Document Type : Original Research Paper

Authors

• Sasan Heydari ¹
• Seyed Hadi Hoseinie ²
• Raheb Bagherpour ¹
• Sasan Ghorbani ³

¹ Department of Mining Engineering, Isfahan University of Technology, Isfahan, Iran,

² Department of Mining Engineering, Isfahan University of Technology, Isfahan, Iran

³ Department of Mining Engineering, Hamedan University of Technology, Hamedan, Iran

Abstract

This study aims to predict the penetration rate (ROP) of jumbo drills in underground mining operations, focusing on drilling and blasting activities in challenging rock environments where tunnel boring machines are impractical. Accurate ROP predictions can help optimize drill performance by adjusting operational parameters. In this study, deep learning models, including Fully Connected Deep Neural Networks (FCDNN), Deep Random Forest (DRF), and Deep Support Vector Regression (DSVR), were trained using a Genetic Algorithm (GA). The dataset comprised 737 samples from various case studies, incorporating key drilling parameters and rock mass characteristics, with the Rock Mass Drillability Index (RDi) being a central variable. The performance of the models was assessed using metrics such as the determination coefficient (R²), variance accounted for (VAF), mean absolute error (MAE), mean absolute percentage error (MAPE), and root mean square error (RMSE). Results indicated that the FCDNN model demonstrated superior prediction accuracy and generalization capabilities compared to the DRF and DSVR models, highlighting its effectiveness in handling complex datasets and contributing to the optimization of drilling processes in underground mining.

Keywords

• Jumbo drill
• Rate of penetration
• FCDNN
• DRF
• DSVR

Main Subjects

• Underground Mining