煤矸石山自燃具有蓄热量大、燃烧时间长、自内向外蔓延和潜在安全隐患等特点，易造成人员伤亡及环境污染。因此，煤矸石山自燃区域热量的移除一直是煤矸石山火灾治理中亟待解决的关键问题。本文采用实验与深度学习算法预测相结合的研究方法，通过开展煤矸石-热管移热降温实验中温度与气体变化规律及基于多信息预测方法的热管移热研究，取得以下成果：

通过搭建煤矸石-热管移热降温实验台，设计正交实验，开展热管在煤矸石堆的移热降温实验，计算不同工况下热管作用24 h的移热性能参数指标（降温幅度、降温率、移热量），掌握煤矸石堆在100 ℃、200 ℃和300 ℃下降温过程温度变化规律。研究表明热管可有效降低煤矸石堆温度，煤矸石堆温度越高，降温效果越显著。热管对靠近热管区域的测点具有较高的敏感性，热管作用下的煤矸石堆降温幅度表现为下层降温幅度>中层降温幅度>上层降温幅度。煤矸石堆内测点的降温幅度和降温率与热管间的距离呈负相关，热管的移热量与煤矸石堆温度呈正相关。各影响因素对热管移热量的敏感性分析及相应的最佳参数组合如下：100 ℃时，工质浓度>长径比>充液率，工质浓度为2%，充液率为25%，长径比为5:1；200 ℃时，充液率>工质浓度>长径比，工质浓度为3%，长径比为5:1，充液率为25%；300 ℃时，长径比>工质浓度>充液率，工质浓度3%，充液率25%，长径比4:1。

通过开展的煤矸石-热管移热降温实验，掌握煤矸石产生的CO和C₂H₄指标气体变化规律。研究表明在热管作用的前600 min是煤矸石各测点气体浓度快速下降阶段，后800 min各测点气体浓度下降速度相比前600 min较缓慢。热管作用下的煤矸石堆气体变化规律表现为下层下降速率>上层下降速率，对不同测点的气体下降速率和距离呈负相关。随着煤矸石堆温度的升高，热管对煤矸石堆测点的气体浓度影响效果越好，相反影响效果越差。

采用CNN-LSTM深度学习算法，建立基于热管降温过程煤矸石温度及气体浓度时序预测模型，通过与CNN、LSTM和GRU三种单一模型对比验证，验证了CNN-LSTM模型的预测性能，确定最佳预测精度模型。结果显示，CNN-LSTM模型在预测煤矸石降温过程中的温度及气体浓度数据时效果最好，MAE、RMSE和MSE最小，R²最大，其预测精度明显优于其他单一模型，为及时防控治理矸石山自燃提供理论支持。

Spontaneous combustion of coal gangue mountain has the characteristics of large heat storage, long combustion time, internal to external spread and potential safety hazards, which are easy to cause casualties and environmental pollution. Therefore, the removal of heat from the spontaneous combustion area of coal gangue mountain has always been a key problem to be solved urgently in the fire control of coal gangue mountain. In this paper, the following results are obtained by combining the research method of experiment and deep learning algorithm prediction, and the study of temperature and gas changes in the coal gangue-heat pipe heat transfer cooling experiment and the heat pipe heat transfer based on the multi-information prediction method are carried out.

By building gangue-heat pipe heat transfer cooling experimental platform, design orthogonal experiments, carry out heat pipe in the gangue heap heat transfer cooling experiments, calculation of different working conditions under the role of the heat pipe 24 h heat transfer performance parameters and indicators (cooling amplitude, cooling rate, the amount of heat shifted), to master the gangue heap in 100 ℃, 200 ℃ and 300 ℃ under the cooling process of the temperature change rule. Research shows that the heat pipe can effectively reduce the temperature of the gangue pile, the higher the temperature of the gangue pile, the more significant cooling effect. The heat pipe has high sensitivity to the measuring point near the heat pipe area, and the cooling amplitude of the gangue pile under the action of the heat pipe is shown as the cooling amplitude of the lower layer > the cooling amplitude of the middle layer > the cooling amplitude of the upper layer. The cooling amplitude and cooling rate of the measuring points in the gangue pile are negatively correlated with the distance between the heat pipes, and the amount of heat transfer from the heat pipes is positively correlated with the temperature of the gangue pile. The sensitivity analysis of each influencing factor on the heat pipe heat transfer effect and the corresponding optimal parameter combinations are as follows: At 100 °C, the work mass concentration >the aspect ratio > filling rate, with a work mass concentration of 2%, a filling rate of 25%, and the aspect ratio of 5:1; at 200 °C, the filling rate > work mass concentration > the aspect ratio, with a work mass concentration of 3%, the aspect ratio of 5:1, and a filling rate of 25%; and at 300 °C, the aspect ratio > work mass concentration > filling rate, with a work mass concentration of 3%, a filling rate of 25%, and the aspect ratio of 4:1.

