A study of the Friction Feeding Servo System of Special Punch for Transformer Insulation bar Cushion Block

摘要

In terms of the friction feeding servo system of high-velocity punch for transformer insulation bar cushion blocks, after discussion of a friction-driven model, this paper establishes a three-layer fuzzy neural network with a hidden layer by adopting a structure which combines fuzzy control, PID control with neural network for the control of AC servo motor. In this system, the fuzzy mapping is realized by using multiple-node networks. The control accuracy of a feeding system is improved by the adaptive friction compensation on single-shaft rotation rate of its motor. The effect of friction nonlinear is inhibited and strong robustness is obtained through simulation research.