中文摘要：

为提高施工企业对安全系统的管理能力,针对施工安全环境的复杂多变性和安全管理的局限性,建立安全系统脆弱性的系统动力学（SD）仿真模型。用误差反向传播（BP）神经网络结合平均影响值（MIV）算法,确定模型中员工、技术及管理系统等相关因素间的权重系数。根据对86个工程项目428份调查问卷数据,完成施工企业安全系统脆弱性仿真。发现员工系统和管理系统对施工企业安全系统应对能力影响较大,技术与设备系统和环境系统在影响安全系统敏感性方面表现显著。因此,施工企业应通过提高员工安全行为能力和管理者的安全管理水平来提高企业安全系统的危险应对能力,通过提高技术与设备以及环境的安全状态水平来降低危险因素对安全系统的干扰强度。

英文摘要：

In order to improve the construction enterprises ’ safety management ability,aiming at construction safety environment complexity and management limitations,a system dynamics simulation model was built for safety system vulnerability. Using BP neural network and MIV algorithm,weights were calculated for factors in the model such as employees,technological system,and management system. A vulnerability simulation was made on basis of data on 86 projects,428 questionnaires. The results show that staff system and management system affect safety system coping capacity significantly,that technology-equipment system and environment system affect safety system sensibility significantly,and that thus the construction enterprises should improve the level of staff safety behavior and manager’s management to increase the coping ability of safety system and improve safety states of technology-equipment system and environment system and improve the safety states of technology-equipment system and environment system to decrease the interference strength by negative factors on safety system.