ML20030A176

NRC Non-LWR Water Reactor Vision Strategy; Volume 1 - Computer Code Suite for Non-LWR Plan Systems Analysis

ML20030A176
Person / Time
Issue date:	01/31/2020
From:	Office of Nuclear Regulatory Research
To:
Webber, Kimberly
Shared Package
ML20030A171	List: ML20030A172 ML20030A174 ML20030A176 ML20030A177 ML20030A178
References
LTR-19-0438-1
Download: ML20030A176 (118)

Text

Revision 1 January 31, 2020 NRC Non-Light Water Reactor (Non-LWR)

Vision and Strategy, Volume 1 - Computer Code Suite for Non-LWR Plant Systems Analysis

This page intentionally blank .

2

TABLE OF CONTENTS LIST OF FIGURES ....................................................................................................................... 4 LIST OF TABLES.......................................................................................................................... 5 NOMENCLATURE ........................................................................................................................ 6 EXECUTIVE

SUMMARY

.............................................................................................................. 9 1.0 Introduction ...................................................................................................................... 10 1.1 Regulatory Application .............................................................................................. 11 2.0 Advanced non-LWR Code Considerations ...................................................................... 13 2.1 Physical Phenomena and Modeling Requirements .................................................. 16 2.2 Multi-Physics Environment Needs .......................................................................... 19 2.3 Code Development Costs and Potential for Cost Sharing ........................................ 19 2.4 Review Schedule Readiness .................................................................................... 20 2.5 Computational Resource Requirements ................................................................... 21 2.6 Impacts on NRC Staff ............................................................................................... 21 3.0 BlueCRAB Code Suite for non-LWRs .............................................................................. 23 3.1 Discussion ................................................................................................................. 27 3.2 Summary of Code Pros and Cons .......................................................................... 39 4.0 Code Suite by Reactor Type ............................................................................................ 45 4.1 Characterization of Code Readiness ........................................................................ 47 4.2 Gas-Cooled Reactors................................................................................................ 49 4.3 Liquid Metal (Fast) Reactors ..................................................................................... 68 4.4 Heat Pipe Cooled Micro Reactors .......................................................................... 79 4.5 Molten Salt Reactors ................................................................................................. 87 4.6 Initial MSR System Inventory .................................................................................. 112 5.0 Summary ........................................................................................................................ 114 6.0 References ..................................................................................................................... 115 3

LIST OF FIGURES Figure 2-1. Code Development Process. ................................................................................... 14 Figure 3-1. The Comprehensive Reactor Analysis Bundle (CRAB) for Analysis of Design Basis Events in non-LWRs. .............................................................................................. 24 Figure 3-2. Example of control rod region from a small SFR. ................................................... 31 Figure 3-3. Example of local finite element mesh for a SFR control rod assembly. ................. 31 Figure 3-4. Tangential stress prediction by BISON in a defective TRISO particle. ................... 34 Figure 3-5. Velocity vectors in the square cavity at = 100000. ............................................ 37 Figure 3-6. Nek5000 prediction of velocities in a random grouping of pebbles. ....................... 38 Figure 4-1. Model of a prismatic HTGR fuel lattice with control rod inserted in the central fuel block using SERPENT. .......................................................................................... 50 Figure 4-2. Modeling Approach for Design Basis Event Simulation in a Prismatic HTGR........ 51 Figure 4-3. Modeling Approach for Design Basis Event Simulation in a Pebble Bed Gas-Cooled Reactor. ..................................................................................................... 58 Figure 4-4. Analysis Approach for Sodium Fast Reactors. ....................................................... 70 Figure 4-5. Model of a Heat Pipe Cooled Reactor for Design Basis Event Analysis. ............... 82 Figure 4-6. Modeling Approach for Design Basis Event Simulation in a molten salt reactor with stationary fuel. ........................................................................................................ 88 Figure 4-7. AHTR Fuel Assembly. ............................................................................................ 91 Figure 4-8. Modeling Approach for Design Basis Event Simulation in a Pebble Bed Salt-Cooled Reactor. .................................................................................................................. 95 Figure 4-9. Modeling Approach for Design Basis Event Simulation in a Molten Fuel Salt Reactor. ................................................................................................................ 101 Figure 4-10. Modeling Approach for Design Basis Event Simulation in a Molten Chloride Fuel Salt Reactor.......................................................................................................... 107 Figure 4-11. Inventory Control Processes in a Molten Salt Reactor. ........................................ 113 4

LIST OF TABLES Table 3-1. Summary of Code Considerations....................................................................... 39 Table 4-1. Generalized Listing of non-LWR Designs ........................................................... 46 Table 4-2. Predictive Capability Maturity Model (PCMM) Matrix .......................................... 48 Table 4-3. PCMM Characterization of Codes for HTGR Analysis ........................................ 56 Table 4-4. PCMM Characterization of Codes for PBMR Analysis ........................................ 64 Table 4-5. PCMM Characterization of Codes for GCFR Analysis ........................................ 67 Table 4-6. PCMM Characterization of Codes for SFR Analysis. .......................................... 74 Table 4-7. PCMM Characterization of Codes for LMR Analysis ........................................... 78 Table 4-8. PCMM Characterization of Codes for HPR Analysis ........................................... 86 Table 4-9. PCMM Characterization of Codes for MSR Analysis .......................................... 93 Table 4-10. PCMM Characterization of Codes for MSPR Analysis ........................................ 99 Table 4-11. PCMM Characterization of Codes for MFSR Analysis ...................................... 105 Table 4-12. PCMM Characterization of Codes for MCSR Analysis ...................................... 111 5

NOMENCLATURE Abbreviation Definition AGREE Advance Gas Reactor Evaluator AHTR Advanced High Temperature Reactor ANL Argonne National Laboratories A&E Aleatory and Epistemic AOO Anticipated Operational Occurrence ATWS Anticipated Transient Without Scram AVR Arbeitsgemeinschaft Versuchsreaktor BDBE Beyond Design Basis Event CASL Consortium of Advanced Simulation of LWRs CEFR Chinese Experimental Fast Reactor CFD Computational Fluid Dynamics CFR Code of Federal Regulations CIET Compact Integral Effects Test CPU Central Processing Unit CR Control Rod CRAB Comprehensive Reactor Analysis Bundle CRBR Clinch River Breeder Reactor D-LOFC Depressurized Loss of Forced Cooling DBE Design Basis Event DOE U.S. Department of Energy ECCS Emergency Core Cooling System EBR Experimental Breeder Reactor EM Evaluation Model FAST Fuel Analysis under Steady-State & Transient FFTF Fast Flux Test Facility FRAPCON Fuel Rod Analysis Program Constant FRAPTRAN Fuel Rod Analysis Program Transient GA General Atomics GCR Gas-Cooled Reactor GDC General Design Criteria 6

Abbreviation Definition HELIOS Heavy Eutectic liquid metal Loop for Integral test of Operability and Safety HPC High Performance Computing HPR Heat Pipe Reactor HTR High Temperature Gas-cooled Reactor Test Module HTGR High-Temperature Gas-Cooled Reactor HTTR High Temperature Test Reactor HTTU High Temperature Test Unit IAP Implementation Action Plan INL Idaho National Laboratories KRUSTY Kilopower Reactor Using Stirling Technology LACANES Lead Alloy-Cooled Advanced Nuclear Energy Systems LES Large Eddy Simulation LMR Liquid Metal Reactor LOFC Loss of Forced Cooling LWR Light Water Reactor MHTGR Modular High Temperature Gas-Cooled Reactor MOOSE Multiphysics Object Oriented Simulation Environment MCNP Monte Carlo N-Particle Transport Code MCSR Molten Chloride Salt Reactor MFSR Molten Fluoride Slat Reactor M&S Modeling and Simulation MSPR Molten Salt Pebble Bed Reactor MSR Molten Salt Reactor MSRE Molten Salt Reactor Experiment NEAMS Nuclear Energy Advanced Modeling and Simulation NGNP Next Generation Nuclear Plant NRC U.S. Nuclear Regulatory Commission NSTF Natural convection Shutdown heat removal Test Facility ORNL Oak Ridge National Laboratories P-LOFC Pressurized Loss of Forced Cooling PARCS Purdue Advanced Reactor Core Solver PARFUME Particle Fuel Model 7

Abbreviation Definition PCMM Predictive Capability Maturity Model PIRT Phenomena Identification and Ranking Tables PRA Probabilistic Risk Assessment PRIME Plume Rise Model Enhancements PRISM Power Reactor Innovative Small Module PTS Pressurized Thermal Shock RANS Reynolds Averaged Navier Stokes RCCS Reactor Cavity Cooling System RIS Regulatory Issue Summary SAM System Analysis Module SANA Selbsttatige Abfuhr der Nachwarme SCALE Standardized Computer Analyses for Licensing Evaluation SFR Sodium Fast Reactor SMR Small Modular Reactor SNAP Symbolic Nuclear Analysis Package SNL Sandia National Laboratories SPH Super-Homogenization SSC Safety Significant Component TAMU Texas A&M University TBD To Be Determined THTR Thorium High Temperature Reactor TRACE TRAC/RELAP Advanced Computational Engine TREAT Transient Test Reactor TRISO Tri-isotropic U/PLOF Unprotected/Protected Loss of Flow U/PLOHS Unprotected/Protected Loss of Heat Sink UQ Uncertainty Quantification UW University of Wisconsin VERA Virtual Environment for Reactor Applications VCS Vessel Cooling System VHTRC Very High-Temperature Reactor Critical ZPPR Zero Power Physics Reactor 8

EXECUTIVE

SUMMARY

This report discusses the code suite proposed for non-light water reactor (non-LWR) systems safety analysis. The Comprehensive Reactor Analysis Bundle (CRAB or alternately BlueCRAB) described in this report can be used for safety analysis of all reactor designs, both light water and non-light water. There are several non-LWR design types for which the Nuclear Regulatory Commission (NRC) needs to be ready to support safety analysis and licensing. Design types include gas-cooled reactors, liquid-metal cooled reactors, molten salt cooled reactors, molten fuel salt reactors, and micro reactors cooled by heat pipes. The CRAB code suite makes use of a coupled computational environment utilizing NRC and DOE codes. The NRC codes will continue to be used for LWR analysis as they have in the past. The set of codes that comprise CRAB apply to all reactor designs regulated by the NRC along with those proposed by non-LWR applicants.

The CRAB code suite can be used for evaluating normal operating conditions and margins to safety limits for accident scenarios not involving gross core disruption and fission product release. The CRAB codes can also be used to simulate accident scenarios of importance to advanced non-LWRs and evaluate the importance and integrity of system and component barriers to fission gas release. The type of safety analyses performed with CRAB will be dependent upon the needs of the regulatory office and could include evaluations of the failure of passive systems and components, the resulting consequences, the safety margins and key uncertainties.

This report describes each of the codes involved, and how they would be applied to each of the principle non-LWR design types. Development status and maturity of the codes are also discussed. Perhaps most important, this report summarizes the tasks necessary to resolve gaps in simulation capabilities. Issues involving code selection, computational hardware requirements, and training needs are discussed.

9

1.0 Introduction As the U.S. Nuclear Regulatory Commission (NRC) prepares to review and regulate a new generation of non-light water reactors (non-LWRs), a vision and strategy has been developed to assure NRC readiness to efficiently and effectively conduct its mission for these technologies, including fuel cycles and waste forms. In December 2016, the NRC published the vision and strategy document for public comment in the Federal Register [1]. The non-LWR vision and strategy document provides a connection to other NRC mission, vision, and strategic planning activities, and describes the objectives, strategies, and contributing activities necessary to achieve non-LWR mission readiness.

The non-LWR vison and strategy approach consists of six different strategies. The main objectives of Strategy 2 are to identify and develop the tools and database that will enable the staff to perform it technical review of a non-LWR application. Central to Strategy 2 are the selection and development of computer codes that may be used to analyze non-LWRs.

One of the primary objectives of Strategy 2 of the Implementation Action Plan (IAP) for non-LWRs is the development of codes suitable for safety analysis of gas-cooled reactors, sodium fast reactors, molten salt reactors, and micro reactors. Modeling and simulation of these designs involve numerous physical processes that generally do not occur in light-water reactors.

These phenomena are discussed in Section 2.1. Therefore, initial efforts have been directed at understanding requirements for modeling and simulation of these new designs, and in identifying codes that either meet or could meet these requirements. Codes used traditionally by the NRC for confirmatory analysis have been developed and assessed for light water reactors and are not immediately extendable to these non-LWR designs. For some designs the NRC codes could be made applicable. However, for other non-LWR designs the NRC codes are not well suited to modeling the dominant phenomena of interest without considerable effort.

While development and modification of NRC codes is a possibility, codes developed outside of the NRC were placed under consideration. In particular, the codes developed under the NEAMS (Nuclear Energy Advanced Modeling and Simulation) possess unique modeling capabilities for non-LWR models making them more readily amenable to non-LWR analysis by the NRC.

The CRAB codes discussed in this report are those intended for plant systems analysis and assessing the adequacy of the emergency core cooling features of a design. Advanced non-LWRs are expected to have fewer active safety systems and rely on passive safety features.

These codes are being prepared to help staff conduct safety analysis to verify safety margins and allowable operational limits. Included are codes for neutronics, fuel performance, and thermal-hydraulics. While the set of accident scenarios for each non-LWR design has not been fully defined and may vary considerably by design type, it is assumed in this report that codes in the CRAB code suite will be used for accident scenarios where there is little or no fission product release from the primary system and the core and vessel geometry remains intact.

Plant systems analysis is expected to show adequacy of safety systems and adequacy of safety significant features of the design. For several of the designs the accident scenarios of most interest involve the loss of forced cooling, or the loss of a heat removal system (either secondary side cooling to a reactor cavity cooling system). Neutronic events of interest are 10

assumed to be those that involve a reactivity insertion or failure of control rods to insert when required.

This report discusses code options for non-LWRs and proposes a set of codes for safety analysis. Because of the large number of design options that the staff may need to review, the designs are discussed in terms of ten general design types. The set of codes consists of a combination of NRC and DOE codes so that a broad range of designs can be simulated, but without introducing such a large number of new codes that the NRC staff becomes overburdened with an excessive amount of training. The proposed set of codes were selected by evaluating against several important considerations. These considerations are discussed in Section 2.

Section 3 of this report provides a description of each code that is proposed for plant systems analysis. The codes proposed are expected to require further development and assessment but have been found to contain most of the fundamental capabilities to begin validation and application to non-LWR designs. Section 3 also includes a discussion on the relative merits of selecting these codes over other alternatives. The proposed set of codes thus serves as a starting point and provides a long-term vision for non-LWR computational analysis.

The readiness of codes for a particular application is presented in Section 4. The state of readiness is characterized using the Predictive Capability Maturity Model (PCMM), which examines six elements associated with code accuracy and credibility. Section 4 describes the PCMM, and then is used to provide an estimate of the current readiness for each of the ten general design types. For each of the designs, a scheme for modeling and simulation is presented. Further, for each design a series of code development and code assessment tasks are listed. The code development and assessment tasks represent, in effect, a gap analysis of necessary work that is needed to make the code(s) applicable to a design with what is expected to be sufficient accuracy for a confirmatory review.

1.1 Regulatory Application Design basis confirmatory analysis is generally used to help the staff evaluate an applicants analysis to show that the regulatory criteria are satisfied. Regulatory Guide 1.232 [2] provides guidance on how the general design criteria (GDC) in Appendix A, General Design Criteria for Nuclear Power Plants, of Title 10 of the Code of Federal Regulations (10 CFR) Part 50, Domestic Licensing of Production and Utilization Facilities, may be adapted for non-light-water reactor (non-LWR) designs. Design Basis Event (DBE) analysis addresses several of these design criteria, generally related to Thermal-Fluid Design, Nuclear Design, and Engineered Safety Systems. For existing LWRs, the general Design Criteria that are the focus of DBE analysis include, but are not limited to:

GDC 10: Reactor design. The reactor core and associated coolant, control, and protection systems shall be designed with appropriate margin to assure that specified acceptable fuel design limits are not exceeded during any condition of normal operation, including the effects of anticipated operational occurrences.

11

GDC 11: Reactor inherent protection. The reactor core and associated systems that contribute to reactivity feedback shall be designed so that, in the power operating range, the net effect of the prompt inherent nuclear feedback characteristics tends to compensate for a rapid increase in reactivity.

GDC 12: Suppression of reactor power oscillations. The reactor core; associated structures; and associated coolant, control, and protection systems shall be designed to ensure that power oscillations that can result in conditions exceeding specified acceptable fuel design limits are not possible or can be reliably and readily detected and suppressed.

GDC 15: Reactor coolant system design. The reactor coolant system and associated auxiliary, control, and protection systems shall be designed with sufficient margin to ensure that the design conditions of the reactor coolant boundary are not exceeded during any condition of normal operation, including anticipated operational occurrences.

GDC 26: Reactivity control systems. A minimum of two reactivity control systems or means shall provide: (1) A means of inserting negative reactivity at a sufficient rate and amount to assure, with appropriate margin for malfunctions, that the design limits for the fission product barriers are not exceeded and safe shutdown is achieved and maintained during normal operation, including anticipated operational occurrences. (2) A means which is independent and diverse from the other(s), shall be capable of controlling the rate of reactivity changes resulting from planned, normal power changes to assure that the design limits for the fission product barriers are not exceeded. (3) A means of inserting negative reactivity at a sufficient rate and amount to assure, with appropriate margin for malfunctions, that the capability to cool the core is maintained and a means of shutting down the reactor and maintaining, at a minimum, a safe shutdown condition following a postulated accident. (4) A means for holding the reactor shutdown under conditions which allow for interventions such as fuel loading, inspection and repair shall be provided.

GDC 28: Reactivity limits. The reactivity control systems shall be designed with appropriate limits on the potential amount and rate of reactivity increase to ensure that the effects of postulated reactivity accidents can neither (1) result in damage to the reactor coolant boundary greater than limited local yielding nor (2) sufficiently disturb the core, its support structures, or other reactor vessel internals to impair significantly the capability to cool the core.

GDC 34: Residual heat removal. A system to remove residual heat shall be provided. For normal operations and anticipated operational occurrences, the system safety function shall be to transfer fission product decay heat and other residual heat from the reactor core at a rate such that specified acceptable fuel design limits and the design conditions of the reactor coolant boundary are not exceeded. Suitable redundancy in components and features and suitable interconnections, leak detection, and isolation capabilities shall be provided to ensure that the system safety function can be accomplished, assuming a single failure.

GDC 35: Emergency core cooling system. A system to assure sufficient core cooling during postulated accidents and to remove residual heat following postulated accidents shall be provided. The system safety function shall be to transfer heat from the reactor core during and 12

following postulated accidents such that fuel and clad damage that could interfere with continued effective core cooling is prevented.

Review of advanced non-LWRs is intended to be risk-informed and performance-based such that the figure(s) of merit for a design are not necessarily specific. To reduce potential difficulty and regulatory uncertainty, the NRC has engaged with stakeholders on two projects relevant to modeling and simulation of non-LWRs; the Licensing Modernization Project (LMP) [3], and the Advanced Reactor Design Criteria (ARDC) [2].

The Licensing Modernization Project is being led by Southern Company, coordinated by the Nuclear Energy Agency (NEI), cost-shared by the Department of Energy (DOE) with active participation by the NRC. The objective of the LMP is to develop technology-inclusive, risk-informed, and performance based regulatory guidance for licensing non-LWRs. The LMP describes a systematic process for identifying and categorizing event sequences as anticipated operational occurrences (AOOs), design basis events (DBEs), or beyond-design-basis events (BDBEs). The primary determinate for categorizing events is the estimated frequency of the event sequence, rather than a deterministic characterization of events. For non-LWRs, the applicant may (but is not required) use the Licensing Modernization Project (LMP). Under the LMP, a Probabilistic Risk Assessment (PRA) is developed to estimate the frequency and consequence of various accident scenarios. The acceptability of a design may depend on the margin between the predicted dose at the site boundary and the proposed frequency -

consequence limit curve. The LMP blurs the traditional distinction between Chapter 15 design basis and Chapter 19 beyond design basis events as some scenarios for non-LWRs may involve multiple failures and yet no result in a source term. Other events, possibly with frequencies >10-6 may have a significant source term. The need and type of safety analysis necessary may therefore depend on the simplicity or complexity of the design, the potential accident scenario and margins to design limits.

For the present report, DBE analysis assumes that at the end of an event the core remains essentially intact while in a BDBE there has been significant core disruption and release of fission products. Because of large anticipated safety margins in non-LWR designs, some events classified as DBE may involve multiple failures in safety functions and still not result in a source term. Thus, DBE analysis may be better characterized as plant systems analysis as the purpose is to confirm the adequacy of safety systems and acceptability of proposed operating conditions. Such an analysis thus ensures that the ARDCs are met and that adequate safety margins do in fact exist in a design.

Depending on the design, acceptance limits for non-LWRs may involve constraints such as maximum vessel wall temperature, the eutectic temperature for metallic fuel, coolant boiling, or avoidance of power excursions. Satisfying these limits help ensure that an accident does not become more severe and otherwise result in a source term release.

2.0 Advanced non-LWR Code Considerations Because of the wide variety of non-LWR designs currently being proposed, the set of codes to be used by the NRC for non-LWR safety analysis must be flexible and relatively straightforward to use. The code suite must be able to accommodate numerous fuel types, core configurations, 13

safety systems, and other design features. The codes selected or developed by the NRC for non-LWR applications have several other requirements and restrictions. For codes not currently used by the NRC, they must be able to execute within the NRCs existing or expected computing environment. New codes represent a challenge to NRC staff because of unfamiliarity and the sometimes steep learning curve that must be overcome. Most important however, codes for non-LWRs must be capable of simulating the physical phenomena associated with the design and important accident scenarios. In some cases, a multi-physics capability may be needed thus requiring a tighter coupling between analysis modules (i.e.,

neutronics, thermal-fluids, fuel performance, and thermo-mechanical effects) than is typical in LWR analysis. The following sub-sections discuss these considerations.

Code development at the NRC generally follows the approach outlined in Regulatory Guide 1.203, Transient and Accident Analysis Methods [4]. Regulatory Guide 1.203 was written under the assumption that it would be for light-water reactors but is general and can be extended to non-LWRs. It is based on the Evaluation Model (EM) concept and establishes the basis for methods used to analyze a particular event or class of events. An evaluation model (EM) is the calculational framework for evaluating the behavior of the reactor system during a postulated transient or design-basis accident. As such, the EM may include one or more computer programs, special models, and all other information needed to apply the calculational framework to a specific event. Figure 2-1 is a flow diagram of the EM development process.

The EM development process is relatively straightforward. However in the case of non-LWRs, there are two initial considerations that both applicants and the NRC will need to address. One Figure 2-1. Code Development Process.

14

of the first steps is identification of the event scenarios that the EM will be applicable to. These event scenarios should be defined at the initiation of the code development process. Central to code development is a clear understanding of the physical phenomena and associated level of fidelity for those phenomena that must be simulated by the codes within the EM. A Phenomena Identification and Ranking Table (PIRT) is therefore important step at the initiation of an EM effort [5] and development of the PIRT requires identification of event scenarios. For LWRs, the accident scenarios are well-known and colloquially defined as Chapter 15 and Chapter 19 events for DBE and BDBE respectively. However, for non-LWRs following the Licensing Modernization Project approach, PRA insights are necessary. Currently, specific sets of accident scenarios are not available for many non-LWR designs.

Event scenarios for non-LWRs are expected to differ from traditional Chapter 15 and Chapter 19 events due to their unique characteristics and the expected large safety margins. The new designs are likely to be able to withstand multiple failures and/or require fewer safety significant components (SSCs) to mitigate an accident. The analysis capability being defined in this report is intended to simulate events in non-LWRs up to those conditions that lead to core disruption and significant release of fission products. While a scenario involving multiple failures is traditionally considered a beyond design basis event, for non-LWRs these scenarios may be classified as design basis events in addition to large safety margins and lack of a source term.

The event scenarios that the codes in this report will be used for will include, but are not limited to the following hypothetical events in gas-cooled, liquid metal cooled, molten salt, and micro reactors:

Gas-Cooled Reactors

• pressurized loss of force cooling (P-LOFC) accident

• de-pressurized loss of force cooling (D-LOFC) accident

• reactivity-induced transients, including ATWS events Events that involve air-ingress and significant oxidation of the graphite, water-ingress, transport and release of graphite dust will be simulated by Volume 3 analysis tools since they involve a significant source term.

Liquid Metal Reactors

• loss of coolant with and without scram

• loss of forced flow

• unprotected loss of flow

• unprotected loss of heat sink

• reactivity-induced transients, including ATWS events Molten Salt Reactors

• loss of forced flow

• unprotected loss of flow

• inadvertent reactivity insertion transients, including ATWS events 15

• loss of coolant

• over-cooling events (leading to partial solidification)

• station blackout

• loss of heat sink Heat Pipe Cooled Micro Reactors

• loss of heat sink

• inadvertent reactivity insertion transients, including ATWS events

• localized heat pipe failure

• cascading loss of heat pipes

• events related to coupling the reactor to the power conversion unit

• monolith temperature and stress under normal operating conditions

• monolith temperature and stress under postulated accident conditions.

Further, external events, such as seismic events, may have different, relevant effects on the designs (such as reactivity changes), and may be considered as appropriate for the codes discussed in this Volume.

2.1 Physical Phenomena and Modeling Requirements First and foremost, the codes must either possess the capability for simulation of non-LWR physical phenomena or be amenable to development and enable the necessary models and features at an adequate level of fidelity for safety analysis. The NRC is expecting a wide variety of design types to be submitted for licensing, and these can be classified into four overall design types: gas-cooled, liquid metal cooled, molten salt, and micro reactors. Phenomena important to each design have been identified for each of these design types, and in most cases applicable PIRTs are available. This section discusses the references used by the NRC to help establish code requirements for non-LWRs.

Phenomena for gas-cooled reactors, both prismatic and pebble bed, were the subject of a comprehensive effort related to the Next Generation Nuclear Plant (NGNP). Expert panels considered five interrelated subject areas including thermal-fluids, neutronics and accident analysis, high temperature materials, nuclear grade graphite, process heat and cogeneration, and fission product transport [6]. With regards to design basis event analysis, phenomena and processes of major concern due to their importance in analysis and relatively low knowledge level included:

• core coolant bypass flows,

• power/flux profiles,

• outlet plenum flows,

• reactivity-temperature feedback coefficients,

• emissivity for the vessel and reactor cavity cooling system,

• reactor vessel cavity air circulation and heat transfer, and

• convection/radiation heating of upper vessel.

