基于非负矩阵分解与主元分析的时频图像识别方法研究

PDF(1551 KB)

振动与冲击 ›› 2012, Vol. 31 ›› Issue (18) : 169-172.

论文

基于非负矩阵分解与主元分析的时频图像识别方法研究

陈李宏坤1,2, 禹臻1，周帅1，张志新1

作者信息 +

Investigation on Time-Frequency Image Classification by Using Non-Negative Matrix Factorization and Principal Component Analysis

Li Hong-Kun, Chen Yu-Zhen, Zhou Shuai, Zhang Zhi-Xin

Author information +

文章历史 +

摘要

本文应用非负矩阵分解与主元分析对时频图像处理，在此基础上进行设备状态识别。论述了对振动信号应用Hilbert谱构建二维时频图像，并用非负矩阵分解对时频图像构造特征向量，应用主元分析对提取的特征向量进行了降维处理，并在三维坐标系中进行表示和状态识别。以滚动轴承不同状态的识别为例，验证方法的有效性。研究表明此方法能够提高设备状态识别的准确性，有利于设备故障诊断的发展。

Abstract

In this paper, a method for machine condition classification is put forward based on Non-negative matrix factorization (NMF) and Principal component analysis (PCA). Vibration signal is used to construct Hilbert two-dimensional time frequency image after pre-processing. Then, NMF is used to determine feature vector for time frequency image. Principal component analysis (PCA) is used to reduce dimension for feature vector, which will be useful for three-dimension condition classification. Rolling bearing different conditions classification is as an example to testify the effectiveness of this method. It can be concluded that this method can improve accuracy for machine condition classification. It will contribute the development for machine fault diagnosis.

导出引用

陈李宏坤;禹臻;周帅;张志新. 基于非负矩阵分解与主元分析的时频图像识别方法研究[J]. 振动与冲击, 2012, 31(18): 169-172

Li Hong-Kun;Chen Yu-Zhen;Zhou Shuai;Zhang Zhi-Xin. Investigation on Time-Frequency Image Classification by Using Non-Negative Matrix Factorization and Principal Component Analysis[J]. Journal of Vibration and Shock, 2012, 31(18): 169-172