几种模态分解方法在矿井涌水量预测中的研究  

Research of Several Modal Decomposition Methods in Predicting Mine Water Inflow

作　　者：李祥鲁[1] 孙文斌[1] 赵吉园 Li Xianglu;Sun Wenbin;Zhao Jiyuan(College of Energy and Mining Engineering,Shandong University of Science and Technology,Qingdao 266590,China)

机构地区：[1]山东科技大学能源与矿业工程学院,山东青岛266590

出　　处：《山东煤炭科技》2025年第4期152-157,共6页Shandong Coal Science and Technology

摘　　要：为提高矿井涌水量的预测准确性,基于矿井实测月度涌水量数据,探讨模态分解方法对时间序列数据的影响,构建经验模态分解(EMD)与长短期记忆神经网络(LSTM)结合的预测模型,并应用于矿井涌水量预测,提出自适应噪声完备集合经验模态分解(CEEMDAN)和变分模态分解(VMD)两种新型模型,进行对比分析。结果表明,VMD-LSTM模型的预测准确率最高,达到94.14%,有效提高了矿井涌水量预测的精度。研究为矿井涌水量的高效预测提供了新的技术路径。To improve the accuracy of predicting mine water inflow,based on the actual measured monthly water inflow data of the mine,the influence of modal decomposition method on time series data is explored.A prediction model combining empirical modal decomposition(EMD)and long short term memory neural network(LSTM)is constructed and applied to mine water inflow prediction.Two type of new models,adaptive noise complete set empirical modal decomposition(CEEMDAN)and variational modal decomposition(VMD),are proposed for comparative analysis.The results show that the VMD-LSTM model has the highest prediction accuracy,reaching 94.14%,effectively improving the accuracy of predicting mine water inflow.The research provides a new technical path for efficient prediction of mine water inflow.

关 键 词：矿井涌水量 时间序列预测 模态分解方法 长短期记忆神经网络 

分 类 号：TD745[矿业工程—矿井通风与安全]