A Study Of Intelligent Self-Tuning PID Controller

A new type of intelligent self-tuning PID controller based on BP network is designed in this paper because the traditional PID controller is not suit for the non-linear system.Two kinds of intelligent self-tuning PID controller based on BP network are mainly studied in this paper: One is the relay-type self-tuning PID controller, the other is the intelligent self-tuning PID controller based on BP network. The effect of the intelligent self-tuning PID controller based on BP network is proved through the tuning effect of relay-type self-tuning PID controller.The relay-type self-tuning PID controller is formed and the intelligent self-tuning PID controller based on BP network, is designed. The experiment hardware is formed and the programs of the intelligent self-tuning PID controller based on BP network and the relay-type self-tuning PID controller are designed. The experiment based on improved PID controller is finished through using the relay-type self-tuning PID controller. The necessary sample is obtained through the experiment.The income sample are divided into the learning sample and the text sample. The part of neural network of the intelligent self-tuning PID controller based on BP network learns the learning sample. A lot of simulation identifies the identification performance, the fault freedom, the generalization of the neural network. The input part of the text sample is input the neural network of the intelligent self-tuning PID controller based on BP network. Then the output is obtained. The output is simulated and is compare with the output of the text sample. The network is identified to be correct and practicable.The neural networks of different degree of difficulty and different parameter are simulated. The simulation identifies that the parameter of the intelligent self-tuning PID controller based on BP network is correct and offer reference to other parameter of network. The result of using the improved BP algorithm is better than the result of using ordinary BP algorithm. So the superiority of the improved BP algorithm is proved.There is very strong actual meaning in this paper and it can offer reference to others because there is a large mumber of esperimental datas and simulations.