精轧机组配电箱的新型跨域开集故障诊断模型

Yonglin Guo; Shihao Li; Tangbin Xia; Di Zhou; Ershun Pan

doi:10.1088/3050-2454/ae4e31

您当前的位置：

首页 >

文章列表页 >

精轧机组配电箱的新型跨域开集故障诊断模型

Papers | 更新时间：2026-04-14

- 精轧机组配电箱的新型跨域开集故障诊断模型
- A novel cross-domain open-set fault diagnosis model for distribution boxes in finishing rolling group
- 可靠性科学与工程学报（英文） 2026年2卷第1期页码：1-17
- 作者机构：
  
  School of Mechanical Engineering, Donghua University, Shanghai 200051, People's Republic of China
  Department of Industrial Engineering, School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai 200240, People's Republic of China
- 作者简介：
  
  [ "Yonglin Guo received the BE degree from the Shandong Jianzhu University, Jinan, China, in 2023. He is currently pursuing his Master's degree at the School of Mechanical Engineering, Donghua University, Shanghai, China. His research interest is fault diagnosis." ]
  [ "Shihao Li is a Master's candidate in the School of Mechanical Engineering at Dong Hua University, Shanghai, China. His research interests focus on reliability systems and failure probability calculations." ]
  [ "Tangbin Xia (Member, IEEE) received the PhD degree in mechanical engineering (industrial engineering) from Shanghai Jiao Tong University, Shanghai, China, in 2014. He is currently a Professor and the Vice Dean of the School of Mechanical Engineering, Shanghai Jiao Tong University. His research interests include intelligent maintenance systems, prognostics and health management, and advanced manufacturing. Dr Xia is a member of IISE, ASME, and INFORMS." ]
  [ "Di Zhou received his PhD degree in Mechanical Engineering and Automation from Northeastern University, China. He is currently an Associate Professor in the School of Mechanical Engineering at Donghua University. His research interests include mechanical reliability analysis and mechanical dynamics." ]
  [ "Ershun Pan received the BS and MS degrees in mechanical design and manufacturing from Northeastern University, Shenyang, China, in 1997, and the PhD degree in mechanical engineering from Shanghai Jiao Tong University, Shanghai, China, in 2000." ]
- 基金信息：
- DOI：10.1088/3050-2454/ae4e31
  中图分类号：
- 收稿：2025-12-01，
  
  修回：2026-02-08，
  
  录用：2026-03-05，
  
  网络首发：2026-03-26，
  
  纸质出版：2026-03
- 稿件说明：
移动端阅览
Yonglin Guo, Shihao Li, Tangbin Xia, et al. 精轧机组配电箱的新型跨域开集故障诊断模型[J]. 可靠性科学与工程学报（英文）, 2026, 2: 015302.

Yonglin Guo, Shihao Li, Tangbin Xia, et al. A novel cross-domain open-set fault diagnosis model for distribution boxes in finishing rolling group[J]. Journal of Reliability Science and Engineering, 2026, 2: 015302.
Yonglin Guo, Shihao Li, Tangbin Xia, et al. 精轧机组配电箱的新型跨域开集故障诊断模型[J]. 可靠性科学与工程学报（英文）, 2026, 2: 015302. DOI： 10.1088/3050-2454/ae4e31.

Yonglin Guo, Shihao Li, Tangbin Xia, et al. A novel cross-domain open-set fault diagnosis model for distribution boxes in finishing rolling group[J]. Journal of Reliability Science and Engineering, 2026, 2: 015302. DOI： 10.1088/3050-2454/ae4e31.

摘要

作为精轧机传动系统的关键环节，配电箱在轧制力传递与分配中起着至关重要的作用。因此，开展配电箱故障诊断研究对于保障工艺运行的安全性与可靠性具有重要意义。然而，当故障特征不显著且故障类型未知时，传统诊断方法可能无法提供可靠的诊断结果。针对这些问题，本文提出一种跨域开集模型，利用对抗学习来增强在模糊故障特征下的特征表示，并融合基于分布的方法用于未知故障识别。考虑到特征表示对于对抗学习和未知故障识别的重要性，本文构建了一种融合深度可分离卷积与图卷积网络的信息交互特征提取器，该结构在提取深度特征的同时，弥补了传统深度可分离卷积在通道独立性方面的局限。为了识别配电箱的未知故障，采用鞍点逼近方法，在训练阶段为每个已知类别建立分布，并在测试阶段利用这些分布识别未知样本。与传统的分布建立方法不同，鞍点逼近方法直接使用数据，无需依赖预设的分布假设或公式。为验证所提方法的优越性，基于配电箱历史故障数据开展了对比实验。结果表明，该方法准确率达94.13%，H分数达91.35%，在已知故障分类与未知故障识别方面均表现出优异性能。

