| A water tank pant is built, and the water level and water temperature are controlled variables, the mechanism model of the plant is given, and the parameters are fixed.A controller based on the combination of Industrial PC (IPC) and Data Acquires and Control (DA & C) card is built, with the tank plant a complete test and control platform is built up. On the platform, conventional PID control experiments are performed, which demonstrated that the integral component make the control performance worse and the PD control has a distinct error.On the platform, manual operation experiments have been carried out time and again, and the expert experiences to keep the water level and temperature constant have been obtained, through the DA & C card sampling, the control data are sorted in the IPC and become the learning sample data.On the basis of conventional control and manual control, the experience data starting-up T-S fuzzy controller design and optimization method is presented and proofed. Experiments demonstrate that, by this means the controller performances excel the conventional one but it is sensitive to the working point and less adaptive.An Artificial Neural Net (ANN) control scheme based on the combination of the supervisor ANN control and ANN inverse control is presented, simulation shows its performances excel the conventional and pure ANN controls, and it is of robustness and adaptive.And a fuzzy neural net control scheme is presented, it is T-S fuzzy controller based and an ANN is used to optimize the T-S fuzzy parameter, and better result than T-S fuzzy control has been obtained. |
PDF Full Text Request