Through the coal gangue-heat pipe heat transfer and cooling experiment, the change law of CO and C₂H₄ index gases produced by coal gangue was mastered. The results show that the gas concentration at each measuring point of coal gangue decreases rapidly in the first 600 min of heat pipe action, and the gas concentration at each measuring point decreases slowly in the second 800 min compared with the first 600 min. The gas variation law of coal gangue pile under the action of heat pipe is as follows that the descending rate of the lower layer > the descending rate of the upper layer, and the gas descending rate and distance of different measuring points are negatively correlated. With the increase of the temperature of the coal gangue pile, the effect of heat pipes on the gas concentration of the coal gangue pile is better, and the opposite effect is worse.

CNN-LSTM deep learning algorithm is used to establish a time series prediction model for the temperature and gas concentration of coal gangue based on the heat pipe cooling process, and the prediction performance of the CNN-LSTM model is verified by comparing and validating it with the three single models of CNN, LSTM, and GRU to determine the best prediction accuracy model. The results show that the CNN-LSTM model is the most effective in predicting the temperature and gas concentration data during the cooling process of coal gangue, with the smallest MAE, RMSE and MSE, and the largest R², and its prediction accuracy is obviously better than that of the other single models, which provides theoretical support for the timely prevention, control and management of spontaneous combustion of gangue mountain.

[1]徐莉, 余红伟, 徐晟. 煤矸石综合利用投入产出模型[J]. 技术经济, 2012, 31(08): 105-110.

[2]田莉, 于晓萌, 秦津. 煤矸石资源化利用途径研究进展[J]. 河北环境工程学院学报, 2020, 30(05): 31-36．

[3]满朝晖, 沈军. 煤矸石山自燃治理方法研究[J]. 煤炭工程, 2014, 46(07): 66-69.

[4]张震, 周少玺, 乔伟. 煤矿矸石山自燃防治技术研究[J]. 内蒙古煤炭经济, 2018 (14): 45-46.

[5]邢永强, 冯进城, 荣晓伟. 河南平煤四矿煤矸石山自燃爆炸成因及防治分析[J]. 中国地质灾害与防治学报, 2007, (02): 145-150.

[6]郭秀荣. 植被恢复在矸石山灾害治理中的作用[D]. 重庆: 重庆大学, 2006.

[7]贾宝山. 煤矸石山自然发火数学模型及防治技术研究[D]. 阜新: 辽宁工程技术大学, 2002.

[8]钟权, 董学强, 赵延兴, 等. 一种比定容热容测量装置[J]. 化工学报, 2019, 70 (S2): 20-24.

[9]Ren T X, Edwards J S. Clarke D. Adiabatic oxidation study on the propensity of pulverized coals to spontaneous combustion[J]. Fuel, 1999, 78(14): 1611-1620.

[10]Beamish B B, Barakat M A, George J D. Adiabatic testing procedures for determining the self-heating propensity of coal and sample ageing effects[J]. Thermochimica Acta, 2000, 362(1-2): 79-87.

[11]李品. 动态采空区煤自燃氧化升温机制的模拟研究[D]. 焦作: 河南理工大学, 2018.

[12]邓军, 徐精彩, 王洪权. 圆柱形煤自然发火实验台的数值模拟研究[J]. 辽宁工程技术大学学报(自然科学版), 2002, (02): 129-132.

[13]张瑞新, 谢和平. 煤堆自然发火的试验研究[J]. 煤炭学报, 2001, 26(02): 168-171.

[14]张瑞新, 谢和平, 谢之康. 露天煤体自然发火的试验研究[J]. 中国矿业大学学报, 2000, 29(03): 13-16.

[15]文虎. 煤自燃过程的实验及数值模拟研究[D]. 西安: 西安科技大学, 2003.

