Neural network based pid control on nc milling process with high-speed

The self-adaptive control strategy, which is applied to NC machining processes, is one of the most important approaches to improve production efficiency. It can be realized by two steps. Firstly, the dynamic model of servo subsystem and milling force is deduced in frequency domain, and is discretized in terms of Z — transform so as to approach the real control situations. Secondly, a PID controller is adopted in tuning the feed velocity to trace the reference value of milling force. The Neural Network (NN) algorithm is used here to change the PID parameters on-line to deal with the real machining situations, such as abrupt or gradual cutting depth changes. The advantages of proposed method are illustrated by comparison to the PID control with fixed gains