The Simulation Study Of Improved GPC Algorithms

Generalized Predictive Control (GPC) is a new computer control algorithm proposed by Clarke and others in the 1980s.It preserves STC's advantages of Identification On-line and Minimum Variance Control, and includes the idea of receding option of DMC and MAC. At the same time it is based on parameter model and introduces vary forecast and control level. So there is good control performance and robustness. It has become one of the most widely used and actively researched predictive control algorithms and been paid more and more attentions and researches and widely applied to industrial process.Here introduce the basic algorithm of GPC, next give the introduction of recursive algorithm of Diophantine equations. And then give an example of SISO System for simulation and study control effect of the parameter's selecting of GPC. Base on the study of a large number of literatures, we could find some disadvantages in GPC, such as huge computation, the tuning parameters are not easy to determine and so on. Therefore, GPC is studied here in the above aspects, and some research has done as follows:First, the implicit GPC algorithm is given. The algorithm does not need recursive Diophantine equation and matrix inversion, so it can overcome the long time in on-line computation of basic GPC. Consider that there is always time-delay in industrial production processes and basic GPC algorithm does not take into account the. A generalized predictive control recursive algorithm with time-delay was given by Simulation studies with MATLAB, and was proved the effective and feasibility. On the basis we seek the qualitative relationship between time-delay and the predictive step and provide some principles of selecting predictive step in the control system with time-delay. Last the step generalized predictive control (SGPC) is proposed in this chapter, which is combined with step-control technology and GPC, studying on application of the algorithm in extending the life of actuating element, and operating steadily. It is a great significance for GPC's applications in real industrial processes.