This study systematically reviews the application of virtual reality (VR) in building evacuation scenarios in disaster contexts, highlighting its transformative potential to enhance preparedness, evacuation strategies, and safety training. Disasters such as fires, earthquakes, and multi-hazard emergencies pose significant challenges in densely populated urban environments, requiring innovative solutions beyond traditional methods. Analyzing 48 peer-reviewed studies (2014–2024) following PRISMA guidelines, this review focuses on VR applications in public buildings, transportation hubs, and high-risk workplaces, with VR simulations emerging as the predominant methodology. Key findings demonstrate VR’s ability to simulate realistic scenarios, improve spatial navigation, and optimize crowd dynamics and mobility accessibility. VR enhances evacuation efficiency and safety compliance by enabling adaptive training for diverse populations, including students, professionals, and vulnerable groups. In public and high-risk environments, VR addresses challenges such as visibility limitations, structural complexity, and the need for customized evacuation protocols. However, gaps remain in exploring multi-hazard environments and mixed-use spaces and ensuring scalability. Future research should integrate VR with artificial intelligence and machine learning for predictive and adaptive evacuation models. Expanding VR applications to underrepresented groups, including individuals with disabilities and the elderly, and collaborating with policymakers and urban planners are vital for translating research into practice. Overall, VR provides a scalable, adaptable, and inclusive solution for building evacuation preparedness, offering actionable insights to enhance resilience and safety in diverse architectural and disaster contexts. Its ability to transform evacuation strategies positions VR as a pivotal tool in advancing disaster management.

Ensuring firefighter safety during oil tank fires is paramount, given the substantial risks posed by thermal radiation. This study employs both the Fire Dynamics Simulator (FDS) and Areal Locations of Hazardous Atmospheres (ALOHA) software to simulate a severe oil tank fire scenario at the Zhushan Branch Power Plant, where two heavy oil tanks and multiple light oil tanks are located. The simulation framework divides the combustion scenario into 22.4 million grids with a grid size of 0.5 m, allowing a fine-resolution assessment of thermal radiation. Assuming a worst-case scenario involving n-Heptane combustion, the FDS simulation calculates essential parameters, including temperature, velocity, and soot distribution fields, and suggests a minimum safe firefighting distance of 22 m (equivalent to one tank diameter, 1D) for those equipped with personal protective equipment when exposed to a 5 kW/m² heat flux. Meanwhile, ALOHA modeling extends the safety assessment, recommending a downwind safety distance of 62 m (approximately 2D) to establish a preliminary exclusion zone, crucial in early emergency response when data may be incomplete. Additionally, a grid sensitivity analysis was conducted to validate the accuracy of the numerical results. This study underscores the importance of coupling FDS and ALOHA outputs to develop a balanced, adaptive approach to firefighter safety, optimizing response protocols for high-risk environments. The results provide essential guidance for establishing safety zones, advancing standards within fire protection and emergency response, and supporting strategy development for large-scale oil and petrochemical storage facilities.