| PID control method is one of the traditional control methods and it can get good effects under many application situations. But with the increase of the complexity of manufacture technics and demands of industrial process performance, the conventional PID control can't meet the requirement of the closed loop optimized controller. Intelligent control independent the model of a plant and based on knowledge offers a new idea for improving the process technologies, has some advantages in many applications. Neural PID controller plays an important role. Neural network is showing its superiority in many applications as one kind of modern information processing technology. And its application result in automation control area—neural network controller is becoming an attracted developing direction. Neural network PID control technology plays a very important role among all kinds of the control methods.This thesis expounds the basic theories of neural network control, analyze the applicability of the neural network controller, discusses the rationality and feasibility to combine the neural network controller and PID controller, concludes the neural network controller's basic form and theoretical principle PID controller based on single neuron, PID controller based on RBF neural network, PID controller based on BP neural network. Simulating the specified controlled object by using these three PID controller, and comparing their simulating results, to analyze their control performance.Based on finishing some kinds of neural network PID control arithmetic simulation study, analyze the process flow of three tank level control system and the external characteristic quieting characteristic of its run process, use conventional digital PID and BP neural network has a better control result, stronger information processing capability,adaptivity and robustness by contrast, and we can get ideal response curve. BP neural network and conventional PID, have a good useful value. |
PDF Full Text Request