the upper plenum is ~ 1,053 gallons. Supply the total volume of the ECT, and the sump recirculation flow rate versus time for all LOCA cases.

47

ERI/NRC 18-204 The response provided the requested information which is that the volume of the ECT is 3,000 gallons. Based upon the information provided by the licensee in the response, the ERI staff finds that RAI-7eS2 is acceptably resolved.

RAI-7fS2

f. Chapter 5 of the PSAR anticipates no changes due to the HEU-LEU conversion and so it contains no technical information. However, to support our review of the LOCA analysis it became necessary to review NBSR-14 Figure 5.2. This graphic shows the discharge from the sump pump going to the D2O storage tank, not directly to the ECT. Confirm the flow path used for recirculation flow and whether all powered components are on the emergency power bus.

The response states that the water from the sump is pumped first to the D2O storage tank which can then route it to the ECT, which feeds the Inner Reserve Tank by gravity. The pumps are powered by the Emergency Power Motor Control Center and commanded from the control room.

This is a more complex flowpath than was described in the previous onsite meeting and it raises additional concerns regarding the ability of this more complex flowpath to provide the recirculation flow since single failures have not been considered. For this reason, the ERI staff finds that the response to RAI-7fS2 does not resolve the information requirement allowing the formation of a basis of acceptability. (GOI-4)

RAI-7gS2

g. PSAR 13.6.5 explains that HEATING-7.3 is used to calculate the fuel temperatures in the quiescent water after the water has drained form the fuel. Explain in more detail where this quiescent water is located. Explain if the quiescent water is assumed to accumulate, exit the bottom of the fuel, or is it allowed to heat up and evaporate. Explain how the flow from the distribution pan into the fuel channels is modeled including how the water flow is distributed over the fuel assemblies, and if there is any allowance for liquid film flow over the plates. Provide assumptions, and the boundary conditions used, including film flow rate, film thickness versus distance from top of fuel plate, specific representation of water film behavior on the fuel plates, etc. used in the analysis.

The response explains that:

the water jet emerging from the distribution pan will hit the inside of the upper end adapter of each fuel element forming a liquid film and continue to flow down one of the side plates (see Figures 13.9 and 13.10 of the PSAR). When the liquid film reaches the section where the fuel plates begin, it is first assumed that the water in the liquid film will distribute evenly among the 18 flow channels. The water flowing down each channel is in contact with three walls, a side plate and two fuel plates (or a fuel plate and an outside plate). Assuming downward channel flow, the film thickness (measured from the inside surface of the side plate) is calculated by a force balance between the gravitational force and wall shear. The friction coefficient for the wall shear can be evaluated using the Blasius equation for open channel flow. The reference cited in Section 13 of the PSAR (Baek, 2014a) discusses how the film thickness is calculated to be 0.12 cm with the concept of open channel flow when the 48

ERI/NRC 18-204 film mass flowrate is 4.2 g/s in a flow channel. It is noted that the thermal analysis of the LOCA assumed a nominal film thickness of 0.1 cm. By reference to Figure 13.11 of the PSAR the liquid film is in contact with solid regions R-9, R-4009, R10 and R-

12. The rest of the fuel plate surface is assumed to be insulated (i.e., no allowance for liquid film flow over the plates). In effect heat is removed by the liquid film only at one end of the fuel plate (R-9 and R-4049) that is adjoining a side plate (regions R-10, R-12 and R-13 in Figure 13.11 of the PSAR).

The basis for assuming an even distribution of flow is unclear. Based upon the information provided the ERI staff expects that the IRT injection flow is highly asymmetric to the cooling of the fuel plates. The submittal and the associated supporting documents do not provide any test data supporting the assumption of evenly distributed flow. There is also no analysis showing whether there is the potential for steam binding that can impact film cooling on which the analysis relies upon to demonstrate fuel integrity. For this reason, the ERI staff finds that the response to RAI-7gS2 does not resolve the information requirement allowing the formation of a basis of acceptability. (GOI-6)

RAI-7hS2

h. Provide the fuel meat, and cladding temperature as a function of time based on the HEATING-7.3 calculations, for at least one of the large LOCA and one of the small LOCA scenarios.

The response states that as a result of relatively good thermal conductivity of aluminum in the fuel meat for HEU fuel, the metal fuel meat for LEU fuel, and aluminum in the cladding, the temperature of the fuel meat and cladding are almost the same at all times. The ERI staff finds that this is a reasonable approximation. Based upon the information provided by the licensee in the response, the ERI staff finds that RAI-7hS2 is acceptably resolved.

RAI-7iS2

i. For cases where the fuel is partially or fully uncovered for any period of time, provide dose calculation to occupational workers and members of the public and indicate over what time interval these exposures apply. Relate these times to activities that are expected to be performed by operators who may be responding to events and compensating for them with operator actions such as opening valves 32-35.

The response states that the top shield plug is adequate to allow work on the reactor top during refueling. The response fails to appreciate the thrust of the RAI. A LOCA event at NBSR that voids the fuel channels will never provide a reflooded core until the break is repaired. Only thereafter can the reactor be placed in normal shutdown condition. NUREG-1537, pg. 13-7, asks the licensee to List the sequence of events, assumed equipment operation and malfunction, and operator actions until a final stabilized condition is reached. The response fails to provide this information. For this reason, the ERI staff finds that the response to RAI-7iS2 does not resolve the information requirement allowing the formation of a basis of acceptability. (GOI-7) 49

ERI/NRC 18-204 LOCA Radiological Consequences PSAR Section 13.6.4 describes the radiological impact of a LOCA event. In such an event the licensee estimates that 3,000 gallons (11,356 liters) of coolant would be discharged from the break into the sump in the first ~20 s of the event. According to TS 3.7.1 the tritium concentration limit is 5 Ci/l. This means that the sump would initially receive 56,780 Ci of tritium. Neither the PSAR, nor the RAI responses explain whether sufficient tritium would vaporize and result in a high radiation alarm resulting in an evacuation that could affect performance of expected manual actions. Site evacuation could alter the SOE as analyzed. (GOI-8)

LOCA Dose Calculations Section 13.6.4 of the PSAR provides an analysis of estimated doses to the public and the workers from releases of the entire contents of the reactor vessel, about 11 cubic meters of D 2O into the process room. This section provides an estimate of doses to the public from the release of tritium to the room and the environment. Even though this portion of the future SAR would not be changed due fuel conversion, the review identifies one anomaly (see below) in the cited information. Because, the dose results have already been reviewed by the NRC and accepted in the previous review, no attempts were made to confirm the cited results.

For the tritium concentration buildup within the process room, the analysis assumes equilibrium between the room air and saturated D2O vapor corresponding its vapor pressure. This analysis determines a derived air concentration (DAC) ratio of 1.25x104, which is confirmed using the related information. However, in the PSAR, the cited DAC ratio is 1.25x10-4 indicating that this value is a typographical error in the PSAR.