[16]Zhang D, Cen X X, Wang W F, et al. The graded warning method of coal spontaneous combustion in Tangjiahui Mine-Science Direct[J]. Fuel. 2021, 15(3): 119635.

[17]Kim A G. Locating fires in abandoned underground coal mines[J]. International Journal of Coal Geology, 2004, 59(1-2): 49-62.

[18]Singh A K, Singh R V K, Singh M P, et al. Mine fire gas indices and their application to Indian underground coal mine fires[J]. International Journal of Coal Geology, 2007, 69(3): 192-204.

[19]王新华. 煤矸石及其掺混生物质木屑的燃烧特性实验研究[D]. 合肥: 中国科学技术大学, 2019.

[20]徐精彩, 余锋, 李树刚, 等. 石嘴山二矿煤样程序升温自燃性测试研究[J]. 辽宁工程技术大学学报, 2003, (02): 148-150.

[21]张嬿妮, 邓军, 金永飞, 等. 煤自燃特性的油浴程序升温实验研究[J]. 煤矿安全, 2010, 41 (11): 7-10.

[22]陆伟, 王德明, 周福宝, 等. 绝热氧化法研究煤的自燃特性[J]. 中国矿业大学学报, 2005, 34(02): 84-88.

[23]张玉涛, 李亚清, 邓军, 等. 煤炭自燃灾变过程突变特性研究[J]. 中国安全科学学报, 2015, 25(01): 78-84.

[24]屈丽娜. 煤自燃阶段特征及其临界点变化规律的研究[D]. 北京: 中国矿业大学, 2013.

[25]张晓昱, 张玉龙, 王俊峰, 等. 外来水分对煤自燃过程影响及作用机制研究[J]. 燃料化学学报, 2020, 48(01): 1-10.

[26]文虎, 陆彦博, 刘文永. 利用热重法研究不同氧浓度对煤自燃特性的影响[J]. 矿业安全与环保, 2021, 48(01): 1-5+10.

[27]王继仁, 邓存宝, 单亚飞, 等. 煤的自燃倾向性新分类方法[J]. 煤炭学报, 2008, (01): 47-50.

[28]冉景煜, 牛奔, 张力, 等. 煤矸石综合燃烧性能及其燃烧动力学特性研究[J]. 中国电机工程学报, 2006, (15): 58-62.

[29]Ren J, Xie C, Lin J Y, et al. Co-utilization of two coal mine residues: Non-catalytic deoxygenation of coal mine methane over coal gangue[J]. Process Safety & Environmental Protection, 2014, 92(6): 896-902.

[30]Grover G M, Cotter T P, Erickson G F. Structures of very high thermal conductance[J]. Journal of Applied Physics, 1964, 36(5): 1990-1991.

[31]Denhartog S L. A thermosyphon for horizontal applications[J]. Cold Regions Science and Technology, 1988, 15(3): 319-321.

[32]Jardine J D, Long E L, Yarmak E. Thermal analysis of forced-air and thermosyphon cooling systems for the Inuvik airport expansion: Discussion[J]. Canadian Geotechnical Journal, 1992, 29(6): 998-1001.

[33]Mochizuki M, Singh R, Nguyen T, et al. Heat pipe based passive emergency core cooling system for safe shutdown of nuclear power reactor[J]. Applied Thermal Engineering, 2014, 73(1): 699-706.

[34]Chiou R Y, Lu L, Chen J S J, et al. Investigation of dry machining with embedded heat pipe cooling by finite element analysis and experiments[J]. International Journal of Advanced Manufacturing Technology, 2007, 31(9-10): 905-914.

[35]Senthilkumar R, Vaidyanathan S, Sivaraman B. Effect of inclination angle in heat pipe performance using copper nanofluid[J]. Procedia Engineering, 2012, 38(11): 3715-3721.

[36]周亚龙, 郭春香, 王旭, 等. 基于热棒功率变化下多年冻土区输电塔热棒桩基的长期降温效果预测[J]. 岩石力学与工程学报, 2019, 38(07): 1461-1469.

[37]郭春香, 吴亚平, 董晟, 等. 热棒填土路基降温效果的三维非线性有限元分析[J]. 中南大学学报(自然科学版), 2014, 45(01): 201-207.

[38]李永强. 青藏铁路多年冻土区热棒直径对降温效果和产冷量的影响分析[J]. 岩土工程学报, 2011, 33(S1): 510-515.

