UC Berkeley

This dissertation treats system design, modeling of transient system response, and characterization of individual phenomena and demonstrates a framework for integration of these three activities early in the design process of a complex engineered system. A system analysis framework for prioritization of experiments, modeling, and development of detailed design is proposed. Two fundamental topics in thermal-hydraulics are discussed, which illustrate the integration of modeling and experimentation with nuclear reactor design and safety analysis: thermal-hydraulic modeling of heat generating pebble bed cores, and scaled experiments for natural circulation heat removal with Boussinesq liquids. The case studies used in this dissertation are derived from the design and safety analysis of a pebble bed fluoride salt cooled high temperature nuclear reactor (PB-FHR), currently under development in the United States at the university and national laboratories level.

In the context of the phenomena identification and ranking table (PIRT) methodology, new tools and approaches are proposed and demonstrated here, which are specifically relevant to technology in the early stages of development, and to analysis of passive safety features. A system decomposition approach is proposed. Definition of system functional requirements complements identification and compilation of the current knowledge base for the behavior of the system. Two new graphical tools are developed for ranking of phenomena importance: a phenomena ranking map, and a phenomena identification and ranking matrix (PIRM). The functional requirements established through this methodology were used for the design and optimization of the reactor core, and for the transient analysis and design of the passive natural circulation driven decay heat removal system for the PB-FHR.

A numerical modeling approach for heat-generating porous media, with multi-dimensional fluid flow is presented. The application of this modeling approach to the PB-FHR annular pebble bed core cooled by fluoride salt mixtures generated a model that is called Pod. Pod was used to show the resilience of the PB-FHR core to generation of hot spots or cold spots, due to the effect of buoyancy on the flow and temperature distribution in the packed bed. Pod was used to investigate the PB-FHR response to ATWS transients. Based on the functional requirements for the core, Pod was used to generate an optimized design of the flow distribution in the core.

An analysis of natural circulation loops cooled by single-phase Boussinesq fluids is presented here, in the context of reactor design that relies on natural circulation decay heat removal, and design of scaled experiments. The scaling arguments are established for a transient natural circulation loop, for loops that have long fluid residence time, and negligible contribution of fluid inertia to the momentum equation. The design of integral effects tests for the loss of forced circulation (LOFC) for PB-FHR is discussed. The special case of natural circulation decay heat removal from a pebble bed reactor was analyzed. A way to define the Reynolds number in a multi-dimensional pebble bed was identified. The scaling methodology for replicating pebble bed friction losses using an electrically resistance heated annular pipe and a needle valve was developed. The thermophysical properties of liquid fluoride salts lead to design of systems with low flow velocities, and hence long fluid residence times. A comparison among liquid coolants for the performance of steady state natural circulation heat removal from a pebble bed was performed.

Transient natural circulation experimental data with simulant fluids for fluoride salts is given here. The low flow velocity and the relatively high viscosity of the fluoride salts lead to low Reynolds number flows, and a low Reynolds number in conjunction with a sufficiently high coefficient of thermal expansion makes the system susceptible to local buoyancy effects Experiments indicate that slow exchange of stagnant fluid in static legs can play a significant role in the transient response of natural circulation loops. The effect of non-linear temperature profiles on the hot or cold legs or other segments of the flow loop, which may develop during transient scenarios, should be considered when modeling the performance of natural circulation loops. The data provided here can be used for validation of the application of thermal-hydraulic systems codes to the modeling of heat removal by natural circulation with liquid fluoride salts and its simulant fluids.