Online Parameter Identification for Model-Based Sensorless Control of Interior Permanent Magnet Synchronous Machine

This paper proposes an online identification method that can accurately estimate the stator resistance and dq-axis stator inductances for the effective model-based sensorless control of interior permanent magnet synchronous motors (IPMSMs). The proposed affine projection algorithms are uniquely designed in the estimated rotating γ-δ frame to precisely identify the parameters mentioned above. The two time-scale approaches are employed in the affine projection algorithms to estimate the three electrical parameters. Despite the electrical parameter variations due to the temperature change and magnetic saturation during operation, the rich enough data are provided to the affine projection algorithms in the discrete-time domain to accurately retrieve the updated parameters. These correctly estimated parameters are adapted to the extended back electromotive force observer for the sensorless control of IPMSM drives. Hence, the adaptation of online updated parameters makes the observer stable and robust to parameter variations as compared to the conventional observer without updated parameters. The MATLAB/Simulink-based simulation results and experimental results via a prototype IPMSM test-bed having TMS320F28335 DSP are given to verify the accurate convergence of the estimated parameters, which results into a stable sensorless control system under various operating conditions.