摘要
针对工业上滚动轴承提取特征难,传统诊断方法需要专业人员的专业知识并且难以保证故障诊断的准确性的问题,采用将卷积神经网络(CNN)和支持向量机分类(SVM)相结合,提出一种用于滚动轴承的故障诊断模型。该模型采用一维卷积神经网络与二维卷积网络相结合,故障轴承原始信号作为输入,通过CNN提取滚动轴承故障信号特征值;并将一维卷积网络输出与二维卷积网络输出拼接在一起,最后通过SVM分类器完成故障分类。为了验证所提出的故障诊断模型,基于SGD搭建网络模型,将CWRU数据集中的电机端故障如内圈故障、外圈故障等典型故障类别进行整理,并利用归一化增加模型泛化能力。通过实验证明了该方法的诊断准确性高达99.25%,可实现滚动轴承故障特征的自适应提取,且与单独使用传统特征提取方法相比,具有更好的识别能力。
It is difficult to extract the characteristics of rolling bearings in industry,and the traditional diagnosis method requires the expertise of professionals and is difficult to ensure the accuracy of fault diagnosis.In order to solve the above problems,this paper combines convolutional neural network(CNN)with support vector machine(SVM)classification to proposes a fault diagnosis model for rolling bearings.The proposed model uses a combination of one-dimensional CNN and two-dimensional CNN,and the original fault bearing signal as input,and the fault signal characteristic value of rolling bearing is extracted by CNN.Then one-dimensional convolutional network output and the two-dimensional convolutional network output are spliced together,finally,the fault classification is completed by SVM classifier.In order to verify the proposed fault diagnosis model,a network model is built based on SGD,and typical fault categories such as inner ring fault and outer ring fault in CWRU data set are sorted,and normalization is used to increase the model generalization ability.Experimental verification is conducted to show that the proposed method has a diagnostic accuracy of up to 99.25%,able to realize the adaptive extraction of rolling bearing fault features,and has better recognition ability compared with the traditional feature extraction method used alone.
作者
刘登
LIU Deng(EVE Energy Co.,Ltd.,Huizhou Guangdong 516006,China;Hubei University of Technology,Wuhan Hubei 430068,China)
出处
《湖北电力》
2024年第2期146-153,共8页
Hubei Electric Power
作者简介
刘登(1998),男,湖北汉川人,研究生,硕士,助理工程师,研究方向为故障诊断。
