煤矿火灾具有突发性、后果严重性，一旦发生极易引起重特大事故，往往还伴随着一系列衍生事件的发生。而事故发生之前都会有前兆因素，从前兆情景的角度探究事故发展的演变规律，有利于快速、高效进行应急决策，从而将事故消灭在萌芽状态或减轻事故带来的损失。因此，探究基于前兆情景推演的煤矿外因火灾应急决策模型与方法，对煤矿火灾应急决策具有重要的意义。

首先，本文从情景分析着手，探讨了煤矿火灾事故的发生规律，依据煤矿外因火灾事故的致因分类进行前兆情景要素的划分，利用jieba分词提取关键的前兆情景，收集了2002-2020年间88例煤矿火灾事故，从机械摩擦、电气设备、明火、爆破、煤尘瓦斯爆炸等方面详细分析事故致因，构建了煤矿外因火灾前兆情景要素的三级指标体系。其次，按照时间顺序研究煤矿外因火灾应急决策前兆情景演变规律，提出将情景要素分为情景状态、前兆要素、应急活动，进行事故的情景推演，分别探讨积极和消极方向上的情景演变路径。再次，借用动态贝叶斯网络方法，确定网络节点概率，建立煤矿外因火灾情景演化动态贝叶斯网络，进而构建煤矿外因火灾情景推演应急决策模型，利用netica软件计算情景节点概率；针对关键情景依据专家经验生成应急决策方案，运用多属性决策的方法，构建煤矿外因火灾事故应急决策综合评价指标体系，构建决策矩阵、确定属性权重，运用基于熵权法改进的T OPSIS法进行应急决策方案的优选。最后，通过重庆松藻煤矿“9·27”重大火灾事故进行实例分析，构建事故情景的贝叶斯网络推演路径图，研究基于情景的动态应急决策过程并进行优选。

本文通过建立前兆情景分析的动态贝叶斯网络模型，对煤矿外因火灾事故的情景演变路径与情景推演应急决策模型进行分析，结合概率帮助决策者清晰直观的从图中看出事故未来发展的趋势、影响事故发展趋势的因素等，基于熵权法-TOPSIS法可辅助应急决策的优选，为决策者在煤矿外因火灾事故的应急决策提供了参考和依据。

Coal mine fires are sudden and have serious consequences, once they occur, they can easily lead to serious accidents and are often accompanied by a series of derivative events. Before an accident occurs, there will always be precursory factors. To explore the evolution law of accident development from the perspective of precursory scenarios is conducive to rapid and efficient emergency decision-making, so as to eliminate the accident in the bud or reduce the losses caused by the accident. Therefore, it is of great significance to explore the emergency decision-making model and method of coal mine external fire emergency based on the deduction of precursory scenarios.

First of all, this paper starts from the scenario analysis, discusses the occurrence law of coal mine fire accidents, divides the precursor scenario elements according to the classification of causes of coal mine external fire accidents, uses jieba word segmentation to extract key precursor scenarios, and collects 88 coal mine cases from 2002 to 2020, the accident causes are analyzed in detail from the aspects of mechanical friction, electrical equipment, open fire, blasting, and coal dust and gas explosionsand, a three-level index system of the precursory scenario elements of coal mine fire accidents is constructed. Secondly, the evolution law of precursory scenarios for emergency decision-making due to external fires in coal mines is studied in chronological order, and it is proposed to divide the scenario elements into scenario states, precursory elements, and emergency activities. Borrowing the dynamic Bayesian network method to determine the probability of network nodes, establish a dynamic Bayesian network for the evolution of externally-caused fire scenarios in coal mines, and then construct an emergency decision-making model for deduction of externally-caused fires in coal mines, and use netica software to calculate the probability of scenario nodes; then, emergency decision-making plans are generated based on expert experience for key scenarios, and a multi-attribute decision-making method is used to construct a comprehensive evaluation index system for emergency decision-making of external fire accidents in coal mines, the decision matrix is constructed, the attribute weight is determined, and the TOPSIS method based on the entropy weight method is used to optimize the emergency decision-making scheme. Finally, through the case analysis of the "9.27" major fire accident in Songzao Coal Mine in Chongqing, the Bayesian network deduction path map of the accident scenario is constructed, and the dynamic emergency decision-making process based on the scenario is studied and optimized.