16

A separate PIRT was also developed for TRISO fuel [7], which discussed mechanisms and phenomena for fission product release and the evaluation of fuel performance. This report is relevant to plant systems analysis due to its identification of fission product release mechanisms. These failure mechanisms help establish conditions where significant fission product release may begin.

For sodium fast reactors, there has been significant efforts made to understand the phenomena and processes of importance to modeling and simulation of various accident scenarios. While not specifically a PIRT, a recent study into technology gaps was completed by a panel of experts [8]. The objective of the study was to evaluate the status of knowledge for accident analysis of sodium fast reactors (SFRs), and the panel considered a broad range of phenomena expected to be important in licensing SFR designs. Because of a fairly large quantity of experimental information, the panel was able to conclude that the knowledge level for SFRs was good and that there are no major technology gaps in preparing a safety case for an advanced SFR, so long as the design stays with known technology. Potential gaps were acknowledged for advanced simulation of coupled neutronic/fluid flow dynamics, supercritical CO2 power conversion, and high minor-actinide content fuel.

The study identified several phenomena important to plant systems analysis of SFRs, including:

• single phase transient sodium flow

• thermal inertia

• pump-coast down

• sodium stratification

• transition to natural convection core cooling core flow

• decay heat generation

• reactivity due to mechanical changes in core structure

• reactivity feedback at high power The NRC sponsored an additional review that considered sodium fast reactors, lead- and lead-bismuth cooled systems [9]. The findings in [8] were confirmed, and largely extended to other liquid metal designs, and developed a comprehensive list of phenomena and processes that must be modeled in analysis of a liquid metal reactor.

While important phenomena and gaps in the technologies are well established for gas-cooled and liquid metal reactors, molten salt reactors represent a new and unique challenge. PIRTs applicable to these designs have recently been developed. For a molten coolant salt design (with liquid coolant and solid fuel), a PIRT was developed considering two events; station blackout and a simultaneous withdrawal of control rods [10]. An accompanying study [11]

considered phenomena involved in the molten salt reactor design and the need for a tight coupling between some phenomena. The two studies [10, 11] discussed physical phenomena and gaps in knowledge for the particular design which utilized a relatively well known fluoride salt and TRISO fuel in the form of plates. Phenomena of importance which were recognized as having gaps in the knowledge base included:

• thermophysical properties of coolant salt (conductivity and viscosity)

• wall friction in the core 17

• core flow asymmetry

• upper and lower plenum mixing

• safety system component performance

• chimney natural circulation and performance These studies were examined by the NRC, which then decided to sponsor an expert panel to consider molten fuel salt reactors (with fissile material dissolved in the salt coolant). This panel investigated and developed PIRTs for both fluoride (thermal spectrum) and chloride (fast spectrum) fuel salt reactors [12-14]. This study concluded that in addition to the phenomena associated with a coolant salt, there are two new challenges introduced with liquid fuel: delayed neutron precursor motion, and a strong coupling of thermal-fluid and neutronic phenomena to salt composition. Hence, it is necessary to take into account the movement of delayed neutron precursors into and out of the core, and the transit times of the fuel, fission products, and transmutation and chemistry products through the primary system. For a fuel salt reactor design the phenomena of importance included:

• delayed neutron precursor motion

• salt chemical composition

• neutron absorption in fuel salt

• physical properties

• convective heat transfer

• primary system flow resistances

• structural material performance (swelling and expansion)

• tritium production and transport (lithium bearing salts)

A major recommendation of this study [12] is that modeling and simulation of MSRs will require development of computational tool(s) capable of tracking chemical inventories of constituents throughout the primary loop of the reactor facility. A comprehensive evaluation of fuel salt composition will likely require the modeling and simulation of salt chemistry, multi-component thermodynamics, mass transport, and the addition and removal of chemical species. A new and separate analysis tool may need to be developed for this purpose.

Finally, for heat pipe cooled micro reactors, four individual PIRTs have been developed for a special purpose 5 MWt design [15]. The PIRTs are for reactor accident and normal operations, heat pipes, materials, and power conversion. These PIRTs expressed concerns with:

• monolith thermal stress

• single heat pipe failure

• machining and inspection of the monolith

• heat pipe performance

• reactivity and core criticality References 6-15 provide a preliminary basis for modeling and simulation requirements for a non-LWR code suite. These PIRTs will likely need to be reconsidered as each of these reactor designs mature and new experimental information is made available.

18

This sub-section has listed a set of phenomena and gaps that are likely to require efforts to resolve in the development of a safety analysis capability. A comprehensive review of these entire PIRTs are beyond the scope of this report, yet there are some findings that are of particular importance with respect to code requirements and future develop for non-LWRs.

Phenomena that are significant and new with increased importance for non-LWRs relative to conventional LWRs include but are not limited to:

- Thermal stratification and thermal striping

- Thermo-mechanical expansion and effect on reactivity

- Large neutron mean-free path length in fast reactors

- Transport of neutron pre-cursors (in fuel salt MSRs)

- Solidification and plate-out (MSRs)

Codes that simulate non-LWRs need to account for these phenomena as well as many others that are important to design basis events. Development and assessment of these tools will certainly be necessary, but the basic code requirements defined in these PIRTs are satisfied with the code suite discussed in this report.

2.2 Multi-Physics Environment Needs One of the complicating factors for advanced non-LWRs is that it may be necessary to have a tight coupling between several analysis codes because of the feedback between physical phenomena. Fast reactors for example, require a tight coupling between neutronics, thermal-hydraulics, and in some cases thermal-mechanical expansion to account for rapid changes in power due to reactivity feedback that can occur with changes in temperature. Thermo-mechanical expansion of the core is an important phenomenon that must be accounted for because of its impact on neutron leakage. In some designs axial and radial leakage can be the largest source of negative reactivity as core temperature increases.

In general, for accident scenarios four types of codes are necessary: fuel performance, neutronics, thermal-hydraulics, and thermomechanical response. Codes for cross-section determination, mesh generation, and possibly computational fluid dynamics may become part of an Evaluation Model for analysis of these designs. Because of the need to enable feedback between the codes, a computing environment in which codes can be efficiently coupled together may be required for some designs. Coupling codes can also reduce development costs by making use of analysis capabilities that are within the set of coupled codes, but not specifically within a given code. The downside of coupling is that time-step size and control can be limited by the slowest or least computationally efficient code in the set. Thus, when code coupling is expected to be used, careful consideration and testing of coupling methods and data transfer techniques are necessary.

2.3 Code Development Costs and Potential for Cost Sharing There are a considerable number of code options and methods available for modeling and simulation of advanced reactors. None of the codes are likely to be acceptable for safety calculations without additional development and assessment. (There are no Evaluation Models currently approved for non-LWRs by the NRC.) The NRC currently maintains a set of codes 19

suitable for LWR safety analysis. With additional development these tools could be made applicable to non-LWRs. However, the costs associated with development, maintenance and assessment of the NRC codes for the broad range of non-LWR designs is expected to greatly exceed the costs of adopting codes that have received significant development and validation for use with GCRs, SFRs, or MSRs. Micro reactors would be a significant challenge to the NRC codes, since thermal-mechanical response and conduction in a multidimensional structure are processes that cannot be simulated with NRC codes. Selecting suitably developed codes produced outside the NRC can significantly reduce costs to the NRC associated with code development and assessment.

A key point in reducing development costs is developing a cooperative agreement with code developers (for those codes not developed by the NRC) such that the NRC has some influence on development priorities. The NRC may at times need to have specialized revisions to a code, and being able to do this quickly during a review will be an important need. The Office of Nuclear Regulatory Research has a Memorandum of Understanding with DOE which provides NRC access to DOE codes for regulatory purposes. It is the NRCs understanding that DOE will have primary responsibility for generically applicable code development, verification and validation activities to support the NRC.

Adoption of a code from outside the NRC is not without costs however. All new codes are expected to require a period of time for staff to become familiar with them. Training from DOE and/or other external organizations may be necessary. An important part of the process of incorporating new codes into the NRC will be providing staff with opportunities to gain experience with the new codes. Therefore, in the tasks that are described in later sections some of the code assessment, and a major portion of plant analysis will be reserved for staff to perform. This is intended to help ensure staff becomes expert in use of all tools involved.

Total life-cycle costs are considered to be those for code development, verification and validation, code maintenance, and staff training. The DOE codes intended for NRC use are expected to be verified and maintained by DOE, and DOE will be expected to perform most of the generic code development and code validation. Some of the development and assessment related tasks are recommended as training exercises for NRC staff. Validation associated with vendor specific, proprietary data will be performed by the NRC.

2.4 Review Schedule Readiness The expected review schedule is highly uncertain. Several potential applicants have indicated that they intend to make submittals for NRC review as early as 2020 and several applicants have suggested commercial operation before 2025. In order to be ready to conduct safety analysis with aggressive schedules such as these, it is important that the codes selected for systems analysis be essentially complete, and because of this code validation began in 2019.

Especially for the expected early applicants, the codes need to be out of the box ready for additional verification and validation. The DOE codes that will be discussed for NRC use satisfy this out of the box readiness to begin significant verification and validation, so that code accuracy can be determined should an aggressive review schedule be necessary.

20

2.5 Computational Resource Requirements The NRC currently has limited access to high-end computational platforms. If new codes, or the nature of non-LWR analysis, requires exceptionally high geometric or temporal resolution then it may be necessary to improve the capability of existing NRC platforms or gain access to high end resources.

There are three options available to the NRC staff, and all have been tested in 2019. One option is to purchase computing time on the cloud. This approach allows considerable flexibility in the number of processors and available memory. The cloud lets the user select the number of computing cores, and allows the user to optimize resources based on size and complexity of the application. Test cases using existing NRC codes and BlueCRAB have demonstrated feasibility. Using the cloud is new and continual changes to the system present a challenge to staff. However, as experience is gained, using the NRC cloud is expected to become an efficient option.

Another option is to use the high performance computing (HPC) resources at the national laboratories. The Memorandum of Understanding with DOE, and subsequent discussion, has resulted in the NRC gaining access to HPC systems at Idaho and Argonne National Labs. Initial testing by NRC has been acceptable and provides an option to simulate very demanding applications if necessary.

Finally, the non-LWR codes discussed in this Volume can be run on NRC provided desktop platforms. This has been demonstrated and is currently the manner in which most cases are run. NRC laptops are capable of running most problems and can easily handle the relatively small applications that make up most of the verification and validation base for the codes.

Larger models can also be run on NRC desktop platforms, but as the application size increases and significant CPU time is needed running on the NRC cloud or on the DOE HPC system will be a more efficient way to run the simulation.

The codes intended for NRC safety analysis as described in this Volume are all capable of executing on computing systems that the NRC currently maintains, or has ready access to.

Many, if not most simulations can be run on NRC desktop platforms, and larger applications can use either the NRC cloud or the DOE HPC system. It is expected that NRC users will develop and debug models on NRC desktop platforms, or on the NRC cloud, and then request additional cloud resources or utilize the DOE system for intensive calculations.

2.6 Impacts on NRC Staff Several of the codes intended for plant systems analysis represent a challenge to effective use of NRC staff. Non-LWR technologies are diverse, and are not as well known to the staff as are those for conventional light-water reactors. There is a learning curve with each technology as well as another learning curve for any code that the staff is not familiar with. In addition, the complexity of using a coupled suite is new and should be considered.

Some of the codes that are part of the BlueCRAB code suite are indeed complex. The complexity is in part due to the flexibility and expanded capabilities of the codes. It is this 21

flexibility however that allows a relatively small number of codes to simulate all of the proposed non-LWR designs. In the case of the DOE codes, part of the learning curve is based on the formulation that differs from that of the NRC codes. The TRACE code is component based and other NRC codes use a finite-volume formulation, while many of the DOE codes are based on a finite-element formulation. This requires use of a finite element mesh, and an understanding of how to generate that mesh to obtain accurate results. In some cases, the NRC codes are moving in this direction, an example being the fuel performance code (FAST) which can now incorporate use of a finite element mesh in order to model and simulate advanced fuel concepts.

Mastering this new complexity is a challenge, and possibly essential for analysis and evaluation of non-LWRs. The added flexibility afforded by a finite element mesh enables a more geometrically accurate modeling of some components (such as the monolith in a micro reactor) and some fuel designs. Some of the potential applicants have indicated that they will be using tools that employ finite element meshing, and thus the staff may eventually need to develop understanding in this technique in order to perform an adequate review.

22

3.0 BlueCRAB Code Suite for non-LWRs The proposed code suite for non-LWR safety analysis makes use of existing NRC codes, and integrates them with several codes developed through the DOE NEAMS program. The codes in some cases have multiple applications. That is, they can be used for more than one design type. There is some redundancy built into the proposed structure to allow for options to meet a currently unspecified pre-applicant schedule and to allow for independence should that become an issue during a review.

Figure 3-1 presents a schematic showing the full suite of non-LWR codes, known collectively as the Comprehensive Reactor Analysis Bundle (CRAB). In Figure 3-1 the NRC developed codes are shown in gold, while those produced by the DOE are shown in light blue. For each reactor design type, only a subset would be active as part of a given calculation which will be explained in Section 4. Codes that are expected to play a role are the NRC developed or sponsored codes TRACE [16], FAST [17], PARCS [18] (and its associated codes for cross sections and depletion, SCALE [19], and codes developed within the MOOSE framework [20] by DOE such as, BISON [21], PRONGHORN [22], SAM [23], and MAMMOTH [24]. Should CFD analysis be necessary, this would be done using the commercially available code FLUENT [25] or possibly the DOE code Nek5000 [26]. SERPENT [27] is a reactor physics code developed at the VTT Technical Research Centre of Finland, capable of calculating cross-sections and performing detailed Monte Carlo simulations.

MELCOR [28] is not part of the BlueCRAB. MELCOR will be used for severe accident and source term analysis and is discussed in Volume 3. For some accident scenarios, it may be prudent for the staff to use both BlueCRAB codes and MELCOR (independent of each other) should there be high uncertainty in some physical processes. This again will depend on the specific design, the scenario and the staffs regulatory need.

A brief description of these codes are as follows:

MOOSE [20]: The Multiphysics Object-Oriented Simulation Environment (MOOSE) is a Multiphysics framework developed by Idaho National Laboratory to provide a high-level interface for computational analysis. MOOSE provides a fully-coupled, fully-implicit solver that allows independent codes to be coupled and exchange information. MOOSE in itself has the capability to independently solve problems involving solid mechanics, phase field modeling, heat conduction, and fluid flow. In the NRCs non-LWR code suite, MOOSE is currently being used to couple other codes, and in some case provide calculations for models not readily available in the other codes.

An important feature of MOOSE is its capability to simulate thermo-mechanical interactions.

MOOSE can simulate thermal expansion in a complex geometry, and heat conduction to the boundaries. This feature makes MOOSE well-suited for some designs such as fast reactors cooled by heat pipes. Thermal expansion affects neutron leakage and thus the neutronic behavior, and thermal conduction from the core supporting structures to the reactor cavity is an important mechanism of heat rejection. Conventional NRC codes do not have this capability.

23

Figure 3-1. The Comprehensive Reactor Analysis Bundle (CRAB) for non-LWR Systems Analysis.

TRACE [16]: The TRAC and RELAP Advanced Computational Engine (TRACE) is a systems thermal-hydraulic code developed by the NRC. TRACE has the capability to simulate fluid flow and heat exchange in a thermal system and has the ability to model a variety of working fluids including helium and sodium. Code updates are available for molten salt coolants as well. Work performed by several CAMP members (an international consortium of TRACE users) has enabled TRACE to simulate many features of SFRs and some MSRs. TRACE can be used to simulate heat transfer loops and heat removal systems in non-LWRs, and will likely be the preferred tool for systems and components involving boiling, condensation and two-phase flow with water as the working fluid.

PARCS [18]: The Purdue Advanced Reactor Core Simulator (PARCS) is a three-dimensional (3D) reactor core simulator which solves the steady-state and time-dependent, multi-group neutron diffusion and low order transport equations in orthogonal and non-orthogonal 24

geometries. PARCS is coupled directly to TRACE and AGREE which provides the temperature and flow field information to PARCS.

SCALE [19]: SCALE provides an integrated framework with computational modules including three deterministic and three Monte Carlo radiation transport solvers that are selected based on the desired solution strategy. SCALE includes current nuclear data libraries and problem-dependent processing tools for continuous-energy (CE) and multigroup (MG) neutronics and coupled neutron-gamma calculations, as well as activation, depletion, and decay calculations.

SCALE includes unique capabilities for automated variance reduction for shielding calculations, as well as sensitivity and uncertainty analysis.

FAST [17]: The Fuel code for Advanced Simulation of Transients (FAST) is a combination of the FRAPCON [29] and FRAPTRAN [30] codes for fuel performance developed by the NRC.

Traditionally used for conventional UO2 fuel, FAST can simulate heat conduction through the fuel and cladding to the coolant, fuel-cladding interaction, elastic-plastic fuel and cladding deformation, fission gas release and rod internal pressure, and cladding oxidation. FAST can be modified to simulate accident tolerant fuel and metallic fuel with relatively little effort.

Modeling of non-cylindrical geometries is being addressed and is expected to be available in the near future. FAST is expected to be the NRCs code of choice for conventional fuel and most accident tolerant fuel products, and represents a potential alternative choice for simulating non-LWR fuel products.

BISON [21]: BISON is a finite element-based nuclear fuel performance code applicable to a variety of fuel forms including light water reactor fuel rods, TRISO particle fuel, and metallic rod and plate fuel. It solves the fully-coupled equations of thermo-mechanics and species diffusion, for either 1D spherical, 2D axisymmetric or 3D geometries. BISON can model an irregular geometry, and can simulate heat conduction through the fuel to the coolant, fuel and cladding deformation, fission gas release, thermal and irradiation creep, and coolant channel hydraulics.

BISON is developed by the Idaho National Laboratory. BISON has been developed as a MOOSE application and can execute massively parallel computations on a high performance computing platform. While BISON has not been subjected to the verification and validation base for conventional fuels that FRAPCON has used, BISON offers significant potential for modeling and simulating several proposed non-LWR fuel products.

PRONGHORN [22]: PRONGHORN is a MOOSE-based multi-physics reactor analysis application developed at INL initially to model the pebble bed gas-cooled reactor. It used a two-group neutron diffusion model and a porous media flow model. Recent development of PRONGHORN has extended the neutron diffusion model to an arbitrary number of groups and extended the thermal fluid model to better capture the physics of a prismatic gas-cooled reactor.

PRONGHORN can be run in serial or on massively parallel computers with one-, two-, three-D, geometries. By solving steady-state and transient-coupled, homogenized, fluid flow-heat transfer problems and standard multi-group diffusion problems (fixed-source, criticality, and time-dependent), PRONGHORN has capabilities equivalent to those of PARCS/AGREE for many of the applications projected for non-LWR designs.

SAM [23]: The System Analysis Module (SAM) is a MOOSE-based system analysis tool being developed at Argonne National Laboratory for advanced non-LWR safety analysis. It aims to 25

provide fast-running, whole-plant transient analyses capability with improved-fidelity for SFR, LFR, and MSR. SAM is being developed as a system-level modeling and simulation tool with higher fidelity yet computationally efficient compared to SAS4A/SASSYS [31], a legacy SFR code. The SAM development effort has been focused on the modeling and simulation capabilities of the heat transfer and single-phase fluid dynamics responses in reactor systems.

The transient simulation capabilities of typical reactor accidents have been demonstrated in the transient simulations of the Advanced Burner Test Reactor and validated against the EBR-II benchmark test results [32, 33].

MAMMOTH [24]: MAMMOTH (sometimes called by its solver routine RATTLESNAKE) is a MOOSE-based application for solving the multigroup radiation transport equation. It provides various discretization schemes. These schemes are leveraged with the multi-scheme capability within MAMMOTH, thus allowing the assignment of more suitable schemes on various subdomains with different levels of resolution in order to optimize the use of computing resources. Rattlesnake solves steady-state, transient and eigenvalue problems with arbitrary order of scattering anisotropy. It also owns all the features provided by the MOOSE framework, including unstructured higher-order meshes, massive parallelization, dimension agnosticism, etc. MAMMOTH is designed for multiphysics simulations. Its applications to the fully-coupled multiphysics simulation and the tightly-coupled multiphysics simulations with data transfer have been both successfully demonstrated. It interacts with a multigroup cross section library management toolkit, for incorporating the cross sections generated with lattice physics codes.

SERPENT [27]: SERPENT is a three-dimensional continuous-energy Monte Carlo reactor physics burnup calculation code, specifically designed for lattice physics applications. The code uses built-in calculation routines for generating homogenized multi-group constants for deterministic reactor simulator calculations. The standard output includes effective and infinite multiplication factors, homogenized reaction cross sections, scattering matrices, diffusion coefficients, assembly discontinuity factors, point-kinetic parameters, effective delayed neutron fractions and precursor group decay constants. User-defined tallies can be set up for calculating various integral reaction rates and spectral quantities.

FLUENT [25]: FLUENT is a commercial general-purpose computational fluid dynamics package that has been used to simulate a wide range of fluid flow and heat transfer problem. The NRC has used FLUENT in numerous nuclear safety related applications including lower plenum boron mixing, pressurized thermal shock (PTS), and spent fuel storage cask evaluations.

FLUENT solves the three-dimensional Reynolds-Averaged Navier-Stokes (RANS) equations for both steady and unsteady compressible or incompressible flow using a finite volume approach.

Species tracking as well as multiphase flow options are also available along with a variety of turbulence and wall modeling options.

Nek5000 [26]: Nek5000 is a spectral element based open source CFD solver designed to simulate laminar, transitional, and turbulent incompressible or low Mach number flows with heat transfer and species transport. For fully developed turbulent flow, turbulence models are provided for the prediction of the mean (time-averaged) flow behavior. Nek5000 solves the unsteady incompressible two-dimensional, axisymmetric, or three-dimensional Navier-Stokes (fluid flow) equations with forced or natural convection heat transfer in both stationary (fixed) or time-dependent geometry. While Nek5000 is currently coupled into the MOOSE framework, 26

the staff does not expect it to be used on a routine basis or as part of an Evaluation Model.

Nek5000 (and/or possibly FLUENT) may be used to verify the accuracy of simplified modeling in a thermal-fluids code.

The Consortium for the Advanced Simulation of LWRs (CASL) has been developing an advanced suite of computer codes designed to perform a highly detailed analysis of a light water reactor. Initial efforts focused on a pressurized water reactor, but in the most recent 5-year plan there are efforts on extending the capabilities to boiling water reactors and small modular (water) reactors. Central to application of the codes being developed is VERA, the Virtual Environment for Reactor Applications which serves to couple individual codes into a coupled multi-physics computing environment. The individual codes model and simulate the physics; neutron transport, thermal-hydraulics, fuel performance, and coolant chemistry. As part of the evaluation of codes for non-LWR application, the CASL codes were considered.

They are capable of high fidelity calculations and have advanced capabilities. However, because the primary mission of CASL has been light-water reactors they are viewed as being less ready for non-LWR application that the codes proposed for CRAB. Thus, the staff does not anticipate using the CASL codes for non-LWR analysis.

3.1 Discussion The codes discussed in the previous subsection provide a number of options and set of capabilities necessary for modeling and simulation of advanced non-LWRs (as well as for conventional LWRs). This section discusses the technical basis for selecting particular codes that are included in CRAB. As outlined in Section 2.2, many of the non-LWR applications are expected to require a multi-physics approach. Feedback between neutronics, thermal-hydraulics, and fuel performance will require a tight coupling between computer codes that are used independently for LWR applications.

The PIRTs also identify several processes that are important for non-LWRs that do not have a comparable process for LWRs. For safety analysis, thermal expansion of the core and its impact on reactivity can be ignored in LWRs for example because it is negligible compared to other components of reactivity. This is not the case however in some fast reactor designs where the negative component of reactivity is dominated by radial and axial expansion that allow neutron leakage from the core. Thus, thermal-mechanical effects in addition to the more traditional areas of neutronics, thermal-hydraulics, and fuel performance must be accounted for.

The following discusses the rationale for codes in the BlueCRAB suite of codes and some of the unique characteristics that make them important for non-LWR analysis.

Code Coupling and Thermal-Mechanical Expansion The multi-physics nature of some non-LWR analysis allows for gains in efficacy if the codes operate in a coupled computational environment. Coupling allows for a rapid feedback if one physical phenomenon affects another. The benefits of tight coupling between physics is likely to be most apparent for molten fuel salt reactors where the reactor kinetics depends on the local temperature for reactivity feedback and the flow field for tracking delayed neutrons. Other reactor designs (gas-cooled and liquid-metal) may not require as tight a temporal coupling 27

however because of the much longer mean free neutron path-length a tight coupling exists in the core-wide power and temperature distributions.

Traditional approaches generally tend to keep all of the required features for an application within one specific code. The TRACE thermal-hydraulic for example has a built-in fuel rod model that incorporates some, but not all of the features of a fuel performance code such as FRAPCON or FRAPTRAN. Differences in models between the code result in a mis-match when calculating parameters of significance for a safety analysis such as fuel centerline and fuel average temperature. A more modern approach is to couple the respective thermal-hydraulic and fuel performance codes so that each code calculates the phenomena for which it is best suited. A long-range objective for TRACE is to couple it to the FAST fuel performance code and eventually end maintenance of the TRACE fuel rod models and enable FAST to make use of TRACE thermal-hydraulics. This is expected to reduce overall development codes and improve accuracy for both TRACE and FAST.

As alluded to previously, thermal-mechanical effects can be important in some non-LWR simulations. If the core expands, changes in the core geometry will affect neutron leakage and the reactivity. The expansion however depends on the temperature distribution in the core and support structures. In some designs, in particular the micro reactors cooled by heat pipes, the support structure temperature depends on thermal conduction from the fuel element.

Both code coupling and evaluation of thermal-mechanical evaluation can be accomplished using MOOSE (Multiphysics Object Oriented Simulation Environment). MOOSE is a finite element based computational platform that enables a user to model and simulate a wide range of problems involving heat transfer and solid mechanics. The finite element approach allows for an unstructured mesh so that any geometric form can be modeled. This formulation is significant in that it allows MOOSE to model geometries than cannot be accurately modeled with conventional NRC codes. Some of the non-LWR applications where MOOSE is necessary include the fuel in some heat-pipe cooled designs and the structures supporting the fuel elements and heat pipes. Supporting structures in liquid-metal fast reactor designs can also be modeled with MOOSE to provide the radial expansion in the affected fuel assemblies.

MOOSE further has been used as a platform to couple other analysis codes. Currently, MOOSE has been used to couple the PRONGHORN and SAM thermal-hydraulic codes and the Nek5000 code for computational fluid dynamics, and to the BISON fuel performance code. In addition, the MAMMOTH code for neutronic analysis has been coupled to these thermal-hydraulic codes through the MOOSE framework.

