Study on the Construction of Safety Management System to Prevent Falling Accidents from High Places—Taking GS Company as an Example

本研究基于我国建筑业高处坠落事故高发的现实困境，结合国际安全生产理论演进脉络，提出“技术–行为–组织”三元协同治理框架。通过解构2018~2023年住建部事故通报数据，发现我国建筑业坠落事故占比(53.7%~58.2%)显著高于全球均值(ILO数据41%)，且呈现防护投入强度(德国水平的63%)与施工复杂度倒挂的典型特征。研究引入复杂系统理论，建立包含4个防御层级、12个关键控制点的坠落事故防控体系，在GS公司开展为期18个月的实证研究。结果显示：该体系使事故直接成本降低78.9万元/年，人员风险感知准确率提升42%，形成可复制的“认知干预–智能监测–文化治理”解决方案。研究创新性体现在：1) 开发基于脑机接口的安全行为训练系统；2) 建立钢结构施工动态防护数学模型；3) 创建安全信用积分制的文化治理模式。Based on the realistic dilemma of the high incidence of fall accidents in China’s construction industry and the evolution of international safety production theory, this study puts forward a three-way collaborative governance framework of “technology-behavior-organization”. By deconstructing the accident notification data of the Ministry of Housing and Urban-Rural Development from 2018 to 2023, it is found that the proportion of fall accidents in China’s construction industry (53.7%~58.2%) is significantly higher than the global average (41% ILO data), and presents a typical feature of the inverse of the intensity of protection input (63% of the German level) and construction complexity. This study introduced complex system theory to establish a fall accident prevention and control system including 4 defense levels and 12 critical control points, and carried out an 18-month empirical study in GS Company. The results show that the system can reduce the direct cost of accidents by 789,000 yuan/year, increase the accuracy of personnel risk perception by 42%, and form a replicable “cognitive intervention—intelligent monitoring—cultural governance” solution. Research innovation is reflected in: 1) Develop a safety behavior training system based on brain-computer interface; 2) Establish the mathematical model of steel structure construction dynamic protection; 3) Create a cultural governance model of the security credit score system.