Robust Adaptive Power Control of a Nuclear Reactor With Control Time Delay

Operating pressurized water reactor (PWR) nuclear power plants requires effective regulation of the control rod speed to achieve the desired reactor power demand. However, the presence of time delay in the control rod speed due to latency in communication and aging can lead to output power oscillation and instability. This article proposes a backstepping fractional-order proportional-integral-derivative (PID)-based sliding mode control (SMC) with a finite-time convergence for power control of a PWR with control time delay compensated using an auxiliary signal to mitigate any power oscillation. Due to the simple structure and approximation accuracy of the multi-dimensional Taylor network (MDTN), it is employed to identify the complex dynamics of the reactor. The MDTN approximation error is attenuated using an adaptive law to improve the approximation accuracy. The bended tracking error (BTE) scheme is used to eliminate the SMC reaching phase so that the output power can track the desired level at the initial time in the sliding phase, thereby ensuring robustness and convergence within a preassigned settling time. The finite-time stability of the closed-loop system is demonstrated through Lyapunov-based analysis. Simulation results have shown that the proposed power control achieved a root-mean-square error (RMSE) of 0.002305, which is superior compared to high-order SMC (HOSMC) and radial basis function neural network-based event-triggered fixed time (RBFNN-ET-FT) control that achieved an RMSE of 1.745 and 0.880, respectively.