A High Precision Initial Position Estimation Method for Low Saliency Ratio Machine Based on Image Tracking

Abstract

In sensorless control of permanent magnet synchronous motor (PMSM) drives, initial rotor position detection is critical to ensure smooth start-up operation. However, extracting initial position information for machines with ultra-low saliency ratios is challenging. To address this issue, this paper proposes an estimation method for rotor position using image tracking based on high-frequency (HF) rotating voltage vector injection. Specifically, this paper extracts positive- and negative-sequence currents in the stationary reference frame and reconstructs the induced current vector. By utilizing the support vector machine (SVM) algorithm, the resulting current vector images, which contain amplitude and phase information of the position, are trained to establish the correlation between the current image and rotor position information. Ultimately, the electric position can be obtained from image recognition. Since this method employs image tracking for position estimation, it accounts for cross saturation and multiple saliencies, leading to the relatively high precision of position estimation. Experimental evaluation shows that the proposed method achieves higher precision compared to conventional observers.