An incremental intelligent fault diagnosis method for marine diesel engines based on CNN-Transformer and cosine similarity

  • Yingying Wu

    Marine Engineering Institute, Jimei University, Xiamen 361021, China
  • Yongjian Wang

    Marine Engineering Institute, Jimei University, Xiamen 361021, China
  • Hangxi Cai

    Marine Engineering Institute, Jimei University, Xiamen 361021, China
  • Guoqiang Li

    Marine Engineering Institute, Jimei University, Xiamen 361021, China
  • Xin Wei

    Marine Engineering Institute, Jimei University, Xiamen 361021, China
Article ID: 3617
DOI: https://doi.org/10.59400/sv3617
Keywords: marine diesel engine; fault diagnosis; incremental fault diagnosis; vibration signals; CNN-Transformer

Abstract

This paper proposes an incremental intelligent fault diagnosis method for marine diesel engines based on a Convolutional Neural Network (CNN)-Transformer architecture and cosine similarity. The method is designed to address critical limitations of conventional supervised diagnostic frameworks, including heavy reliance on labeled data, weak cross-condition generalization, and the inability to identify new or evolving fault types. The model first employs CNN to extract local temporal features from vibration signals and then uses a Transformer to learn high-level semantic representations of fault attributes. During the incremental learning phase, known fault classes—such as exhaust valve failures—are used to train the model. In the testing phase, the model calculates the cosine similarity between feature embeddings of unseen samples and the prototypes of known classes in the attribute space to determine their classification or novelty. This mechanism enables effective identification of both known and novel faults, including those in cylinder liners and piston rings, without requiring prior labeled data for the latter. Experimental results demonstrate that the proposed approach achieves superior classification accuracy, robustness, and adaptability compared to traditional supervised methods, offering a scalable and generalizable solution for intelligent marine diesel engine fault diagnostics.

Published
2025-12-01
How to Cite
Wu, Y., Wang, Y., Cai, H., Li, G., & Wei, X. (2025). An incremental intelligent fault diagnosis method for marine diesel engines based on CNN-Transformer and cosine similarity. Sound & Vibration, 59(6). https://doi.org/10.59400/sv3617
Issue
Vol. 59 No. 6 (2025)
Section
Article

Copyright (c) 2026 Yingying Wu, Yongjian Wang, Hangxi Cai, Guoqiang Li, Xin Wei

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.

References

[1]Wang, L. Research on Health State Assessment and Fault Diagnosis of Marine Diesel Engine Based on Performance Parameters [Master's Thesis]. Harbin Engineering University; 2023. (in Chinese)

[2]Wang C, Lu N, Cheng Y, et al. A Data-Driven Aero-Engine Degradation Prognostic Strategy. IEEE Transactions on Cybernetics. 2021; 51(3): 1531–1541. doi: 10.1109/TCYB.2019.2938244

[3]Wang X, Jiang B, Ding SX, et al. Extended Relevance Vector Machine-Based Remaining Useful Life Prediction for DC-Link Capacitor in High-Speed Train. IEEE Transactions on Cybernetics. 2022; 52(9): 9746–9755. doi: 10.1109/TCYB.2020.3035796

[4]Chen C, Shi J, Shen M, et al. Pseudo-Label Guided Sparse Deep Belief Network Learning Method for Fault Diagnosis of Radar Critical Components. IEEE Transactions on Instrumentation and Measurement. 2023; 72: 1–12. doi: 10.1109/TIM.2023.3256474

[5]Bai H, Zhan X, Yan H, et al. Research on Diesel Engine Fault Diagnosis Method Based on Stacked Sparse Autoencoder and Support Vector Machine. Electronics. 2022; 11(14): 2249. doi: 10.3390/electronics11142249

[6]Wang S, Lei Y, Lu N, et al. Graph Continual Learning Network: An Incremental Intelligent Diagnosis Method of Machines for New Fault Detection. IEEE Transactions on Automation Science and Engineering. 2024; 1–11. doi: 10.1109/TASE.2024.3417208

[7]Li X, Zhang W, Li X, et al. Partial Domain Adaptation in Remaining Useful Life Prediction With Incomplete Target Data. IEEE/ASME Transactions on Mechatronics. 2024; 29(3): 1903–1913. doi: 10.1109/TMECH.2023.3325538

