Metric learning-based two-stage imbalanced fault diagnosis model for water injection pump

Zhiwen Yao; Xiwen Cui; Nagi Gebraeel; Tangbin Xia; Lifeng Xi

doi:10.1088/3050-2454/ae227a

您当前的位置：

首页 >

文章列表页 >

Metric learning-based two-stage imbalanced fault diagnosis model for water injection pump

Papers | 更新时间：2026-02-13

- Metric learning-based two-stage imbalanced fault diagnosis model for water injection pump
- 基于度量学习的两阶段不平衡水注泵故障诊断模型
- Journal of Reliability Science and Engineering Issue 4, Pages: 1-15(2025)
- 作者机构：
  
  State Key Laboratory of Mechanical System and Vibration, School of Mechanical Engineering, Shanghai Jiao Tong University, SJTU-Fraunhofer Center, Shanghai 200240, People’s Republic of China
  H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology, 765 Ferst Drive, Atlanta, GA 30332, United States of America
  Special Environment Digital Manufacturing Equipment Technology Invitation Center, Mianyang, Sichuan 621900, People’s Republic of China
- 作者简介：
  
  [ "MrZhiwen Yao Zhiwen Yao received the BS degree in Mechanical Engineering from Shanghai Jiao Tong University, Shanghai, China, in 2024, where he is currently pursuing a Master’s degree in Industrial Engineering and Management with the School of Mechanical Engineering." ]
  [ "His research interests include fault diagnosis and prognostics and health management of industrial equipment using deep learning." ]
  [ "Her research interests focus on applying deep learning methods for fault detection, prognosis, and health management of industrial machinery." ]
  [ "ProfTangbin Xia Tangbin Xia received a PhD degree in Mechanical Engineering (Industrial Engineering) from Shanghai Jiao Tong University in 2014 and was a co-cultured PhD with the University of Michigan. Currently, he is a Professor and serves as Vice Dean of the School of Mechanical Engineering at Shanghai Jiao Tong University, State Key Laboratory of Mechanical System and Vibration. His major research interests are intelligent maintenance systems, prognostics & health management, and advanced manufacturing. He is a member of IEEE, ASME, IISE and INFORMS." ]
  [ "ProfLifeng Xi Lifeng Xi received a PhD degree in Mechanical Engineering from Shanghai Jiao Tong University in 1995. He is currently a Professor and serves as the Executive Vice President of Shanghai Jiao Tong University. He also serves as the Fellow of State Key Laboratory of Mechanical System and Vibration, and the Executive Director of the Chinese Quality Association. He has published more than 200 papers in international journals. His interest areas are in the fields of production system design and planning, quality management, and reliability engineering." ]
- 基金信息：
- DOI：10.1088/3050-2454/ae227a
  CLC：
- Received：25 June 2025，
  
  Revised：2025-10-26，
  
  Accepted：20 November 2025，
  
  Online First：03 December 2025，
  
  Published：2025
- 稿件说明：
移动端阅览
Zhiwen Yao, Xiwen Cui, Nagi Gebraeel, et al. Metric learning-based two-stage imbalanced fault diagnosis model for water injection pump[J]. Journal of Reliability Science and Engineering, 2025, (4): 1-15.
DOI：

Zhiwen Yao, Xiwen Cui, Nagi Gebraeel, et al. Metric learning-based two-stage imbalanced fault diagnosis model for water injection pump[J]. Journal of Reliability Science and Engineering, 2025, (4): 1-15. DOI： 10.1088/3050-2454/ae227a.

摘要

针对柱塞式注水泵故障诊断中数据严重不平衡及特征提取困难的问题，本文提出了一种融合度量学习与采样自适应机制的故障诊断算法。首先，构建了一种基于拉普拉斯小波卷积与多尺度注意力机制的特征提取器，用于捕获关键故障特征并自适应地分配特征权重。其次，引入由双中心损失驱动的度量学习模块，在高维嵌入空间中同时施加类间与类内距离约束，以促进同类故障样本的聚类，从而提升故障识别能力。最后，采用类别均衡的批次采样机制对模型进行自适应强化，克服数据不平衡带来的性能限制，并进一步提高整体诊断精度。在真实的不平衡注水泵数据集上的验证结果表明，所提出的方法实现了 97.11% 的诊断准确率。

Abstract

To address the severe data imbalance and the difficulty of feature extraction in fault diagnosis of plunger-type injection pumps

we propose a fault diagnosis algorithm that integrates metric learning with a sampling-adaptation mechanism. First

we design a feature extractor based on Laplacian wavelet convolution and a multi-scale attention mechanism to capture critical fault features and adaptively assign feature weights. Second

we introduce a metric-learning module driven by dual-center loss

imposing inter-class and intra-class distance constraints in a high-dimensional embedding space to promote clustering of samples from the same fault class and thus enhance fault recognition capability. Finally

we adopt a class-balanced batch sampling mechanism to adaptively strengthen the model

overcoming the performance limitations imposed by data imbalance and further improving overall accuracy. Validated on a real

imbalanced injection-pump dataset

the proposed method achieves a diagnostic accuracy of 97.11%.