LOCA - Confirmatory Analysis A total of six scenarios are presented in PSAR Section 13.6 to describe the NBSR response to various breaks causing a LOCA. The limiting event is case 2 in which a double-ended guillotine break occurs on the 14-inch outer plenum inlet pipe. It is simulated by closing VALVE-51 at the outer plenum inlet pipe and opening VALVE-I and VALVE-2 near VALVE-51 (see PSAR Figure 13.7).

This accident results in draining all the water from the inside of the assembly and reducing the resulting level in the core region external to the assemblies to the level of the holdup pan as illustrated in Figure 2-22. In this analysis, TRACE is used to model only the drain-down portion of the event. Comparisons are then made to the equivalent RELAP5 results.

50

ERI/NRC 18-204 Figure 2-22 Confirmatory Case-2 LOCA Graphic - Post Blowdown The confirmatory analysis shows the break flow from pump side and vessel side from both TRACE (ERI staff) and RELAP5 (licensee) in Figure 2-23. The comparison shows reasonable agreement in general behavior. However, there is two noticeable differences found near the beginning and the end of calculation period.

As seen in Figure 2-23, the RELAP5 maximum break flow (620 kg/s) from pump side is noticeably larger than that from TRACE (530 kg/s). It is noted that this break flow is almost the same as the maximum pump flow (625 kg/s) predicted from TRACE. However, the break flow from pump side cannot be the same as the pump flow since the pump flow is split into two flow paths: one is for the break and the other is to the inner plenum piping. In that sense, the PSAR maximum break flow from pump side seems to be overpredicted by RELAP. Figure 2-23 results are truncated at approximately the time of the complete core uncovery since analysis of film flow cooling following IRT initiation cannot be modeled accurately in TRACE.

51

ERI/NRC 18-204 Figure 2-23 Core Water Level in the Hot Plate The ERI staff finds that the confirmatory results for the LOCA draindown in case 2 indicates some issues with the differences in the predicted behavior when comparing the RELAP5 to the TRACE confirmatory predictions. Overall, the drain-down time is quite similar to that provided by the licensee response and it is considered acceptable (See Table 2-11).

52

ERI/NRC 18-204 Table 2-11 TRACE and RELAP Comparison of Draindown Analysis for Case 2 Event (case 2) RELAP TRACE Guillotine break takes place at the 14-inch pipe between the control valve 0.0 0.0 DWV-1 and the outer plenum.

Flowrate at the inner plenum inlet pipe decreases to the setpoint of low inner 0.4 1.55 plenum flow (<1,200 gpm).

First reactor scram signal is generated due to low inner plenum flow. 0.8 1.95 LOCA signal is generated due to low level. 1.5 1.5 Main coolant pumps are tripped.

Valves at the main coolant pump outlets are completely closed. 4.5 4.5 The hot fuel plate starts to be uncovered upper section (Node-407). 8.6 9.3 The hot fuel plate starts to be uncovered lower section (Node-403). 8.7 9.4 The hot fuel plate is uncovered upper section (Node-407). 12.3 12.8 The hot fuel plate is uncovered lower section (Node-403). 12.4 12.9 Simulation ends 30.0 30.0 LOCA Technical Specification Issues As stated earlier in this document, the IRT, the ECT, and the recirculation sump provide the primary success path for delivering cooling water to the fuel post-LOCA. This terminology fits exactly with the terminology in 10 CFR 50.36 regarding technical specifications. Therefore, the ERI staff expects that all the elements in this success path are appropriately incorporated into the TS. However, there are instances where they are not.

The licensee response to RAI-7aS1 specifically credits the contribution of cooling water to the fuel from the IRT consequential to a LOCA. However, there is no mention of the IRT in the existing TS or in the proposed changes to the TS in PSAR Chapter 14. TS 3.3.2 Emergency Core Cooling establishes no limiting condition for operation (LCO) for the IRT and TS 4.3.2 Emergency Core Cooling establishes no surveillance requirement. This conflicts with the requirement of 10 CFR 50.36 Criterion 3 that requires LCOs for:

A structure, system, or component that is part of the primary success path and which functions or actuates to mitigate a design basis accident or transient that either assumes the failure of or presents a challenge to the integrity of a fission product barrier.

The success path for emergency core cooling is stated clearly in the response to RAI-7eS2 and it includes the requirement for 3,000 gallons of D2O water, supplied to the IRT at 40 gpm for 2.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />. TS 3.3.2 does not include specifications requiring the ECT to have 3,000 gallons of coolant to be available, nor does it require delivery at 40 gpm. TS 4.3.2 does not verify this volume. It does not specify a delivery rate in the specification, but in the basis requires 25 gpm only. Transfer of 3,000 gallons at 40 gpm terminates after 1.25 hours2.893519e-4 days <br />0.00694 hours <br />4.133598e-5 weeks <br />9.5125e-6 months <br />. At 25 gpm it terminates at 2.0 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />.

53

ERI/NRC 18-204 There is no commitment for sump recirculation in the LOCA analysis or in the TS. Therefore, the analysis should support the success path identified and the TS should codify that success path.

The TSs for the IRT, ECT, and recirculation cooling do not satisfy the requirements of 10 CFR 50.36. (GOI-9)

The PSAR documents the improved understanding regarding the behavior of NBSR to certain LOCA events. This understanding is markedly different from the understanding presented in NBSR-14 (Ref. 2) that supported the issuance of the current SER (Ref. 3).

LOCA Conclusions The supplied LOCA analysis does not:

• provide a comprehensive SOE,

• analyze long term cooling of NBSR,

• indicate whether a LOCA event as described would result in radiological consequences that would result in a site evacuation alarm and when would such an evacuation occur in the SOE,

• provide convincing evidence that the manual operations required to engage the ECT or sump recirculation can be accomplished and show such operation in the SOE,

• provide LCOs for the components of the LOCA accident success path and SRs that ensure that the performance levels of these components are maintained.

2.7.6 Natural Circulation at Low Power Event This event is discussed in PSAR Section 13.7. To analyze this event, RELAP5 is used to simulate operation at low power without forced-flow cooling. The LSSS for reactor power with natural circulation cooling is 10 kW. To demonstrate the conservatism in this TS, the analysis is performed at a power level of 100 kW. Since low power operation without forced flow is an allowed configuration under the TS, further information on this event is documented in Section 2.6 of this report.

The ERI staff concludes that the results of events involving the natural circulation at low power event continues to be acceptable with the conversion to LEU fuel.

2.7.7 Flow Blockage Accident - the Maximum Hypothetical Accident NUREG-1537 states:

In general, the escape of fission products from fuel or fueled experiments and their release to the unrestricted environment would be the most hazardous radiological accident conceivable at a non-power reactor. However, non-power reactors are designed and operated so that a fission product release is not credible for most.

Therefore, this release under accident conditions can reasonably be selected as the 54

ERI/NRC 18-204 MHA, which bounds all credible accidents and can be used to illustrate the analysis of events and consequences during the accidental release of radioactive material.

