Error Compensation in Initial Temperature Estimation of Electric Motors using a Kalman Filter

To generate high accuracy torque by adapting the permanent magnet flux, it is vital to estimate the temperature of permanent magnets. These are nowadays extensively used in electric motors assembled in electric vehicle. In this paper, a short introduction on consequences of neglecting permanent magnet temperature estimation is explained followed by the existing research on permanent magnet temperature estimation methods. This paper focuses on the most ignored topic in majority of research articles based on thermal modelling of electric motor, i.e. initial temperature estimation of each thermal node, its algorithm and its error compensation methods. It is proposed that the error generated from incorrect reference temperature can be compensated with the help of a Kalman filter integrated in thermal network. The proposed method overcomes the challenges involved in an automotive application. Simulation results obtained using experimental data reveal that the proposed theory can be used for practical implementation.