关键词

Keywords

references

Zhou X , Van Gelder P H A J M , Liang Y and Zhang H 2020 An integrated methodology for the supply reliability analysis of multi-product pipeline systems under pumps failure Reliab. Eng. Syst. Saf. 204 107185

Huang H et al 2025 Theory and application of possibility and evidence in reliability analysis and design optimization J. Reliab. Sci. Eng. 1 15007

Zio E 2022 Prognostics and health management (PHM): where are we and where do we (need to) go in theory and practice Reliab. Eng. Syst. Saf. 218 108119

Jin T and Abedin A 2025 Co-allocation of reliability-redundancy, spare parts and spacecraft trajectory for lunar habitation J. Reliab. Sci. Eng. 1 15005

Xia T , Dong Y , Xiao L , Du S , Pan E and Xi L 2018 Recent advances in prognostics and health management for advanced manufacturing paradigms Reliab. Eng. Syst. Saf. 178 255 – 68

Xia T , Xing X , Yan T , Wang D , Pan E and Xi L 2024 Interpretable temporal degradation state chain based fusion graph for intelligent bearing fault detection Adv. Eng. Inform. 59 102342

Wang J , Zhang Z , Liu Z , Han B , Bao H and Ji S 2023 Digital twin aided adversarial transfer learning method for domain adaptation fault diagnosis Reliab. Eng. Syst. Saf. 234 109152

Tan Q , Mu X , Fu M , Yuan H , Sun J , Liang G and Sun L 2022 A new sensor fault diagnosis method for gas leakage monitoring based on the naive Bayes and probabilistic neural network classifier Measurement 194 111037

Tang S , Zhu Y and Yuan S 2022 A novel adaptive convolutional neural network for fault diagnosis of hydraulic piston pump with acoustic images Adv. Eng. Inform. 52 101554

Wang S and Xiang J 2020 A minimum entropy deconvolution-enhanced convolutional neural networks for fault diagnosis of axial piston pumps Soft Comput. 24 2983 – 97

Li G , Hu J , Shan D , Ao J , Huang B and Huang Z 2023 A CNN model based on innovative expansion operation improving the fault diagnosis accuracy of drilling pump fluid end Mech. Syst. Signal Process. 187 109974

Tang S , Zhu Y and Yuan S 2022 Intelligent fault diagnosis of hydraulic piston pump based on deep learning and Bayesian optimization ISA Trans. 129 555 – 63

Mei J , Zhu M , Liu W , Fu M and Tang Q 2023 Conditional variational encoder classifier for open set fault classification of rotating machinery vibration signals IEEE Trans. Ind. Inform. 20 3038 – 49

Haixiang G , Yijing L , Shang J , Mingyun G , Yuanyue H and Bing G 2017 Learning from class-imbalanced data: review of methods and applications Expert Syst. Appl. 73 220 – 39

Mao W , Liu Y , Ding L and Li Y 2019 Imbalanced fault diagnosis of rolling bearing based on generative adversarial network: a comparative study IEEE Access 7 9515 – 30

Wang J , Wei J , Huang H , Wen L , Yuan Y , Chen H , Wu R and Wu J 2024 IMWMOTE: a novel oversampling technique for fault diagnosis in heterogeneous imbalanced data Expert Syst. Appl. 251 123987

Wang H , Li P , Lang X , Tao D , Ma J and Li X 2023 FTGAN: a novel GAN-based data augmentation method coupled time–frequency domain for imbalanced bearing fault diagnosis IEEE Trans. Instrum. Meas. 72 1 – 14

Zhou Y , Gao L , Wang D , Wu W , Zhou Z and Ye T 2023 Imbalanced multi-fault diagnosis via improved localized feature selection IEEE Trans. Instrum. Meas. 72 1 – 11

Jian C and Ao Y 2023 Imbalanced fault diagnosis based on semi-supervised ensemble learning J. Intell. Manuf. 34 3143 – 58

