基于PSO-WNN的无刷直流电机转子位置检测方法

A new position detection based on PSO-wavelet neural network method for brushless DC motors

通过分析无刷直流电机间接位置检测原理,提出了一种新的方法来检测转子位置.该方法首先推导出转子位置可以通过以相磁通和相电流来决定,结合小波函数多尺度多分辨率的优点以及神经网络的非线性求解特点,通过构建小波神经网络模型,并采用粒子群算法来训练网络参数而得出转角位置.仿真结果表明该模型能有效地控制电机换相.

The paper analyzed the principle of position sensorless control for brushless DC Motors （BLDCM）, and a new position detection method for BLDCM was proposed. This method build a wavelet neural network which use phase flux linkages and phase currents as the input of network, then to estimate the rotor position. A wavelet neural network model was built whose parameters were trained based on particle swarm optimizer algorithm The Simulation results show that the given modeling method can control the commutation.