| Machine Learning-Based Prediction of TBM Penetration Rate Using Geological and Operational Data from the Karaj Water Conveyance Tunnel |
| کد مقاله : 1038-NIEGC2025 |
| نویسندگان |
|
علی اصغر قائدی وانانی *1، Jafar Hassanpour2، Mehdi Alihkani3، Mohamad Eslami4 1دانش آموخته دکتری تخصصی زمین شناسی، دانشگاه تربیت مدرس 2Assistant Professor, Faculty of Geology, Basic Sciences, University of Tehran 3PhD student in Water Resources Engineering, Islamic Azad University, Shahrekord Branch 4Master's degree in Tectonic Geology, Faculty of Basic Sciences, Shiraz University |
| چکیده مقاله |
| This study employs the machine learning algorithms Multiple Linear Regression (LMR) and Artificial Neural Network (ANN) to develop predictive models for estimating the Penetration Rate (PR) of a Tunnel Boring Machine (TBM). A total of 161 actual data samples obtained from the excavation face of the Karaj Water Conveyance Tunnel Project (Sections 1 and 2, total length 22.7 km) were analyzed. The input parameters included geological features (Uniaxial Compressive Strength (UCS), Rock Quality Designation (RQD%), Geological Strength Index (GSI), and joint spacing (Sj)) and operational TBM parameters (including rotation speed (RPM), normal cutter force (Fn), and rolling force (Fr)). The results indicated that UCS, RQD, GSI, Sj, RPM, Fn, and Fr were the most influential factors affecting TBM penetration. Among the models, ANN achieved the highest accuracy (R² = 0.93, MAE = 0.02), followed by LMR (R² = 0.76, MAE = 0.07), which exhibited lower predictive capability. The superior performance of ANNs is attributed to their ability to capture complex nonlinear interactions between geological and operational parameters. The findings highlight the strong potential of ANNs for accurate TBM performance prediction and their practical value in optimizing operational parameters, enhancing planning efficiency, and reducing tunneling costs. |
| کلیدواژه ها |
| KEYWORDS: TUNNEL BORING MACHINE, PENETRATION RATE, MACHINE LEARNING, KARAJ WATER CONVEYANCE TUNNEL. |
| وضعیت: پذیرفته شده برای ارسال فایل های ارائه پوستر |