摘要
为了提高机械故障诊断的正确率,针对当前机械故障诊断方法存在局限性,设计了基于传感器器信号融合的机械故障诊断方法。首先采用多个传感器对机械工作状态的信号进行采集,并从信号中抽取机械故障信息特征,然后采用主成分分析选择比较重要的机械故障信息特征,并采用支持向量机对机械故障信息特征向量进行学习,建立机械故障诊断的分类器,最后采用实际的机械故障数据进行了性能测试,结果表明,本文方法解决了当前机械故障诊断中的难题,获得了十分理想的机械故障诊断效果,而且机械故障诊断正确率要高于其它方法,具有广泛的应用前景。
In order to improve the accuracy of mechanical fault diagnosis for the current limitation, a new method based on sensor fusion was proposed in this paper. First, collect the signal of mechanical working situation from several sensors and extract the information features of mechanical fault from them. Then, select important ones by main components analyzing and study on these vectors of mechanical information feature through support vector machine ( SVM ) to build the classifier of mechanical fault diagnosis. Finally, do the performance testing through actual mechanical fault data. The experimental results show that using this way to solve the current problems of mechanical diagnosis has an ideal effect. Moreover, its accuracy is higher than other ways and it has wide application prospect.
出处
《激光杂志》
北大核心
2017年第7期184-187,共4页
Laser Journal
基金
2015年河北省科技厅项目(15274514)
2015年河北省自然科学基金资助项目(B2015110004)
关键词
传感器
信号采集
机械故障诊断
特征选择
sensor
signal acquisition
mechanical fault diagnosis
feature selection