In PSAR Section 13.8, the licensee describes the flow blockage as a postulated event that assumes that there is a complete blockage of flow to one fuel element leading to complete melting of 34 fuel plates (i.e., 17 upper and 17 lower). They characterize this event as the maximum hypothetical accident (MHA) which establishes that this event provides the most significant radiological hazard to the facility. This event is analyzed for both the HEU and the LEU cores.

The SOE is that after the occurrence of the blockage, boiling would begin in the coolant channels.

This results in a decrease in reactivity which can also cause fluctuations in power. It is assumed that these fluctuations would not cause a reactor scram. The fuel element can then heat up to the point at which the fuel plates melt, releasing radionuclides into the D2O coolant. The radionuclides in the reactor core cover gas will rapidly reach the stack and that will initiate a reactor scram. According to PSAR Section 13.8 the melting of fuel occurs far below 1,000°C and so metal-water reactions do not need to be considered. The ERI staff agrees with this SOE.

Nuclide Inventory PSAR Table 13.21 provides the activities of select iodine and noble gas isotopes in a half-element for both the HEU and LEU fuel elements. RAI-11S1 the following issues that require additional information (Ref. 2). In the response to RAI-11S1 (Ref. 10) the licensee describes the inventory generation process in some detail.

Previous analysis (Ref. 10) developed the assumed inventory for HEU fuel using ORIGEN and assumed that the activity resulted from burning a fuel element for eight cycles with an eleven-day cooling period between cycles. The power in a fuel element was the core average power during the irradiation period. Those calculations used one-group cross sections from a library generated for a heavy water CANDU reactor. The PSAR analysis uses the deletion module within MCNPX.

The analysis is based on the actual NSBR fuel definition for the equilibrium HEU and LEU cores.

The compositions in any half-element consider the history of that half-element, rather than assuming an average power. Fuel elements are specifically tracked by location and therefore, the exposure is much more accurate than previously described. A 63-group flux spectrum is generated by MCNPX and used with the CINDER 63-group library to generate appropriate one-group fluxes to solve the burnup equations. The PSAR calculation uses a 38.5-day cycle and a 10.5-day cooling period which is realistic. The composition of a limiting half element is then depleted in CINDER for five days to obtain the activity in Curies. The final calculation has the equilibrium distribution of isomers which is important because of their high activity. Overall the applicant concludes that this methodology is more rigorous and realistic than the original ORIGEN analysis.

The ERI staff reviewed the licensees MHA source term radionuclide inventory analysis and performed confirmatory inventory calculations for the selected isotopes using information on fission yields (Ref. 21) in conjunction with a maximum fuel element power of 417 kW. This fuel element power is an average power from the startup to the end of cycle as depicted in PSAR Table 4.13. Because the NSBR is continuously operating and the half-lives of the listed isotope, except for the krypton-85 (Kr-85), are much smaller than the irradiation time, these isotopes should be saturated.Table 2-12 below provides a comparison of radionuclide inventories for select 55

ERI/NRC 18-204 halogens and noble gases, as provided by the licensee; and those calculated using typical fission product yields for U-235 by the ERI staff in the confirmatory analysis. (Note, the inventory calculation focuses on the LEU fuel only.) Even though there are differences in the estimated inventories, the overall results are very similar for many of the more radiologically important isotopes. However, there are differences in the estimated inventories for some radiologically important isotopes, such I-131. This isotope is the dominant contributor to the occupational dose during an accident and the consequential releases into the experimental and operation rooms.

Table 2-12 Estimates of MHA Source Term Nuclide Inventory Isotope Half-life Licensee Estimate of MHA Confirmatory Estimate of MHA Nuclides (s) Source Term in Half Element (Ci) Source Term in Half Element (Ci)

I-130 4.45x104 7.24x101 7.65x101 I-130m 5.30x102 5.35x101 2.40x101 I-131 6.95x105 8.68x103 1.04x104 I-132 8.28x103 1.52x104 1.56x104 I-132m 5.02x103 6.21x101 3.29x101 I-133 7.49x104 2.44x104 2.42x104 I-134 3.16x103 2.88x104 2.82x104 I-134m 2.16x102 1.53x103 1.31x103 I-135 2.38x104 2.32x104 2.27x104 Kr-83m 6.59x103 1.93x103 1.93x103 Kr-85 a 3.39x108 3.12x101 1.02x103 Kr-85m 1.61x104 4.27x103 4.66x103 Kr-87 4.58x103 9.20x103 9.23x103 Kr-88 1.02x104 1.23x104 1.28x104 Xe-131m 1.02x106 2.42x101 1.46x102 Xe-133 4.53x105 1.88x104 2.42x104 Xe-133m 1.89x105 4.93x102 7.03x102 Xe-135 3.29x104 1.82x103 2.36x104 Xe-135m 9.17x102 4.43x103 3.97x103 Xe-137 2.29x102 2.25x104 2.21x104 Xe-138 8.45x102 2.29x104 2.27x104 Total Inventory 2.01x105 2.30x105 a Kr-85 with a half-life of 10.7 years will never saturate given the NBSR cycle length.

The bold values represent a difference of more than 10 percent.

The bold italic values represent the differences of a factor of 5 or more 56

ERI/NRC 18-204 Release Fractions The licensee assumes that all of the noble gases in one fuel element are released into the reactor vessel. Because the noble gases are insoluble in water they quickly collect in the helium cover gas space that has a volume of about 0.7 cubic meters at the top of the reactor vessel. Based on the analysis performed for the 2010 SAR (Ref. 22), the licensee assumes that about 3 percent of the iodine that enters the cover gas will be in the form of I2. The amount of I2 release into the helium space is based on the assumption that 50 percent of the iodine that is in the fuel element is released into the reactor vessel followed by the 95 percent absorption in the reactor coolant, which leads to an overall iodine release of 2.5 percent (rounded to 3 percent). The ERI staff finds that these assumptions are consistent with the current NRC guidance.

Releases to the Confinement Buildings The noble gas and iodine isotopes are released into the confinement building along with helium cover gas at a rate characteristic of the tightness of the primary system under emergency ventilation conditions. These rates to the different areas within the confinement building are shown in Table 2-13 along with the rate of removal due to the emergency ventilation system which has been activated. The rates are assumed to be constant with time.

Table 2-13 Leak Rates to Confinement and Release Rates to Stack Confinement Area Room Volume Floor In-leak Rate Removal Rate to the Stack (m3) (m3/s) (m3/s)

Experimental Floor (C-100) 4.5x103 1.2x10-6 9.4x10-3 Operations Level (C-200) 8.3x103 8.1x10-6 1.9x10-2 Process Room (Basement) 2.0x103 2.3x10-6 7.1x10-3 Given the release rates in Table 2-13, the licensee then calculates the time-dependent accumulation of released iodine and noble gas isotopes within the confinement areas and determines the quantities that are released to the environment.

