Towards intelligent PID control

Abstract

Autotuners for PID controllers have been commercially available since 1981. These controllers automate some tasks normally performed by an instrument engineer. The autotuners include methods for extracting process dynamics from experiments and control design methods. They may be able to decide when to use PI or PID control. To make systems with a higher degree of automation it is desirable to also automate tasks normally performed by process engineers. To do so, it is necessary to provide the controllers with reasoning capabilities. This seems technically feasible with the increased computing power that is now available in single-loop controllers. This paper describes some features that may be included in the next generation of PID controllers.