基于稀疏编码的振动信号特征提取算法与实验研究

PDF(1738 KB)

振动与冲击 ›› 2014, Vol. 33 ›› Issue (15) : 76-81.

论文

基于稀疏编码的振动信号特征提取算法与实验研究

苗中华1，周广兴1，刘海宁2，刘成良2

作者信息 +

Study on Experiments and Feature Extraction Algorithm of vibration signals Based on Sparse Coding

Miao Zhonghua 1 Zhou Guangxing 1 Liu Haining2 Liu Chengliang 2

Author information +

文章历史 +

摘要

针对海量冗余数据中设备状态信息特征提取问题，借鉴生物感知系统“冗余度压缩”的信息处理原则，基于神经科学研究中的稀疏编码算法，提出了连续长时间采样时振动信号有效特征提取方法。介绍了稀疏编码算法及其模型，详细研究了稀疏编码的系数求解和字典学习两大问题。基于人工轴承故障数据集进行了实验研究，实验表明：基于稀疏编码的振动信号特征提取算法不仅能有效提取设备状态特征，而且稀疏特征具有良好的可分性。该方法可用于设备故障诊断，为基于状态的设备智能维护提供有效工具。

Abstract

To solve the problem of status information feature's extraction of the mass redundant data in device, an effective feature extraction method of vibration signal which is continuously sampled in a long time is presented. This method is rooted in the principle of information processing in the biological perceptual system's redundancy compression, and based on sparse coding algorithm in neuroscience research. The two problems of the sparse coding algorithm (Coefficient solving and Dictionary learning) are researched in detail. Experiments about the proposed method have been accomplished based on the artificial bearing fault data sets. The results indicate that the proposed vibration signal feature extraction method not only can effectively extract status characteristics from the device information, but also has good separability in the sparse feature. This method can be used for equipment's fault diagnosis, and also provides an effective tool for intelligent maintenance of equipment based on state.

导出引用

苗中华;周广兴;刘海宁;刘成良. 基于稀疏编码的振动信号特征提取算法与实验研究[J]. 振动与冲击, 2014, 33(15): 76-81

Miao Zhonghua Zhou Guangxing Liu Haining Liu Chengliang . Study on Experiments and Feature Extraction Algorithm of vibration signals Based on Sparse Coding[J]. Journal of Vibration and Shock, 2014, 33(15): 76-81