Accuracy of Online Estimation Methods of Stator Resistance and Rotor Flux Linkage in PMSMs

• Published in: IEEE journal of emerging and selected topics in power electronics 2024-10, Vol.12 (5), p.4941-4955
• Main Authors: Brescia, Elia, Pio Savastio, Luigi, Di Nardo, Mauro, Leonardo Cascella, Giuseppe, Cupertino, Francesco
• Format: Article
• Language: English
• Subjects: Algorithms; Condition monitoring; Couplings; Error analysis; Estimation error; Fault diagnosis; Monte Carlo simulation; Motors; online parameter estimation; Parameter estimation; Parameter robustness; permanent magnet; Permanent magnets; Probabilistic analysis; rank deficiency; Resistance; Robust control; Rotors; Stators; Steady state models; synchronous machines; Synchronous motors
• ISSN: 2168-6777; 2168-6785
• DOI: 10.1109/JESTPE.2024.3453949

Online parameter estimation algorithms of electrical motors are widely employed to achieve robust and high-performance control as well as condition monitoring and fault diagnosis. This article proposes an error analysis to compare the performance of four different online estimation methods of the stator resistance and rotor flux linkage in nonsalient permanent magnet synchronous motors (PMSMs). Starting from the dq-axis steady-state model of the PMSM, including the voltage source inverter (VSI) nonlinearity and the presence of measurement errors, a formulation of the estimation errors is outlined. A probabilistic analysis is then proposed to determine the expected values and the percentile ranges of the estimation errors on the basis of the previously introduced formulations. The theoretical predictions have been verified with an extensive experimental test campaign carried out on an instrumented test rig in multiple operating conditions (OCs) of two different motors. The experimentally validated results show that the considered parameter estimation methods exhibit different performances depending on the motor speed and torque levels. No parameter estimation technique performs notably better than any other under all OCs. As a consequence, the proposed analysis constitutes a useful tool to select the most appropriate estimation method.