Model predictive control strategy based on prediction error feedforward for superheating process

Superheating process in thermal power plant is a two-stage series system. In this paper, model predictive control strategy based on prediction error feedforward is proposed for superheating process. In this series system, the prediction error of the former unit is the estimation of unmeasured disturbance that will further affect the subsequent unit. Consequently, this prediction error can be constructed as an external measured feedforward variable for MPC design of the subsequent unit. The model identification method for this external feedforward variable is also presented. Simulation results indicate that the model prediction accuracy of steam temperature has been improved effectively. This method can enhance the control performance of disturbance rejection for the series system.