基于支持向量机的船舶感应电机轴承故障在线诊断方法

The Online Diagnosis Method of Marine Induction Motor Bearing Fault Based on SVM

提出一种基于WELCH分析法结合支持向量机(SVM)的船舶感应电机轴承故障在线诊断方法,通过数字信号处理器(DSP)在线采集感应电动机的定子电流,使用WELCH法对电机的定子电流信号进行预处理,从预处理频谱分析中提取峰度、偏度、波峰因数、间隙和形状因子5个参数,形成支持向量机的特征向量,通过实验优化支持向量机的核函数参数,完成故障的识别。设计基于DSP的在线轴承故障诊断试验系统,实验结果证明,所提方法计算相对简单,可有效地在线识别电机轴承的外沟道故障。

An online diagnostic method for bearing fault information of marine asynchronous motor based on WELCH analysis combined with support vector machine(SVM)was proposed.The stator current of induction motor was collected and preprocessed online by digital signal processor(DSP),and the WELCH analysis and SVM method was used.Five parameters including kurtosis,skew-ness,crest factor,clearance and shape factor were extracted from the WELCH analysis to form the SVM feature vector.The kernel function parameters of SVM were optimized by the experiment in order to complete the fault identification.An online bearing fault diagnosis experiment system based on DSP was designed.The experimental results showed that the proposed method is relatively simple in calculation and effective in identification of the motor bearing outer raceway fault online.