Dose Calculations The licensee calculated the potential MHA total effective dose equivalent (TEDE) for an occupational worker in two locations within the confinement areas (Experimental floor, and operation level), and a public individual at the 400-meter exclusion zone outside the reactor building. The conditions for these calculations include the assumptions:

a. the failure of the hottest fuel element and

b. nuclide decay.

Other parameters used in the dose calculations include a breathing rate of 0.02 m3/minute (333 cubic centimeters per second), which is consistent with the value given in Appendix B to 10 CFR Part 20, and a stay-time of 10 minutes.

57

ERI/NRC 18-204 For the public doe, the licensee used the general Gaussian plume diffusion model to calculate nuclide concentrations at a 400-meter downwind distance, using the HotSpot computer program (Ref. 23).

Occupational Dose The PSAR analysis assumes that the occupational workers would evacuate the building within 10 minutes after an MHA. Their TEDE is determined by converting the concentration of radioactivity, conservatively assumed to be uniform throughout the specified space, into dose, considering both immersion and inhalation. The analysis utilizes the radiation dosimetry Dose Conversion Factors (DCF) from the Hotspot computer program. The results are listed in Table 2-14. The annual occupational dose limit for workers is 5-rem from 10 CFR 20.1201, which is considerably higher than the expected doses shown in the table.

Table 2-14 Ten-Minute Dose (TEDE) to NBSR Staff after MHA Location HEU Fuel (rem) LEU Fuel (rem)

C-100 0.3 0.3 C-200 2.1 2.1 The ERI staff performed confirmatory analyses for the MHA involving LEU fuel using the nuclide inventories provided by the licensee and those calculated using the fission yields (see Table 2-14). These analyses use the same assumptions as used by the licensee. The ERI staff finds that the DCFs used in the HotSpot computer program are acceptable based upon ERI staff experience in reviewing similar applications. Table 2-15 provides a comparison of the calculated occupational doses. Even though the NRCs confirmatory analyses show higher occupational doses during the 10-minute exposure before the evacuation, the calculated doses are lower than the dose limit of 5-rem in 10 CFR 20.1201. The ERI staff finds that this information satisfies the request in RAI-8eS2.

Table 2-15 MHA Occupational Worker Dose Estimates in Restricted Locations ERI Confirmatory NIST LEU (rem) 10 CFR 20.1201 Limit Location (rem) (rem)

NIST Inventory Saturated Inventory C-100 0.3 0.71 0.77 5

C-200 2.1 2.6 2.8 Public Dose Public dose is determined using Hotspot computer code. In the PSAR, the licensee added that the following parameters were used in the HotSpot model: low wind speed, highly stable atmospheric conditions, and no change in wind direction. Additionally, the licensee assumes that for the two-hour dose, the concentration of material is what is obtained at the end of two hours, 58

ERI/NRC 18-204 considering an average one-hour decay time. The instantaneous release for the iodines is reduced by the factor of 0.999 to consider the efficiency of the filters in the ventilation system.

Similarly, for the release used in Hotspot for the period 2-24 hours, the amount of instantaneously released material is that accumulated over the 22-hour period considering an average 11-hour decay. The approach for the 1 to 30-day dose is similar. In each case the material is released at the top of the 30-m high stack. The licensee added that of the original inventory, more than 80%

of the gaseous material will leak from the building within 30 days.

RAI-8S2 requests additional clarifications on the specific assumptions related to wind speed, atmospheric stability, elevated release, fraction of iodine that is organic, and the use of HotSpot code for a long duration release. In the response to RAI-8S2, the licensee provides the following clarifications:

• It was assumed that 0.15 percent of the releases iodide is in organic form consistent with the guidance in Regulatory Guide 1.183 (Ref. 24).

• The calculations assumed a wind speed of 1 meter per second and performed the calculations using both stability class-F (moderately stable) and class-A (very unstable).

• Because the stack height is not 2.5 times the height of adjacent solid structure, the analysis used the virtual source model in the HotSpot code, which puts the release point upstream of the stack release that corresponds to the lateral and vertical plume spread equal to the width and height of the building (considered to be 27 meters. This model creates building wake effect leading to a larger dose near the structures.

• With respect to the use of HotSpot for a long duration release (more than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />), a yearly met data from the nearby BWI airport is used in conjunction with the met data option in the HotSpot code. (The licensee provided two HotSpot output results, one for a 2-hour release and one for a 1 to 30-day release, both with the stability class-F meteorology.)

Based on the above considerations, the licensee revised the TEDE to an individual at the 400-meter exclusion area for the LEU fuel. The results as indicated in Table 1 of the response and reproduced in Table 2-16. Similar changes for the HEU fuel should also be made and included in the revised PSAR Table 13.24.

Based on the results in Table 1 of the response, the licensee states that the total dose to a member of the public is acceptable, because it is less than 100 mrem (10 CFR 20.1301). The results given in the table in mrem show a significant margin to this limit. They also show that there is no need to calculate the total dose at any other location or for longer periods.

During the meeting at NIST (Ref. 20), the licensee provided additional insights on the revised dose calculations and provided additional information on the approach in determining the time-dependent activities within the various locations in the restricted areas, and a yearly meteorology data used in the analysis for a 1 to 30-day release period.

59

ERI/NRC 18-204 Table 2-16 Dose (TEDE) to an Individual after MHA Stability Location LEU Fuel (mrem)

Class (m) 0-2 hr 2-24 hr 1-30 days Total F 400 0.7 9.0 9.7 19.4 210/max dose 1.0 12.0 13.0 26.0 A 400 0.2 2.1 2.3 4.6 55/max dose 2.0 24.0 26.0 52.0 The ERI staff performed confirmatory analyses for the both in-building and external activity concentrations. In addition, based on these concentrations it calculated the occupational and public TEDEs. The occupational doses are listed above Table 2-15.

As stated previously, the licensee used the HotSpot code in conjunction with virtual source to determine the dose to a maximally exposed individual at 400-meter exclusion area with an elevated release (i.e., out the building stack). This option simulates inclusion of the building wake.

Because the NBSR stack height is less than 2.5 times the height of the nearby structures, Regulatory Guide 1.145, position 1.3.1.a provides an approach to determine the relative dispersion concentration using the effect of the building wake (Ref. 25). The ERI staff performed a confirmatory calculation following the guidance in Regulatory Guide (RG) 1.145 for a 0-2 hours release using the licensee assumptions. The ERI staff calculation of the concentration (e.g., x/Q value) at a distance of 400 meters finds that the RG 1.145 approach results in a lower value than that generated by the HotSpot code (6.3x10-5 vs. 7.6x10-5). This results in a calculated TEDE of 0.64 mrem compared to the HotSpot calculated dose of 0.71 mrem. Based on these calculations, the ERI staff finds that the analysis approach is acceptable.