Coupling to other codes is relatively straight-forward with MOOSE. As a demonstration, the NRCs TRACE code was coupled to BISON in late 2017. Coupling TRACE with other codes that utilize an unstructured mesh is a challenge because of the adaptive fine-mesh renodalization scheme that TRACE uses to track a quench front. The current coupling approach has resolved that difficulty, and coupling at locations where nodalization remains fixed remains to be verified but is not expected to pose a problem.

The recommended approach is to also couple the fuel performance code FAST into the MOOSE framework. Fuel performance has two important function in plant systems analysis, 28

and in many accident scenarios. First, the fuel performance code provides accurate models for fuel thermal conductivity, gap conductance (for fuel products where gaps are present) and initial stored energy. These parameters may be functions of burnup. These ensure the initial temperature distribution is estimated at the start of an event and that the worst-time-in-life can be identified. Second, the fuel performance code provides failure mechanisms and helps determine conditions at which fission product release may begin or become escalated.

In many scenarios, the fuel maximum temperature is often the main concern. However, for non-LWR fuels safety concerns may involve other phenomena such as the eutectic formation for metallic fuels or fission product diffusion in TRISO particles. Because of the number of potential fuel products and variety of phenomena, two fuel performance codes are part of the BlueCRAB code suite, BISON and FAST. BISON is already coupled into the MOOSE framework and has undergone significant development and validation for metallic and TRISO fuels. Because of these features, BISON is available for near-term analyses. FAST is currently under development and the staff plans to couple into the MOOSE framework so it can be used with the other codes.

The goal is to make these codes interchangeable so that differences in fuel behavior can be identified and explained. Use of two fuel performance codes introduces an uncertainty and also an independence in the regulatory use of the CRAB code suite. The two fuel performance codes allow the staff additional independence and options for confirmatory evaluation. Thus, coupling of FAST into the MOOSE framework is considered an important and necessary step as part of BlueCRAB. Additional information on the fuel performance codes can be found in Volume 2 of this report.

Neutronics Codes Two neutronics codes are considered as part of the CRAB code suite; MAMMOTH and PARCS.

PARCS is a diffusion based neutronics code and is well-known to the NRC staff having been used extensively for LWR analysis. Historically, diffusion based methods have been widely used. However, more recently due to the rapid increase in computational capabilities transport calculations are being used in place of diffusion calculations where detailed local information is needed. Transport avoids the limitations and approximations involved with diffusion and may be necessary to obtain sufficient accuracy in non-LWR application where the neutron mean-free path is significantly larger than in LWRs.

Generalized equivalence theory has been very successful when applied to LWRs due to the relatively short neutron mean free path as evidenced by the usage of assembly discontinuity factors in PARCS. However, for small fast reactors, assembly-based equivalence factors are not applicable, rather a full 3-D equivalence scheme such as the super homogenization (SPH) of MAMMOTH is required. Thus, the MAMMOTH code which allows for both transport and diffusion within a calculation is a viable tool. MAMMOTH also includes the capability for both macro- and micro-depletion calculations. This section discusses neutronics codes in the CRAB code suite.

For application to advanced non-LWRs, MAMMOTH (also referred to by its neutronics solver called Rattlesnake) has several advantages over traditional LWR-centric codes. In particular:

29

• Code Coupling: as MAMMOTH is a MOOSE-based code, it can couple to any other MOOSE-based code (e.g., PRONGHORN, SAM and BISON) without any modifications to any of the codes. All of the necessary coupling features such as data transfers and time step control are handled solely through the input models of the individual codes.

• Multi-Scheme Capability: not only does MAMMOTH contain discretization schemes that cover the gamut from diffusion to full transport using discrete ordinates, it can also use multiple schemes within one simulation. For example, in a non-LWR geometry (see example below), the control rods may require a different modeling approach than the rest of the core. MAMMOTHs multi-scheme approach would allow for the control rod subassembly to be modeled with a full transport method while the bulk of the core could use the more computationally efficient diffusion scheme with super-homogenization (SPH).

• Flexible Meshing: advanced non-LWRs contain non-standard geometric features that are not easily treated in traditional codes. For example, a micro reactor may have a core composed of hexagonal assemblies contained within a cylindrical reflector that in turn contains large cylindrical control rods. The finite element meshing employed in MAMMOTH allows for these important geometry features to be explicitly modeled without compromise. To illustrate, a super-cell model of the near control rod region from a small SFR is shown in Figure 3-2.

As opposed to an LWR design, the control rods are isolated in their own subassembly rather than being tightly integrated with the fuel assemblies. Homogenization of the control rod absorber material across the subassembly, as would typically be done in a diffusion code, leads to a dramatic over-prediction of the control rod worth. MAMMOTH can model each of the seven control rods individually, as shown in Figure 3-3, as well as apply a higher resolution transport scheme locally for this region.

This blending of multi-scale and multi-scheme approaches within one simulation allows for a more realistic modeling of non-LWR type geometries without the enormous computational penalty that would be incurred should higher resolution methods be applied to the entire core.

• Parallel Computation: Due to being MOOSE-based, MAMMOTH runs efficiently in a parallel mode either for shared memory desktop machines or on a high-performance cluster. For example, on a 12-core desktop, MAMMOTH runs about 8 times faster than in serial model. This makes higher resolution and fidelity calculations practicable should they be necessary. Legacy codes such as PARCS are only capable of running in a serial mode.

30

Figure 3-2. Example of control rod region from a small SFR.

Figure 3-3. Example of local finite element mesh for a SFR control rod assembly.

31

• Validation: MAMMOTH has been assessed against a variety of reactor data and benchmarks with applicability to non-LWRs. These include HTR-10 and OECD MHTGR-350 benchmark. In progress or planned are simulations of ATR, Godiva, and the CEFR startup tests.

Also included in the CRAB suite is the NRCs currently used PARCS code. PARCS is a nodal diffusion based code and was the focus of development for NGNP. Because that work ended in about 2010, several development tasks and much of the code assessment was not completed.

However, because of its potential application to gas-cooled reactors PARCS is considered as a backup approach for the following reasons:

• Code Coupling: PARCS is coupled to the AGREE [34] code for gas-cooled reactor analysis and is also coupled to TRACE.

• Validation: For non-LWR application PARCS has been assessed using Peach Bottom and HTTR gas-cooled reactors, and assessment using EBR-II is in progress.

Cross-Sections as Input to Neutronic Codes As with neutronic codes, there are options for determining cross-sections and reactivity coefficients. The important technical concern is that accurate and reliable information is supplied to the neutronics code(s) and for reactivity coefficients when point kinetics models are applied. There are two main options: SERPENT developed by VTT, and SCALE developed by ORNL. For the NEAMS codes, the SERPENT code has been the preferred code to develop this input. Because of this, and with the use of NEAMS codes SERPENT is the tool of choice - at least initially, for the following reasons:

1) SERPENT is already part of the MAMMOTH/Rattlesnake workflow. Scripts that transfer information from SERPENT to MAMMOTH already exist and have been extensively tested and verified. Implementing SCALE/Shift is possible, but only with additional cost and scheduling uncertainty.

2) In addition to providing macroscopic cross-sections, SERPENT uses Monte Carlo methods to generate a full 3-D reference solution. This reference solution is used not only for checking the validity of deterministic solutions but, as part of the MAMMOTH/Rattlesnake workflow, for the generation of super-homogenization (SPH) factors that allow computationally efficient diffusion calculations to have the accuracy of transport schemes.

3) SERPENT has a unique capability for pebble bed reactors. Specifically, it can generate random pebble and random coated fuel particle distributions for its Monte Carlo analysis. This treats the "double-heterogeneity" problem directly without requiring modeling assumptions and compromises.

4) For small fast reactors, the entire core is coupled as the neutron mean free path is on the order of the core size. For accurate analyses, this requires a fully 3-D solution rather than the traditional pin-cell approach used so successfully for LWRs. SERPENT is well-suited for this.

32

5) With respect to other Monte Carlo methods, such as MCNP, SERPENT is much faster computationally. Similar execution speed with SCALE/Shift is only expected if a large number of CPUs are used, and this may be well beyond the capability of NRCs computers.

6) Ease of use. It is relatively very easy to learn to use SERPENT which is a concern for staff and the need to rapidly become proficient in the usage of multiple analysis tools for non-LWRs.

7) Due to INL's knowledge and usage of SERPENT, it becomes a "one-stop shopping" for reactor physics. INL alone can provide the necessary support for MAMMOTH/Rattlesnake and SERPENT, whereas the usage of another code for cross-section generation would require interfacing with two contractors rather than one.

8) Maturity. SERPENT has been in active use for several years, while SCALE/Shift has yet to release a beta version that has SERPENT-like capabilities.

Fuel Performance For steady-state and design basis accident analysis, the purpose of the fuel performance code is to establish the initial condition of the core (fuel temperatures and stored energy content) and simulate fuel related processes during a hypothetical transient (barrier breach, dimensional changes with temperature excursion, fission gas release). For non-LWRs, several types of fuels have been proposed ranging from UO2, uranium nitride, metallic, and TRISO. TRISO fuel may have UO2 kernels or UCO kernels. Because of this broad range of fuel types two fuel performance codes are recommended for development and validation as part of CRAB, BISON and FAST. BISON is the DOE code developed under both CASL and NEAMS for advanced fuel analysis, and FAST is an NRC code under development. Both codes can and should be used as part of non-LWR fuel analysis, given the importance of fuel behavior during a hypothetical event. Overlap in code capability allows the staff to better understand uncertainties in predictions and provides an option should an applicant also use one these tools.

To avoid significant duplication, the approach in CRAB is to utilize FAST for fuels that are closer to conventional fuel products (oxide with Zr or metallic cladding) while BISON becomes the primary tool for TRISO. Metallic fuels such as those proposed for micro reactors and liquid metal reactors can be simulated with either code. In all cases significant code validation is necessary. Tasks identified in following sections reflect this use of the two fuel performance codes.

Inclusion of BISON in the CRAB code suite is due to:

• Code Coupling: as BISON is a MOOSE-based code, it can couple to any other MOOSE-based code (e.g., PRONGHORN, SAM and MAMMOTH) without any modifications to any of the codes. All of the necessary coupling features such as data transfers and time step control are handled solely though the input models of the individual codes. BISON is also now coupled to TRACE through MOOSE, and validation can be initiated.

33

• Flexible Meshing: advanced non-LWRs contain non-standard geometry features that are not easily treated in traditional fuel performance codes. FRAPCON and FRAPTRAN for example can only model a cylindrical geometry. However, several designs intend to use fuel products that are not circular cylinders. The finite element meshing employed in BISON allows for these important geometry features to be explicitly modeled without compromise. To illustrate, a model of defective TRISO particle and stress concentrations calculated by BISON are shown in Figure 3-4:

Figure 3-4. Tangential stress prediction by BISON in a defective TRISO particle.

• Models and Correlations for Metallic Fuel: fuel performance models and correlations are available in BISON for U-Zr and U-Pu-Zr metallic fuels. These models have been obtained from the LIFE-METAL code previously developed and used by Argonne National Lab for EBR-II analysis. While the correlations have been implemented in BISON, testing and validation however remains necessary.

• Models and Correlations for TRISO Fuel: fuel performance models and correlations are available in BISON for TRISO with UO2 kernels. Many of these models originated in the PARFUME code that was under development for TRISO fuel prior to 2010. These models will need to be revised and updated for UCO kernels as data supporting that development are produced.

• Parallel Computation: Due to being MOOSE-based, BISON runs efficiently in a parallel mode either for shared memory desktop machines or on a high-performance cluster.

This enables faster calculations to be performed on multi-CPU clusters, although BISON can execute rapidly on a single CPU.

34

• Verification of BISON and code revisions is supported by a 2000+ case regression test suite that is run on a regular basis.

Inclusion of FAST in the CRAB code suite is due to:

• Models and Correlations for Oxide Fuel: fuel performance models and correlations are available in FAST for conventional oxide fuels. These models have been obtained from the FRAPCON and FRAPTRAN codes, both of which have been extensively validated for conventional fuel performance. Revisions to models to simulate accident tolerant fuel (ATF) are currently planned.

• Models and Correlations for Metallic Fuel: addition of fuel performance models and correlations in FAST for metallic fuels are planned.

• Flexible Meshing: advanced non-LWRs contain non-standard geometry features that are not easily treated in traditional fuel performance codes. However, since FRAPCON and FRAPTRAN can only model a cylindrical geometry, revisions are being made to FAST to enable it to model non-cylindrical geometries.

A development plan for FAST for non-LWR fuel products is contained in Volume 2 of this report.

Thermal-Hydraulics Several codes are proposed and recommended to be part of CRAB in order to accommodate the wide range of advanced non-LWR designs and scale of the physical phenomena. By scale of physical phenomena, it is meant that the length scale of importance can vary considerably in an advanced reactor. Analysis capability for very large scale, often one-dimensional regions can be addressed with application of a systems level thermal-hydraulic code. With a systems code, large regions can be approximated without significant loss in accuracy. However, when dealing with recirculating flow and details, such as the velocity field in a pebble bed or plumes in an upper plenum, a greater resolution of the flow field may be required.

To accomplish this range of scale, four codes are recommended for CRAB: TRACE and SAM for system thermal-hydraulics, PRONGHORN for engineering scale thermal hydraulics, and Nek5000 (or commercial CFD code) for applications requiring computational fluid dynamics analysis. The ability to couple various codes as part of a multi-physics environment is very important in addressing this range of scale to allow for an efficient transfer of information.

Thus, part of the rationale for these codes is:

• Code Coupling: SAM and PRONGHORN are both MOOSE-based codes and can couple to any other MOOSE-based code (e.g., MAMMOTH, BISON) without any modifications to any of the codes. All of the necessary coupling features such as data transfers and time step control are handled solely though the input models of the individual codes. TRACE and Nek5000 are considered MOOSE-wrapped codes, with the essential elements of information transfer now available. While not as efficient as transfer between MOOSE-based codes, the coupling between TRACE, Nek5000 is sufficient for most purposes.

35

Inclusion of SAM in the CRAB code suite is due to:

• System Hydraulics for LMRs and MSRs: SAM can be considered as a replacement for, or modernization of the SAS4A/SASSYS code that was developed by Argonne National Lab for sodium fast reactor analysis. Most of the features available in SAS4A/SASSYS have been incorporated into SAM, enabling it to model and simulate a complex one-dimensional flow network. (Features associated with core relocation and severe accidents are not available in SAM.) Fluid properties and equation of state are available for sodium, lead, air, and several salts including FliBe and FliNaK. Models and correlations suitable for simulation of low Pr number fluids are available in SAM as are models for wire-wrapped fuel rods. SAM has been updated to simulate neutron precursor drift, which is an important process in molten fuel salt reactors.

• Validation: SAM has been validated against EBR-II and benchmarked against SAS4A/SASSYS with excellent agreement. The DOE is continuing to assess and benchmark SAM using integral effects tests from EBR-II, FFTF, Phenix, and Monju in addition to several separate effects tests for liquid metal reactors. For molten salt reactors validation is in progress for the MSRE molten salt experiment, and for the UCB-CIET integral tests.

• Reduced Order 3D Modeling: Most modern liquid-metal and molten salt reactors are pool as opposed to loop designs. That is, flow circulation and heat transfer takes place within the reactor vessel rather than through an external flow loop. The reactor vessels often contain large regions in the plena where thermal stratification and internal recirculation can be important. Thermal-hydraulic systems codes generally have difficulty simulating these processes because they lack the models and detailed nodalization required for fluid mixing. SAM however has a reduced order modeling capability that enables it to simulate these processes and approximate the simulation produced by CFD. Figure 3-5 shows an example of SAM simulation of flow in a square cavity at low Ra.

Inclusion of TRACE in the CRAB code suite is due to:

• System Hydraulics for Secondary Systems: TRACE is now a MOOSE-wrapped application and can exchange information with other codes in the MOOSE framework.

Because of this coupling and the large NRC user base for TRACE, it can be used to simulate the secondary and tertiary systems in non-LWR designs. TRACE contains fluid properties for several fluids and updates are available for some molten salts and supercritical CO2. TRACE also provides the capability to simulate thermodynamic cycles involving two-phase water.

• System Hydraulics with Two-Phase Mixtures: TRACE has the capability to track a non-condensable two-phase mixture such as air and water. This feature may be used to investigate conditions within a non-LWR at very low void fraction.

36

Figure 3-5. Velocity vectors in the square cavity at = 100000.

Inclusion of PRONGHORN in the CRAB code suite is due to:

• Engineering Scale Hydraulics for Primary Systems: PRONGHORN was developed starting in 2008 primarily for the representation of porous media in gas-cooled reactors.

For these applications an intermediate-scale representation is needed to simulate flow and heat transfer in regions where conduction, convection and thermal radiation are important. PRONGHORN can model both pebble bed and prismatic cores. Fluid properties and equation of state models are available for helium, air, and several molten salt coolants making it applicable to pebble bed cores in both gas-cooled and molten salt cooled designs.

• Validation: Assessment against the SANA tests demonstrate the applicability of PRONGHORN to simulate conduction cool-down in pebble bed cores, with good accuracy. Validation is currently in progress for other important tests including HTR-10 and HTR-PM, and planned for AVR, THTR-300, and HTTR.

• Coupling to CFD: Because PRONGHORN is a MOOSE-based application it is coupled with the other codes in the MOOSE framework. Of particular interest is the coupling with Nek5000, which can provide a detailed prediction of the local flow pattern in a pebble bed and when coupled to BISON a detailed prediction of the temperature distribution on the pebble compact. Figure 3-6 shows the Nek5000 prediction of velocities in a random 37

Figure 3-6. Nek5000 prediction of velocities in a random grouping of pebbles.

pebble experiment conducted at Texas A&M University (TAMU) that compared favorably with Particle Image Velocity (PIV) measurements.

Inclusion of Nek5000 in the CRAB code suite is due to:

• Microscale Thermal-Hydraulic Simulation: Nek5000 is a computational fluid dynamics code that solves the unsteady incompressible two-dimensional, axisymmetric, or three-dimensional Stokes or Navier-Stokes equations with forced or natural convection heat transfer. Because it is a MOOSE-wrapped application, it couples with the other codes in the MOOSE framework. For problems requiring a close coupling between CFD and other tools, Nek5000 will be a strong consideration. For problems that can be addressed without coupling, FLUENT could be used instead.

Computational Fluid Dynamics Computational Fluid Dynamics (CFD) is expected to be utilized in the staffs evaluation of non-LWRs. However, the expectation is that CFD will be used to augment and in some cases benchmark the simulation results obtained by systems level codes (SAM, PRONGHORN, and TRACE). As in light-water reactor analysis, CFD is frequently applied in order to examine specific regions of a reactor system where turbulent mixing, stratification, plumes, and complex geometry make a systems level code inaccurate or impractical.

38

As previously mentioned, the Nek5000 code is a MOOSE-wrapped application and thus is coupled to the other NEAMS codes. This enables a ready transfer of information between the codes, and an existing capability. Nek5000 however uses the large-eddy simulation (LES) approach which while highly detailed is difficult to apply and computationally expensive. Thus, the FLUENT code which often uses the less-detailed, less computationally expensive Reynolds Average Navier-Stokes (RANS) approach may be more practical when coupling to another code is not necessary. (STAR-CCM and other commercial CFD codes could also provide this capability.)

3.2 Summary of Code Pros and Cons It is informative and important to consider the pros and cons of various codes that could be applied to non-LWR analysis. This sub-section provides some of these pros-cons and highlights issues involved. In Section 2, this report discussed some of the considerations involved with code selection. This sub-section summarizes these considerations.

Table 3-1 lists the pros & cons involved with the codes intended for CRAB and other codes that have been considered.

Table 3-1. Summary of Code Considerations Analysis Main Pro(s) Con(s)

Code Application TRACE Thermal-

• Familiar to relatively large

• Cost. Extensive code Hydraulics number of NRC staff. development and

• Well validated for systems validation for non-LWRs using water as coolant. is necessary.

• Can simulate transient two-

• Heat structure model not phase flow. amenable to irregular

• Can simulate multi-dimensional geometries.

flows.

• No capability to model

• Includes properties for several entire vessel as a single fluids (other than H2O). 3-D heat structure

• Some code updates applicable (passive decay heat to non-LWRs available from removal systems).

CAMP members.

• Not capable of simulating

• Modernized programming. pools or thermal

• Extensive verification in place. stratification due to lack

• Currently coupled to PARCS of fluid-fluid shear.

neutronics code.

• Lacks validation for

• Currently coupled to MOOSE GCRs.

framework and NEAMS codes.

• Lacks validation for MSRs.

RELAP Thermal-

• Familiar to some NRC staff.

• Cost. All development Hydraulics

• Easy to use. and assessment costs would be funded by NRC.

• Offers little improvement relative to TRACE.

39

Analysis Main Pro(s) Con(s)

Code Application

• Poor V&V relative to TRACE for water as coolant.

• Lacks validation for GCRs.

• Lacks validation for SFRs.

• Lacks validation for MSRs.

• Little/no capability for multidimensional flow.

• No longer supported by NRC, and being phased out.

• Relatively few expert Users or Developers.

• Not capable of simulating pools or thermal stratification due to lack of fluid-fluid shear.

• Heat structure model not amenable to irregular geometries.

• No capability to model entire vessel as a single 3-D heat structure (passive decay heat removal systems).

• Lacks modern programming.

PRONGHORN Thermal-

• Developed specifically for

• Not familiar to staff.

Hydraulics GCRs.

• Coupling to TRACE

• Includes models and requires verification.

correlations specific to both

• Additional validation pebble bed and prismatic needed.

GCRs.

• Existing validation demonstrates good agreement with data.

• High degree of verification, exceeding NRC software requirements.

• Modern programming.

• Includes capability for simulation of pebble bed with molten salt coolant.

• High degree of parallelization to permit medium to high fidelity 40

Analysis Main Pro(s) Con(s)

Code Application modeling with acceptable CPU resources.

• DOE responsible for most development and maintenance costs.

SAM Thermal-

• Developed specifically for

• Not familiar to staff.

Hydraulics SFRs.

• Lacks compressibility

• Includes models & correlations term.

specific to sodium and other

• Coupling to TRACE liquid metals. requires verification.

• Includes properties for molten

• Additional validation salts and ability to track neutron needed.

precursors.

• Poor User Manual

• High degree of verification, exceeding NRC software requirements.

• Capable of CFD style simulation of pools and complex flows expected in SFR plena, MSR plena and core, and RCCS.

• Modern programming.

• High degree of parallelization to permit medium to high fidelity modeling with acceptable CPU resources.

• DOE responsible for most development and maintenance costs.

AGREE Thermal-

• Some validation exists with

• Limited staff familiarity, Hydraulics good agreement with data. with almost no expert

• Coupled with PARCS for Users.

neutronic feedback.

• No developers on staff.

• Insufficient software verification.

• Theory and User manuals inadequate.

• Additional validation needed.

CTF Thermal-

• Can simulate subchannel

• Limited to pre-CHF Hydraulics hydraulics. regimes (with water as

• Modern programming used. coolant).

• Several staff are expert in

• Can only simulate core COBRA-TF development and and vessel. Lacks use. models for loops and

• Available at no cost to NRC. components such as pumps or valves.

41

Analysis Main Pro(s) Con(s)

Code Application

• Uncertain plans for continued support of CASL codes.

• Most capabilities already in TRACE.

• Not coupled to MOOSE framework.

• Significant validation needed for non-LWRs.

MPACT Neutronics

• Provides extensive amount of

• Not familiar to staff.

detail for core power

• Expected to require distributions. excessive high

• Tested and applicable to LWRs. performance computing

• Good validation basis for resources, above NRC LWRs. capability to attain.

• Available at no cost to NRC.

• Extensive detail not expected to be required for safety analysis.

• Uncertain plans for continued support of CASL codes.

• Not coupled to MOOSE framework.

• Significant validation needed for non-LWRs.

PARCS Neutronics

• Familiar to staff.

• Limited to problems

• Relatively fast running. where diffusion theory

• Well validated for LWRs. applies.

• Often requires correction factors that must be obtained from experiment.

• Not applicable to non-LWRs w/o extensive revisions to model geometries involved.

• Limited/no basis for fast reactors.

• Lacks capability to do transport calculations.

• Not coupled to MOOSE framework.

MAMMOTH Neutronics

• Integrated with NEAMS codes

• Not familiar to staff.

for GCR analysis.

• Capable of modeling the double heterogeneity of pebbles.

• Good validation set already exists.

42

Analysis Main Pro(s) Con(s)

Code Application

• Can simultaneously model core regions with diffusion or transport as needed.

• Flexible meshing allows rapid re-nodalization.

• Ongoing validation work for fast reactors.

• High degree of parallelization to permit medium to high fidelity modeling with acceptable CPU resources.

• DOE responsible for most development and maintenance costs.

• Available at no cost to NRC.

PROTEUS Neutronics

• Well validated for fast reactors.

• Not familiar to staff.

• Available at no cost to NRC.

• Expected to require extensive HPC requirements.

• Not coupled to MOOSE framework.

BISON Fuel

• Provides 3D simulation of fuel.

• Limited familiarity with performance

• Can model any fuel geometry. staff.

• Includes models for metallic

• Validation considered fuel. less extensive than

• Includes models for TRISO FRAPCON for UO2 fuels.

(UO2 kernel).

• Validation incomplete for

• Coupled to MOOSE framework. metallic and TRISO fuel.

• Available at no cost to NRC.

• Theory and User manuals available.

FAST Fuel

• Limited familiarity with staff.

• Theory and User performance

• Combines capabilities of manuals under FRAPCON and FRAPTRAN. development, not

• Soon to be capable of modeling complete.

3D geometry.

• Lacks models and

• Extensive validation for ceramic correlations for metallic fuel. and TRISO fuels.

• Coupling to MOOSE framework in progress.

FRAPCON / Fuel

• Familiar to staff.

• Being phased out.

FRAPTRAN performance

• Extensive validation performed.

• Not coupled to MOOSE

• Fast running. framework.

• Limited to cylindrical fuel geometries LIFE-METAL Fuel

• Includes models for metallic

• Not familiar to staff.

performance fuel.

43

Analysis Main Pro(s) Con(s)

Code Application

• Some validation performed.

• No longer supported by DOE.

• Not coupled to MOOSE framework.

PARFUME Fuel

• Includes models for TRISO.

• Not familiar to staff.

performance

• Some validation performed.

• No longer supported by DOE.

• Not coupled to MOOSE framework.

Nek5000 CFD

• Uses Large Eddy Simulation

• Not familiar to staff.

(LES), which offers more

• No clear technical accuracy than Reynolds advantage over FLUENT Averaged Navier Stokes if RANS is sufficiently (RANS) accurate.

