Equivalent circuit model estimation of induction machines under elevated temperature conditions

Abstract

This paper explores performing identifiability analysis on the equivalent circuit model (ECM) parameters of an electric machine (EM) using its impedance response. When modelling the ECM of an EM for room temperature conditions, some of the ECM parameters can be obtained from the manufacturer's data. However, as the temperature of an EM increases this significantly changes the underlying physics (resistivity, capacitance and inductance) of machine parameters, therefore the manufacturers data become inaccurate for equivalent circuit modelling purposes. ECM parameters need to be obtained from the frequency response under the different temperature conditions. To achieve this a nonlinear optimisation scheme with constraints is proposed for the purpose of ECM parameter identification, whereby a temperature-dependent ECM is derived. This work has important applications in EM design and condition monitoring and provides a valuable precursor towards developing age-dependent models.