Search Results (251)

Effective earthquake early warning (EEW) is essential for disaster prevention in the built environment, enabling a rapid structural response, system shutdown, and occupant evacuation to mitigate damage and casualties. However, most current EEW systems lack rigorous reliability analyses of their predictive outcomes, limiting their effectiveness in real-world scenarios—especially for on-site warnings, where data are limited and time is critical. To address these challenges, we propose a Bayesian neural network (BNN) framework based on Stein variational gradient descent (SVGD). By performing Bayesian inference, we estimate the posterior distribution of the parameters, thus outputting a reliability analysis of the prediction results. In addition, we incorporate a continual learning mechanism based on elastic weight consolidation, allowing the system to adapt quickly without full retraining. Our experiments demonstrate that our SVGD-BNN model significantly outperforms traditional peak displacement (Pd)-based approaches. In a 3 s time window, the Pearson correlation coefficient R increases by 9.2% and the residual standard deviation $SD$ decreases by 24.4% compared to a variational inference (VI)-based BNN. Furthermore, the prediction variance generated by the model can effectively reflect the uncertainty of the prediction results. The continual learning strategy reduces the training time by 133–194 s, enhancing the system’s responsiveness. These features make the proposed framework a promising tool for real-time, reliable, and adaptive EEW—supporting disaster-resilient building design and operation. Full article

Early evacuation during bushfires remains the safest strategy; however, in many realistic scenarios, timely evacuation is challenging, making safe sheltering a last-resort option to reduce risk compared to late evacuation attempts. However, most Australian homes in bushfire-prone areas are neither designed nor retrofitted to provide adequate protection against extreme bushfires, raising safety concerns. This study addresses this gap by investigating the concept of retrofitting a part of the residential buildings as attached safe rooms for sheltering and protection of valuables, providing a potential last-resort solution for bushfire-prone communities. Numerical simulations were conducted using the Fire Dynamics Simulator to assess heat transfer and internal temperature conditions in a representative residential building under bushfire exposure conditions. The study investigated the impact of the placement of the safe room relative to the fire front direction, failure of vulnerable building components, and the effectiveness of steel shutters in response to internal temperatures. The results showed that the strategic placement of safe rooms inside the building, along with adequate protective measures for windows, can substantially reduce internal temperatures. The findings emphasised the importance of maintaining the integrity of openings and the external building envelope, demonstrating the potential of retrofitted attached safe rooms as a last-resort solution for existing residential buildings in bushfire-prone areas where the entire building was not constructed to withstand bushfire conditions. Full article

In this article, two generalized net models (GNMs) are proposed to study the behavior and effectiveness of intelligent virtual agents (IVA) working in a risky environment under different scenarios and training algorithms. The proposed GNMs allow for the selection of machine learning algorithms such as intensity of characteristics Q-learning (InCh-Q), as well as the modification of multi-plan reinforcement learning (RL), proximal policy optimization (PPO), soft actor–critic (SAC), the generative adversarial imitation learning (GAIL) algorithm, and behavioral cloning (CB). The choice of action, the change in priorities, and the achievement of goals by the IVA are studied under different scenarios, such as fire extinguishing, rescue operations, evacuation, patrolling, and training. Transitions in the GNMs represent the scenarios and learning algorithms. The tokens that pass through the GNMs can be the GNMs of the IVA architecture or the IVA memory model, which are enriched with knowledge and experience during the experiments, as the scenarios develop. The proposed GNMs are formally correct and, at the same time, understandable, practically applicable, and convenient for interpretation. Achieving GNMs that meet these requirements is a complex problem. Therefore, issues related to the design and use of GNMs for the reliable modeling and analysis of the behavior and effectiveness of IVAs operating in a dynamic and risky environment are discussed. Some advantages and challenges in using GNMs compared to other classical models used to study IVA behavior are considered. Full article

Emergencies can occur unexpectedly and require immediate action, especially in aviation, where time pressure and uncertainty are high. This study focused on improving emergency evacuation in airport and aircraft scenarios using real-time decision-making support. A system based on the Asynchronous Advantage Actor–Critic (A3C) algorithm, an advanced deep reinforcement learning method, was developed to generate faster and more efficient evacuation routes compared to traditional models. The A3C model was tested in various scenarios, including different environmental conditions and numbers of agents, and its performance was compared with the Deep Q-Network (DQN) algorithm. The results showed that A3C achieved evacuations 43.86% faster on average and converged in fewer episodes (100 vs. 250 for DQN). In dynamic environments with moving threats, A3C also outperformed DQN in maintaining agent safety and adapting routes in real time. As the number of agents increased, A3C maintained high levels of efficiency and robustness. These findings demonstrate A3C’s strong potential to enhance evacuation planning through improved speed, adaptability, and scalability. The study concludes by highlighting the practical benefits of applying such models in real-world emergency response systems, including significantly faster evacuation times, real-time adaptability to evolving threats, and enhanced scalability for managing large crowds in high-density environments including airport terminals. The A3C-based model offers a cost-effective alternative to full-scale evacuation drills by enabling virtual scenario testing, supports proactive safety planning through predictive modeling, and contributes to the development of intelligent decision-support tools that improve coordination and reduce response time during emergencies. Full article

