Rapid tuning of controllers by IFT for profile cutting machines

The tuning of controllers for metal profile cutting machines can be time consuming and costly requiring a skilled operator to intuitively select and adjust control parameters. Plants contain variations due to construction, load, and cutting speed. In this paper a new Iterative Feedback Tuning technique is developed for implementation on a metal cutting machine with a plasma torch. It is compared with the original iterative tuning technique and shows an improvement of 40% less error in cutting head position over 10 iterations. The improvements are made by providing a step size in each controller parameter update direction and by monitoring for an increase in design criterion suggesting parameters are moving away from the optimised controller. A set of rules are defined that allow the step size to be reduced and repeat the previous update if an increase in the design criterion is detected. In this way beginning with a coarse search that is slowly refined over the iterations allows parameters to rapidly reach a local minimum in the design criterion.