Wind Tunnels Design Construction, Types and Usage Limitations Chaplin

@seismicisolation
@seismicisolation
@seismicisolation
@seismicisolation
MECHANICAL ENGINEERING THEORY AND APPLICATIONS
WIND TUNNELS
DESIGN/CONSTRUCTION,
TYPES AND USAGE LIMITATIONS
No part of this digital document may be reproduced, stored in a retrieval system or transmitted in any form or
by any means. The publisher has taken reasonable care in the preparation of this digital document, but makes no
expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No
liability is assumed for incidental or consequential damages in connection with or arising out of information
contained herein. This digital document is sold with the clear understanding that the publisher is not engaged in
rendering legal, medical or any other professional services.
@seismicisolation
@seismicisolation
MECHANICAL ENGINEERING
THEORY AND APPLICATIONS
Additional books in this series can be found on Nova’s website

under the Series tab.
Additional e-books in this series can be found on Nova’s website

under the e-book tab.
@seismicisolation
@seismicisolation
MECHANICAL ENGINEERING THEORY AND APPLICATIONS
WIND TUNNELS
DESIGN/CONSTRUCTION,
TYPES AND USAGE LIMITATIONS
SUSAN B. CHAPLIN
EDITOR
New York
@seismicisolation
@seismicisolation
Copyright © 2013 by Nova Science Publishers, Inc.
All rights reserved. No part of this book may be reproduced, stored in a retrieval system or
transmitted in any form or by any means: electronic, electrostatic, magnetic, tape,
mechanical photocopying, recording or otherwise without the written permission of the
Publisher.
For permission to use material from this book please contact us:
Telephone 631-231-7269; Fax 631-231-8175
Web Site: http://www.novapublishers.com
NOTICE TO THE READER
The Publisher has taken reasonable care in the preparation of this book, but makes no
expressed or implied warranty of any kind and assumes no responsibility for any errors or
omissions. No liability is assumed for incidental or consequential damages in connection
with or arising out of information contained in this book. The Publisher shall not be liable
for any special, consequential, or exemplary damages resulting, in whole or in part, from
the readers’ use of, or reliance upon, this material. Any parts of this book based on
government reports are so indicated and copyright is claimed for those parts to the extent
applicable to compilations of such works.
Independent verification should be sought for any data, advice or recommendations

