摘要
将小波包变换与LDA算法相结合,提出了一种基于LDA模型的滚动轴承故障类型检测新方法。首先通过小波包变换提取轴承振动信号的能量特征及其所包含的故障信息特征,并用"词袋"模型将故障信息特征表示成视觉词向量,然后利用LDA模型对轴承故障类型进行判别。试验表明,该方法能精确提取轴承的故障信息特征,快速检测出轴承的故障类型,与SVM等方法相比检测精度更高,鲁棒性更强,具有很好的故障检测效果。
A new method for fault type detection on rolling bearings based on LDA model is proposed by combining the LDA algorithm with a wavelet packet transform. Firstly the wavelet packet transform is used to extract the energy characteristics and containing fault information characteristics of vibration signals for rolling bearings,and the fault information characteristics is expressed as visual word vector by using the " bag of words" model. Then the fault types of rolling bearings are identified with the LDA model. The experiments show that the method can accurately extract the fault information characteristics of rolling bearings and rapidly obtain detection of faults and their types of bearings. Compared with SVM and other methods,it features higher accuracy,stronger robustness and better detection results.
出处
《轴承》
北大核心
2014年第7期42-46,共5页
Bearing
基金
湖南省科技计划项目(2014FJ3057)
湖南省教育科学"十二五"规划课题(XJK012CGD022)
湖南省普通高等学校教学改革研究课题(湘教通[2012]401号544)
关键词
滚动轴承
故障检测
小波包分解
LDA模型
rolling bearing
fault detection
wavelet packet decomposition
LDA model