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

doi:10.12141/j.issn.1000-565X.220593

Abstract

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

CLC Number:

TH17

ZHAO Rongchao, WU Baili, CHEN Zhuyun, et al. Graph Neural Network for Fault Diagnosis with Multi-Scale Time-Spatial Information Fusion Mechanism[J]. Journal of South China University of Technology(Natural Science Edition), 2023, 51(12): 42-52.

Figures/Tables 16

Fig.1

Fig.2

Fig.3

Fig.4

Table 1

Parameters of the proposed model"

网络结构	网络参数	输出大小
输入层		64 $×$ 1 024
多尺度卷积层	8 $×$ 7 $×$ 1，8 $×$ 15 $×$ 1，8 $×$ 31 $×$ 1，8 $×$ 63 $×$ 1	64 $×$ 32 $×$ 1 024
注意力层	32 $×$ 16，16 $×$ 32	64 $×$ 32 $×$ 1 024
卷积层	32 $×$ 3 $×$ 1	64 $×$ 32 $×$ 1 022
池化层	2 $×$ 1，最大池化	64 $×$ 32 $×$ 511
卷积层	64 $×$ 3 $×$ 1	64 $×$ 64 $×$ 509
卷积层	128 $×$ 3 $×$ 1	64 $×$ 128 $×$ 507
池化层	4 $×$ 1，自适应池化	64 $×$ 128 $×$ 4
全连接层	512 $×$ 256	64 $×$ 256
图生成层	256 $×$ 10	64 $×$ 64
图卷积层	256 $×$ 256	64 $×$ 256
图卷积层	256 $×$ 256	64 $×$ 256
全连接层	256 $×$ 5	64 $×$ 5

