Adaptive control for metal cutting

Adaptive control software developed by Jones & Shipman for use in grinding has fared extremely well in practical industry tests. Using a Supramat machine to grind a long, slender drive spindle, Slack & Parr was able to reduce a plunge-grind cycle of 90s to 22s.

The advantages of simple adaptive controls for metal cutting operations have been recognised for some time. Now, after two years of research and development carried out by Jones & Shipman and Liverpool John Moores University, high-production adaptive control available on the company’s new range of grinders.

A programme which entailed writing intelligent algorithms for Jones & Shipman PC software in conjunction with GE Fanuc open CNC systems has resulted in the identification of two adaptive control strategies: `auto dwell selection’ and `feed rate update’.

It was found that during the spark out period of a cylindrical grinding cycle the forces encountered at the grinding interface decrease in a predictable fashion until the part reaches the designated diameter. The `auto dwell’ strategy makes use of the knowledge that the power needed by the grinding wheel motor is proportional to the grinding forces encountered.

By measuring the wheel power at the start of the dwell, and applying a mathematically calculated curve fit, engineers can predict the length of time for the exponential decay curve to reach no-load power.

It thus, becomes possible to make an in-cycle determination of the dwell period required to bring the part to size, ensuring roundness of the part and improving the surface finish.