摘要
结合可视化技术和隐马尔可夫模型,提出一种基于可视化重心特征提取的轴承故障诊断新模型。首先,将轴承多维音频数据表示为雷达图的形式,再从雷达图中提取可视化的重心特征;然后,将重心特征进行可视化绘制的同时送入HMM分类器进行故障诊断,实现了可视化诊断和自动诊断2种诊断方式。试验结果表明,该方法不仅能够直观显示轴承信号,而且诊断精度可达97%,平均诊断时间为0.08 s。
A new method of bearing fault diagnosis based on barycenter feature of visualization is presented,combined with visualization and Continuous Gaussian Mixture Hidden Markov Model (HMM).Firstly,the acoustic signal of bearing is showed in radar chart,from which barycenter feature can be exacted;then barycenter feature is visually drawn and sent to HMM classifier for fault diagnosing simultaneously,so as to achieve two diagnosis modes:visual di-agnosis and automatic diagnosis.The result of experiments shows this new method could not only show bearing signal visually,but also has higher diagnosis accuracy (97%)and faster calculating speed (the average time of diagnosis is 0.08 s).
出处
《轴承》
北大核心
2014年第12期54-57,共4页
Bearing
基金
国家自然科学基金项目(61103108)
湖南省教育厅科研项目(13C879)
湘南学院[2012]125号NO2计算机应用技术创新训练中心项目
湘南学院"十二五"重点学科计算机应用技术学科项目
关键词
滚动轴承
故障诊断
雷达图
重心特征
可视化
rolling bearing fault diagnosis radar chart barycenter feature visualization