Scholarlink

<p>Understanding the development of water-conducting fissure zones in the Huanglong Jurassic coalfield and accurately predicting the height of these fissures are crucial to preventing and controlling damage to the overlying sandstone aquifer of the Luohe Formation. To develop a predictive model applicable to the Huanglong Jurassic coalfield, data from measurements of 27 water-conducting fissure zones in the coalfield were used as samples, and the coal seam burial depth, coal seam mining thickness, and oblique length of the working face were used as training indicators. The whale optimisation algorithm (WOA), back-propagation neural network (BPNN), and AdaBoost algorithm were combined to develop the AdaBoost–WOA–BPNN model for predicting the height of water-conducting fissure zones. The accuracies of the models were compared, and the height of the water-conducting fissure zone in the 4105 working face of the Wenjiapo coal mine in the Binchang mining area of the Huanglong Coalfield was predicted. The AdaBoost–WOA–BPNN model outperformed the other models in terms of error, predictive accuracy, and applicability. Moreover, this predictive accuracy met the requirements of engineering practice. The results of this study provide a valuable reference for predicting the height of the water-conducting fissure zones and for preventing and controlling roof water hazards.</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

Predicting the Height of Water-Conducting Fissure Zones in a Jurassic Coalfield Based on AdaBoost-WOA-BPNN

  • 期刊论文

  • 发表日期：

    2025年2月28日

查看全文
  • Enke Hou,
  • Meng Bi,
  • Tianwen Long,
  • Xiaoshen Xie,
  • Pengfei Hou,
  • Qianlong Li

摘要

Understanding the development of water-conducting fissure zones in the Huanglong Jurassic coalfield and accurately predicting the height of these fissures are crucial to preventing and controlling damage to the overlying sandstone aquifer of the Luohe Formation. To develop a predictive model applicable to the Huanglong Jurassic coalfield, data from measurements of 27 water-conducting fissure zones in the coalfield were used as samples, and the coal seam burial depth, coal seam mining thickness, and oblique length of the working face were used as training indicators. The whale optimisation algorithm (WOA), back-propagation neural network (BPNN), and AdaBoost algorithm were combined to develop the AdaBoost–WOA–BPNN model for predicting the height of water-conducting fissure zones. The accuracies of the models were compared, and the height of the water-conducting fissure zone in the 4105 working face of the Wenjiapo coal mine in the Binchang mining area of the Huanglong Coalfield was predicted. The AdaBoost–WOA–BPNN model outperformed the other models in terms of error, predictive accuracy, and applicability. Moreover, this predictive accuracy met the requirements of engineering practice. The results of this study provide a valuable reference for predicting the height of the water-conducting fissure zones and for preventing and controlling roof water hazards.