1 College of Earth and Environmental Sciences, Anhui University of Science and Technology, Huainan 232001, China
2 Xieqiao Coal Mine, Huainan Mining Group, Fuyang 236221, China
Show less
文章历史+
Received
Published
Revised
2026-01-23
2026-10-18
2026-04-07
Online Date
2026-10-30
PDF (9249K)
摘要
采矿扰动条件下,精准预测煤矿区地下水位动态对矿山突水预测与水害防治具有重要的指导意义。现有地下水位预测方法存在精度低、准确性差等问题,难以满足矿区安全生产需求。以淮南谢桥煤矿石炭系太原组灰岩含水层为研究对象,利用2013—2024年的地下水位埋深观测数据及涌水量数据,采用深度学习长短期记忆网络(long short-term memory,LSTM)模型和卷积神经网络-长短期记忆网络(convolutional neural networks-LSTM,CNN-LSTM)混合模型对补ⅢC3-Ⅱ孔进行水位埋深时序预测。结果表明:C3Ⅱ组灰岩含水层是太原组灰岩含水层关键过渡层,不同井孔的水位埋深变化具有明显差异。多输入特征的LSTM模型具有较强的预测能力,测试集均方根误差(root mean square error,RMSE)达到0.63 m,决定系数(R2)达到0.86,具有较好的预测精度。多输入特征的CNN-LSTM混合模型相对于单一模型表现精度更高,测试集RMSE达到0.55 m,R2达到0.9。LSTM模型与CNN-LSTM模型均表现出良好的预测性能,但CNN-LSTM模型展现出更强的预测能力,研究结果为精准预测采矿扰动水位及矿山水害防治提供重要科学依据。
Abstract
Accurate prediction of groundwater level dynamics in coal mining areas is essential for providing critical guidance for water inrush forecasting and water disaster prevention under mining-disturbed conditions. Existing methods for groundwater level prediction often suffer from low precision and poor accuracy, making it difficult to meet the safety production requirements of mining areas. The limestone aquifer of the Carboniferous Taiyuan Formation in the Xieqiao Coal Mine, Huainan, was selected as the research object. Groundwater depth time series data and water inflow records from 2013 to 2024 were utilized. Into deep learning models, including the long short-term memory (LSTM) network and a convolutional neural network-LSTM (CNN-LSTM) hybrid model, time series forecasting was conducted to predict the water table level depth at the borehole BuⅢC3-Ⅱ. The results indicate that the C3Ⅱ Group Limestone Aquifer served as a key transitional layer for the Taiyuan Formation limestone water, with distinct variations in the response patterns of recharge and discharge observed across boreholes in different sub-areas. The LSTM model with multiple input features demonstrated strong predictive capability, achieving a root mean square error (RMSE) of 0.63 m on the test set and a coefficient of determination (R2) of 0.86, indicating excellent prediction accuracy. The CNN-LSTM hybrid model with multiple input features showed higher accuracy compared to single models, achieving RMSE of 0.55 m on the test set and R2 of 0.9. Both the LSTM model and CNN-LSTM model proposed show satisfactory predictive performance. However, CNN-LSTM model demonstrated superior predictive capability. The results provide a crucial scientific basis for accurately predicting groundwater level forecasting under mining disturbance and mine water disaster prevention and control.
YI Si-tong, LI Xu, ZHANG Gang, ZHANG Hai-tao, XU Guang-quan, HAO Ruo-nan, DAI Xin.
Time Series Prediction of Groundwater Level under Mining Conditions in Huainan Coalfield Based on Deep Learning[J].
Science Technology and Engineering, 2026, 26(29): 12521-12529 DOI:10.12404/j.issn.1671-1815.2600490
ZhangJixiong, JuFeng, LiMeng, et al. Research progress on theories and methods of strata control in deep mining with backfill[J]. Bulletin of National Natural Science Foundation of China, 2024, 38(6): 1043-1051.
LianHuiqing, YanTao, YinShangxian, et al. Research on early warning of roof water inrush in working faces based on a transparent hydrogeological model[J]. Coal Science and Technology, 2025, 53(1): 259-271.
LiShaoyu, YanWei, YuDan. Analysis of human causes based on coal mine flood accident report[J]. Science Technology and Engineering, 2023, 23(10): 4457-4462.
ZengYifan, YuChao, WuQiang, et al. “Three zones” method for coal mine water hazard control and its significance[J]. Journal of China Coal Society, 2024, 49(8): 3605-3618.
SunWenjie, LiWenjie, NingDianyan, et al. Current states, prediction and prevention suggestions for water hazard accidents in China's coal mines[J]. Coal Geology and Exploration, 2023, 51(12): 185-194.
[11]
HuZ, ZhangT, ZhanM. Multi-attribute analysis of transmission channel waves: applications in mine water damage prevention[J]. Water, 2025, 17(7): 1018.