• Coupled into the MOOSE framework.

• Available at no cost to NRC.

FLUENT / CFD

• Familiar to staff.

• Expensive annual STAR-CCM

• Uses RANS, which is fast licensing fee.

running.

44

4.0 Code Suite by Reactor Type This section discusses how the code suite is expected to be used for a given application.

Assumed, is that because of the multi-physics nature of non-LWR accident scenarios, two or more codes will need to be used as part of a coupled calculation. No individual code is suitable for all reactor design types, but some codes have applicability to more than one design type.

The codes for each application thus retain some flexibility, and the code(s) best suited for some applications may need additional evaluation.

In this report, no clear distinction is being made between accident scenarios traditionally considered design basis Chapter 15 events and beyond design basis Chapter 19 events.

Event selection will eventually be based on specifics of the individual designs and the risk-informed approach that the NRC will determine from IAP Strategies 3 and 5. For the codes proposed here, the assumption is made that design basis events will not result in significant core disruption. Events that result in widespread core relocation and fission product transport will be simulated by MELCOR or other severe accident analysis code.

Several generic applications are discussed in this section and the codes applied for these general designs identified. The designs currently under consideration can be characterized into ten generic design types. Table 4-1 lists these designs, and the expected fuel type for each design. The examples represent possible applicants, with those listed with an asterisk (*) being those that have submitted a Regulatory Issues Summary (RIS) to the NRC. These designs can be further categorized into four general design type: gas-cooled reactors, liquid-metal reactors, heat-pipe cooled micro reactors, and molten salt reactors.

The following sections discuss the modeling and simulation approach for each of these ten design types. The approach described is intended to achieve readiness at an early date with minimal development costs. For each design type some efforts are needed to improve code capabilities and validate the codes that are applied. Therefore, for each design type a set of gaps to be resolved are listed as well as a minimal set of validation tests considered necessary. Resolution of the gaps and completion of the validation is expected to provide the tools appropriate for modeling and simulation of each design, and thus be ready to fully support licensing - subject to additional code development and assessment, if necessary, to address design specific features that will only be known following a submittal.

45

Table 4-1. Generalized Listing of non-LWR Designs Plant Description Example(s) Fuel Type No.

1 HTGR; prismatic core, thermal spectrum Framatome TRISO (rods or plates) 2 PBMR; pebble bed core, thermal spectrum X-energy* TRISO (pebbles)

Starcore 3 GCFR; prismatic core, fast spectrum GA SIC clad UC (plates) 4 SFR; sodium cooled, fast spectrum PRISM Metallic (U-10Zr)

ARC TerraPower 5 LMR; lead cooled, fast spectrum Westinghouse Not available.

Columbia Basin (Possibly nitride Hydromine fuel.)

6 HPR; heat pipe cooled, fast or epithermal Oklo* TBD spectrum Westinghouse 7 MSR; prismatic core, thermal spectrum AHTR TRISO (plates) 8 MSPR; pebble bed, thermal spectrum Kairos* TRISO (pebbles) 9 MFSR; fluoride fuel salt, Terrestrial*

thermal/epithermal spectrum Thorcon FliBe Fuel salt 10 MCSR; chloride fuel salt, fast spectrum TerraPower* Fuel salt Elysium 46

4.1 Characterization of Code Readiness Qualification of a code for a complex application depends on several factors. These factors range from the ability of the code to simulate a given phenomenon to uncertainties associated with User choices and numerical methods. Quantification of code accuracy has been a topic of significant interest over the past two decades with the rapid increase in computing capacity.

In the development of analysis methods for advanced non-LWRs, the state of readiness can be a strong function of design type. Because of previous work, both by the NRC and elsewhere the capability to simulate some designs are better than others. Molten (fuel) salt reactors for example lack an extensive experimental database and have not been investigated nearly as much as gas-cooled and sodium fast reactors.

To characterize the state of readiness for the NRCs tools for advanced non-LWRs, a method known as the Predictive Capability Maturity Model (PCMM) is employed. The PCMM [35] was developed by Sandia National Laboratory and has been used in some of the DOE code development activities. The PCMM examines a complex analysis tool in terms of six fundamental modeling and simulation elements; (1) representation and geometric fidelity, (2) physics and material model fidelity, (3) code verification, (4) solution verification, (5) model validation, and (6) uncertainty quantification and sensitivity analysis. These six elements are important in judging the credibility of a complex simulation tool such as an Evaluation Model for reactor safety.

Each of these six modeling and simulation elements was rated by NRC staff on a maturity level ranging from 0 to 3. A maturity level of 0 implies a low state of readiness while a level of 3 indicates a highly advanced capability. Table 2 provides general descriptions of the elements and maturity level assessment for each element.

The maturity scale is a subjective measure of the maturity of a Modeling and Simulation (M&S) effort for an engineering system of interest. The maturity scale, however, does not assess whether the modeling tools, accuracy of the predictions, or the performance of the engineering system satisfies a set of regulatory requirements.

The summary information in the PCMM table is expected to be beneficial in a number of ways.

As suggested in [35]:

• Conducting a PCMM assessment and sharing it with interested parties and stakeholders engenders discussions that would not have occurred without the assessment. Such communication is a highly significant consequence of an M&S maturity assessment.

• By using the PCMM over time, progress in the M&S effort can be tracked. This is useful for managers, decision makers using the results of the modeling and simulation effort, and the funding sources to determine progress.

For each design type a PCMM estimate is presented. The purpose is to provide an estimate of the maturity for each design and assist in planning for future development.

47

Table 4-2. Predictive Capability Maturity Model (PCMM) Matrix Element \ Maturity Maturity Level Maturity Level 1 Maturity Level 2 Maturity Level 3 0

Representation and

• Judgement only

• Significant

• Limited simplification

• Essentially no

• Little or no simplification of the of major components simplifications made Geometric Fidelity representation or system

• Geometry is well

• Geometry of all What features are neglected geometric fidelity defined for major components because of simplifications? for the system components and represented as built some minor

• Independent peer review components conducted

• Some peer review conducted Physics and Model

• Judgement only

• Some models and

• Physics based

• All models and

• Model forms are correlations are models and correlations are physics Fidelity unknown or ad physics based and correlations for all based How fundamental are the hoc calibrated to data important processes

• Sound physical basis for physics and calibration of the

• Few physics

• Minimal or ad hoc

• Significant calibration extrapolation models? informed models coupling of models using SETs and IETs

• Full coupling of models

• No coupling of

• Some peer review

• Independent peer review models conducted conducted Code Verification

• Judgement only

• Code is managed by

• Some algorithms are

• All of the important Are software errors and poor

• Minimum testing SQA procedures tested to determine algorithms tested to quality assurance practices? of software

• Unit and regression convergence determine convergence elements testing performed

• Some features are

• All features and

• Little or no SQA tested with capabilities tested with benchmark solutions rigorous benchmark

• Some peer review solutions conducted

• Independent peer review conducted Solution Verification

• Judgement only

• Numerical effects

• Numerical effects

• Numerical effects are Are numerical errors

• Numerical errors are qualitatively quantitatively determined to be small corrupting the results? are unknown or estimated estimated to be small

• Important simulations have large effect

• Input/output (I/O)

• I/O independently can be independently on results verified only by verified reproduced analysis

• Some peer review

• Independent peer review conducted conducted Model Validation

• Judgement only

• Quantitative

• Quantitative

• Quantitative assessment How carefully is the accuracy

• Few, if any assessment of assessment of of predictive accuracy of the simulation and comparisons to accuracy not directly predictive accuracy for all important figures experimental results measurements in relevant for some key figures of merit from SETs and assessed? similar systems or

• Large or unknown of merit from SETs IETs at applications experimental and IETs conditions/geometries uncertainties

• Experimental directly relevant to the uncertainties well application characterized

• Experimental

• Some peer review uncertainties well conducted characterized

• Independent peer review conducted Uncertainty

• Judgement only

• Aleatory and

• A&E uncertainties

• A&E uncertainties

• Only deterministic epistemic (A&E) propagated and comprehensively treated Quantification and analyses uncertainties identified and properly interpreted Sensitivity Analysis conducted propagated, but

• Quantitative

• Comprehensive How thoroughly are

• Uncertainties and without distinction sensitivity analyses sensitivity analyses uncertainties and sensitivities sensitivities not

• Informal sensitivity conducted conducted characterized? addressed. studies only

• Numerical

• Numerical propagation propagation errors are demonstrated to be estimated small

• Some strong

• No significant assumptions made assumptions

• Some peer review

• Independent peer review conducted conducted 48

4.2 Gas-Cooled Reactors Three types of gas-cooled reactors must be addressed; prismatic gas-cooled reactors (HTGR),

pebble bed gas-cooled reactors (PBMR), and gas-cooled fast reactors (GCFR). Design basis evaluation is expected to focus on steady-state operation and accidents such as pressurized and de-pressurized loss of forced cooling events. Oxidation of graphite, nor chemical reactor due to water ingress are not expected as part of these scenarios. Consistent with the approach initiated for NGNP, beyond design basis events would be simulated with MELCOR. In each of the subsections that follow the modeling and simulation approach is described, along with the gaps that exist in code development and assessment. Each gap is assigned a unique Task number that will be used for tracking progress and associated resource requirements. Tasks are also specified for development of a Reference Plant model for each design. The Reference Plant can be exercised through simulation of normal operation and several hypothetical transients to verify that the model and code(s) are ready for safety analysis. The Reference Plant is intended to approximate, but not be fully representative, of design(s) that may be submitted to the NRC for Design Certification review.

1. HTGR (prismatic TRISO fuel, thermal spectrum): At least one potential applicant is proposing a gas-cooled reactor with a prismatic core. TRISO fuel is to be contained within graphite blocks that also have bypass holes for the helium coolant. Design basis events of main interest are pressurized and depressurized loss of forced cooling events (P-LOFC and D-LOFC).

Technical Development Issues The modeling and simulation needs associated with gas-cooled reactors are relatively well known due to previous work on NGNP. To address the analytical tools and data that would be needed, the NRC conducted a Phenomena Identification and Ranking Table (PIRT) exercise in major topical areas of NGNP including accident analysis and thermal-fluids including neutronics [6]. The main report (Volume 1 of NUREG/CR-6944) summarized the important findings. Previously, a separate PIRT was conducted on TRISO-coated particle fuel for high temperature gas-cooled reactor (HTGR) technology and documented in a NUREG report (NUREG/CR-6844) [7].

Key phenomena during P-LOFC and D-LOFC include conduction through the graphite block and the effective thermal conductivity, bypass flow, and performance of the reactor cavity cooling system. Reactivity temperature feedback coefficients are important in the neutronic calculations. NUREG/CR-6944 documents several PIRTs for gas-cooled reactor analysis.

Primary Modeling and Simulation Approach The PRONGHORN code contains models for a prismatic core, and the porous media approach in PRONGHORN makes it well suited for this design. The constitutive models for the prismatic core existed in a previous version of PRONGHORN and have been retained in most recent version. PRONGHORN will be used to simulate the core and reactor vessel.

49

Neutronic calculations can be performed using MAMMOTH. Cross-sections for a prismatic gas-cooled reactor have been generated previously using SERPENT demonstrating its feasibility for this design type (see Figure 4-1). Since MAMMOTH is already coupled to PRONGHORN through the MOOSE framework, there should be an efficient transfer of information between the neutronic and thermal-fluid codes.

The secondary system, steam supply system and RCCS can be simulated with TRACE as necessary. (Simulation of loss of forced cooling events may not require modeling of the secondary system since flow into and out of the vessel can be treated with a simple boundary condition.) While TRACE has been coupled to the MOOSE framework, some testing will be necessary to validate fluid cell to fluid cell coupling to PRONGHORN and coupling at multiple locations (RCCS and secondary system).

Figure 4-2 shows a simple schematic of the codes for simulation of a prismatic gas-cooled reactor.

Figure 4-1. Model of a prismatic HTGR fuel lattice with control rod inserted in the central fuel block using SERPENT.

Significant Code Development Gaps While the codes (PRONGHORN, MAMMOTH, TRACE) can be used to simulate a prismatic core, several gaps exist that should be eliminated in order to make it more capable to address technical issues that are expected in evaluation of normal operation and various loss-of-forced cooling (LOFC) events. The PIRTs for gas-cooled reactors [6, 50

Figure 4-2. Modeling Approach for Design Basis Event Simulation in a Prismatic HTGR.

7] point out several processes that are difficult to resolve including the prediction of bypass flow and of the power/flux distribution. The thermal-hydraulics panel was concerned with the possibility of higher than expected core temperatures in normal operation, RCCS performance, and peak fuel temperatures in depressurized loss of cooling events.

To improve the ability of codes intended for simulation of prismatic gas-cooled reactors, the technical gaps to be resolved are to:

(1) Develop a reactor cavity cooling system model (RCCS) for radiation/convection from the vessel wall to the heat exchanger panel applicable to TRACE and PRONGHORN. [Tasks H3, H4, H14]

(2) Add the capability to model a 1D flow channel embedded in a 3D solid such as in a graphite block. This is to allow a more accurate representation of bypass flow. [Task H5, H14]

51

(3) Complete coupling to MOOSE framework. Provide verification on coupling of TRACE to PRONGHORN at heat exchange interfaces (and fluid cell to fluid cell if necessary). [Tasks H1, H14]

(4) Develop the capability in MAMMOTH to calculate heat generation rate in graphite moderator due to gamma absorption and neutron scattering. [Task H8]

(5) Add properties for nitrogen gas and implement the property relations for gas mixtures to PRONGHORN. [Tasks H6]

(6) Develop verification cases/sample problems for the simulation of thermal resistance between graphite block (in both radial and axial directions), and for thermal radiation across a stagnant gap. [Tasks H48, H49]

(7) Improved Control Rod (CR) treatment - develop heterogeneous CR models and test cusping treatment. In particular, ensure that SPH can be used for CR withdrawal events. [Task H50]

Assessment Assessment to demonstrate the basic applicability of a code suite for prismatic gas-cooled reactor (normal operation and design basis event) analysis must show that power and temperature distributions, as well as the important heat exchange processes can be simulated. This is complicated by the uncertainty in predicting coolant bypass, convection/conduction through a prismatic core, and heat exchange with the RCCS.

Assessment considered essential for demonstrating the adequacy of simulation codes include:

(1) HTTR: The HTTR is a helium gas cooled and graphite-moderated HTGR with the rated thermal power of 30 MW and the maximum reactor outlet coolant temperature of 950 °C. Tests, including loss of forced cooling tests, were conducted to demonstrate shutdown and inherent safety. [Task H11]

(2) OSU-HTTF: An integral test facility at Oregon State University is currently being used to produce thermal-fluid data applicable to a prismatic gas-cooled reactor.

These data are expected to be available soon. [Task H12]

(3) ANL RCCS Tests: The Natural convection Shutdown heat removal Test Facility (NSTF) is a large scale thermal hydraulics test facility that has been built at Argonne National Laboratory (ANL). The facility has principally been designed for testing of Reactor Cavity Cooling System (RCCS) concepts that rely on natural convection cooling for either air or water-based systems. [Task H44]

(4) UW RCCS Tests: An RCCS 1/4 scale air-cooled facility constructed at the University of Wisconsin - Madison has been used to compliment the tests from the Argonne NSTF. [Task H2]

(5) Ft. Saint Vrain: Ft. Saint Vrain was an operating gas-cooled reactor with a core power of 842 MWt. Operational data is available to assess neutronics codes. [Task H39]

52

(6) Peach Bottom: Peach Bottom was the first experimental gas-cooled reactor built in the U.S., and had a rated power of 40 MWe. Measurements included power distributions, control rod worths, and temperature coefficients. [Task H20]

(7) VHTRC: The VHTRC is a graphite-moderated critical assembly that has a core loaded with pin-in-block LEU fuel and a graphite reflector. [Task H40]

(8) MHTGR-350: MHTGR-350 is a benchmark problem requiring an infinite lattice calculation to be performed on a single fuel block. The prismatic HTGR design represents the 350 MW General Atomics Modular High Temperature Gas-cooled Reactor. [Task H41]

(9) HTTR-VCS: The vessel cooling system (VCS) in the HTTR corresponds to the reactor cavity cooling system (RCCS), and it mainly cools the reactor pressure vessel (RPV) by thermal radiation. Tests with the VCS compliment LOFC test without VCS. [Task H45]

(10) HTTR Neutronics: Data from HTTR provides a means to validate neutronics models for a prismatic core. [Task H19]

(11) TREAT M8CAL: TREAT is a homogeneous, air-cooled, graphite-moderated and graphite-reflected reactor. The TREAT calibration test provides neutronic validation relevant to prismatic gas-cooled cores. [Task H42]

(12) Develop gap analysis and assessment plan for TRISO fuels. [Task H26]

(13) Perform assessment of BISON for TRISO fuel as outlined in the gap analysis and assessment plan. [Task H26]

(14) Complete Code Assessment for HTGR Report. Include verification and validation to demonstrate simulation of thermal-fluids, RCCS, neutronics. [Task H32]

HTGR Reference Plant Model Development (1) Develop reference (prismatic) plant cross-sections, reactivity coefficients. [Task H16]

(2) Develop reference (prismatic) plant model. The reference model is to be used to verify all aspects of the model and perform initial scoping studies and sensitivity analysis.

[Task H25]

(3) Complete reference plant model simulations, document in a report with statement on code maturity for HTGR analysis with identification of plant specific modeling and remaining assessment requirements. (To include both prismatic and pebble bed reactors with thermal spectrum.) [Task H33]

(4) Complete Code Adequacy to HTGRs Report. Include sections on each code utilized.

Verify that code functionality, models and correlations are documented in Theory and User Manuals. [Task H31]

53

PCMM Characterization for HTGR Geometric Representation: In general, the primary code system contains models to simulate the most important phenomena. PRONGHORN and MAMMOTH were developed specifically for gas-cooled reactor cores and thus are capable of a detailed geometric representation of the core and vessel. TRACE is capable of modeling a complex flow system such as the cross-connection pipes, steam generator, and secondary systems. If a Rankine cycle is used for the secondary side and needs to be modelled, TRACE should be applicable without any additional code development.

Models for helical coiled tubes have been developed for TRACE as part of the NuScale review, should those be used in the circulator.

For PRONGHORN and TRACE, the remaining challenge is modeling the fluid flow and heat exchange in the reactor cavity and RCCS. Natural circulation in this region may require a CFD-like simulation, depending on the specific design. Neither TRACE nor PRONGHORN are well-suited for the simulation of natural circulation in a high aspect ratio enclosure. It may be possible to develop a suitable nodalization and calibrate PRONGHORN and/or TRACE using some of the assessment cases. Until this is accomplished the set of codes is rated at a maturity level of 2 for geometric representation.

Physics and Model Fidelity: Both TRACE and PRONGHORN contain models and correlations that are appropriate for system hydraulics and fluid flow/heat transfer in a pebble bed, respectively. Models and correlations are available for flow and heat transfer in the reactor cavity in both TRACE and PRONGHORN. The heat transfer package in PRONGHORN has been assessed using the SANA data, and preliminary results appear to be satisfactory. However, validation to date has been limited and the range of thermal conditions in the actual plant design are not yet available. Thus, the maturity level is rated as a 2 since significant calibration has not been performed, nor has there been an internal peer review.

Likewise, assessment of SERPENT and PRONGHORN using the MHTGR-350 benchmark, TREAT M8 calibration test, as well as HTR-10 was found to be very good.

Both codes appear to contain sufficient and appropriate models. Due to limited assessment, calibration is not possible and therefore the maturity level is currently rated as a 2 pending additional assessment and peer review.

Code Verification: TRACE, BISON, PRONGHORN, and MAMMOTH are each subjected to a verification test suite consisting of (at least) several hundred cases to test new updates and revisions to the codes. Plans are in place to qualify PRONGHORN and MAMMOTH as satisfying NQA-1 criteria within the next two years. Cases for verification of code coupling have not been established. Code verification maturity is rated a 2 pending completion of NQA-1 qualification and establishment of code coupling verification cases.

Solution Verification: Numerical convergence and sufficiency of nodalization schemes has not been examined for the coupled set of codes. On an individual code basis, 54

TRACE has been examined in various time-step size and nodalization studies and could be assigned a maturity level of at least 2 for its use in the primary and secondary side simulations. (A value of 3 for TRACE will depend on nodalization and solution convergence for RCCS simulation.) The maturity of BISON, PRONGHORN and MAMMOTH has not been investigated and therefore is conservatively rated as 0.

Model Validation: Code assessment, and calibration of some models (bypass and effective conduction in a 3D graphite block) is not complete. Thermal-hydraulic assessment for large scale facilities such as the OSU-HTTF and HTTR are not complete. The neutronics assessment using MAMMOTH showed excellent agreement with data and Monte Carlo techniques. Additional assessment is necessary to demonstrate the ability of PRONGHORN and TRACE for RCCS performance, which is considered an important need. BISON needs additional validation for TRISO fuel, and may need to wait for additional test to be completed. The maturity rating is conservatively rated a 1 due to the current lack of assessment for RCCS and for additional neutronic benchmarks against tests such as AVR, ASTRA and HTR-PROTEUS.

Uncertainty Quantification and Sensitivity: Currently not addressed. Long range plans are to use the DAKOTA based uncertainty plug-in available in SNAP to perform uncertainty quantification and automate sensitivity studies. Development of a User interface for MAMMOTH and PRONGHORN is being considered but does not appear to represent a difficult task. Because no work has been initiated, the maturity is conservatively rated 0.

The following table provides a quick view of maturity by individual code. Maturity for the set of codes for this application should be based on the lowest rated code for a given element.

55

Table 4-3. PCMM Characterization of Codes for HTGR Analysis Element \ Maturity Maturity Maturity Maturity Comments Maturity Level 0 Level 1 Level 2 Level 3 Representation

• Limited simplification and Geometric TRACE MAMMOTH required. Geometry PRONGHORN BISON and major Fidelity components can be represented.

• Modeling of the RCCS has uncertainty.

• Physics based Physics and TRACE models available, but MAMMOTH significant calibration Model Fidelity PRONGHORN using SETs and IETs BISON remains necessary.

• Verification standards Code TRACE generally applied for MAMMOTH each code. Code Verification PRONGHORN coupling verification BISON not yet addressed.

• Solution convergence Solution MAMMOTH and nodalization PRONGHORN TRACE sufficiency not yet Verification BISON established.

• Solution convergence is not yet known.

• Quantitative Model MAMMOTH TRACE assessment of PRONGHORN predictive accuracy Validation BISON for some SETs and IETs.

• Large or unknown uncertainties.

Uncertainty

• Only deterministic Quantification MAMMOTH analyses possible.

PRONGHORN TRACE Uncertainties are not and Sensitivity BISON addressed.

Analysis

• TRACE has UQ functionality built in through SNAP using DAKOTA.

56

2. PBMR (pebble bed TRISO fuel, thermal spectrum): Two potential applicants, X-Energy and Starcore, are designing pebble bed gas-cooled reactors. TRISO fuel kernels are contained within pebbles, and helium coolant flows through the pebble bed.

Design basis events of main interest are expected to be pressurized and depressurized loss of forced cooling events (P-LOFC and D-LOFC). Rod ejection represents a neutronic event of potential interest.

Technical Development Issues The modeling and simulation needs associated with gas-cooled reactors are relatively well known due to previous work on NGNP. Key phenomena during P-LOFC and D-LOFC include conduction and thermal radiation through the pebble bed and the effective thermal conductivity, bypass flow and pressure drop through the pebble bed, and performance of the reactor cavity cooling system. Reactivity temperature feedback coefficients are important in the neutronic calculations. NUREG/CR-6944 [6] documents several PIRTs for gas-cooled reactor analysis. These PIRTs provide a basis for code selection to support the safety analysis of gas-cooled reactors.

Primary Modeling and Simulation Approach The PRONGHORN code contains models for a pebble bed core, and the porous media approach in PRONGHORN makes it well suited for this design. PRONGHORN contains thermal-fluid models for flow through a porous bed (Ergun equation for pressure drop),

and for simulation of convection and thermal radiation within the bed (Zehner-Schlunder). Recent code assessment for PRONGHORN found excellent agreement between predicted and measured temperature distributions in the SANA experiments

[36]. In models of pebble bed reactors, PRONGHORN will be used to simulate the core and reactor vessel.

Neutronic calculations can be performed using MAMMOTH and SERPENT, which contain features that make them uniquely qualified for pebble bed reactor analysis.

High-temperature gas-cooled reactor geometries differ significantly from conventional light water reactors. The fissile material is encapsulated inside microscopic tristructural-isotropic (TRISO) fuel particles, randomly dispersed in fuel compacts or pebbles made of graphite. In pebble-bed type geometries the spherical fuel pebbles are piled inside the reactor core, which brings another level of random heterogeneity in the calculation.

The explicit HTGR geometry model in SERPENT reads the coordinates of fuel particles or pebbles from a separate input file, and generates the geometry as it is defined, without any approximations. The model works on several levels (particles inside a pebble and pebbles inside the core) and it has been tested in realistic double-heterogeneous reactor configurations consisting of over 60 million randomly positioned units [37, 38]. The computational overhead from handling the unstructured configuration is moderate compared to a similar regular-lattice model. The routine also enables the calculation of pebble-wise power distributions over the reactor core without defining 57

additional tallies. Recent assessment for MAMMOTH found excellent agreement between predicted and benchmark results for the HTR-10 pebble bed reactor [39].

Since MAMMOTH is already coupled to PRONGHORN through the MOOSE framework, there should be an efficient transfer of information between the neutronic and thermal-fluid codes.

While PRONGHORN is well-suited for simulation of the reactor core and vessel, the remainder of the primary cooling circuit could also be handled by an engineering level systems code. (This may not be necessary for conduction controlled cooldowns.) By coupling TRACE to the MOOSE framework, it is possible to couple PRONGHORN to TRACE at the cross-connect pipe and allow TRACE to simulate both the primary and secondary sides of the steam generator. The secondary system, steam supply system and RCCS can be simulated with TRACE. While TRACE has been coupled to the MOOSE framework, some testing will be necessary to validate coupling to PRONGHORN at multiple separate locations (cross-connect pipe, RCCS and in-vessel heat exchanger). Figure 4-3 shows a schematic of modeling for a pebble bed gas-cooled reactor.

Figure 4-3. Modeling Approach for Design Basis Event Simulation in a Pebble Bed Gas-Cooled Reactor.

