All Issue

2025 Vol.27, Issue 5 Preview Page

Research Paper

Data-augmented machine learning approach for abnormal wear prediction in TBM disc cutters

TBM 디스크 커터 비정상 마모 예측을 위한 데이터 증강 기반 머신러닝 연구

Yuri Yeom1
,
Hangseok Choi2
,
Yerim Yang3
,
Jinsoo Park4
,
Kibeom Kwon5*
염 유리1
,
최 항석2
,
양 예림3
,
박 진수4
,
권 기범5*
1Master Student, Dept. of Civil, Environmental and Architectural Engineering, Korea University
2Professor, School of Civil, Environmental and Architectural Engineering, Korea University
3Ph.D. Candidate, Dept. of Civil, Environmental and Architectural Engineering, Korea University
4Research Professor, Future and Fusion Lab of Architectural, Civil and Environmental Engineering, Korea University
5Researcher, Future and Fusion Lab of Architectural, Civil and Environmental Engineering, Korea University
1정회원, 고려대학교 건축사회환경공학과 석사과정
2정회원, 고려대학교 건축사회환경공학부 교수
3정회원, 고려대학교 건축사회환경공학과 박사과정
4정회원, 고려대학교 미래건설환경융합연구소 연구교수
5정회원, 고려대학교 미래건설환경융합연구소 연구원
*Corresponding Author
30 September 2025. pp. 367-383
PDF XML Citation
Abstract

Although numerous studies have addressed disc cutter wear prediction, most have focused on normal wear due to its higher occurrence, resulting in relatively limited research on abnormal wear. This study developed a machine learning model for predicting abnormal wear using data augmentation and analyzed the influence of features on the model predictions based on a dataset obtained from an earth pressure balance (EPB) shield TBM site. Cutter wear was classified into two categories: normal and abnormal wear. To address the class imbalance between the two wear types, the Synthetic Minority Oversampling Technique (SMOTE) and BorderlineSMOTE were applied. A total of six models were constructed using Random Forest (RF) and eXtreme Gradient Boosting (XGB). The prediction results showed that models trained on the augmented datasets improved abnormal wear detection by approximately 18 to 38% compared to those trained on the original imbalanced dataset, while also reducing the performance disparity between the two wear types. In particular, the RF-BorderlineSMOTE model achieved comparable recall values of 0.842 and 0.836 for normal and abnormal wear, respectively, indicating that oversampling near the decision boundary effectively enhanced abnormal wear detection. Furthermore, SHapley Additive exPlanation (SHAP) analysis identified chamber pressure, thrust force and elastic modulus as key predictive features. Increases in chamber pressure and thrust force were associated with a higher risk of abnormal wear, whereas higher elastic modulus was found to reduce this risk.

Keywords
Tunnel boring machine
Disc cutter
Abnormal wear
Machine learning
Data augmentation

디스크 커터의 마모 예측에 관한 연구는 지속적으로 이루어져 왔으나, 정상 마모와 비정상 마모 간 발생 빈도 차이로 인해 대부분 정상 마모 예측에 초점을 두고 있으며, 비정상 마모에 대한 연구는 상대적으로 부족한 실정이다. 본 연구에서는 데이터 증강을 활용하여 비정상 마모 예측을 위한 머신러닝 모델을 구축하고, 모델 내 특성 중요도 및 그 영향 양상을 분석하였다. 토압식 쉴드 TBM 현장에서 수집된 데이터를 기반으로 마모 유형을 정상 마모와 비정상 마모로 분류하여 데이터세트를 구축하였다. Synthetic Minority Oversampling Technique (SMOTE)와 BorderlineSMOTE의 2가지 데이터 증강 기법을 도입하여 마모 유형 간의 불균형이 해소된 증강 데이터세트를 구축하였다. Random Forest (RF)와 eXtreme Gradient Boosting (XGB)의 머신러닝 기법을 기존 데이터세트와 2가지 증강 데이터세트에 각각 적용하여 총 6가지의 예측 모델을 개발하였다. 예측 성능을 비교한 결과, 증강 데이터 기반 모델은 기존 불균형 데이터 기반 모델에 비해 비정상 마모 탐지 성능이 약 18%에서 38%까지 크게 향상되었으며, 정상 마모와 비정상 마모 간의 성능 차이 또한 감소하는 경향을 보였다. 특히, 최적 모델인 RF-BorderlineSMOTE 모델의 경우에는 정상 마모와 비정상 마모에 대한 재현율이 각각 0.842와 0.836으로 유사하게 나타났으며, 이는 두 마모 유형 간의 결정 경계 인근 데이터를 중심으로 한 증강 방식이 비정상 마모 탐지에 효과적임을 시사한다. 또한, SHapley Additive exPlanation (SHAP) 분석에 따라 챔버압, 추력, 탄성계수가 예측 모델 내 주요 특성으로 식별되었으며, 챔버압과 추력은 비정상 마모 발생 가능성과 양의 상관성을, 탄성계수는 음의 상관성을 보였다.

