Chen Juntao, Fan Mingjin
College of Energy and Mining Engineering, Shandong University of Science and Technology, Qingdao, 266590, China.
National Demonstration Centre for Experimental Mining Engineering Education, Shandong University of Science and Technology, Qingdao, 266590, China.
Sci Rep. 2025 Feb 5;15(1):4378. doi: 10.1038/s41598-024-82580-8.
Water damage accidents occur frequently in mines in China, and accurate prediction of incoming water has become an important guarantee for the safe and efficient mining of coal resources. To improve the accuracy of mine water prediction, this paper proposes the VMD-iCHOA-GRU mine water prediction model by selecting and improving it according to the previous research results in decomposition method, time series prediction model and optimization algorithm. After processing the raw data and setting the model parameters, MAE, RMSE, MAPE and R are selected as the evaluation indexes of prediction accuracy, and VMD-GRU model, iCHOA-GRU model, CHOA-GRU model and GRU model are selected as the comparison models to validate the advantages of the VMD-iCHOA-GRU model in the prediction of mine inrush water. The results show that the VMD-iCHOA-GRU model has the best prediction effect on the trend of water inflow, with the evaluation index values of 0.00862, 0.01059, 0.02189%, 0.87079, respectively, and with the smallest MAE, RMSE, MAPE, and the largest R, and the highest prediction accuracy of the VMD-iCHOA-GRU model.
我国煤矿水害事故频发，准确预测突水量已成为煤炭资源安全高效开采的重要保障。为提高矿井水害预测精度，本文在前人对分解方法、时间序列预测模型和优化算法的研究成果基础上进行筛选和改进，提出了VMD-iCHOA-GRU矿井水害预测模型。对原始数据进行处理并设置模型参数后，选取平均绝对误差（MAE）、均方根误差（RMSE）、平均绝对百分比误差（MAPE）和相关系数（R）作为预测精度评价指标，并选取VMD-GRU模型、iCHOA-GRU模型、CHOA-GRU模型和GRU模型作为对比模型，验证VMD-iCHOA-GRU模型在矿井突水预测中的优势。结果表明，VMD-iCHOA-GRU模型对涌水量变化趋势的预测效果最佳，其评价指标值分别为0.00862、0.01059、0.02189%、0.87079，MAE、RMSE、MAPE最小，R最大，预测精度最高。
