Page 114 - 《振动工程学报》2026年第5期

1318 振 动 工 程 学 报 第 39 卷

speed conditions[J]. Mechanical Systems and Signal Process- ing based on WHVG and GCN[J]. IEEE Transactions on
ing，2023，189：110109. Instrumentation and Measurement，2021，70：3519811.
[9] HONG D， KIM B. 1D convolutional neural network-based [21] YE L X，KEOGH E J. Time series shapelets：a new primi-
adaptive algorithm structure with system fault diagnosis and tive for data mining[C]//Proceedings of the 15th ACM
signal feature extraction for noise and vibration enhancement SIGKDD International Conference on Knowledge Discovery
in mechanical systems[J]. Mechanical Systems and Signal and Data Mining. ACM，2009：947-956.
Processing，2023，197：110395. [22] CHENG Z Q，YANG Y，JIANG S，et al. Time2Graph+：
[10] LI S W，YANG Y C. A recurrent neural network framework bridging time series and graph representation learning via
with an adaptive training strategy for long-time predictive multiple attentions[J]. IEEE Transactions on Knowledge and
modeling of nonlinear dynamical systems[J]. Journal of Sound Data Engineering，2023，35（2）：2078-2090.
and Vibration，2021，506：116167. [23] HU W J，YANG Y，CHENG Z Q，et al. Time-series event
[11] DONG G S，WAN H P，LUO Y Z，et al. A fast sparsity- prediction with evolutionary state graph[C]//Proceedings of the
free compressive sensing approach for vibration data recon-
14th ACM International Conference on Web Search and Data
struction using deep convolutional GAN[J]. Mechanical
Mining. ACM，2021：580-588.
Systems and Signal Processing，2023，188：109937.
[24] 王达，荣冈. 时间序列的模式距离 [J]. 浙江大学学报 (工学
[12] QIAO L K，ZHANG Y X，WANG Q S. Fault detection in
版)，2004，38（7）：795-798.
wind turbine generators using a meta-learning-based convolu-
WANG Da，RONG Gang. Pattern distance of time series[J].
tional neural network[J]. Mechanical Systems and Signal
Journal of Zhejiang University (Engineering Science)，
Processing，2023，200：110528.
2004，38（7）：795-798.
[13] KIM Y，NA K，YOUN B D. A health-adaptive time-scale
[25] ZHANG D，GAO J Y，LI X L. Multivariate time series clas-
representation （HTSR） embedded convolutional neural network
sification with crucial timestamps guidance[J]. Expert Systems
for gearbox fault diagnostics[J]. Mechanical Systems and
with Applications，2024，255：124591.
Signal Processing，2022，167：108575.
[26] VELIČKOVIČ P，CUCURULL G，CASANOVA A，et al.
[14] SHI J C，PENG D K，PENG Z X，et al. Planetary gearbox
Graph attention networks[EB/OL]. (2018-02-04)[2024-03-20].
fault diagnosis using bidirectional-convolutional LSTM
https://arXiv.org/abs/1710.10903v3.
networks[J]. Mechanical Systems and Signal Processing，
[27] GAO H Y，JI S W. Graph U-Nets[J]. IEEE Transactions on
2022，162：107996.
Pattern Analysis and Machine Intelligence， 2022， 44（ 9） ：
[15] QIU C C，LI K，LI B，et al. Semi-supervised graph convo-
4948-4960.
lutional network to predict position- and speed-dependent tool
[28] CAI B R，HUANG G Y，YANG S Q，et al. SE-shapelets：
tip dynamics with limited labeled data[J]. Mechanical Systems
semi-supervised clustering of time series using representative
and Signal Processing，2022，164：108225.
shapelets[J]. Expert Systems with Applications，2024，240：
[16] LI C Y，MO L F，YAN R Q. Rolling bearing fault diagnosis
122584.
based on horizontal visibility graph and graph neural
[29] JI C，HU Y P，LIU S J，et al. Fully convolutional networks
networks[C]//Proceedings of International Conference on
with shapelet features for time series classification[J]. Informa-
Sensing，Measurement & Data Analytics in the Era of Artifi-
tion Sciences，2022，612：835-847.
cial Intelligence (ICSMD). IEEE，2020：275-279.
[30] LIANG Z Y，WANG H Z. Efficient class-specific shapelets
[17] YIN P Z，NIE J，LIANG X Y，et al. A multiscale graph
learning for interpretable time series classification[J]. Informa-
convolutional neural network framework for fault diagnosis of
tion Sciences：An International Journal，2021，570：428-
rolling bearing[J]. IEEE Transactions on Instrumentation and
450.
Measurement，2023，72：2520713.
[18] GAO Y， WU H W， LIAO H Q， et al. A fault diagnosis [31] DEFFERRARD M, BRESSON X, VANDERGHEYNST P，
et al. Convolutional neural networks on graphs with fast local-
method for rolling bearings based on graph neural network
with one-shot learning[J]. EURASIP Journal on Advances in ized spectral filtering[C]//Proceedings of the 30th Interna-
Signal Processing，2023，2023(1)：101. tional Conference on Neural Information Processing Systems.
[19] YU Z D，ZHANG C H，DENG C. An improved GNN using ACM，2016：3844-3852.
dynamic graph embedding mechanism： a novel end-to-end
framework for rolling bearing fault diagnosis under variable 第一作者：郭海宇（1986—），女，博士，副教授。
working conditions[J]. Mechanical Systems and Signal E-mail：ghy@sut.edu.cn
Processing，2023，200：110534. 通信作者：张晓光（1987—），男，博士，主任工程师。
[20] LI C Y，MO L F，YAN R Q. Fault diagnosis of rolling bear- E-mail：shuming6245@163.com