Nuclear Thermal Rocket Engine Control Autonomy Via Embedded Decision

This doctoral dissertation presents an investigation of embedded decision capabilities as a means for developing nuclear reactor autonomous control. Nuclear thermal propulsion (NTP) is identified as a high priority technology for development, and is the focus of this research. First, a background investigation is presented on the state of the art in nuclear thermal rocket (NTR) engine control and modeling practices, resulting in the development of a low order NTR engine dynamic model based on the literature. The engine model was used to perform the following investigation, and is intended to serve as a research platform for the future development of autonomous control in NTR engines. Next, since embedded decision techniques are argued to be the basis for autonomous control, additional background is provided on autonomy and embedded decision. This background served as the basis for this investigation, resulting in the first application of utility based decision in an NTR engine control system (ECS). The decision system was developed to accommodate faults detected in the primary control system, and is based on expected actuator availability. The behavior resulting from the actuator availability attribute was considered overly sensitive to perceived faults, and attributes based on engine performance are recommended for further development of the utility based decision approach. In addition to the research platform and the findings it produced, this work resulted in a collection of tools for future use in further research of engine modeling, control, and applications of embedded decision to NTR engine control.