Research On Distributed Predictive Functional Control And Its Improved Algorithm

Predictive functional control(PFC)can make the input of the control system structured,resulting in small amount of calculation,good rapidity,strong anti-interference ability etc,and it is widely used in industrial control.Furthermore,with the development of modern technologies,PFC method has been greatly developed.It has been extended to an organic combination with distributed control(DC),namely,distributed predictive functional control(DPFC).The DPFC algorithm has a better control effect for some large-scale coupling systems.However,due to the influence of people and the actual environment,there exist interference and constraints in the actual production processes,which cannot be ignored and will greatly affect the performance of DPFC algorithm.How to eliminate the influence of disturbance on the performance of DPFC and effectively deal with the restriction of constraint are the main focus.The research of this thesis can be divided into three parts.1.For some large-scale systems with strong coupling,the DPFC algorithm based on Nash strategy is proposed.The large-scale system with strong coupling is distributed to each subsystem for online optimization by Nash strategy,where the distributed predictive functional controller is obtained.2.Considering that the performance of DPFC algorithm needs to be further improved in the case of introducing irresistible factors such as interference,this section designs a basic function PID distributed predictive functional control algorithm(BFPID-DPFC).The algorithm uses PID factors and basic function weighting coefficients to reconstruct the performance index of the new distributed predictive functional control,the performance is improved by adjusting PID and the weighting coefficients of basic functions.Finally,through simulation comparisons between DPFC algorithm and BFPID-DPFC algorithm,it is proved that the overall performance of BFPID-DPFC algorithm is better than DPFC algorithm.3.Considering the constraints,this section designs an explicit distributed predictive functional control algorithm(EDPFC).In order to eliminate the constraints of the system,the value domain and kernel space of the constraint matrix are decomposed,then according to the optimization strategy of Nash game theory,thecoupling effect of the system is eliminated,so as to obtain the explicit distributed predictive functional controller.Finally,the feasibility of the EDPFC is verified by comparing the EDPFC algorithm with traditional distributed predictive functional control(DPFC)algorithm.