Self-Supervised Representation Learning for Remaining Useful Life Prediction: A Taxonomy and Triplet-Loss Case Study
##plugins.themes.bootstrap3.article.main##
##plugins.themes.bootstrap3.article.sidebar##
Abstract
Remaining useful life (RUL) prediction is fundamentally constrained by the scarcity of labeled run-to-failure data, motivating growing use of self-supervised learning (SSL) to exploit abundant unlabeled condition monitoring data. However, the literature lacks a systematic comparison of the loss functions underlying these SSL methods, as existing surveys organize the field by architecture rather than by the training objective itself. This paper reviews self-supervised loss functions for RUL prediction, spanning temporal-order, contrastive, metric-learning, multi-view, and meta-learning objectives, summarizing each family's assumptions, requirements, and limitations. To ground this review empirically, we present a framework pre-training a Transformer encoder with a triplet margin loss and evaluating it under few-shot fine-tuning on the C-MAPSS turbofan benchmark. Across 2 to 50 labeled fine-tuning sequences, the pre-trained model consistently outperforms an identically structured supervised baseline, with the gap narrowing as labeled data becomes abundant, confirming that SSL pre-training is most valuable precisely in the label-scarce regime.
How to Cite
##plugins.themes.bootstrap3.article.details##
RUL Prediction, Self-Supervised Learning, Prognostics, Triplet, Representation Learning
Dong, J., Wu, H., Zhang, H., Zhang, L., Wang, J., & Long, M. (2023). Simmtm: A simple pre-training framework for masked time-series modeling. Advances in Neural Information Processing Systems, 36, 29996–30025.
Eldele, E., Ragab, M., Chen, Z., Wu, M., Kwoh, C. K., Li, X., & Guan, C. (2021). Time-series representation learning via temporal and contextual contrasting. arXiv preprint arXiv:2106.14112.
Faysal, A., Ngui, W., Lim, M., & Leong, M. (2023). Ensemble augmentation for deep neural networks using 1-d time series vibration data. Journal of vibration engineering & technologies, 11(5), 1987–2011.
Franceschi, J.-Y., Dieuleveut, A., & Jaggi, M. (2019). Unsupervised scalable representation learning for multivariate time series. Advances in neural information processing systems, 32.
Guo, H., Zhu, H., Wang, J., Vadakkepat, P., Ho, W. K., & Lee, T. H. (2022). Masked self-supervision for remaining useful lifetime prediction in machine tools. In 2022 ieee 20th international conference on industrial informatics (indin) (pp. 353–358).
He, K., Chen, X., Xie, S., Li, Y., Dollar, P., & Girshick, R. (2022). Masked autoencoders are scalable vision learners. In Proceedings of the ieee/cvf conference on computer vision and pattern recognition (pp. 16000– 16009).
Hou, G., Xu, S., Zhou, N., Yang, L., & Fu, Q. (2020). Remaining useful life estimation using deep convolutional generative adversarial networks based on an autoencoder scheme. Computational Intelligence and Neuroscience, 2020(1), 9601389.
Kong, Z., Jin, X., Xu, Z., & Chen, Z. (2023). A contrastive learning framework enhanced by unlabeled samples for remaining useful life prediction. Reliability Engineering & System Safety, 234, 109163.
Krokotsch, T., Knaak, M., & G¨uhmann, C. (2022). Improving semi-supervised learning for remaining useful lifetime estimation through self-supervision. International Journal of Prognostics and Health Management, 13(1).
Lai, J.-J., Pan, S.-J., Ong, C.-W., & Chen, C.-L. (2022). Weibull reliability regression model for prediction of bearing remaining useful life. In Computer aided chemical engineering (Vol. 49, pp. 829–834). Elsevier.
Li, Y., Bai, X., Liu, C., Chu, J., & Tan, J. (2025). A selfsupervised assisted label-efficient method for online remaining useful life prediction. Measurement, 242, 115902.
Liu, B., Xu, Z., Kang, Y., Cao, Y., & Zhao, Y. (2025). Multisensor contrast neural network for remaining useful life prediction of rolling bearings under scarce labeled data. Frontiers of Information Technology & Electronic Engineering, 26(7), 1180–1193.
Liu, J., & Chen, S. (2024). Timesurl: Self-supervised contrastive learning for universal time series representation learning. In Proceedings of the aaai conference on artificial intelligence (Vol. 38, pp. 13918–13926).
Ma, L., & Zhang, Z. (2026). A meta multi-scale spatial attention network for interpretable few-shot remaining useful life prediction. Engineering Research Express, 8(7), 075206.
Mo, R., Zhou, H., Yin, H., & Si, X. (2025). A survey on few-shot learning for remaining useful life prediction. Reliability Engineering & System Safety, 257, 110850.
Mo, Y., Li, L., Huang, B., & Li, X. (2023). Few-shot rul estimation based on model-agnostic meta-learning. Journal of Intelligent Manufacturing, 34(5), 2359–2372.
Morales-Espejel, G., & Gabelli, A. (2020). A model for rolling bearing life with surface and subsurface survival: Surface thermal effects. Wear, 460, 203446.
Nectoux, P., Gouriveau, R., Medjaher, K., Ramasso, E., Chebel-Morello, B., Zerhouni, N., & Varnier, C. (2012). Pronostia: An experimental platform for bearings accelerated degradation tests. In Ieee international conference on prognostics and health management, phm’12. (pp. 1–8).
