Dynamic assessment of Construction and operation Risks of Swivel Bridges Based on Bayesian Networks and deep learning

In view of the complexity and dynamics of structural safety risks during the construction and operation stages of swivel bridges, this study proposes an intelligent safety monitoring method based on the integration of Bayesian network (BN) and deep learning. By constructing a three-layer index system of "basic geology-rotation process-real-time monitoring", combining expert knowledge and engineering data to complete the learning of the network structure, and using the temporal convolutional network (TCN) to extract temporal features such as the rotation attitude and structural stress, the mapping of risk states is achieved. The (60+64) m swivel bridge of the North Extension project of Xi 'an Xingfu Road was taken as a case for verification. The results show that the fusion model is significantly superior to the single model in terms of indicators such as precision rate (0.91), recall rate (0.88), and AUC value (0.93), and the recognition ability for high-risk working conditions has improved by 32%. The research results provide quantitative tools for the precise control of structural safety during the construction and operation stages of swivel bridges, especially having significant advantages in dynamic risk prediction and multi-source monitoring data fusion under complex rotating processes.