摘要
在分析反电动势过零检测原理的基础上,推导出线反电动势过零点与电机换相点及线电压与线反电动势间的关系,从而在线电压与转子位置间建立了联系。但由于无刷直流电机(BLDCM)是一个复杂的非线性系统,并且电机的一些参数在运行过程中也是变化的,因此直接通过线电压获得准确的转子位置比较困难。为此提出利用自适应小波神经网络将3个线电压作为输入信号来辨识电机转子位置的方法,并采用差分进化(DE)算法优化小波神经网络结构,从而提高了转子位置辨识的精度和收敛速度。最后通过仿真和实验证明所提出的方法辨识转子位置精度很高,且具有很强的自适应能力,可准确获得BLDCM换相信号。
Based on analysis of the back-EMF zero crossing detection principle, the relationship between motor back- EMF zero-crossing point and motor commutation point is derivated, and the relationship between line voltage and line EMF,the relationship between line voltage and rotor position is established.However,brushless DC motor （BLDCM） is a complex nonlinear systems, and some parameters of the motor are also changing during operation, and therefore to obtain accurate rotor position directly through the line voltage is more difficult.So the paper propos- es adaptive wavelet neural network using three line voltage as the input signal to identify the motor rotor position, and uses of differential evolution（DE） algorithm to optimize wavelet neural network structure, thus the accuracy and the rotor position identification convergence speed are improved.Finally the simulation and experiment show that this paper puts forward of the method recognize to rotor position accuracy very high, and has very strongly from orientation ability, and can obtain the exact communication signals.
出处
《电力电子技术》
CSCD
北大核心
2012年第4期87-89,共3页
Power Electronics
关键词
无刷直流电机
转子位置
自适应小波神经网络
brushless direct current motor
rotor position
adaptive wavelet neural network
