基于峭度原则的VMD-SVD微型电机声音信号降噪方法  

A VMD-SVD micro-motor sound signal noise reduction method based on the kurtosis principle

作　　者：李伟光[1] 兰钦泓 马贤武 LI Weiguang;LAN Qinhong;MA Xianwu(School of Mechanical and Automotive Engineering,South China University of Technology,Guangzhou 510641,China)

机构地区：[1]华南理工大学机械与汽车工程学院,广东广州510641

出　　处：《中国测试》2023年第1期111-118,共8页China Measurement & Test

基　　金：国家自然科学基金项目(51875216);广东省自然科学基金项目(2017A050501004);广东省自然资源厅项目(2020030);广东省重点领域研发计划(2019B090918003)。

摘　　要：微型电机运转时的声音信号包含丰富的状态信息,可用于生产线上电机的快速检测,但由于待测电机体积小、声音能量低,采集过程中声音信号易与环境噪声耦合,导致声音信号提取和检测不准确。该文通过研究电机组成结构,分析声音信号频率成分与成因,得到该文研究电机的声音信号3倍频谐波特点,提出一种基于峭度原则的VMDSVD算法对电机声音信号进行提纯降噪,该算法采用VMD分段原理,对各分段信号进行SVD分解,提取谐波特征,利用峭度原则优化VMD参数选取。首先通过仿真信号对比实验,验证了该文算法具有更好的降噪效果和降噪性能指标。而后,将该方法应用于微型电机实测声音信号,测试结果表明提出的基于峭度原则VMD-SVD算法具有良好降噪效果,能够显著提高原始信号信噪比,更利于后续特征提取和故障检测工作。The sound signal of micro motor is rich in state information,which can be used for the rapid detection of motor on the production line.However,due to the small volume and low sound energy of the motor to be tested,the sound signal is easy to be coupled with the environmental noise in the process of acquisition,which leads to the inaccurate extraction and detection of sound signal.This paper studies the structure of motor,analyzes the frequency components and causes of sound signal,obtains the 3 times harmonic characteristics of the sound signal of the motor,a VMD-SVD algorithm based on the kurtosis principle is proposed to purify and reduce the noise of the motor sound signal.The algorithm uses the VMD segmentation principle to decompose each segmented signal by SVD,extract the harmonic characteristics,and use the kurtosis principle to optimize the selection of VMD parameters.Firstly,the simulation results show that the algorithm has better noise reduction effect and performance index.Then,the method is applied to the sound signal measured by micro motor.The test results show that the VMD-SVD algorithm based on the kurtosis principle proposed in this paper has a good noise reduction effect can significantly improve the signal-to-noise ratio of original signal,and is more conducive to the subsequent feature extraction and fault detection.

关 键 词：微型电机 声音信号降噪 变分模态分解(VMD) 奇异值分解(SVD) 

分 类 号：TP206.1[自动化与计算机技术—检测技术与自动化装置] TB9[自动化与计算机技术—控制科学与工程]