Fast Generalized Predictive Control Algorithm Study And Application

Generalized Predictive Control (GPC) is, which appeared in 80's,a new typeof computer control method and one of the most representation algorithm. It hasreceived increasing attention in the field of control and industry. But for the hugecomputation, especially for multiple systems, it has much huger computation. Sothe virtue of predictive control will not be embodied in the real time systems ofindustry. In order to cope with these basic problems, several new FastGeneralized Predictive Control Algorithms are proposed in this paper and somesimulation studies were done for the typical industrial models.Firstly, for the limits of the standard generalized predictive control, we putforward an improved generalized predictive control algorithm. It can restrainover-shoot very well, at the same time, also increase the responding speed of thesystems, but the effect is not distinct for the huge computation. As a new controlalgorithm, Predictive Functional Control has some characteristic, the simplecontrol equations,less calculation of real time control. So we introduce the ideaof the basic function of PFC in the Generalized Predictive Control to produce afast control study based on predictive function. This strategy overcame theshortcoming of unfavorable real-time for high precision tracking controlperformance.Secondly, some industry processes not only demand high precision trackingbut also smoothing input, avoiding sharp changes of input signals to damageequipments. For these, a constraining input fast GPC algorithm is presented. Itsatisfied high precision tracking need, also restrained sharp changes of inputsignals.The paper makes the full research from control theory, design method andthe application of the theory to several typical industrial models. Several noveldesign methods about GPC have been mentioned, proved in theory and computersimulation experiments have been successfully implemented. The simulationresults show that these algorithms can obtain better control effect thanconventional GPC algorithm. The contribution of this dissertation lies inaccelerating the application of GPC control algorithm in the industrial controlfield.