Identification of diffusion processes using fractional non commensurate order models

Abstract

This article deals with heat transfer modeling and identification, based on fractional non commensurate order state space models. An output error method is developed to estimate the model parameters including the fractional order and an iterative non-linear programming method is used, based on a novel fractional multivariable sensitivity functions model. Numerical simulation data of the heat transfer problem are performed using finite differences, and the suggested method is applied to identify the thermal process. Two fractional non commensurate models are tested in order to analyze their fitting ability to approximate the interface diffusion in the frequency domain.