Multivariate model predictive control study of large scale PWR nuclear power plant based on the asymptotic method

Third-generation PWR nuclear power plants face increased grid peaking tasks, demanding enhanced load following capabilities of the reactor control system. The MPC algorithm, owing to its advantage, more effectively addresses grid peaking challenges. Therefore, a model predictive control algorithm based on the asymptotic method is utilized to research advanced control of PWR models. Initially, the paper conducts identification experiments based on asymptotic identification method and the characteristics of PWR. A reduced-order model is derived from a higher-order model based on identification data and objective function minimization. Model order selection is guided by the asymptotic criterion. The identified model demonstrates good performance in the frequency domain according to the model validation theory of asymptotic method. In time domain validation, the error percentage of the model output is low and the step response exhibits correct direction. Subsequently, a model predictive controller is designed based on the identification model, and simulation analysis is conducted for four working conditions respectively. Under load step conditions, the MPC controller markedly reduced both regulation time and steady-state error compared to the original controller, achieving a complete elimination of steady-state error and over 75% reduction in regulation time for average coolant temperature control. Concurrently, under conditions of load ramping, the MPC controller achieved a significant reduction of over 99% in the ITSE as compared to the original controller. However, it should be noted that the output frequency and intensity are higher for MPC controller compared to the original controller.