System and method for closed loop autotuning of pid controllers

Abstract

A system and method for automatically tuning a PID controller resident within a PID control loop. The PID control loop includes a PID controller and a process. The process supplies a process variable which is compared to the loop input. The result of the comparison is supplied to the PID controller, and the PID controller drives the process. A relay is applied to the loop input. The relay compares a set point value to the process variable. If the set point value is greater than the process variable, the relay drives the loop input with a first amplitude value. If the set point value is less than the process variable, the relay drives the loop input with a second amplitude value. In response to the set point relay, the process variable develops a sustained oscillation. The period and amplitude of the sustained oscillation are measured. A new set of PID controller parameters are calculated from the period and amplitude of sustained oscillation. In particular, the oscillation period and amplitude are used to calculated a time constant and dead time for a standard process model. The time constant and dead time are used to calculate the new PID controller parameters either (a) directly through the formulae associated with the Ziegler-Nichols reaction curve method, or (b) through the intermediate step of calculating an ultimate period and frequency from the time constant and deadtime.