By establishing a dynamic Bayesian network model for precursory scenario analysis, this paper analyzes the scenario evolution path and scenario deduction emergency decision-making model of coal mine external fire accidents, combining the probability helps decision makers to clearly and intuitively see the future development trend of the accident and the factors affecting the development trend of the accident from the graph,based on the entropy weight method-TOPSIS method, it can assist the selection of emergency decision-making, and provide a reference and basis for decision-makers to make emergency decision-making due to mine external fire.

[1]余明高.矿井火灾防治[M].北京:国防工业出版社,2013.

[2]Bill Reid. New training Standard for Mine Fires[N] . COAL LEADER, 2002, 5.

[3]R.E.Greuer,Transient State Simulation of Ventilation System in Fires Condition, 3rd International Mine Ventilation Congress,1984.

[4]谢之康,王省身.矿井外因火灾计算机控制及有待解决的若干问题[J].中国安全科学学报,1998(01): 67-71.

[5]Wu Z, Li H. Simulation of mine ventilation under the influence of mine fires. Proceedings of the 6th US Mine Ventilation Symposium. Salt Lake City. USA. 21~23June. 1993.

[6]Melody S M, Johnston F H. Coal mine fire sand human health:What do we know?[J]. International Journal of Coal Geology, 2015, 152:1-14.

[7]Larry G R, Kinilakodi H, Kecojevic V. Pilot sample risk analysis for underground coal mine fire sand explosion susing MSHA citationdata[J]. Safety Science, 2009,47(10):1371-1378.

[8]Wu J J, Liu X C. Risk assessment of underground coal fire development atregional scale[J]. International Journal of Coal Geology, 2011, 86(1):87-94.

[9]Spada M, Burgherr P. An after math analysis of the 2014 coal mine accident in Soma, Turkey: Use of risk performance in dicators based on historical experience[J]. Accident Analysis and Prevention, 2016, 87(2):134-140.

[10]Cole H P, VaughtC,Wiehagen W J, et al. Decision making during asimulated mine fire escape[J]. IEEET ransactions on Engineering Equipment, 1998, 45(2):153-162.

[11]王德明.矿井火灾学[M].徐州:中国矿业大学出版社,2008.

[12]张国枢.通风安全学[M].徐州:中国矿业大学出版社,2011.

[13]高尚青.矿用智能防灭火钻孔温度测量仪的开发与研制[D].太原:太原理工大学,2005.

[14]孙继平.煤矿电气安全关键技术研究[J].工矿自动化,2011,1(1):1-4.

[15]余明高,潘荣锟,黄照,等.乌达煤矿采空区煤自燃影响因素评价分析[J].中国安全科学学报,2006,16(4):97-101.

[16]许波波,张人伟,闫明仁,等.层次分析法在矿井火灾防治中的应用[J].煤矿现代化,2009(5):45-46.

[17]王桂明,张嘉勇,巩学敏,等.煤矿火灾事故评价方法研究[J].能源技术与管理,2010(1):52-54.

[18]丁大同,邸学勤.矿井自然火灾危险性多级模糊综合评价方法研究[J].煤炭科学技术,2005,33(3):58-60.

[19]刘业娇,曹庆贵,林晓飞.煤矿内因火灾事故的2级模糊综合评判[J].安全与环境学报,2006,6(S1):156-158.

[20]褚廷湘,杨胜强,刘国忠,等.模糊综合评价对采空区自然发火的评价及应用[J].中国安全生产科学技术,2008,4(1):124-127.

[21]张龙.五沟矿俯采综采面采空区煤炭自燃危险性模糊综合评价[D].淮南:安徽理工大学,2014.

[22]Harries C. Correspondence to what?Coherence to what?What is good scenario-baseddecision making?[J]. Technological Forecasting and Social Change, 2003, 70(8):797-817.

[23]KUL’BA, ZAIKINO, SHELKOVA, et al.Scenario analysis in the management of technogenic safety[J]. Ifac-Papers online, 2016, 49(12):1267-1271.

[24]Wang Y, Zhang L. Research on Scenario Deduction of Unconventional Emergency Based on Knowledge-Unit[C]//IEEE. IEEE, 2016.

