Rolling Window Deep Learning for Hard-rock TBM Penetration Rate Prediction

Nantapol Monthanopparat; Tawatchai Tanchaisawat

doi:doi:10.11648/j.ajma.20261301.12

Research Article |

| Peer-Reviewed

Rolling Window Deep Learning for Hard-rock TBM Penetration Rate Prediction

Nantapol Monthanopparat^*

, Tawatchai Tanchaisawat

Published in American Journal of Mechanics and Applications (Volume 13, Issue 1)

Received: 4 May 2026 Accepted: 14 May 2026 Published: 26 May 2026

Views: Downloads:

Download PDF

Share This Article

Twitter
Linked In
Facebook

Abstract

Penetration rate prediction for hard rock tunnel boring machines (TBMs) remains challenging because rock mass conditions and machine operating regimes vary continuously along the tunnel alignment. Conventional static prediction models may not adequately represent such non-stationary construction conditions, particularly when geotechnical measurements are incomplete or obtained with delay during excavation. This study aims to develop a construction-phase prediction framework for ring-scale TBM penetration rate in a granite tunnel drive by integrating geotechnical data completion, sequence deep learning, and rolling-window model evaluation. A dataset of 1,000 consecutive rings was compiled, including boring-only penetration rate, thrust, torque, cutterhead rotational speed, rock mass type, and uniaxial compressive strength (UCS). Missing UCS measurements were completed using inverse distance weighting within a block model representation, resulting in estimated UCS values of approximately 33–177 MPa for rock masses dominated by massive and fractured granite. Three sequence deep learning models, namely long short-term memory (LSTM), gated recurrent unit (GRU), and temporal convolutional network (TCN), were evaluated using root mean square error (RMSE), mean absolute error (MAE), and a symmetric ±10% tolerance band adapted from accuracy-band concepts used in AACE-based project controls. The proposed rolling protocol used 100-ring validation and 100-ring test blocks to assess predictive performance under changing ground conditions. The results show that the optimized GRU model provided the most robust overall performance, achieving a mean test RMSE of approximately 0.229 m/h and a mean within-band compliance of approximately 54% across rolling folds. These findings indicate that rolling-window sequence learning can provide a practical and adaptable framework for construction-phase TBM performance prediction under evolving geological and operational conditions.

Published in	American Journal of Mechanics and Applications (Volume 13, Issue 1)
DOI	10.11648/j.ajma.20261301.12
Page(s)	10-21
Creative Commons	This is an Open Access article, distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium or format, provided the original work is properly cited.
Copyright	Copyright © The Author(s), 2026. Published by Science Publishing Group

Keywords

Tunnel Boring Machine, Penetration Rate Prediction, Hard Rock Tunnelling, Rolling Window Evaluation, Gated Recurrent Unit

