Abstract: Unanticipated wear of tunnel boring machine (TBM) disc cutters is a critical factor causing project delays and cost overruns in tunneling engineering. Accurate, real-time prediction of the cutter’s wear state is therefore essential for enabling predictive maintenance. Data-driven methods, particularly deep learning, have shown promise for this task, but their performance is constrained by the scarcity of high-quality labeled data in practical industrial settings. To address this challenge, we propose a novel, decoupled semi-supervised framework called PL-HLANet. The first component of this framework is a multi-view pseudo-labeling (PL) module, which mines high-confidence supervisory signals from massive unlabeled data by leveraging heterogeneous views derived from feature engineering and diverse model architectures; it is followed by a consistency check to ensure label quality. This process effectively augments the training set while correcting for sampling bias. Subsequently, a specialized hierarchical hybrid attention network (HLANet) is used to make predictions. The HLANet organically integrates a temporal convolutional network (TCN) for local feature extraction, a bidirectional long short-term memory (Bi-LSTM) network for capturing temporal dynamics, and a custom attention mechanism for focusing on critical information. Experiments on a real-world tunneling dataset show that PL-HLANet significantly outperforms both supervised and mainstream semi-supervised baselines, such as the Mean Teacher and FixMatch. The framework’s effectiveness is further substantiated by validations of its architectural design and data-driven selection of hyperparameters. Moreover, PL-HLANet has a high inference speed, showcasing its practicality for real-world scenarios. Our work provides an effective solution for machining equipment monitoring in data-scarce industrial environments.

PL-HLANet：一种用于隧道掘进机滚刀磨损预测的半监督方法

[1]AgrawalAK, MurthyVMSR, ChattopadhyayaS, et al., 2022. Prediction of TBM disc cutter wear and penetration rate in tunneling through hard and abrasive rock using multi-layer shallow neural network and response surface methods. Rock Mechanics and Rock Engineering, 55(6):3489-3506.

[2]BaiLP, MoDY, LiHS, et al., 2025. Optimized data preprocessing and model selection for TBM cutter wear prediction. Coatings, 15(5):564.

[3]BarzegariG, KhodayariJ, RostamiJ, 2021. Evaluation of TBM cutter wear in Naghadeh water conveyance tunnel and developing a new prediction model. Rock Mechanics and Rock Engineering, 54(12):6281-6297.

[4]ChenBX, JiangJG, WangXM, et al., 2022. Debiased self-training for semi-supervised learning. Proceedings of the 36th International Conference on Neural Information Processing Systems, article 2349.

[5]ChengQ, ChenYX, XiaoYT, et al., 2022. A dual-stage attention-based Bi-LSTM network for multivariate time series prediction. The Journal of Supercomputing, 78(14):16214-16235.

[6]DingXB, XieYX, XueHW, et al., 2022. A new approach for developing EPB-TBM disc cutter wear prediction equations in granite stratum using backpropagation neural network. Tunnelling and Underground Space Technology, 128:104654.

[7]DouYM, LiKW, LvWJ, et al., 2024. ContrasInver: ultra-sparse label semi-supervised regression for multidimensional seismic inversion. IEEE Transactions on Geoscience and Remote Sensing, 62:5917613.

[8]FanJ, ZhangK, HuangY, et al., 2023. Parallel spatio-temporal attention-based TCN for multivariate time series prediction. Neural Computing and Applications, 35(18):13109-13118.

[9]GhorbaniE, YagizS, 2024. Predicting disc cutter wear using two optimized machine learning techniques. Archives of Civil and Mechanical Engineering, 24(2):106.

[10]HewageP, BeheraA, TrovatiM, et al., 2020. Temporal convolutional neural (TCN) network for an effective weather forecasting using time-series data from the local weather station. Soft Computing, 24(21):16453-16482.

[11]JiaLX, PuXB, ShangKD, et al., 2022. A deep learning combination model to predict TBM disc-cutter wear status. Proceedings of the IEEE 17th Conference on Industrial Electronics and Applications, p.1469-1474.

[12]KaramiM, ZareS, RostamiJ, 2021. Introducing an empirical model for prediction of disc cutter life for TBM application in jointed rocks: case study, Kerman water conveyance tunnel. Bulletin of Engineering Geology and the Environment, 80(5):3853-3870.

[13]LiL, LiuZB, ZhouHY, et al., 2022. Prediction of TBM cutterhead speed and penetration rate for high-efficiency excavation of hard rock tunnel using CNN-LSTM model with construction big data. Arabian Journal of Geosciences, 15(3):280.

[14]LuDC, LiuYH, KongFC, et al., 2024. A novel Bi-LSTM method fusing current and historical data for tunnelling parameters of shield tunnel. Transportation Geotechnics, 49:101402.

[15]MahmoodzadehA, MohammadiM, Hashim IbrahimH, et al., 2021. Machine learning forecasting models of disc cutters life of tunnel boring machine. Automation in Construction, 128:103779.

[16]MoDY, BaiLP, HuangWR, et al., 2024. TBM disc cutter wear prediction using stratal slicing and IPSO-LSTM in mixed weathered granite stratum. Tunnelling and Underground Space Technology, 148:105745.

[17]PuXB, JiaLX, ShangKD, et al., 2022. A new strategy for disc cutter wear status perception using vibration detection and machine learning. Sensors, 22(17):6686.

[18]ShenX, ChenXS, FuYB, et al., 2022. Prediction and analysis of slurry shield TBM disc cutter wear and its application in cutter change time. Wear, 498-499:204314.

[19]ShiG, QinCJ, ZhangZN, et al., 2024. A novel decomposition and hybrid transfer learning-based method for multi-step cutterhead torque prediction of shield machine. Mechanical Systems and Signal Processing, 214:111362.

[20]ShinYJ, KwonK, BaeA, et al., 2024. Machine learning-based prediction model for disc cutter life in TBM excavation through hard rock formations. Tunnelling and Underground Space Technology, 150:105826.

[21]SunJD, ShangY, WangK, et al., 2023. A new prediction model for disc cutter wear based on Cerchar Abrasivity Index. Wear, 526-527:204927.

[22]WangHD, QinCJ, YuHG, et al., 2025. A real-time multi-head mixed attention mechanism-based prediction method for tunnel boring machine disc cutter wear. Science China Technological Sciences, 68(1):1120302.

[23]WangKY, WuXG, ZhangLM, et al., 2023. Data-driven multi-step robust prediction of TBM attitude using a hybrid deep learning approach. Advanced Engineering Informatics, 55:101854.

[24]WenXY, LiWB, 2023. Time series prediction based on LSTM-attention-LSTM model. IEEE Access, 11:48322-48331.

[25]WuYH, WuGQ, LinJX, et al., 2025. Role exchange-based self-training semi-supervision framework for complex medical image segmentation. IEEE Transactions on Neural Networks and Learning Systems, 36(5):8372-8386.

[26]YuFX, WangD, ChenYP, et al., 2021. Unsupervised domain adaptation for object detection via cross-domain semi-supervised learning. arXiv:1911.07158.

[27]YuHG, TaoJF, HuangS, et al., 2021. A field parameters-based method for real-time wear estimation of disc cutter on TBM cutterhead. Automation in Construction, 124:103603.

[28]ZhangN, ShenSL, ZhouAN, 2023. A new index for cutter life evaluation and ensemble model for prediction of cutter wear. Tunnelling and Underground Space Technology, 131:104830.

Similar articles