基于BP神经网络左逆的无轴承永磁薄片电机无速度传感器控制研究

Speed-Sensorless Operation of Bearingless Permanent Magnet Slice Motor Based on Artificial Neural Network Left-inverse System

在无轴承永磁薄片电机的稳定运行中,实时精确地检测转子速度起着关键性的作用,一般的是使用相关的传感器检测转子速度,但是传统的速度传感器增大了电机的体积、提高了系统的成本,降低了电机在高速运行情况下的可靠性。因此本文提出一种基于神经网络左逆的速度检测方法。基于神经网络原理和左逆原理,设计出速度观测器实现对转速的观测。构建出无轴承永磁薄片电机无速度传感器控制系统,对所提出的速度检测方法进行了仿真和实验研究。结果表明,该方法可以快速准确地识别转速大小,实现无速度传感器下电机的稳定悬浮运行。

The real-time and accurate detection the rotor speed plays a key role in the steady operation ofbearingless permanent magnet slice motors,but the traditional speed sensor has many disadvantages,a speed-sensorless control method based on artificial neural network left-inverse system was proposed.Firstly,the subsystem including the rotor speed and the currents in torque windings was built,and the left-invertibility of the subsystem was discussed.Then,neural network was applied to build the left-inverse and by connecting the left-inverse system with the subsystem in series,the rotor speed was observed effectively.Finally,the simulation and experiment were carried out.The results show that,with the proposed control method,the operation speed of the BPMSM can be estimated precisely and quickly and the stable suspension operation can also be realized.