Construction and application of a TCN-LSTM-SVM-based time series prediction model for water inflow in coal seam roofs

Authors

LIU Xuan, China Coal Research Institute, Beijing 100013, China; CCTEG Xi’an Research Institute (Group) Co., Ltd., Xi’an 710077, China; State Key Laboratory of Coal Mine Disaster Prevention and Control, Xi’an 710077, China; Shaanxi Key Laboratory of Prevention and Control Technology for Coal Mine Water Hazard, Xi’an 710077, ChinaFollow
JI Yadong, China Coal Research Institute, Beijing 100013, China; CCTEG Xi’an Research Institute (Group) Co., Ltd., Xi’an 710077, China; State Key Laboratory of Coal Mine Disaster Prevention and Control, Xi’an 710077, China; Shaanxi Key Laboratory of Prevention and Control Technology for Coal Mine Water Hazard, Xi’an 710077, ChinaFollow
ZHU Kaipeng, CCTEG Xi’an Research Institute (Group) Co., Ltd., Xi’an 710077, China; State Key Laboratory of Coal Mine Disaster Prevention and Control, Xi’an 710077, China; Shaanxi Key Laboratory of Prevention and Control Technology for Coal Mine Water Hazard, Xi’an 710077, China
ZHAO Chunhu, China Coal Research Institute, Beijing 100013, China; CCTEG Xi’an Research Institute (Group) Co., Ltd., Xi’an 710077, China; State Key Laboratory of Coal Mine Disaster Prevention and Control, Xi’an 710077, China; Shaanxi Key Laboratory of Prevention and Control Technology for Coal Mine Water Hazard, Xi’an 710077, China
LI Kai, CCTEG Xi’an Research Institute (Group) Co., Ltd., Xi’an 710077, China; State Key Laboratory of Coal Mine Disaster Prevention and Control, Xi’an 710077, China; Shaanxi Key Laboratory of Prevention and Control Technology for Coal Mine Water Hazard, Xi’an 710077, China
LI Chaofeng, CCTEG Xi’an Research Institute (Group) Co., Ltd., Xi’an 710077, China; State Key Laboratory of Coal Mine Disaster Prevention and Control, Xi’an 710077, China; Shaanxi Key Laboratory of Prevention and Control Technology for Coal Mine Water Hazard, Xi’an 710077, China
YUAN Chenhan, CCTEG Xi’an Research Institute (Group) Co., Ltd., Xi’an 710077, China; State Key Laboratory of Coal Mine Disaster Prevention and Control, Xi’an 710077, China; Shaanxi Key Laboratory of Prevention and Control Technology for Coal Mine Water Hazard, Xi’an 710077, China
LI Panpan, CCTEG Xi’an Research Institute (Group) Co., Ltd., Xi’an 710077, China; State Key Laboratory of Coal Mine Disaster Prevention and Control, Xi’an 710077, China; Shaanxi Key Laboratory of Prevention and Control Technology for Coal Mine Water Hazard, Xi’an 710077, China
YAN Pengzhen, China Coal Research Institute, Beijing 100013, China; CCTEG Xi’an Research Institute (Group) Co., Ltd., Xi’an 710077, China; State Key Laboratory of Coal Mine Disaster Prevention and Control, Xi’an 710077, China; Shaanxi Key Laboratory of Prevention and Control Technology for Coal Mine Water Hazard, Xi’an 710077, China