키워드
Tunnel boring machine
디스크 커터
비정상 마모
머신러닝
데이터증강
References
1

Agrawal, A.K., Murthy, V.M.S.R., Chattopadhyaya, S., Raina, A.K. (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, Vol. 55, No. 6, pp. 3489-3506.

10.1007/s00603-022-02834-7
2

Bae, G.J., Choi, S.W., Chang, S.H., Lee, G.P., Song, B.C., Kim, K.B. (2015), “Analysis of inner parts in the disc cutters applied to the field tests”, Journal of Korean Tunnelling and Underground Space Association, Vol. 17, No. 4, pp. 473-485.

10.9711/KTAJ.2015.17.4.473
3

Breiman, L. (2001), “Random forests”, Machine Learning, Vol. 45, pp. 5-32.

10.1023/A:1010933404324
4

Bruland, A. (2000), Hard rock tunnel boring, Doctoral Thesis, Norwegian University of Science and Technology, pp. 14-31.

5

Chang, S.H., Choi, S.W., Park, Y.T., Lee, G.P., Bae, G.J. (2012), “Characterization of the deformation of a disc cutter in linear rock cutting test”, Journal of Korean Tunnelling and Underground Space Association, Vol. 14, No. 3, pp. 197-213.

10.9711/KTAJ.2012.14.3.197
6

Chawla, N.V., Bowyer, K.W., Hall, L.O., Kegelmeyer, W.P. (2002), “SMOTE: Synthetic minority over-sampling technique”, Journal of Artificial Intelligence Research, Vol. 16, pp. 321-357.

10.1613/jair.953
7

Chen, T., Guestrin, C. (2016), “XGBoost: A scalable tree boosting system”, Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, USA, pp. 785-794.

10.1145/2939672.2939785
8

Frenzel, C., Käsling, H., Thuro, K. (2008), “Factors influencing disc cutter wear”, Geomechanics and Tunnelling, Vol. 1, No. 1, pp. 55-60.

10.1002/geot.200800006
9

Gao, Y., Chen, R., Qin, W., Wei, L., Zhou, C. (2023), “Learning from explainable data-driven tunneling graphs: A spatio-temporal graph convolutional network for clogging detection”, Automation in Construction, Vol. 147, 104741.

10.1016/j.autcon.2023.104741
10

Garcia-Gutiérrez, S., Esteban-Aizpiri, C., Lafuente, I., Barrio, I., Quiros, R., Quintana, J.M., Uranga, A. (2022), “Machine learning-based model for prediction of clinical deterioration in hospitalized patients by COVID 19”, Scientific Reports, Vol. 12, No. 1, 7097.

10.1038/s41598-022-09771-z35501359PMC9059444
11

Ge, Y., Liu, Y., Lin, P., Xu, Z. (2022), “Effects of rock properties on the wear of TBM disc cutter: A case study of the Yellow River diversion project, China”, International Journal of Geomechanics, Vol. 22, No. 4, 04022011.

10.1061/(ASCE)GM.1943-5622.0002344
12

Gehring, K. (1995), “Prognosis of advance rates and wear for underground mechanized excavations”, Felsbau, Vol. 13, No. 6, pp. 439-448.

13

Gong, Q., Yin, L., Ma, H., Zhao, J. (2016), “TBM tunnelling under adverse geological conditions: An overview”, Tunnelling and Underground Space Technology, Vol, 57, pp. 4-17.

10.1016/j.tust.2016.04.002
14

Han, H., Wang, W.Y., Mao, B.H. (2005), “Borderline-SMOTE: A new over-sampling method in imbalanced data sets learning”, Proceedings of the International Conference on Intelligent Computing, Springer Berlin Heidelberg, Heidelberg, Berlin, pp. 878-887.

10.1007/11538059_91
15

Hassanpour, J., Rostami, J., Tarigh Azali, S., Zhao, J. (2014), “Introduction of an empirical TBM cutter wear prediction model for pyroclastic and mafic igneous rocks; a case history of Karaj water conveyance tunnel, Iran”, Tunnelling and Underground Space Technology, Vol. 43, pp. 222-231.

10.1016/j.tust.2014.05.007
16

Hwang, B., Kang, M., Kwon, K., Yang, J., Choi, H. (2023), “Evaluating rheological properties of excavated soil for EPB shield TBM with foam and polymer”, Journal of Korean Tunnelling and Underground Space Association, Vol. 25, No. 5, pp. 387-401.