Table 1

Fig.5

Table 2

Table 3

Fig.6

Table 4

Fig.7

Table 5

Fig.8

Fig.9

Table 6

Fig.10

References 24

1	唐宏宾，傅政，邓习树，等．工程机械柱塞泵变载荷工况故障诊断方法［J］．华南理工大学学报（自然科学版），2021，49（2）：110-119，139.
	TANG Hongbin， FU Zheng， DENG Xishu，et al ．Fault diagnosis method of piston pump in construction machinery under variable load condition［J］．Journal of South China University of Technology （Natural Science Edition），2021，49（2）：110-119，139.
2	陈雪峰，郭艳婕，许才彬，等．风电装备故障诊断与健康监测研究综述［J］．中国机械工程，2020，31（2）：175-189.
	CHEN Xuefeng， GUO Yanjie， XU Caibin，et al ．Review of fault diagnosis and health monitoring for wind power equipment［J］．China Mechanical Engineering，2020，31（2）：175-189.
3	刘畅，伍星，刘韬，等．基于近似等距投影和支持向量机的滚动轴承故障诊断［J］．振动与冲击，2018，37（5）：234-239.
	LIU Chang， WU Xing， LIU Tao，et al ．Fault diagnosis of rolling bearings based on near-isometric projection and support vector machine［J］．Journal of Vibration and Shock，2018，37（5）：234-239.
4	王栋璀，丁云飞，朱晨烜．基于中智KNN的齿轮箱故障诊断方法［J］．振动与冲击，2019，38（20）：148-153.
	WANG Dongcui， DING Yunfei， ZHU Chenxuan ．A fault diagnosis method for gearbox based on neutrosophic K-nearest neighbor［J］．Journal of Vibration and Shock，2019，38（20）：148-153.
5	周璇，王晓佩，梁列全，等．基于随机森林算法的制冷剂充注量故障诊断［J］．华南理工大学学报（自然科学版），2020，48（2）：16-24.
	ZHOU Xuan， WANG Xiaopei， LIANG Liequan，et al ．Random forests algorithm-based fault diagnosis for refrigerant charge［J］．Journal of South China University of Technology （Natural Science Edition），2020，48（2）：16-24.
6	陈祝云，钟琪，黄如意，等．基于增强迁移卷积神经网络的机械智能故障诊断［J］．机械工程学报，2021，57（21）：96-105.
	CHEN Zhuyun， ZHONG Qi， HUANG Ruyi，et al ．Intelligent fault diagnosis for machinery based on enhanced transfer convolutional neural network［J］．Journal of Mechanical Engineering，2021，57（21）：96-105.
7	杨平，苏燕辰．基于卷积门控循环网络的滚动轴承故障诊断［J］．航空动力学报，2019，34（11）：2432-2439.
	YANG Ping， SU Yanchen ．Fault diagnosis of rolling bearing based on convolution gated recurrent network［J］．Journal of Aerospace Power，2019，34（11）：2432-2439.
8	侯文擎，叶鸣，李巍华．基于改进堆叠降噪自编码的滚动轴承故障分类［J］．机械工程学报，2018，54（7）：87-96.
	HOU Wenqing， YE Ming， LI Weihua ．Rolling element bearing fault classification using improved stacked de-noising auto-encoders［J］．Journal of Mechanical Engineering，2018，54（7）：87-96.
9	李益兵，黄定洪，马建波，等．基于深度置信网络与信息融合的齿轮故障诊断方法［J］．振动与冲击，2021，40（8）：62-69.
	LI Yibing， HUANG Dinghong， MA Jianbo，et al ．A gear fault diagnosis method based on deep belief network and information fusion［J］．Journal of Vibration and Shock，2021，40（8）：62-69.
10	ZHANG W， PENG G， LI C，et al ．A new deep learning model for fault diagnosis with good anti-noise and domain adaptation ability on raw vibration signals［J］．Sensors，2017，17（2）：425/1-21.
11	WEN L， LI X， GAO L，et al ．A new convolutional neural network-based data-driven fault diagnosis method ［J］．IEEE Transactions on Industrial Electronics，2017，65（7）：5990-5998.
12	ZHAO R， WANG D， YAN R，et al ．Machine health monitoring using local feature-based gated recurrent unit networks［J］．IEEE Transactions on Industrial Electronics，2017，65（2）：1539-1548.
13	de BRUIN T， VERBERT K， BABUSKA R ．Railway track circuit fault diagnosis using recurrent neural networks［J］．IEEE Transactions on Neural Networks and Learning Systems，2016，28（3）：523-533.
14	ZHANG D， STEWART E， ENTEZAMI M，et al ．Intelligent acoustic-based fault diagnosis of roller bearings using a deep graph convolutional network［J］．Measurement，2020，156：107585.
15	YU X， TANG B， ZHANG K. Fault diagnosis of wind turbine gearbox using a novel method of fast deep graph convolutional networks［J］．IEEE Transactions on Instrumentation and Measurement，2021，70：6502714/1-14.
16	TANG Y， ZHANG X， ZHAI Y，et al ．Rotating machine systems fault diagnosis using semisupervised conditional random field-based graph attention network［J］．IEEE Transactions on Instrumentation and Measurement，2021，70：3521010/1-10.
17	LI T， ZHAO Z， SUN C，et al ．Multireceptive field graph convolutional networks for machine fault diagnosis［J］．IEEE Transactions on Industrial Electronics，2020，68（12）：12739-12749.
18	YANG C， ZHOU K， LIU J ．SuperGraph：spatial-temporal graph-based feature extraction for rotating machinery diagnosis［J］．IEEE Transactions on Industrial Electronics，2021，69（4）：4167-4176.
19	LIU X， TIAN H， DAI Z ．Bearing fault diagnosis based on multi-scale convolution neural network and dropout［C］∥Proceedings of 2020 IEEE 4th Information Technology，Networking，Electronic and Automation Control Conference （ITNEC）．Shanghai：IEEE，2020：1401-1406.
20	JIANG G， HE H， YAN J，et al ．Multiscale convolutional neural networks for fault diagnosis of wind turbine gearbox［J］．IEEE Transactions on Industrial Electronics，2018，66（4）：3196-3207.
21	吴静然，丁恩杰，崔冉，等．采用多尺度注意力机制的旋转机械故障诊断方法［J］．西安交通大学学报，2020，54（2）：51-58.
	WU Jingran， DING Enjie， CUI Ran，et al ．A diagnostic approach for rotating machinery using multi-scale feature attention mechanism［J］．Journal of Xi’an Jiaotong University，2020，54（2）：51-58
22	SCARSELLI F， GORI M， TSOI A C，et al ．The graph neural network model［J］．IEEE Transactions on Neural Networks，2008，20（1）： 61-80.
23	KIFI T N， WELLING M ．Semi-supervised classification with graph convolutional networks［J］．arXiv preprint arXiv：，2016.
24	ZHAO Z， LI T， WU J，et al ．Deep learning algorithms for rotating machinery intelligent diagnosis：an open source benchmark study［J］．ISA Transactions，2020，107：224-255.

任务	训练集	训练集数量	测试集	测试集数量
A	Load 0	2 500	Load 2	2 500
B	Load 0	2 500	Load 4	2 500
C	Load 2	2 500	Load 0	2 500
D	Load 2	2 500	Load 4	2 500
E	Load 4	2 500	Load 0	2 500
F	Load 4	2 500	Load 2	2 500

任务	训练集	训练集数量	测试集	测试集数量
G	Speed 480，600	5 000	Speed 720	2 500
H	Speed 480，720	5 000	Speed 600	2 500
I	Speed 600，720	5 000	Speed 480	2 500

诊断方法	平均正确率/%	平均F₁分数
CNN	92.63	0.911
多尺度CNN	96.62	0.946
多尺度CNN+通道注意力	97.42	0.971
多尺度CNN+GCN	97.72	0.976
所提方法	98.85	0.986

诊断方法	平均正确率/%	平均F₁分数
WDCNN	82.29	0.819
LeNet	86.05	0.863
LSTM	88.97	0.885
ResNet	93.77	0.928
MSCNN	91.24	0.912
所提方法	98.85	0.986

诊断方法	平均正确率/%	平均F₁分数
WDCNN	74.05	0.736
LeNet	75.93	0.743
LSTM	87.09	0.869
ResNet	86.72	0.865
MSCNN	87.11	0.875
所提方法	91.29	0.895