Research On Distributed Model Predictive Control And Its Improved Algorithm

Due to the advantages of controlling complex large-scale systems,research on distributed model predictive control has been a hot direction.With the control accuracy of industrial process and other requirements are constantly improving,the traditional control methods may not meet the relevant requirements for industrial processes with input-output constraints and high requirements for dynamic performance and robustness,especially in the presence of external disturbances,the output usually cannot achieve good tracking performance.Therefore,for different complex large-scale systems,through combining with the relevant advanced control algorithm,it is particularly important to improve the existing distributed model predictive control.Based on the theory of distributed model predictive control,the algorithm is improved in the following two aspects: On the one hand,for complex large-scale systems,by combining fractional order PID with distributed dynamic matrix control,a distributed fractional order PID type dynamic matrix control method is proposed.The fractional order PID operator is introduced into the objective function of the distributed dynamic matrix control,which inherits the conventional PID characteristics and further expands the parameter tuning range.The gain and phase margin methods are used to adjust the parameters,and the Nash optimal strategy is used to solve the optimal control law of the whole control system.The simulation results verify the superiority of the improved algorithm.On the other hand,an explicit distributed model predictive control method based on extended non-minimal state space is proposed to overcome the shortcomings of distributed model predictive control in dealing with constraints and multivariable process control.In this method,the influence of output error,measurement output and input on the design of model predictive controller is considered.The traditional state space model is transformed into an extended non-minimal state space model with state variables and output tracking errors.The characteristics of the state space model are inherited,and the overall control performance index of the system is also considered.In addition,the explicit predictive control method is used to deal with the system constraints,which makes up for the shortcomings of the traditional distributed model predictive control in high-precision industrial process,and improves the overall control performance of the system.Compared with the traditional distributed model predictive control algorithm,the feasibility of the improved method is verified by simulation,and the application of distributed model predictive control in constrained system is promoted.