Parameter Identification of PMSMs Considering VSI Nonlinearity with Coupled Adaline NNs

This paper proposes a novel parameter identification method of permanent magnet synchronous motors (PMSMs). The parameters of the PMSM are identified by using Adaline neural networks designed to operate with two steady states obtained without signal injection or ad hoc tests. Furthermore, the voltage-source inverter nonlinearity is compensated to increase the accuracy of the parameter identification. The proposed identification procedure is suitable for offline as well as online implementation, overcoming the drawbacks of conventional parameter identification methods. The method is validated with both simulation and experimental data.