[8]Lee D, Lee JG, Choi M, et al. Multi-fidelity sub-label-guided transfer network with physically interpretable synthetic datasets for rotor fault diagnosis. Engineering Applications of Artificial Intelligence. 2025; 148: 110467. doi: 10.1016/j.engappai.2025.110467

[9]Lei Y, Li N, Li X. Big Data-Driven Intelligent Fault Diagnosis and Prognosis for Mechanical Systems. Springer; 2023. doi: 10.1007/978-981-16-9131-7

[10]Chen X, Li X, Yu S, et al. Dynamic Vision Enabled Contactless Cross-Domain Machine Fault Diagnosis with Neuromorphic Computing. IEEE/CAA Journal of Automatica Sinica. 2024; 11(3): 788–790. doi: 10.1109/JAS.2023.124107

[11]Dong H, Tian Z, Spencer JW, et al. Bilevel Optimization of Sizing and Control Strategy of Hybrid Energy Storage System in Urban Rail Transit Considering Substation Operation Stability. IEEE Transactions on Transportation Electrification. 2024; 10(4): 10102–10114. doi: 10.1109/TTE.2024.3385821

[12]Yang B, Xu S, Lei Y, et al. Multi-source transfer learning network to complement knowledge for intelligent diagnosis of machines with unseen faults. Mechanical Systems and Signal Processing. 2022; 162: 108095. doi: 10.1016/j.ymssp.2021.108095

[13]Chen J, Wang G, Lv J, et al. Open-Set Classification for Signal Diagnosis of Machinery Sensor in Industrial Environment. IEEE Transactions on Industrial Informatics. 2023; 19(3): 2574–2584. doi: 10.1109/TII.2022.3169459

[14]Meire M, Van Baelen Q, Ooijevaar T, et al. Constraint Guided Autoencoders to Enforce a Predefined Threshold on Anomaly Scores: An Application in Machine Condition Monitoring. Journal of Dynamics, Monitoring and Diagnostics. 2023; 2(2). doi: 10.37965/jdmd.2023.234

[15]Zhong Z, Mi J, Zhao Y, et al. Coordinated Control of the Onboard and Wayside Energy Storage System of an Urban Rail Train Based on Rule Mining. Urban Rail Transit. 2024; 10(3): 232–247. doi: 10.1007/s40864-024-00223-7

[16]Zhang X, Wang B, Chen X. Intelligent fault diagnosis of roller bearings with multivariable ensemble-based incremental support vector machine. Knowledge-Based Systems. 2015; 89: 56–85. doi: 10.1016/j.knosys.2015.06.017

[17]Shi M, Ding C, Chang S, et al. Cross-Domain Class Incremental Broad Network for Continuous Diagnosis of Rotating Machinery Faults Under Variable Operating Conditions. IEEE Transactions on Industrial Informatics. 2024; 20(4): 6356–6368. doi: 10.1109/TII.2023.3345449

[18]Fu Y, Cao H, Chen X, et al. Broad auto-encoder for machinery intelligent fault diagnosis with incremental fault samples and fault modes. Mechanical Systems and Signal Processing. 2022; 178: 109353. doi: 10.1016/j.ymssp.2022.109353

[19]Li M, Zhu Z. Spatial-Temporal Fusion Graph Neural Networks for Traffic Flow Forecasting. Proceedings of the AAAI Conference on Artificial Intelligence. 2021; 35(5): 4189–4196. doi: 10.1609/aaai.v35i5.16542

[20]Wang S, Liu T, Luo K, et al. Identification of engine faults based on acoustic emission signals using a 1DCNN-ViT ensemble model. Measurement Science and Technology. 2023; 34(2): 024007. doi: 10.1088/1361-6501/aca041

[21]Vaswani A, Shazeer N, Parmar N, et al. Attention Is All You Need. arXiv preprint. 2017. doi: 10.48550/ARXIV.1706.03762

[22]Niu Z, Zhong G, Yu H. A review on the attention mechanism of deep learning. Neurocomputing. 2021; 452: 48–62. doi: 10.1016/j.neucom.2021.03.091

[23]Zhang YH, Shao F, Zhao XP, et al. Rolling bearing fault diagnosis based on multi-label zero-shot learning. Journal of Vibration and Shock. 2022; 41(11): 55–64. (in Chinese)