Application of Neural Network Models in Predicting TBM Performance and Managing Excavation Costs: A Case Study of the Mechanized Excavation of the Twin Tunnels of Tabriz Urban Railway Line 1

Masoud  Haghkish; Seyed Azim Hosseini; Ghodratollah  Mohammadi; Abbas  Akbarpour Nikghalb

Authors

Masoud Haghkish Ph.D. Student, Department of Civil Engineering, ST.C., Islamic Azad University, Tehran, Iran Author https://orcid.org/0009-0001-1353-6387
Seyed Azim Hosseini Department of Civil Engineering, ST.C., Islamic Azad University, Tehran, Iran Corresponding author https://orcid.org/0000-0002-3691-0204
Ghodratollah Mohammadi Department of Mining Engineering, ST.C., Islamic Azad University, Tehran, Iran Author https://orcid.org/0000-0002-6769-5930
Abbas Akbarpour Nikghalb Department of Civil Engineering, ST.C., Islamic Azad University, Tehran, Iran Author https://orcid.org/0000-0002-4219-681X

Keywords:

metro, mechanized tunnel excavation, failure, artificial neural network, , locomotive, linear regression, Pareto analysis

Abstract

During mechanized tunnel excavation operations, support locomotives refer to a fleet assigned such tasks as transporting foam and grout—the injection slurry used to fill the annular void between the excavated tunnel surface and the rear face of the segmental lining—delivering precast reinforced-concrete segments into the tunnel and to the excavation face, and removing excavated spoil from the tunnel. Accordingly, this fleet plays a key role and has a direct impact on the tunnel excavation phase. This study investigated and predicted the performance of the support locomotive fleet used in the tunnel boring machine operations of Tabriz Urban Railway Line 1. In the first stage, each locomotive was divided into four principal systems: mechanical (engine), electrical, hydraulic, and pneumatic. In the subsequent stages, each of these four systems was further decomposed into smaller-scale subsystems and components. Field studies were then conducted, and failure data from five locomotives were collected using methods endorsed by internationally recognized project management standards, including expert judgment. Following data integration, more than 30,000 failure-related observations were obtained. Pareto analysis, linear regression, and artificial neural networks were examined as statistical and intelligent analytical methods. The findings indicated that the neural network method generated predictions that closely approximated the actual values, with low error and high accuracy.

References

[1] L. Yang, Y. Zhu, M. Chatzimichailidou, and X. Liu, "Assessing Human Emotional Responses to the Design of Public Spaces around Subway Stations: A Human Factors Research," Urban Design International, vol. 28, no. 4, pp. 285-303, 2023, doi: 10.1057/s41289-023-00219-y.

[2] H. Haghkish, H. Chakeri, and A. Mousavi, Underground Structures in Subways. 2020.

[3] A. G. Benardos and D. C. Kaliampakos, "Modelling TBM Performance with Artificial Neural Networks," Tunnelling and Underground Space Technology, vol. 19, pp. 597-605, 2004.

[4] P. Blanco-Valcarcel, A. Castro, and M. E. Vázquez-Méndez, "On the numerical computation of earthworks in road and railway projects," Results in Engineering, vol. 27, p. 106824, 2025, doi: 10.1016/J.RINENG.2025.106824.

[5] I. Kaastra and M. Boyd, "Designing a Neural Network for Forecasting Financial and Economic Time Series," Neurocomputing, vol. 10, pp. 215-236, 1996.

[6] I. Kanellopoulos and G. G. Wilkinson, "Strategies and Best Practice for Neural Network Image Classification," International Journal of Remote Sensing, vol. 18, pp. 711-725, 1997.

[7] H. B. Demuth and M. H. Beale, Neural Network Toolbox for Use with MATLAB: User's Guide. USA: The MathWorks, Inc., 2004.

[8] S. Yagiz, C. Gokceoglu, E. Sezer, and S. Iplikci, "Application of Two Non-Linear Prediction Tools to the Estimation of Tunnel Boring Machine Performance," Engineering Applications of Artificial Intelligence, vol. 22, no. 4-5, pp. 808-814, 2009.

[9] J. Gholamnejad and N. Tayarani, "Application of Artificial Neural Networks to the Prediction of Tunnel Boring Machine Penetration Rate," Mining Science and Technology (China), vol. 20, pp. 727-733, 2010.

[10] H. Nejati and M. Ahmadi, "Estimation of Tunnel Boring Machine (TBM) Efficiency Using Artificial Neural Network," Environmental Education and Sustainable Development, vol. 12, pp. 63-72, 2010.

[11] S. R. Torabi, H. Shirazi, H. Hajali, and M. Monjezi, "Study of the Influence of Geotechnical Parameters on the TBM Performance in Tehran-Shomal Highway Project Using ANN and SPSS," Arabian Journal of Geosciences, vol. 6, pp. 1215-1227, 2013.

[12] M. Zahiri, K. Gashtasbi, J. Khademi, and K. Ahangari, "Providing an Experimental Model for Predicting the Consumption Rate of Cutters in Full-Face Tunnel Boring Machines (TBMs)," Iranian Journal of Rock Mechanics, vol. 2, pp. 1-9, 2018.

[13] Schoma, "Technical Report on the Operation of Locomotives," 2016.

[14] F. Aslani, M. Ayat, S. Goharimehr, and M. Mohammadi, "The effect of meaning in life on job performance with the mediating role of extrinsic and intrinsic motivation in employees of Isfahan Urban Railway Organization," Quarterly Journal of Work Psychology, vol. 4, no. 1, pp. 9-20, 2025, doi: 10.30473/jpsy.2025.75139.1096.

[15] M. Lin, C. Chen, and D. Ji, "Mathematical Modelling and Dynamic Characteristics Analysis of Railway Stabilizing Operation on Different Status Ballasted Track Railways," Proceedings of the Institution of Mechanical Engineers Part C Journal of Mechanical Engineering Science, 2025, doi: 10.1177/09544062251365538.

[16] S. Roy, "FedGIS-Water: A Federated Learning-Enhanced GIS Framework for Circular Water Infrastructure Assessment in Railway Stations," 2026, doi: 10.21203/rs.3.rs-8740219/v1.