58

Significant Code Development Gaps Modeling and simulation gaps for gas-cooled reactors are discussed by Ball [40]. Both pebble bed and prismatic designs were addressed, with the most significant gaps for thermal-fluids simulation in pebble bed reactors being identified as:

(a) Core coolant bypass flow (normal operation),

(b) Core effective thermal conductivity (D-LOFC),

(c) Reactor Cavity Cooling System (RCCS) performance [during loss of forced circulation (Air ingress in confinement was also identified. However, this process is considered as part of severe accident phenomena.) While these gaps were identified for NGNP, they also apply to the capabilities required for the most recent gas-cooled designs.

To complete the ability of codes intended for simulation of pebble bed gas-cooled reactors, technical gaps to be resolved are to:

(1) Develop a reactor cavity cooling system model (RCCS) for radiation/convection from the vessel wall to the heat exchanger panel applicable to TRACE and PRONGHORN.

[Tasks H3, H4, H14]

(2) Complete coupling to MOOSE framework. Provide verification on coupling of TRACE to PRONGHORN at heat exchange interfaces and fluid to fluid coupling at the cross-connection pipe. [Task H1, H14]

(3) Develop the capability in MAMMOTH and PRONGHORN to calculate heat generation rate in graphite moderator due to gamma absorption and neutron scattering. [Task H8]

(4) Develop a sample problem demonstrating the explicit modeling of the pebble transient temperature response using Nek5000 and BISON. [Tasks H7, H30]

(5) Verify coupling with MOOSE, demonstrate ability to calculate transient, local temperatures of individual pebble(s) with BISON. [Task H29]

(6) Add properties for nitrogen gas and implement the property relations for gas mixtures to PRONGHORN. [Task H6]

(7) Develop verification cases/sample problems for the simulation of thermal radiation across a stagnant gap. [Tasks H49, H14]

(8) Improved Control Rod (CR) treatment - develop heterogeneous CR models and test cusping treatment. In particular, ensure that SPH can be used for CR withdrawal events. [Task H50]

59

(9) Complete Code Assessment for PBMR Report. Include verification and validation to demonstrate simulation of thermal-fluids, RCCS, neutronics and CFD. [Task H32]

The first three gaps are common to prismatic gas-cooled reactor designs. They help ensure the energy deposition in regions surrounding the fuel are correct and that the RCCS is sufficiently well modeled. The fourth and fifth tasks are unique to a pebble bed design. These enable a detailed analysis of individual pebbles, including what might be the hottest pebble in the core as well as the possibility for non-uniform temperatures around a pebble.

Assessment Code verification and validation, also referred to as code assessment, is essential to development of codes for analysis of complex phenomena. For advanced gas-cooled reactors with pebble bed cores this will involve:

(1) SANA: The SANA tests simulated heat transfer and coolant flow in a pebble bed geometry and represents an important assessment of the simulation of convection/conduction/radiation in a packed bed. [Task H9]

(2) HTTU: The HTTU tests involved heat transfer in a pebble bed, and were conducted by Northwest University in South Africa. These tests also represent an important assessment of heat transfer in a pebble bed. [Task H10]

(3) PBMR-400, -268: The PBMR-268 and -400 benchmarks were an international exercise that concerns Pebble Bed Modular Reactor (PBMR) coupled neutronics/thermal-hydraulics transients based on the PBMR-400MW design.

[Tasks H13, H43]

(4) ANL RCCS Tests: The Natural convection Shutdown heat removal Test Facility (NSTF) is a large scale thermal hydraulics test facility that has been built at Argonne National Laboratory (ANL). The facility has principally been designed for testing of Reactor Cavity Cooling System (RCCS) concepts that rely on natural convection cooling for either air or water-based systems. [Task H2]

(5) UW RCCS Tests: An RCCS 1/4 scale air-cooled facility constructed at the University of Wisconsin - Madison has been used to compliment the tests from the Argonne NSTF. [Task H2]

(6) HTR-10: HTR-10 is a 10 MWt pebble bed reactor in China and provides startup test data that is important for neutronics code assessment. (Preliminary results have been produced using SERPENT/RATTLESNAKE.) [Task H18]

(7) ASTRA: ASTRA is a zero-power critical facility located at the Kurchatov Institute in Russia. It has been configured to represent a PBR core for the South African company PBMR. These data would provide a useful assessment of neutronics codes, if available. [Task 23]

60

(8) AVR: AVR was a prototype PBR that operated for over 21 years, at Forschungszentrum Jülich (FZJ) and provided a great deal of operating experience and experimental data, including some accident simulation data. It used a variety of fuel pebbles, which eventually evolved into the TRISO pebbles. While a good deal of the data obtained from AVR can be found in archival journals, much of the information needed for benchmarking is contained in FZJ laboratory reports. These data would provide a useful assessment of neutronics codes. [Task 21]

(9) HTTR-VCS: The vessel cooling system (VCS) in the HTTR corresponds to the reactor cavity cooling system (RCCS), and it mainly cools the reactor pressure vessel (RPV) by thermal radiation. Tests with the VCS compliment LOFC test without VCS. [Task H45]

(10) TAMU P: Pebble bed pressure drop tests were conducted at Texas A&M University that provide data on hydraulic losses for flow through a porous matrix.

[Task H46]

(11) HTR-PROTEUS: PROTEUS was a zero-power research reactor located at the Paul Scherrer Institute (PSI) configured as a pebble bed reactor (PBR) critical facility and given the designation HTR-PROTEUS. It provided experimental benchmark data to support computational assessment of high-temperature gas-cooled reactors (HTGRs). The series consisted of 17 critical configurations and various reactor physics measurements. [Task H47]

(12) Develop gap analysis and assessment plan for TRISO fuels. [Task H26]

(13) Perform assessment of BISON for TRISO fuel as outlined in the fuel gap analysis and assessment plan (which is discussed in Volume 2). [Task H26]

Plant Model Development (1) Develop reference (pebble bed) plant cross-sections, reactivity coefficients. [Task H15]

(2) Develop a sample problem illustrating the algorithm to find the equilibrium core burnup distribution for a pebble bed reactor. [Task H17]

(3) Develop reference (pebble bed) plant model. The reference model is to be used to verify all aspects of the model and perform initial scoping studies and sensitivity analysis. [Task H24]

(4) Complete reference plant model simulations, document in a report with statement on code maturity for PBMR analysis with identification of plant specific modeling and remaining assessment requirements. (To include both prismatic and pebble bed reactors with thermal spectrum.) [Task H33]

61

(5) Complete Code Adequacy to PBMR Report. Include sections on each code utilized.

Verify that code functionality, models and correlations are documented in Theory and User Manuals. [Task H31]

PCMM Characterization for PBMRs Geometric Representation: In general, the primary code system contains models to simulate the most important phenomena. PRONGHORN and MAMMOTH were developed specifically for pebble bed cores and thus are capable of a detailed geometric representation of the core and vessel. TRACE is capable of modeling a complex flow system such as the cross-connection pipes, steam generator, and secondary systems.

Since a Rankine cycle is intended, TRACE should be applicable without additional code development. Models for helical coiled tubes have been developed for TRACE as part of the NuScale review. For this set of codes, the remaining challenge is modeling the fluid flow and heat exchange in the reactor cavity and RCCS. Natural circulation in this region may require a CFD-like simulation, depending on the specific design. Neither TRACE nor PRONGHORN are well-suited for the simulation of natural circulation in a high aspect ratio enclosure. It may be possible to develop a suitable nodalization and calibrate PRONGHORN and/or TRACE using some of the assessment cases. Until this is accomplished the set of codes is rated at a maturity level of 2 for geometric representation.

Physics and Model Fidelity: Both TRACE and PRONGHORN contain models and correlations that are appropriate for system hydraulics and fluid flow/heat transfer in a pebble bed, respectively. Models and correlations are available for flow and heat transfer in the reactor cavity in both TRACE and PRONGHORN. The heat transfer package in PRONGHORN has been assessed using the SANA data, and preliminary results appear to be satisfactory. However, in validation to date has been limited and the range of thermal conditions in the actual plant design are not yet available. Thus, the maturity level is rated as a 2 since significant calibration has not been performed, nor has there been an internal peer review.

Likewise, assessment of SERPENT and PRONGHORN using HTR-10 was found to be very good. Both codes appear to contain sufficient and appropriate models. Due to limited assessment, calibration is not possible and therefore the maturity level is currently rated as a 2 pending additional assessment and peer review.

Code Verification: TRACE, PRONGHORN, and MAMMOTH are each subjected to a verification test suite consisting of (at least) several hundred cases to test new updates and revisions to the codes. Plans are in place to qualify PRONGHORN and MAMMOTH as satisfying NQA-1 criteria within the next two years. Cases for verification of code coupling has not been established. Code verification maturity is rated a 2 pending completion of NQA-1 qualification and establishment of code coupling verification cases.

Solution Verification: Numerical convergence and sufficiency of nodalization schemes has not been examined for the coupled set of codes. On an individual code basis, TRACE has been examined in various time-step size and nodalization studies and could 62

be assigned a maturity level of at least 2 for its use in the primary and secondary side simulations. (A value of 3 for TRACE will depend on nodalization and solution convergence for RCCS simulation.) The maturity of BISON, PRONGHORN and MAMMOTH for solution verification has not been investigated and therefore is conservatively rated as 0.

Model Validation: Code assessment, and calibration of some models (bypass and effective conduction in a pebble bed) is not complete. Initial efforts have used the SANA and PBMR thermal-hydraulic tests, and results are promising. The neutronics assessment using MAMMOTH showed excellent agreement with data and Monte Carlo techniques. Additional assessment is necessary to demonstrate the ability of PRONGHORN and TRACE for RCCS performance, which is considered an important need. The maturity rating is conservatively rated a 1 due to the current lack of assessment for RCCS and for additional neutronic benchmarks against tests such as AVR, ASTRA and HTR-PROTEUS.

Uncertainty Quantification and Sensitivity: Currently not addressed. Long range plans are to use the DAKOTA based uncertainty plug-in available in SNAP to perform uncertainty quantification and automate sensitivity studies. Development of a User interface for MAMMOTH and PRONGHORN was been considered but does not appear to represent a difficult task. Because no work has been initiated, the maturity is conservatively rated 0.

The following table provides a quick view of maturity by individual code. Maturity for the set of codes for this application should be based on the lowest rated code for a given element.

63

Table 4-4. PCMM Characterization of Codes for PBMR Analysis Element \ Maturity Maturity Maturity Maturity Comments Maturity Level 0 Level 1 Level 2 Level 3 Representation

• Limited simplification and Geometric TRACE MAMMOTH required. Geometry PRONGHORN and major Fidelity BISON components can be represented.

• Simulation of RCCS may require development.

• Physics based Physics and MAMMOTH TRACE models available, but PRONGHORN significant calibration Model Fidelity BISON using SETs and IETs remains necessary.

• Verification standards Code TRACE generally applied for MAMMOTH each code. Code Verification PRONGHORN coupling verification BISON not yet addressed.

• Solution convergence Solution MAMMOTH TRACE and nodalization PRONGHORN sufficiency not yet Verification BISON established.

• Solution convergence is not yet known.

• Quantitative Model MAMMOTH assessment of PRONGHORN predictive accuracy Validation TRACE for some SETs and BISON IETs.

• Large or unknown uncertainties.

Uncertainty

• Only deterministic Quantification MAMMOTH TRACE analyses possible.

PRONGHORN Uncertainties are not and Sensitivity BISON addressed.

Analysis

• TRACE has UQ functionality built in through SNAP using DAKOTA.

64

3. GCFR (gas-cooled, fast spectrum): General Atomics is developing a gas-cooled fast reactor with SiC clad UC fuel in addition to conventional spent fuel. A Brayton cycle gas-turbine is expected to provide high thermal efficiency.

Technical Development Issues Gas-cooled fast reactor designs are preliminary, and not all technical issues have been identified. Thermal-fluid analysis of a prismatic core is not expected to be a technical challenge. Simulation of the core neutronics, because of the proposed use of a new fuel type and re-use of spent conventional fuel may be a challenge, especially in development of cross-sections and reactivity coefficients. Details on the secondary and emergency core cooling systems have not been specified.

Primary Modeling and Simulation Approach The primary approach to gas-cooled fast reactor analysis is expected to be similar to that for gas-cooled reactors with a thermal spectrum using PRONGHORN for core and vessel thermal-fluid analysis and SERPENT and MAMMOTH for neutronic evaluation.

TRACE should be adequate for secondary and other safety systems. Either BISON or FAST would be used for simulation of the fuel, however the fuel product has not been specified.

Significant Code Development Gaps A PIRT applicable strictly to gas-cooled fast reactors is not available at this time.

Modeling and simulation gaps for gas-cooled reactors in general is discussed by Ball

[27]. Both pebble bed and prismatic designs were addressed, with the most significant gaps for thermal-fluids simulation in pebble bed reactors being identified as:

(a) Core coolant bypass flow (normal operation),

(b) Core effective thermal conductivity (D-LOFC),

(c) Reactor Cavity Cooling System (RCCS) performance [during loss of forced circulation.

(Air ingress in confinement was also identified. However, this process is considered as part of severe accident phenomena.) While these gaps were identified for NGNP, they also apply to the capabilities required for the most recent gas-cooled designs.

These gaps likely exist for a fast spectrum gas-cooled reactor. However, given the lack of design details, a specific set of gaps to be resolved is not defined at this time, except for development of an applicable PIRT:

65

(1) A phenomena identification and ranking table (PIRT) for a fast spectrum gas-cooled reactor should be developed to establish modeling and simulation needs. Additional gaps will be identified following the PIRT effort. [Task H34]

(2) Develop verification cases/sample problems for the simulation of thermal resistance between graphite block (in both radial and axial directions), and for thermal radiation across a stagnant gap. [Tasks H48, H49, H14]

(3) Improved Control Rod (CR) treatment - develop heterogeneous CR models and test cusping treatment. In particular, ensure that SPH can be used for CR withdrawal events. [Task H50]

(4) Other tasks TBD as design specific information is obtained.

Assessment Code assessment for fast spectrum gas-cooled reactors is expected to be similar to that for prismatic gas-cooled reactors. However, additional and or new benchmarks for neutronics will be necessary and there may be a lack of experimental data. Therefore, assessment for gas-cooled fast reactors remains to-be-determined (TBD). Results of code assessment will be documented in a separate report. [Task H37]

Plant Model Development (1) Develop reference (GCFR) plant cross-sections, reactivity coefficients. [Task H35]

(2) Develop reference (GCFR) plant model. The reference model is to be used to verify all aspects of the model and perform initial scoping studies and sensitivity analysis.

[Task H35]

(3) Complete reference plant model simulations, document in a report with statement on code maturity for GCFR analysis with identification of plant specific modeling and remaining assessment requirements. (To include both prismatic and pebble bed reactors with thermal spectrum.) [Task H38]

(4) Complete Code Adequacy for GCFR Report. Include sections on each code utilized.

Verify that code functionality, models and correlations are documented in Theory and User Manuals. [Task H36]

PCMM Characterization for GCFR Because of the lack of design information, little judgement can be made regarding code readiness. While it is likely that codes for prismatic gas-cooled reactors can be used for gas-cooled fast reactors, the maturity levels for representation and physics are conservatively rated as 0 primarily due to the lack of design information. Since verification processes are in place, these warrant a higher rating, although it is not clear what new and specific verification is necessary for gas-cooled fast reactors.

66

The following table provides a quick view of maturity by individual code. Maturity for the set of codes for this application should be based on the lowest rated code for a given element.

Table 4-5. PCMM Characterization of Codes for GCFR Analysis Element \ Maturity Maturity Maturity Maturity Comments Maturity Level 0 Level 1 Level 2 Level 3 Representation

• Judgement only and Geometric MAMMOTH BISON

• Little or no PRONGHORN representation or Fidelity TRACE geometric fidelity for the FAST system

• No design information

• Judgement only MAMMOTH

• Model forms are Physics and unknown or ad hoc PRONGHORN Model Fidelity TRACE

• No design information BISON FAST

• Verification standards Code TRACE generally applied for MAMMOTH each code. Code Verification PRONGHORN coupling verification BISON not yet addressed.

FAST

• Judgement only MAMMOTH

• Numerical errors are Solution unknown PRONGHORN Verification TRACE BISON FAST

• Judgement only MAMMOTH

• Few, if any comparisons Model to measurements in PRONGHORN Validation TRACE similar systems or applications BISON FAST Uncertainty

• Only deterministic Quantification MAMMOTH TRACE analyses possible.

PRONGHORN Uncertainties are not and Sensitivity BISON addressed.

Analysis FAST

• TRACE has UQ functionality built in through SNAP using DAKOTA.

67

4.3 Liquid Metal (Fast) Reactors Fast reactors often use a liquid-metal coolant because of their excellent heat removal capability and relatively low moderating capacity. Currently both sodium cooled and lead-cooled designs have been proposed to the NRC for possible review. This section discusses the modeling approach to these designs. While lead-bismuth cooled fast reactors have not been proposed for licensing, the approach described in this section is expected to be applicable. In each of the subsections that follow the modeling and simulation approach is described, along with the gaps that exist in code development and assessment. Each gap is assigned a unique Task number that will be used for tracking progress and associated resource requirements. Tasks are also specified for development of a Reference Plant model of each design. The Reference Plant is a can be exercised through simulation of normal operation and several hypothetical transients to verify that model and code(s) are ready for safety analysis. The Reference Plant is intended to approximate, but not be fully representative, of design(s) that may be submitted to the NRC for Design Certification review.

4. SFR (sodium cooled, fast spectrum): At least two potential applicants are proposing sodium cooled reactors with a fast neutron spectrum. These reactors are expected to operate at near atmospheric pressure and have secondary and tertiary loops to transfer heat away from the core. In the U.S., pool type as opposed to loop type designs appear to be preferred. (In a pool type SFR the components such as primary heat exchanger and coolant pumps are inside the reactor vessel.) Because of the fast spectrum, the core is densely packed with fuel and has a high power density. Design basis accident scenarios are expected to include loss of flow, loss of shutdown cooling, and inadvertent reactivity insertions.

Technical Development Issues Simulation of sodium cooled fast reactors must take into account several processes and physical phenomena associated with the high power density core and fast spectrum.

Sodium is an effective coolant because of its high thermal conductivity, and thus simulation of a sodium cooled system must be able to model axial conduction and have correlations appropriate for a low Prandtl number fluid. Neutron leakage from the core is important and represents a significant source of (negative) reactivity while voids can be a source of a large positive reactivity.

Primary Modeling and Simulation Approach While sodium fast reactors represent a relatively new technical analysis area for the NRC, there has been significant code development at the National Labs and internationally. Argonne National Laboratory in particular has continued development of codes applicable to liquid metal reactors. The most modern systems code for liquid metal reactors, SAM, builds upon the capabilities of the legacy SAS4A/SASSYS code.

SAM is a MOOSE-based application and therefore can be readily coupled with other analysis tools in the proposed tool set. Simulation capabilities of SAM have been 68

demonstrated in the transient simulations of the Advanced Burner Test Reactor and validated against the EBR-II benchmark test results.

Modeling of the primary system and core will be done using SAM, which was developed specifically for liquid-metal reactors. The SAM model will account for heat transfer and primary coolant flow in the reactor vessel, where in pool type designs there are regions where the CFD-like capability (i.e., Reduced Order Model) in SAM is necessary.

Heat exchange and simulation of the secondary and tertiary systems will be performed using TRACE. TRACE currently has the capability to simulate liquid sodium, but lacks models and correlations for simulating heat transfer in the reactor core (i.e., models for wire wrapped fuel). TRACE is however capable of modeling the secondary and tertiary heat transfer loops where sodium and/or other coolants may be used.

For neutronics, MAMMOTH can be used. MAMMOTH and SAM have already been coupled as part of the MOOSE framework and can share a common mesh. This feature is expected to be very important in designs and transients where the mesh can be displaced due to thermal expansion.

For fuel performance, either BISON or FAST could be used. BISON currently contains models and correlations for metallic fuel and FAST is scheduled to include those models in the very near future. Neither code however has been verified and validated for metallic fuel and a final decision as to which code is left open at this time. Tasks are included for this plant design to couple FAST into the MOOSE framework, which when complete will allow either code to be used.

Simulation of design basis events using SAM, MAMMOTH, and TRACE is considered the primary approach because nearly all of the generic code development for SFRs is complete. Coupling of TRACE into the MOOSE framework is also sufficiently complete and verified such that using TRACE for the secondary and tertiary systems is clearly feasible. Essentially, the approach using this set of codes is ready for code assessment.

Figure 4-4 shows a schematic of this analysis approach. Coupling between SAM and TRACE occurs at the primary heat exchanger outer surface with SAM transferring heat to the wall (or tube) and TRACE transferring heat from the HTSTR to a PIPE or other component in the flow loop.

An attractive alternative to the primary approach is to use TRACE and PARCS to simulate a sodium fast reactor. International members from CAMP have performed code development and contributed code updates for TRACE and PARCS that should enable those codes to simulate SFRs. To use this alternative approach, the code updates would have to be tested within the NRC quality assurance framework, and applied to TRACE and PARCS. A review of these updates and TRACE is necessary to ensure that features necessary to model and simulate a SFR core are complete. Work on PARCS would need to be completed in order for it to model a hexagonal core. Cross sections could be generated by SERPENT.

69

Figure 4-4. Analysis Approach for Sodium Fast Reactors.

Significant Code Development Gaps Significant work has gone into development of modeling and simulation capabilities for sodium fast reactors in the U.S. and most technical issues are well understood. Several references [8, 9, 39] discuss phenomena, gaps and modeling requirements for sodium fast reactors. While the system of codes intended for SFR analysis appear to be well suited to the task, several gaps should be resolved to improve their adequacy. Gaps to be resolved are:

(1) Verify application of sodium as a coolant has been completed in TRACE, and review the availability of appropriate convective heat transfer correlations for low Pr fluids.

[Task S1]

(2) Add coupling for 2D heat structure to 3D porous medium in SAM. [Task S6]

(3) Develop reactor cavity cooling system model (RCCS) for heat exchange from panel to secondary. Perform assessment using Wisconsin and ANL tests (with SAM and TRACE). Test and verify the coupling of TRACE and SAM. [Tasks S2, S4]

70

(4) Add reactivity feedback parameters for both axial and radial expansion effects to SAMs point kinetics model. [Task S5]

(5) Couple FAST into MOOSE framework. This is to allow the use of either FAST or BISON for evaluation of SFR fuel (assumed to be metallic). [Task S40]

(6) Complete finite element modeling implementation into FAST, include geometry independent solvers, and couple to SCALE. [Task S46]

(7) Development of an updated Material Library that allows for easy implementation of new material properties, including properties for metallic fuel (U10Zr), cladding (HT9) and sodium coolant. [Task S50]

(8) Develop a MAMMOTH sample problem for the calculation of decay heat in metallic fuel. [Task S49]

(9) Assessment and implementation (if needed) of physical models for zirconium redistribution and mechanistic fission gas release. [Task S51]

(10) Modify FAST to provide capability to simulate multiple rods. [Task S52]

(11) Develop a gap analysis, development and assessment plan for metallic fuels.

[Task S20]

(12) Improvements to constitutive models identified during assessment. [Task S12]

Assessment Code assessment remains as the major task in completing the codes for analysis of sodium fast reactor analysis. Assessment cases that appear to have the most value are:

(1) EBR-II: The unprotected loss-of-force-cooling tests in EBR-II are extremely valuable assessment cases. This will need to be simulated with final codes versions, however, SAM has already been used to simulate these tests with favorable results.

[Task S7, S15]

(2) FFTF: Loss of forced cooling tests were conducted in the FFTF facility, and will be the subject of an IAEA international benchmark exercise starting late 2018. [Task S8]

(3) Monju: Natural circulation tests were conducted in the Japanese Monju SFR.

Simulation of these tests can help assess codes for low flow conditions in a sodium loop. [Task S9]

(4) CEFR: The Chinese Experimental Fast Reactor (CEFR) is a 65 MWt sodium cooled fast reactor. The reactor was used to develop several sets of data that can be used to assess a neutronics code for a fast reactor. These tests are part of an IAEA benchmark that started in mid-2018. [Task S17]

71

(5) ZPPR: A series of criticality tests were performed by Argonne National Lab in the 1960s. Tests ZPPR-6, -15, and -21 are recommended. Test ZPPR-15 appears to be suitable for fuel with enrichments expected for modern fast reactors. [Task S18]

(6) Godiva: Godiva was an assembly at LANL that provides data for criticality calculations and has been used in several neutronic code assessments. It was fueled with 93% HEU. [Task S16]

(7) Phenix: Phenix was a 250 MWe pool-type SFR in France. The end-of-life benchmark provides data on control rod efficiency and core power shape due to insertion and withdrawal of control rods. [Tasks S11, S14]

(8) BN-600: The Russian BN-600 SFR was used for an IAEA benchmark involving MOX. Parameters evaluated included core power distribution and sodium density coefficient. [Task S10, S13]

(9) Perform assessment of FAST for metallic fuel. [Task S45]

(10) Perform assessment of BISON for metallic fuel. [Task S19]

(11) Complete metallic fuel performance assessment report. [Tasks S22, S44]

(12) Complete SFR assessment report, to include both thermal-fluid and neutronic assessments. [Task S24]

Plant Model Development (1) Develop reference (SFR) plant model. The reference model is to be used to verify all aspects of the model and perform initial scoping studies and sensitivity analysis.

[Task S25]

(2) Complete reference plant model simulations, document in a report with statement on code maturity for SFR analysis with identification of plant specific modeling and remaining assessment requirements. [Task S26]

(3) Complete Code Adequacy to SFR Report. Include sections on each code utilized.

Verify that code functionality, models and correlations are documented in Theory and User Manuals. [Task S23, S53]

PCMM Characterization Many of the models in the SAM code originated with the SAS4A/SASSY code for SFR analysis, and the code was designed specifically for SFRs. While TRACE must be verified for a sodium coolant, there is experience in using TRACE for SFRs by international partners. Thus, the ability to model a SFR is very mature. Some limited physics must be implemented, notably an accounting for the radial and axial expansions in the core.

Geometric Representation: In general, the primary code system contains models to simulate the most important phenomena. SAM was developed specifically for liquid metal reactors and is are capable of a detailed geometric representation of the core and 72

vessel. Moreover, the reduced order CFD capability in SAM enables it to simulate large fluid regions, or pools that are part of modern liquid metal reactor designs. TRACE is capable of modeling a complex flow system such as the secondary and steam generation system. TRACE should be applicable without additional code development. Some efforts are necessary to model the RCCS system and refine multidimensional heat structures. Until this is accomplished the set of codes is rated at a maturity level of 2 for geometric representation.

