Supervisory Control of Nuclear Power Plants Using Finite State Automata

Reducing operation and maintenance expenses associated with energy generation costs for small reactors and microreactors is necessary for their economic viability. One avenue for reducing these expenses is the implementation of autonomous reactor control systems. Advanced reactor designs inherently support reduced human operator interventions and incorporate enhanced passive safety features, reduced source terms, and digital instrumentation and control systems. Thus, the introduction of supervisory control systems (SCSs) for dynamic operations creates an opportunity to further reduce dependence on human operators and fosters confidence in advanced control systems. SCSs, as a transition technology, serve as a strategic pathway toward fully autonomous nuclear power plant operations. Finite state automata provide a fully verifiable and validatable supervisory control framework and thereby facilitate the transition to autonomous control systems, serving as a foundational mathematical tool for modeling discrete event systems (DESs). Furthermore, a DES can directly model both nuclear power plant systems and their operational procedures, which can be arranged within a hierarchical architecture using a graded functionalization approach already employed in complex dynamic systems like self-driving vehicles. Here, low-level actuation is accomplished via feedback controllers, and high-level state transitions are controlled by the supervisory controller. In the absence of licensed advanced reactor designs and approved nuclear autonomous control systems, legacy light water reactors (LWRs) can be used as a case study for the phased-implementation of advanced control systems because they provide realistic feedback and controls within a more conservative operating margin than the proposed advanced reactors. A generic supervisory controller was developed as a transition technology toward autonomous nuclear power plant operations. This controller was then tailored and scaled for implementation on LWR simulators. The lessons learned from incorporating automated processes can guide the framework development for autonomous processes. Incorporating a holistic and scalable approach via transitional steps toward autonomy provides an opportunity for validating these processes while actively engaging the nuclear industry in the shift away from reliance on human operators.