International Journal of Ventilation ISSN 1473-3315 Volume 6 No 3
________________________________________________________________________________________________________________________
Book Review
Computational Fluid Dynamics in Ventilation Design
REHVA Guidebook No 10
Peter V Nielsen (editor), Francis Allard, Hazim B Awbi, Lars Davidson and Alois Schälin
Publisher REHVA: Federation of European Heating and Air-Conditioning Associations, ISBN 2-9600468-9-7
Computational Fluid Dynamics in Ventilation
Design is a new title in the REHVA guidebook
series. This guidebook is written for people who
need to use and discuss results based on CFD
predictions, and it gives insight into the subject for
those who are not used to working with CFD. It is
also written for users working with CFD who have
to be more aware of how this numerical method is
applied in the area of ventilation. The guidebook
has, for example, chapters that are very important
for CFD quality control in general and for the
quality control of ventilation related problems in
particular.
A large number of CFD predictions are made
nowadays and it is often difficult to judge the
quality level of these predictions. The guidebook
introduces rules for good quality prediction work. It
is the purpose of the guidebook to improve the
technical level of CFD work in ventilation.
The book contains the following main chapters:
• Mathematical background
• Turbulence models
• Numerical methods
• Boundary conditions
• Quality control
• CFD combined with other prediction models
• Application of CFD codes in building design
• Case studies
• Benchmark tests
The chapters Mathematical background and
Turbulence models give a short introduction to the
theory behind the methods used in CFD modelling.
The fundamental transport equations are discussed
with emphasis on ventilation applications.
Descriptions are given for two-dimensional
geometry to simplify the concepts. A user of CFD
predictions must have some knowledge of fluid
mechanics. It is important to understand conditions
such as laminar flow, turbulent flow, steady flow,
time dependent flow etc. both in connection with
________________________________________________________________________________________________________________________
291
CFD and also with measurements in rooms for
validating the predictions.
The turbulence model is a particularly important
aspect of CFD. It is obvious that room airflow will
be turbulent because of geometry and practical
velocity levels, but it will not always be a fully
developed turbulent flow. Some of the widely used
models are discussed such as the k-ε model, the SST
model and the Reynolds Stress model. The Large
Eddy simulation is also briefly mentioned.
The chapter on Numerical method illustrates the
structure of a CFD program and it demonstrates
much of the experience a user will have in using a
commercial program. Most of this chapter is based on
one-dimensional theory. The use of one-dimensional
analysis makes it possible to understand, by hand
calculation, many concepts and issues such as order
of accuracy, necessary number of grid points,
oscillations in the solution, iterations, divergence, etc.
This is demonstrated with a convection-diffusion
equation which is solved with the use of different
schemes at different velocity levels.
The chapter on Boundary conditions is especially
important in the ventilation area. Often the flow in a
room is determined by small details in the diffuser
design. This means that a numerical prediction
method should be able to handle small details in
dimensions of one or two millimetres, as well as
dimensions of several metres. This wide range of the
geometry necessitates a large number of cells in the
numerical scheme, which increases the prediction
cost and computing time to a rather high level. The
problem is overcome by applying different
simplifications such as simplified boundary
conditions, box method, prescribed velocity method
or momentum method. Continuous development of
computational capacity and speed will undoubtedly
make the direct methods with local grid refinements
or multigrid solution possible. This is illustrated in
Figure 1 in which a diffuser consists of 12 small slots
which can be adjusted to different flow directions.
________________________________________________________________________________________________________________________

Figure 1. A direct simulation of a diffuser consisting of 12 small slots which can be adjusted to different flow
directions. Development of computer capacity will make direct simulations of diffusers possible in the future.

This chapter also discusses other boundary
conditions such as the surface boundary which is
important for heat transfer predictions, free
boundary, plane of symmetry, air exit opening and
obstacle boundary.
The chapter on Quality Control is one of the more
important chapters in this guidebook. Quality
control consists of four major steps: to recognize
possible error sources, to check for them in the
simulation, to estimate the accuracy of the
simulations, and to improve the simulation
whenever possible. Two of the many examples
given in the guidebook are shown here.
There is a strong reduction in computing cost by
working with two-dimensional flow instead of
three-dimensional flow but is it possible in all
situations where the boundary conditions are “two-
dimensional”? Figure 2 shows a long hall with a
shed roof. There are complaints about strong
downdraught, but a two-dimensional prediction is
not able to show this effect. It is necessary to use a
three-dimensional transient approach to predict the
downdraught.
For each set of problems a grid independence study
should be performed. Figure 3 shows an example of
such a study, where the same case (a transient fire

Figure 2. Hall with shed roof, two-dimensional prediction and transient three-dimensional prediction.