Zhao Y , Yang X , Huang J , Gao J and Cui J 2024 Improved weighted extreme learning machine with adaptive cost-sensitive strategy for imbalanced fault diagnosis of rotating machinery Mech. Syst. Signal Process. 217 111526

Mao G , Li Y , Cai Z , Qiao B and Jia S 2023 Transferable dynamic enhanced cost-sensitive network for cross-domain intelligent diagnosis of rotating machinery under imbalanced datasets Eng. Appl. Artif. Intell. 125 106670

Gao W , Xu Z and Akoudad Y 2025 Metric-based meta-learning relation network for cross-domain few-shot bearing fault diagnosis IEEE Sens. J. 25 13632 – 47

Li T , Zhao Z , Sun C , Cheng L , Chen X , Yan R and Gao R X 2021 WaveletKernelNet: an interpretable deep neural network for industrial intelligent diagnosis IEEE Trans. Syst. Man Cybern. Syst. 52 2302 – 12

Jiang G , Wang J , Wang L , Xie P , Li Y and Li X 2023 An interpretable convolutional neural network with multi-wavelet kernel fusion for intelligent fault diagnosis J. Manuf. Syst. 70 18 – 30

Chen J , Huang R , Zhao K , Wang W , Liu L and Li W 2021 Multiscale convolutional neural network with feature alignment for bearing fault diagnosis IEEE Trans. Instrum. Meas. 70 1 – 10

Xu Y , Chen Y , Zhang H , Feng K , Wang Y , Yang C and Ni Q 2023 Global contextual feature aggregation networks with multiscale attention mechanism for mechanical fault diagnosis under non-stationary conditions Mech. Syst. Signal Process. 203 110724

Li F , Wang L , Wang D , Wu J and Zhao H 2023 An adaptive multiscale fully convolutional network for bearing fault diagnosis under noisy environments Measurement 216 112993

Woo S , Park J , Lee J Y and Kweon I S 2018 CBAM: convolutional block attention module Proc. European Conf. on Computer Vision (ECCV) ( Munich ) ( Springer ) pp 3 – 19

Zhang Y , Liu Z and Huang Q 2023 A contrastive learning-based fault diagnosis method for rotating machinery with limited and imbalanced labels IEEE Sens. J. 23 16402 – 12

Zhang K , Chen J , He S , Li F , Feng Y and Zhou Z 2022 Triplet metric driven multi-head GNN augmented with decoupling adversarial learning for intelligent fault diagnosis of machines under varying working condition J. Manuf. Syst. 62 1 – 16

Elharrouss O , Mahmood Y , Bechqito Y , Adel Serhani M , Badidi E , Riffi J and Tairi H 2025 Loss functions in deep learning: a comprehensive review ( arXiv:2504.04242 )

Qian Z et al 2025 Enhanced diagnosis of thyroid-associated eye diseases based on deep learning: a novel triplet loss design strategy Biomed. Signal Process. Control 100 107161

Liu Z , Fu Y , Pan Q and Zhang Z W 2022 Orientational distribution learning with hierarchical spatial attention for open set recognition IEEE Trans. Pattern Anal. Mach. Intell. 45 8757 – 72

Zhong H , Yu S , Trinh H , Lv Y , Yuan R and Wang Y 2023 Fine-tuning transfer learning based on DCGAN integrated with self-attention and spectral normalization for bearing fault diagnosis Measurement 210 112421

Wang Y , Lei Y , Li N , Li X , Yang B , Gao X and Liu X 2025 Multimodal parallel bilinear co-attention fusion with full-dimensional dynamic wavelet-kernel structure for machinery RUL prediction Measurement 247 116685

Li W , Lan H , Chen J , Feng K and Huang R 2023 WavCapsNet: an interpretable intelligent compound fault diagnosis method by backward tracking IEEE Trans. Instrum. Meas. 72 1 – 11

Witmer A , Theagarajan R and Bhanu B 2023 Triplet-net classification of contiguous stem cell microscopy images IEEE/ACM Trans. Comput. Biol. Bioinform. 20 2314 – 27

Liu Z , Zhang J and He X 2023 A discrimination-guided active learning method based on marginal representations for industrial compound fault diagnosis IEEE Trans. Autom. Sci. Eng. 21 6411 – 22

Liu Z , Zhang J , He X , Zhang Q , Sun G and Zhou D 2022 Fault diagnosis of rotating machinery with limited expert interaction: a multicriteria active learning approach based on broad learning system IEEE Trans. Control Syst. Technol. 31 953 – 60

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

No data

Related Author

No data

Related Institution

No data

⁰