Choosing the Right Fire Engineering Method

Prescriptive Provisions, Zone Models, CFD and Experiments

There is no single best method in fire engineering. The right method depends on the question, the available information, the level of detail required and the consequences of being wrong.

In practice, a project often starts with a simple question:

Can the applicable fire safety rules already demonstrate compliance?

If the answer is yes, a prescriptive solution may be the most suitable route. If the design falls outside those provisions, or if the team needs to compare alternative solutions, a performance-based assessment may be needed. That assessment can use a zone model, CFD, experiments, or a combination of methods.

The aim is not to select the most advanced tool. The aim is to produce evidence that is suitable for the decision.

1. Define what must be demonstrated

Before choosing a model, write the design question in one sentence. For example:

Will the proposed smoke control system keep the evacuation balcony tenable for the required period under the defined design-fire scenarios?

This sentence identifies the outputs that matter. They may include smoke-layer height, temperature, visibility, toxic species, pressure, flow rate, heat flux or time to activation.

The following points should be clear:

What decision will the analysis support?
Who will review or use the result?
Which quantity controls acceptance?
Where must that quantity be evaluated?
What time period and fire scenarios are relevant?
How much uncertainty is acceptable?

This step is important because different methods answer different questions. A compartment average may be enough for one decision. A local velocity or visibility field may be necessary for another.

2. Check the prescriptive route first

Fire safety regulations often provide a complete or partial design route. This may include required fire resistance, compartmentation, travel distance, opening protection, separation distance, smoke-control provisions, structural protection, or active-system requirements.

These provisions are not “less engineering”. They are a design method based on rules, classifications, test evidence and established assumptions. When the building fits within the scope of the provisions, following them can be clearer and more robust than creating a new performance-based argument.

Some prescriptive methods also require calculations. For example, a code may provide tables or formulas to determine a separation distance, a required fire resistance, or an acceptable level of radiant heat. The calculation is still part of the prescribed route. It is not the same as building a general-purpose fire model from first principles.

Examples from different countries

The exact terminology changes between jurisdictions, but the basic structure is common.

In Italy, the Codice di prevenzione incendi distinguishes between soluzioni conformi, soluzioni alternative and soluzioni in deroga. A conforming solution does not require an additional technical demonstration for the related performance level. An alternative solution requires the designer to demonstrate that the relevant fire safety objectives are achieved, using one of the accepted fire safety design methods.¹

In England, Approved Document B provides statutory guidance for satisfying the fire safety requirements of the Building Regulations. It contains prescriptive guidance for issues such as means of escape, fire separation, structural fire protection and access for the fire service. The document also recognises BS 7974 as a code of practice for applying fire safety engineering principles to building design.²

In Australia, the National Construction Code separates Deemed-to-Satisfy Provisions from a Performance Solution. A Performance Solution must demonstrate compliance with the relevant Performance Requirements through an accepted assessment method.³

These systems are not identical, and the approval process depends on the jurisdiction. The common lesson is simple: understand the prescribed route before deciding that a simulation is necessary.

3. Match the method to the question

Method	Typical use	Strength	Main limitation
Prescriptive provisions and codified calculations	Standard building situations and defined design checks	Clear, efficient and familiar to reviewers	Limited flexibility outside the stated scope
Zone models	Smoke and heat movement between compartments	Fast and useful for many scenarios	Limited local and three-dimensional detail
CFD / field models	Complex flow paths and local conditions	Spatially and temporally resolved results	High demands on inputs, numerics and interpretation
Experiments	Project-specific behaviour, validation and commissioning	Direct evidence for the tested arrangement	Costly and limited to the tested conditions

The methods can be combined. A typical performance-based study may use prescribed provisions for the parts of the design that already have an accepted solution, a zone model for scenario screening, CFD for a local flow question, and an experiment to reduce uncertainty in a critical physical input.

A simple decision map

The diagram is a guide, not a mandatory sequence. The decision may move between methods as the project develops.

