基于基座多传感核主元分析的故障诊断.pdfVIP

学兔兔 第32卷 第7期 仪 器 仪 表 学 报 V01．32 No．7 2011年7月 Chinese Journal of Scientific Instrument Ju1．2011 基于基座多传感核主元分析的故障诊断米 李学军 ，杨大炼 ，郭灯塔 ，蒋玲莉 (1湖南科技大学，湖南省机械设备健康维护重点实验室 湘潭 411201；2湘潭电机股份有限公司 湘潭 411101) 摘 要：传统振动测试中传感器一般安装在靠近振源位置，此类方式采集的振动信号的故障特征相对明显，但实际应用中常存 在传感器安装不便的矛盾。本文提出基于基座的故障诊断方法，以多个传感器采集的基座上的微弱故障特征信号为原始信号， 通过核主元分析(kernel principal component analysis，KPCA)进行信息融合，利用支持向量机(suppo~vector machine，SVM)进行 模式识别。文章还探析了核参数选取对KPCA性能的影响，最后实例验证了基于基座故障诊断方法的可行性。 关键词：基座；核主元分析；故障诊断；信息融合；支持向量机 中图分类号：TP206 ．3文献标识码：A 国家标准学科分类代码：520．2040 Fault diagnosis method based on multi—sensors installed on the base and KPCA Li Xuejun’，Yang Dalian ，Guo Dengta ，Jiang Lingli (J．Hunan Provincial Key Laboratory ofHealth Maintenancefor Mechanical Equipment，Hunan University of Science and Technology，Xiangtan 411201，China；2．Xiangtan Electric Co．Ltd，Xiangtan 411101，China) Abstract：Sensors are usually installed near the vibration source location in traditional vibration testing；in this way， the fault features of vibration signal are relatively obvious．However，in practical applications，it is often ineonven— ience to install the sensors in this way．A fault diagnosis method based on sensors installed on the base is proposed in this paper；multiple sensors are used to obtain weak fault signals on the base as the original signals and KPCA is used for information fusion，then SVM is used for pattern recognition．This paper also analyzes the effects of the ker— nel parameters 0n KPCA performance；experiment examples verify that the proposed fault diagnosis method is feasi- ble． Key words：base；KPCA；fault diagnosis；information fusion；SVM 诊断技术