Root mean square based identification of FOPTD models from step response data

In this paper an identification technique of first order plus time delay (FOPTD) process model from open loop step response data is proposed. The output response data from practical systems always contains noise. To remove the effect of noise in estimation of model parameters, root mean square (RMS) of few data points are considered. First delay of the model is obtained and then root mean square (RMS) of few data points is used to estimate system gain and the time constant of the model. The accuracy of identified model is evaluated in terms of % - fit. The proposed technique is compared with few well established techniques and compared in terms of % - fit value and found to give better model than the existing one. The proposed method has the advantage of being simple, intuitive and easy to understand and implement.