[25]Qian J, Liu Y, Wang G, et al. A Cross-Reasoning Method for Scenario Awareness and Decision-Making Support in Earthquake Emergency Response[M]//Web-Age Information Management. 2014.

[26]Liu L, Liu L, Wang B. The Research of Effectiveness Assessment of Unconventional Emergency Plans Based on Scenario[C]. IEEE International Conferenceon Emergency Management&Management Sciences. 2010.

[27]Bood R, Postma T. Strategi clearning with scenarios[J]. European Management Journal,1997,15(6):633-647.

[28]Danner A G, Safeeq M, Grant G E, et al. Scenario-Based and Scenario-Neutral Assessment of Climate Change Impactson Operational Performance of a Multipurpose Reservoir[J]. JAWRA Journal of the American Water Resources Association, 2017,53(6):1467-1482.

[29]Sinha, R. Impact of Experience on Decision-making in Emergency Situation[R]. C/D Extended Essay, 2005:15.

[30]付娉娉.基于情景推演的非常规突发事件应急决策研究[D].哈尔滨工业大学,2015.

[31]王循庆,李勇建,孙华丽.基于情景推演的群体性突发事件演化博弈分析[J].管理科学,2015,28(06):133-143.

[32]张磊,王延章,陈雪龙.基于知识元的非常规突发事件情景模糊推演方法[J].系统工程学报,2016,31(06):729-738.

[33]孙超,钟少波,邓羽.基于暴雨内涝灾害情景推演的北京市应急救援方案评估与决策优化[J].地理学报,2017,72(05):804-816.

[34]巩前胜.基于动态贝叶斯网络的突发事件情景推演模型研究[J].西安石油大学学报(自然科学版),2018,33(02):119-126.

[35]宋英华,刘含笑,蒋新宇,等.基于知识元与贝叶斯网络的食品安全事故情景推演研究[J].情报学报,2018,37(07):712-720.

[36]张志霞,郝纹慧,张二双.网络舆情驱动下突发事件情景推演研究[J].情报科学,2020,38(05):141-147.

[37]郄子君.基于关键承灾体的区域复杂灾害情景建模研究[D].大连理工大学,2018.

[38]钱静,刘奕,刘呈,等.案例分析的多维情景空间方法及其在情景推演中的应用[J].系统工程理论与实践,2015,35(10):2588-2595.

[39]高山,王晗奕.基于动态贝叶斯网络的突发公共卫生事件情景推演[J].灾害学,2021,36(03):28-34.

[40]黄湘皖,程浩,陈小刚,等.基于动态贝叶斯网络的战场工程抢修情景分析[J].防护工程,2021,43(05):42-47.

[41]Antonioni G, Spadoni G, Cozzani V. A methodology for the quantitative risk assessment of major accidents triggered by seismic events[J]. Journal of Hazardous Materials, 2007, 147(1):48-59.

[42]Mietzner D, Reger G. Advantages and Disadvantages of Scenario Approaches for Strategic Foresight[J]. Social Science Electronic Publishing, 2005, 1(2).

[43]Cosgrave, John. Decision making in emergencies[J]. Disaster Prevention and Mana-gement, 1996, 5(4): 28-35.

[44]Cheung K F, Phadke AC, Wei Y, et al. Modeling of storm-induced coastal flooding foremergency management[J]. Ocean Engineering, 2003, 30(11): 1353-1386.

[45]Jie G, Zeshui X, Peijia R, et al. Anemergency decision making method based on the multiplicative consistency of probabilistic linguistic preference relations[J]. International Journal of Machine Learning and Cybernetics, 2018.

[46]张奕,卜凡亮.基于CBR与RBR的应急决策支持系统设计与实现[J].软件导刊,2019,18(02)：55-59.

[47]杨乃定,吴静杰.基于直觉模糊软集的煤矿应急救援决策方法[J].运筹与管理,2019,28(01):54-60.

[48]刘洋,樊治平,尤天慧,等.事前-事中两阶段突发事件应急决策方法[J].系统工程理论与实践,2019,39(01):215-225.