Abstract

Background The accurate prediction of mine water inflow plays a significant role in the prevention and control of water hazards and the safe and efficient production in coal mines. Methods To construct a prediction model of water inflow in mines under threat of water hazards from extremely thick sandstones aquifer in West China, this study investigated a typical mine (also referred to as the studied mine) severely affected by such water hazards in the Binchang mining area of Shaanxi Province. The correlation between the mining footage and water inflow of the mining face was selected as the characteristic variable for the time series prediction of mine water inflow. Accordingly,this study proposed a prediction model for water inflow along the mining face in the studied mine based on the temporal convolutional network (TCN), long short-term memory (LSTM), and support vector machine (SVM)—the TCN-LSTM-SVM model. First, by raw data processing using the TCN framework, this model extracted the dependency between mining footage and water inflow and its dynamic characteristics. Subsequently, the extracted characteristics were output to the LSTM-SVM model to further capture the time series relationship between mining footage and water inflow and its characteristics. Results The training and prediction results indicate that the TCN-LSTM-SVM model yielded mean absolute errors ($ {E}_{{\mathrm{MA}}} $) ranging from 56.02 m³/h to 129.89 m³/h, mean absolute percentage errors ($ {E}_{{\mathrm{MAP}}} $) from 3 % to 7 %, root mean square errors ($ {E}_{{\mathrm{RMS}}} $) from 82.60 m³/h to 162.61 m³/h, and coefficients of determination ($ {R}^{2} $) from 0.81 to 0.98 based on the training, validation, and test sets. This model exhibited more accurate prediction results compared to the commonly used prediction models like backpropagation neural network (BPNN), random forest (RF), and Transformer while avoiding excessive errors produced by most of these models on the validation and test sets. The results indicate that the TCN-LSTM-SVM model integrated the parallel processing advantages and multi-scale feature extraction capacity of the TCN model while also enjoying the excellent prediction performance and generalization capability of the LSTM-SVM model. Compared to previously developed models, the TCN-LSTM-SVM model demonstrated certain superiority and applicability in the prediction of water inflow along the mining face in the studied mine. Conclusions The results of this study provide a new approach to water inflow prediction for mines with similar geological conditions to those in the Binchang mining area. Therefore, this study holds practical implications for water inflow prediction and water prevention and control in mining faces with similar geological conditions to those in the studied mine.

Keywords

mine water hazard, coal seam roof, water inflow prediction, temporal convolutional network, long short-term memory (LSTM) network, support vector machine (SVM), Binchang mining area in Shaanxi

DOI

10.12363/issn.1001-1986.25.03.0198

Recommended Citation

LIU Xuan, JI Yadong, ZHU Kaipeng, et al. (2025) "Construction and application of a TCN-LSTM-SVM-based time series prediction model for water inflow in coal seam roofs," Coal Geology & Exploration: Vol. 53: Iss. 6, Article 17.
DOI: 10.12363/issn.1001-1986.25.03.0198
Available at: https://cge.researchcommons.org/journal/vol53/iss6/17

Reference

[1] 王双明，申艳军，宋世杰，等. “双碳”目标下煤炭能源地位变化与绿色低碳开发[J]. 煤炭学报，2023，48(7)：2599−2612.

WANG Shuangming，SHEN Yanjun，SONG Shijie，et al. Change of coal energy status and green and low–carbon development under the “dual carbon” goal[J]. Journal of China Coal Society，2023，48(7)：2599−2612.

[2] 杨建，蒲悦. 我国西部矿井涌水分布特征及充水模式研究[J/OL]. 煤炭科学技术，2025：1–13 (2025-01-16) [2025-03-18]. https://kns.cnki.net/KCMS/detail/detail.aspx?filename=MTKJ20241219002&dbname=CJFD&dbcode=CJFQ.

YANG Jian，PU Yue. Research on the distribution characteristics and water filling modes of mine water inflow in western mining areas of China[J/OL]. Coal Science and Technology，2025：1–13 (2025-01-16) [2025-03-18]. https://kns.cnki.net/KCMS/detail/detail.aspx?filename=MTKJ20241219002&dbname=CJFD&dbcode=CJFQ.

[3] 孙文洁，李文杰，宁殿艳，等. 我国煤矿水害事故现状、预测及防治建议[J]. 煤田地质与勘探，2023，51(12)：185−194.

SUN Wenjie，LI Wenjie，NING Dianyan，et al. Current states，prediction and prevention suggestions for water hazard accidents in China’s coal mines[J]. Coal Geology & Exploration，2023，51(12)：185−194.

[4] 顾大钊，李井峰，曹志国，等. 我国煤矿矿井水保护利用发展战略与工程科技[J]. 煤炭学报，2021，46(10)：3079−3089.