Chemical leakage accidents in chemical industrial parks pose significant threats to personnel safety, particularly during evacuation processes, where individual behavior and evacuation strategies have a considerable impact on overall efficiency. This study takes a leakage incident at an alkylation unit as a case study. First, ALOHA5.4.7 software was used to simulate the influence of meteorological conditions across different seasons on the dispersion range of toxic gases, thereby generating an annual comprehensive risk zone distribution map. Subsequently, different evacuation scenarios were constructed in Pathfinder2024.1.0605, with the integration of trigger mechanisms to simulate individual behaviors during evacuation, such as variations in risk perception and peer influence. Furthermore, this study expanded the conventional application scope of Pathfinder—typically limited to small-scale building evacuations—by successfully adapting it for large-scale evacuation simulations in chemical industrial parks. The feasibility of such simulations was thereby demonstrated, highlighting the software’s potential. According to the simulation results, exit configuration, shelter placement, and individual behavior modeling significantly affect the total evacuation time. This study provides both theoretical insights and practical guidance for emergency response planning in chemical industrial parks. Full article

The increasing number of natural and man-made disasters registered at the global level is causing a significant amount of damage. This represents one of the main sustainability challenges at the global level. The collapse of the Twin Towers, Hurricane Katrina, and the nuclear accident at the Fukushima power plant are some of the most representative disaster events that occurred at the beginning of the third millennium. These relevant disasters need an enhanced level of preparedness to reduce the gaps between the plan and its implementation. Among these actions, training and exercises play a relevant role because they increase the capability of planners, managers, and the people involved. By focusing on the exposure risk component, the general objective of the research is to obtain quantitative evaluations of the exercise’s contribution to risk reduction through evacuation. The paper aims to analyze serious games using a set of methods and models that simulate an urban risk reduction plan. In particular, the paper proposes a transparent framework that merges transport risk analysis (TRA) and transport system models (TSMs), developing serious game activities with the support of emerging information and communication technologies (e-ICT). Transparency is possible through the explicitation of reproducible analytical formulations and linked parameters. The core framework of serious games is constituted by a set of models that reproduce the effects of players’ choices, including planned actions of decisionmakers and travel users’ choices. The framework constitutes the prototype of a digital platform in a “non-stressful” context aimed at providing more insights about the effects of planned actions. The proposed framework is characterized by transparency, a feature that allows other analysts and planners to reproduce each risk scenario, by applying TRA and relative effects simulations in territorial contexts by means of TSMs and parameters updated by e-ICT. A basic experimentation is performed by using a game, presenting the main results of a prototype test based on a reproducible exercise. The prototype experiment demonstrates the efficacy of increasing preparedness levels and reducing exposure by designing and implementing a serious game. The paper’s methodology and results are useful for policymakers, emergency managers, and the community for increasing the preparedness level. Full article

In this article, the application of virtual reality technology for the realistic and immersive visualization of various tasks and scenarios in fields such as power engineering and fire safety has been examined in order to help prepare students and professional electrical engineers with electrical safety, the operation of electrical substations, potential emergencies, injury prevention, fire safety, and others. Additionally, the use of machine learning algorithms to guide evacuations from hazardous environments, fault prevention, fire prediction, and discovery of conductive materials has been examined. The most frequently used algorithms in these areas have also been described and summarized, and conclusions have been made about the combined advantages of using VR and ML algorithms. Finally, the needs, contributions, and challenges of using machine learning in virtual reality projects have been examined. Full article

Underground metro stations are integral to urban transit infrastructure, and ensuring their safety during fire emergencies is crucial. This study proposes a novel evacuation strategy for underground metro stations wherein a segment of evacuees descends to the platform level via train, while the remaining individuals evacuate upward to the ground level through station exits. A novel safety assessment methodology is established to evaluate fire evacuation efficacy, incorporating the cumulative effects of smoke, elevated temperatures, carbon dioxide, and reduced oxygen levels. Employing an actual underground metro station in Guangzhou, China, as a case study, fire and evacuation models were developed to compare the traditional upward evacuation method with the proposed partial downward evacuation strategy. The analysis reveals that both evacuation strategies are effective under the assessed fire scenario. However, the partial downward evacuation is completed more swiftly—in 385.5 s compared to 494.8 s for upward evacuation—thereby mitigating smoke inhalation risks, as the smoke height remains above the critical threshold of 1.8 m for a longer duration than observed in the upward evacuation scenario. Simulations further indicate that neither high temperatures nor carbon monoxide concentrations reach hazardous levels in either evacuation mode, ensuring evacuee safety. The study concludes that, with appropriate training arrangements and under specific fire and evacuation conditions, the partial downward evacuation strategy is safer and more efficient than upward evacuation. Full article

