The main purpose of fire simulations is to predict how smoke and heat will spread within a structure during a fire. This allows designers to evaluate, at the design stage, where smoke will accumulate, how it will affect evacuation routes, and how ventilation systems will perform under a real fire scenario.
Understanding how smoke moves within a building and how the stack effect governs this movement is critically important for effective fire safety design.
In one of our analyses, when the outdoor air temperature was 30 °C and the indoor environment remained cooler, it was observed that hot smoke rose throughout the building and was discharged through the upper-floor windows. Open windows at the ground level allowed fresh air to enter, helping to keep the lower floors smoke-free and contributing to the formation of safe evacuation routes.
CFD-based smoke simulations make these behaviors visible, enabling more informed engineering decisions. They also allow for the visualization of smoke concentration and support visibility assessments.
It can be clearly observed that, as if a line were drawn through the center of the building, hot air tends to rise and exit from the upper levels, while cooler air tends to enter from the lower levels.
One Fire, Two Models: Which One is Closer to Reality?
In fire and smoke propagation simulations, the choice of turbulence model directly determines the quality of the results. At this point, LES (Large Eddy Simulation) provides more realistic results compared to URANS, as it directly resolves the large-scale turbulent structures within the flow. This is because fire-induced flows are not steady; they are highly time-dependent and inherently unsteady.
The main reasons for preferring the LES approach are:
- It captures time-dependent vortex formation and turbulence interactions more accurately
- It resolves smoke plume oscillations and fluctuations
- It represents buoyancy effects caused by fire more realistically
- It provides a more accurate representation of the smoke layer formation beneath the ceiling
- It predicts smoke propagation speed and mixing physics more reliably
URANS, on the other hand, offers lower computational cost but models most of the turbulence through averaged behavior. As a result, it suppresses important physical details in highly transient flows such as fires. The inability to resolve large eddy structures, the damping of plume fluctuations, and the weak representation of sudden spread phenomena prevent it from fully capturing reality.
For this reason, LES is a much more powerful and reliable approach for fire simulations from an engineering perspective.
