Sensorless Control of Dual Three-Phase Permanent Magnet Synchronous Motor Based on Phase Current Reconstruction and GA-SMO

This paper presents a sensorless control approach for Dual Three-phase Permanent Magnet Synchronous Motors (DTP-PMSMs), focusing on parameter tuning of a Sliding Mode Observer (SMO) based on phase current reconstruction. The method uses cross-branch sampling and Space Vector Pulse Width Modulation (SVPWM) to reconstruct six-phase currents with only two sensors. A Genetic Algorithm (GA) is employed to optimize key SMO parameters—sliding mode gain, current weighting factors, and phase-locked loop (PLL) compensation terms— to minimize estimation error. Simulation results confirm that the optimized SMO maintains sufficient accuracy in electrical angle and speed estimation, satisfying control requirements. Although based on existing work, the reconstruction scheme reduces sensor count compared to conventional designs, lowering hardware costs. The proposed strategy eliminates manual tuning and enhances observer robustness under varying conditions, offering a practical and cost-effective sensorless control solution for DTP-PMSMs.