基于网络层次分析-灰色聚类法的高层建筑火灾风险分析  

Fire risk assessment of the high-rise buildings based on the network hierarchical analysis-gray clustering method

作　　者：贾进章 陈怡诺 JIA Jin-zhang;CHEN Yi-nuo(College of Safety Science and Engineering,Key Laboratory of Mine Thermal Power Disaster and Prevention,Ministry of Education,Liaoning Technical University,Fuxin 123000,Liaoning,China)

机构地区：[1]辽宁工程技术大学安全科学与工程学院,矿山热动力灾害与防治教育部重点实验室,辽宁阜新123000

出　　处：《安全与环境学报》2020年第4期1228-1235,共8页Journal of Safety and Environment

基　　金：国家自然科学基金项目(51374121);辽宁特聘教授资助项目(551710007007);辽宁省百千万人才工程项目(2019-45-15)。

摘　　要：为了探索高层建筑火灾的众多影响因素,降低城市火灾的风险等级,针对高层建筑火灾影响因素相互关联及因素信息不确定的问题,将定性描述与定量计算相结合,提出了一种基于网络层次分析–灰色聚类法的高层建筑火灾风险评价方法。结合调查研究和相关法律法规建立了高层建筑火灾风险评价指标体系,据此确定了目标层因素、5个准则层因素和23个网络层因素,利用Super Decision软件的[Design]、[Cluster]、[Node]、[Node conations from]、[priorities]等功能构建了高层建筑火灾网络层次分析模型并计算各指标的权重。运用灰色聚类方法处理指标中不确定和不完整的信息,确定了灰类、阈值和风险等级,构建了相应的白化权函数,聚类分析出评价目标所属的风险等级。利用提出的方法确定沈阳某高层建筑火灾的风险等级为"较安全状态",与传统的模糊综合评价结果一致,表明该方法合理可行。其中火灾自动报警系统、周边救援、逃生规划路线和管理水平4个风险因素需重点关注,同时针对安全疏散能力和安全管理能力等薄弱环节提出了相应对策。The present paper intends to make an exploration of the influential factors of the high-rise building fire disasters so as to reduce such fire risk probability in the cities.For this purpose,we have established an evaluation index system for the high-rise building fire risks in accordance with the specific features of the fire causes.Aiming at the interaction among the factors of the said fire safety evaluation system and the difficulty for precise delimitation of such factors,the paper has combined the methods of qualitative description and the quantitative calculation so as to be able to propose a high-rise building fire risk evaluation method based on the analytic network process grey clustering method.And,then,with the help of the index system of the said high-rise building fire risks,we have ascertained the target layer factors,the 5 criterion layer factors and the 23 network layer factors.Furthermore,we have also adopted necessary functions of some kinds of super decision-making software with superior features,known as the Design,Cluster,Node,Node Conations to establish an analytic network process model for the high-rise building fire disasters to assess the weights of each evaluation index.And,next,the grey clustering method has been taken to deal with the uncertain and incomplete pieces of information in the index so as to confirm the corresponding whitening weight functions constructed by the gray class,the threshold and the risk levels,to identify the risk levels of the indicators through clustering and regulation.In so doing,we have managed to apply the above mentioned methods to evaluate and analyze the fire risks of the high-rise buildings in Shenyang.Besides,we have also compared the traditional evaluation methods,such as the fuzzy comprehensive evaluation.And,all the results we have gained prove to be in the‘safety state’and in a position to verify the effectiveness and conceivability of the proposed evaluation method.Among them,much more attention should be paid to the4 risk factors,that is,th

关 键 词：安全工程 高层建筑 火灾 网络层次分析 灰色聚类法 风险分析 

分 类 号：X93[环境科学与工程—安全科学]