Metropolitan city express line tunnels are fully enclosed and often span long distances between stations, allowing multiple trains within a single interval. Traditional segmented ventilation ensures only one train per section, but ultra-long tunnels with shaftless designs introduce new challenges under fire conditions. This study investigates smoke behavior in an ultra-long inter-district tunnel during multi-train blockage scenarios. A numerical model evaluates the effects of train spacing, fire source location, and receding spacing on smoke back-layering, temperature distribution, and flow velocity. Results indicate that when train spacing exceeds 200 m and longitudinal wind speed is above 1.2 m/s, the impact of train spacing on smoke back-layering becomes negligible. Larger train spacing increases back-layering under constant wind speed, while higher wind speeds reduce it. Fire source location and evacuation spacing affect the extent and pattern of smoke spread and high-temperature zones, especially under reverse ventilation conditions. These findings provide quantitative insights into fire-induced smoke dynamics in ultra-long tunnels, offering theoretical support for optimizing ventilation control and evacuation strategies in urban express systems. Full article

This article discusses and compares approaches to the visualization of the danger zone formed as a result of spreading toxic substances during a fire at an industrial enterprise, to create predictive models and scenarios for evacuation and environmental protection measures. The purpose of this study is to analyze the features and conditions for the application of algorithms for predicting the spread of a danger zone, based on the Gauss equation and the probabilistic algorithm of a cellular automaton. The research is also aimed at the analysis of the consequences of a fire at an industrial enterprise, taking into account natural and climatic conditions, the development of the area, and the scale of the fire. The subject of this study is the development of software and algorithmic support for the visualization of the danger zone and analysis of the consequences of a fire, which can be confirmed by comparing a computational experiment and actual measurements of toxic substance concentrations. The main research methods include a Gaussian model and probabilistic, frontal, and empirical cellular automation. The results of the study represent the development of algorithms for a cellular automation model for the visual forecasting of a dangerous zone. They are characterized by taking into consideration the rules for filling the dispersion ellipse, as well as determining the effects of interaction with obstacles, which allows for a more accurate mathematical description of the spread of a cloud of toxic combustion products in densely built-up areas. Since the main problems of the cellular automation approach to modeling the dispersion of pollutants are the problems of speed and numerical diffusion, in this article the frontal cellular automation algorithm with a 16-point neighborhood pattern is used, which takes into account the features of the calculation scheme for finding the shortest path. Software and algorithmic support for an integrated system for the visualization and analysis of fire consequences at an industrial enterprise has been developed; the efficiency of the system has been confirmed by computational analysis and actual measurement. It has been shown that the future development of the visualization of dangerous zones during fires is associated with the integration of the Bayesian approach and stochastic forecasting algorithms based on Markov chains into the simulation model of a dangerous zone for the efficient assessment of uncertainties associated with complex atmospheric processes. Full article

This paper presents a detailed analysis of the dynamics of indoor environmental parameters under three simulated fire scenarios in a multi-story building, using the PyroSim platform (based on the Fire Dynamics Simulator—FDS). The study compares two smoke control strategies, organized natural ventilation (a passive system) and mechanical pressurization (an active system), evaluating their influence on temperature, differential pressure, air velocity, heat release rate (HRR), and toxic gas distribution. The simulations revealed that passive systems, relying on the stack effect and vertical natural ventilation, do not ensure the effective control of smoke infiltration into evacuation routes, allowing significant heat accumulation and reduced visibility. The results highlight the superior effectiveness of unidirectional mechanical pressurization in maintaining a stable flow regime, functional visibility, and a safe evacuation environment. A key finding is the transition from static pressure control to velocity-based flow control at the moment of door opening toward the fire source. The results confirm that a dynamically adapted application of mechanical pressurization—synchronized with the opening of access pathways—not only reinforces existing principles for protecting egress routes, but also provides a precise operational approach for optimizing emergency responses in high-rise buildings. Full article

