基于速度环扩展卡尔曼滤波的无位置传感器电机控制

Sensorless Motor Control Based on Speed Loop Extended Kalman Filter

针对大负载摇臂机构,提出了一种基于速度环扩展卡尔曼滤波的无位置传感器电机控制方案。首先根据电机的电压方程,在全速度范围内使用凸优化方法来估计转子位置,该方法在低速情况下需要注入高频信号。针对低速重载情况下铁心饱和导致位置可观测性下降的问题,结合电机机械运动模型,使用扩展卡尔曼滤波算法,对基于凸优化的估计进行修正。此外由于EKF对参数较为敏感,使用离线最小二乘法辨识系统的惯量、阻尼和转矩常数,并将结果带入EKF中。经过EKF滤波之后,极大地提高了速度估计质量,尤其是在饱和发生时。最后对所提出的方法,进行了仿真研究,并在由表贴式永磁同步电机驱动的电机控制实验平台上进行了实验验证。

Aiming at the heavy-load swing servo mechanism, a position sensorless motor control scheme based on the speed loop extended Kalman filter was proposed. The rotor position was first estimated from the motor’s voltage equation using a convex optimization method over the full speed range, which required injection of high-frequency signals at low speeds. Aiming at the problem that the core saturation led to the decrease of position observability under the condition of low speed and heavy load, combined with the mechanical motion model of the motor, the extended Kalman filter algorithm was used to correct the estimation based on convex optimization. In addition, since the EKF is sensitive to parameters, the inertia, damping and torque constants of the system were identified using the offline least squares method, and the results were brought into the EKF. After EKF filtering, the speed estimation quality was greatly improved, especially when saturation occurs. Finally, the proposed method was simulated and verified by experiments on a motor control experimental platform driven by a surface-mounted permanent magnet synchronous motor.