多度量下ResGAT的风力发电机齿轮箱故障诊断

李明; 曹洁; 刘宗礼; 王进花

doi:doi:10.19912/j.0254-0096.tynxb.2024-0173

English Version Reader LoginJournal Login

您当前的位置：

首页 >

文章列表页 >

多度量下ResGAT的风力发电机齿轮箱故障诊断

更新时间：2025-12-02

- 多度量下ResGAT的风力发电机齿轮箱故障诊断
- Vol. 46, Issue 6, Pages: 683-690(2025)
- 作者机构：
- 作者简介：
- 基金信息：
- DOI：doi:10.19912/j.0254-0096.tynxb.2024-0173
  CLC： TP277
- Published Online：07 July 2025，
  
  Published：2025
- 稿件说明：
移动端阅览
李明, 曹洁, 刘宗礼, et al. 多度量下ResGAT的风力发电机齿轮箱故障诊断[J]. 2025, 46(6): 683-690.
DOI：

李明, 曹洁, 刘宗礼, et al. 多度量下ResGAT的风力发电机齿轮箱故障诊断[J]. 2025, 46(6): 683-690. DOI： doi:10.19912/j.0254-0096.tynxb.2024-0173.

摘要

针对现有深度学习方法在风力发电机齿轮箱故障诊断方面的特征提取和样本相似性建模局限性

提出一种多种距离度量下残差连接的图注意力网络（ResGAT）。该方法构建全连接图以生成邻接矩阵

并结合多种距离度量方法

充分挖掘样本之间的相似性。利用图注意力网络进行节点特征聚合

结合残差连接以减轻模型梯度消失风险。进一步地

在Adam优化器中融入L2正则化及偏置校正

以降低过拟合问题。实验结果显示

ResGAT方法在WT-Planetary gearbox dataset齿轮箱数据集上能有效提取样本间相似性

并在风力发电机齿轮箱故障诊断上展现出优异性能。

Abstract

To address the limitations of existing deep learning methods in feature extraction and sample similarity modeling for wind turbine gearbox fault diagnosis

we propose a Residual Connected Graph Attention Network （ResGAT） that incorporates multiple distance metrics. This approach constructs a fully connected graph to generate an adjacency matrix and integrates various distance metric methods to fully explore the similarity between samples. Utilizing graph attention networks for node feature aggregation

coupled with residual connections

mitigates the risk of gradient vanishing. Furthermore

L2 regularization and bias correction are incorporated into the Adam optimizer to mitigate overfitting issues. Experimental results demonstrate that the ResGAT method effectively captures sample similarity on the WT-Planetary gearbox dataset and exhibits outstanding performance in fault diagnosis.

关键词

Keywords

references

张建超, 陈湘松, 郝如江, 等. 高速动车齿轮箱故障诊断技术研究综述[J]. 机械传动, 2023, 47(3): 133-140.

郭冲. 基于图神经网络的滚动轴承剩余寿命预测[D]. 北京: 北京化工大学, 2023.

谭启瑜, 马萍, 张宏立, 等. 基于图注意力网络的风力发电机齿轮箱故障诊断[J]. 太阳能学报, 2024, 45(1): 265-274.

LI T F, ZHAO Z B, SUN C, et al.Multireceptive field graph convolutional networks for machine fault diagnosis[J]. IEEE Transactions on industrial electronics, 2020, 68(12): 12739-12749.

HE K, YU D J, WANG D, et al.Graph attention network-based fault detection for UAVs with wultivariant time series flight data[J]. IEEE transactions on instrumentation and measurement, 2022, 71: 3530213.

EPISHIN V I.Studying fault tolerance of bipartite homogeneous minimal quasi-complete graphs using cisco packet tracer[C]//2021 IEEE Conference of Russian Young Researchers in Electrical and Electronic Engineering (ElConRus). St. Petersburg, Moscow, Russia, 2021: 318-320.

LI Q K, MA R X, LIAO M F, et al.A novel fault evaluation method based on nonlinear vibration features and Euclidean distance measurement for grid-like structures[J]. Structural health monitoring, 2023, 22(6): 4131-4148.

LIU X, LIU X M, ZHANG R L, et al.Securely computing the manhattan distance under the malicious model and its applications[J]. Applied sciences, 2022, 12(22): 11705.

WU C Z, WANG B.Extracting topics based on Word2Vec and improved jaccard similarity coefficient[C]//2017 IEEE Second International Conference on Data Science in Cyberspace (Dsc). Shenzhen, China, 2017: 389-397.

SHOPON M, BARI A S M H, GAVRILOVA M L. Residual connection-based graph convolutional neural networks for gait recognition[J]. The visual computer, 2021, 37(9): 2713-2724.

徐硕, 邓艾东, 杨宏强, 等. 基于改进残差网络的旋转机械故障诊断[J]. 太阳能学报, 2023, 44(7): 409-418.

赵小强, 罗维兰, 梁浩鹏. 基于SIR多级残差连接密集网络的轴承故障诊断[J]. 兰州理工大学学报, 2022, 48(6): 46-54.

白铂, 刘玉婷, 马驰骋, 等. 图神经网络[J]. 中国科学: 数学, 2020, 50(3): 367-384.

刘德澳, 董明刚, 叶威, 等. 基于图神经网络的非均匀风电场功率预测方法[J]. 太阳能学报, 2023, 44(11): 279-286.

KINGMA D, BA J.Adam: a method for stochastic optimization[J]. Computer science, 2014, 48(2): 2159-2167.

黄宝洲, 杨俊华, 沈辉, 等. 基于FFT的直驱式波浪发电系统功率优化控制[J]. 太阳能学报, 2021, 42(3): 206-213.

LIU D D, CUI L L, WANG H Q.Rotating machinery fault diagnosis under time-varying speeds: a review[J]. IEEE Sensors Journal, 2023, 23(24): 29969-29990.

曲建岭, 余路, 袁涛, 等. 基于一维卷积神经网络的滚动轴承自适应故障诊断算法[J]. 仪器仪表学报, 2018, 39(7): 134-143.

LIU Y X, JAFARPOUR B.Graph attention network with Granger causality map for fault detection and root cause diagnosis[J]. Computers & chemical engineering, 2024, 180: 108453.

QIN J H, HUANG Y J, WEN W S.Multi-scale feature fusion residual network for single image super-resolution[J]. Neurocomputing, 2020, 379: 334-342.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

基于改进MW-KECA算法的风电齿轮箱故障诊断研究

信息融合的NRP-AlexNet-SENet风电齿轮箱故障诊断

FAULT DIAGNOSIS METHOD OF WIDN TURBINE GEARBOX BASED ON MULTI-CORE PARELLEL RFECV-GNB

FAULT DIAGNOSIS OF WIND POWER GEARBOX BASED ON SHSVD-AS

Related Author

刘轩

霍忠堂

李冰

龙霞飞

何志成

曾进辉

周凌

梁凯

Related Institution

School of Electrical Engineering and Information Engineering, Lanzhou University of Technology

School of Information Engineering, Lanzhou City University

National Engineering Research Center of Power Generation Control and Safety, School of Energy and Environment, Southeast University

China Energy Jiangsu Power Co., Ltd.

National Engineering Research Center of Power Generation Control and Safety

AI问答

⁰