contained in this book. In addition, no responsibility is assumed by the publisher for any
injury and/or damage to persons or property arising from any methods, products,
instructions, ideas or otherwise contained in this publication.
This publication is designed to provide accurate and authoritative information with regard
to the subject matter covered herein. It is sold with the clear understanding that the
Publisher is not engaged in rendering legal or any other professional services. If legal or any
other expert assistance is required, the services of a competent person should be sought.
FROM A DECLARATION OF PARTICIPANTS JOINTLY ADOPTED BY A
COMMITTEE OF THE AMERICAN BAR ASSOCIATION AND A COMMITTEE OF
PUBLISHERS.
Additional color graphics may be available in the e-book version of this book.
Library of Congress Cataloging-in-Publication Data
ISBN: (eBook)
Published by Nova Science Publishers, Inc. † New York
@seismicisolation
@seismicisolation
CONTENTS
Preface vii
Chapter 1 The Characteristics of the ABRI Wind Tunnel 1
J. J. Miau, Z. L. Chen and C. C. Hu
Chapter 2 Analysis of Wind Environment and Air
Quality in Densely Populated Areas using
Wind Tunnel Experiments 35
Mahmoud Bady
Chapter 3 Wind Tunnel Investigation into the
Drag Characteristics of Catamaran Form 73
I. K. A. P. Utama and A. Jamaluddin
Chapter 4 Environmental Wind Tunnels 105
Z. Janour and K. Jurcakova
Chapter 5 Use of Wind-tunnels to Analyse Wind
Loading on Scaffold Structures 125
H. Irtaza and R. G. Beale
Index 149
@seismicisolation
@seismicisolation
@seismicisolation
@seismicisolation
PREFACE
In this book, the authors discuss the design and construction, types and
usage limitations of wind tunnels. Topics include the use of wind tunnels to
analyze wind loading on scaffold structures; the characteristics of the
Architecture and Building Research Institute wind tunnel project; analysis of
the wind environment and air quality in densely populated areas using wind
tunnel experiments; wind tunnel investigation into the drag characteristics of
catamaran form; and environmental wind tunnels designed and constructed to
investigate a wide range of aerodynamic tasks.
Chapter 1 - In 2001, the ABRI (Architecture and Building Research
Institute, Ministry of the Interior, Taiwan) initiated a project to establish an
environmental wind tunnel in the Kui-Ren campus of National Cheng Kung
University. This project was completed in 2004. The wind tunnel is of a
closed-return type, featuring two test sections in series. The primary test
section is 3 m by 2.6 m in cross section and 36.5 m in length; the secondary
test section is 6 m by 2.6 m in cross section and 21 m in length. The wind
tunnel uses a 500 KW axial fan. Immediately after the completion of the wind
tunnel, the wind tunnel was calibrated. The results of the calibration indicate
that the maximum speed at the inlet of the primary test section is 36 m/s and
that the turbulence intensity and non-uniformity of the flow measured at the
inlet of the main test section are less than 0.3% and 0.4%, respectively. At a
flow of speed more than 30 m/s, the energy ratio calculated from the fan test
data, is about 0.94, which almost coincides with that predicted by the design
data. The flow quality of the wind tunnel was further examined using an
experiment with two circular cylinders at Reynolds numbers in the critical
regime. The base pressure coefficients measured were found to be in good
agreement with those reported in the literature. The oil-film flow visualization
@seismicisolation
@seismicisolation
viii Susan B. Chaplin
performed in this experiment further provides information regarding the flow

patterns corresponding to different flow states in the critical regime.
Chapter 2 - This chapter presents an experimental investigation of wind
flow characteristics and air quality along a street canyon located within a
dense urban area. Four typical models of a highly populated urban area are
studied and wind tunnel experiments are carried out over an extended range of
the applied wind directions. The building patterns are represented by 1:100
scale models, where wind velocity, wind pressure and tracer gas
concentrations are measured along the two sides of the street.
A serious problem associated with wind tunnel tests on the flow around
buildings is that of blockage. To overcome such problem, an experiment for
determining the building blocking effect was carried out and a correction
factor was estimated and considered in the measured pressures. Details of such
experiment are given at the end of this chapter.
The study results provide evidence that building configurations and wind
directions are very important factors in determining both wind flow and
pollutant dispersion characteristics within urban domains. Also, the results
demonstrate that gaps between buildings are a very important factor to be
considered by urban planners and designers, because, for a given building
height, larger gaps induce more wind in urban canyons, thus improving the
ventilation process.
Chapter 3 - The breakdown into the resistance or drag components of
catamarans has been widely discussed worldwide in the last 30 years. The
resistance interference (both wave and viscous parts) has been the major part
among the components. Wave resistance interference can be rather easily
estimated using tank test, whilst the viscous component is rather complicated
to determine. Tank test can be used to estimate the skin friction, but correction
should be made attributed to interference of wave resistance on skin friction or
viscous resistance. In order to isolate the viscous resistance, hence free from
wave component, wind tunnel test was carried out. A series of tests of
catamaran forms were carried out using low speed wind tunnel. Various
configurations of slender catamaran were made in order to identify the viscous
resistance (hence the form factor) and viscous resistance interference.
The Chapter discusses the construction and experimental procedure of the
wind tunnel test. Verifications were made with tank test and CFD
(computational fluid dynamics) data in order to examine the accuracy of using
wind tunnel. Overall results demonstrate the effectiveness of using wind
tunnel test to estimate viscous resistance and its interference component.
@seismicisolation
@seismicisolation
Preface ix
Chapter 4 - A considerable variety of wind tunnels have been designed