For the public dose, the ERI staff was able to confirm the doses calculated by HotSpot for short-term release duration. The ERI staff finds that the yearly meteorology data provided by the licensee has a format inconsistency with HotSpot Version 3.1 and this data is not properly processed by the code. After resolving this format issue, the results indicate that the model includes a windspeed of zero meter per second, which results in code execution problems. After changing to a wind speed of 0.1 meter per second, the ERI staff was able to reproduce the wind speed grouping fractions as provided in the response to RAI-8S2, however, the code produces dose results that are not consistent with those provided by the licensee. The licensees cited dose at 400 meters in Table 1 of RAI 8S2 response for the 1-30-day release (i.e., 9.7 mrem) cannot be found in the corresponding information in the attachment Hotspot_output_1_30_d_stab_F.txt to the RAI response. Because the doses are calculated for an individual at all sectors and distances specified, at each distance the cited dose cannot be the sum of doses in all sectors. It appears that the licensee has used the weighted sum of doses in all sectors as indicated (e.g., 13.0 mrem with the 95th percentile results) for the maximum dose at about 200 meters for the 1-30-day exposure. The licensee needs to revise the cited dose results for the long duration releases to reflect the maximum dose in a sector at a given distance from the point of release. (COI-8) 60

ERI/NRC 18-204 At 400 meters from the reactor building, the ERI analysis finds a maximum dose of 8.35 mrem and 42.4 mrem at East direction for the 50th and 95th percentile dose distribution results. The maximum dose occurs at about 200 meters at East direction with values ranging from 11 to 55.7 mrem with 50th and 95th percentile results.

Finally, the ERI staff performed a public dose calculation for the 1 to 30-day release by conservatively assuming the weather condition remains unchanged with a windspeed of 1 meter per second and stability class-F. This dose will then be considered as the maximum dose.

Table 2-17 summarizes the offsite dose results calculated as part of the confirmatory analysis or provided by the licensee. It should be noted that only one met file with stability class-F was provided by the licensee; therefore, most analyses were performed using the class-F stability. In addition, it was assumed that the weather condition during the analysis period (one month) will remain unchanged. This analysis will be similar as considering the release occurs over a brief period which is the main premise behind the development of HotSpot code, as indicated in its user manual (Ref. 23).

Table 2-17 Offsite Dose Consequences LEU Fuel (mrem)

Stability Location (meters)/source Class 0-2 hr 2-24 hr a 1-30 days a Total 400/NIST 0.7 9.0 9.7 19.4 400/ Confirmatory 0.71 2.59/13.0/14 8.35/42.4/25 12/46/40 F

Max dose b / NIST 1.0 12.0 13.0 26.0 Max dose / Confirmatory 0.95 3.42/17.2/19 11/55.7/33 15/74/43 400/NIST 0.2 2.1 2.3 4.6 400/ Confirmatory 0.17 NA d/NA/3 NA/NA/6.1 NA/NA/9 A

Max dose c / NIST 2.0 24.0 26.0 52.0 Max dose / Confirmatory 1.95 NA/NA/39 NA/NA/68 NA/NA/119e a For the confirmatory calculations 3 different values are cited, they are 50th/95th dose percentiles when the historical site met data is used and location dose data if we were to assume the weather condition remains as that of the 02-hour release.

b Maximum dose occurs at about 200 meters from the point of release under stability class-F.

c Maximum dose occurs at about 55 meters from the point of release under stability class-A.

d NA means not analyzed, because no historical met data with stability class-A is provided.

e This dose value is not valid because the location (55 meters) from the building is within the facility control area and it is very unlikely that the weather condition would be very unstable for a month.

The ERI staff finds that the results of the MHA with the reactor fueled with LEU fuel is similar to that for the HEU fueled reactor and that the information requested in RAI-8S2 and RAI-11S1 is acceptable. Because the doses are within the limits established by 10 CFR 20.1301, the results of the MHA continue to be acceptable.

61

ERI/NRC 18-204 2.7.8 Mishandling, Malfunction, or Misloading of Fuel The PSAR describes some aspects of some potential fuel malfunction events. RAI-12S1 requests clarification regarding certain aspects of such events. In the response (Ref. 10) the licensee describes the considerations that protect against radionuclide releases from such events. These include shield plug placement, release monitors, and their previous analysis of heavy load consequences. Based upon the information provided by the licensee in the response, the ERI staff finds that RAI-12S1 is acceptably resolved.

Four separate scenarios involving mishandling of fuel were extensively analyzed during license renewal (Ref. 3) and shown to present no significant risks. These events were:

• a refueling event involving a dropped fuel element

• dropping of a fuel element into the storage pool

• dropping of a heavy object onto the fuel rack in the storage pool

• dropping of the spent fuel cask during a shipping operation The licensee states that there are no anticipated changes in any of these events that are expected for the LEU-fuel and the previous analysis applies also to the LEU fuel.

However, during a facility walk-down with site personnel (Ref. 20) the licensee described the process that they use to modify the length of the burned fuel assemblies and cut the non-fueled end pieces so that the fuel elements then fit at a higher density within the shipping casks. The licensee provides no discussion of this issue in the PSAR, the NBSR-14, or the SAR. RAI-11S2 requests that the licensee to clarify consideration of this event and contributing factors such as debris impingement, cladding longitudinal tears, cutting blade fracture, or any other credible failure that could lead to breaching the cladding-fuel interface. In the response the licensee states that a complete transverse cutting of the fueled section of a fuel element would result in fission product release from the region of the fuel plate near the cut and it is estimated that the release would be less than 1% of that predicted for the MHA, under the worst conditions. From the ERI staff consideration of the reported information, it is assumed that if the meat of the fuel plate is accidentally cut during fuel element shortening activities, then the amount of the fuel-plate fission product activity that would be available to be released would depend on the temperature of the fuel, the burnup of the fuel, the surface area of the fuel that is exposed, and the location where it occurs. As explained by the licensee, these activities are conducted after the fuel is decayed in the spent fuel pool. Therefore, the fuel temperature would be very low and diffusion from the fuel matrix would be essentially negligible. Consequently, the only radionuclides that would be available for release from the cut surface would be due to the kinetic energy associated with fission fragment recoil from the fuel surface that is exposed by the cutting.

The noble gases and iodine from the exposed fuel plate that are within the recoil distance are released into the pool. Using NUREG/CR-2079 (Ref. 26) the recoil distance is 1.37 millimeters.

The incorrect cutting of the fuel element could lead to a severance of 17 fuel plates, exposing the fuel meat either in the upper or the lower section of the fuel element. Using the recoil distance and the dimensions of the LEU fuel plate, the recoil volume of the 17 fuel plates cut surfaces would be is about 0.5 percent of the fuel meat in a fuel element. Therefore, about 0.5 percent of 62

ERI/NRC 18-204 the noble gases and iodine that are in a fuel element will be released to the pool. This amount is smaller than the release fractions resulting from an MHA. Given that any fuel element shortening activities would occur long after the fuel is discharged from the reactor, the gaseous releases would be limited to long-half-life isotopes. These include I-131 (8.04 days), Kr-85 (10.7 years),

Xe-131m (11.8 days), Xe-133 (5.24 days), and Xe-133m (2.19 days). In all likelihood, except for Kr-85, all other isotopes of interest have already been decayed before the fuel element shortening activities could occur. Therefore, the ERI staff concludes that the consequences of such accident are bounded by the MHA.

