An Inertia Identification Method Based on Adaptive Linear Disturbance Torque Observer for PMSM Drives

The inertia identification effectiveness depends on the accuracy of the disturbance torque estimation. In this paper, an inertia identification method based on an adaptive linear disturbance torque observer is proposed to avoid the potential oscillation of the traditional reduced-order observer. Based on the adaptive linear neuron network, the disturbance torque estimation is adjusted automatically according to the speed error. The proposed method could accurately identify the disturbance torque without any assumptions and the inertia identification precision is improved. The proposed method is verified on a 2.2-kW permanent magnet synchronous motor drive platform.