The Impact of Architectural Design Parameters on Fire Risk in Double-Skin Facades: Fire Simulations and Regression-Based Assessment

Double skin façades (DSF) are widely preferred in contemporary architecture due to their advantages, such as natural ventilation and maximum daylight utilization in sustainable building design. However, vertical gaps, characteristic features of these systems, pose significant risks regarding fire safety and can accelerate the vertical spread of fire along the façade. The strong stack effect in vertical gaps during a fire can increase the air movement and cause uneven and turbulent spread of smoke and hot gases in the gap and surrounding areas. This study aims to investigate the effects of natural ventilation-based design parameters on fire behavior by analyzing the relationship between the geometric and natural ventilation configuration of the DSF interior gap and fire development with numerical models and statistical methods. In this context, a total of 400 scenarios, including different gap widths, heights, window areas, and air inlet-exit openings, were modeled using computational fluid dynamics (CFD) based fire simulations. The 400 simulations were systematically constructed to cover a wide range of natural ventilation design parameter combinations, enabling robust statistical analysis through regression modeling to quantify their influence on fire development. Multivariate regression analysis was used to quantitatively determine the effect levels of design parameters on the dependent variables of heat flow, heat release rate, fire room temperature, and velocity obtained from the simulations. The results of the analyses revealed that gap width and window area have a high degree of influence on fire development. An increase in gap width leads to a higher heat release rate (HRR) but reduces flame and smoke density and air velocity in the gap, which can delay vertical fire spread. In contrast, increasing window area decreases the fire room temperature, smoke and flame concentration, and airflow velocity, contributing to better containment during the early fire stages. These findings provide important guidance for optimizing fire safety in the early design phase of DSF systems.