With the rapid urbanization and development of metropolises, urban road tunnels have been constructed at an increasing rate, significantly alleviating urban traffic pressure, and improving urban resilience. Fire hazards have become a major threat to modern road tunnels due to the growing popularity of electric vehicles and high-density transportation of goods, particularly flammable materials. Asphalt pavements, as an essential component of road tunnels, may release harmful effluences and smoke under high temperatures, exacerbating the fire and adding risk to life safety. It is hence critical to investigate fire-retarding asphalt materials and their potential use in urban road tunnels pavements. This paper provides a comprehensive review of fire-retarding asphalt pavements for urban road tunnel pavements. The review covers tunnel fire generation mechanisms, evaluation methods, flame retardants for asphalt pavements, and recent developments in flame retardant technologies. By investigating these aspects, this paper aims to better understand the flammability of asphalt mixtures and asphalt pavements in urban road tunnels, promote the research of flame-retardant technology, and ultimately reduce the damage and loss caused by asphalt road tunnel fire accidents. Additionally, this study identifies the limitations of current research and provides an outlook for future research to contribute to the resilience of urban road tunnel structures and the longer service life of asphalt pavement in semi-closed road tunnels.@en

Deep learning has been widely applied to automated leakage detection and location of natural gas pipe networks. Prevalent deep learning approaches do not consider the spatial dependency of sensors, which limits leakage detection performance. Graph deep learning is a promising alternative to prevailing approaches as it can model spatial dependency. However, the challenge of collecting real-world anomaly data for training limits the accuracy and robustness of currently used graph deep learning approaches. This study proposes a deep probabilistic graph neural network in which attention-based graph neural network is built to model spatial sensor dependency. Variational Bayesian inference is integrated to model the posterior distribution of sensor dependency so that the leakage can be localized. An urban natural gas pipe network experiment is employed to construct the benchmark dataset, in which normal time-series data is applied to develop our proposed model while anomaly leakage data is used for performance comparison between our model and other state-of-the-art models. The results demonstrate that our model exhibits competitive detection accuracy (AUC) = 0.9484, while the additional uncertainty interval provides more comprehensive leakage detection information compared to state-of-the-art deep learning models. In addition, our model's posterior distribution enhances the leakage localization with the accuracy of positioning (PAc) = 0.8, which is higher than that of other state-of-the-art graph deep learning models. This study provides a comprehensive and robust alternative for subsequent decision-making to mitigate natural gas leakage from pipe networks.@en

Long short-term memory (LSTM) has been widely applied to real-time automated natural gas leak detection and localization. However, LSTM approach could not provide the interpretation that this leak position is localized instead of other positions. This study proposes a leakage detection and localization approach by integrating the attention mechanism (AM) with the LSTM network. In this hybrid network, a fully-connected neural network behaving as AM is first applied to assign initial weights to time-series data. LSTM is then used to discover the complex correlation between the weighted data and leakage positions. A labor-scale pipeline leakage experiment of an urban natural gas distribution network is conducted to construct the benchmark dataset. A comparison between the proposed approach and the state-of-the-arts is also performed. The results demonstrate our proposed approach exhibits higher accuracy with AUC = 0.99. Our proposed approach assigns a higher attention weight to the sensor close to the leakage position, indicating the variation of data from the sensor has a significant influence on leakage localization. It corresponds that the closer to the leakage position, the larger variation of monitoring pressure after leakage, which enhances the detection results’ trustiness. This study provides a transparent and robust alternative for real-time automatic pipeline leak detection and localization, which contributes to constructing a digital twin of emergency management of urban pipeline leakage.@en

The ‘travelling fire’ models have been used to describe the localised and travelling burning of uniform fuel bed in large open-plan building space. However, fuel is typically distributed non-uniformly in the built environment, leading to complex fire spread behaviours. This paper investigates the effect of non-uniform fuel load distribution on fire development in a sufficiently-ventilated space. A series of fire tests up to 3.5 MW with different wood crib layouts are categorised into two types, i.e., non-uniform and continuous, and non-uniform and discontinuous. The leading and trailing edges of the flame, height of flame, and fire spread rates are estimated using visual evidence. The non-uniform fuel load distribution fundamentally changes the spreading behaviour of fire. On a continuous wood crib, the fire spread rate and fire size are generally proportional to the fuel load density when the arrangement of the wood crib is similar. However, when wood cribs are discontinuous, the fire dynamics depend more on the localised burning size and gaps between fuels. Furthermore, very distinct fire behaviours were observed for fuel loads with different porosity. This work reveals the possible under-estimation of fire hazards of assuming evenly distributed fuel load and suggests considering design fire scenarios of non-uniform fuel load distribution in the performance-based fire safety design.@en

High traffic flow in a confined tunnel makes fire safety a critical issue. This paper proposed a digital twin framework for tunnel fire safety management in real-time, driven by dynamic sensor data and AIoT technologies. A deep learning model trained by the Transformer network and simulation dataset is used to predict real-time fire location and size. Then, the AI model is integrated into a 3D digital twin platform developed by the game engine Unity 3D. The performance of the proposed digital twin framework is demonstrated using numerical experiments and large-scale tunnel fire tests. Results show that the established AI model achieved promising accuracy in predicting fire location and power for both numerical and experimental data. The digital twin platform can also visualize the 3D fire scene that supports evacuation, firefighting, and emergency rescue. This research demonstrates the feasibility of using a 3D environment and digital twin in real-time fire safety management.@en

This study analyses the structural response of an aluminium reticulated roof structure that is constructed at Sichuan Fire Research Institute (Sichuan, China), and to be tested in fire. The structural fire behaviour under 960 localised fire scenarios is considered first, and then used to construct a database for training a modular artificial intelligence (AI) system for real-time forecasting. The system consists of several AI models, each of which predicts the displacement at a specific monitoring point. These individual predictions are then combined to generate a comprehensive forecast of the global structural-fire behaviour. The individual AI model utilized is a Long Short-Term Memory Recurrent Neural Network (LSTM RNN). The modular design allows different models to be modified or added as needed, making the system flexible and adaptable, and improving the accuracy and reliability of the predictions. The results demonstrate the effectiveness of the modular AI approach in accurately forecasting fire-induced structural collapses as indicated by the sensitivity the local models can have. The key objective of this research is to help to make informed decisions and prioritize efforts to minimize the risk of structural collapse in fire.@en

First respondents to fires in structures face severe risks as both the fire and structural behaviour are unpredictable. While structural collapse may manifest some warning signs, these signs are not always easily identified which has led to the death of many fire fighters over the years. Both fire and structural fire simulation have come a long way and are now capable of assessing the thermomechanical behaviour of structures to a good degree of accuracy. However, such simulations take hundreds or thousands of engineering and computation hours. This paper explores performing these analyses a priori and using the generated database to train a recurrent neural network for real time prediction of potential failure. The analysis is performed on an aluminium reticulated roof structure that is constructed in Sichuan Fire Research Institute (Sichuan, China) and is expected to be tested to failure in fire in 2023. One hundred localised fire scenarios were used to cover the potential fire that will be used to induce the failure of the test roof. Heat transfer analyses for each section were then performed in OpenSEES followed by thermomechanical analysis in the same software. The generated results database was then cleaned and the data at several key locations were extracted and used to train a long short term memory recurrent neural network. The results of the predictions show that the artificial intelligence model can infer results with increasing accuracy the closer the structure is to failure. The real test of the accuracy of the model, however, will be during the fire experiment on the real structure. This would be the first time an artificial intelligence model for rapid forecasting of structural response in fire is built a priori and tested against a real fire.@en