Hence, the fuel cutting activity creates a possibility of a type of accident leading to a radionuclide release that is not discussed in NBSR-14, the revised SAR, the PSAR, or the TS and has not been reviewed by the NRC staff for its acceptability. The ERI staff finds that fuel cutting accidents are events that have potential for accidental radionuclide releases, they should be discussed in the PSAR, and the activity should be controlled in the NBSR TS. (GOI-9) 2.7.9 Experiment Malfunction PSAR Section 13.10 states that this event was previously analyzed and there is no expected change as a consequence of the conversion. RAI-14S1 requests the licensee to confirm whether there are hot cells, fume hoods, or glove boxes at NBSR connected to a ventilation system that can exhaust to public receptors. The response to RAI-14S1 (Ref. 10) confirms that NBSR has chemical fume hoods and a glove box for handling experiments on the reactor license which are connected to the ventilation system.

The current and proposed TS for NBSR allow the licensee to perform fueled experiments.

However, neither NBSR-14, the SAR, nor the PSAR provide any discussion of the potential for such failures and the dose consequences. RAI-13S1 requests clarification on fueled experiments in NBSR. In the response to this RAI (Ref. 10), the licensee states that as indicated in the response to RAI-11, there never has been, and there are no plans for, any fueled experiments, as that is not part of the mission of the NCNR. The licensee adds that the amount of radioactivity involved in any experiment is controlled by TSs 3.8.2 and 6.5, which require a thorough hazards review before an experiment can be initiated.

The ERI staff finds that since the TSs allow a fueled experiment the consequence of a fueled experiment malfunction within the closed reactor primary system should be shown to be bounded by the analyzed MHA. Furthermore, if such experiment can be then removed from the vessel the dose in the limiting location (e.g., in transit, fume hood, or glove box) and the consequences of a malfunction in a fueled experiment should be established by the review process and controlled so that they are less than that evaluated in the MHA.

The TSs, those in use now and those provided for this review, include no provision requiring that irradiated experiments be limited so that the consequences of experiment failure are bounded by the MHA. (GOI-9) 63

ERI/NRC 18-204 2.7.10 Loss of Normal Power The bounding scenario of a loss of normal power is the LOSP event as described in Section 2.7.4 of this report.

The ERI staff concludes that the results of events involving loss of normal power to be acceptable with the conversion to LEU fuel.

2.7.11 Equipment Malfunction The ERI staff concludes that the results of events involving equipment malfunction to be acceptable with the conversion to LEU fuel.

2.7.12 External Events The primary event of concern to the ERI review is the response of the revised fuel design to seismic events. The previous fuel design response to seismic events (e.g., deflection) are a function of material properties (e.g., stiffness, yield strength, etc.), geometry, and mass. The anticipated mass of the LEU fuel is greater than the HEU mass. Other properties may change also. It is recommended to provide a technical evaluation of the LEU fuel response to seismic input in the PSAR.

2.7.13 Accident Analysis Conclusions The ERI staff finds that the event analysis provided in the PSAR does not provide sufficient information to form a basis of acceptability for several topics. The issues listed in Table 2-18 need to be resolved.

2.8 Fuel Storage The licensee response to RAI-10S2 states that:

The analysis for the previous 2004 SAR has been updated and is referenced in two reports that are available upon request. The reports describe a conservative analysis for HEU fuel that shows there is no possibility of an inadvertent criticalityif the elements and fuel pieces are secured properly Calculations are also presented showing the impact of loading spent LEU fuel and again the margin to criticality is large.

These reports need to be reviewed. When they are reviewed, it will be to the TS 3.9.1 criterion that the calculated keff will be less than 0.90 under optimum conditions or water moderation and reflection. The statement in the response that there is no possibility of an inadvertent criticality is insufficient. For this reason, the ERI staff finds that the response to RAI-10S2 does not resolve the information requirement allowing the formation of a basis of acceptability. (COI-7) 64

ERI/NRC 18-204 Table 2-18 Accident Analysis Summary Topic Information or Action Needed Reactivity Insertion Events The licensee needs to identify the limiting single-failure event and demonstrate that the resulting safety margins are acceptable.

Loss-of-Flow Events The licensee needs to identify the limiting single-failure event and demonstrate that the resulting safety margins are acceptable.

Loss-of-Coolant Accident (LOCA) The licensee needs to identify the limiting single-failure event and demonstrate that the resulting safety margins are acceptable.

The licensee needs to reconsider pursuing change to the facility to change LOCA behavior under 10 CFR 50.59.

The licensee needs to develop more fully the SOE for LOCA including consideration of radiological consequences resulting in site evacuation and long-term cooling.

The licensee needs to demonstrate acceptable long-term cooling.

The licensee needs to justify more fully the methodology for cooling the fuel plates by the flow from the distribution pan.

The licensee needs to modify the TSs to provide LCOs and SRs for all components that are part of the LOCA success path.

Flow Blockage Accident - the Maximum The licensee needs to confirm the acceptability of Hypothetical Accident some of the high-importance isotopes in the inventory.

The licensee needs to revise the noted input errors in the HotSpot model and confirm the acceptability of the results.

Mishandling, Malfunction, or Misloading of Fuel The licensee needs to fully disclose and discuss the process of cutting fuel assemblies into pieces in the future SAR, consider such events for their potential to be event initiators, and impose appropriate controls in the TSs.

Experiment Malfunction The licensee needs to revise the TSs to include a provision requiring that irradiated experiments be limited so that the consequences of experiment failure are bounded by the MHA.

65

ERI/NRC 18-204 2.9 LEU Startup Plan In response to RAI-11S2 the licensee defers the final response requesting the startup plan until the fuel design activities are further along.

2.10 Proposed Changes to License Conditions and Technical Specifications 2.10.1 Proposed Changes to License Conditions None identified.

2.10.2 Proposed Changes to Technical Specifications In PSAR Chapter 14 the licensee proposes certain changes to the TS. As previously mentioned in Section 2.4, the licensee has stated that they plan to transition to LEU over several cycles.

With this in mind the proposed TS will not be suitable as they address a LEU-only core.

Technical Specification 5.3, Reactor Core and Fuel:

4. The 20 MW reactor core may consist of 30 3.0x3.3-inch (7.6x8.4 cm) MTR curved plate-type fuel elements. The NBSR MTR-type fuel elements shall be such that the central 7 inches of the fuel element contains no fuel. The middle 6 inches of the aluminum in the unfueled region of each plate shall have been removed.