10.9711/KTAJ.2023.25.5.387
17

Kang, M., Lee, J., Kwon, K., Park, S., Choi, H. (2023), “Laboratory simulations on hybrid non-destructive survey of electrical resistivity and induced polarization to predict geological risks ahead of a TBM tunnel”, Tunnelling and Underground Space Technology, Vol. 135, 105066.

10.1016/j.tust.2023.105066
18

Kang, T.H., Choi, S.W., Lee, C., Chang, S.H. (2022), “A study on the prediction of disc cutter wear using TBM data and machine learning algorithm”, Tunnel and Underground Space, Vol. 32, No. 6, pp. 502-517.

10.7474/TUS.2022.32.6.502
19

Kim, Y., Shin, J., Kim, B. (2024), “Analysis of disc cutter replacement based on wear patterns using artificial intelligence classification models”, Geomechanics and Engineering, Vol. 38, No. 6, pp. 633-645.

10.12989/GAE.2024.38.6.633
20

Kwon, K., Choi, H., Jung, J., Kim, D., Shin, Y.J. (2025a), “Prediction of abnormal TBM disc cutter wear in mixed ground condition using interpretable machine learning with data augmentation”, Journal of Rock Mechanics and Geotechnical Engineering, Vol. 17, No. 4, pp. 2059-2071.

10.1016/j.jrmge.2024.05.027
21

Kwon, K., Choi, H., Pham, K., Kim, S., Bae, A. (2024), “Influence analysis of pavement distress on international roughness index using machine learning”, KSCE Journal of Civil Engineering, Vol. 28, No. 10, pp. 4344-4355.

10.1007/s12205-024-0093-9
22

Kwon, K., Kang, M., Kim, D., Choi, H. (2023), “Prioritization of hazardous zones using an advanced risk management model combining the analytic hierarchy process and fuzzy set theory”, Sustainability, Vol. 15, No. 15, 12018.

10.3390/su151512018
23

Kwon, K., Kang, M., Shin, Y.J., Ahn, B., Choi, H. (2025b), “An interpretable framework for risk management in TBM excavation using expert elicitation integrated with fuzzy set theory”, Scientific Reports, Vol. 15, No. 1, 23030.

10.1038/s41598-025-07514-440594655PMC12215182
24

Kwon, K., Yeom, Y., Shin, Y.J., Bae, A., Choi, H. (2025c), “Machine learning models for predicting the International Roughness Index of asphalt concrete overlays on Portland cement concrete pavements”, Computer-Aided Civil and Infrastructure Engineering, Vol. 40, No. 21, pp. 3385-3402.

10.1111/mice.13524
25

Lee, S., Choi, H., Kwon, K., Lee, D., Hwang, B. (2024), “A lab-scale screw conveyor system for EPB shield TBM: System development and applicability assessment”, Journal of Korean Tunnelling and Underground Space Association, Vol. 26, No. 5, pp. 533-549.

10.9711/KTAJ.2024.26.5.533
26

Li, Q., Du, L., Yang, Y., Ni, Z., Zhao, X. (2024), “Intelligent identification and warning method of disc cutter abnormal wear in TBM construction based on extreme learning machine”, Scientific Reports, Vol. 14, No. 1, 30655.

10.1038/s41598-024-76172-939730385PMC11680588
27

Liu, Q., Huang, X., Gong, Q., Du, L., Pan, Y., Liu, J. (2016), “Application and development of hard rock TBM and its prospect in China”, Tunneling and Underground Space Technology, Vol. 57, pp. 33-46.

10.1016/j.tust.2016.01.034
28

Liu, Q., Liu, J., Pan, Y., Zhang, X., Peng, X., Gong, Q., Du, L. (2017), “A wear rule and cutter life prediction model of a 20-in. TBM cutter for granite: A case study of a water conveyance tunnel in China”, Rock Mechanics and Rock Engineering, Vol. 50, pp. 1303-1320.

10.1007/s00603-017-1176-4
29

Lundberg, S.M., Lee, S.I. (2017), “A unified approach to interpreting model predictions”, Advances in Neural Information Processing Systems, Vol. 30.

30

Mahmoodzadeh, A., Mohammadi, M., Ibrahim, H.H., Abdulhamid, S.N., Ali, H.F.H., Hasan, A.M., Khishe, M., Mahmud, H. (2021), “Machine learning forecasting models of disc cutters life of tunnel boring machine”, Automation in Construction, Vol. 128, 103779.

