中国安全科学学报 ›› 2021, Vol. 31 ›› Issue (11): 163-170.doi: 10.16265/j.cnki.issn 1003-3033.2021.11.023

• 公共安全 • 上一篇    下一篇

基于动静态混合数据分析的误报警判定方法:面向建筑信息模型的运维平台*

邹荣伟, 杨启亮 教授, 邢建春 教授, 李苏亮, 陈文杰, 孔琳琳   

  1. 陆军工程大学 国防工程学院,江苏 南京 210007
  • 收稿日期:2021-08-14 修回日期:2021-10-09 出版日期:2021-11-28 发布日期:2022-05-28
  • 作者简介:邹荣伟 (1996—),女,安徽蚌埠人,硕士研究生,主要研究方向为建筑信息化。E-mail:zrwlyf@163.com。
  • 基金资助:
    江苏省自然科学基金资助(BK20201335);国家重点研发计划(2017YFC0704100)。

False alarm judgment method based on dynamic and static mixed data analysis: for building BIM operation and maintenance platforms

ZOU Rongwei, YANG Qiliang, XING Jianchun, LI Suliang, CHEN Wenjie, KONG Linlin   

  1. College of Defense Engineering, PLA Army Engineering University, Nanjing Jiangsu 210007, China
  • Received:2021-08-14 Revised:2021-10-09 Online:2021-11-28 Published:2022-05-28

摘要: 为解决消防误报警消息的频繁出现给建筑运维工作带来极大不便的问题,充分利用建筑信息模型(BIM)运维平台的数据资源优势,提出一种基于动静态混合数据分析的消防误报警判定方法。在对建筑火灾风险及消防误报警成因研究分析基础上,首先,从建筑物联网获取报警点所属空间的温度、烟雾浓度、CO浓度等参数构成动态数据集,采用遗传算法优化的BP神经网络(GA-BP)分析动态数据集,获得建筑空间存在明火的概率;然后,从BIM模型提取报警点空间位置与建筑材料参数构成静态数据集,采用提出的模糊贝叶斯网络评估该空间火灾风险概率;最后,加权融合以上2个概率值,最终得到报警点所属空间发生火灾的概率,进而反向推断出误报警的概率。研究结果表明:该方法可以判定建筑BIM运维平台的消防误报警概率,为建筑运维人员提供决策依据。

关键词: 动静态混合数据, 消防误报警, 建筑信息模型(BIM), 运维平台, 遗传算法-BP(GA-BP)

Abstract: In order to address the great inconvenience caused by frequent occurrence of false fire alarm messages to the construction operation and maintenance work, a judgement method of false alarms based on dynamic and static mixed data analysis was proposed by utilizing advantages of data resources of BIM operation and maintenance platforms. Firstly, building on the research and analysis on building fire risks and causes of false alarms, parameters of the space where alarm points belonged, including temperature, smoke concentration, and CO concentration were obtained from building Internet of things to form a dynamic data set, and optimized BP neural network by genetic algorithm(GA-BP) was used to analyze the data set so as to obtain probability of open fire in building space. Then, spatial position of alarm points and building materials' parameters were extracted from BIM model to form a static data set, and the proposed fuzzy Bayesian network was applied to evaluate fire risk probability of the space. Finally, values of the two probability were weighted and merged to obtain probability of fire in the space where alarm points were located, and then that of false alarm was deduced. The results show that this method can decide false fire alarm probability of BIM operation and maintenance platforms, and provide a basis for related personnel to make a decision.

Key words: dynamic and static mixed data, false fire alarm, building information model (BIM), operation and maintenance platform, genetic algorithm-BP (GA-BP)

中图分类号: