Scholarlink

<p>Blasting operations in mining significantly impact subsequent processes such as loading, transportation, crushing, and pulverizing. It plays a crucial role in the mining efficiency and economic benefits of open-pit mines. During blasting, part of the blast energy is converted into the kinetic energy of rock fragments, causing the ore and waste distribution in the muckpile to differ greatly from the geological model. It leads to ore loss and dilution during shovel loading. Precise evaluation of blast-induced rock movement and effective delineation of ore and waste boundaries after blasting are critical for enhancing recovery rates. Based on a database from the Husab Uranium Mine in Namibia, the Mirador Copper Mine in Ecuador and the Phoenix Mine in the USA, this study developed three hybrid models based on the light gradient-boosting machine (LGBM) with genetic algorithm, gray wolf optimizer, and Kepler optimization algorithm (KOA) for predicting blast-induced rock movement, and compared their performance with other prevalent models. Among them, the KOA-LGBM model demonstrated the best predictive performance, achieving coefficients of determination (<i>R</i><sup>2</sup>) of 0.874 and 0.852 for the training and testing datasets, respectively, identifying it as the optimal prediction model. Moreover, an interpretability analysis was performed to evaluate how input variables influence blast-induced rock movement, revealing that the depth coefficient and powder factor were the primary contributing factors. This study provides reliable prediction techniques and promotes broader integration of artificial intelligence into mining engineering.</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

Evaluation and Interpretation of Blast-Induced Rock Movement During Open-Pit Bench Blasting: Use of a LGBM Hybrid Model

  • 期刊论文

  • 发表日期：

    2026年5月7日

查看全文
  • Zhi Yu,
  • Linfeng Du,
  • Xiaofeng Huo,
  • Song He,
  • Jianhua Hu,
  • Jian Zhou

摘要

Blasting operations in mining significantly impact subsequent processes such as loading, transportation, crushing, and pulverizing. It plays a crucial role in the mining efficiency and economic benefits of open-pit mines. During blasting, part of the blast energy is converted into the kinetic energy of rock fragments, causing the ore and waste distribution in the muckpile to differ greatly from the geological model. It leads to ore loss and dilution during shovel loading. Precise evaluation of blast-induced rock movement and effective delineation of ore and waste boundaries after blasting are critical for enhancing recovery rates. Based on a database from the Husab Uranium Mine in Namibia, the Mirador Copper Mine in Ecuador and the Phoenix Mine in the USA, this study developed three hybrid models based on the light gradient-boosting machine (LGBM) with genetic algorithm, gray wolf optimizer, and Kepler optimization algorithm (KOA) for predicting blast-induced rock movement, and compared their performance with other prevalent models. Among them, the KOA-LGBM model demonstrated the best predictive performance, achieving coefficients of determination (R2) of 0.874 and 0.852 for the training and testing datasets, respectively, identifying it as the optimal prediction model. Moreover, an interpretability analysis was performed to evaluate how input variables influence blast-induced rock movement, revealing that the depth coefficient and powder factor were the primary contributing factors. This study provides reliable prediction techniques and promotes broader integration of artificial intelligence into mining engineering.