Application of neural network sliding controller in ac servo systems

Some ac position servo system possesses abominable load properties, i.e. largely changing moment of inertia and moment of imbalance. It´s difficult for traditional PID controller to achieve the satisfying effects. In order to realize high speed and high precisive position control of the ac servo system, aiming at uncertainties in actual system, a neural network sliding controller is put forward, which takes the advantages of excellences such as rapid response and consistence for parameters variation and foreign disturbance, and takes advantage of online learning function of neural network to eliminate chattering in sliding control to enhance steady performance of system. Simulation results indicate this controller can restrain disturbance effectively, decrease sensitivity of system to parameters variation, improve dynamic performance and steady precision markedly, fulfill the goal of rapid tracking. It has practical application value.