5. The side plates, unfueled outer plates, and end adaptor castings of the fuel element shall be aluminum alloy.

6. The fuel plates shall be uranium-molybdenum alloy foils clad with aluminum alloy with a zirconium interlayer between foil and clad.

Basis:

3. The neutronic and thermal hydraulic analysis was based on the use of 30 NBSR MTR-type thirty-four (34) plate fuel elements. The NBSR fuel element has a 7 inch centrally located unfueled area, in the open lattice array. The middle 6 inches of aluminum in the unfueled region has been removed. The analysis requires that the fuel be loaded in a specific pattern. Significant changes in core loading patterns would require a recalculation of the power distribution to ensure that the CHFR would be within acceptable limits.

4. and 3. The fuel element with aluminum alloy clad and uranium-molybdenum alloy foils have been qualified for use in the NBSR.

Technical Specification 2.1 Specification:

The reactor fuel cladding temperature shall not exceed 716°F (380°C) for any operating conditions of power and flow.

66

ERI/NRC 18-204 Basis:

Maintaining the integrity of the fuel cladding requires that the cladding remain below its blistering temperature of 716°F (380°C). For all reactor operating conditions that avoid either a departure from nucleate boiling (DNB), or exceeding the Critical Heat Flux (CHF), or the onset of flow instability (OFI), cladding temperatures remain substantially below the fuel blistering temperature.

Conservative calculations have shown that limiting combinations of reactor power and reactor coolant system flow and temperature will prevent DNB and thus fuel blistering.

LCO 3.1.2, Reactivity Limitations Specifications:

3. The maximum available excess reactivity for the reference core conditions shall not exceed 15% (approximately $20)

4. The reactor shall not be operated unless shutdown margin provided by the shim arm is greater than 0.68% ($1.0) with:

c) The reactor in any core condition, and d) All movable experiments in their most reactive condition.

LCO 3.1.4, Fuel Burnup Specification:

The average fission density shall not exceed 4.5x1027 fissions/m3.

Basis:

Fuel elements in the NBSR are burned for seven or eight cycles. An eight-cycle fuel element has an average fission density of approximately 4.1x1027 fissions/m3 (Brown 2014). Allowing for a 10% increase provides the specification.

The ERI staff review of these proposed TSs finds that, subject to formal approval of the fuel design, they are acceptable.

2.10.3 Conclusions Although the ERI staff finds the TSs submitted in the PSAR to be acceptable to support the HEU-LEU conversion, this review documents several other TS issues that are preventing the formation of a basis of acceptability. The ERI staff finds that the following issues require TS under 10 CFR 50.36, but, no such TSs exist or are proposed (GOI-9):

• There is no cooldown period identified for leaving the facility unmanned.

• The IRT has no LCO or surveillance requirement (SR).

• The ECT has no LCO/SR for the volume requirements.

• TS 4.3.2(1) requires valves to be exercised but there is no corresponding LCO.

67

ERI/NRC 18-204

• TS 3.3.2(2) requires a source for ECT makeup but there is no LCO/SR identifying the sump pumps that provide such service or their requirements.

• The basis for TS 3.3.2 identifies service water as an alternative to the sump pumps, but there is no LCO/SR for its use.

• There are no TSs controlling the dismantling of fuel.

68

ERI/NRC 18-204

3. CONCLUSIONS The ERI staff has reviewed and evaluated the operational and safety factors affected by the use of LEU fuel in place of HEU fuel in NBSR. This concludes, on the basis of the information provided that the proposal by the licensee for conversion of the reactor to LEU fuel is consistent with and in furtherance of the requirements of 10 CFR 50.64.

However, the ERI staff also finds that there are several issues identified during the review of this submittal that need to be more acceptably resolved by the licensee prior to submittal of an application supporting the use of LEU fuel. These are currently identified as conversion open items and summarized in Table 3-1.

Table 3-1 Conversion Open Items Open Item Description

1. Resolve the effect of fuel plate mass difference on seismic analysis as part of fuel qualification program.

2. Provide a bounding limiting core configuration (LCC) for the HEU transition to LEU to support the T&H and safety analysis.

3. Revise the neutronics parameters using the LCC.

4. Ensure that the MCNPX model is suitably predictive before using it to analyze the LCC.

5. Utilize the results from the Brown analysis to ensure that the power distribution used for T&H and safety analysis is suitably conservative.

6. It is unclear whether the limiting power distribution is used to populate the RELAP5 model and determine the acceptability of the NBSR under the stated conditions.

7. The statement in one of the responses that there is no possibility of an inadvertent criticality is insufficient to satisfy the TS 3.9.1 criterion.

8. The licensee needs to revise the dose results for the long duration releases to reflect the maximum dose in a sector at a given distance from the point of release.

In addition, the ERI staff also finds that certain issues have arisen because of the review of the conversion PSAR, but which are broader than conversion. These are currently identified as general open items and summarized in Table 3-2.

69

ERI/NRC 18-204 Table 3-2 General Open Items Open Item Description

1. The decision to not perform a single failure analysis means that the analysis is not compliant with the NUREG-1537 guidance.

2. The licensee has the responsibility to notify the NRC staff regarding the erroneous LOCA analysis under 10 CFR 50.9.

3. The licensee is making changes to their LOCA instrumentation and system behavior under 10 CFR 50.59. The ERI concern is that our interpretation of the regulation indicates that 10 CFR 50.90 is the appropriate venue for these changes.

4. The LOCA SOE does not identify the time for ECT actuation, it does not address the question of whether operator action is or is not required, and it does not address the timing of when such operator actions can be achieved.

5. The time scale for the LOCA analysis does not show a complete SOE and the long-term cooling (draining of the IRT, switchover to the ECT, and initiation of sump recirculation) is not indicated.

6. The cooling of the fuel plates is based upon injection flow that is highly asymmetric. The PSAR does not provide any test data supporting the analysis of fuel plate cooling. There is also no analysis showing whether there is the potential for steam binding that can impact film cooling.

7. The final stabilized condition cited in NUREG-1537 post LOCA is not acceptably described and may not be achievable.

8. It is not clear that the vaporization of the tritiated water is acceptably considered for its effect on evacuation alarms and their effect on the LOCA SOE.

9. The technical specifications NBSR do not satisfy the requirements of 10 CFR 50.36.

• There is no cooldown period identified for leaving the facility unmanned.

• The IRT has no LCO or SR.

• The ECT has no LCO/SR for the volume requirements.

• TS 4.3.2(1) requires valves to be exercised but there is no corresponding LCO.

• TS 3.3.2(2) requires a source for ECT makeup but there is no LCO/SR identifying the sump pumps that provide such service or their requirements.

• The basis for TS 3.3.2 identifies service water as an alternative to the sump pumps, but there is no LCO/SR for its use.

• There are no TSs controlling the dismantling of fuel.

