Automatic Tuning of Pid Controllers Through Response Optimization Over a Finite Time Horizon

Abstract

The input-output functional relationship of a plant is identified and expressed in explicit form over a finite time horizon, Several schemes (orthogonal expansions, ARMA and MA models) are shown to result in a similar problem formulation. The identified model is operated on the fixed structure of a discrete time PID controller, and the parameters of the latter calculated as to minimize a performance index that penalizes both the quadratic departure of the error from the desired value and large variations in the error signal therefore avoiding abrupt controller actions.Results from simulation are given in terms of tuning the PID controller of a triple effect evaporator that simulates a. real equipment used at the industry to obtain juice concentrate.

Best achievable low order decentralized performance

“…A very popular class of low order controllers is the class of PI and PID controllers. The design University of Texas, Austin, TX 78712 of such controllers has been studied extensively and several tuning methodologies have been developed [4], [2].…”