4. Prescriptive provisions and codified calculations

The prescriptive route is usually the first option to check because it can answer the question directly. It is especially useful when:

the building type and geometry are within the scope of the code;
the required fire resistance or separation is defined;
the fuel and occupancy assumptions match the code basis;
the proposed systems follow recognised standards;
the authority expects a prescribed compliance route.

The main advantage is transparency. A reviewer can trace the design from the rule to the selected construction, system or dimension. The method is also efficient when the project is standard and the assumptions are already built into the provisions.

The main limitation is scope. A prescribed solution may not cover an unusual atrium, a complex smoke-control arrangement, a non-standard façade, a novel material, or an existing building with difficult constraints. In that situation, forcing the design into a table can be less defensible than using a performance-based method.

The right question is therefore not “prescriptive or engineering?” Prescriptive rules are themselves an engineering basis. The real question is whether they are applicable to the building and to the decision.

5. Zone models

A zone model represents a compartment with a small number of regions that have averaged properties. In a typical two-zone model, the upper region contains hot smoke and the lower region contains cooler air. The model calculates how these regions change with time and how mass and energy move through doors, vents and other openings.

NIST describes CFAST as “a two-zone fire model capable of predicting the environment in a multi-compartment structure subjected to a fire.”⁴

Zone models are often suitable for questions about:

smoke-layer height and temperature;
smoke movement between connected compartments;
approximate pressure and flow rates;
the effect of vents and mechanical extraction;
detector or sprinkler activation;
comparison of many fire sizes and ventilation conditions.

Their main advantage is speed. They allow the engineer to test many scenarios and perform sensitivity studies without the computational cost of a field model.

Their main limitation is resolution. The model does not describe every local flow around a balcony, beam, screen, stair opening or exhaust grille. If a local condition controls the decision, a zone model may not be enough.

This does not mean that CFD is automatically required. First decide whether the local effect matters to the acceptance criteria.

6. CFD and fire-specific field models

CFD divides the domain into many cells and calculates the evolution of flow, temperature and species. FDS is one of the best-known fire-specific field models. NIST describes FDS as “a large-eddy simulation (LES) code for low-speed flows, with an emphasis on smoke and heat transport from fires.”⁵

CFD is justified when the decision depends on effects that a prescriptive method or zone model cannot represent with enough detail. Examples include:

complex three-dimensional smoke movement;
atria, tunnels and large halls;
local conditions near balconies or openings;
interaction between natural and mechanical ventilation;
non-uniform smoke layers and strong mixing;
local visibility, temperature, velocity or species concentration;
detailed smoke-control layouts;
transient control logic with important spatial effects.

The justification should be linked to the decision. “The building is complex” is not enough. A stronger statement is:

The acceptance criterion concerns visibility at a balcony edge, where a local recirculation region may form. A compartment-average model cannot provide this quantity.

CFD provides more spatial detail. It does not automatically provide more accurate results. The result still depends on the design fire, material properties, soot yield, ventilation, boundary conditions, mesh, time step and model assumptions.

For this reason, a CFD assessment should normally include:

a clear design-fire basis;
a description of the relevant physical assumptions;
mesh and time-step checks;
sensitivity studies for important inputs;
relevant verification and validation evidence;
defined output metrics and acceptance criteria.

Smokeview images can help explain a simulation. They are not, by themselves, evidence that a design is safe.

7. Experiments and physical tests

Experiments answer a different question from models: what happened under the tested physical conditions?

They are useful when:

a material or burning behaviour is unknown;
a critical input is project-specific;
a smoke-control or suppression system must be tested;
existing validation evidence is not relevant to the application;
the cost of an incorrect assumption is high;
commissioning evidence is required.

The test may be small-scale, component-scale, reduced-scale or full-scale. The most useful test is not necessarily the largest one. It is the test that measures the quantity controlling the decision.

Experiments also have limits. A test at one fire size, wind condition, opening configuration or control sequence does not validate every other scenario. Scale effects, instrumentation, repeatability and boundary conditions must be considered.