XieGuoqiang, HeKaikai, YiSihai, et al. Response mechanism of vegetation ecology to groundwater hydrological variations and climate change in the Shendong mining area[J]. Science Technology and Engineering, 2026, 26(4): 1745-1756.
[14]
LiB, LuY, ZhangS, et al. Adapting newly constructed well depth to groundwater level changes[J]. Water, 2025, 17(14): 2066.
[15]
MondalC N, BegumF. Retraction Note: hydraulic connectivity influenced by rainfall, reservoir water level, and groundwater dynamics: insights of statistical analysis in Warna region, Maharashtra, India[J]. Environmental Earth Sciences, 2025, 85(1): 20.
LiRui, ZhouHongfu, LiShuwu, et al. Prediction of water inflow after excavation of underground chamber group with long span and high buried depth[J]. Hydrogeology and Engineering Geology, 2025, 52(1): 179-189.
[18]
SundharS, ShashankS, NandhiniD, et al. Ground water quality assessment and forecasting using attention-based mechanisms[J]. Environmental Sciences Europe, 2025, 37(1): 232.
[19]
SuescumC L, CookP, PartingtonD, et al. Influence of aquifer properties on optimal location and duration of groundwater level monitoring to predict aquifer recovery after mine closure[J]. Journal of Hydrology, 2025, 661: 133686.
WangSiying, HanQi, ZhangLiqiu, et al. Research progress and prospect of groundwater pollution prevention and control technology in China[J]. Environmental Pollution and Control, 2024, 46(12): 1834-1840.
XuYuanwei, LiZuoyong, WangJiayang. Similar forecast models of underground water level based on set pair analysis[J]. Journal of China Hydrology, 2015, 35(6): 6-10.
[24]
WeiA, LiX, YanL, et al. Machine learning models combined with wavelet transform and phase space reconstruction for groundwater level forecasting[J]. Computers and Geosciences, 2023, 177: 105386.
WangWenxue, WangBosen, GaoYanwei, et al. Law governing variations in water-pressure transmission rate in confined aquifers[J]. Coal Geology and Exploration, 2024, 52(3): 79-88.
[27]
LiH, GuoZ, ChenK, et al. Modeling land subsidence under future water stress: the influence of groundwater exploitation, climate change, and inter-basin water diversion[J]. Water Resources Research, 2025, 61(11): e2025WR041105.
[28]
GaoW, LiaoY, ChenY, et al. Enhancing transparency in data-driven urban pluvial flood prediction using an explainable CNN model[J]. Journal of Hydrology, 2024, 645: 132228.
[29]
YangY, FengD, BeckE, et al. Global daily discharge estimation based on grid long short-term memory (LSTM) model and river routing[J]. Water Resources Research, 2025, 61(6): e2024WR039764.
QinHao, PengYuting, ZhangWei, et al. Overview of the development of world network information frontier technology in 2022[J]. Journal of China Academy of Electronics and Information Technology, 2023, 18(6): 578-585.
[32]
LiM, ZhouQ, HanX, et al. Prediction of reference crop evapotranspiration based on improved convolutional neural network (CNN) and long short-term memory network (LSTM) models in Northeast China[J]. Journal of Hydrology, 2024, 645: 132223.
[33]
OokoO, CheptegeiL, KarumeM. Application of machine learning for real-time water quality monitoring in developing countries: a review[J]. Sustainable Futures, 2025, 10: 100984.
BianHuiyuan, JiJiajun, DuanChaowei, et al. Coal body structure identification method based on Bayesian-optimized CNN-BiLSTM-Attention[J]. Coal Geology and Exploration, 2025, 53(11): 34-43.
GaoMing, LiHu, LiuXinjin, et al. Prediction method of groundwater in karst strata based on distance-attribute hybrid clustering combined with ConvLSTM model[J]. Science Technology and Engineering, 2025, 25(20): 8424-8434.
LiXiaogen, LiuHongchen, FuJingbao, et al. Groundwater level prediction based on CNN-Attention-LSTM model[J]. Water Resources Protection, 2025, 41(4): 228-235, 243.
ZhangHaitao, ChenXiaoqing, XuGuangquan, et al. Hydrogeochemical characteristics and thermal indication significance of Cambrian and Ordovician karst groundwater in Huainan Coalfield[J]. Journal of China Coal Society, 2026, 51(2): 1517-1533.
[42]
YanL, JinJ, ZhangR, et al. Interpretable soil moisture prediction based on the SABO-CONV1D-BiLSTM model on the Tibetan Plateau[J]. Advances in Climate Change Research, 2025, 16(5): 1022-1031.
[43]
LiuJ, WuY. A hybrid time series forecasting CGAN model based on LSTM and Conv1D and its application for load forecasting[J]. Electrical Engineering, 2025, 107(11): 1-18.