Wang Lu, Cheng Yong, Yin Wei
College of Energy and Power Engineering, Shandong University, Jinan 250061, China.
Sensors (Basel). 2023 Jul 22;23(14):6604. doi: 10.3390/s23146604.
Accurate knowledge of the rotor position is essential for the control of brushless DC motors (BLDCM). Any deviation in this identification can cause fluctuations in motor current and torque, increase noise, and lead to reduced motor efficiency. This paper focused on a BLDCM equipped with a three-phase binary Hall sensor. Based on the principle of minimum deviation, this paper estimated the relative installation offset between the Hall sensors. It also provided a clear method for ideal phase commutation position recognition and eliminated the Hall sensor installation position deviation. The proposed pre-calibration method identified and eliminated the offset of the permanent magnet poles, the delay time caused by the Hall signal conditioning circuit, and the offset of the sensor signal identification due to armature response under different loads. Based on the pre-calibration results, a correction strategy for correcting the rotor position information of BLDCMs was proposed. This paper presented a self-adaptive position information prediction algorithm based on the Sage-Husa method. This filters out rotor position information deviations that are not eliminated in pre-calibration. Experimental results on a hydrogen circulation pump motor showed that, after the pre-calibration method was adopted, the Mean Square Error (MSE) of motor speed fluctuations decreased by 92.0%, motor vibration was significantly reduced, average phase current decreased by 62.8%, and the efficiency of the hydrogen circulation pump system was significantly improved. Compared to the traditional KF prediction algorithm, the Sage-Husa adaptive position information prediction algorithm reduced the speed fluctuation during the uniform speed operation stage and speed adjustment stage, the speed curve overshoot, and the commutation time deviation throughout the process by 44.8%, 56.0%, 54.9%, and 14.7%, respectively. This indicates a higher disturbance rejection ability and a more accurate and stable prediction of the commutation moment.
准确掌握转子位置对于无刷直流电机（BLDCM）的控制至关重要。这种识别中的任何偏差都可能导致电机电流和转矩波动、增加噪声并降低电机效率。本文聚焦于一款配备三相二元霍尔传感器的无刷直流电机。基于最小偏差原理，本文估算了霍尔传感器之间的相对安装偏移。它还提供了一种清晰的方法来识别理想的换相位置并消除霍尔传感器安装位置偏差。所提出的预校准方法识别并消除了永磁极的偏移、霍尔信号调理电路引起的延迟时间以及不同负载下电枢响应导致的传感器信号识别偏移。基于预校准结果，提出了一种用于校正无刷直流电机转子位置信息的校正策略。本文提出了一种基于Sage-Husa方法的自适应位置信息预测算法。这滤除了预校准中未消除的转子位置信息偏差。在氢气循环泵电机上的实验结果表明，采用预校准方法后，电机速度波动的均方误差（MSE）降低了92.0%，电机振动显著降低，平均相电流降低了62.8%，氢气循环泵系统的效率显著提高。与传统的卡尔曼滤波（KF）预测算法相比，Sage-Husa自适应位置信息预测算法在匀速运行阶段和速度调整阶段分别将速度波动、速度曲线超调量和整个过程中的换相时间偏差降低了44.8%、56.0%、54.9%和14.7%。这表明其具有更高的抗干扰能力以及对换相时刻更准确、稳定的预测。
