安全与环境学报

为了更合理地对建筑施工高处作业进行综合风险分析，提出了一种基于改进贝叶斯理论的风险评价方法。在基本贝叶斯理论的基础上，利用动态赋权修正原始计算权重，用评价指标的实际“分值距离”取代原始加权距离，并引入“得分均衡度”。以西北某建筑施工高处作业项目为案例进行风险评价，结果表明，该方法能较为准确地得出风险因素和施工风险变化趋势，指标间具有耦合作用。与传统方法相比，该方法改进了处于临界状态下风险等级划分的问题，实现了贝叶斯模型权重的动态更新，使指标权重更加准确、合理，且改进后的方法不仅可以计算出某个时间阶段的风险值，还可以对项目不同时间阶段的安全风险进行对比，解决了风险评价不准确的问题。

To conduct a comprehensive risk analysis of high-altitude construction work more reasonably, this research proposes a safety assessment method based on Bayesian theory. In view of working conditions at heights, based on case analysis, literature research and on-site monitoring, risk factors are identified from the four aspects of human, material, environment and management, and a preliminary establishment of a safety evaluation index system for working at heights, and qualitative and quantitative evaluation standards and indicators are proposed. The basis for judging the level of risk factors provides a theoretical basis for the combination of qualitative and quantitative risk assessment. Based on the basic Bayesian theory, comprehensive use of difference degree ranking technology, extension model theory based on tomographic analysis(AHP), and dynamic weighting method, quantitative and qualitative data are combined and analyzed, and the dynamic index weight is realized renew. Finally, the combined weighting of the Bayesian model is determined. By changing the parameter “weighted distance”, the risk assessment results are more accurate and reasonable. By introducing the parameter “score equilibrium”, the gap between the probability values of the project at different risk levels and the gap between the probability values of the risk levels at different time points is widened, and the problem of inaccurate risk evaluation caused by close or equal probabilities is solved. Taking a building construction high-altitude operation project in Northwest China as an example, the intelligent safety monitoring equipment is developed to monitor the construction site, and the sensitivity analysis of the monitoring indicators that characterize the weather and physical conditions is carried out to obtain the sensitive factors in the high-altitude operation project. The results show that the key factors considered during construction at heights in winter in Northwest China are human heart rate, environmental humidity, noise, and PM_2.5. Compared with the traditional method, this method can more accurately obtain the risk factors and the changing trend of construction risk and improve the problem of risk classification in a critical state. By obtaining the dynamic update of the combined weights of the Bayesian model, the index weights are more accurate and reasonable. The improved method can not only calculate the risk value of a certain period but also compare the safety risks of different periods of the project, which solves the problem of inaccurate risk assessment. The research results are important for solving the safety production problems of high-altitude operations.