Evaluation of order and the spread of time constants for aperiodic processes using step response

A simple and effective method of parameter identification for a linear model of aperiodic processes with known time delay is presented in this paper. The process time lag order and, subsequently, the remaining model parameters are determined from a few samples of the initial phase of the process step response. The method enables one to evaluate the gain and the mean time constant (when using Strejc's model), as well as the spread of time constants and, subsequently, the individual time constants over the range (T-min, T-max) (when using the multi-time-lag model with different time constants). Test results regarding the identification of selected electroheated control plants, as well those regarding the autotuning of PID controllers, are given. Copyright (C) 1997 Elsevier Science Ltd.