Physics and Model Fidelity: Both SAM and TRACE contain models and correlations that are appropriate for system hydraulics and fluid flow/heat transfer in a liquid metal primary and secondary, respectively. Models and correlations are available for flow and heat transfer in the reactor cavity in both TRACE and SAM. However, in validation to date has been limited and the range of thermal conditions in the actual plant design are not yet available. Thus, the maturity level is rated as a 2 since significant calibration has not been performed, nor has there been an internal peer review.

Code Verification: TRACE, SAM, and MAMMOTH are each subjected to a verification test suite consisting of (at least) several hundred cases to test new updates and revisions to the codes. Plans are in place to qualify SAM and MAMMOTH as satisfying NQA-1 criteria within the next two years. Cases for verification of code coupling have not been established. Code verification maturity is rated a 2 pending completion of NQA-1 qualification and establishment of code coupling verification cases.

Solution Verification: Numerical convergence and sufficiency of nodalization schemes has not been examined for the coupled set of codes. On an individual code basis, TRACE has been examined in various time-step size and nodalization studies and could be assigned a maturity level of at least 2 for its use in the primary and secondary side simulations. (A value of 3 for TRACE will depend on nodalization and solution convergence for RCCS simulation.) The maturity of SAM and MAMMOTH has not been investigated and therefore is conservatively rated as 0 simply due to lack of information.

Model Validation: Code assessment is not complete however initial efforts have used the EBR-II tests, and results have shown excellent agreement with data and SAS4A.

Additional assessment against integral experiments is on-going or planned. Additional assessment is necessary to demonstrate the ability of SAM and TRACE for RCCS performance, which is considered an important need. The maturity rating is rated a 2 due to the current lack of assessment for RCCS.

The neutronics assessment for fast spectrums is also in progress, but with limited current results it is premature to give as high a rating as for hydraulics. Therefore, MAMMOTH is conservatively rated as a 1 pending completion of neutronic benchmarks applicable to fast spectrums. Likewise, for fuel performance assessment a 1 is assigned pending validation against applicable metallic fuel data.

Uncertainty Quantification and Sensitivity: Currently not addressed. Long range plans are to use the DAKOTA based uncertainty plug-in available in SNAP to perform 73

uncertainty quantification and automate sensitivity studies. Development of a User interface for MAMMOTH and SAM has been considered but does not appear to represent a difficult task. Because no work has been initiated, the maturity is conservatively rated 0.

The following table provides a quick view of maturity by individual code. Maturity for the set of codes for this application should be based on the lowest rated code for a given element.

Table 4-6. PCMM Characterization of Codes for SFR Analysis.

Element \ Maturity Maturity Maturity Maturity Comments Maturity Level 0 Level 1 Level 2 Level 3 Representation

• Limited simplification and Geometric MAMMOTH required. Geometry TRACE SAM and major Fidelity BISON components can be FAST represented.

• Physics based Physics and BISON MAMMOTH models available, but FAST SAM significant calibration Model Fidelity TRACE using SETs and IETs remains necessary.

• Verification standards Code BISON TRACE generally applied for FAST MAMMOTH each code. Code Verification SAM coupling verification not yet addressed.

• Solution convergence Solution MAMMOTH and nodalization SAM TRACE sufficiency not yet Verification BISON established.

FAST

• Solution convergence is not yet known.

• Quantitative Model assessment of predictive accuracy Validation MAMMOTH SAM for some SETs and BISON TRACE IETs.

FAST

• Good agreement with EBR-II, which may be the most important validation case.

• Only deterministic Uncertainty analyses possible.

MAMMOTH TRACE Uncertainties are not Quantification SAM addressed.

and Sensitivity BISON

• TRACE has UQ Analysis FAST functionality built in through SNAP using DAKOTA.

74

5. LMR (Pb cooled, fast spectrum): There are three potential applicants proposing lead-cooled fast reactors including Westinghouse, Hydromine, and Columbia Basin. While the U.S. has no operational experience in lead, or lead-bismuth cooled designs they are of interest because these coolants, unlike sodium, are chemically inert. These designs are similar to sodium cooled fast reactors.

Technical Development Issues The technical issues and modeling challenges for lead- and lead-bismuth cooled designs are expected to be similar to those in sodium fast reactors. Lead has a melting point significantly higher than sodium, and solidification may be a unique issue to lead-cooled designs.

Primary Modeling and Simulation Approach Simulation of design basis events using SAM, MAMMOTH, and TRACE is considered the primary approach because nearly all of the generic code development for lead- and lead-bismuth cooled reactors is complete. Coupling of TRACE into the MOOSE framework is also sufficiently complete and verified such that using TRACE for the secondary and tertiary systems is clearly feasible. Essentially, the approach using this set of codes is ready for code assessment.

The alternative approach is to use TRACE and PARCS to simulate a lead- or lead-bismuth cooled fast reactor. Contributions by CAMP members to TRACE and PARCS development provide some of the code development necessary to simulate lead-cooled systems. In particular, NUREG/IA-0421 [41] reports some of those contributions applicable to lead-cooled systems.

Significant Code Development Gaps Lead cooled and lead-bismuth cooled reactors have not received significant attention with regards to modeling and simulation tools. It is expected that the modeling requirements for sodium fast reactors will encompass most if not all of the modeling needs for these other liquid metal cooled designs. The first step in identifying modeling and simulation gaps would be to develop a PIRT once sufficient design information is available. Thus, gaps on a preliminary basis are:

(1) A phenomena identification and ranking table (PIRT) for a fast spectrum lead-cooled reactor should be developed to establish modeling and simulation needs. [Task S43]

(2) Implementation and testing of code updates for TRACE to simulate lead as a coolant. [Task S41]

75

Assessment The database to assess codes for a lead or lead-bismuth cooled fast reactor are very limited. An initial assessment could be performed with:

(1) HELIOS: The LACANES benchmark deals with the characterization of the local pressure losses, due to form and frictional losses, at the HELIOS facility for lead alloy coolants. Several benchmark participants calculated the losses measured during an isothermal forced convection case. [Task S42]

Plant Model Development To be identified at a later date.

PCMM Characterization Because of the lack of design information, little judgement can be made regarding code readiness. While it is likely that codes for sodium fast reactors can be used for lead and lead-bismuth cooled fast reactors, the maturity levels for representation and physics are conservatively rated as 0. Since verification processes are in place, these warrant a higher rating, although it is not clear what new and specific verification is necessary for lead-cooled fast reactors.

Geometric Representation: In general, the primary code system contains models to simulate the most important phenomena. SAM was developed specifically for liquid metal reactors and is capable of a detailed geometric representation of the core and vessel. Moreover, the reduced order CFD capability in SAM enables it to simulate large fluid regions, or pools that are part of modern liquid metal reactor designs. TRACE is capable of modeling a complex flow system such as the secondary and steam generation system. TRACE should be applicable without additional code development. Some efforts are necessary to model the RCCS system and refine multidimensional heat structures. Until this is accomplished the set of codes is rated at a maturity level of 2 for geometric representation.

Physics and Model Fidelity: Both SAM and TRACE contain models and correlations that are appropriate for system hydraulics and fluid flow/heat transfer in a liquid metal primary and secondary, respectively. Models and correlations are available for flow and heat transfer in the reactor cavity in both TRACE and SAM (assuming that for lead-cooled reactor design include an RCCS). However, in validation to date has been limited and the range of thermal conditions in the actual plant design are not yet available. Thus, the maturity level is rated as a 2 since significant calibration has not been performed, nor has there been an internal peer review.

Code Verification: TRACE, SAM, BISON, and MAMMOTH are each subjected to a verification test suite consisting of (at least) several hundred cases to test new updates and revisions to the codes. Plans are in place to qualify SAM and MAMMOTH as 76

satisfying NQA-1 criteria within the next two years. Cases for verification of code coupling have not been established. Code verification maturity is rated a 2 pending completion of NQA-1 qualification and establishment of code coupling verification cases.

Solution Verification: Numerical convergence and sufficiency of nodalization schemes has not been examined for the coupled set of codes. On an individual code basis, TRACE has been examined in various time-step size and nodalization studies and could be assigned a maturity level of at least 2 for its use in the primary and secondary side simulations. (A value of 3 for TRACE will depend on nodalization and solution convergence for RCCS simulation.) The maturity of SAM and MAMMOTH has not been investigated and therefore is conservatively rated as 0 simply due to lack of information.

Model Validation: Code assessment, and calibration of some models (bypass and effective conduction in a pebble bed) is not complete. Initial efforts have used the EBR-II tests, and results have shown excellent agreement with data and SAS4A. Additional assessment against integral experiments is on-going or planned. Additional assessment is necessary to demonstrate the ability of SAM and TRACE for RCCS performance, which is considered an important need. The maturity rating is rated a 2 due to the current lack of assessment for RCCS.

The neutronics assessment for fast spectrums is also in progress, but with limited current results it is premature to give as high a rating as for hydraulics. Therefore, MAMMOTH is conservatively rated as a 1 pending completion of neutronic benchmarks applicable to fast spectrums.

Uncertainty Quantification and Sensitivity: Currently not addressed. Long range plans are to use the DAKOTA based uncertainty plug-in available in SNAP to perform uncertainty quantification and automate sensitivity studies. Development of a User interface for MAMMOTH and SAM was been considered but does not appear to represent a difficult task. Because no work has been initiated, the maturity is conservatively rated 0.

The following table provides a quick view of maturity by individual code. Maturity for the set of codes for this application should be based on the lowest rated code for a given element.

77

Table 4-7. PCMM Characterization of Codes for LMR Analysis Element \ Maturity Maturity Maturity Maturity Comments Maturity Level 0 Level 1 Level 2 Level 3 Representation

• Limited simplification and Geometric SAM MAMMOTH required. Geometry TRACE BISON and major Fidelity FAST components can be represented.

• Physics based Physics and BISON MAMMOTH models available, but FAST SAM significant calibration Model Fidelity TRACE using SETs and IETs remains necessary.

TRACE

• Verification standards Code MAMMOTH generally applied for SAM each code. Code Verification BISON coupling verification FAST not yet addressed.

• Solution convergence Solution MAMMOTH and nodalization SAM TRACE sufficiency not yet Verification BISON established.

FAST

• Solution convergence is not yet known.

• Quantitative Model BISON MAMMOTH assessment of FAST SAM accuracy not directly Validation TRACE relevant

• Only deterministic Uncertainty analyses possible.

MAMMOTH TRACE Uncertainties are not Quantification SAM addressed.

and Sensitivity BISON

• TRACE has UQ Analysis SAM functionality built in through SNAP using DAKOTA.

78

4.4 Heat Pipe Cooled Micro Reactors Recently, there has been interest in small transportable reactors for application in remote areas.

These designs have a low power levels compared to conventional power reactors, and make effective use of heat pipes to transport heat from the core to the secondary system. While the heat pipes contain sodium, they are not conveniently characterized as sodium fast reactors as in the previous section because the sodium is contained within several dozen heat pipes and is not pumped through a primary system. Technical issues associated with simulation of heat pipe cooled designs are unique. Therefore, modeling and simulation of heat pipe cooled designs is treated separately. In each of the subsections that follow the modeling and simulation approach is described, along with the gaps that exist in code development and assessment.

Each gap is assigned a unique Task number that will be used for tracking progress and associated resource requirements. Tasks are also specified for development of a Reference Plant model of each design. The Reference Plant is a can be exercised through simulation of normal operation and several hypothetical transients to verify that model and code(s) are ready for safety analysis. The Reference Plant is intended to approximate, but not be fully representative, of design(s) that may be submitted to the NRC for Design Certification review.

6. HPR (heat pipe cooled, fast reactor): Two potential applicants are considering reactors that employ heat pipes to removed energy from the core; Oklo, and the Westinghouse eVinci reactor. Modeling of heat pipe cooled micro reactors are therefore based on reactors for space based applications.

Technical Development Issues Unique to these designs are the use of heat pipes to transfer heat from the core to the secondary system. Heat pipes are very effective in transferring heat and do not depend on gravity. Fluid within the heat pipe (expected to be sodium at sub-atmospheric pressure) evaporates at the heat source and condenses at the heat sink causing the heat pipe to act as if it were a superconductor. Currently no nuclear safety systems codes simulate heat pipe performance.

An important neutronic issue is that of radial and axial expansion of the core and its supporting structure. Radial expansion increases the leakage of fast neutrons from the core, and this leakage is expected to represent the largest component of negative reactivity. Thus, simulation of, or at least an estimate of the radial expansion is necessary for an accurate determination of keff. The heat pipes also represent a path for neutron leakage from the core, and this must also be accounted for.

Thermo-mechanical response of the supporting structures and components in the core are also important in heat removal during many accident scenarios such as a loss of heat removal to the secondary side which renders the heat pipe ineffective. Heat must then be conducted through the various structures to the vessel wall where heat is then removed by a Reactor Cavity Cooling System (RCCS) or by natural convection to the ambient. Thus, the heat conduction pathways and resistances to heat removal (such as the contact resistance between fuel elements) must be modeled.

79

Primary Modeling and Simulation Approach Accident scenarios of interest to a heat pipe cooled reactor are expected to involve the loss of secondary flow and inadvertent increase in reactivity due to an unintended positioning of the control drums that surround the reactor.

The heat pipe cooled designs do not rely on flowing sodium as do typical Gen-IV sodium fast reactors. Heat removal occurs through the heat pipe, and if there is a loss of heat sink for the heat pipe, by conduction to the reactor cavity. Thus, there is no primary fluid system to be represented by system thermal-hydraulic code. Because conduction heat transfer and radial expansion of the core are important, possibly dominant physical processes, MOOSE will be used to model the structures outside of the fuel. Estimating contact resistances between various core components and structures may be a challenge, as will the thermo-mechanical interactions that arise from thermal expansion of these components and structures.. The active fuel region would be simulated with MOOSE/BISON as appropriate. FAST is an alternate means to simulate the fuel. However, because the intended burnup level is expected to be low fuel performance is not an issue rather than being able to assign the region occupied by fuel with the correct thermal and mechanical properties. The main goal is to simulate heat generation within the fuel and the expansion of fuel.

The heat pipe will be simulated using a new component developed for the SAM code.

This component will be simple, and approximate heat pipe performance as a superconductor with a very high effective thermal conductivity. Vendor specific heat pipe performance data will be required eventually in order to benchmark the SAM component. Should a more advanced heat pipe simulation capability be required, the SOCKEYE code developed by Los Alamos and INL could be coupled into the MOOSE framework or used as a stand-alone code to benchmark the SAM component.

SAM will also be used to simulate heat transfer and fluid flow in the gas space between the reflector region and the reactor cavity cooling system heat exchange panels. Heat transfer in this region is expected to be by thermal radiation and natural convection, with the flow depending on the aspect ratio of the gas space. The flow is likely to be multidimensional, requiring the CFD-like capabilities in SAM.

Core physics will be simulated using MAMMOTH. SERPENT will be used to provide a Monte Carlo reference solution and cross-sections for MAMMOTH/Rattlesnake calculations.

TRACE will be used to simulate heat removal from the heat pipe upper plenum and the associated secondary system, and also to simulate the reactor cavity cooling system (RCCS). Simulation of the heat exchange effectiveness from the heat pipe to the upper plenum will require design specific performance data for a range of conditions. Likewise, simulation of the RCCS will require design specific information.

80

Figure 4-5 shows a schematic of the modeling approach for a heat pipe cooled reactor.

Note that coupling between TRACE and SAM (heat pipe) occurs at the heat exchanger plenum near the top of the heat pipe. In the MOOSE model of the core, there will be N individual heat pipes each coupled to BISON at the active fuel and at the upper plenum where the N heat pipes dump their heat into the TRACE (PIPE) component representing a common secondary side plenum. Since the heat pipe is part of SAM, the N heat pipe components must couple to both BISON and TRACE.

Coupling between TRACE and SAM also occurs at the interface between the reactor cavity gas space and the RCCS. The TRACE representation of the RCCS may involve more than one set of loop components if multiple trains of RCCS must be simulated. In the reactor cavity gas space SAM will need to be coupled to the TRACE heat structure (HTSTR) representing the tubes or plates for the RCCS heat exchanger.

An alternate approach to modeling the thermal conduction and radial/axial expansion of the core would be to model the core and supporting structures with a commercial code such as COMSOL. COMSOL would also be used to model the heat pipe as a superconductor. Heat transfer across the gap between the reflector and the RCCS would be estimated through the selection of appropriate correlation(s) for convective heat transfer and radiation heat transfer. TRACE would be used to simulate the secondary system and the RCCS.

Significant Code Development Gaps A heat pipe cooled reactor presents several unique challenges to modeling and simulation. Heat pipe performance is an obvious difference between micro and conventional reactors. Because the micro reactor intend to have a fast spectrum (or possibly epithermal), and the heat pipes provide an axial pathway for neutron leakage the radial and axial expansion must be simulates to get reactivity correct.

(1) Develop a PIRT for heat pipe cooled reactors (or evaluate and augment [15]). [Task S36]

(2) Develop update(s) to simulate supercritical CO2 as a secondary side coolant in TRACE. [Task S27]

(3) Develop a thermo-mechanical example demonstrating the calculation of the radial expansion of a core support plate using the MOOSE mechanics module and 3D conduction capability. [Tasks S28, S29]

(4) Develop a heat pipe component model for use in SAM (or incorporate SOCKEYE).

[Task S33]

81

Figure 4-5. Model of a Heat Pipe Cooled Reactor for Design Basis Event Analysis.

(5) Develop a single fuel element model of a heat pipe cooled micro reactor to test code coupling and perform sensitivity studies. [Task S31]

(6) Develop a reactor cavity cooling system model (RCCS) for radiation/convection from the vessel wall to the heat exchanger panel applicable to TRACE and SAM. [Tasks S2, S4]

(7) Develop a sample problem for a small fast reactor demonstrating the reactivity effects associated with axial thermal expansion of a metallic fuel and radial thermal expansion of the core support plate. [Task S30]

82

(8) Couple FAST into MOOSE framework. This is to allow the use of either FAST or BISON for evaluation of SFR fuel (assumed to be metallic). [Task S40]

(9) Complete finite element modeling implementation into FAST, include geometry independent solvers, and couple to SCALE. [Task S46]

(10) Development of an updated Material Library that allows for easy implementation of new material properties, including properties for metallic fuel (U10Zr), cladding (HT9) and sodium coolant. [Task S50]

(11) Assessment and implementation (if needed) of physical models for zirconium redistribution and mechanistic fission gas release. [Task S51]

(12) Modify FAST to provide capability to simulate multiple rods. [Task S52]

(13) Develop a gap analysis, development and assessment plan for metallic fuels.

[Task S20]

Assessment Code assessment for a heat pipe cooled system represents a major challenge due to the lack of publicly available and applicable data. No integral facilities or prototypes exist, and design specific heat pipe performance information has not been provided as of yet.

Assessment should include:

(1) CEFR: The Chinese Experimental Fast Reactor (CEFR) is a 65 MWt sodium cooled fast reactor. The reactor was used to develop several sets of data that can be used to assess a neutronics code for a fast reactor. These tests are part of an IAEA benchmark that started in mid-2018. [Task S17]

(2) KRUSTY: The Kilopower Reactor Using Stirling Technology (KRUSTY) experiment provides data for a heat pipe cooled reactor. KRUSTY was operated at powers up to 5.5 kWt and includes data for several transients. [Task S32]

(3) EBR-II: The unprotected loss-of-force-cooling tests in EBR-II are extremely valuable assessment cases. This will need to be simulated with final codes versions.

However, SAM has already been used to simulate these tests with favorable results.

[Tasks S7, S15]

(4) UWisc RCCS: An RCCS 1/4 scale air-cooled facility constructed at the University of Wisconsin - Madison has been used to compliment the tests from the Argonne NSTF. [Task H2]

(5) ANL RCCS: The Natural convection Shutdown heat removal Test Facility (NSTF) is a large scale thermal hydraulics test facility that has been built at Argonne National Laboratory (ANL). The facility has principally been designed for testing of Reactor Cavity Cooling System (RCCS) concepts that rely on natural convection cooling for either air or water-based systems. [Task H44]

83

(6) HTTR-VCS: The vessel cooling system (VCS) in the HTTR corresponds to the reactor cavity cooling system (RCCS), and it mainly cools the reactor pressure vessel (RPV) by thermal radiation. Tests with the VCS compliment LOFC test without VCS. [Task H45]

(7) Heat Pipe Performance: The heat pipe model should be benchmarked against heat pipe performance data. (May be proprietary.) [Task TBD]

(8) Perform assessment of FAST for metallic fuel. [Task S45]

(9) Perform assessment of BISON for metallic fuel. [Task S19]

(10) Complete metallic fuel performance assessment report. [Tasks S22, S44]

Plant Model Development (1) Develop reference (HPR) plant model. The reference model is to be used to verify all aspects of the model and perform initial scoping studies and sensitivity analysis.

[Task S34]

(2) Complete reference plant model simulations, document in a report with statement on code maturity for HPR analysis with identification of plant specific modeling and remaining assessment requirements. [Task S39]

(3) Complete Code Adequacy to SFR Report. Include sections on each code utilized.

Verify that code functionality, models and correlations are documented in Theory and User Manuals. [Task S37]

PCMM Characterization Although there is a lack of design information, judgement can be made regarding code readiness due to the simplicity of the system. Heat transfer of a heat pipe cooled reactor is assumed to be by conduction and thermal radiation from the structures surrounding the fuel to an RCCS. Heat transfer from the heat pipes to the secondary system may contribute to cooling during some design basis events, but may not be essential.

Thermal conduction, and thermal expansion of the core and fuel is expected to be important processes for both cooling of the system and neutronic behavior.

Geometric Representation: In general, geometric representation of the fuel, core and supporting structures can be accomplished through a meshing tool such as CUBIT.

This mesh can be used by MOOSE to simulate conduction throughout the core and support structures providing thermal-mechanical expansion and the temperature distribution (at the outer edge of the reflector to serve as the boundary condition to the RCCS). The heat pipe can be simulated as a superconducting material as an initial approach. This could also be accomplished by defining a special component in SAM or by eventually adopting the SOCKEYE code which provides a detailed simulation of heat pipe performance. The reduced order CFD capability in SAM enables it to simulate large fluid regions, such as that in the reactor cavity. TRACE is capable of modeling a complex flow system such as the secondary and steam generation system. TRACE 84

should be applicable without additional code development. Some efforts are necessary to model the RCCS system and refine multidimensional heat structures. Until this is accomplished the set of codes is rated at a maturity level of 2 for geometric representation. MAMMOTH has already been used to simulate the power distribution in a heat pipe cooled design demonstrating its geometric fidelity.

Physics and Model Fidelity: Both SAM and TRACE contain models and correlations that are appropriate for system hydraulics and fluid flow/heat transfer in the secondary (an update is available for supercritical CO2). Models and correlations are available for flow and heat transfer in the reactor cavity in both TRACE and SAM. However, validation to date has been limited and the range of thermal conditions in the actual design are not yet available. Thus, the maturity level is rated as a 2 since significant calibration has not been performed, nor has there been an internal peer review.

Code Verification: TRACE, SAM, and MAMMOTH are each subjected to a verification test suite consisting of (at least) several hundred cases to test new updates and revisions to the codes. Plans are in place to qualify SAM and MAMMOTH as satisfying NQA-1 criteria within the next two years. Cases for verification of code coupling have not been established. Code verification maturity is rated a 2 pending completion of NQA-1 qualification and establishment of code coupling verification cases.

Solution Verification: Numerical convergence and sufficiency of nodalization schemes has not been examined for the coupled set of codes. On an individual code basis, TRACE has been examined in various time-step size and nodalization studies and could be assigned a maturity level of at least 2 for its use in the primary and secondary side simulations. (A value of 3 for TRACE will depend on nodalization and solution convergence for RCCS simulation.) The maturity of SAM and MAMMOTH has not been investigated and therefore is conservatively rated as 0 simply due to lack of information.

Model Validation: Code assessment, and calibration of some models (bypass and effective conduction in a pebble bed) is not complete. Initial efforts have used the EBR-II tests, and results have shown excellent agreement with data and SAS4A. Additional assessment against integral experiments is on-going or planned. Additional assessment is necessary to demonstrate the ability of SAM and TRACE for RCCS performance, which is considered an important need. The maturity rating is rated a 2 due to the current lack of assessment for RCCS.

The neutronics assessment for fast spectrums is also in progress, but with limited current results it is premature to give as high a rating as for hydraulics. Therefore, MAMMOTH is conservatively rated as a 1 pending completion of neutronic benchmarks applicable to fast spectrums.

Uncertainty Quantification and Sensitivity: Currently not addressed. Long range plans are to use the DAKOTA based uncertainty plug-in available in SNAP to perform uncertainty quantification and automate sensitivity studies. Development of a User interface for MAMMOTH and SAM was been considered but does not appear to 85

represent a difficult task. Because no work has been initiated, the maturity is conservatively rated 0.

The following table provides a quick view of maturity by individual code. Maturity for the set of codes for this application should be based on the lowest rated code for a given element.

Table 4-8. PCMM Characterization of Codes for HPR Analysis Element \ Maturity Maturity Maturity Maturity Comments Maturity Level 0 Level 1 Level 2 Level 3 Representation

• Limited simplification and Geometric required. Geometry SAM MOOSE and major Fidelity FAST TRACE BISON components can be MAMMOTH represented.

• Some uncertainty in modeling of RCCS for SAM, TRACE.

• Physics based Physics and MAMMOTH models available, but SAM MOOSE significant calibration Model Fidelity TRACE using SETs and IETs BISON remains necessary.

FAST

• Verification standards Code TRACE generally applied for MAMMOTH MOOSE each code. Code Verification SAM BISON coupling verification FAST not yet addressed.

• Solution convergence Solution MAMMOTH and nodalization SAM TRACE sufficiency not yet Verification MOOSE established.

BISON FAST

• Solution convergence is not yet known.

MAMMOTH

• Quantitative Model SAM assessment of TRACE accuracy not Validation MOOSE directly relevant BISON FAST

• Only deterministic Uncertainty MAMMOTH analyses possible.

SAM TRACE Uncertainties are not Quantification MOOSE addressed.

and Sensitivity BISON

• TRACE has UQ Analysis FAST functionality built in through SNAP using DAKOTA.

86

4.5 Molten Salt Reactors Molten salt reactors represent the most difficult challenge to modeling and simulation of the various non-LWR designs. The difficulty in modeling MSRs is twofold; first there are a large number of MSR variants and second some MSRs have the fissile material held within the coolant. The MSR designs can be generalized into four types; MSRs in which the fuel is stationary in rods or plates and the coolant is a molten salt, MSRs in which the salt is a coolant but the fuel is TRISO in a pebble bed, MSRs using a fluoride fuel salt in which graphite in the core serves as a moderator for a thermal spectrum, and MSRs with a chloride fuel salt utilizing a fast neutron spectrum. The analytical approach for design basis events for each design type follows. In each of the subsections that follow the modeling and simulation approach is described, along with the gaps that exist in code development and assessment. Each gap is assigned a unique Task number that will be used for tracking progress and associated resource requirements. Tasks are also specified for development of a Reference Plant model of each design. The Reference Plant can be exercised through simulation of normal operation and several hypothetical transients to verify that model and code(s) are ready for safety analysis.