________________________________________________________________________________________________________________________

292
International Journal of Ventilation ISSN 1473-3315 Volume 6 No 3
________________________________________________________________________________________________________________________
Figure 3. Simulation of a transient fire with 50,000 and 200,000 cells in the solution domain.
simulation) is run for various grids from coarse to
fine, and with homogeneous grids and other grids
with mixtures of prisms, tetra- and hexahedral cells.
The latter performs best. Figure 3 shows the
temperature distribution for 50,000 cells and 20,000
cells.
The use of a CFD program in connection with other
programs is discussed in the chapter CFD combined
with other prediction models. Of highest interest and
importance to ventilation design are the following:
coupling of airflow and multi-zone dynamic thermal
simulation where, especially, energy storage is an
important issue; coupling of airflow, moisture and
energy transport through walls; coupling of airflow
and multi-zone flow simulation, where the zonal
flow simulation also handles the transport of
additional components such as contaminants (e.g.
smoke, CO2, odours, moisture); and, lastly, coupling
of the airflow and emission from building materials.
The possibilities of applying CFD for simulating
airflow in a building are discussed in the chapter
Figure 4. Velocity distribution in the centre plane of an
office room ventilated by a ceiling mounted textile
terminal giving vertical ventilation.
________________________________________________________________________________________________________________________
293
Application of CFD codes in building design. A
number of areas such as: room air movement,
concentration distribution, emission from materials,
thermal comfort assessment, ventilation
effectiveness prediction and smoke management can
be evaluated by a CFD program from the conceptual
design to the preliminary design and right through to
the final detail design stages.
The chapter, Case studies, shows different practical
applications of predictions made at different stages
in the initial design, detail design and
commissioning phases. In particular, four different
air distribution systems are studied by CFD and
compared to measurements. The four systems are
mixing ventilation with a wall-mounted diffuser,
vertical ventilation, displacement ventilation, and
mixing ventilation generated by a ceiling mounted
radial diffuser. All systems are designed to handle
the same load in the same room. Figure 4 shows the
results for vertical ventilation.
The final chapter is a small chapter discussing
different types of benchmark tests. A benchmark
test can be used for new beginners in CFD to obtain
a fast insight into different problems for the
prediction of ventilation and to obtain an initial
experience by comparing CFD outputs.
REHVA - the Federation of European Heating and
Air-Conditioning Associations represents more than
110000 building engineers in 29 European
countries. REHVA Guidebooks are appreciated for
their solid scientific quality and are also easily
understood by practitioners. All REHVA
Guidebooks can be ordered online on www.rehva.eu
Peter V. Nielsen
________________________________________________________________________________________________________________________
International Journal of Ventilation - Notes to Authors
The International Journal of Ventilation aims to accept and
publish papers of the highest quality. Papers submitted to the
journal should fulfil one or more of the following categories:
• Introduce a novel idea concerned with the development or
application of ventilation;
• Present a validated case study demonstrating the
performance of a ventilation strategy;
• Studies on ventilation needs and solutions for specific
building types including: offices, dwellings, schools,
hospitals, parking garages, urban buildings and recreational
buildings;
• Numerical methods for application in design;
• Practical measurement techniques;
• Related issues in which the impact of ventilation plays an
important role (e.g. the interaction of ventilation with indoor
air quality, health and comfort);
• Energy issues related to ventilation (e.g. low energy systems,
ventilation heating and cooling loss);
• Driving forces (weather data, fan performance etc).
Submission Details
Papers should not exceed 4MB in size and should be submitted,
in electronic form (e.g. MSWord, WordPerfect, Star/Open Office
or ‘txt’), by email, to: mliddament@ijovent.org. It is essential that
papers comply with the preparation instructions presented below.
Preparation Instructions
Size of Paper
A maximum of 12 A4 or Letter format pages should be submitted
based on:
• 11pt Times Roman font (approx 4500 – 6000 words);
• Figures submitted in final print size (see layout of submitted
papers);
• Tables submitted in final print size and fully formatted (see
layout of submitted papers).
Exceptionally, longer papers may be accepted subject to content
and space availability.
Layout of Submitted Papers
Flexibility in presentation is permitted subject to the following:
• Numbered headings with a maximum of 2 levels e.g.
1. ……..
1.1 ………
• Abstract: This should not exceed 150 words and should
explain both the objectives of the work and outline the
results and conclusions. This section should end with a
selection of key words.
• Introduction: This should introduce and explain the concept
of the work. It should also explain the structure of the paper
and outline its uniqueness;
Author Copyright and Declaration: Please visit www.ijovent.org.uk/AuthorCopyright for copyright and declaration details.
________________________________________________________________________________________________________________________
294
• Conclusions: This should summarise the essential results
and benefits of the work. This Section should be in the
region of 200 to 300 words in length.
• References: Within the body of the text references should
be made by first author (et al) and date, e.g. (Waters et al
1999). In the reference section, references should be
presented in alphabetical order in the following format:
Waters JR, Simons MW and Grazebrook J. (1999) “Air
distribution and air quality in a large open space”, Building
Serv. Eng Res Technol. 20 pp195 – 200.
• Acknowledgements (Optional): To be included, as
necessary, to acknowledge funders and co-workers etc.
• Appendix (Optional): An appendix may be included,
especially to outline a complex numerical procedure that
may be of value to researchers but would upset the flow of
the paper.
• Figures: These should be referenced in the text as Figure 1,
Figure 2 etc; Figures must be presented, in the size in which
they are to be published, in electronic format. For each
paper, there is a maximum of 5 pages available for figures
(A4 or Letter size). Authors may submit figures in 1 page,
half page and quarter page format subject to this 5-page
quota. The submitted figures must be legible (line thickness,
style etc) and the smallest print font, at final size, should not
be less than 8pt. Colour should be avoided. Colour graphics
must translate into unique grey scale.
• Tables: These must be submitted, fully formatted, in the size
in which they are to be published.
Papers submitted in ‘txt’ format must also be sent as a hard
copy.
Review Process
An aim of the journal is to publish papers as rapidly as possible.
To achieve this, papers will undergo a two-stage review, i.e:
Pre Review: This process is aimed at filtering papers that do not
fit within the specified format. It is essential that papers be
submitted to the correct format (especially in relation to
overall size and legibility of figures and tables). Authors of
papers rejected at this stage will be encouraged to resubmit.
Peer Review: After screening, papers will be sent for peer
review. This review will be based on technical content, accuracy
and appropriateness for the journal. This process will sort papers
according to:
• Accepted with none or only minor revision;
• Accepted subject to substantial revision;
• Rejected (major errors, duplicated work or does not fit the
ethos of the journal).
Fast Tracking: Papers deemed by reviewers to be of high quality
and requiring no more than minor revision will be fast-tracked
into publication.