|
标题：基于神经网络的集中绕组BLDC转子位置估计 |
|
作者：钟黎萍,江恩祝,张水平，等 |
|
| 2013年第4期 访问次数：265次 |
|
摘 要：集中绕组的BLDC的高频电流响应中含有多重位置信息，直接运用外差法进行位置估计存在较大误差。由于这些位置信息的频率与转速相关，在低速和零速下无法用传统的滤波器将有用信号分离。利用神经网络进行处理，则无需考虑信号的频率结构，可直接映射出转子位置。实验结果证明了该方法的可行性。
关键词：神经网络；旋转变换；位置估计
Abstract: There are multiple rotor position information in the high frequency current response of BLDC with concentrated winding, and it would produce a large errors if directly using the method of heterodyne to estimate the rotor position. Furthermore, it is impossible to extract useful information through typical filter at low-speed and zero-speed because the frequencies of the signals are relative to the motor speed. In this condition, neural network is a good option for it can directly map out rotor position regardless of the response’s frequency structure. The related experiments were made and the results shown that the method was practicable.
Key words: neural network; rotation transformation; position estimation
|
|
| |
|