Abstract

As a critical link in the drive system of the finishing mill

distribution boxes play a vital role in rolling force transmission and distribution. Therefore

researching fault diagnosis for distribution boxes is essential for ensuring the security and reliability of process operation. However

when fault characteristics are indistinct and the fault type is unknown

traditional diagnostic methods may fail to provide reliable diagnosis. To address these challenges

this paper proposes a cross-domain open-set model that leverages adversarial learning to enhance feature representations under indistinct fault characteristics

and integrates distribution-based methods for identifying unknown faults. Considering the importance of feature representation for adversarial learning and unknown fault identification

we build an information-exchange feature extractor integrates depthwise separable convolution and GCN. This structure extracts depth features and addresses the channel independence limitation of traditional depthwise separable convolution. To identify unknown faults in distribution boxes

the saddle-point approximation (SPA) method is employed. This method is used during training to establish a distribution for each known class

and these distributions are utilized in testing to identify unknown samples. Unlike traditional methods for establishing distributions

the SPA method directly uses data without relying on predefined distribution ass

umptions or formulas. To verify the superiority of the proposed method

we conducted a comparative experiment based on historical fault data of distribution boxes. The results demonstrate that the proposed method achieves an accuracy of 94.13% and an

-score of 91.35%

which demonstrates the superior performance in classifying known faults and identifying unknown faults in distribution boxes.

关键词

Keywords

references

Zhao S , Bao L , Hou C , Bai Y and Yu Y 2024 Multi-source domain adversarial graph convolutional networks for rolling mill health states diagnosis under variable working conditions Struct. Health Monit. 23 3505 – 24

Yildiz S K , Forbes J F , Huang B , Zhang Y , Wang F , Vaculik V and Dudzic M 2009 Dynamic modelling and simulation of a hot strip finishing mill Appl. Math. Modelling. 33 3208 – 25

Zhou X and Ben X 2023 Maintenance modelling for work rolls in hot finishing mill group with constraint of thermal character Int. J. Prod. Res. 62 1846 – 61

He H , Shao J , Wang X , Yang Q , Liu Y , Xu D and Sun Y-Z 2021 Research and application of approximate rectangular section control technology in hot strip mills J. Iron Steel Res. Int. 28 279 – 90

Zhang W , He J , Ma C , Gao W and Li G 2025 Compound fault diagnosis method of rotating machinery using multi-view multi-label feature selection based on label compression and local label correlation Adv. Eng. Inf. 65 103310

Li W , Liu D , Li Y , Hou M , Liu J , Zhao Z , Guo A , Zhao H and Deng W 2025 Fault diagnosis using variational autoencoder GAN and focal loss CNN under unbalanced data Struct. Health Monit. 24 1859 – 72

Zhao C , Sun J , Lin S and Peng Y 2021 Fault diagnosis method for rolling mill multi row bearings based on AMVMD-MC1DCNN under unbalanced dataset Sensor 21 5494

Yu Y , Shi P , Tian J , Xu X and Hua C 2023 Rolling mill health states diagnosing method based on multi−sensor information fusion and improved DBNs under limited datasets ISA Trans. 134 529 – 47

Shi P , Gao H , Yu Y , Xu X and Han D 2022 Intelligent fault diagnosis of rolling mills based on dual attention−guided deep learning method under imbalanced data conditions Measurement 204 111993

Guo Y , Zhou D , Chen H , Yue X and Cheng Y 2024 Fault intelligent diagnosis for distribution box in hot rolling based on depthwise separable convolution and Bi−LSTM Processes 12 1999

Peng R , Zhang X and Shi P 2022 Bearing fault diagnosis of hot-rolling mill utilizing intelligent optimized self-adaptive deep belief network with limited samples Sensors 22 7815

Li X , Jin W , Xu X and Yang H 2022 A domain-adversarial multi-graph convolutional network for unsupervised domain adaptation rolling bearing fault diagnosis Symmetry 14 2654

Hou D , Shi P , Chen J and Shi P 2024 Improved GNN based on graph−transformer: a new framework for rolling mill bearing fault diagnosis Trans. Inst. Meas. Control. 46 2804 – 15

Wang C , Xin C and Xu Z 2021 A novel deep metric learning model for imbalanced fault diagnosis and toward open-set classification Knowl.-Based Syst. 220 106925

Wang L , Gao Y , Li X and Gao L 2024 Self-supervised-enabled open-set cross-domain fault diagnosis method for rotating machinery IEEE Trans. Ind. Inform. 20 10314 – 24

Lundgren A and Jung D 2022 Data-driven fault diagnosis analysis and open-set classification of time-series data Control Eng. Pract. 121 105006

Razavi-Far R , Hallaji E , Saif M and Ditzler G 2019 A novelty detector and extreme verification latency model for nonstationary environments IEEE Trans. Ind. Electron. 66 561 – 70

Tian Y , Wang Z , Zhang L , Lu C and Ma J 2018 A subspace learning-based feature fusion and open-set fault diagnosis approach for machinery components Adv. Eng. Inf. 36 194 – 206

Peng P , Lu J , Xie T , Tao S , Wang H and Zhang H 2023 Open-set fault diagnosis via supervised contrastive learning with negative out-of-distribution data augmentation IEEE Trans. Ind. Inform. 19 2463 – 73

Jung D 2020 Data-driven open-set fault classification of residual data using Bayesian filtering IEEE Trans. Control. Syst. Technol. 28 2045 – 52

