Expert tuning of automatic controllers

Techniques for expert tuning of automatic controllers are presented, with special attention devoted to knowledge acquisition by combining offline and online techniques. These techniques have been implemented with low computational requirements. Two applications are included. In the first the authors consider the supervision of cutting conditions in numerical control machines. The system has been implemented for the supervision of an industrial vertical machining center. The second is the expert tuning of the proportional, integral, and derivative (PID) bands of a PID controller. In this case the authors illustrate the method with the automatic tuning of a PID controller used for temperature control of a heated air stream. Both applications have been performed with very low computational requirements.<
>