References

[1]	Paraskevopoulou C & Boutsis G, Cost overruns in Tunnelling Projects: Investigating the impact of geological and geotechnical uncertainty using case studies, infrastructures, 2020. https://doi.org/10.3390/infrastructures5090073
[2]	Pandey P. K. et al., Influence of geology on tunnel boring machine perfromance - A review, Journal of Mining and Metallurgy, pp. 1-14, 2020.
[3]	Rostami J., Performance prediction of hard rock Tunnel Boring Machine (TBMs) in difficult ground, Tunnelling and underground space technology, no. 57, pp. 173-182, 2016. https://doi.org/10.1016/j.tust.2016.01.009
[4]	Farrokh E. et al., Study of various models for estimation of penetration rate of hard rock TBMs, Tunnelling and Underground Space Technology incorporating Trenchless Research, no. 30, pp. 110-123, 2012. https://doi.org/10.1016/j.tust.2012.02.012
[5]	Macias F., Hard rock tunnel boring: performance predictions and cutter life assessment, PhD thesis, Norwegian University of Science and Technology, Faculty of Engineering Science and Technology, Department of Civil and Transport Engineering, 2016.
[6]	Monthanopparat N & Tanchaisawat T, Advance Tunnel Boring Machine Performance Prediction in Massive and Highly Fractured Granite: Integrating Innovative Deep Learning and Block Model Technique, Geotechnical Engineering Journal of the SEAGs & AGSSEA, vol. 55, no. 3, pp. 26-34, 2024. https://doi.org/10.14456/seagj.2024.17
[7]	Koopialipoor M et al., Development of a new hybrid ANN for solving a geotechnical problem related to tunnel boring machine performance, Engineering with Computers, no. 36, pp. 345-357, 2020. https://doi.org/10.1007/s00366-019-00701-8
[8]	Poniewierski J, Block Model Knowledge for Mining Engineers - An Introduction, ACADEMIA, 2019.
[9]	AACE, TCM Framwork: 7.3 - Cost Estimating and Budgeting, in COST ESTIMATE CLASSIFICATION SYSTEM - AS APPLIED IN ENGINEERING, PROCUREMENT, AND CONSTRUCTION FOR THE PROCESS INDUSTRIES, 2005.
[10]	Rostami J, Development of a force estimation model for rock fragmentation with disc cutters through theoretical modeling and physical measurement of crushed zone pressure, PhD thesis dissertation, Colorado School of Mines, Golden, Colorado, USA, 1997.
[11]	Yagiz S, Development of rock fracture and brittleness indiced to qualify the effects of rock mass fracture and toughness in the CSM model basic penetration for hard rock tunnelling machines, thesis, Colorado School of Mines, Golden, Colorado, USA, 2002.
[12]	Yagiz S, TBM performance prediction based on rock properties, EUROCK, 2006. https://doi.org/10.1201/9781439833469.ch97
[13]	Bruland A, Vol. 7: The Boring Process, in Hard rock tunnel boring, Trondheim, Norway, Norwegian University of Science and Technology, Department of Building and Construction Engineering, 2000.
[14]	Gerhing K, Leistungs- und Verschleiβprognose im maschinellen Tunnelbau, Felsbau Magazin,no. 13, 1995.
[15]	Wilfing L, The influence of geotechnical parameters on penetration prediction in TBM tunelling in hard rock, PhD, Technische Universität München, 2016.
[16]	Kaewkongkaew K. et al., Geologcal model of Mae Tang-Mae Ngad Diversion tunnel project, Northern Thailand, Open Journal of Geology, no. 3, pp. 340-351, 2013. https://doi.org/10.4236/ojg.2013.35039
[17]	Hewamalage H et al., Forecast evaluation for data scientists: common pitfalls and best practices, Data Mining and Knowledge Discovery, vol. 37, pp. 788-832, 2023. https://doi.org/10.1007/s10618-022-00894-5
[18]	Hochreiter S & Schmidhuber J, Long Short-Term Memory, Neural Computation, vol. 9, no. 8, pp. 1735-1780, 1997. https://doi.org/10.1162/neco.1997.9.8.1735
[19]	Cho K et al., Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation, in Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), Association for Computational Linguistics, 2014, pp. 1724-1734. https://doi.org/10.3115/v1/D14-1179
[20]	Bai S et al., An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling, 2018. https://doi.org/10.48550/arXiv.1803.01271

Cite This Article

Plain Text BibTeX RIS

APA Style

Monthanopparat, N., Tanchaisawat, T. (2026). Rolling Window Deep Learning for Hard-rock TBM Penetration Rate Prediction. American Journal of Mechanics and Applications, 13(1), 10-21. https://doi.org/10.11648/j.ajma.20261301.12

Copy | Download

ACS Style

Monthanopparat, N.; Tanchaisawat, T. Rolling Window Deep Learning for Hard-rock TBM Penetration Rate Prediction. Am. J. Mech. Appl. 2026, 13(1), 10-21. doi: 10.11648/j.ajma.20261301.12

Copy | Download

AMA Style

Monthanopparat N, Tanchaisawat T. Rolling Window Deep Learning for Hard-rock TBM Penetration Rate Prediction. Am J Mech Appl. 2026;13(1):10-21. doi: 10.11648/j.ajma.20261301.12

Copy | Download

@article{10.11648/j.ajma.20261301.12,
  author = {Nantapol Monthanopparat and Tawatchai Tanchaisawat},
  title = {Rolling Window Deep Learning for Hard-rock TBM Penetration Rate Prediction},
  journal = {American Journal of Mechanics and Applications},
  volume = {13},
  number = {1},
  pages = {10-21},
  doi = {10.11648/j.ajma.20261301.12},
  url = {https://doi.org/10.11648/j.ajma.20261301.12},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ajma.20261301.12},
  abstract = {Penetration rate prediction for hard rock tunnel boring machines (TBMs) remains challenging because rock mass conditions and machine operating regimes vary continuously along the tunnel alignment. Conventional static prediction models may not adequately represent such non-stationary construction conditions, particularly when geotechnical measurements are incomplete or obtained with delay during excavation. This study aims to develop a construction-phase prediction framework for ring-scale TBM penetration rate in a granite tunnel drive by integrating geotechnical data completion, sequence deep learning, and rolling-window model evaluation. A dataset of 1,000 consecutive rings was compiled, including boring-only penetration rate, thrust, torque, cutterhead rotational speed, rock mass type, and uniaxial compressive strength (UCS). Missing UCS measurements were completed using inverse distance weighting within a block model representation, resulting in estimated UCS values of approximately 33–177 MPa for rock masses dominated by massive and fractured granite. Three sequence deep learning models, namely long short-term memory (LSTM), gated recurrent unit (GRU), and temporal convolutional network (TCN), were evaluated using root mean square error (RMSE), mean absolute error (MAE), and a symmetric ±10% tolerance band adapted from accuracy-band concepts used in AACE-based project controls. The proposed rolling protocol used 100-ring validation and 100-ring test blocks to assess predictive performance under changing ground conditions. The results show that the optimized GRU model provided the most robust overall performance, achieving a mean test RMSE of approximately 0.229 m/h and a mean within-band compliance of approximately 54% across rolling folds. These findings indicate that rolling-window sequence learning can provide a practical and adaptable framework for construction-phase TBM performance prediction under evolving geological and operational conditions.},
 year = {2026}
}

