Scholarlink

<p>Blast-induced ground vibrations significantly affect surrounding structures, leading to structural damage and safety concerns. Therefore, accurately predicting and evaluating blasting vibrations is essential for mitigating these adverse effects. This study developed two machine learning (ML) models—Support Vector Regression (SVR) and Artificial Neural Network (ANN)—to predict blast-induced ground vibrations at an Open Pit Mine, situated in Pakistan, using data from 90 blast events. The dataset includes six blast design parameters as model inputs: hole depth (HD), burden (B), stemming length (SL), spacing (S), blast-to-monitoring distance (D), and maximum explosive charge per delay (MCPD), with peak particle velocity (PPV) as the output. Model performance was evaluated using statistical metrics, including the coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE). While the ANN model demonstrated satisfactory predictive capability, the SVR model outperformed ANN, achieving an excellent R² value of 0.91 and lower RMSE (0.029) and MAE (0.023) for the test dataset. Sensitivity analysis indicated that burden (B) had the greatest influence, while distance (D) had the least influence on blast-induced ground vibration. Therefore, the SVR model proves to be a reliable tool for predicting blasting vibrations in open-pit mines and other engineering applications with similar blasting and geotechnical characteristics.</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

Machine Learning-Based Prediction of Blast-Induced Ground Vibration in Open-Pit Mining

  • 期刊论文

  • 发表日期：

    2025年5月9日

查看全文
  • Sami Ullah,
  • Gaofeng Ren,
  • Yongxiang Ge,
  • Yewuhalashet Fissha,
  • Eric Munene Kinyua,
  • Luwei Zhang

摘要

Blast-induced ground vibrations significantly affect surrounding structures, leading to structural damage and safety concerns. Therefore, accurately predicting and evaluating blasting vibrations is essential for mitigating these adverse effects. This study developed two machine learning (ML) models—Support Vector Regression (SVR) and Artificial Neural Network (ANN)—to predict blast-induced ground vibrations at an Open Pit Mine, situated in Pakistan, using data from 90 blast events. The dataset includes six blast design parameters as model inputs: hole depth (HD), burden (B), stemming length (SL), spacing (S), blast-to-monitoring distance (D), and maximum explosive charge per delay (MCPD), with peak particle velocity (PPV) as the output. Model performance was evaluated using statistical metrics, including the coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE). While the ANN model demonstrated satisfactory predictive capability, the SVR model outperformed ANN, achieving an excellent R² value of 0.91 and lower RMSE (0.029) and MAE (0.023) for the test dataset. Sensitivity analysis indicated that burden (B) had the greatest influence, while distance (D) had the least influence on blast-induced ground vibration. Therefore, the SVR model proves to be a reliable tool for predicting blasting vibrations in open-pit mines and other engineering applications with similar blasting and geotechnical characteristics.