A good test plan should therefore define:

the question being tested;
the acceptance criteria;
the measurements required;
the relevant boundary conditions;
how the results will be compared with a model or design target.

8. Verification, validation and uncertainty

Verification and validation apply to all methods, not only to CFD.

Verification asks whether the calculation or model has been implemented and solved correctly. For a codified calculation, this may involve checking units, inputs, arithmetic and limiting cases. For a CFD model, it may include code benchmarks, input-file checks, mesh studies and time-step studies.

Validation asks whether the method represents the physical phenomenon adequately for the intended application. It is not a universal property of a model. ASTM E1355-23 states that “Validation for one application or scenario does not imply validation for different scenarios.”⁶

The engineer should also consider user effects. These include the selected geometry, fire location, heat-release history, material properties, leakage, boundary conditions, output locations and scenario set. A detailed model with weak assumptions can be less reliable than a simple model with well-supported inputs.

For each method, document:

the assumptions;
the source and quality of the inputs;
the output quantities and acceptance criteria;
the basis for selecting the method;
the sensitivity of the result;
the remaining uncertainty and limitations.

SFPE recommends a similar process: define the problem, select a candidate model, verify and validate it, address user effects, and document the work.⁷

9. Example: smoke control in an atrium

Consider a three-storey atrium with natural smoke exhaust. The design team wants to maintain a tenable layer above an evacuation balcony for a defined period.

First, check the applicable fire safety provisions. They may already define requirements for compartmentation, smoke exhaust, openings, construction and means of escape.

Next, use simple codified checks and a zone model to compare the main design-fire and ventilation scenarios. This stage is useful for identifying the controlling cases.

CFD becomes useful if the result depends on local effects that the zone model cannot resolve, such as smoke recirculation near the balcony, the position of exhaust openings, or the interaction between the plume and the roof geometry.

An experiment or commissioning test may be justified if the main uncertainty concerns the actual exhaust flow, façade leakage, damper operation or control sequence.

The final assessment may therefore use several methods:

prescriptive provisions for the standard parts of the design;
a zone model for scenario screening;
CFD for a defined local question;
a test to confirm a critical physical input or system behaviour.

This is usually more efficient and more transparent than using CFD for every question from the beginning.

Conclusion

The choice of fire engineering method should follow the decision.

Use the prescriptive route when it applies and gives a clear answer. Use codified calculations when they are part of that route. Use zone models for compartment-scale behaviour and efficient scenario studies. Use CFD when local three-dimensional effects control the result. Use experiments when important physical behaviour is unknown or when a system must be confirmed in practice.

The best assessment is often hybrid. It uses different methods for different parts of the problem and makes the assumptions visible.

As SFPE notes, “A more complicated solution is not always a better solution unless the problem definition has confirmed that greater detail is needed.”⁷

The objective is not to use the most impressive tool. It is to provide the right evidence for the right engineering decision.

References

Corpo Nazionale dei Vigili del Fuoco, Codice di prevenzione incendi, D.M. 3 August 2015, sections G.2 and G.2.6. Official document. ↩
UK Government, Approved Document B: Fire safety, Volume 2, Buildings other than dwellings. Official guidance. ↩
Australian Building Codes Board, NCC: Application and Performance Solutions. Official NCC guidance. ↩
Walter W. Jones, Richard D. Peacock, Glenn P. Forney and Paul A. Reneke, CFAST: Consolidated Model of Fire Growth and Smoke Transport (Version 6), Technical Reference Guide, NIST SP 1026, 2009. NIST publication. ↩
National Institute of Standards and Technology, Fire Dynamics Simulator and Smokeview. Official FDS-SMV website. ↩
ASTM International, ASTM E1355-23: Standard Guide for Evaluating the Predictive Capability of Deterministic Fire Models. ASTM standard. ↩
Craig E. Hofmeister and Stephen M. Hill, “SFPE Guidelines for Substantiating a Fire Model for a Given Application,” SFPE Fire Protection Engineering Magazine, Issue 54. SFPE article. ↩