The Reference Plant is intended to approximate, but not be fully representative, of design(s) that may be submitted to the NRC for Design Certification review.

7. MSR (solid fuel in plates or pins, thermal spectrum): The Advanced High Temperature Reactor (AHTR) is a molten sat reactor with solid fuel in the form of plates and is being proposed for MSR insights by ORNL. There are no potential applicants, however the design provides a way to test analysis codes that can then be used for other designs that use a molten salt. The fuel in the AHTR is TRISO contained in a flat plate lattice. The coolant is a molten fluoride salt (FLiBe).

Technical Development Issues The AHTR has been considered by a PIRT panel of experts, and technical issues for simulation of both thermal-fluid phenomena and neutronics identified [11]. In general, the technical issues related to coolant salt reactors with fixed fuel are those of TRISO fuel performance, and thermophysical properties of the coolant salt. Thermo-physical property uncertainty in MSRs is not a code development problem, but rather a problem due to lack of experimental data for the intended range of application.

Primary Modeling and Simulation Approach For this MSR (fixed fuel with a coolant salt), the most straight-forward and cost effective approach would be to use TRACE for the thermal-fluid simulation and PARCS for neutronic evaluation. Cross-sections could be obtained from either SERPENT or SCALE. What makes this option attractive is that updates are available for TRACE for properties of FLiBe and PARCS is already coupled to TRACE. Figure 4-6 shows a schematic of this modeling approach.

87

Figure 4-6. Modeling Approach for Design Basis Event Simulation in a molten salt reactor with stationary fuel.

The alternative approach for a molten salt reactor with fixed fuel would be to use SAM for thermal-fluid analysis and MAMMOTH for neutronics. TRACE would be used for secondary systems, the RCCS and/or DRACS. This alternative approach would be similar to that for the molten salt reactor with pebble bed fuel (i.e. Kairos) without the complication of buoyant pebbles and their possible migration. While the primary approach for this design takes advantage of in-house expertise and full control of code development, this alternative approach is effectively ready for code assessment.

Significant Code Development Gaps The two significant gaps affecting this design are the lack of molten salt properties in TRACE and simulation of the RCCS. Depending on the reactor cavity design and configuration, it may be necessary to couple TRACE with a CFD code in order to simulate natural circulation and heat exchange between the reactor vessel and the RCCS. The unique fuel design (series of flat plates in a hexagonal array pattern) may require some revisions to PARCS. Thus, some of the gaps are:

88

(1) Verify and implement the ORNL provided code updates to TRACE properties for FliBe, and add properties for additional salts if needed for secondary systems. [Task M1]

(2) Update TRACE with properties to include salts intended for secondary side fluids.

[Task M2]

(3) Develop a reactor cavity cooling system model (RCCS) for radiation/convection from the vessel wall to the heat exchanger panel applicable to TRACE. [Tasks S2, S4, M8]

(4) Modify PARCS to model the AHTR fuel geometry. [Task M29]

(5) Development/addition of models to account for salt solidification. [Tasks M30, M31]

(6) Development/addition of models and correlations appropriate for convective heat transfer and hydraulic losses in a molten salt. [Task M3]

Assessment For a molten coolant salt reactor with stationary fuel, codes will need to be assessed against relevant separate effects and integral effects test facilities with salt coolant(s).

There are very few tests that have been conducted with molten salts, and it remains to be shown that these facilities are appropriately scaled to the AHTR. The only large scale facility operated with molten salt was the MSRE. This however was for a fuel salt, rather than a coolant salt.

Assessment for the AHTR should include:

(1) UWisc Flow Loop: Small flow loop(s) in operation at Univ. of Wisconsin may provide data on heat transfer and flow in a molten salt loop. [Task M33]

(2) UMich Flow Loop: A flow loop is under construction at Univ. of Michigan to examine flow and heat transfer in a molten salt loop. These data, when available, can be used to validate hydraulic codes. [Task M34]

(3) ANL RCCS Tests: The Natural convection Shutdown heat removal Test Facility (NSTF) is a large scale thermal hydraulics test facility that has been built at Argonne National Laboratory (ANL). The facility has principally been designed for testing of Reactor Cavity Cooling System (RCCS) concepts that rely on natural convection cooling for either air or water-based systems. [Task H44]

(4) UW RCCS Tests: An RCCS 1/4 scale air-cooled facility constructed at the University of Wisconsin - Madison has been used to compliment the tests from the Argonne NSTF. [Task H2]

(5) HTTR-VCS: The vessel cooling system (VCS) in the HTTR corresponds to the reactor cavity cooling system (RCCS), and it mainly cools the reactor pressure vessel (RPV) by thermal radiation. Tests with the VCS compliment LOFC test without VCS. [Task H45]

89

Plant Model Development (1) Complete SCALE/Shift calculations for cross-sections and reactivity coefficients for fluoride fuel salt. (For ATHR) [Task M16]

(2) Develop reference (AHTR) plant model. The reference model is to be used to verify all aspects of the model and perform initial scoping studies and sensitivity analysis.

[Task M35]

(3) Complete reference plant model simulations, document in a report with statement on code maturity for AHTR analysis with identification of plant specific modeling and remaining assessment requirements. [Task M36]

(4) Complete Code Adequacy to AHTR Report. Include sections on each code utilized.

Verify that code functionality, models and correlations are documented in Theory and User Manuals. [Task M37]

PCMM Characterization A molten salt cooled reactor with stationary fuel is the simplest of MSRs to model and simulate. Thermal-fluid codes must be revised and updated for salt coolant properties and heat transfer and fluid flow correlations specific to salts. Neutronic and fuel modeling can be accomplished using methods being developed for thermal neutron spectrums and TRISO fuel. Because of the relative simplicity of this design, codes could mature quickly and assessment started. However due to the lack of basic information such as salt properties over a wide range and separate effects test data it is currently not possible to assess or calibrate models to match data. Thus, while the intended codes have the maturity to geometrically model the system they cannot be effectively assessed.

Geometric Representation: TRACE is capable of modeling a complex flow system including the secondary system and steam generation loop. Simulation of the AHTR fuel elements (see Figure 4-7) would be difficult using the heat structure modeling currently available in TRACE, however BISON could be used to model the geometry (and TRISO fuel). Some efforts are necessary to model the RCCS system and multidimensional heat structures. Until this is accomplished the set of codes is rated at a maturity level of 2 for geometric representation. PARCS is currently not configured to simulate this particular geometry and therefore is rated at a maturity level of 1.

Physics and Model Fidelity: Updates are available for TRACE to simulate FLiBe and could be updated to include models and correlations that are appropriate for system hydraulics and fluid flow/heat transfer in a molten salt coolant. Models and correlations are available for flow and heat transfer in the reactor cavity in TRACE, although simulation of convection in the reactor cavity will likely require benchmarking against a CFD simulation. No validation has been performed to date and the range of thermal 90

Figure 4-7. AHTR Fuel Assembly.

conditions in the actual design are not yet available. Thus, the maturity level is rated as a 1 since significant calibration has not been performed, nor has there been an internal peer review.

Code Verification: TRACE and PARC are each subjected to a verification test suite consisting of (at least) several hundred cases to test new updates and revisions to the codes. BISON verification appears to satisfy NQA-1. Cases for verification of code coupling have not been established. Code verification maturity is rated a 2 pending completion of NQA-1 qualification and establishment of code coupling verification cases.

Solution Verification: Numerical convergence and sufficiency of nodalization schemes has not been examined for the coupled set of codes. On an individual code basis, each code has been examined in various time-step size and nodalization studies and could be assigned a maturity level of at least 2 for its use in the primary and secondary side simulations. Solution verification for BISON is not known at this time and conservatively ranked as 0 Model Validation: Model assessment does not currently exist for salt coolants. Limited assessment is available for TRISO fuel.

91

Uncertainty Quantification and Sensitivity: Currently not addressed in PARCS or BISON. Long range plans are to use the DAKOTA based uncertainty plug-in available in SNAP to perform uncertainty quantification and automate sensitivity studies. Because no work has been initiated, the maturity is conservatively rated 0. TRACE already has uncertainty and sensitivity capabilities and is rated 2.

The following table provides a quick view of maturity by individual code. Maturity for the set of codes for this application should be based on the lowest rated code for a given element.

92

Table 4-9. PCMM Characterization of Codes for MSR Analysis Element \ Maturity Maturity Maturity Maturity Comments Maturity Level 0 Level 1 Level 2 Level 3 Representation

• BISON able to model and Geometric fuel geometry.

PARCS TRACE BISON Fidelity

• TRACE able to model system, but with possible difficulty in RCCS.

• PARCS development necessary to model geometry.

• Some models and Physics and TRACE correlations are PARCS physics based and Model Fidelity BISON calibrated to data

• Verification standards Code TRACE generally applied for PARCS TRACE/PARCS, but Verification BISON not to NQA-1 standards.

• Coupling needs to be verified

• Numerical effects quantitatively Solution estimated to be small BISON TRACE Verification PARCS

• Coupling needs to be verified.

• Quantitative assessment of Model accuracy not directly TRACE BISON Validation PARCS relevant

• Large or unknown experimental uncertainties

• Only deterministic Uncertainty analyses possible.

PARCS TRACE Uncertainties are not Quantification BISON addressed.

and Sensitivity

• TRACE has UQ Analysis functionality built in through SNAP using DAKOTA.

93

8. MSPR (pebble bed, fluoride coolant salt, thermal spectrum): One applicant (Kairos) is pursuing a novel design that has TRISO fuel in the form of pebbles in a salt cooled reactor with a thermal neutron spectrum. The TRISO pebbles (which will be smaller than those in gas-cooled pebble bed designs) are buoyant in the coolant salt.

Technical Development Issues This type of design is closely related to other coolant salt reactors such as the Advanced High Temperature Reactor (AHTR). A PIRT has been developed for (fluoride) coolant salt MSRs [11] and issues associated with neutronics and thermal-fluid behavior discussed, however there are no PIRTs specifically for a pebble bed reactor cooled by a fluoride salt coolant. In general, the technical issues related to coolant salt reactors with fixed fuel are those of TRISO fuel performance [7], and thermophysical properties of the coolant salt. Thermo-physical property uncertainty in MSRs is not a code development problem, but rather a problem due to lack of experimental data for the intended range of application. Currently however, there are no PIRTs specific to a molten salt cooled pebble bed reactor and PIRT is recommended once the design specifics are known.

Primary Modeling and Simulation Approach For an advanced non-LWR with a liquid coolant (molten salt or liquid metal) the SAM code is well suited to simulate thermal-fluid phenomena. Similar to the approach for pebble bed gas-cooled reactors, updates are available for SAM in order to a pebble bed core. The models are similar to the Ergun and Zehner-Schlunder correlations in PRONGHORN, but are more consistent with models by IAEA for Gen-IV applications.

An important distinction exists for the molten salt cooled pebble bed reactor. This is due to the buoyancy of the TRISO pebbles in the molten salt. The pebbles can be mobile and this mobility may affect the pressure drop through the bed. The component of heat transfer by thermal radiation may be small or negligible in a coolant salt.

SAM already has properties for LiFBeF2 (same as TRACE) and others could be added as needed. SAM also includes properties for Dowtherm A, which has been shown to be an excellent simulant fluid for a high temperature molten salt. In models of pebble bed reactors with a coolant salt, SAM will also be used to simulate the core and reactor vessel.

For core neutronics, MAMMOTH can be used in a manner similar to the approach for gas-cooled reactors. At least initially, we would assume that the pebble bed is static within the molten salt. The salt flows around the pebbles, but the pebbles remain in the same location. Sensitivity studies will likely be needed to investigate the effect of pebble motion.

The secondary system, steam supply system and RCCS can be simulated with TRACE.

While TRACE has been coupled to the MOOSE framework, some testing will be necessary to validate coupling to SAM at two separate locations (RCCS and in-vessel 94

Figure 4-8. Modeling Approach for Design Basis Event Simulation in a Pebble Bed Salt-Cooled Reactor.

heat exchanger). Figure 4-8 shows a schematic of modeling for a pebble bed salt-cooled reactor.

For a pebble bed, molten salt cooled reactor such as this, there are two alternative approaches if the primary approach is determined to be unfeasible. One approach would be to use PRONGHORN for the core and vessel thermal-fluid simulation.

PRONGHORN currently has the capability to simulate molten salts, with the properties for LiFBeF2 having been added. Since PRONGHORN is already coupled to SERPENT/RATTLESNAKE, the capability to simulate core neutronics is feasible.

In addition, coupling to a CFD code such as FLUENT or Nek5000 may be necessary to simulate regions that are pools and lack significant wall drag.

Significant Code Development Gaps Simulation of a pebble bed molten (coolant) salt reactor combines many of the requirements and gaps associated with pebble bed gas-cooled reactors as well as those for molten salt (coolant) reactors. This includes, but is not limited, to thermo-physical properties for molten salts, and pressure drop and heat transfer through a pebble bed core. Buoyancy driven pebble flow for example, is a process that may need to be 95

considered. Additional gaps may be identified as the MSR designs become more clearly defined. Some of the gaps are:

1) Develop a PIRT for a pebble bed molten salt cooled reactor. Insights from the molten salt reactor PIRTs and the TRISO fuel PIRT provide initial identification of important processes, however because of the buoyancy of pebbles in a molten salt, a separate PIRT should be developed. [Task M32]

2) Verify and implement the ORNL provided code updates to TRACE properties for FliNak, and add properties for additional salts if needed for secondary systems.

[Task M1, M2]

3) Develop a reactor cavity cooling system model (RCCS) for radiation/convection from the vessel wall to the heat exchanger panel applicable to TRACE. Couple TRACE and a CFD code if necessary to properly simulate natural circulation and heat exchange between the reactor vessel and the RCCS. [Tasks S2, S4, M8]

4) Evaluation and possible development of models and correlations for heat transfer and fluid flow in a pebble bed composed of buoyant fuel pebbles. (The pebbles intended for Kairos are smaller than those of other TRISO pebble bed designs, and it is not readily clear if conventional correlations for pebble bed flow are applicable.)

[Task M3, M23]

5) Scaling evaluation to justify the use of Dowtherm-A as a simulant fluid to represent a high temperature molten salt. This is necessary to justify CIET as a valued assessment case. [Task M11]

6) Development/addition of models to account for salt solidification. [Task M30, M31]

7) Improvement to SAM heat transfer modeling capabilities (connect 1D pipe to 3D structure, couple porous medium fluid equation to solid), effective thermal conductivity model. [Task M6, H5]

Assessment Assessment for a molten coolant pebble bed salt reactor should include:

(1) CIET: CIET is a scaled model of a system close to that of the Kairos design. It was scaled using Dowtherm-A as the coolant, and is reportedly a close simulant of a high temperature molten salt. [Task M11]

(2) UWisc Flow Loop: Small flow loop(s) in operation at Univ. of Wisconsin may provide data on heat transfer and flow in a molten salt loop. [Task M9]

(3) UMich Flow Loop: A flow loop is under construction at Univ. of Michigan to examine flow and heat transfer in a molten salt loop. These data, when available can be used to validate hydraulic codes. [Task M10]

96

(4) ANL RCCS Tests: The Natural convection Shutdown heat removal Test Facility (NSTF) is a large scale thermal hydraulics test facility that has been built at Argonne National Laboratory (ANL). The facility has principally been designed for testing of Reactor Cavity Cooling System (RCCS) concepts that rely on natural convection cooling for either air or water-based systems. [Task H44]

(5) UW RCCS Tests: An RCCS 1/4 scale air-cooled facility constructed at the University of Wisconsin - Madison has been used to compliment the tests from the Argonne NSTF. [Task H2]

(6) HTTR-VCS: The vessel cooling system (VCS) in the HTTR corresponds to the reactor cavity cooling system (RCCS), and it mainly cools the reactor pressure vessel (RPV) by thermal radiation. Tests with the VCS compliment LOFC test without VCS. [Task H45]

(7) SANA: The SANA tests simulated heat transfer and coolant flow in a pebble bed geometry and represents an important assessment of the simulation of convection/conduction/radiation in a packed bed. [Task H9]

Plant Model Development (1) Complete SERPENT calculations for cross-sections and reactivity coefficients for fluoride fuel salt. (For MSPR) [Task M38]

(2) Develop reference (MSPR) plant model. The reference model is to be used to verify all aspects of the model and perform initial scoping studies and sensitivity analysis.

[Task M23]

(3) Complete reference plant model simulations, document in a report with statement on code maturity for MSPR analysis with identification of plant specific modeling and remaining assessment requirements. [Task M39]

(4) Complete Code Adequacy to MSPR Report. Include sections on each code utilized.

Verify that code functionality, models and correlations are documented in Theory and User Manuals. [Task M22]

PCMM Characterization Although there is a lack of design information, judgement can be made regarding code readiness based on available information.

Geometric Representation: In general, geometric representation of the fuel and core region can be accomplished as in the simulation of the HTR-10 assessment case using PRONGHORN and MAMMOTH. The reduced order CFD capability in SAM enables it to simulate large fluid regions, such as that in the reactor cavity. TRACE is capable of modeling a complex flow system such as the secondary and steam generation system.

TRACE should be applicable without additional code development. Some efforts are necessary to model the RCCS system and refine multidimensional heat structures.

Until this is accomplished the set of codes is rated at a maturity level of 2 for geometric 97

representation. As in HTR-10 MAMMOTH can be used to simulate the power in a pebble bed.

Physics and Model Fidelity: Both SAM, PRONGHORN and TRACE contain models and correlations that are appropriate for system hydraulics and fluid flow/heat transfer in a loop. Both SAM and PRONGHORN contain models for coolant salts, while updates are available but not yet installed into TRACE. Some modification will be necessary to simulate convective heat transfer and fluid flow in a molten salt, and the porous media model will need to be examined for application to a pebble bed where the pebbles are buoyant. Due to uncertainty in models and correlations the maturity level is rated as a 1.

Code Verification: TRACE, SAM, PRONGHORN, and MAMMOTH are each subjected to a verification test suite consisting of (at least) several hundred cases to test new updates and revisions to the codes. Plans are in place to qualify SAM, PRONGHORN, and MAMMOTH as satisfying NQA-1 criteria within the next two years. Cases for verification of code coupling have not been established. Code verification maturity is rated a 2 pending completion of NQA-1 qualification and establishment of code coupling verification cases.

Solution Verification: Numerical convergence and sufficiency of nodalization schemes has not been examined for the coupled set of codes. On an individual code basis, TRACE has been examined in various time-step size and nodalization studies and could be assigned a maturity level of at least 2 for its use in the primary and secondary side simulations. (A value of 3 for TRACE will depend on nodalization and solution convergence for RCCS simulation.) The maturity of SAM, PRONGHORN, and MAMMOTH has not been investigated and therefore is conservatively rated as 0 simply due to lack of information.

Model Validation: Model assessment does not currently exist for salt coolants. Limited assessment is available for TRISO fuel and flow in a porous bed so that PRONGHORN, MAMMOTH, and BISON can be rated at 1. The maturity rating is rated a 1 due to the some applicable assessment for SAM, while for secondary systems sufficient assessment justifies a 2 for TRACE.

Uncertainty Quantification and Sensitivity: Currently not addressed. Long range plans are to use the DAKOTA based uncertainty plug-in available in SNAP to perform uncertainty quantification and automate sensitivity studies. Development of a User interface for MAMMOTH and SAM was been considered but does not appear to represent a difficult task. Because no work has been initiated, the maturity is conservatively rated 0.

The following table provides a quick view of maturity by individual code. Maturity for the set of codes for this application should be based on the lowest rated code for a given element.

98

Table 4-10. PCMM Characterization of Codes for MSPR Analysis Element \ Maturity Maturity Maturity Maturity Comments Maturity Level 0 Level 1 Level 2 Level 3 Representation

• Limited simplification and Geometric required. Geometry SAM BISON and major Fidelity TRACE PRONGHORN components can be MAMMOTH represented, with some uncertainty in RCCS modeling.

MAMMOTH

• Some models and SAM correlations are Physics and physics based and PRONGHORN Model Fidelity TRACE calibrated to data BISON

• Verification standards Code TRACE generally applied for MAMMOTH each code. Code Verification SAM coupling verification PRONGHORN not yet addressed.

BISON

• Solution convergence Solution MAMMOTH and nodalization SAM TRACE sufficiency not yet Verification PRONGHORN established.

BISON

• Solution convergence is not yet known.

• Quantitative Model MAMMOTH assessment of PRONGHORN TRACE accuracy not Validation BISON directly relevant SAM

• Only deterministic Uncertainty analyses possible.

MAMMOTH TRACE Uncertainties are not Quantification SAM addressed.

and Sensitivity BISON

• TRACE has UQ Analysis functionality built in through SNAP using DAKOTA.

99

9. MFSR (fluoride fuel salt, thermal/epithermal spectrum): Several potential applicants are designing molten fuel salt reactors including Terrestrial, Flibe, and Thorcon. Terrestrial, Thorcon and Flibe intend to operate with a thermal neutron spectrum.

Technical Development Issues Fuel salt reactors have fissile material as part of the coolant with a number of unique technical challenges. Unlike conventional fuel, the delayed neutrons in a molten fuel salt reactor are not emitted at the location where fission occurred. The delayed neutrons, important to control of a thermal reactor, can be emitted in the core or elsewhere in the primary system or vessel. The salt itself presents several challenges, starting with thermophysical properties that have high uncertainty. Composition of the salt can change during operation, as fission products accumulate and insoluble species separate from the liquid. Salts have relatively high melting points, and solidification may be an issue for overcooling events. Technical issues associated with molten fuel salt reactors are discussed in several references [12-14].

Primary Modeling and Simulation Approach The primary approach is to use SAM for the core and vessel thermal-fluid analysis. For an advanced non-LWR with a liquid coolant (molten salt or liquid metal) the SAM code is well suited to simulate thermal-fluid phenomena. SAM already has properties for LiFBeF2 (same as TRACE) and others could be added as needed. Some of the molten fuel salt designs contain relatively open regions which may be simulated with the CFD-like capability in SAM.

Since MAMMOTH has already been coupled to SAM through the MOOSE framework, it can be used for neutronic evaluation of the core and primary vessel. Additional modifications may be needed to MAMMOTH (or any other neutronics code) and SAM so that pre-cursor drift is accounted for. (Pre-cursor drift can currently be simulated using the point kinetics model in SAM.) While the initial approach may be to use point kinetics, eventually it will be necessary to evaluate the 3D power distribution. Because of pre-cursor drift fission may not be limited to the core where graphite structures serve as the moderator.

TRACE, currently coupled through the MOOSE framework, can be used to simulate the secondary systems and the RCCS. Coupling to SAM will need to be tested, but since the nodalization where coupling occurs is static (i.e., fine mesh renodalization is not activated), there should not be major problems.

Figure 4-9 shows a schematic of the modeling approach for a molten fuel salt reactor.

100

Figure 4-9. Modeling Approach for Design Basis Event Simulation in a Molten Fuel Salt Reactor.

Significant Code Development Gaps Molten fuel salt reactors represent a significant modeling and simulation challenge. No codes have previously been developed to simulate processes in a flowing, critical molten fuel. There are several technical gaps common to fuel salt MSRs including:

(1) Implement the ability to use the local concentration of delayed neutron pre-cursors for the analysis of molten salt-fueled reactors (in SAM, and/or MAMMOTH). [Task M14]

(2) Continue development and testing of SAMs reduced-order 3D flow module for the analysis of flow and heat transfer in a fast spectrum MSR core. [Task M8]

(3) Improvement to heat transfer modeling capabilities (connect 1D pipe to 3D structure, couple porous medium fluid equation to solid), effective thermal conductivity model. [Task M6]

(4) Implement molten salt properties into TRACE for secondary side analysis. [Task M1, M2]

(5) Review models and update models and correlations for salt coolants. [Task M3]

101

(6) Perform Monte Carlo calculations (SERPENT and/or SCALE/Shift) to obtain cross-sections and reactivity coefficients for fluoride salt(s). [Task M12]

(7) Development/addition of models to account for salt solidification. [Task M30]

Assessment Assessment for a molten fluoride fuel salt reactor should include:

(1) MSRE: The Molten-Salt Reactor Experiment (MSRE) was a 7.4 MWt experimental molten salt reactor at the Oak Ridge National Laboratory (ORNL). Operational data from MSRE can be used to assess thermal-fluid and neutronic codes for molten salts. [Task M7]

(2) UWisc Flow Loop: Small flow loop(s) in operation at Univ. of Wisconsin may provide data on heat transfer and flow in a molten salt loop. [Task M9]

(3) UMich Flow Loop: A flow loop is under construction at Univ. of Michigan to examine flow and heat transfer in a molten salt loop. These data, when available can be used to validate hydraulic codes. [Task M10]

(4) ANL RCCS Tests: The Natural convection Shutdown heat removal Test Facility (NSTF) is a large scale thermal hydraulics test facility that has been built at Argonne National Laboratory (ANL). The facility has principally been designed for testing of Reactor Cavity Cooling System (RCCS) concepts that rely on natural convection cooling for either air or water-based systems. [Task H44]

(5) UW RCCS Tests: An RCCS 1/4 scale air-cooled facility constructed at the University of Wisconsin - Madison has been used to compliment the tests from the Argonne NSTF. [Task H2]

(6) HTTR-VCS: The vessel cooling system (VCS) in the HTTR corresponds to the reactor cavity cooling system (RCCS), and it mainly cools the reactor pressure vessel (RPV) by thermal radiation. Tests with the VCS compliment LOFC test without VCS. [Task H45]

Plant Model Development (1) Complete SERPENT calculations for cross-sections and reactivity coefficients for fluoride fuel salt. (For MFSR) [Task M12]

(2) Develop reference (MFSR) plant model. The reference model is to be used to verify all aspects of the model and perform initial scoping studies and sensitivity analysis.

[Task M24]

(3) Complete reference plant model simulations, document in a report with statement on code maturity for MSPR analysis with identification of plant specific modeling and remaining assessment requirements. [Task M28]

102

(4) Complete Code Adequacy to MSPR Report. Include sections on each code utilized.

Verify that code functionality, models and correlations are documented in Theory and User Manuals. [Task M21]

PCMM Characterization Because of the lack of design information, uniqueness of the design and its components, only preliminary judgement can be made regarding code readiness. Codes intended for analysis of fuel salt MSRs with a thermal neutron spectrum are currently un-validated.

