Time scale analysis and synthesis for Model Predictive Control under stochastic environments

This paper presents a method of time-scale analysis and synthesis for Model Predictive Control (MPC) under stochastic environment. A high-order plant is decoupled into slow and fast subsystems using time-scale method with high-order accuracy. Based on the two subsystems, Kalman filters and sub-controllers are designed separately for the subsystems. Then a composite model predictive controller is obtained. The method is illustrated by applying the proposed method to wind energy conversion system. The response of the output from the composite model predictive controller is compared to that of the original MPC showing the simplicity and reduction in computation effort of the proposed method for Model Predictive Control.