Initial Position Identification of Multiphase PMSM Based on Dual-Winding Symmetrical Pulse Voltage Excitation Using Boundary-Constrained PSO

For the position sensorless system of dual threephase permanent magnet synchronous motor (DTP-PMSM), accurate initial position information is crucial for the smooth startup. When the conventional high-frequency (HF) signal injection method is applied, the detection precision will be issue, an initial position detection method based on dual-winding symmetrical pulse voltage excitation using boundary-constrained particle swarm optimization (PSO) is proposed, which is a special design tailored for DTP-PMSM. Based on the principle of saturation nonlinearity, the initial rotor position can be estimated by tracking the position corresponding to the maximum amplitude difference between two sets of induced symmetrical HF pulse currents. The optimization searching process is carried out using a boundary-constrained PSO algorithm, which has a superior detection accuracy and speed compared to the traditional fixed-step scanning approach. Through dual-axis symmetrical injection, the proposed method mitigates the influence of coupling effect, and is independent of the motor saliency and the spatial winding distribution compared with the conventional HF injection methods. This method has a good universality and adaptability. The effectiveness is validated by experiments on the developed DTP-PMSM drive platform.