Forced Alignment-based Motor Position Sensor Error Identification and Compensation Technique to Reduce Second Order Torque Ripple

This paper proposes a method to extract and compensate for the angular error of a motor position sensor by performing forced alignment to mitigate torque ripple caused by the second-order mechanical angular error of the sensor. To identify the root causes of the second-order mechanical angular error, the operating principles of the motor position sensor are explained, and the factors contributing to angular errors are analyzed. A mathematical analysis was conducted to derive the relationship between angular error and induced torque, demonstrating that angular error components become as torque ripple. Subsequently, the conventional angular error identification technique, which requires additional testing equipment, was introduced, and its potential issues were analyzed mathematically. The proposed method eliminated the need for external test equipment by extracting angular errors through forced alignment and conducting a frequency analysis to determine the magnitude and phase of each order component. Specifically, the inverse-phase signal of the second-order angular error component was compensated in the estimated position. The effectiveness of the proposed approach was experimentally validated using an electric steering motor, demonstrating a reduction in second-order torque ripple compared to conventional methods. The proposed technique is not only applicable to steering systems but can also be extended to braking systems and general motor control applications.