Simon Decker, Johannes Stoss, Andreas Liske, Matthias Brodatzki, Johannes Kolb, Michael Braun

2019 21th European Conference on Power Electronics and Applications (EPE'19 ECCE Europe)

This paper shows a fast online parameter identification method for permanent magnet synchronous machines (PSMs), which uses the PWM (pulse width modulation) excitations and no additional test signals. Based on the inverter induced current slopes, the corresponding applied voltages, the rotor angle and a precise machine model, the model's parameters are calculated. Thereby the machine model consists of the dq-system equations, linearized within one PWM period. For stable and precise identification of the PSM machine parameters, the machine model has to be modified with a regularization approach. The identified parameters enable self-commissioning, tuning of the control parameters, condition monitoring or inner fault detection. In this paper the theoretical approach of the suggested method and simulation results of an equivalent test-bench system are presented.