GU Dazhao，LI Jingfeng，CAO Zhiguo，et al. Technology and engineering development strategy of water protection and utilization of coal mine in China[J]. Journal of China Coal Society，2021，46(10)：3079−3089.

[5] 韦华鹏，罗奇斌，康卫东，等. 基于不同开采方式的煤矿涌水量预测及其环境影响分析[J]. 水文地质工程地质，2023，50(1)：21−31.

WEI Huapeng，LUO Qibin，KANG Weidong，et al. Prediction of coal mine water inflow by different mining methods and environment impact analyses[J]. Hydrogeology & Engineering Geology，2023，50(1)：21−31.

[6] 周振方，董书宁，董阳，等. 蒙陕接壤区典型煤层开采顶板周期性变形破坏及涌水响应特征[J]. 煤田地质与勘探，2024，52(8)：101−110.

ZHOU Zhenfang，DONG Shuning，DONG Yang，et al. Periodic roof deformation and failure and associated water inflow characteristics during the mining of typical coal seams in the Inner Mongolia–Shaanxi contiguous area[J]. Coal Geology & Exploration，2024，52(8)：101−110.

[7] 李超峰，刘业献，张金魁，等. 基于双Packer抽水试验的洛河组水文地质特征垂向变异性研究[J]. 煤田地质与勘探，2023，51(11)：55−64.

LI Chaofeng，LIU Yexian，ZHANG Jinkui，et al. Vertical variability of hydrogeological characteristics of Luohe Formation by double packer system pumping test[J]. Coal Geology & Exploration，2023，51(11)：55−64.

[8] 冯龙飞，王双明，王海，等. 彬长矿区典型顶板水害矿井洛河组砂岩微观孔隙特征研究[J]. 煤炭科学技术，2023，51(8)：208−218.

FENG Longfei，WANG Shuangming，WANG Hai，et al. Micro pore characteristics of Luohe aquifer sandstone in Binchang mining area typical roof water hazard mines[J]. Coal Science and Technology，2023，51(8)：208−218.

[9] 郭小铭. 彬长矿区洛河组沉积控水及开采扰动流场响应特征研究[D]. 北京：煤炭科学研究总院，2022：13–14.

GUO Xiaoming. Sedimentary controlling groundwater in Luohe Formation and the water flow field response regularity after mining disturbance in Binchang mining area[D]. Beijing：China Coal Research Institute，2022：13–14.

[10] 周全超，王丹丹，党志伟，等. 矿井涌水量预测方法及对比分析：以新疆东沟煤矿为例[J]. 煤炭科学技术，2024，52(增刊1)：211−220.

ZHOU Quanchao，WANG Dandan，DANG Zhiwei，et al. Research and comparative analysis of mine water inflow prediction method：A case study of Donggou coal mine in Xinjiang[J]. Coal Science and Technology，2024，52(Sup.1)：211−220.

[11] 周全超，刘玲，李继升，等. 基于多种解析法和数值法求解巨厚砂岩水文地质参数[J]. 煤炭科学技术，2024，52(增刊1)：174−182.

ZHOU Quanchao，LIU Ling，LI Jisheng，et al. Solving hydrogeological parameters of thick sandstone based on various analytical and numerical methods[J]. Coal Science and Technology，2024，52(Sup.1)：174−182.

[12] 胡兵，刘会月，李媛，等. Modflow中关键参数取值对矿井涌水量预测的影响研究[J]. 煤炭工程，2024，56(10)：247−253.

HU Bing，LIU Huiyue，LI Yuan，et al. Influence of key parameter values in Modflow on mine water inflow prediction[J]. Coal Engineering，2024，56(10)：247−253.

[13] 李建林，王路扬，李松营，等. 基于灰色–多步马尔科夫模型的矿井涌水量预测[J]. 河南理工大学学报(自然科学版)，2023，42(5)：66−71.