Zhang X , Wang J , Han B , Zhang Z , Yan Z , Jia M and Guo L 2022 Feature distance-based deep prototype network for few-shot fault diagnosis under open-set domain adaptation scenario Measurement 201 111522

Cui Z , Lu Y , Yan X and Cui S 2024 Compound fault diagnosis of diesel engines by combining generative adversarial networks and transfer learning Expert Syst. Appl. 251 123969

Wang H , Xu Z , Tong X and Song L 2023 Cross-domain open set fault diagnosis based on weighted domain adaptation with double classifiers Sensors 23 2137

Zhu J , Huang C , Shen C and Shen Y 2022 Cross-domain open-set machinery fault diagnosis based on adversarial network with multiple auxiliary classifiers IEEE Trans. Ind. Inform. 18 8077 – 86

Zhao C and Shen W 2022 Adaptive open set domain generalization network: learning to diagnose unknown faults under unknown working conditions Reliab. Eng. Syst. Saf. 226 108672

Chen Z , Liao Y , Li J , Huang R , Xu L , Jin G and Li W 2023 A multi-source weighted deep transfer network for open-set fault diagnosis of rotary machinery IEEE Trans. Cybern. 53 1982 – 93

Zhang Z , Nie G , Shao M , Li L , Zhou J and Shao S 2023 Multi-sample-distances-fusion- and generalized-Pareto-distribution-based open-set fault diagnosis of rolling bearing Nonlinear Dyn. 111 11407 – 28

Wang X , Zhang H , Qiao X , Ma K , Tao S et al 2024 Generalized out-of-distribution fault diagnosis (GOOFD) via internal contrastive learning IEEE Trans. Ind. Inform. 20 9987 - 9996

Yu X et al 2022 Deep-learning-based open set fault diagnosis by extreme value theory IEEE Trans. Ind. Inform. 18 185 – 96

Zhang B , Zhou C , Li W , Ji S , Li H , Tong Z and Ng S-K 2022 Intelligent bearing fault diagnosis based on open set convolutional neural network Mathematics 10 3953

Gao Y , Ma A J , Gao Y , Wang J and Pan Y 2020 Adversarial open set domain adaptation via progressive selection of transferable target samples Neurocomputing 410 174 – 84

Fan Y , Guo L , Sun Y , Luo T , Hou K et al 2025 Wasserstein bi-classifier adversarial learning network for machinery fault diagnostics Struct. Health Monit. 24 3363 – 81

Verstraete D B , Droguett E L , Meruane V , Modarres M and Ferrada A 2019 Deep semi-supervised generative adversarial fault diagnostics of rolling element bearings Struct. Health Monit. 19 390 – 411

Mao W , Liu Y , Ding L , Safian A and Liang X 2021 A new structured domain adversarial neural network for transfer fault diagnosis of rolling bearings under different working conditions IEEE Trans. Instrum. Meas. 70 1 - 13

Zhou D , Pan E , Zhang X and Zhang Y 2020 Dynamic model-based saddle-point approximation for reliability and reliability-based sensitivity analysis Reliab. Eng. Syst. Saf. 201 106972

Huang B and Du X 2006 Uncertainty analysis by dimension reduction integration and saddlepoint approximations J. Mech. Des. 128 26

He J , Huang W , Liu Y , Qian C , Ma C , Gao W and Jin X 2025 Data imbalance fault diagnosis method based on an ensemble multi-scale convolutional attention network Mech. Syst. Signal Process. 236 112934

Zhou D , Zhang X and Zhang Y 2016 Dynamic reliability analysis for planetary gear system in shearer mechanisms Mech. Mach. Theory 105 244 – 59

Wang C , Li X , Yuan P , Su K , Xie Z and Wang J 2025 An integrated approach for mechanical fault diagnosis using maximum mean square discrepancy representation and CNN-based mixed information fusion Struct. Health Monit. 24 3674 – 87

Liu T , Zhang Y , Huang G , Liu C and Jiang D 2025 Fault identification of aero-engine central rod-fastened rotor system based on multi-source information fusion and graph convolutional network Struct. Health Monit. 0

Xiang Q , Xiang Q , Lai J , Lei L , Song Y , He J and Li R 2024 Quadruplet depth-wise separable fusion convolution neural network for ballistic target recognition with limited samples Expert Syst. Appl. 235 121182

Bendale A and Boult T E 2016 Towards open set deep networks Proc. IEEE Conf. on Computer Vision and Pattern Recognition pp 1563 – 72

Qin Y , Zhang T , Qian Q and Mao Y 2024 Large model for rotating machine fault diagnosis based on a dense connection network with depthwise separable convolution IEEE Trans. Instrum. Meas. 73 3520512

Ganin Y , Ustinova U , Ajakan H , Germain P , Larochelle H , Laviolette F , Marchand M and Lempitsky V 2016 Domain-adversarial training of neural networks ( arXiv: 1505.07818 )

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

A neuro-fuzzy approach with hypergraph convolution for fault diagnosis in industrial devices