Adaptive predictive control of average coolant temperature based on chaotic particle swarm optimization

Abstract

Aiming at the problem that the average temperature control system of nuclear reactor coolant under variable conditions is nonlinear and time-varying, and the traditional fixed model cannot meet the control requirements, an improved generalized predictive control method with adaptive identification of model parameters is proposed. First, the structure of the coolant average temperature model is determined through mechanism derivation, and the forgetting factor recursive least square algorithm is used to identify the model parameters online and in real time. Secondly, the chaotic particle swarm algorithm is used to optimize the control increment during rolling optimization. The simulation is carried out on the MATLAB platform, and the results show that the tracking ability of the average coolant temperature is significantly better than the traditional PI control under variable conditions.