and constructed to investigate a wide range of aerodynamic tasks. Nowadays,
environmental wind tunnels designed to simulate Atmospheric Boundary
Layer are very attractive.
They are used to determine air pollution, wind loads on buildings and
constructions, snowdrift, accident with the discharge of harmful substances,
pedestrian wind comfort, etc. Several requirements for experiments in
meteorological or environmental tasks have to be fulfilled in order to transfer
results from small scale wind tunnel experiments to full scale: 1) proper
scaling: matching the scale of the model and the boundary layer scale; 2)
matching dimensionless similarity numbers, especially the Reynolds numbers;
3) proper simulation of air flow, including distribution of velocity and
turbulence characteristics within the boundary layer; 4) acquiring the zero
pressure gradient for equilibrium boundary layer. The similarity requirements
are the starting point for the environmental wind tunnel design. Some types of
these wind tunnels are introduced in this chapter, and their construction is
briefly described. Integral part of environmental wind tunnel laboratories is
their experimental equipment.
Special devices are described, including systems for flow visualization,
for turbulent characteristics and concentrations measurements. Some examples
of the tasks solved in the environmental wind tunnels are mentioned at the end.
Chapter 5 - This chapter describes the use of wind-tunnels to investigate
the wind pressures applied to sheeted scaffold structures. A brief description of
the problems of simulating wind in a tunnel is outlined. The tests required to
calibrate a wind-tunnel are described with sample results presented. Wind-
tunnel tests on a model of a clad cubical building and on a model of the Silsoe
Experimental Building are presented together with pressure results showing
the application of the procedures. The chapter shows that providing the
procedures described are carried out that good agreement between wind-tunnel
simulations and full-scale tests can be achieved. Finally, the method of
deriving coefficients required for computational fluid dynamics calculations
for permeable netting from experimental pressure-velocity measurements
obtained in a small wind-tunnel is given.
@seismicisolation
@seismicisolation
@seismicisolation
@seismicisolation
In: Wind Tunnels ISBN: 978-1-62618-396-4
Editor: Susan B. Chaplin © 2013 Nova Science Publishers, Inc.
Chapter 1
THE CHARACTERISTICS
OF THE ABRI WIND TUNNEL
J. J. Miau1,, Z. L. Chen1 and C. C. Hu2

1
National Cheng Kung University, Tainan, Taiwan, ROC
2
Kao Yuan University, Kaohsiung,, Taiwan, ROC
ABSTRACT
In 2001, the ABRI (Architecture and Building Research Institute,
Ministry of the Interior, Taiwan) initiated a project to establish an
environmental wind tunnel in the Kui-Ren campus of National Cheng
Kung University. This project was completed in 2004. The wind tunnel is
of a closed-return type, featuring two test sections in series. The primary
test section is 3 m by 2.6 m in cross section and 36.5 m in length; the
secondary test section is 6 m by 2.6 m in cross section and 21 m in
length. The wind tunnel uses a 500 KW axial fan. Immediately after the
completion of the wind tunnel, the wind tunnel was calibrated. The
results of the calibration indicate that the maximum speed at the inlet of
the primary test section is 36 m/s and that the turbulence intensity and
non-uniformity of the flow measured at the inlet of the main test section
are less than 0.3% and 0.4%, respectively. At a flow of speed more than
30 m/s, the energy ratio calculated from the fan test data, is about 0.94,
which almost coincides with that predicted by the design data. The flow
quality of the wind tunnel was further examined using an experiment with