[49]张恭孝,杨荣华,崔萌,等.基于“情景-应对”模式的危险化学品泄漏应急决策系统的构建[J].化工环保,2018,38(05):609-614.

[50]樊自甫,田苡毓,吕浪,等.不确定环境下网络舆情突发事件应急方案评估决策研究[J].数学的实践与认识,2019,49(08):9-17.

[51]张国军,李浩,凌云翔.基于案例推理的地震应急救援决策方法[J].火力与指挥控制,2019,44(04):60-64.

[52]刘建国.关于应急管理和决策[J].现代营销(经营版),2019(04):95.

[53]陈祖琴,蒋勋.基于情报流程优化的应急决策生成与优化研究[J].情报理论与实践,2018,41(12):133-138.

[54]张志霞,许梦乐.基于云模型的天然气管道突发事件应急决策方案[J].消防科学与技术,2019,38(02):288-291+298.

[55]金卫健,徐浩,黄传峰,等.基于前景理论的大规模传染疫情应急管理决策研究[J].中国管理科学,2020,1:1-13.

[56]陈刚,张悦旺,谢科范.大型体育赛事突发事件应急群决策的相对熵变权集结模型[J].中国管理科学,2014,22:121-125.

[57]陈孝国,母丽华,杜红,等.煤矿突发事件应急救援的群决策方法[J].灾害学,2015,1,167-170.

[58]陈兴,王勇,吴凌云,等.多阶段多目标多部门应急决策模型[J].系统工程理论与实践,2010,30(11):1977-1985.

[59]Fu P P, Wu C, Tang J. Unconventional emergency management based on intelligent group decision-making methodology[J]. Advances in Information Sciences&Service Sciences, 2012, 4(7): 208-216.

[60]刘霞,严晓.非常规突发事件动态应急群决策:"情景-权变"范式[J].云南社会科学,2010,5:5-8.

[61]吴静杰,杨乃定.一种基于直觉模糊数的应急群决策方法[J].统计与决策,2015,17:77-79.

[62]Xu X H, Du Z J, Chen X H. Consensus model for multi-criterialarge-group emergency decision making considering non-cooperative behavior sand minority opinions[J]. Decision Support Systems, 2015, 79: 150-160.

[63]张倩生,谢柏林,张新猛.网络舆情突发事件的应急群决策方法[J].统计与决策,2016(10):38-41.

[64]郑晶,王应明.面向多部门多属性的群决策应急方案调整方法[J].运筹与管理,2020,29(02):66-72.

[65]Wack, P. Scenarios: Shooting the Rapids[J]. Harvard Business Review,1985,63(6):139-150.

[66]姜春光,郑鑫慧.基于熵权-TOPSIS模型的电网企业应急能力建设评价[J].工业安全与环保,2021,47(12):50-53.

[67]郭雨. 煤矿机电设备[M]. 徐州:中国矿业大学出版社,2005.

[68]巩前胜. “情景—应对”型应急决策中情景识别关键技术研究[D].西安科技大学,2018.

[69]Zheng Y, Zhao F, Wang Z. Fault diagnosis system of bridge crane equipment based on fault tree and Bayesian network[J]. International Journal of Advanced Manufacturing Technology, 2019(2).

[70]贾晓惠,王化姗,崔用梅.基于贝叶斯网络的公交车事故外因分析[J].科学技术与工程,2021,21(21):9116-9122.

[71]Sirra Pugh. Spatio-temporal Lagged Models for Variable Rate Irrigation in Agriculture[J]. Journal of Agricultural, Biological, and Environmental Statistics, 2019, 24(4):634-650.

[72]Ullah I, Mengersen K. Bayesian mixture models and their Big Data implementations with application to invasive species presence-only data[J]. Journal of Big Data,2019, 6(1):1-25.

[73]姜波,陈涛,袁宏永,等.基于情景时空演化的暴雨灾害应急决策方法[J].清华大学学报(自然科学版),2021,62(01):52-59.

[74]王志平,傅敏,王沛文.概率犹豫模糊环境下基于前景理论和TOPSIS法的多属性群决策模型[J].科学技术与工程,2022,22(04):1329-1337.

[75]李家伟,于忠霞. 基于情景分析的高速公路事件检测决策系统研究[J]. 2007, (7): 26-28.