Multi-parameter Identification of PMSM Servo Systems Under Nonstationary Conditions Based on Dynamic Average Model

Most existing parameter identification methods are developed under stable conditions, which are inevitably vulnerable under nonstationary conditions, such as variable load or variable speed conditions. In this paper, a novel parameter identification method of permanent magnet synchronous motor (PMSM) servo systems is proposed, by which the PMSM parameters and voltage source inverter (VSI) nonlinearity can be identified under nonstationary conditions. Firstly, a novel dynamic average model is established, which only focuses on the dynamic average values of nonstationary vibration signals, instead of their instantaneous sampling values. Due to the neglect of disturbances such as high- frequency components and measurement noise, the parameter identification method based on this model has strong robustness and stability under nonstationary conditions. Secondly, according to the identifiability analysis and proposed information extracted method, the dynamic average values which make the model full rank can be extracted from the nonstationary vibration signals. Finally, the parameters in the model are solved by multi-innovation least squares (MILS) algorithm. The proposed parameter identification method is tested on a laboratory PMSM servo system. The experimental results demonstrate that this method can accurately identify parameters under various nonstationary conditions.