Adaptive Full-Order Observer for Sensorless Variable Flux Reluctance Motor Drives With Dual-Parameter Alternating Adaptation

Direct-current excited variable flux reluctance motors (DC-VFRMs) offer advantages of the simple rotor structure and no rare earth. Sensorless control further reduces system costs and improves reliability. However, the stator resistance and mutual inductance of VFRMs are affected by temperature and magnetic saturation, respectively, reducing the accuracy of position estimation for sensorless control. Conventional parameter adaptation methods face challenges with rank deficiency when attempting to simultaneously identify these two parameters. To solve the above problems, an improved full-order observer is proposed to enable identification of stator resistance and mutual inductance, respectively. Then, an alternating adaptation strategy is proposed, which adjusts the field current to generate two operating points: The stator resistance is identified at operating point 1, and the mutual inductance is identified at operating point 2. Theoretical analysis of steady-state error in stator resistance and mutual inductance estimation is derived, along with design guidance of field current injection. Finally, the validity of the proposed method is verified through the 12/10 DC-VFRM experimental platform, and the results demonstrate improved accuracy in position and torque estimation, confirming the effectiveness of the proposed method.