中国机械工程 ›› 2013, Vol. 24 ›› Issue (24): 3345-3348.

• 机械基础工程 • 上一篇    下一篇

一种基于LLE特征融合的故障识别方法

胡建中;吴瑶;谢小欣   

  1. 东南大学,南京,211189
  • 出版日期:2013-12-25 发布日期:2013-12-27
  • 基金资助:
    国家自然科学基金资助项目(51075069,51075070)

A Method for Fault Recognition Based on LLE Feature Fusion

Hu Jianzhong;Wu Yao;Xie Xiaoxin   

  1. Southeast University,Nanjing,211189
  • Online:2013-12-25 Published:2013-12-27
  • Supported by:
    National Natural Science Foundation of China(No. 51075069,51075070)

摘要:

针对传统的故障识别中未能充分利用特征信息的问题,提出一种基于局部线性嵌入(LLE)特征融合的故障识别方法,通过初步提取信号时域和时频域的特征获得原始特征集,利用LLE算法对原始特征集进行二次特征提取,进一步融合两组特征集并使用KNN算法进行故障识别。仿真信号数据分析与实际故障分析证明了所提方法对故障样本识别的可行性和有效性。

关键词: 特征提取, 局部线性嵌入(LLE), 特征融合, 故障识别

Abstract:

Aiming at the problem of traditional fault recognition which failed to make full use of the feature information, a method which made feature fusion for fault recognition based on the LLE algorithm was presented, an initial extraction was obtained by extracting the time domain features and time-frequency domain features of signals. A secondary feature extraction for the initial feature sets was obtained by LLE algorithm, then a fusion of  these two groups of feature set was made and the KNN algorithm was used for fault recognition. The simulation data analysis and experiments show the feasibility and effectiveness of this method for fault sample recognition.

Key words: feature extraction, locally linear embedding(LLE), feature fusion, fault recognition

中图分类号: