Computational Fluid Dynamics allows fire engineers to simulate fire behaviour in virtual buildings before they're constructed. The technology is powerful — but understanding its limitations is essential.. Virtual Fires, Real Decisions Computational Fluid Dynamics (CFD) fire modelling has revolutionised fire engineering. Instead of relying solely on prescriptive rules and simplified calculations, engineers can now simulate fire behaviour in complex three dimensional spaces — predicting smoke movement, temperature distribution, visibility, and toxicity with remarkable accuracy. But CFD modelling is a tool, not an oracle. Understanding what it can and cannot do is essential for anyone commissioning, conducting, or reviewing fire modelling studies. What CFD Fire Modelling Does CFD solves the fundamental equations of fluid dynamics (Navier Stokes equations) to simulate: Fire growth and development : Heat release rate over time Smoke production and movement : Layer depth, temperature, visibility, toxicity Air flow patterns : Natural ventilation, mechanical ventilation, wind effects Temperature distribution : Gas temperatures, surface temperatures, structural member temperatures Radiation : Radiant heat flux from flames and hot gas layers Sprinkler activation : Predicting when and which sprinkler heads activate Detector activation : Predicting smoke and heat detector response times Common CFD Software in UK Practice Fire Dynamics Simulator (FDS): Developed by NIST (US National Institute of Standards and Technology) Open source and freely available The most widely used CFD fire model globally Well validated for most fire scenarios Computationally intensive — large models can take days to run SMARTFIRE: Developed by the University of Greenwich User friendly interface with automated meshing Good for evacuation studies combined with fire modelling Used extensively in UK practice PyroSim: Graphical interface for FDS Simplifies model creation and results visualisation Commercial product (Thunderhead Engineering) Does not change the underlying FDS calculations When CFD Modelling Is Needed CFD is not needed for every project. It's most valuable when: 1. Complex geometries : Atria, interconnected floors, complex ceiling shapes 2. Smoke ventilation design : Verifying that smoke control systems maintain tenable conditions 3. Extended travel distances : Justifying escape distances beyond prescriptive limits 4. Unusual fire scenarios : External fire spread, car park fires, industrial process fires 5. Structural fire engineering : Determining thermal exposure to structural elements 6. Performance based design : When departing from prescriptive guidance requires evidence 7. Dispute resolution : Providing quantitative evidence in design disputes Limitations and Common Errors Model limitations: CFD results are only as good as the input assumptions Fire growth rate assumptions significantly affect results Ventilation conditions may change during a fire (windows breaking, doors opening) Turbulence modelling introduces uncertainty at small scales Combustion chemistry is simplified compared to real fires Common user errors: Mesh too coarse — insufficient resolution to capture important flow features Incorrect boundary conditions — wall materials, leakage, ventilation openings Unrealistic fire size — using design fires that don't represent credible worst case scenarios Insufficient simulation time — stopping the simulation before steady state conditions are reached Cherry picking results — presenting favourable time steps while ignoring adverse conditions Quality assurance: Models should be verified (correct implementation) and validated (correct physics) Sensitivity analysis on key parameters (fire size, ventilation, material properties) Peer review by an independent fire engineer Results presented with appropriate caveats and uncertainty bounds The Future: AI Enhanced Fire Modelling Emerging developments include: Machine learning to accelerate CFD solving times (from days to minutes) Generative design : AI exploring thousands of fire safety configurations to optimise designs Real time digital twins : CFD models updated in real time from building sensor data Automated compliance checking : AI comparing modelling results against performance criteria Magnus Opifex maintains an in house CFD fire modelling capability using FDS, SMARTFIRE, and proprietary tools. Contact us for fire modelling consultancy.