Scholarlink

<p>A rapid and accurate prediction of the spontaneous combustion of coal is required to ensure safety in coal mines. To this end, this paper presents an intuitive prediction model established by combining the improved particle swarm optimization (IPSO) algorithm with the backpropagation neural network (BPNN), referred to as the IPSO–BPNN model. First, IPSO was enhanced by introducing logistic chaotic mapping and Lévy flight to improve its performance. Subsequently, IPSO was used to optimize the BPNN. A performance evaluation of the proposed models was conducted using common metrics, including the accuracy, precision, recall, and F1 score. The experimental results demonstrated that the accuracy, precision, recall, and F1 score of IPSO-BPNN model were 0.95, 0.97, 0.96, and 0.96, respectively. Compared to other models, the accuracy of the IPSO-BPNN model exhibited improvements of 0.25, 0.10, 0.10, 0.10, and 0.10, while precision increased by 0.18, 0.08, 0.08, 0.08, and 0.07, recall improved by 0.17, 0.03, 0.17, 0.05, and 0.17, and the F1 score rose by 0.17, 0.05, 0.12, 0.06, and 0.12. These results unequivocally indicate that the IPSO-BPNN model outperformed all comparative models in coal spontaneous combustion risk prediction. The proposed model was empirically demonstrated to have a high accuracy in predicting the spontaneous combustion risk of coal. A dynamic warning model established using the IPSO–BPNN was applied to other coal seam from the mine, and it showed a superior predictive performance over the existing models, further confirming its generality and stability.</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

Dynamic Early Warning Model for Coal Spontaneous Combustion Based on Machine Learning

  • 期刊论文

  • 发表日期：

    2026年3月5日

查看全文
  • Gang Bai,
  • Ran Liang,
  • Wei Wang,
  • Yun Qi,
  • Xiaowen Zhang,
  • Xueming Li,
  • Xun Zhang

摘要

A rapid and accurate prediction of the spontaneous combustion of coal is required to ensure safety in coal mines. To this end, this paper presents an intuitive prediction model established by combining the improved particle swarm optimization (IPSO) algorithm with the backpropagation neural network (BPNN), referred to as the IPSO–BPNN model. First, IPSO was enhanced by introducing logistic chaotic mapping and Lévy flight to improve its performance. Subsequently, IPSO was used to optimize the BPNN. A performance evaluation of the proposed models was conducted using common metrics, including the accuracy, precision, recall, and F1 score. The experimental results demonstrated that the accuracy, precision, recall, and F1 score of IPSO-BPNN model were 0.95, 0.97, 0.96, and 0.96, respectively. Compared to other models, the accuracy of the IPSO-BPNN model exhibited improvements of 0.25, 0.10, 0.10, 0.10, and 0.10, while precision increased by 0.18, 0.08, 0.08, 0.08, and 0.07, recall improved by 0.17, 0.03, 0.17, 0.05, and 0.17, and the F1 score rose by 0.17, 0.05, 0.12, 0.06, and 0.12. These results unequivocally indicate that the IPSO-BPNN model outperformed all comparative models in coal spontaneous combustion risk prediction. The proposed model was empirically demonstrated to have a high accuracy in predicting the spontaneous combustion risk of coal. A dynamic warning model established using the IPSO–BPNN was applied to other coal seam from the mine, and it showed a superior predictive performance over the existing models, further confirming its generality and stability.