• There are no TSs requiring that irradiated experiments be limited so that the consequences of experiment failure are bounded by the MHA 70

ERI/NRC 18-204

4. REFERENCES

1. Conversion Preliminary Safety Analysis Report for the NIST Research Reactor, BNL-107265-2015-IR, dated December 30, 2014, submitted January 7, 2015, ADAMS Accession No. ML15028A135.

2. Safety Analysis Report for the National Institute of Standards and Technology Reactor -

NBSR, NBSR 14, April 9, 2004, ADAMS Accession No. ML041120167.

3. Safety Evaluation Report Related to the Renewal of Facility Operating License No. TR-5 for the National Bureau of Standards Test Reactor, National Institute of Standards and Technology, June 2009, ADAMS Accession No. ML090990135.

4. Safety Analysis Report for the National Institute of Standards and Technology Reactor -

NBSR, NBSR 14, Revision 6, May 10, 2013.

5. NIST-NRC, National Institute of Standards and Technology- Staff Assessment of Applicability of Fukushima Lessons Learned to the National Institute of Standards and Technology Center for Neutron Research Test Reactor, January 11, 2017, ADAMS Accession No. ML15153A407.

6. Staff Evaluation of Applicability of Lessons Learned From the Fukushima Dai-Ichi Accident to Facilities Other Than Operating Power Reactors, SECY-15-0081, June 9, 2015, ADAMS Accession No. ML15050A066.

7. Guidelines for Preparing and Reviewing Applications for the Licensing of Non-Power Reactors, Standard Review Plan and Acceptance Criteria, NUREG 1537, Parts 1 and 2, U.S. Nuclear Regulatory Commission, February 1996.

8. U.S. Nuclear Regulatory Commission-NIST, National Institute of Standards and Technology - Preliminary Safety Analysis Report for the National Bureau of Standards Reactor (TAC No. MF7235), April 25, 2016, ADAMS Accession No. ML16103A140.

9. U.S. Nuclear Regulatory Commission-NIST, National Institute of Standards and Technology - Preliminary Safety Analysis Report for the National Bureau of Standards Reactor (TAC No. MF7235), August 10, 2017, ADAMS Accession No. ML17187B012.

10. NIST-NRC, Response to Request for Additional Information on Preliminary Safety Analysis Report (TAC no. MF7235), July 21, 2016, ADAMS Accession No. ML16211A064, including:

a. Response to RAls

b. Report BNL-99145-2013-IR, Brown, Hanson, and Diamond (RAI #2 and #12)

c. Sudo and Kaminaga CHF paper (1993)

d. Kaminaga, Yamamoto, and Sudo technical report on CHF Correlation (1998)

e. Saha and Zuber paper on vapor generation (1974)

f. RELAP5 input

g. HOTSPOT output

h. NBS 1980 71

ERI/NRC 18-204

11. NIST-NRC, Response to Request for Additional Information on Preliminary Safety Analysis Report (TAC no. MF7235), October 5, 2017, ADAMS Accession No. ML17284A181.

12. M. Libby, R. Karimi and M. Khatib-Rahbar (ERI) to X. Yin (NRC), Conversion Open Items Based on the Review of the NIST PSAR, ERI/NRC 2018-05-04 COI, May 4, 2018.

13. M. Libby, R. Karimi and M. Khatib-Rahbar (ERI) to X. Yin (NRC), General Open Items Based on the Review of the NIST PSAR, ERI/NRC 2018-05-04 GOI, May 4, 2018.

14. Bess, J., "September 2011 Status Update for the NRAD Reactor Benchmark Models,"

Proceedings of the TRTR Meeting, Idaho Falls, ID, 2011.

15. Brown et al, Local Burn-Up Effects in the NBSR Fuel Element, BNL-99145-2013-IR, January 2013.

16. RELAP5 MOD3.3 supporting documents, (ADAMS Accession No. ML110330203):

a. RELAP5/MOD3.3 Code Manual Volume I: Code Structure, System Models, and Solution Methods, NUREG/CR-5535/Rev P4-Vol I.

b. RELAP5/MOD3.3 Code Manual Volume II: Users Guide and Input Requirements, NUREG/CR-5535/Rev P4-Vol II.

c. RELAP5/MOD3.3 Code Manual Volume II: Appendix A Input Requirements, NUREG/CR-5535/Rev P4-Vol II App. A.

d. RELAP5/MOD3.3 Code Manual Volume III: Developmental Assessment Problems, NUREG/CR-5535/Rev P4-Vol III.

e. RELAP3.3 MOD3.3 Code Manual Volume IV: Models and Correlations, NUREG/CR-5535/Rev P4-Vol IV.

f. RELAP5/MOD3.3 Code Manual Volume V: Users Guidelines, NUREG/CR-5535/Rev P4-Vol V.

g. RELAP5/MOD3 Code Manual Volume 6: Validation of Numerical Techniques in RELAP5/MOD3.0, NUREG/CR-5535/Rev P3-Vol VI.

h. RELAP5/MOD3.3 Code Manual Volume VII: Summaries and Reviews of Independent Code Assessment Reports, NUREG/CR-5535/Rev P4-Vol VII.

i. RELAP5/MOD3.3 Code Manual Volume VIII: Programmers Manual,

j. NUREG/CR-5535/Rev P4-Vol VIII.

17. TRACE V5.0 Theory Manual, U.S. Nuclear Regulatory Commission (Undated).

18. Baek, Cheng, Diamond, Analysis of Loss-of-Coolant Accidents in the NBSR, BNL-105287-2014-IR, May 2014.

19. Baek, Cheng, Diamond, LOCA Analysis for the NIST Research Reactor, Transactions of the American Nuclear Society, Vol. 116, San Francisco, California, June 11-15, 2017.

20. Meeting at NIST Site, December 6, 2017.

72

ERI/NRC 18-204

21. National Nuclear Data Center, Evaluated Nuclear Data File (ENDF) Retrieval & Plotting, Periodic Table Browse, ENDF/B-VII.1 (2011) Library, cumulative fission yield, Brookhaven National Laboratory, Accessed on April 2, 2018, at http://www.nndc.bnl.gov/sigma/index.jsp?as=235&lib=endfb7.1&nsub=11.

22. Safety Analysis Report (SAR) for License Renewal for the National Institute of Standards and Technology Reactor - NBSR; NSBR 14, Rev 4, National Institute of Standards and Technology (NIST), Gaithersburg, MD, 2010a.

23. Homann, S.G. and Aluzzi, F., "HotSpot - Health Physics Codes, Version 3.0, User's Guide,"

LLNL-SM-636474, Lawrence Livermore National Laboratory, August 27, 2014.

24. Regulatory Guide 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors, USNRC, July 2000.

25. Regulatory Guide 1.145, Revision 1, "Atmospheric Dispersion Models for Potential Accident Consequence Assessments at Nuclear Power Plants," USNRC, November 1982.

26. NUREG/CR-2079, "Analysis of Credible Accidents for Argonaut Reactors," PNL-3691, April 1981.

73

ERI/NRC 18-204 Intentionally left blank 74