基于神经网络的发动机故障诊断分类器设计  

Engine fault diagnosis classifier design based on Neural Network

作　　者：于柏森[1] 张云鹏[1] 

机构地区：[1]长春工业大学基础科学学院,吉林长春130012

出　　处：《长春工业大学学报》2010年第2期143-147,共5页Journal of Changchun University of Technology

摘　　要：建立了类间与类内距离之比作为特征提取中的可分判据,设计了分类器的拓扑结构,采用动量梯度下降法作为分类器的训练算法。结果表明,该分类器能够满足工程实际的需要。The distance ratios between the classes and within the classes are built as a divide criterion for the feature extraction, and the topology of the classifier is designed. We apply the momentum gradient descent method as the classifier training algorithm and the results show that the classifier can meet the practical needs.

关 键 词：故障诊断 分类器 神经网络 

分 类 号：TP806.3[自动化与计算机技术—检测技术与自动化装置]