Structural safety risk prediction method for terminal building steel roof construction considering spatial and temporal variations

• Published in: Journal of constructional steel research 2025-01, Vol.224, p.109126, Article 109126
• Main Authors: Liu, Zhansheng, Ji, Chengkuan, Shi, Guoliang, Mo, Yanchi
• Format: Article
• Language: English
• Subjects: Digital twin; Steel-mesh roof; Structural safety risk; XGBoost
• ISSN: 0143-974X
• DOI: 10.1016/j.jcsr.2024.109126

In the field of aviation, terminal building renovation and expansion projects frequently involve complex construction processes and the structure is in a non-static state during the construction process, resulting in significant safety risks accompanying these processes. In order to visualize and efficiently demonstrate the construction of airport terminal building renovation and expansion, and to provide a rapid decision-making analysis tool for the control of structural safety risks, a digital twin model for safety risk prediction of steel mesh frame structures is proposed. On this basis, taking the construction situation of an airport as an example, sensors were installed according to the results of the analysis of the finite element model, and a prediction model was established using the height parameters of twelve hydraulic jacks, the ambient temperature data and the stress data, in order to predict the stress changes of the key rods of the steel mesh frame. The results show that the proposed prediction model can accurately capture the stress distribution of the steel mesh frame during the overall jacking process. The stress prediction accuracy can reach 98 % for the key bars. Through this method combining digital twin and machine learning, this study provides a new perspective for safety risk management in airport terminal building renovation and expansion projects, and verifies its applicability and effectiveness in actual projects through empirical studies. •Applying Digital Twin Technology to Airport Terminal Renovation and Expansion Project.•The integration of multi-source data provides more comprehensive data support for stress prediction.•Responds in real time to changes in height during jacking and dynamically predicts construction safety.