Anisotropy-based position estimation for highly-saturated permanent magnet synchronous machines in slowly-sampled control systems

This paper presents an anisotropy-based position estimation approach taking advantage of saturation effects in permanent magnet synchronous machines (PMSM). Due to magnetic anisotropies of the electrical machine, current responses to high-frequency voltage excitations contain rotor position information. Therefore, the rotor position can be estimated by means of these current responses. The relation between the high-frequency current changes, the applied phase voltages and the rotor position is given by the inverse inductance matrix of the machine. In this paper, an analytical model of the inverse inductance matrix considering secondary anisotropies and saturation effects is proposed. It is shown that the amount of rotor position information contained in these current changes depends on the direction of the voltage excitation and the operating point. By means of this knowledge, a position estimation approach for slowly-sampled control systems is developed. Experimental results show the effectiveness of the proposed approach.