Estimation of Process Step Response Weights based on Closed-Loop Step Response Data

Abstract

Dynamic matrix controllers (DMC) and model predictive controllers (MPC) have found wide application in the process industries. In order to design such a controller, a step response model of the process is often used. This paper will provide an algorithm for obtaining the step response model of a process based on closed-loop operating data generated by a step change in the setpoint. This algorithm is intended for application to closed-loop systems where a DMC or MPC is being designed to replace an arbitrary existing controller.