嵌入NLB模块的FCN在轴承信号降噪中的应用  

Application of FCN embedded in NLB module for bearing signal noise reduction

作　　者：范啸宇 刘韬[1,2] 王振亚 陶佳 朱振军 Fan Xiaoyu;Liu Tao;Wang Zhenya;Tao Jia;Zhu Zhenjun(Faculty of Mechanical and Electrical Engineering,Kunming University of Science and Technology,Kunming 650500,China;Engineering Research Center for Intelligent Maintenance of Advanced Equipment of Yunnan Province,Kunming 650500,China;SAIC-GM,Shanghai 201200,China)

机构地区：[1]昆明理工大学机电工程学院,昆明650500 [2]云南省先进装备智能制造技术重点实验室,昆明650500 [3]上汽通用汽车有限公司,上海201200

出　　处：《电子测量与仪器学报》2024年第4期55-65,共11页Journal of Electronic Measurement and Instrumentation

基　　金：云南省重大科技专项计划(202202AC080003,202202AC080008)资助。

摘　　要：深度学习在故障诊断取得了显著的进展,然而其多为端到端的智能诊断,在信号降噪方面的应用较少。本文提出了一种基于全卷积神经网络(fully convolutional network,FCN)的降噪方法。首先,模型整体采用了encoder-decoder架构,其中encoder部分由三层卷积层组成,decoder部分由四层反卷积层组成。其次,引入了残差连接对模型的学习目标进行了约束,使得模型在传播过程中更多地关注噪声信息。并且为了增强模型的特征提取能力,在encoder和decoder中引入了非局部块(non-local block,NLB)。然后,通过仿真信号对比实验选择网络的超参数,与目前主流的降噪方法进行对比,初步验证模型的降噪效果。最后,通过实际案例对所提方法的降噪效果进行对比验证,结果表明本文提出的方法在直观观察和降噪性能指标方面均取得了良好的应用效果,能够有效提高故障诊断的准确率。Deep learning has made significant progress in fault diagnosis,but it is mostly an end-to-end intelligent diagnosis with limited application in signal denoising.This article proposes a denoising method based on fully convolutional network(FCN).Firstly,the overall model adopts the encoder decoder architecture,where the encoder part consists of three convolutional layers and the decoder part consists of four deconvolution layers.Secondly,residual connections were introduced to constrain the learning objectives of the model,allowing the model to focus more on noise information during propagation.And in order to enhance the feature extraction ability of the model,non-local blocks(NLB)are introduced in the encoder and decoder.Then,through simulation signal comparison experiments,select the hyperparameters of the network and compare them with current mainstream noise reduction methods to preliminarily verify the noise reduction effect of the model.Finally,the denoising effect of the proposed method was compared and verified through practical cases.The results showed that the method proposed in this paper achieved good application effects in both intuitive observation and denoising performance indicators,and can effectively improve the accuracy of fault diagnosis.

关 键 词：全卷积神经网络 残差连接 反卷积 降噪 故障诊断 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]