[39]徐兵魁, 熊治文. 青藏高原多年冻土区热棒路基设计计算[J]. 中国铁道科学, 2006, 27(05): 17-22.

[40]杨永平, 魏庆朝, 张鲁新, 等. 青藏铁路多年冻土地区热管路基三维数值分析[J]. 中国铁道科学, 2005, 26(02): 23-27.

[41]侯江丽, 伍玲玲, 罗旭照, 等. 基于直膨式热泵技术的深井集热降温实验系统[J]. 有色金属工程, 2020, 10(03): 62-68.

[42]李注江. 热管技术在煤矿工作面冷却水回收方面的研究[D]. 邯郸: 河北工程大学, 2014.

[43]Lenhard R, Katarína K, Štefan P. Analysis of the fill amount influence on the heat performance of heat pipe[C]. XIX the Application of Experimental & Numerical Methods in Fluid Mechanics & Energetics: International Conference. American Institute of Physics, 2014, 1608(1): 146-152.

[44]Streltsov A I. Theoretical and experimental investigation of optimum filling for heat pipes[J]. Heat Transfer-Soviet Research, 1975.

[45]徐礼华. 煤垛热管降温的实验研究[J]. 能源研究与利用, 1991(02): 11-15.

[46]董现锋, 谷明川, 王会勤, 等. 热管深部移热技术在煤矿矸石山的应用研究[J]. 中州煤炭, 2008(01): 8-10.

[47]Schmidt M, Suhendra S, Rüter H. Heat pipes-suitable for extinguishing underground coal fire[C]. Proceedings of Second International Conference on Coal Fire Research, Berlin, Germany, 2010: 433-437.

[48]屈锐. 重力热管提取储煤堆自燃热量的实验研究[D]. 西安: 西安科技大学, 2014.

[49]邓军, 李贝, 马砺. 用热棒技术强化煤堆降温幅度试验[J]. 中国安全科学学报, 2015, 25(06): 62-67.

[50]马砺, 李贝, 邓军, 等. 地面储煤堆(矸石山)自然发火蓄热高温区域的热棒深部移热技术[J]. 科技导报, 2014, 32(17): 76-80.

[51]Kamyar A, Ong K S, Saidur R. Effects of nanofluids on heat transfer characteristics of a two-phase closed thermosyphon-ScienceDirect[J]. International Journal of Heat and Mass Transfer, 2013, 65(7): 610-618.

[52]冯乾, 王景刚, 李敏婕, 等. 热管治理煤矸石山自燃技术的实验研究[J]. 煤炭与化工, 2019, 42(08): 110-114.

[53]王建国, 郑晨光, 王延秋. 冷凝段翅片类型对热管抑制煤自燃的降温效应影响研究[J]. 矿业安全与环保, 2020, 47(05): 13-17.

[54]尹岚. 煤火热能热棒提取与温差发电利用实验研究[D]. 西安: 西安科技大学, 2019.

[55]叶星星. L型热虹吸器强化煤火热能移取性能研究[D]. 西安: 西安科技大学, 2022.

[56]付华. 煤矿瓦斯灾害特征提取与信息融合技术研究[D]. 阜新: 辽宁工程技术大学, 2006.

[57]何萍, 王飞宇, 唐修义, 等. 煤氧化过程中气体的形成特征与煤自燃指标气体选择[J]. 煤炭学报, 1994, 19(06): 635-643.

[58]徐杨, 周延. 煤自然发火预报的人工神经网络模型[J]. 西安科技大学学报, 2009, 29(04): 410-414.

[59]高平, 边冰, 刘志坚. 基于BP神经网络的煤自然发火预报系统[J]. 煤炭技术, 2014, 33(09): 60-62.

[60]何启林, 彭伟. 利用束管监测系统预报采空区遗煤氧化情况[J]. 安徽理工大学学报(自然科学版), 2011, 31(03): 31-34.

[61]王福生, 韩慧兰, 柳晓莉, 等. 煤自然发火预测预报模型的构建[J]. 矿业安全与环保, 2011, 38(01): 21-22+6.

[62]陈晓坤. 煤自燃多源信息融合预警研究[D]. 西安: 西安科技大学, 2012.

[63]马文龙. 多传感器信息融合方法研究及在火灾预测中的应用[D]. 沈阳: 东北大学, 2015.

