摘要
针对传统神经网络用于开关磁阻电动机转子位置间接检测时存在网络结构确定困难和训练过程过于复杂的问题,将利用回声状态网络来实现转子位置检测。这种新型的网络利用储备池和线性回归算法简化了网络设计和训练过程,使得模型具有良好的收敛速度和实用性。利用离线获取的磁特性数据建立的转子位置预测模型,与基于BP和RBF神经网络的预测模型相比,在保证良好预测精度的前提下,具有计算简单,收敛速度快等优势。实验结果表明该模型可以快速准确地实现转子位置检测,为开关磁阻电动机的转子位置检测提供了一种新方法。
Focusing on the problems that the network structure is difficult to determining and the training process is too complex in the rotor position detection methods of switched reluctance motor using the traditional neural network, a rotor po- sition estimation method based on echo state network was proposed. The echo state network simplified the network design and the training process using the reserve pond and the linear regression algorithm, enables the model to have the good con- vergence rate and practicability. Under the premise to ensure good prediction accuracy, the rotor position prediction models using off-line magnetic characteristics data, compared with BP and RBF neural networks, has the advantages of simple cal- culation and fast convergence speed. Experimental results show that this model can realize rotor position detection of switched reluctance motor rapidly and accuracy, and is a new method for the rotor position estimation.
出处
《微特电机》
北大核心
2015年第6期58-61,共4页
Small & Special Electrical Machines
基金
中央高校基本科研业务费专项资金项目(CHD2011JC131)
陕西省微特电机及驱动技术重点实验室开放基金项目(2013SSJ2003)
关键词
开关磁阻电动机
递归神经网络
转子位置检测
回声状态网络
switch reluctance motor ( SRM )
recurrent neural network
rotor position detection
echo state network(ESN)