A direct estimation method of second order plus time delay model parameters for PID controller autotuning

Abstract

In this paper, we propose an identification method to estimate the second order plus time delay (SOPTD) model parameters. The proposed method obtains the parameters of the SOPTD model directly using a time weighted integral transform and a least square method, so model reduction steps can be avoided to tune the PID controller. It can also be applied regardless of the initial states while almost all previous identification methods require the assumption of zero initial states. Through extensive simulations, it is proved clearly that the proposed method gives acceptable estimation performances whether initial states are zero or not. Moreover, it shows an acceptable robustness to the disturbances and measurement noise.