Historical districts face pressing disaster preparedness challenges due to their special spatial properties—risks compounded by static approaches that overlook dynamic fire–pedestrian interactions. This study employs an agent-based model (ABM) for fire simulations and AnyLogic pedestrian dynamics to address these gaps in Dukezong Ancient Town, Yunnan Province, China, considering diverse ignition points, seasonal temperatures, and wind conditions. Dynamic simulations of 16 scenarios reveal critical spatial impacts: within 30 min, ≥28% of streets became impassable, with central ignition points causing faster obstructions. Static models underestimate evacuation durations by up to 135%, neglecting early stage congestions and detours caused by high-temperature zones. Congestions are concentrated along main east–west arterial roads, worsening with longer warning distances. A mismatch between evacuation flows and shelter capacity is found. Thus, a three-stage interaction simplification is derived: localized detours (0–10 min), congestion-driven delays on critical roads (11–30 min), and prolonged structural damage afterward. This study challenges static approaches by highlighting the “fast alert-fast congestion” paradox, where rapid alerts overwhelm narrow pathways. Solutions prioritize multi-route guidance systems, optimized shelter access points, and real-time information dissemination to reduce bottlenecks without costly infrastructure changes. This study advances disaster modeling by bridging disaster development with dynamic evacuation, offering a replicable framework for similar environments. Full article

As the attention to using hydrogen as a potential energy storage medium for power generation and mobility increases, hydrogen production, storage, and transportation safety should be considered. For instance, hydrogen’s extreme physical and chemical properties and the wide range of flammable concentrations raise many concerns about the current safety measures in processing other flammable gases. Hydrogen cloud accumulation in the case of leakage in confined spaces can lead to reaching the hydrogen lower flammability limit (LFL) within seconds if the hydrogen is not properly evacuated from the space. At Jülich Research Centre, hydrogen mixed with natural gas is foreseen to be used as a fuel for the central heating system of the campus. In this work, the release, dispersion, formation, and spread of the hydrogen cloud in the case of hydrogen leakage inside the central utility building of the campus are numerically simulated using the OpenFOAM-based containmentFOAM CFD codes. Additionally, different ventilation scenarios are simulated to investigate the behavior of the hydrogen cloud. The results show that locating exhaust openings close to the ceiling and the potential leakage source can be the most effective way to safely evacuate hydrogen from the building. Additionally, locating the exhaust outlets near the ceiling can decrease the combustible cloud volume by more than 25% compared to side openings far below the ceiling. Also, hydrogen concentrations can reach the LFL in case of improper forced ventilation after only 8 s, while it does not exceed 0.15% in the case of natural ventilation under certain conditions. The results of this work show the significant effect of locating exhaust outlets near the ceiling and the importance of natural ventilation to mitigate the effects of hydrogen leakage. The approach illustrated in this study can be used to study hydrogen dispersion in closed buildings in case of leakage and the proper design of the ventilation outlets for closed spaces with hydrogen systems. Full article

Existing studies on ontology evolution lack automated mechanisms to balance semantic coherence and adaptability under real-time uncertainties, particularly in resolving spatiotemporal asymmetry and multidimensional coupling imbalances in dam-break scenarios. Traditional methods such as WordNet’s tree symmetry and FrameNet’s frame symmetry fail to formalize dynamic adjustments through quantitative metrics, leading to path dependency and delayed responses. This study addresses this gap by introducing a novel symmetry-entropy-constrained matrix fusion algorithm, which integrates algebraic direct sum operations and Hadamard product with entropy-driven adaptive weighting. The original contribution lies in the symmetry entropy metric, which quantifies structural deviations during fusion to systematically balance semantic stability and adaptability. This work formalizes ontology evolution as a symmetry-driven optimization process. Experimental results demonstrate that shared concepts between ontologies (s = 3) reduce structural asymmetry by 25% compared to ontologies (s = 1), while case studies validate the algorithm’s ability to reconcile discrepancies between theoretical models and practical challenges in evacuation efficiency and crowd dynamics. This advancement promotes the evolution of traditional emergency management systems towards an adaptive intelligent form. Full article

The use of CFRP composite increased significantly since the last 40 years for aircraft structure. Unfortunately, such structures are subjected to significant damages if struck by lightning compared to metallic structure. This is mainly due to the low conductivity of this material, which cannot evacuate the current without high Joule heating. Lightning strike-induced damage in a composite laminate is composed of in-depth delamination, fibre breakage, and resin deterioration due to the surface explosion and the core current flow linked to interaction of the arc with the surface. But very rare previous studies dedicated to the analysis of damage as a direct effect of lightning have considered the spurious effect of the paint that always covers real aeronautic structures neither on the thermal nor the mechanical loads that are the root cause of these damages. We present in this paper a coupled non-destructive and destructive damage analysis to support the proposition of damage scenarios depending on the presence and thickness of the paint. The mechanical and thermal sources contribution in the global loading on the core damage is discussed, which confirms previous studies’ analysis and modelling and is in accordance with existing works in the literature. Full article