On the practical implementation of a noncausal feedforward technique for PID control

In order to improve the set-point following performance of a PID controller, a stable input-output inversion procedure can be employed. It allows to synthesise the analytical expression of the command input to be applied to the closed-loop system in order to obtain a desired transition of the process variable from a set-point value to another one. However, since the system to be inverted is nonminimum-phase, the synthesised command function is noncausal. In this paper we show how an approximated solution can be effectively causally implemented in an industrial context (namely, with a Distributed Control System) with no significant loss of performance.