摘要
针对风力发电机是一个复杂的时变非线性系统难以提取有效故障特征的问题,首次提出一种优化的局部判别基(LDB)算法结合SOM-BP混合网络进行故障诊断与定位的新方法。首先利用改进的LDB算法提取初始的故障特征,为进一步提高类间可分离度,将这个初始的故障特征通过自组织特征映射(SOM)网络映射到一个类别可分性更高的特征空间,最后利用反向传播(BP)网络根据映射后的特征实现非线性分类,完成故障诊断与定位。
As wind power generation is a complicated nonlinear time-varying system, it' s hard to extract effective fault feature. A novel algorithm that combined the modified local discriminant basis (LDB) algorithm and SOM-BP network is proposed to fault diagnosis and isolation. Extracting primal fault feature by improved LDB algorithm, then map this incipient fault feature into a new feature space with high class separability via self-organizing feature map (SOM) nonlinearly transform, finally BP is used as the nonlinear classifier to implement fault diagnosis and isolation.
出处
《电工技术学报》
EI
CSCD
北大核心
2009年第4期224-228,共5页
Transactions of China Electrotechnical Society
基金
广东省科技计划(0711050600004)
广东省自然科学基金(032030)资助项目
关键词
小波包
局部判别基
神经网络
故障诊断
Wavelet packet, local discriminant basis, neural network, fault diagnosis