Copy | Download

TY  - JOUR
T1  - Rolling Window Deep Learning for Hard-rock TBM Penetration Rate Prediction
AU  - Nantapol Monthanopparat
AU  - Tawatchai Tanchaisawat
Y1  - 2026/05/26
PY  - 2026
N1  - https://doi.org/10.11648/j.ajma.20261301.12
DO  - 10.11648/j.ajma.20261301.12
T2  - American Journal of Mechanics and Applications
JF  - American Journal of Mechanics and Applications
JO  - American Journal of Mechanics and Applications
SP  - 10
EP  - 21
PB  - Science Publishing Group
SN  - 2376-6131
UR  - https://doi.org/10.11648/j.ajma.20261301.12
AB  - Penetration rate prediction for hard rock tunnel boring machines (TBMs) remains challenging because rock mass conditions and machine operating regimes vary continuously along the tunnel alignment. Conventional static prediction models may not adequately represent such non-stationary construction conditions, particularly when geotechnical measurements are incomplete or obtained with delay during excavation. This study aims to develop a construction-phase prediction framework for ring-scale TBM penetration rate in a granite tunnel drive by integrating geotechnical data completion, sequence deep learning, and rolling-window model evaluation. A dataset of 1,000 consecutive rings was compiled, including boring-only penetration rate, thrust, torque, cutterhead rotational speed, rock mass type, and uniaxial compressive strength (UCS). Missing UCS measurements were completed using inverse distance weighting within a block model representation, resulting in estimated UCS values of approximately 33–177 MPa for rock masses dominated by massive and fractured granite. Three sequence deep learning models, namely long short-term memory (LSTM), gated recurrent unit (GRU), and temporal convolutional network (TCN), were evaluated using root mean square error (RMSE), mean absolute error (MAE), and a symmetric ±10% tolerance band adapted from accuracy-band concepts used in AACE-based project controls. The proposed rolling protocol used 100-ring validation and 100-ring test blocks to assess predictive performance under changing ground conditions. The results show that the optimized GRU model provided the most robust overall performance, achieving a mean test RMSE of approximately 0.229 m/h and a mean within-band compliance of approximately 54% across rolling folds. These findings indicate that rolling-window sequence learning can provide a practical and adaptable framework for construction-phase TBM performance prediction under evolving geological and operational conditions.
VL  - 13
IS  - 1
ER  -

Copy | Download

Author Information

Nantapol Monthanopparat

Department of Civil Engineering, Chiang Mai University, Chiang Mai, Thailand

Contact Email

http://orcid.org/0009-0005-4399-2761
Tawatchai Tanchaisawat

Department of Civil Engineering, Chiang Mai University, Chiang Mai, Thailand

Download PDF

Submit an Article

Sections

Plain Text BibTeX RIS

APA Style

Monthanopparat, N., Tanchaisawat, T. (2026). Rolling Window Deep Learning for Hard-rock TBM Penetration Rate Prediction. American Journal of Mechanics and Applications, 13(1), 10-21. https://doi.org/10.11648/j.ajma.20261301.12

Copy | Download

ACS Style

Monthanopparat, N.; Tanchaisawat, T. Rolling Window Deep Learning for Hard-rock TBM Penetration Rate Prediction. Am. J. Mech. Appl. 2026, 13(1), 10-21. doi: 10.11648/j.ajma.20261301.12

Copy | Download

AMA Style

Monthanopparat N, Tanchaisawat T. Rolling Window Deep Learning for Hard-rock TBM Penetration Rate Prediction. Am J Mech Appl. 2026;13(1):10-21. doi: 10.11648/j.ajma.20261301.12

