Investigation of core-power control based on fuzzy predictive control and Kalman filter for a fast-spectrum lithium-cooled space nuclear reactor

The intelligent control of reactor core power is critical for the safety and reliability of space nuclear reactors, which operate in highly uncertain cosmic environments. Focusing on a lithium-cooled space reactor, this study proposed a fuzzy-model-based model predictive controller (FM-MPC-KF) based on a Takagi-Sugeno (T-S) fuzzy model and Kalman filter for core power regulation. To address system nonlinearities, the T-S fuzzy approach was employed to approximate the core dynamics. A Kalman filter was incorporated to estimate the reactor core states under limited sensor availability and the presence of measurement noise. The FM-MPC-KF controller was evaluated through comprehensive simulations covering ramp and step transients, and compared with Fuzzy Model Predictive Control (FMPC). The results demonstrate that FM-MPC-KF offers superior dynamic performance and enhanced load-following capability compared to the standard FMPC. Furthermore, the unified fuzzy prediction framework conceptually simplifies the multi-controller architecture, establishing its proof-of-concept algorithmic feasibility as a potential candidate for space reactor control. Moreover, with the integrated Kalman filter, FM-MPC-KF exhibits notably better baseline noise immunity than FMPC against Gaussian, Poisson, and salt-and-pepper measurement noise. This study underscores the potential of combining T-S fuzzy modeling, MPC, and the Kalman filter for the advanced control of space nuclear reactors under complex and uncertain conditions.