卢绮雯1  林声宇2  邓昌2  刘夏丽2  黄志海2

（1.广州智柔智能科技有限公司，广东 广州，510006

2.广东工业大学机电工程学院，广东 广州，510006）

摘要：通过对电磁继电器异响声音信号进行时频分析和异响特征提取方法的研究，选择梅尔频谱来提取异响特征。首先，对异响声音信号进行预处理；然后，通过短时FFT变换得到异响声音信号的频谱图；接着，通过梅尔滤波器组得到梅尔频谱；最后，经过实验验证，梅尔频谱可以较好地表达电磁继电器异响特征，具有较强的抗噪声干扰能力，且数据大小适中，可为后续的故障诊断提供有效信息。

关键词：电磁继电器；内部异响；梅尔频谱；特征提取；时频分析

中图分类号：TP391.4             文献标志码：A        文章编号：1674-2605(2023)06-0007-07

DOI：10.3969/j.issn.1674-2605.2023.06.007

Extraction of Internal Abnormal Noise Features of 

Electromagnetic Relay Based on Mel Spectrum 

LU Qiwen1  LIN Shengyu2  DENG Chang2  LIU Xiali2  HUANG Zhihai2

(1.Guangzhou Zhirou Intelligent Technology Co., Ltd., Guangzhou 510006, China

2.School of Mechanical and Electrical Engineering, Guangdong University of Technology,           Guangzhou 510006, China)

Abstract: By conducting time-frequency analysis on the abnormal sound signal of electromagnetic relay and studying the method of extracting abnormal sound features, Mel spectrum is selected to extract abnormal sound features. Firstly, preprocess the abnormal sound signal; Then, the frequency spectrum of the abnormal sound signal is obtained through short-term FFT transformation; Next, obtain the Mel spectrum through the Mel filter bank; Finally, through experimental verification, the Mel spectrum can better express the abnormal noise features of electromagnetic relay, has strong anti noise interference ability, and the data size is moderate, which can provide effective information for subsequent fault diagnosis.

Keywords: electromagnetic relay; internal abnormal noise; Mel spectrum; feature extraction; time-frequency analysis