Corresponding author Email address: [email protected].
@seismicisolation
@seismicisolation
2 J. J. Miau, Z. L. Chen and C. C. Hu
two circular cylinders at Reynolds numbers in the critical regime. The

base pressure coefficients measured were found to be in good agreement
with those reported in the literature. The oil-film flow visualization
performed in this experiment further provides information regarding the
flow patterns corresponding to different flow states in the critical regime.
Keywords: Environmental wind tunnel, closed-return loop, circular cylinder,

critical regime
1. INTRODUCTION
The effect of wind on buildings, civil structures and bridges is a great
concern for public safety and the economy. From the viewpoint of fluid
dynamics, the flow over a structure or multiple structures induces pressure
forces upon the surfaces and produces complex phenomena in the
surroundings. Strong wind may cause damage to buildings and structural
elements and can induce motion in tall buildings and bridges (Cermak et al.,
1966; Cermak, 1975; Cermak 1976; Houghton and Carruther, 1976).
However, the effects of light wind are predominant considerations in studies of
air pollution.
The ABRI, a governmental agency, must evaluate the building codes of
the country and propose modifications, if necessary. In 2000, ABRI made a
general agreement with National Cheng Kung University to establish several
large-scale facilities in the Kuei-Ren campus of the university, in Tainan, in
order to fulfill this mission. One of the facilities was the environmental wind
tunnel, addressed in this study.
In order to establish the wind tunnel facility, in 2001, ABRI formed a
committee with members from several universities in Taiwan, namely,
Professors C. M. Cheng and K. C. Woo from Tamkang University, Professor
C. R. Chu from National Central University, Professor S. K. Ren from Cheng-
Shiung University and Professors J. H. Chou and J. J. Miau from National
Cheng-Kung University. This committee was chaired by Professor J. J. Miau.
The committee was tasked to provide the technical assistance for the project,
including the design of the wind tunnel, the monitoring the integration process
for the wind tunnel, and calibration measurements to validate the design of the
wind tunnel.
This facility was completed at the end of 2004. It is of a closed-return
circuit and can be switched to an open loop to allow flow visualization and
@seismicisolation
@seismicisolation
The Characteristics of the ABRI Wind Tunnel 3
particle dispersion experiments. The wind tunnel features two rectangular test
sections in series. The primary test section is 4 m by 3 m in cross section and
is used mainly for testing building models; the secondary test section is 6m by
2.6m in cross section and is used mainly for testing bridge models.
This study provides a summary of the experience gleaned in the past ten
years in the design and validation of the wind tunnel. In the following, the
design aspects of the wind tunnel are given firstly. The construction and
integration of the wind tunnel are then described, followed by the calibration
results for the wind tunnel. Finally, the results of an experiment using two
circular cylinders at critical Reynolds numbers are given and compared with
those reported in literature.
2. DESIGN ASPECTS OF THE ABRI WIND TUNNEL

The ABRI environmental wind tunnel was designed mainly for the testing
of building and bridge models. However, in the initial phase of the design
(Miau et al., 2004), it was realized that the requirements for building and
bridge testing are in conflict, in regard to the cross-sectional shapes required
for the test sections. A test section for the testing of building models normally
has a low aspect ratio in cross section, whereas a test section for the testing of
bridge models is preferred to be of a high aspect ratio. Unless the test section
could be large enough to accommodate both requirements, multiple test
sections were the only option.
This wind tunnel uses two test sections in series, in a closed loop. One of
the two test sections is 4 m by 2.6 m in cross section, in which the flow speed
ranges from 1 and 30 m/s, and the other is 6 m by 2.6 m in cross section, for
which the maximum speed is 20 m/s. The former test section, called the
primary test section, was designed mainly for the testing of building models,
whereas the latter, called the secondary test section, was for the testing of
bridge models. The wind tunnel was also designed to be operated in an open-
circuit mode, in order to conduct experiments for flow visualization and
particle dispersion in the primary test section.
Gross Features of the Wind Tunnel
Figure 1 shows a schematic drawing of the wind tunnel, i.e., a closed

single-return circuit with two test sections.
@seismicisolation
@seismicisolation
(
1)
(
( (
9)
17) 16)
Figure 1. A schematic drawing of ABRI wind tunnel. (Miau et al., 2003).
Figure 2. A perspective view of the wind tunnel and the building. (Miau et al., 2003).
The primary test section is 4 m by 2.6 m in cross section and 36.5 m in

