The Research And Application Of Modified Algorithms About Fuzzy Predictive Functional Control

As the continuous development and wide application in industry of control technology, the controlled plants of actual industrial processes become more and more complex, such as time-delay, nonlinear, uncertainty and high intertia. For this complicated system, traditional control theory is too difficult to give a very good control effect as it is difficult to establish the precise mathematical model. Although predictive functional control does not demand high-accurate prediction model, the control effect is still limited. Studying predictive functional control in fuzzy environment can extend the applied scope of predictive functional control.Based on the previous research work,this dissertation is further researching on fuzzy predictive functional control. The main work of the dissertation is illustrated as follows:1. Research and basis theory of fuzzy predictive functional control algorithms Analyze the development status of fuzzy predictive control and predictive functional control, and introduce simply predictive functional control algorithm, fuzzy inference and T-S fuzzy model, it can provide theoretical references for the further research on the dissertation.2. Self-tuning scale factor fuzzy predictive functional controla. By considering that the control parameters of conventional fuzzy controller is fixed and unable to adapt to continuous process change, a self-tuning scale factor fuzzy predictive functional control is given out, making self-tuning scale factor K_u changing through adjusting factorδ, aiming to adjust the output of fuzzy controller, then to control the object with predictive functional control increment and controlled amount at the moment before.b. The proposed algorithm is applied in main steam temperature control, PID control law is adopted in inner-loop and self-tuning fuzzy predictive functional control scheme is used in outer-loop.3. Generalized T-S fuzzy model predictive functional control based on genetic algorithm a. By considering that fuzzy membership functions and fuzzy rules of general T-S fuzzy model is in large measured by people, a new algorithm is given, firstly to choose clustering center by subtraction clustering, secondly, to identify the parameters of premise and conclusion of generalized T-S model fuzzy by genetic algorithm, then to control the system by predictive functional control.b. The proposed method is applied in PH process. Firstly, PH process's generalized T-S model is obtained by identifying with genetic algorithm, then to control the PH process with predictive functional control.