Hybrid core power control using PI, Fuzzy and MPC for TRIGA nuclear reactor

Abstract

A nuclear reactor is a complex system that equipped highly complicated controllers to maintain multiple parameters within the safety limit and the core power output continuously stable. Even though many controllers have been designed to enhance a nuclear reactor’s power tracking performance, the hybrid-switched control structure has yet to be extensively explored. An inherent limitation in single-type controller utilization can be overcome by the integration of multiple types of controllers within a single control system. The present study aimed to investigate a hybrid switching controller for controlling the core power at the reactor using a combination of PI, Fuzzy, and MPC. A static PI controller is insufficient to handle the non-linear plant, a fuzzy rule-base relies on tuning optimization, which takes a long time to perform manually, and model predictive control (MPC) usually requires a significant amount of computational time. With the 10th order linearized reactor core model, the scope is limited to a 30% error deviation for switching activation. The performance of the controllers is compared via simulation in terms of tracking performance including settling time, rise time, per cent overshoot, chattering error, workload, and energy released. Overall, the results show that the response from the single and hybrid Fuzzy controllers offers better results than the PI/hybrid PI, which reduces the chattering error during steady-state by up to 29%, the settling time by up to 14%, and the per cent overshoot up to 12%.