length. The test section is long enough for the natural development of a thick
turbulent boundary near the downstream end, in order to simulate an
atmospheric boundary layer. Artificial roughness elements with spires can also
be installed at the inlet of the test section to produce an even thicker boundary
layer, to simulate the atmospheric boundary layers under different specified
@seismicisolation
@seismicisolation
environmental conditions. On the lower level of the circuit, the secondary test
section is 6 m by 2.6 m in cross section and 21 m in length. As shown in
Figure 1, the total length of the wind tunnel body is 77.9 m and its maximum
width and height are 7.4 m and 15.2 m, respectively. For the convenience of
later description, each of the wind tunnel components is assigned a number
indicated in the figure. The primary test section is denoted as component (1).
Following the flow direction, the remainder of the components are numbered
accordingly. For instance, the diffuser immediately downstream of the primary
test section is denoted as component (2).
Figure 2 shows that the wind tunnel and the laboratory building are
integrated. The wind tunnel circuit consists of two floor levels. The primary
test section is situated on the upper level and the secondary test section is
situated on the lower level. There are two control rooms near the two test
sections. A 90o corner in component (4) is actually situated on a rail, so that
the corner could be moved aside during open-circuit operation.
The wind tunnel is driven by an axial fan of 500 kW, noted as component
(13) in Figure 1. It is 4.75 m in diameter, with a center body 1.9 m in diameter.
Figure 3 shows a photo of the fan. The fan is situated on an isolated foundation
on the lower level floor. Two flexible joints are connected to the inlet and
outlet of the fan section, to prevent fan vibration propagating to other
components. Since the pitch angle of the fan blades is fixed, the speed of the
airflow is solely controlled by the rotational speed of the fan. At the maximum
flow rate, the total pressure produced by the fan was required to be at least
1000 Pa and the flow velocity in the primary test section to be no less than 30
m/s. However, the fan was also required to deliver a stable volume flow rate at
1 m/s in the primary test section, which defines the lowest flow rate to be
delivered by the fan. In addition, a 1.5 hp blower is installed outside the fan, to
cool the fan motor.
A closed-loop wind tunnel generally requires cooling devices to
continuously extract heat from the air in the tunnel circuit. At the initial stage
of the design, a trade-off study led to a decision to install a water-sprinkling
system around two 90o corners outside the building; components (17) and (18)
in Figure 1, respectively. This design is capable of cooling the wind tunnel
body effectively and produces no additional pressure loss in the tunnel circuit.
During operation, the cooling water is collected and recycled, for continuous
use. Two adjustable breather slots are located immediately downstream of the
primary test section and immediately upstream of the fan section. By adjusting
the openings of these breather slots, the amount of air exchanged between the
interior and exterior of the tunnel can be controlled.
@seismicisolation
@seismicisolation
Primary Test Section
The primary test section has dimensions of 4m by 2.6m in cross section

and 36.5m in length, which is specifically designed for testing building models
at flow speeds up to 30 m/s. Because of the thickening of boundary layers on
the walls and the presence of test models, an undesirable streamwise pressure
gradient was possible in the test section. In order to compensate for the
pressure gradient, the ceiling of the primary test section was designed to be
flexible, such that its vertical position could be varied in a range of ±300 mm.
In order to allow viewing of the model in the test section and taking
photographs during an experiment, some sections of the sidewalls have
transparent glass windows in the lower halves, as shown in Figure 4.
Figure 3. A frontal view of the fan.
Figure 4. The primary test section with a 3-D traversing mechanism. (Kao, 2005).
@seismicisolation
@seismicisolation
These windows can be opened easily, using weight-balancing