LI Jianlin，WANG Luyang，LI Songying，et al. Forecast of mine water inflow based on grey–multi–step Markov model[J]. Journal of Henan Polytechnic University (Natural Science)，2023，42(5)：66−71.

[14] 朱宗奎，黄鑫磊. 基于Visual MODFLOW的煤层底板突水量预测研究[J]. 煤炭科学技术，2020，48(8)：157−163.

ZHU Zongkui，HUANG Xinlei. Study on water inrush quantity prediction from coal seam floor based on Visual MODFLOW[J]. Coal Science and Technology，2020，48(8)：157−163.

[15] 徐智敏，陈天赐，陈歌，等. 煤层采动顶板水文地质参数演化与矿井涌水量动态计算方法[J]. 煤炭学报，2023，48(2)：833−845.

XU Zhimin，CHEN Tianci，CHEN Ge，et al. Hydrogeological parameter evolution of coal seam roof and dynamic calculation method of mine water inflow[J]. Journal of China Coal Society，2023，48(2)：833−845.

[16] 董书宁. 人工智能技术在煤矿水害防治智能化发展中的应用[J]. 煤矿安全，2023，54(5)：1−12.

DONG Shuning. Application of artificial intelligence technology in intelligent development of coal mine water disaster prevention and control[J]. Safety in Coal Mines，2023，54(5)：1−12.

[17] 姚辉，尹慧超，梁满玉，等. 机器学习方法在矿井水防治理论体系研究中的应用思考[J]. 煤田地质与勘探，2024，52(5)：107−117.

YAO Hui，YIN Huichao，LIANG Manyu，et al. Some reflections on the application of machine learning to research into the theoretical system of mine water prevention and control[J]. Coal Geology & Exploration，2024，52(5)：107−117.

[18] 李建林，高培强，王心义，等. 基于混沌–广义回归神经网络的矿井涌水量预测[J]. 煤炭科学技术，2022，50(4)：149−155.

LI Jianlin，GAO Peiqiang，WANG Xinyi，et al. Prediction of mine water inflow based on chaos–generalized regression neural network[J]. Coal Science and Technology，2022，50(4)：149−155.

[19] 侯恩科，夏冰冰，吴章涛，等. 基于CEEMDAN–BO–BiGRU的矿井涌水量预测研究[J]. 科学技术与工程，2023，23(28)：12012−12019.

HOU Enke，XIA Bingbing，WU Zhangtao，et al. Mine water inflow prediction based on CEEMDAN–BO–BiGRU[J]. Science Technology and Engineering，2023，23(28)：12012−12019.

[20] 侯恩科，徐林啸，荣统瑞. 彬长大佛寺矿井涌水量时序预测[J]. 西安科技大学学报，2024，44(3)：490−500.

HOU Enke，XU Linxiao，RONG Tongrui. Time series prediction of mine water inflow from Binchang Dafosi mine[J]. Journal of Xi’an University of Science and Technology，2024，44(3)：490−500.

[21] 王飞，荣统瑞，侯恩科，等. 基于VMD–BiLSTM组合模型的矿井涌水量时间序列预测方法研究[J]. 矿业研究与开发，2024，44(3)：143−151.

WANG Fei，RONG Tongrui，HOU Enke，et al. Research on time series prediction method of mine water inflow based on VMD–BiLSTM combined model[J]. Mining Research and Development，2024，44(3)：143−151.

[22] 丁莹莹，尹尚先，连会青，等. 基于SSA–CG–Attention模型的多因素采煤工作面涌水量预测[J]. 煤田地质与勘探，2024，52(4)：111−119.

DING Yingying，YIN Shangxian，LIAN Huiqing，et al. Prediction of mine water inflow along mining faces using the SSA–CG–Attention multifactor model[J]. Coal Geology & Exploration，2024，52(4)：111−119.

[23] SHI Junwei，WANG Shiqi，QU Pengfei，et al. Time series prediction model using LSTM–Transformer neural network for mine water inflow[J]. Scientific Reports，2024，14：18284.

