中国安全科学学报 ›› 2019, Vol. 29 ›› Issue (10): 147-153.doi: 10.16265/j.cnki.issn1003-3033.2019.10.023

• 公共安全 • 上一篇    下一篇

基于Cloud-BN的装配式住宅构件吊装安全评价

申玲 教授, 唐晔文, 牟月   

  1. 南京工业大学 土木工程学院,江苏 南京 211816)
  • 收稿日期:2019-07-04 修回日期:2019-09-25 出版日期:2019-10-28 发布日期:2020-10-27
  • 作者简介:申 玲 (1966—),女,重庆人,博士,教授,硕士生导师,从事于工程项目管理方面的研究。E-mail: sl7455@sina.com。唐晔文 (1994—),男,江苏南通人,硕士研究生,研究方向为工程项目管理方面的研究。E-mail: 939798509@qq.com。
  • 基金资助:
    国家自然科学基金资助(71601095)。

Safety assessment of component hoisting for prefabricated residence based on Cloud-BN model

SHEN Ling, TANG Yewen, MU Yue   

  1. School of Civil Engineering, Nanjing Tech University, Nanjing Jiangsu 211816, China
  • Received:2019-07-04 Revised:2019-09-25 Online:2019-10-28 Published:2020-10-27

摘要: 为降低装配式住宅构件吊装事故发生率,构建基于云贝叶斯网络(Cloud-BN)的安全风险评价模型,动态评价构件吊装作业安全风险。首先,分析云模型与贝叶斯网络(BN)结合的可行性与合理性;其次,结合工程实际确定构件吊装施工安全风险指标体系与BN结构,通过参数学习算法获取各节点间的条件概率分布;然后利用云模型转换,将获取的观测节点先验概率作为证据输入模型,得到构件吊装风险处于各风险状态的概率,并利用综合云生成算法确定最终的风险等级;最后,以某装配式住宅项目为例,验证该模型的有效性。研究表明:该模型评价结果与实际情况基本一致,平均相对误差控制在5%以内;与传统BN模型相比,结果更加客观准确。

关键词: 云贝叶斯网络(Cloud-BN), 装配式住宅, 构件吊装, 施工安全, 风险评价

Abstract: In order to reduce accident rates of component hoisting construction of prefabricated residence, a risk assessment model based on Cloud-BN was constructed to dynamically evaluate these risks. Firstly, feasibility and rationality of the combination of cloud model and Bayesian networks (BN) were analyzed. Secondly, risk index system and BN structure of component hoisting construction were established according to engineering practice, and conditional probability distribution among nodes was obtained through parameter learning algorithm. Then, the obtained prior probability was input into the model through cloud model transformation to reason out risk possibilities at various risk states before final risk levels were obtain through comprehensive cloud generation method. Finally, a prefabricated residence project was taken as an example to verify this model's effectiveness and advantages. It is proved that its results are basically consistent with reality with an average relative error rate within 5%, which is more objective and accurate compared with conventional BN models.

Key words: cloud-Bayesian network(Cloud-BN), prefabricated residence, component hoisting, construction safety, risk assessment

中图分类号: