首页|变工况下动态卷积域对抗图神经网络故障诊断

变工况下动态卷积域对抗图神经网络故障诊断

扫码查看
原文链接
  • 国家科技期刊平台
  • NETL
  • NSTL
  • 万方数据
针对基于无监督域自适应故障诊断方法忽略了域间数据结构信息、传统域对齐平均最大差异法全局泛化能力差等问题,本文提出一种基于无监督域自适应理论的动态卷积域对抗图神经网络故障诊断模型,分别通过对数据的类别标签、域标签和数据结构信息进行建模.通过分类器和域鉴别器分别建模类别标签和域标签,通过图神经网络将数据结构信息嵌入到实例图节点中,利用高斯Wasserstein距离来度量不同领域的实例图之间的差异.本文对比了不同工况下共 14 种迁移任务在各模型下故障识别的准确率.实验结果表明:基于动态卷积的域对抗图神经网络模型在变工况下的故障诊断效果均优于其他对比模型,且模型性能稳定.
Fault diagnosis of dynamic convolutional domain adversarial graph neural networks under variable working conditions
This paper proposes a dynamic convolutional domain adversarial graph neural network fault diagnosis model based on unsupervised domain adaptation theory.Traditional domain alignment methods,such as the average maximum difference method,often ignore inter-domain data structure information and exhibit poor global generaliza-tion ability.This model addresses these issues by separately modeling the category labels,domain labels,and data structure information.The approach embeds data structure information into instance graph nodes using a graph neu-ral network.Category labels and domain labels are modeled using classifiers and domain discriminators,respective-ly.The Gaussian Wasserstein distance is used to measure the differences between instance graphs in different do-mains.This paper compares the accuracy of fault identification across 14 migration tasks using various models and under different operating conditions.Experimental results show that the domain adversarial graph neural network model,based on dynamic convolution,outperforms other comparative models in fault diagnosis under varying oper-ating conditions and maintains stable performance.

unsupervised domain adaptationdynamic convolutiondomain adversarialgraph neural networkgraph generationGaussian Wasserstein distancefault diagnosisvariable working condition

王桐、王晨程、邰宇、欧阳敏、陈立伟

展开 >

哈尔滨工程大学 信息与通信工程学院,黑龙江 哈尔滨 150001

哈尔滨工程大学 先进船舶通信与信息技术工业和信息化部重点实验室,黑龙江 哈尔滨 150001

黑龙江省政务大数据中心,黑龙江 哈尔滨 150028

无监督域自适应 动态卷积 域对抗 图神经网络 图生成 高斯Wasserstein距离 故障诊断 变工况

国家自然科学基金项目国防科技重点实验室基金项目先进船舶通信与信息技术工业和信息化部重点实验室项目中央高校基本科研业务费项目

611021056142209190107AMCIT2101-083072022QBZ0806

2024

10.11990/jheu.202208005

哈尔滨工程大学学报
哈尔滨工程大学

哈尔滨工程大学学报

CSTPCD北大核心
影响因子：0.655
ISSN：1006-7043
年,卷(期)：2024.45(7)