Parameter Identification for Non Integer Order, Discrete, State Space Model of Heat Transfer Process Using CFE Approximation

Abstract

In the paper the parameters identification problem for a new, non integer order, discrete, state space model of heat transfer process is presented. The proposed model employes Continuous Fraction Expansion (CFE) approximation what assures good accuracy with relatively low order and short memory length. The parameters of the model are estimated via numerical minimization of the Mean Square Error (MSE) cost function. The proposed model is compared to discrete model using Power Series Expansion (PSE) approximation. Results of the comparison show, that the proposed model assures the same accuracy using significantly smaller memory length than model employing PSE approximation.