Time Delay Compensation And Control Of Network Control System Based On Generalized Predictive Control

In recent years,thanks to the rapid development of communication technology,the control system has undergone major changes,and its internal not only achieves the most decentralized closed-loop control,but also the data transmission efficiency has been improved.The research on network control systems has also been rapidly carry out.Because the original control system is placed in the network environment,the network delay and packet loss problem will be generated,which greatly affects the performance of the system.The traditional control method is not applicable to the current network control system.In order to achieve better control of the network control system,this paper uses the improved BP neural network to predict it and combines the improved generalized predictive control to compensate the delay in the network.For the nonlinear time-delay control system,the neural network is used to fit the system,a multi-step predictive model is established.Aiming at the existence of constraints in the control system,a PI-type implicit generalized predictive control algorithm is proposed.The main contents of this paper are as follows:(1)In view of the problem that the controllability of the system is reduced due to the random communication delay,this paper uses the improved neural network to predict the random communication delay,solves the problem that the parameters are difficult to select,the generalization is difficult,and the prediction is improved.On this basis,combined with the improved generalized prediction algorithm,the random delay is actively compensated,which solves the limitation that the original algorithm can only compensate for the fixed delay.Finally,the correctness of the algorithm is verified by MATLAB simulation experiments.(2)Aiming at a class of nonlinear time-delay systems in industrial production process,this paper proposes a generalized predictive control algorithm based on extreme learning machine.Firstly,the improved differential algorithm is used to optimize the neural network of extreme learning machine.Then the optimized extreme learning machine is used to establish the prediction model of the nonlinear time-delay controlled object.Finally,the prediction model is brought into the implicit generalized predictive controller.The algorithm better solves the problem that the nonlinear system is difficult to model and improves the accuracy of the prediction.MATLAB simulation confirmed the correctness of the algorithm.(3)Aiming at the general constraints of the actual control system,this paper designs a PI-type implicit generalized predictive control algorithm with constraints and analyzes its stability.The algorithm adds the PI feedback structure to the original implicit generalized prediction and improves the selection rule of the control quantity in generalized prediction.It not only can solve the constraint problem well,but also combines the advantages of PI type implicit generalized prediction and has good control performance.The MATLAB simulation results show the correctness of the algorithm.