Modeling of the flow loop and major components is likely to be feasible, however, the main difficulty is in simulation of the power distribution and pre-cursor location in the primary. Since verification processes are in place for some of the codes, these warrant a higher rating, although it is not clear what new and specific verification is necessary the coupling between the neutronic and thermal-fluid codes.

Geometric Representation: In general, geometric representation of the fuel, core, vessel and secondary side can be accomplished with SAM, MAMMOTH, and TRACE. Some development is needed to model the 3D structures that have embedded 1D flow channels. The reduced order CFD capability in SAM enables it to simulate large fluid regions, such as that in the reactor cavity. TRACE is capable of modeling a complex flow system such as the secondary and steam generation system. TRACE should be applicable without significant additional code development. Some efforts are necessary to model the RCCS system and refine multidimensional heat structures. Until this is accomplished the set of codes is rated at a maturity level of 2 for geometric representation.

Physics and Model Fidelity: Both SAM and TRACE contain models and correlations that are appropriate for system hydraulics and fluid flow/heat transfer in a loop. SAM does contain models for coolant salts, while updates are available but not yet installed into TRACE. Some modification will be necessary to simulate convective heat transfer and fluid flow in a molten salt, and models to simulate solidication are necessary. While pre-cursor drift can be simulated with the point kinetics model in SAM, there is high uncertainty due to the lack of experimental information. Due to high uncertainty in models and correlations the maturity level is rated as a 1.

Code Verification: TRACE, SAM, and MAMMOTH are each subjected to a verification test suite consisting of (at least) several hundred cases to test new updates and revisions to the codes. Plans are in place to qualify SAM, and MAMMOTH as satisfying NQA-1 criteria within the next two years. Cases for verification of code coupling have not been established. Code verification maturity is rated a 2 pending completion of NQA-1 qualification and establishment of code coupling verification cases.

Solution Verification: Numerical convergence and sufficiency of nodalization schemes has not been examined for the coupled set of codes. On an individual code basis, TRACE has been examined in various time-step size and nodalization studies and could be assigned a maturity level of at least 2 for its use in the primary and secondary side simulations. (A value of 3 for TRACE will depend on nodalization and solution 103

convergence for RCCS simulation.) The maturity of SAM, and MAMMOTH has not been investigated and therefore is conservatively rated as 0 simply due to lack of information.

Model Validation: Model assessment does not currently exist for fuel salt. The maturity rating is rated a 0 due to the current lack of applicable assessment for SAM and MAMMOTH, while for secondary systems sufficient assessment justifies a 2 for TRACE.

Uncertainty Quantification and Sensitivity: Currently not addressed. Long range plans are to use the DAKOTA based uncertainty plug-in available in SNAP to perform uncertainty quantification and automate sensitivity studies. Development of a User interface for MAMMOTH and SAM was been considered but does not appear to represent a difficult task. Because no work has been initiated, the maturity is conservatively rated 0.

The following table provides a quick view of maturity by individual code. Maturity for the set of codes for this application should be based on the lowest rated code for a given element.

104

Table 4-11. PCMM Characterization of Codes for MFSR Analysis Element \ Maturity Maturity Maturity Maturity Comments Maturity Level 0 Level 1 Level 2 Level 3 Representation

• Limited simplification SAM required. Geometry and and Geometric major components can TRACE Fidelity MAMMOTH be represented, with some uncertainty in RCCS modeling.

MAMMOTH

• Some models and SAM correlations are Physics and physics based and TRACE Model Fidelity calibrated to data

• Verification standards TRACE generally applied for Code each code. Code MAMMOTH Verification SAM coupling verification not yet addressed.

• Solution convergence MAMMOTH and nodalization Solution sufficiency not yet SAM TRACE Verification established.

• Solution convergence is not yet known.

• Judgement only SAM

• Few, if any comparisons Model to measurements in MAMMOTH TRACE Validation similar systems or applications

• Only deterministic analyses possible.

Uncertainty Uncertainties are not MAMMOTH TRACE Quantification SAM addressed.

and Sensitivity

• TRACE has UQ Analysis functionality built in through SNAP using DAKOTA.

105

10. MCSR (chloride fuel salt, fast spectrum): Two potential applicants (TerraPower and Elysium) are proposing a molten fuel salt reactor with a fast neutron spectrum. While details on this design are scarce, it is expected that the core region will contain few structures in order to avoid moderation. Fissile material will be held within a chloride salt.

Technical Development Issues As with thermal spectrum molten fuel salt reactors, molten chloride fast reactors have a number of unique technical challenges. Unlike conventional fuel, the delayed neutrons in a molten fuel salt reactor are not emitted at the location where fission occurred although in a fast reactor the delayed neutron fraction is smaller. The chloride salt presents several challenges, starting with thermophysical properties that have high uncertainty. Composition of the salt can change during operation, as fission products accumulate and insoluble species separate from the liquid. Salts have relatively high melting points, and solidification may be an issue for overcooling events. Technical issues associated with molten fuel salt reactors are discussed in several references [12-14]. Fast spectrum chloride fuel salt reactors differ from other MSRs in that the core is closer to a pool as opposed to a wall-bounded region. Since graphite or another moderator is not needed for a fast spectrum reactor, the special region that is critical lacks solid structures. Simulation of fluid flow requires a near-CFD capability.

Primary Modeling and Simulation Approach The primary approach is to use SAM for the core and vessel thermal-fluid analysis. For an advanced non-LWR with a liquid coolant (molten salt or liquid metal) the SAM code is well suited to simulate thermal-fluid phenomena. SAM will need to be updated with chloride salt properties as those properties become available. Because a MCSR is expected to have a core relatively free of structural material, the CFD-like capability in SAM is necessary.

Since MAMMOTH has already been coupled to SAM through the MOOSE framework, it can be used for neutronic evaluation of the core and primary vessel. Modifications recently made to SAM and MAMMOTH (or any other neutronics code) so that pre-cursor drift is appropriately accounted for must be checked. While the initial approach may be to use point kinetics, eventually it will be necessary to evaluate the 3D power distribution.

TRACE, currently coupled through the MOOSE framework, can be used to simulate the secondary systems and the RCCS. Coupling to SAM will need to be tested, but since the nodalization where coupling occurs is static (i.e., fine mesh renodalization is not activated), there should not be major problems.

Figure 4-10 shows a schematic of the modeling approach for a molten fuel salt reactor.

106

Figure 4-10. Modeling Approach for Design Basis Event Simulation in a Molten Chloride Fuel Salt Reactor.

A alternative option would be to simulate the core and primary vessel with a CFD code such as FLUENT or Nek5000 and couple it to a neutronics code. This would enable the simulation of open regions dominated by turbulent shear while obtaining the power distribution and pre-cursor drift.

Significant Code Development Gaps Molten fuel salt reactors represent a significant modeling and simulation challenge. No codes have previously been developed to simulate processes in a flowing, critical molten fuel. There are several technical gaps common to fuel salt MSRs including:

(1) Implement the ability to use the local concentration of delayed neutron pre-cursors for the analysis of molten salt-fueled reactors (in SAM and/or MAMMOTH). [Task M14]

(2) Continue development and testing of SAMs reduced-order 3D flow module for the analysis of flow and heat transfer in a fast spectrum MSR core. Perform CFD 107

simulations as necessary to verify and validate the reduced-order module. [Task M8, Task M40 ]

(3) Implement molten salt properties into TRACE for secondary side analysis. [Tasks M1, M2]

(4) Perform Monte Carlo calculations (SERPENT and/or SCALE/Shift) to obtain cross-sections and reactivity coefficients for chloride salt(s). [Task M13, M15]

(5) Development/addition of models to account for salt solidification for TRACE and SAM. [Task M30, M31]

Assessment Assessment for a molten chloride fuel salt reactor should include:

(1) MSRE: The Molten-Salt Reactor Experiment (MSRE) was a 7.4 MWt experimental molten salt reactor at the Oak Ridge National Laboratory (ORNL). Operational data from MSRE can be used to assess thermal-fluid and neutronic codes for molten salts. [Task M7]

(2) UWisc Flow Loop: Small flow loop(s) in operation at Univ. of Wisconsin may provide data on heat transfer and flow in a molten salt loop. [Task M9]

(3) UMich Flow Loop: A flow loop is under construction at Univ. of Michigan to examine flow and heat transfer in a molten salt loop. These data, when available can be used to validate hydraulic codes. [Task M10]

(4) ANL RCCS Tests: The Natural convection Shutdown heat removal Test Facility (NSTF) is a large scale thermal hydraulics test facility that has been built at Argonne National Laboratory (ANL). The facility has principally been designed for testing of Reactor Cavity Cooling System (RCCS) concepts that rely on natural convection cooling for either air or water-based systems. [Task H44]

(5) UW RCCS Tests: An RCCS 1/4 scale air-cooled facility constructed at the University of Wisconsin - Madison has been used to compliment the tests from the Argonne NSTF. [Task H2]

(6) HTTR-VCS: The vessel cooling system (VCS) in the HTTR corresponds to the reactor cavity cooling system (RCCS), and it mainly cools the reactor pressure vessel (RPV) by thermal radiation. Tests with the VCS compliment LOFC test without VCS. [Task H45]

Plant Model Development (1) Complete SERPENT calculations for cross-sections and reactivity coefficients for chloride fuel salt. (For MCSR) [Task M13]

108

(2) Develop reference (MCSR) plant model. The reference model is to be used to verify all aspects of the model and perform initial scoping studies and sensitivity analysis.

[Task M25]

(3) Complete reference plant model simulations, document in a report with statement on code maturity for MCSR analysis with identification of plant specific modeling and remaining assessment requirements. [Task M28]

(4) Complete Code Adequacy to MCSR Report. Include sections on each code utilized.

Verify that code functionality, models and correlations are documented in Theory and User Manuals. [Task M21]

PCMM Characterization Because of the lack of design information, uniqueness of the design and its components, only preliminary judgement can be made regarding code readiness. Codes intended for analysis of fuel salt MSRs with a thermal neutron spectrum are currently un-validated.

Modeling of the flow loop and major components is likely to be feasible, however, the main difficulty is in simulation of the power distribution and pre-cursor location in the primary. Since verification processes are in place for some of the codes, these warrant a higher rating, although it is not clear what new and specific verification is necessary the coupling between the neutronic and thermal-fluid codes.

Geometric Representation: In general, geometric representation of the fuel, core, vessel and secondary side can be accomplished with SAM, MAMMOTH, and TRACE. The reduced order CFD capability in SAM enables it to simulate large fluid regions, such as the core and the reactor cavity. TRACE is capable of modeling a complex flow system such as the secondary and steam generation system. TRACE should be applicable without significant additional code development. Some efforts are necessary to model the RCCS system and refine multidimensional heat structures. Until this is accomplished the set of codes is rated at a maturity level of 2 for geometric representation.

Physics and Model Fidelity: Both SAM and TRACE contain models and correlations that are appropriate for system hydraulics and fluid flow/heat transfer in a loop. SAM does contain models for coolant salts, while updates are available but not yet installed into TRACE. Some modification will be necessary to simulate convective heat transfer and fluid flow in a molten salt, and models to simulate solidification are necessary.

While pre-cursor drift can be simulated with the point kinetics model in SAM, there is high uncertainty due to the lack of experimental information. Due to high uncertainty in models and correlations the maturity level is rated as a 1.

Code Verification: TRACE, SAM, and MAMMOTH are each subjected to a verification test suite consisting of (at least) several hundred cases to test new updates and revisions to the codes. Plans are in place to qualify SAM, and MAMMOTH as satisfying NQA-1 criteria within the next two years. Cases for verification of code coupling have not been established. Code verification maturity is rated a 2 pending completion of NQA-1 qualification and establishment of code coupling verification cases.

109

Solution Verification: Numerical convergence and sufficiency of nodalization schemes has not been examined for the coupled set of codes. On an individual code basis, TRACE has been examined in various time-step size and nodalization studies and could be assigned a maturity level of at least 2 for its use in the primary and secondary side simulations. (A value of 3 for TRACE will depend on nodalization and solution convergence for RCCS simulation.) The maturity of SAM, and MAMMOTH has not been investigated and therefore is conservatively rated as 0 simply due to lack of information.

Model Validation: Model assessment does not currently exist for fuel salt. The maturity rating is rated a 0 due to the current lack of applicable assessment for SAM and MAMMOTH, while for secondary systems sufficient assessment justifies a 2 for TRACE.

Uncertainty Quantification and Sensitivity: Currently not addressed. Long range plans are to use the DAKOTA based uncertainty plug-in available in SNAP to perform uncertainty quantification and automate sensitivity studies. Development of a User interface for MAMMOTH and SAM was been considered but does not appear to represent a difficult task. Because no work has been initiated, the maturity is conservatively rated 0.

The following table provides a quick view of maturity by individual code. Maturity for the set of codes for this application should be based on the lowest rated code for a given element.

110

Table 4-12. PCMM Characterization of Codes for MCSR Analysis Element \ Maturity Maturity Maturity Maturity Comments Maturity Level 0 Level 1 Level 2 Level 3 Representation

• Limited simplification SAM required. Geometry and and Geometric major components can TRACE Fidelity MAMMOTH be represented, with some uncertainty in RCCS modeling.

MAMMOTH

• Some models and SAM correlations are Physics and physics based and TRACE Model Fidelity calibrated to data

• Verification standards TRACE generally applied for Code each code. Code MAMMOTH Verification SAM coupling verification not yet addressed.

• Solution convergence MAMMOTH and nodalization Solution sufficiency not yet SAM TRACE Verification established.

• Solution convergence is not yet known.

• Judgement only SAM

• Few, if any comparisons Model to measurements in MAMMOTH TRACE Validation similar systems or applications

• Only deterministic analyses possible.

Uncertainty Uncertainties are not MAMMOTH TRACE Quantification SAM addressed.

and Sensitivity

• TRACE has UQ Analysis functionality built in through SNAP using DAKOTA.

111

4.6 Initial MSR System Inventory Accident scenarios generally involve over-power or under-cooling conditions that are short relative to the period of operation of a plant. For conventional LWRs for example, which operate months between shutdown, a loss-of-coolant accident takes place over only a couple of hours. Some MSR designs however operate in conjunction with fuel cycle operations. Fission products can be withdrawn from the primary while fresh fissile material is added. These chemical batch operations are expected to occur frequently during operation, although no scheme has been proposed to the NRC so far. As operation continues, fission of fissile material produces more fission products. Some fission products are gaseous and collect in the cover gas. Other fission products remain within the liquid salt, unless their concentration exceeds their solubility limit causing them to plate out.

This initial inventory is needed to start a transient calculation. Because the worst time in life for a reactivity related event may not be the same as that for an event involving dispersal of fission products, an inventory control analysis tool is necessary. Figure 4-11 presents a schematic of the primary system - fuel cycle interaction and the processes involved in quantification of the primary side inventory. It is not clear if such a code or tool exists presently.

This code will need to estimate the change in inventory due to decay of fission products in addition to the chemical reactions that may occur.

The key decision involving inventory control is first the identification of a code or tool that can estimate the primary inventory, and an evaluation of that tool to ensure that all of the major processes are accounted for. Currently no codes are available that account for all processes.

A script known as ChemTriton [42] has been developed to simulate equilibrium MSR fuel cycle performance by modeling the changing isotopic composition of an irradiated fuel salt using SCALE for neutron transport and depletion calculations. ChemTriton has been used to simulate the startup of a thorium-based MSR fuel cycle and can predict depletion of the fuel salt and has the capability to model feed and separation operations. It is not clear to the extent that ChemTrition simulates chemical reactions or determines if a species exceeds its solubility limit.

Further, some isotopes may change phase as they decay. Gaseous components are likely to separate into the cover gas system that is intended for some MSRs.

An additional item that ChemTrition may be useful for is quantification of the decay heat loading in gaseous, liquid soluble, and plated out components. Unlike a conventional fuel rod where all of the fission products and all of the decay heat is retained within the cladding, in an MSR the decay heat will be distributed. Xe and Kr will likely separate and migrate into the cover gas, where the head load (and temperature) in the storage tank must be determined. Molten salts freeze at relatively high temperatures (> 700 K). Decay heat may be relied upon to prevent freezing during over-cooling events and/or at cooler regions of the system. A determination of the decay heat within the liquid is necessary.

112

Figure 4-11. Inventory Control Processes in a Molten Salt Reactor.

113

5.0 Summary The proposed code suite for non-LWR analysis has been designed to maximize flexibility and minimize expected code development cost to the NRC. By adopting several DOE developed codes the NRC can leverage the investment by DOE into technologies that have not been central to the NRCs mission to this point. The plan presented in this report has identified and selected codes that can serve multiple purposes, thus allowing the NRC to minimize the impact on staff resources.

This plan identifies many, but not all, of the gaps in code development and assessment for non-LWRs. Because many of the designs are pre-mature, details on the safety systems and on safety significant components are limited. Once the details of a design are known to the staff and information on the safety systems are examined, some additional code development and assessment may be warranted. This is especially true for molten salt cooled reactors in general, and molten fuel salt reactors in particular.

114

6.0 References

[1] USNRC, NRC Vision and Strategy: Safely Achieving Effective and Efficient Non-Light Water Reactor Mission Readiness, ADAMS ML16356A670, December 2016.

[2] USNRC, GUIDANCE FOR DEVELOPING PRINCIPAL DESIGN CRITERIA FOR NON-LIGHT-WATER REACTORS, Regulatory Guide 1.232, ADAMS ML17375A611, April 2018.

[3] USNRC, GUIDANCE FOR A TECHNOLOGY-INCLUSIVE, RISK-INFORMED, AND PERFORMANCE-BASED APPROACH TO INFORM THE CONTENT OF APPLICATIONS FOR LICENSES, CERTIFICATIONS, AND APPROVALS FOR NON-LIGHT-WATER REACTORS, Draft Regulatory Guide 1353, 2018.

[4] USNRC, Transient and Accident Analysis Methods, Regulatory Guide 1.203, ADAMS ML053500170, 2002.

[5] G. E. Wilson and B. E. Boyack, The Role of the PIRT Process in Experiments, Code Development, and Code Applications Associated with Reactor Safety Analysis, Nuclear Engineering and Design, 186, pp 23-27, 1998.

[6] S.J. Ball and S.E. Fisher, Next Generation Nuclear Plant Phenomena Identification and Ranking Tables (PIRTs), ORNL/TM-2007/147, NUREG/CR-6944, Vol. 1, March 2008.

[7] USNRC, TRISO-Coated Particle Fuel Phenomenon Identification and Ranking Tables (PIRTs) for Fission Product Transport Due to Manufacturing, Operations, and Accidents, NUREG/CR-6844, 2004.

[8] R. Schmidt, et al., Sodium Fast Reactor Gaps Analysis of Computer Codes and Models for Accident Analysis and Reactor Safety, SAND2011-4145, 2011.

[9] Lap-Yan Cheng, Michael Todosow, and David Diamond, Phenomena Important in Liquid Metal Reactor Simulations, BNL-207816-2018-INRE, ADAMS Accession No. ML18291B305 (2018).

[10] Hsun-Chia Lin, Sheng Zhang, David Diamond, Stephen Bajorek, Richard Christensen, Yujun Guo, David Holcomb, Graydon Yoder, Shanbin Shi, Qiuping L., Xiaodong Sun, Phenomena Identification and Ranking Table Study for Thermal Hydraulics for Advanced High Temperature Reactor, Annals of Nuclear Energy, 124 (2019) 257-269.

[11] Farzad Rahnema, Xiaodong Sun, Bojan Petrovic, David Diamond, Stephen Bajorek, Yujun Guo, Graydon Yoder, Dingkang Zhang, Paul Burke, Phenomena identification and categorization by the required level of multiphysics coupling in FHR modeling and simulation, Annals of Nuclear Energy, 121 (2018) 540-551.

[12] D. J. Diamond, N. R. Brown, R. Denning and S. Bajorek, Phenomena Important in Modeling and Simulation of Molten Salt Reactors, BNL-114869-2018-IR, Brookhaven National Laboratory, 115

ADAMS ML18124A330, April 23, 2018.

[13] Diamond, D. J., Brown, N. R., Denning, R., Bajorek, S., 2018. Neutronics Phenomena Important in Modeling and Simulation of Liquid-Fuel Molten Salt Reactors, Advances in Thermal Hydraulics (ATH 2018); Nov. 11-15; Orlando, Florida, USA: American Nuclear Society.

[14] Bajorek, S., Diamond, D. J., Brown, N. R., Denning, R., 2018. Thermal-Hydraulics Phenomena Important in Modeling and Simulation of Liquid-Fuel Molten Salt Reactors, Advances in Thermal Hydraulics (ATH 2018); Nov. 11-15; Orlando, Florida, USA: American Nuclear Society.

[15] J. W. Sterbentz, J. E. Werner, M. G. McKellar, A. J. Hummel, J. C. Kennedy, R. N. Wright, J. M. Biersdorf, "Special Purpose Nuclear Reactor (5 MW) for Reliable Power at Remote Sites Assessment Report; Using Phenomena Identification and Ranking Tables (PIRTs)," INL/EXT-16-40741, Revision 1, April 2017.

[16] USNRC, TRACE V5.0 Developmental Assessment Manual, ADAMS ML120060208, ML120060187, ML120060191, ML120060172, 2008.

[17] I. Porter, K. Geelhood, D. Richmond, D. Colameco, E. Torres, W. Luscher and L.

Kyriazidis, "FAST-1.0: A Computer Code for Thermal-Mechanical Nuclear Fuel Analysis under Steady-state and Transients," U.S. Nuclear Regulatory Commission, Washington, DC, Publication Pending.

[18] T. Downar, V. Seker, and A. Ward, "PARCS: Purdue Advanced Reactor Core Simulator,"

Proc. Topl. Mtg. Advances in Nuclear Analysis and Simulation (PHYSOR-2006), Vancouver, Canada, September 10-14, 2006, American Nuclear Society (2006).

[19] B. Rearden and M. Jessee, "SCALE Code System, ORNL/TM-2005/39, Version 6.2.3,"

UT-Battelle, LLC, Oak Ridge National Laboratory, 2018.

[20] D. Gaston, C. Newman, G. Hansen, and D. Lebrun-Grandie. MOOSE: A parallel computational framework for coupled systems of nonlinear equations. Nuclear Engineering and Design, 239:1768-1778, 2009.

[21] R. L. Williamson, J. D. Hales, S. R. Novascone, G. Pastore, D. M. Perez, B. W. Spencer, and R. C. Martineau, Overview of the BISON multidimensional fuel performance code, IAEA Technical Meeting: Modeling of Water-Cooled Fuel Including Design-Basis and Severe Accidents, Chengdu, China, October 28-November 1, 2013.

[22] Idaho National Laboratory, PRONGHORN Manual, July 2017.

[23] Rui Hu, SAM Theory Manual, Argonne National Laboratory, ANL/NE-17/4, March 2017.

[24] Derek R. Gaston, Cody J. Permann, John W. Peterson, Andrew E. Slaughter, David Andrs, Yaqi Wang, Michael P. Short, Danielle M. Perez, Michael R. Tonks, Javier Ortensi, Ling Zou, 116

Richard C. Martineau, Physics-based multiscale coupling for full core nuclear reactor simulation, Annals of Nuclear Energy 84 (Oct. 2015), pp. 45-54.

[25] ANSYS, Inc., ANSYS FLUENT User's Guide, November 2011.

[26] Paul Fischer, et al., Nek5000 User Documentation, ANL/MCS-TM-351, 2015.

[27] Leppanen, J., "SERPENT - A Continuous Energy Monte Carlo Reactor Physics Burnup Calculation Code," VTT Technical Research Centre of Finland, 2015.

[28] Sandia National Laboratories, MELCOR Computer Code Manuals, Vol. 2: Reference Manual, Version 2.2.9541, SAND 2017-0876 O, January 2017 (ADAMS Accession No. ML17040A420).

[29] K. J. Geelhood, et al, FRAPCON-3.4: A Computer Code for the Calculation of Steady-State, Thermal-Mechanical Behavior of Oxide Fuel Rods for High Burnup, PNNL-19418, NUREG/CR-7022, 2010.

[30] K.J. Geelhood, et al, FRAPTRAN 1.4: A Computer Code for the Transient Analysis of Oxide Fuel Rods, PNNL-19400, NUREG/CR-7023, 2011.

[31] T. H. Fanning and J. E. Cahalan, The SAS4A/SASSYS-1 Safety Analysis Code System, Version 5, ANL/NE-16/19, 2017.

[32] Rui Hu and Tyler Sumner, Benchmark Simulations of The Thermal-Hydraulic Responses During EBR-II Inherent Safety Tests Using SAM, Proceedings of the 2016 International Congress on Advances in Nuclear Power Plants, San Francisco, CA, April 2016, 1632-1641.

[33] G. Hu, G. Zhang, and R. Hu, Reactivity Feedback Modeling in SAM, ANL-NSE-19/1, February 2019.

[34] T. Downar, V. Seker, T. Drzewiecki, B. Kochunas, AGREE/TRIPEN v1.0: U.S. NRC Core Thermal Fluids and Neutronics Simulator for NGNP. Theory Manual, October 23, 2015.

ML15300A234.

[35] W. L. Oberkampf et al., Predictive Capability Maturity Model for Computational Modeling and Simulation, SAND2007-5948, 2007.

[36] Zou et al., Validation of Pronghorn with the SANA Experiments, INL/EXT-17-44085, Dec.

2017.

[37] Suikkanen, H., Rintala, V. and Kyrki-Rajamki, R. (2010) "An approach for detailed reactor physics modelling of randomly packed pebble beds." In proc. HTR 2010, Prague, Czech Republic, Oct. 18-20, 2010.

117

[38] Rintala, V., Suikkanen, H., Leppnen, J. and Kyrki-Rajamki, R. (2015) "Modeling of realistic pebble bed reactor geometries using the Serpent Monte Carlo code." Ann. Nucl.

Energy, 77 (2015) 223-230.

[39] Ortensi et al., Benchmark Analysis of the HTR-10 with the MAMMOTH Reactor Physics Application, INL/EXT-18-45453, June 2018.

[40] S. J. Ball, Next Generation Nuclear Plant GAP Analysis Report, ORNL/TM-2007/228, 2008.

[41] Wadim Jaeger, Victor Hugo Sanchez Espinoza, Improvements and Validation of the System Code TRACE for Lead and Lead-Alloy Cooled Fast Reactors Safety-Related Investigations, NUREG/IA-0421, ADAMS ML13044A244, 2013.

[42] Benjamin R. Betzler, Jeffrey J. Powers, Andrew Worrall, Molten salt reactor neutronics and fuel cycle modeling and simulation with SCALE, Annals of Nuclear Energy 101 (2017) 489-503.

118