Scholarlink

Computer-assisted drilling rigs have been implemented in recent tunnel construction projects. They can realize automated drilling of charge holes as designed beforehand and contribute to labor-saving in drilling operations, in case the proper and reasonable blasting designs are provided. The authors, therefore, have been developing a blasting design method based on automatically collected Measurement While Drilling (MWD) data by the rigs which consist of positions and loads on drifters. Specifically, we statistically analyze the correlation of the MWD data obtained in some tunneling projects with blasting results including explosive volume used and over-excavation profile measured by 3D laser scanner. The estimation methods with machine learning (ML) algorithms are then constructed in which the required explosive volume and over-excavation volume could be predicted. This paper reports two case studies: an explosive volume estimation based on the MWD data obtained by XE3C (Epiroc) and an over-excavation volume estimation based on the MWD data collected by DT1131i (Sandvik).

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Study on Machine Learning Algorithm for Optimal Tunnel Blasting Design

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Kazuo Sakai,
  • Shuntaro Miyanaga,
  • Masahito Yamagami,
  • Alla Sapronova,
  • Abdallah Ahmed Fouad Elsayed Soliman

摘要

Computer-assisted drilling rigs have been implemented in recent tunnel construction projects. They can realize automated drilling of charge holes as designed beforehand and contribute to labor-saving in drilling operations, in case the proper and reasonable blasting designs are provided. The authors, therefore, have been developing a blasting design method based on automatically collected Measurement While Drilling (MWD) data by the rigs which consist of positions and loads on drifters. Specifically, we statistically analyze the correlation of the MWD data obtained in some tunneling projects with blasting results including explosive volume used and over-excavation profile measured by 3D laser scanner. The estimation methods with machine learning (ML) algorithms are then constructed in which the required explosive volume and over-excavation volume could be predicted. This paper reports two case studies: an explosive volume estimation based on the MWD data obtained by XE3C (Epiroc) and an over-excavation volume estimation based on the MWD data collected by DT1131i (Sandvik).