This paper mainly researched the problems emerging from the neural network application in the parameter tuning and optimizing of PJD controller. A lot of simulations and experiments have proved the following conclusion: (l)The NN system identification must be further completed in order to provide a more accurate object model. (2) Because of the character of PID controller and the chosen of objective function, the direct optimizing method based on objective function such as pure-form method is still most commonly used. (3) The NN Intelligence PID controller has a good control effect in the on-line parameter tuning and optimizing. (4) Besides of B-P network, the application of Hopfield network in the parameter tuning and optimizing of PID controller is a promising method. Guo Peng (Control Theory and Control Engineering) Directed by prof. Han Pu...
|