UC Berkeley

The development of advanced nuclear reactor technology requires understanding of complex, integrated systems that exhibit novel phenomenology under normal and accident conditions. The advent of passive safety systems and enhanced modular construction methods requires the development and use of new frameworks to predict the behavior of advanced nuclear reactors, both from a safety standpoint and from an environmental impact perspective. This dissertation introduces such frameworks for scaling of integral effects tests for natural circulation in fluoride-salt-cooled, high-temperature reactors (FHRs) to validate evaluation models (EMs) for system behavior; subsequent reliability assessment of passive, natural- circulation-driven decay heat removal systems, using these validated models; evaluation of life cycle carbon dioxide emissions as a key environmental impact metric; and recommendations for further work to apply these frameworks in the development and optimization of advanced nuclear reactor designs. In this study, the developed frameworks are applied to the analysis of the Mark 1 pebble-bed FHR (Mk1 PB-FHR) under current investigation at the University of California, Berkeley (UCB).

The capability to validate integral transient response models is a key issue for licensing new reactor designs. This dissertation presents the scaling strategy, design and fabrication aspects, and startup testing results from the Compact Integral Effects Test (CIET) facility at UCB, which reproduces the thermal hydraulic response of an FHR under forced and natural circulation operation. CIET provides validation data to confirm the performance of the direct reactor auxiliary cooling system (DRACS) in an FHR, used for natural-circulation-driven decay heat removal, under a set of reference licensing basis events, as predicted by best-estimate codes such as RELAP5-3D. CIET uses a simulant fluid, Dowtherm A oil, which at relatively low temperatures (50-120°C) matches the Prandtl, Reynolds, Froude and Grashof numbers of the major liquid salts simultaneously, at approximately 50% geometric scale and heater power under 2% of prototypical conditions. The studies reported here include isothermal pressure drop tests performed during startup testing of CIET, with extensive pressure data collection to determine friction losses in the system, as well as subsequent heated tests, from parasitic heat loss tests to more complex feedback control tests and natural circulation experiments. For initial code validation, coupled steady-state single-phase natural circulation loops and simple forced cooling transients were conducted in CIET. For various heat input levels and temperature boundary conditions, fluid mass flow rates and temperatures were compared between RELAP5- 3D results, analytical solutions when available, and experimental data. This study shows that RELAP5-3D provides excellent predictions of steady-state natural circulation and simple transient forced cooling in CIET. The code predicts natural circulation mass flow rates within 8%, and steady-state and transient fluid temperatures, under both natural and forced circulation, within 2°C of experimental data, suggesting that RELAP5-3D is a good EM to use to design and license FHRs.

A key element in design and licensing of new reactor technology lies in the analysis of the plant response to a variety of potential transients. When applicable, this involves understanding of passive safety system behavior. This dissertation develops a framework to assess reliability and propose design optimization and risk mitigation strategies associated with passive decay heat removal systems, applied to the Mk1 PB-FHR DRACS. This investigation builds upon previous detailed design work for Mk1 components and the use of RELAP5-3D models validated for FHR natural circulation phenomenology. For risk assessment, reliability of the point design of the passive safety system for the Mk1 PB-FHR, which depends on the ability of various structures to fulfill their safety functions, is studied. Whereas traditional probabilistic risk assessment (PRA) methods are based on event and fault trees for components of the system that perform in a binary way – operating or not operating –, this study is mostly based on probability distributions of heat load compared to the capacity of the system to remove heat, as recommended by the reliability methods for passive safety functions (RMPS) that are used here. To reduce computational time, the use of response surfaces to describe the system in a simplified manner, in the context of RMPS, is also demonstrated. The design optimization and risk mitigation part proposes a framework to study the elements of the design of the reactor, and more specifically its passive safety cooling system, which can contribute to enhanced reliability of heat removal under accident conditions. Risk mitigation measures based on design, startup testing, in-service inspection and online monitoring are proposed to narrow probability distributions of key parameters of the system and increase reliability and safety.

Another major aspect in the development of novel energy systems is the assessment of their impacts on the environment compared to current technologies. While most existing life cycle assessment (LCA) studies have been applied to conventional nuclear power plants, this dissertation proposes a framework to extend such studies to advanced reactor designs, using the example of the Mk1 PB-FHR. The Mk1 uses a nuclear air-Brayton combined cycle designed to produce 100 MWe of base-load electricity when operated with only nuclear heat, and 242 MWe using natural gas co-firing for peaking power. The Mk1 design provides a basis for quantities and costs of major classes of materials involved in building the reactor and fabricating fuel, and operation parameters. Existing data and economic input-output LCA models are used to calculate greenhouse gas emissions per kWh of electricity produced over the life cycle of the reactor. Baseline life cycle emissions from the Mk1 PB-FHR in base-load configuration are 26% lower than average Generation II light water reactors in the U.S., 98% lower than average U.S. coal plants and 96% lower than average U.S. natural gas combined cycle plants using the same turbine technology. In peaking configuration, due to its nuclear component and higher thermal efficiency, the Mk1 plant only produces 32% of the emissions of average U.S. gas turbine simple cycle peaking plants. One key contribution to life cycle emissions results from the amount and type of concrete used for reactor construction. This is an incentive to develop innovative construction methods using optimized steel-concrete composite wall modules and new concrete mixes to reduce life cycle emissions from the Mk1 and other advanced reactor designs.