【作者】 李宏坤； 马孝江；

【Author】 Hongkun Li and Xiaojiang Ma (Key Laboratory for Precision and Non-traditional Machining Technology of Ministry of Education, Dalian University of Technology, Dalian 116024)

【机构】 大连理工大学精密与特种加工教育部重点实验室；

【摘要】 为了有效对往复机进行故障诊断,本文根据其振动信号和状态识别的特点,采用主元分析(PCA)和支持向量机(SVM)共同诊断方法。首先,对采集得到的振动信号从多个方面进行特征提取,获得更多状态信息。然后对提取的特征向量进行主元分析,从而得到两个新的更能反映设备状态的特征参数,组成一个新的特征向量。最后,将新的特征向量输入到训练好的支持向量机结构中进行压缩机的状态识别。以某石化公司一台往复式压缩机气阀磨损的状态识别为例,验证了此方法对于压缩机故障诊断的有效性。更多还原

【Abstract】 To effectively recognize reciprocating compressor’s working condition, this paper introduced a new method by combining principle component analysis (PCA) and support vector machine (SVM) for statistical recognition based on the vibration signal and statistical recognition characteristics for reciprocating compressor. Firstly, the features of monitored signal are extracted from different aspects. A multiple-parameter vector is obtained. Then, two new main parameters are taken out from the feature vector by using PCA. A new feature vector is obtained which can demonstrate the engine working condition. In the end, the condition of reciprocating engine is recognized by input the new feature vector to the trained SVM. The valve pattern recognition of a reciprocating compressor from a petroleum company is used as an example to testify this method. It can be concluded that this method is very effectiveness for reciprocating engine pattern recognition and fault diagnosis. It is a promising method and can contribute the development of reciprocating compressor pattern recognition and preventative maintenance.更多还原