Scholarlink

Abstract <p>The deepest-level underground structures of various purpose are mines, tunnels, hydropower stations and underground research laboratories. The authors describe a possible classification of rockbursts by their initiation mechanism: ahead of a mined-out space, in a pillar, nearby a fault. The authors review the common applied systems of rockburst hazard prediction and select continuous measurement systems which produce data suitable for machine processing. The geodynamic situation prediction at a mineral deposit using artificial neural networks is described. The comparative test of machine learning methods for the analysis of geodynamic phenomena is carried out. The structure of artificial neural networks for the prediction of geodynamic phenomena and stability of underground openings is described. The seismic events in the Sheregesh Mine are selected for the analysis. It is found that different models accurately determine clusters of seismic events. k-means clustering produces the best results (97.92%).</p>

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

Artificial Intelligence in Prediction of Geodynamic Phenomena in Rock Masses

  • 期刊论文

  • 发表日期：

    2025年6月1日

查看全文
  • A. I. Konurin,
  • D. V. Orlov

摘要

Abstract

The deepest-level underground structures of various purpose are mines, tunnels, hydropower stations and underground research laboratories. The authors describe a possible classification of rockbursts by their initiation mechanism: ahead of a mined-out space, in a pillar, nearby a fault. The authors review the common applied systems of rockburst hazard prediction and select continuous measurement systems which produce data suitable for machine processing. The geodynamic situation prediction at a mineral deposit using artificial neural networks is described. The comparative test of machine learning methods for the analysis of geodynamic phenomena is carried out. The structure of artificial neural networks for the prediction of geodynamic phenomena and stability of underground openings is described. The seismic events in the Sheregesh Mine are selected for the analysis. It is found that different models accurately determine clusters of seismic events. k-means clustering produces the best results (97.92%).