华南理工大学学报(自然科学版) ›› 2023, Vol. 51 ›› Issue (12): 42-52.doi: 10.12141/j.issn.1000-565X.220593

所属专题: 2023年机械工程

• 机械工程 • 上一篇    下一篇

多尺度时空信息融合驱动的图神经网络故障诊断方法

赵荣超1 吴百礼1,5 陈祝云1,2,3 温楷儒1 张绍辉4 李巍华1,2   

  1. 1.华南理工大学 机械与汽车工程学院,广东 广州 510640
    2.琶洲实验室,广东 广州 510335
    3.北京信息科技大学重点科研机构 北京 100192
    4.东莞理工学院 机械工程学院,广东 东莞 523808
    5.广东石油化工学院 广东省石化装备 故障诊断重点实验室,广东 茂名 525000
  • 收稿日期:2022-09-09 出版日期:2023-12-25 发布日期:2023-04-24
  • 通信作者: 陈祝云(1990-),男,博士,副研究员,主要从事机械动态信号处理、装备健康管理与智能运维等研究。 E-mail:mezychen@scut.edu.cn
  • 作者简介:赵荣超(1987-),男,博士,副教授,主要从事汽车热管理和故障诊断研究。E-mail: merczhao@scut.edu.cn
  • 基金资助:
    国家自然科学基金资助项目(52205101);广东省基础与应用基础研究基金区域联合基金青年基金资助项目(2021A1515110708);广州市基础研究计划基础与应用基础研究项目(202201010615);北京信息科技大学重点科研机构项目(KF20212223204)

Graph Neural Network for Fault Diagnosis with Multi-Scale Time-Spatial Information Fusion Mechanism

ZHAO Rongchao1 WU Baili1,5 CHEN Zhuyun1,2,3 WEN Kairu1 ZHANG Shaohui4 LI Weihua1,2   

  1. 1.School of Mechanical and Automotive Engineering,South China University of Technology,Guangzhou 510640,Guangdong,China
    2.PAZHOU LAB,Guangzhou 510335,Guangdong,China
    3.Beijing Key Laboratory of Measurement Control of Mechanical and Electrical System Technology,Beijing Information Science Technology University,Beijing 100192,China
    4.College of Mechanical Engineering,Dongguan University of Technology,Dongguan 523808,Guangdong,China
    5.Guangdong Provincial Key Laboratory of Petrochemical Equipment Fault Diagnosis,Guangdong University of Petrochemical Technology,Maoming 525000,Guangdong,China
  • Received:2022-09-09 Online:2023-12-25 Published:2023-04-24
  • Contact: 陈祝云(1990-),男,博士,副研究员,主要从事机械动态信号处理、装备健康管理与智能运维等研究。 E-mail:mezychen@scut.edu.cn
  • About author:赵荣超(1987-),男,博士,副教授,主要从事汽车热管理和故障诊断研究。E-mail: merczhao@scut.edu.cn
  • Supported by:
    the National Natural Science Foundation of China(52205101);the Guangdong Basic and Applied Basic Research Foundation under Grant(2021A1515110708)

摘要:

行星齿轮箱作为机械系统中常见的减速装置,由于长期在强噪声环境和变工况工作条件下运行,导致采集到的振动信号故障特征微弱、信号模式多变难以识别,针对行星齿轮箱故障诊断效果不佳,泛化能力差的问题,提出一种多尺度时空信息融合驱动的图神经网络故障诊断方法来提高故障诊断模型准确率和泛化能力。该方法首先构建多尺度卷积核对原始时序信号进行不同尺度特征提取,削弱强噪声信号对有效信息的掩盖作用并增强故障特征的表达能力;然后再构造通道注意力机制,根据通道特征重要程度,给不同尺度卷积核提取的特征自适应分配不同权重,对含有关键故障特征的信息片段进行特征强化;最后对卷积输出的多尺度特征,构造空域下的图数据并通过图卷积网络聚合多尺度特征,从而有效利用数据的时序多维信息和空域结构关联信息,实现多尺度下时空域故障信息的深度融合,提高诊断的准确精度和模型的泛化性能。通过利用具有行星齿轮箱结构的风电装备故障数据集对所提方法进行验证,并与其他深度学习方法(第一层宽卷积核深度卷积神经网(WDCNN)、长短时记忆网络(LSTM)、残差网络(ResNet)、多尺度卷积神经网络(MSCNN))进行比较,结果表明:本研究提出的方法在跨负载和跨转速两种工况下的平均诊断准确率分别可以达到98.85%与91.29%,明显优于其他对比方法,验证了本研究提出的方法的强泛化性能和优越性。

关键词: 故障诊断, 多尺度注意力机制, 图神经网络, 深度学习

Abstract:

Due to the long-term operation of planetary gearboxes in strong noise environments and changing working conditions, the collected vibration signals exhibit weak fault characteristics and variable signal patterns, making them difficult to identify. Intelligent fault diagnosis of planetary gearboxes under these conditions remains a challenging task. In order to achieve high diagnostic accuracy and strong model generalization performance, a fault diagnosis method using a graph neural network with a multi-scale time-spatial information fusion mechanism is proposed. The method first uses convolution kernels of different scales to extract features from the original vibration signal, reducing the masking effect of strong noise signals on valuable information and enhancing its feature expression ability. A channel attention mechanism is then constructed to adaptively assign different weights among different channels to features of different scales, enhancing features in segments of information containing crucial fault characteristics. Finally, the multi-scale features of the convolution module output are used to construct graph data with spatial structure information for graph convolution learning. This approach allows for the full utilization and deep fusion of multi-dimensional time domain information and spatial correlation information, effectively improving the accuracy of diagnosis and the generalization performance of the model. The proposed method was verified using a fault dataset of wind power equipment with planetary gearbox structure. The average diagnosis accuracy of the proposed method was found to reach 98.85% and 91.29% under cross-load and cross-speed conditions, respectively. These results are superior to other intelligent diagnosis methods, including deep convolutional neural networks with wide first-layer kernels (WDCNN), long short-term memory network (LSTM), residual network (ResNet), and multi-scale convolution neural network (MSCNN). Therefore, the strong generalization performance and superiority of the proposed method were confirmed.

Key words: fault diagnosis, multi-scale attention mechanism, graph neural network, deep learning

中图分类号: