Assessment of smoke movement from potential fires in enclosed spaces, such as underground stations, airport terminals, office buildings or shopping malls, is essential to ensure that they are safe for their occupants and in particular that, in the event of a fire, there will be sufficient time for everyone to evacuate. 

However, such an assessment is not straightforward. Different methods are available but each has its limitations.

Actual fire tests, or hot smoke trials, can be invaluable for establishing the propagation of smoke in an enclosed space, or for evaluating and demonstrating the performance of smoke control measures. But safety, cost and/or political considerations often make these impractical or limited in extent. Measurements may also be carried out using “cold” smoke - from smoke bombs or theatrical smoke generators. Thyer¹ described such tests, but highlighted the potential differences in behaviour between cold and hot smoke. Extrapolation of the findings of cold smoke tests to a hot smoke (fire) scenario is difficult, because the latter is dominated by buoyancy, whilst the former is not. Of course, if the assessment is to be undertaken before the building is completed, or at the design stage, then reliance will largely have to be placed on modelling techniques, possibly substantiated by limited fire trials upon completion of the building.

Empirical models can be employed to help identify the potential routes of transport of smoke from a fire. However, these models are based on experiments in geometrically simple configurations. Therefore, they cannot readily be used to extrapolate beyond the configuration in which the experiments were carried out.

Zone models allow some of the complexity of the configuration to be taken into account. They are relatively simple mathematical models which divide the problem into a small number of “zones” - such as the fire plume, a hot and cold layer. Typically empirical models are used to model some of these zones. They are applicable to simple spaces, such as regular-shaped rooms. For this application they are a good compromise between speed and accuracy. However, if used in the context of complex enclosed spaces, or even certain large regular spaces, e.g. modern warehousing, then the basic assumption that the flow can, a priori, be divided into a small number of distinct zones, breaks down. In this case other techniques should be employed.

In principle, CFD (Computational Fluid Dynamics) modelling provides a means to predict smoke movement without the constraints imposed by the above simpler modelling techniques. In particular, it can be used to represent complex geometries and can be readily adapted to model varying levels of physics.

CFD can be applied at the design stage as well as in assessing a finished structure. It can, potentially, be used to evaluate the effects of changes in structural design and ventilation, and to assess performance of safety measures for a range of fire scenarios. CFD results contain a wealth of information about the predicted flows; which may include velocity, temperature, smoke and gas concentrations at hundreds of thousands or even millions of points within the space.

However, CFD does have limitations. It is a complex technique to apply and requires a knowledgeable approach to produce reliable results: ill-considered application of CFD can produce predictions that may differ greatly from reality. It should be remembered that any modelling technique is, at best, an approximation.

Arguably the most credible and useful assessment of smoke movement in a building can be gained from a combination of CFD modelling, and measurements made during actual fire tests or hot smoke trials. Measurements can be used to check the validity of the simulation results. CFD modelling can then be used to explore a far wider range of scenarios than is practicable in what are usually limited fire trials.

Scope of this book

This book provides guidance on the application of CFD to the analysis of smoke movement from fires in large complex enclosed spaces such as underground stations, shopping malls, and airport terminals. The built environment is the primary area of application for CFD modelling of smoke movement, and it is against this background that the guidance is given.

Although many of the issues and techniques outlined in this book are relevant to the modelling of fire and smoke movement in simple closely confined spaces, such as a single compartment, the focus of the book is smoke movement in large and complex spaces. Therefore, some of the considerations which are very relevant to simple compartment fires, such as the downward radiation of heat from a high temperature smoke layer, are of lesser concern, and as a consequence are either covered briefly or not at all.

The main interest is the CFD modelling of smoke transport, rather than smoke production. Therefore, only relatively simple means to model the fire source are described.

The guidance covers the use of Reynolds-averaged Navier Stokes (RANS) CFD modelling of smoke movement. This approach is widely used for smoke movement analysis. It provides a reasonable compromise between accuracy and the computational resources required.

The guidance is of a practical nature. It is written from the perspective of use within fire safety engineering for assessment of life safety. The main aim of such an assessment should be to demonstrate that occupants can be protected from smoke as they make their way to safety, since it is widely accepted that most fire fatalities are due to exposure to smoke and its toxic products of combustion. 

The guidance is specific to smoke movement analysis. Thus general advice on, for example, mesh design is not covered. Instead, guidance on those aspects of mesh design which are specific to smoke movement analysis are addressed.

This book does not aim to provide a detailed description of all the modelling issues that are relevant to the prediction of smoke movement. Instead, the book provides an overview and introduction to the issues involved. A wide range of references is included. It is recommended that these are consulted to obtain a deeper understanding, and the knowledge necessary, to successfully utilise CFD in what is a challenging area of application.

Target Audience

This book is aimed at the reader who has some understanding and experience in the use of CFD but is not familiar with its application to the analysis of smoke movement. Some prior knowledge of fire safety science is assumed and recommended – it would be unwise to attempt CFD analysis of smoke movement without either a basic knowledge of fire safety science or good access to fire safety specialists. An understanding of fluid mechanics is also presumed.

Contents

1.    Introduction.

1.1      Why do CFD?.
1.2      Scope of this book.
1.3      Target Audience.

2.    Guidance.

2.1      Key steps.
2.2      Expertise of the CFD user
2.3      CFD tools.
2.4      Computational domain.
2.5      Mesh.
2.6      Physical sub-models.
2.7      Fire source specification.
2.8      Smoke generation and transport
2.9      Smoke visibility and toxicity.
2.10    Boundary conditions.
2.11    Numerical aspects.
2.12    Validation.

3.    Examples of Application.

3.1      Smoke movement in a multi-storey building.
3.2      Smoke movement in a lab-scale inclined tunnel
3.3      Smoke movement in a lab-scale atrium.
3.4      Smoke movement in a building under construction.
3.5      Smoke movement in an offshore accommodation module.
3.6      Smoke movement in an underground station.
3.7      Conclusion.

4.    References.

© British Crown Copyright 2007