基于CNN-SVM的核电厂轴承故障诊断方法  

Bearing fault diagnosis method in nuclear power plants based on CNN-SVM

作　　者：尹文哲 夏虹[1,2] 彭彬森 朱少民 王志超[1,2] 张汲宇 姜莹莹 YIN Wenzhe;XIA Hong;PENG Binsen;ZHU Shaomin;WANG Zhichao;ZHANG Jiyu;JIANG Yingying(Key Laboratory of Nuclear Safety and Advanced Nuclear Energy Technology,Ministry of Industry and Information Technology,Harbin Engineering University,Harbin 150001,China;Fundamental Science on Nuclear Safety and Simulation Technology Laboratory,Harbin Engineering University,Harbin 150001,China)

机构地区：[1]哈尔滨工程大学核安全与先进核能技术工业和信息化部重点实验室,黑龙江哈尔滨150001 [2]哈尔滨工程大学核安全与仿真技术重点学科实验室,黑龙江哈尔滨150001

出　　处：《哈尔滨工程大学学报》2023年第3期410-417,共8页Journal of Harbin Engineering University

基　　金：国家自然科学基金项目(51379046);黑龙江省自然科学基金项目(E2017023).

摘　　要：为提升核电厂旋转机械部件的故障诊断准确率,以及增强诊断模型泛化能力,本文提出了一种基于卷积神经网络和支持向量机的滚动轴承故障诊断方法。对轴承原始振动信号进行连续小波变换,得到其时频图;然后,使用预训练好的卷积基对小波时频图进行特征提取,获取深层特征,并将这些深层特征正则化处理后,使用主成分分析法对其进行降维;将得到的特征数据输入到基于粒子群优化的支持向量机中,从而实现滚动轴承的故障诊断。实验结果表明:该方法对不同负载工况下的多类滚动轴承故障具有良好的诊断效果,并且在噪声干扰下也能保持较好的效果,与其他方法相比,其抗噪稳定性更好,泛化能力更强。To improve the accuracy of fault diagnosis of rotating mechanical components in a nuclear power plant and to enhance the generalization ability of diagnostic models,this paper presents a rolling bearing fault diagnosis method based on a convolutional neural network and support vector machine(SVM).First,a continuous wavelet transform is performed on the original vibration signals of the rolling bearing to obtain its time-frequency map.Then,the pre-trained convolutional base is used to extract features from the wavelet time-frequency map to obtain deep features,which are further regularized.Furthermore,the principal component analysis method is employed to reduce its dimensionality.Finally,the obtained feature data are inputted into the SVM based on particle swarm op-timization to realize the fault diagnosis of the rolling bearing.Experimental results show that the proposed method has good performance in fault diagnosis under different load conditions.Comparisons with other methods highlight the advantages of the proposed method in anti-noise robustness and generalization ability.

关 键 词：核电厂 滚动轴承 故障诊断 深度学习 卷积神经网络 支持向量机 粒子群优化 数据驱动 

分 类 号：TK05[动力工程及工程热物理]