The Study Of Compound Control Based On Artificial Neural Network And PID

With simple structure and easy debugging, conventional PID controller is still widely used on the industrial area today, especially suit for the deterministic system with precise mathematical model. In the actual application, many controlled process are high nonlinear, time-variation, probabilistic and lag, the procedure parameters even the model structure vary along with the time and work environment, under the influence of noise and load disturbance. In this case, it can not obtain satisfied control effect using conventional PID controller.Artificial neural network is an important branch of intelligent control, it has the ability of self study, self adapting and self-organizing. So combining the ANN and the PID, make the compound ANN-PID controller identify and set the controlled process parameter automatic, can solve the problem of conventional PID controller with difficult parameter setting, unforcement robustness and can not tune the parameter real-time.On the basis of investigation on the ANN-PID compound control scheme at home and abroad, this paper puts emphasis on studying two compound control schemes with the method of theory and simulation. One is tuning the PID parameter online using ANN and the other is the compound control based on CMAC and PID.The main works and conclusions in this paper are as follows:(1) The control scheme of tuning the PID parameter online using ANN, the ANN in this scheme is used global ANN (e.g. BP ), the drawbacks of the global ANN determines the drawbacks of the ANN-PID compound control scheme, the study speed is low and easy trapped in local minimum. Under the simulation on a specific example , the simulation result indicates that when this compound control scheme tracing the square wave signal, its response is obtuse and can not use for real-time control.(2) A new conceptual mapping algorithm of CMAC is proposed.(3) Modified CMAC-PID controller is proposed. In the conventional CMAC-PID controller , for the conflict between CMAC and PID, the system may become unstable a long time later. In the modified CMAC-PID controller , lead the system error to the study algorithm of CMAC to tune the weights of the association units. In the new scheme, the study of CMAC implements the inverse dynamic model of the whole system including the PID controller. The simulation results indicate that the modified CMAC-PID controller avoids the unstable phenomenon , when tracing the square wave signal the error is smaller than under control of conventional CMAC-PID controller and general PID controller. The modification is effective.(4) The diesel generator set is controlled using modified CMAC-PID controller which is high nonlinear and time-variation. The diesel generator set model in the heat-engine plant is developed under the environment of MATLAB SIMULINK. The simulation results indicate that the dynamic performance of emergency diesel generator set under the control of modified CMAC-PID controller is better than under the control of conventional CMAC-PID controller and general PID controller, with smaller transient adjustment rate and stabilization time.The modified compound scheme avoids the drawbacksof the general PID controller and increases he control effect.