Utilizing ARMA Models for System Identification in Stirred Tank Heater: Different Approaches

Abstract

Since system identification is a substantial part of many control algorithms, attaining a suitable model is necessary. In this paper, as a novelty, six standard identification methods are used so as to identify the linear ARMA forms of the proposed, widely used model of stirred tank heater. We have chosen these six methods because these methods fit to our desire for the parameter estimation of the ARMA model and they are new and applicable for modeling of all fields of engineering and systems. Beyond that, the mathematics of these methods are so easy to apply and they provide enough data for identifying a wide variation of systems. Furthermore, these algorithms are brought up in a simple and concise manner and easy to implement. In addition, the mathematical form of the suggested stirred tank heater is presented. Furthermore, identification methods are compared in terms of some measures, such as rate of convergence and produced error. Finally, the effectiveness of the six presented identification methods for estimating the stirred tank heater model is illustrated.