Multivariable generalized predictive control based on receding feedback correction in binary distillation process

摘要

An improved generalized predictive control (GPC) algorithm has been proposed in this paper. In order to cut off computation burden and get good performance of disturbance rejection, receding feedback correction mechanism is used to reject the disturbance instead of the traditional online identification mechanism. Furthermore, two factors are presented to tune for good dynamic performance. Simulation has been carried out with a binary distillation process model, named wood& berry model, which is a typical multivariable process. The simulation result proves the efficiency of proposed algorithm.