[24] 杨鑫源，范传光，叶佳锐，等. 基于TCN–Attention模型的电力负荷短期概率预测方法[J]. 供用电，2025，42(4)：91−98.

YANG Xinyuan，FAN Chuanguang，YE Jiarui，et al. A short–term probability prediction method of power load based on TCN–Attention model[J]. Distribution & Utilization，2025，42(4)：91−98.

[25] 李士哲，张天宇，谢家乐. 基于电池老化趋势重构与TCN–GRU–Attention网络的SOH估计[J]. 电力科学与工程，2025，41(3)：38−45.

LI Shizhe，ZHANG Tianyu，XIE Jiale. SOH estimation based on battery aging trend reconstruction and TCN–GRU–Attention network[J]. Electric Power Science and Engineering，2025，41(3)：38−45.

[26] 陆睿，尹尚先，王玉国，等. 基于GMS的深部煤层开采工作面涌水量预测[J]. 煤矿安全，2025，56(1)：164−170.

LU Rui，YIN Shangxian，WANG Yuguo，et al. Prediction of water inflow in deep coal seam mining face based on GMS[J]. Safety in Coal Mines，2025，56(1)：164−170.

[27] 张琴琴，刘文强，陈之鸿，等. 基于LSTM–SVM模型的河流流量预测[J]. 天津师范大学学报(自然科学版)，2023，43(6)：45−52.

ZHANG Qinqin，LIU Wenqiang，CHEN Zhihong，et al. Forecasting of river flow based on LSTM–SVM model[J]. Journal of Tianjin Normal University (Natural Science Edition)，2023，43(6)：45−52.

[28] 王发麟，俞威. 基于TCN–LSTM神经网络的线缆性能衰退预测方法[J]. 科学技术与工程，2023，23(35)：15109−15116.

WANG Falin，YU Wei. Prediction method research of cable performance degradation state for complex mechatronic products based on TCN–LSTM[J]. Science Technology and Engineering，2023，23(35)：15109−15116.

[29] 赵玉，陈丽霞，梁梦姣. 基于LSTM_TCN 模型的降雨型滑坡时间概率预测及气象预警建模[J]. 地质科技通报，2024，43(2)：201−214.

ZHAO Yu，CHEN Lixia，LIANG Mengjiao. Temporal probability prediction and meteorological early warning modeling of rainfall-induced landslide based on LSTM_TCN model[J]. Bulletin of Geological Science and Technology，2024，43(2)：201−214.

[30] 王超，赵永良，王杰，等. 基于TCN–LSTM–SE神经网络的低压有源配电网电压控制方法研究[J]. 供用电，2025，42(4)：48−58.

WANG Chao，ZHAO Yongliang，WANG Jie，et al. Research on voltage control method of low voltage active distribution network based on TCN–LSTM–SE neural network[J]. Distribution & Utilization，2025，42(4)：48−58.

[31] 苏连成，朱娇娇，李英伟. 基于时间卷积网络残差校正的短期风电功率预测[J]. 太阳能学报，2023，44(7)：427−435.

SU Liancheng，ZHU Jiaojiao，LI Yingwei. Short–term wind power prediction based on temporal convolutional network residual correction model[J]. Acta Energiae Solaris Sinica，2023，44(7)：427−435.

[32] 王舒雨，李豪，马刚，等. 基于TCN和DLinear的光伏发电功率多步预测模型[J]. 电力建设，2025，46(4)：173−184.

WANG Shuyu，LI Hao，MA Gang，et al. Multistep prediction model for photovoltaic power generation based on time convolution and DLinear[J]. Electric Power Construction，2025，46(4)：173−184.

[33] 闫昭阳，朱景宝，刘赫奕，等. CEEMDAN–LSTM框架下的天然地震与矿震区分技术研究[J]. 煤炭学报，2024，49(增刊2)：731−745.

YAN Zhaoyang，ZHU Jingbao，LIU Heyi，et al. CEEMDAN–LSTM framework for distinguishing natural and mining–induced seismicity[J]. Journal of China Coal Society，2024，49(Sup.2)：731−745.