Reduced-order electrical model extensions for induction machine temperature monitoring with extended Kalman filter

The paper presents an electrical reduced-order model for real-time purposes. This model allows taking easily into account iron losses. Several model extensions for a Kalman filter are defined, in the objective of the thermal monitoring of the induction machine, in real-time. An accurate experimental set-up is presented and machine model identification is described. Experimental results are presented and the Kalman filter convergence is discussed, considering the different model extensions, in the aim of temperature estimation