摘要
为解决当前高层住宅火灾风险评估缺乏统计分析与结果验证的问题,构建了一种百分制扣分式高层住宅火灾风险调研评估体系,运用该体系进行区域内高层住宅的科学抽样与192栋抽样样本的统计分析,建立支持向量机模型验证了评估工作的准确性。结果表明:不同高层住宅小区的火灾风险值存在较大差异,其原因主要是消防教育培训、消防设施维护保养等;同一小区内高层住宅火灾风险水平相差不大,但住户消防安全意识、电气隐患增多等造成部分楼栋火灾风险值差别大于10分;最佳优化的SVM模型证实了模型准确率在94%以上。
In order to solve the problems of lacking statistical analysis and result verification in fire risk assessment of high-rise residential buildings,a percentage-deduction system for fire risk investigation was constructed.The system was used to carry out the statistical analysis of 192 high-rise residential samples after scientific sampling analysis.A SVM model was established to verify the accuracy of the assessment work.The results showed that the fire risk values of high-rise residential communities are quite different,and main reasons are fire education and training,fire facility maintainance,etc;the fire risk values of buildings in the same high-rise residential community showed small differences,yet the fire safety awareness of the household,the increase of electrical hidden dangers cause some buildings has more than risk value difference more than 10;the accuracy of optimized SVM model for the assessment work is 94%.
作者
卢颖
张秀玲
吕希凡
姜学鹏
LU Ying;ZHANG Xiu-ling;LV Xi-fan;JIANG Xue-peng(School of Resource&Environmental Engineering,Wuhan University of Science and Technology,Hubei Wuhan 430081,China;Hubei Research Center of Industrial Safety Engineering Technology,Hubei Wuhan 430081,China;Institute of Fire Technology,Wuhan University of Science and Technology,Hubei Wuhan 430081,China)
出处
《消防科学与技术》
CAS
北大核心
2019年第11期1636-1640,共5页
Fire Science and Technology
基金
湖北省自然科学基金项目(2018CFB186)
湖北省冶金矿产资源高效利用与造块重点实验室开放基金项目(2017zy004)
国家安全生产重特大事故防治关键技术科技项目(hubei-0002-2017AQ)
关键词
高层住宅
火灾风险评估
统计分析
支持向量机
验证分析
high-rise apartment
fire risk assessment
statistical analysis
support vector machine
verification analysis