Nie, Y., Nguyen, N. H., Sinthong, P., & Kalagnanam, J. (2022). A time series is worth 64 words: Long-term forecasting with transformers. arXiv preprint arXiv:2211.14730.
Paris, P., & Erdogan, F. (1963). A critical analysis of crack propagation laws. Journal of Basic engineering, 85(4), 528–533.
Saxena, A., Goebel, K., Simon, D., & Eklund, N. (2008). Damage propagation modeling for aircraft engine runto- failure simulation. In 2008 international conference on prognostics and health management (pp. 1–9).
Shang, X., Shang, J., Li, M., Qiu, H., Gao, L., & Xu, D. (2025). A cross-domain few-shot remaining useful life estimation framework based on model-agnostic metalearning with task embeddings. Computers in Industry, 173, 104396.
Shen, Z., Yang, C., Cheng, L., Kong, X., Liu, Z., & Su, K. (2026). A novel dual-dimensional contrastive self-supervised learning-based framework for rolling bearing remaining useful life prediction. Scientific Reports.
Shi, Z., & Chehade, A. (2021). A dual-lstm framework combining change point detection and remaining useful life prediction. Reliability Engineering & System Safety, 205, 107257.
Song, L., Zheng, C., Lu, X., Liu, J., & Wang, H. (2026). A multi-task prediction network with dynamic convolution and self-supervised contrastive learning in mechanical remaining useful life prediction. Journal of Vibration and Control, 10775463261450874.
Tonekaboni, S., Eytan, D., & Goldenberg, A. (2021). Unsupervised representation learning for time series with temporal neighborhood coding. arXiv preprint arXiv:2106.00750.
Vermelin, W. S., L¨ovberg, A., & Kyprianidis, K. (2022). Self-supervised learning for efficient remaining useful life prediction. In Annual conference of the phm society (Vol. 14).
Wang, B., Lei, Y., Li, N., & Li, N. (2018). A hybrid prognostics approach for estimating remaining useful life of rolling element bearings. IEEE Transactions on reliability, 69(1), 401–412.
Wang, L., Cao, H., Xu, H., & Liu, H. (2022). A gated graph convolutional network with multi-sensor signals for remaining useful life prediction. Knowledge-Based Systems, 252, 109340.
Wu, C., He, J., Wang, L., Ma, C., & Liu, Y. (2025). A dual-path architecture based on time series decomposition and degradation correction for remaining useful life prediction of aero-engine. Computers & Industrial Engineering, 203, 110964.
Yang, C., Cheng, L., Shen, Z., Liu, Z., Su, K., Kong, X., . . . Wang, Z. (2026). Iaa-crl: A universal informationadaptive augmentation based contrastive representation learning framework for remaining useful life prediction of rolling bearings. Mechanical Systems and Signal Processing, 254, 114357.
Yang, J., & Wang, X. (2024). Meta-learning with deep flow kernel network for few shot cross-domain remaining useful life prediction. Reliability Engineering & System Safety, 244, 109928.
Yang, J., Wang, X., & Luo, Z. (2024). Few-shot remaining useful life prediction based on meta-learning with deep sparse kernel network. Information Sciences, 653, 119795.
Yue, Z., Wang, Y., Duan, J., Yang, T., Huang, C., Tong, Y., & Xu, B. (2022). Ts2vec: Towards universal representation of time series. In Proceedings of the aaai conference on artificial intelligence (Vol. 36, pp. 8980–8987).
Zhang, J., Luo, H., Hu, J., Cheng, C., Wu, S., Yan, P., & Tian, J. (2025). A transferable topology-aware graph pooling network for remaining useful life prediction under cross-domain conditions. IEEE Transactions on Reliability.
Zhang, W., Li, C., Peng, G., Chen, Y., & Zhang, Z. (2018). A deep convolutional neural network with new training methods for bearing fault diagnosis under noisy environment and different working load. Mechanical systems and signal processing, 100, 439–453.
Zhao, Y., & Wang, Y. (2021). Remaining useful life prediction for multi-sensor systems using a novel end-to-end deeplearning method. Measurement, 182, 109685.
Zhou, L., & Wang, H. (2024). Mst-gat: A multi-perspective spatial-temporal graph attention network for multisensor equipment remaining useful life prediction. Information Fusion, 110, 102462.
Zhu, Z., Lv, Z., & Xie, T. (2025). Trend contrast featuresbased bearing remaining useful life prediction method. Control Engineering Practice, 162, 106358.
This work is licensed under a Creative Commons Attribution 3.0 Unported License.
The Prognostic and Health Management Society advocates open-access to scientific data and uses a Creative Commons license for publishing and distributing any papers. A Creative Commons license does not relinquish the author’s copyright; rather it allows them to share some of their rights with any member of the public under certain conditions whilst enjoying full legal protection. By submitting an article to the International Conference of the Prognostics and Health Management Society, the authors agree to be bound by the associated terms and conditions including the following:
As the author, you retain the copyright to your Work. By submitting your Work, you are granting anybody the right to copy, distribute and transmit your Work and to adapt your Work with proper attribution under the terms of the Creative Commons Attribution 3.0 United States license. You assign rights to the Prognostics and Health Management Society to publish and disseminate your Work through electronic and print media if it is accepted for publication. A license note citing the Creative Commons Attribution 3.0 United States License as shown below needs to be placed in the footnote on the first page of the article.
First Author et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 United States License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.