An AIoT-based smart digital twin for real-time tunnel fire safety monitoring
Publication details: Borås : RISE Research Institutes of Sweden AB, 2023Description: s. 216-223Subject(s): Online resources: In: Proceedings from the Tenth International Symposium on Tunnel Safety and Security, Stavanger, Norway, April 26-28, 2023Abstract: Road tunnels have played significant roles in the model transportation system with the development of the economy and urbanization. Fire accidents occurring in the tunnel are fatal and destructive, which may pose great threats to the trapped person and firefighters. This work proposes an AIoT-based smart digital twin to predict the real-time fire risk such as the fire size based on deep learning algorithms. The numerical model is first validated by the full-scale tunnel fire test, and then a numerical database of 30 tunnel-fire scenarios is established under different conditions. The Transformer architecture is adopted to construct the AI model, and the AI model attains an accuracy of 98% in predicting the fire information. Then the digital twin system was developed by the game engine and trained AI model. Finally, the system is demonstrated on a full-scale tunnel to predict the real-time fire size. The results show that the fire size can be accurately predicted.Road tunnels have played significant roles in the model transportation system with the development of the economy and urbanization. Fire accidents occurring in the tunnel are fatal and destructive, which may pose great threats to the trapped person and firefighters. This work proposes an AIoT-based smart digital twin to predict the real-time fire risk such as the fire size based on deep learning algorithms. The numerical model is first validated by the full-scale tunnel fire test, and then a numerical database of 30 tunnel-fire scenarios is established under different conditions. The Transformer architecture is adopted to construct the AI model, and the AI model attains an accuracy of 98% in predicting the fire information. Then the digital twin system was developed by the game engine and trained AI model. Finally, the system is demonstrated on a full-scale tunnel to predict the real-time fire size. The results show that the fire size can be accurately predicted.