Research Of Predictive Control In The Large Time Delay System

Large time-delay system is a typical system for those complex processes such as petrochemical, brewery, papermaking or metallurgical, and this kind of system is still as a hard nut to crack in automation field. The characteristics of such kind of process can be summarized: The system's key parameters can not meet the requirements as the large time-delay problem and even worse the system will become unstable. As the conventional control methods such as PID algorithm or Smith predictor can't act a good performance at the time-delay systems, therefore, there are a large number of scholars focusing on these problems and many improved algorithms have been proposed. In addition to traditional strategies, several advanced control algorithms such as fuzzy control or robust control are also used extensively in this field. Although the control for large time-delay system has been significantly improved, there is still a long way to go.A comprehensive analysis to the existing control methods for time-delay systems has been shown in this paper at first. Based on this analysis we know that the predictive control algorithm has some shortcomings at interference suppression and on the control of the time variable delay system, although it has some advantages for time-delay system. Based on these shortages, the improved Dynamic Matrix Control (DMC) algorithm with the combination of Genetic Algorithm (GA) has been given in this paper, which can not only suppress interference immediately but also get a good performance on the time variable delay system.Beer fermentation is a typical large time-delay system, and it still can't be controlled well due to its complex biochemical process. As we all know that the test of beer and the production efficiency of beer fermentation are closely related to fermentation temperature, so the experimental model of fermentation temperature is given as a target plant to implement the control strategy of this paper. The result proved that the improved strategy can get a good performance on the control of temperature plant.