Studies Of The Multivariable Multirate Predictive Control Systems

In this dissertation, the multirate predictive control for multivariable linear system is studied. Firstly, the multirate sampling systems are classified according to the sampling rates of the outputs and performing rates of the inputs. Then, the multirate predictive control algorithm based on the multirate models and the single-rate models and their characteristic are analyzed, respectively. From the rolling optimization time and the performance index, the algorithms are improved and the simulations are done. Secondly, during the study of multirate self-tuning generalized predictive control (GPC), applying the transition between the multirate model and the single-rate model, taking full use of the message of the controlled variable, the improved multirate GPC algorithm is proposed. Thirdly, by transforming the multirate predictive control system into internal model control (IMC) and using the IMC analysis method, its stability and zero steady-state error characteristics are studied. At last, by extending the input and the output vectors, the multirate predictive control systems are changed into extended single-rate predictive control systems. The conditions on stability of the multirate predictive control based on the multirate model and the single-rate model are obtained, respectively. A stability theorem of the multirate GPC system is proposed.The main contribution and innovation in this dissertation are as following:1) The multirate predictive control algorithms being available now are extended. For the multirate sampling system in which the sampling rates of the inputs and the outputs are all different, taking full use of the messages which have been obtained, modifying the rolling optimization time, the improved algorithms are proposed: the improved multirate predictive control algorithm based on the multirate model and the single-rate model. The simulation results show that the dynamic qualities are improved with quicker response and less oscillation of the inputs.2) In order to reduce the overshoot and shorten the steady state time without affecting the response speed, the multirate predictive control algorithm based...