If you've ever seen a video of smoke curling across a ceiling or fire racing up through the external building facade, you have seen what Fire Engineers try to predict every day. But in complex buildings—think tunnels, high-rises, or shopping centres—smoke does not behave in simple ways. It moves based on how the building is shaped, how the air flows, and how the fire spreads.

To predict all of the above, we use Computational Fluid Dynamics (CFD) or numerical analysis.

At its core, CFD gives insight into how air, smoke, heat, and gases move in space and time. It’s a powerful way to understand fire and smoke behaviour before anything is built. Typically, we use CFD to answer questions like:

• How will smoke and hot gases migrate through the building or enclosure?
• Can occupants reach a place of safety before tenability limits are exceeded?
• What temperature levels will develop in key zones, and how might they affect structural or system performance?
• Will visibility, temperature, or toxicity conditions allow safe firefighter access and intervention?

This article breaks down how CFD works, why it’s important in performance-based fire engineering design, and how it helps us make informed decisions.

Fires behave like fluids. Hot gases rise, cool air moves in to replace them, and smoke travels through paths of least resistance. CFD is a way of simulating that fluid movement using the laws of Physics.

Fires behave like fluids. Hot gases rise, cool air moves in to replace them, and smoke travels through paths of least resistance. CFD is a way of simulating that fluid movement using the laws of Physics.

The basic idea behind CFD is to numerically solve the equations that govern fluid motion—known as the Navier-Stokes equations. These equations describe how air, smoke, and heat move due to forces like pressure, buoyancy, and turbulence.

FDS simplifies these equations using what's called the low Mach number approximation. Fires in buildings rarely involve high-speed gas flows. By assuming that velocities remain well below the speed of sound, FDS can ignore acoustic waves and focus computation on the thermally driven flow—where buoyancy, heat release, and ventilation matter most​.

The fire space is divided into a 3D grid of small rectangular cells. Within each cell, properties like velocity, temperature, and gas concentration are calculated. The finer the grid, the more detail—but also more computation.

Here’s the key: FDS doesn’t just guess what happens. It computes it, timestep by timestep, using Physics.

Fires are chemical reactions, so a CFD fire model needs a way to represent combustion. FDS simplifies this using the mixture fraction model, which assumes that combustion occurs in a thin reaction zone where fuel and air mix. The system tracks the mass fractions of “lumped species” (air, fuel, and combustion products), and solves transport equations for these components​.

This avoids the need to simulate detailed chemical kinetics for every gas species involved. Instead, a single reaction step is assumed: fuel + air → products. The mixture fraction method is computationally efficient and sufficiently accurate for most fire engineering applications.

The amount of soot generated, and the fraction of heat released as radiation can be adjusted through empirical parameters, allowing calibration with experimental results.

FDS doesn't just simulate gases. It also models solid materials, allowing surfaces to absorb heat, burn, and release flammable gases. Each surface can consist of multiple layers, with unique thermal properties and degradation reactions. The model calculates one-dimensional heat conduction into the material and predicts when it will start to pyrolyse or ignite​

Surface boundary conditions include convective and radiative heat fluxes. This makes it possible to simulate things like flashover, fire spread across linings, and structural heat damage.

Case Study:

Imagine a vehicle catches fire in a long tunnel. The fire starts releasing thick black smoke and intense heat. Within seconds, smoke begins to fill the tunnel. Ventilation fans kick in, but is it fast enough? Can people safely evacuate?

This is exactly the kind of situation CFD is built to analyse.

In a recent project, our team used FDS to simulate several tunnel fire scenarios. We wanted to know:

• How quickly smoke would fill the space?
• What visibility levels occupants would experience during evacuation?
• How long tenable conditions would last?
• Whether jet fans should be activated immediately, or delayed to avoid spreading smoke toward people evacuating

Using FDS, we created a digital model of the tunnel geometry, added ventilation systems, and defined the fire properties based on vehicle size and fuel load. We then simulated fire growth and smoke spread over a 10-minute period.

The simulation also revealed temperature hotspots that would challenge structural elements, helping the design team reinforce those areas. None of this could have been fully predicted as accurately without CFD.

In traditional fire engineering, prescriptive codes are typically used. These codes are conservative and designed for standard buildings, but modern structures are far from standard. Today's buildings feature complex designs such as open atriums, natural ventilation, double-skin façades, interconnected basements, and hybrid occupancy types. CFD gives us a way to go beyond code. It lets us test different fire sizes, ventilation strategies, and evacuation times before construction even begins. We can see how a fire behaves in a specific building layout and make evidence-based decisions.

This matters not just for life safety, but for first responders too. By understanding where smoke will collect and how fast it will travel, we can help fire brigades plan more effectively.

CFD is powerful, but it’s not magic. The results are only as good as the inputs. That means:

• You need realistic assumptions about fire size, fuel load, and ventilation.
• The grid size must be fine enough to capture detail, but not so fine that simulations take weeks.
• Boundary conditions—like door positions or wind effects—must be chosen carefully.

It also takes engineering judgment to interpret results. Just because smoke reaches a corridor in 6 minutes, doesn’t mean evacuation is unsafe. It depends on who’s in the building, what the alarms do, and how the systems interact.

So, we treat CFD as part of a bigger toolbox. It works alongside evacuation modelling, risk analysis, and compliance pathways to build a full fire safety strategy.

At Lote, our goal is simple: to make the world safer and more secure. CFD helps us do that by giving us insight—before there's ever a flame. We use it to challenge assumptions, test alternatives, and support designs that are smarter, safer, and tailored to the real world.

Whether it’s a tunnel, a tower, or a transport hub, the question is always the same: what if there’s a fire? With CFD, we don’t just ask that question—we simulate it, analyse it, and find answers.