Copy | Download

@article{10.11648/j.ajma.20261301.12,
  author = {Nantapol Monthanopparat and Tawatchai Tanchaisawat},
  title = {Rolling Window Deep Learning for Hard-rock TBM Penetration Rate Prediction},
  journal = {American Journal of Mechanics and Applications},
  volume = {13},
  number = {1},
  pages = {10-21},
  doi = {10.11648/j.ajma.20261301.12},
  url = {https://doi.org/10.11648/j.ajma.20261301.12},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ajma.20261301.12},
  abstract = {Penetration rate prediction for hard rock tunnel boring machines (TBMs) remains challenging because rock mass conditions and machine operating regimes vary continuously along the tunnel alignment. Conventional static prediction models may not adequately represent such non-stationary construction conditions, particularly when geotechnical measurements are incomplete or obtained with delay during excavation. This study aims to develop a construction-phase prediction framework for ring-scale TBM penetration rate in a granite tunnel drive by integrating geotechnical data completion, sequence deep learning, and rolling-window model evaluation. A dataset of 1,000 consecutive rings was compiled, including boring-only penetration rate, thrust, torque, cutterhead rotational speed, rock mass type, and uniaxial compressive strength (UCS). Missing UCS measurements were completed using inverse distance weighting within a block model representation, resulting in estimated UCS values of approximately 33–177 MPa for rock masses dominated by massive and fractured granite. Three sequence deep learning models, namely long short-term memory (LSTM), gated recurrent unit (GRU), and temporal convolutional network (TCN), were evaluated using root mean square error (RMSE), mean absolute error (MAE), and a symmetric ±10% tolerance band adapted from accuracy-band concepts used in AACE-based project controls. The proposed rolling protocol used 100-ring validation and 100-ring test blocks to assess predictive performance under changing ground conditions. The results show that the optimized GRU model provided the most robust overall performance, achieving a mean test RMSE of approximately 0.229 m/h and a mean within-band compliance of approximately 54% across rolling folds. These findings indicate that rolling-window sequence learning can provide a practical and adaptable framework for construction-phase TBM performance prediction under evolving geological and operational conditions.},
 year = {2026}
}

Copy | Download

TY  - JOUR
T1  - Rolling Window Deep Learning for Hard-rock TBM Penetration Rate Prediction
AU  - Nantapol Monthanopparat
AU  - Tawatchai Tanchaisawat
Y1  - 2026/05/26
PY  - 2026
N1  - https://doi.org/10.11648/j.ajma.20261301.12
DO  - 10.11648/j.ajma.20261301.12
T2  - American Journal of Mechanics and Applications
JF  - American Journal of Mechanics and Applications
JO  - American Journal of Mechanics and Applications
SP  - 10
EP  - 21
PB  - Science Publishing Group
SN  - 2376-6131
UR  - https://doi.org/10.11648/j.ajma.20261301.12
AB  - Penetration rate prediction for hard rock tunnel boring machines (TBMs) remains challenging because rock mass conditions and machine operating regimes vary continuously along the tunnel alignment. Conventional static prediction models may not adequately represent such non-stationary construction conditions, particularly when geotechnical measurements are incomplete or obtained with delay during excavation. This study aims to develop a construction-phase prediction framework for ring-scale TBM penetration rate in a granite tunnel drive by integrating geotechnical data completion, sequence deep learning, and rolling-window model evaluation. A dataset of 1,000 consecutive rings was compiled, including boring-only penetration rate, thrust, torque, cutterhead rotational speed, rock mass type, and uniaxial compressive strength (UCS). Missing UCS measurements were completed using inverse distance weighting within a block model representation, resulting in estimated UCS values of approximately 33–177 MPa for rock masses dominated by massive and fractured granite. Three sequence deep learning models, namely long short-term memory (LSTM), gated recurrent unit (GRU), and temporal convolutional network (TCN), were evaluated using root mean square error (RMSE), mean absolute error (MAE), and a symmetric ±10% tolerance band adapted from accuracy-band concepts used in AACE-based project controls. The proposed rolling protocol used 100-ring validation and 100-ring test blocks to assess predictive performance under changing ground conditions. The results show that the optimized GRU model provided the most robust overall performance, achieving a mean test RMSE of approximately 0.229 m/h and a mean within-band compliance of approximately 54% across rolling folds. These findings indicate that rolling-window sequence learning can provide a practical and adaptable framework for construction-phase TBM performance prediction under evolving geological and operational conditions.
VL  - 13
IS  - 1
ER  -

Copy | Download