摘要
将SVM的分类算法应用于齿轮小样本故障诊断中。选取识别能力好的时域无量纲指标和频域中的9段频谱融合作为支持向量机的特征矢量,对齿轮的三种典型故障进行分类,结果表明:SVM在解决小样本情况下的机械故障诊断的分类问题中具有良好的应用前景。
SVM classification algorithms will been applied to gear fault diagnosis in small samples. Some dimensionless factors which have good identification ability and 9 frequency spectrum are selected as the input of Support Vector Machine in order to classify three kinds of fault. The result showed that SVM has good application prospects among the classification problem in small samples in mechanical fault diagnosis.
出处
《冶金设备》
2006年第5期48-51,共4页
Metallurgical Equipment
基金
湖北省机械传动与制造工程重点实验室资助项目(2005A04)
武汉科技大学科学研究发展基金资助(2005XY17)
关键词
统计学习理论
支持向量机
小样本
Statistical learning theory Support vector machine Small sample
