摘要
轴心轨迹是反映转子轴振的重要特征,其中隐含着转子的各种故障信息。提出了利用转子轴心轨迹图像不变矩与SVM的转子轴心轨迹自动识别方法。该方法首先对转子轴心轨迹图形进行填充处理,得到转子轴心轨迹图像,然后计算图像的7个不变矩。将提出的7个不变矩作为特征向量结合SVM分类方法对故障进行诊断。经实验分析和验证,该方法可以有效实现转子故障的诊断。
Axis orbit of rotor is a reflection of the important characteristics of shaft vibration, which implies the rotor fault information. This paper proposes an intelligent dentification method of shaft orbit based on support vector machines and invariant moments. The first step is filling The rotor orbit graphics, get rotor orbit image. The second step is calculating the image of 7 invariant moments, combined with SVM to diagnosis fault. According to experimental analysis and validation, this method can effectively achieve the rotor fault diagnosis.
出处
《汽轮机技术》
北大核心
2012年第6期469-472,414,共5页
Turbine Technology
关键词
支持向量机
不变矩
特征提取
轴心轨迹
SVM
support vector machine
invariant moments
feature extraction
shaft orbit
SVM