[64]Ruilin Z, Lowndes I S. The application of a coupled artificial neural network and fault tree analysis model to predict coal and gas outbursts[J]. International Journal of Coal Geology, 2010, 84(2): 141-152.

[65]Cao S G, Liu Y B, Wang Y P. A forecasting and forewarning model for methane hazard in, working face of coal mine based on LS-SVM[J]. Journal of China University of Mining & Technology, 2008, 18(02): 172-176.

[66]邓军, 雷昌奎, 曹凯, 等. 采空区煤自燃预测的随机森林方法[J]. 煤炭学报, 2018, 43(10): 2800-2808.

[67]王丽贤. 时间序列预测技术研究[D]. 天津: 天津理工大学, 2012.

[68]王宇, 李治平, 刘超. 基于分形与ARIMA的煤层气产量预测[J]. 天然气与石油, 2011, 29(03): 45-48+87.

[69]Sánchez A B, Ordóñez C, Lasheras F S, et al. Forecasting SO2 pollution incidents by means of Elman artificial neural networks and ARIMA models[J]. Abstract and Applied Analysis, 2013, 2013: 1-6.

[70]Smagulova K, James A P. A survey on LSTM memristive neural network architectures and applications[J]. The European Physical Journal Special Topics, 2019, 228(10): 2313-2324.

[71]Chambon S, Galtier M N, Arnal P J, et al. A deep learning architecture for temporal sleep stage classification using multivariate and multimodal time series[J]. IEEE Transactions on Neural Systems & Rehabilitation Engineering, 2018, 26(4): 758-769.

[72]Zhang T, Song S, Li S, et al. Research on Gas Concentration Prediction Models Based on LSTM Multidimensional Time Series[J]. Energies, 2019, 12(1): 161.

[73]Jia P, Liu H, Wang S, et al. Research on a mine gas concentration forecasting model based on a GRU network[J]. IEEE Access, 2020(8): 38023-38031.

[74]Lyu P, Chen N, Mao S, et al. LSTM based encoder-decoder for short-term predictions of gas concentration using multi-sensor fusion[J]. Process Safety and Environmental Protection, 2020(137): 93-105.

[75]Hong H, Zheng L, Zhu J, et al. Automatic recognition of coal and gangue based on convolution neural network[J]. ArXiv, 2017.

[76]Fu Y, Aldrich C. Quantitative ore texture analysis with convolutional neural networks[J]. IFAC-Papers OnLine, 2019, 52(14): 99-104.

[77]蒋卫祥. 增量式矿石自动化分拣系统研究[J]. 矿业研究与开发, 2020, 40(11): 150-155.

[78]Parker W J. Jenkins R J. Thermal conductivity measurements on bismuth telluride in the presence of a 2 MeV electron beam[J]. Advanced Energy Conversion, 1962, 2(62): 87-103.

[79]Gosset D, Guillois O, Popular R, Thermal diffusivity of compacted coal powders[J]. Carbon, 1996, 34(3): 369-373.

[80]Abdulagatov I M , Emirov S N, Abdulagatova Z Z , et al, Effect of pressure and temperature on the thermal conductivity of rock[J]. Journal of Chemical and Engineering Data: the ACS Journal for Data, 2006, 51(1): 22-33.

[81]甄强, 王方方, 王亚丽, 等. 煤矸石制备多孔复相陶瓷材料及导热系数研究[J]. 硅酸盐学报, 2014, 33(11): 2797-2801+2808.

[82]Wen H, Lu J H, Xiao Y, et al. Temperature dependence of thermal conductivity, diffusion and specific heat capacity for coal and rocks from coalfield[J]. Thermochimica Acta, 2015, 619: 41-47.

[83]李贝. 煤矸石山非控自燃热动力学特征及移热方法研究[D]. 西安: 西安科技大学, 2017.

[84]白子明. 防治煤自燃的微胶囊化复合阻化剂研究[D]. 徐州: 中国矿业大学, 2019.

[85]Rosenblatt F. The perceptron: A probabilistic model for information storage and organization in the brain[J]. Psychological Review, 1958, 65(6): 386-408.

[86]Rumelhart D E, Hinton G E, Wiliams R J. Learning representations by back-propagating errors[J] Nature, 1986, 323(6088): 533-536.

[87]Hochreiter S, Schmidhuber J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780.