10.1016/j.autcon.2021.103779
31

Ren, D.J., Shen, S.L., Arulrajah, A., Cheng, W.C. (2018), “Prediction model of TBM disc cutter wear during tunnelling in heterogeneous ground”, Rock Mechanics and Rock Engineering, Vol. 51, pp. 3599-3611.

10.1007/s00603-018-1549-3
32

Rostami, J. (1997), Development of a force estimation model for rock fragmentation with disc cutters through theoretical modeling and physical measurement of crushed zone pressure, Doctoral Thesis, Colorado School of Mines, Golden, Colorado, USA, pp. 55-125.

33

Shen, X., Chen, X., Fu, Y., Cao, C., Yuan, D., Li, X., Xiao, Y. (2022), “Prediction and analysis of slurry shield TBM disc cutter wear and its application in cutter change time”, Wear, Vol. 498-499, 204314.

10.1016/j.wear.2022.204314
34

Shin, Y.J., Kwon, K., Bae, A., Choi, H., Kim, D. (2024), “Machine learning-based prediction model for disc cutter life in TBM excavation through hard rock formations”, Tunnelling and Underground Space Technology, Vol. 150, 105826.

10.1016/j.tust.2024.105826
35

Su, W., Li, X., Jin, D., Yang, Y., Qin, R., Wang, X. (2020), “Analysis and prediction of TBM disc cutter wear when tunneling in hard rock strata: A case study of a metro tunnel excavation in Shenzhen, China”, Wear, Vol. 446-447, 203190.

10.1016/j.wear.2020.203190
36

Yang, H., Liu, B., Wang, Y., Li, C. (2021), “Prediction model for normal and flat wear of disc cutters during TBM tunneling process”, International Journal of Geomechanics, Vol. 21, No. 3, 06021002.

10.1061/(ASCE)GM.1943-5622.0001950
37

Yang, J.H., Zhang, X.P., Ji, P.Q., Liu, Q.S., Lu, X.J., Wei, J.P., Qi, S.H., Fang, H.G., Fang, J.N., Geng, Y.J. (2019), “Analysis of disc cutter damage and consumption of TBM1 section on water conveyance tunnel at Lanzhou water source construction engineering”, Tunnelling and Underground Space Technology, Vol. 85, pp. 67-75.

10.1016/j.tust.2018.11.040
38

Yang, Y., Hong, K., Sun, Z., Chen, K., Li, F., Zhou, J., Zhang, B. (2018), “The derivation and validation of TBM disc cutter wear prediction model”, Geotechnical and Geological Engineering, Vol. 36, No. 6, pp. 3391-3398.

10.1007/s10706-018-0540-9
39

Yeom, Y., Choi, H., Yang, Y., Kwon, K. (2025), “Application of data augmentation for enhancing machine learning-based TBM disc cutter wear prediction”, Journal of Korean Tunnelling and Underground Space Association, Vol. 27, No. 3, pp. 217-232.

10.9711/KTAJ.2025.27.3.217
40

Yoon, Y., Choi, H., Kwon, K., Hwang, B., Kang, M. (2023), “Optimization of electrical resistivity survey utilizing modified harmony search algorithm to predict anomalous zone ahead of tunnel faces”, Measurement, Vol. 223, 113747.

10.1016/j.measurement.2023.113747
41

Zhang, N., Shen, S.L., Zhou, A. (2023), “A new index for cutter life evaluation and ensemble model for prediction of cutter wear”, Tunnelling and Underground Space Technology, Vol. 131, 104830.

10.1016/j.tust.2022.104830
42

Zhao, J., Gong, Q.M., Eisensten, Z. (2007), “Tunnelling through a frequently changing and mixed ground: A case history in Singapore”, Tunnelling and Underground Space Technology, Vol. 22, No. 4, pp. 388-400.

10.1016/j.tust.2006.10.002
Information
  • Publisher :Korean Tunneling and Underground Space Association
  • Publisher(Ko) :한국터널지하공간학회
  • Journal Title :Journal of Korean Tunnelling and Underground Space Association
  • Journal Title(Ko) :한국터널지하공간학회 논문집
  • Volume : 27
  • No :5
  • Pages :367-383
  • Received Date : 2025-08-04
  • Revised Date : 2025-09-02
  • Accepted Date : 2025-09-02
  • DOI :https://doi.org/10.9711/KTAJ.2025.27.5.367
Journal Informaiton Journal of Korean Tunnelling and Underground Space Association Journal of Korean Tunnelling and Underground Space Association
Online Submission & Review System
  • NRF
  • KOFST
  • KISTI Current Status
  • KISTI Cited-by
  • crosscheck
  • orcid
  • open access
  • ccl
Journal Informaiton Journal Informaiton - close