mechanisms. The test model can be placed on one of three turntables, located
at 2.8 m, 25.5 m and 31.5 m downstream of the inlet of the test section. The
farthest upstream is used for the experiments that require a uniform incoming
flow.
The other two turntables are mainly for model testing in a thickened
boundary layer. A three-dimensional traversing mechanism is provided in the
test section for the convenience of a flow field survey, as seen in Figure 4. The
traversing mechanism can be moved manually in the streamwise direction, for
a distance of 8 to 34 m downstream of the inlet. The traversing mechanism is
controlled to within an accuracy of 1mm by servo motors in a three-
dimensional region of 4m, 3.8m and 2.6m in the streamwise, spanwise and
vertical directions, respectively.
A turbulence generation device can be installed at the inlet of the primary
test section to produce variable turbulence intensities in the free stream. These
screens can be inserted or removed using a slide mechanism.
Secondary Test Section
The secondary test section has dimensions of 6m by 2.6m in cross section

and 21m in length and flow speeds up to 20m/s are possible. It is designed
mainly for experiments with bridge models. Similarly to the primary test
section, the lower halves of the sidewalls have transparent glass windows,
which can be opened easily, using weight-balancing mechanisms.
However, unlike the primary test section, the upper ceiling is not
adjustable. Only one turntable is provided in this test section. It was
anticipated that the flow quality in this test section would be inferior to that in
the primary test section, because of the limited streamwise extent of flow
management upstream.
Diffusers
As seen in Figure 1, three diffusers are provided in the wind tunnel circuit,
labeled as components (2), (10) and (15), also named diffuser #1, diffuser #2
and diffuser #3, respectively, below.
The equivalent diverging angles of the three diffusers are 4.6o, 4.2o, and
4.6 , respectively. These angles are larger than the value of 3.0o, which is
o
@seismicisolation
@seismicisolation
generally recommended to avoid the flow separation that occurs in a diffuser

(Gorlin and Slezing, 1966; Rae and Pope, 1984).
As seen in Figure 1, the two sidewalls of diffuser #2 diverge widely,
whereas the top and bottom walls appear to converge. This is due to the
geometrical constraints at the inlet and exit of this diffuser. Therefore, two
split plates are inserted vertically in diffuser #1, to divide the passage into
three partitions; in diffuser #2, two split plates, perpendicular to each other, are
inserted to equally divide the passage into four partitions. In diffuser #3, which
is 16.5 m long, no split plates are installed, because the pronounced frictional
loss might be caused by split plates. Since this diffuser is located immediately
downstream of the fan section, the flow at the inlet of the diffuser contains
fluctuations at high level, in addition to rotation, which actually hinders flow
separation.
Flow Management Sections
Flow management sections upstream of the two contraction sections

function to ensure that the flow at the inlets of the two test sections is uniform
and steady. In the section labeled (19) in Figure 1, a honeycomb made of
circular cells 19 mm in diameter and 200 mm in length, i.e., a length-to-
diameter ratio of about 10, is installed at the inlet to straighten the flow in the
streamwise direction. Downstream of the honeycomb, four screens of different
mesh-sizes are installed with sufficient spacing to reduce the intensity of the
turbulence of the flow. The screens have mesh sizes of 2, 5, 10, and 16,
respectively.
In the flow management section, labeled (7) in Figure 1, three turbulence
screens are arranged. These are mesh sizes 5, 10, and 16, respectively.
Contractions
The components immediately downstream of the flow management

sections mentioned are the contraction sections, respectively labeled (20) and
(8) in Figure 1, and also named as contraction sections #1 and #2 below. Note
that the contraction ratios of these two contraction sections are 4.71 and 1.54,
respectively, which are considerably smaller than the typical value of 6 to10
(Rae and Pope, 1984), because of the limits of the dimensions of the circuit.
Only the top and bottom walls of contraction section #2 converge, so non-
@seismicisolation
@seismicisolation
uniformity in the mean flow and the turbulence intensity at the exit of
contraction #2 are anticipated to be higher than those at the exit of contraction
#1. A design requirement was that the non-uniformity of the velocity
distribution at the exit of contraction #1 be less than 0.5%, with a turbulence
level of less than 1%. This requirement is later verified using the calibration
data.
Corners
Figure 1 shows four 90o corner sections, which are indispensable to a

closed-return circuit, in which turning vanes are installed to suppress the
secondary flow motions.
The turning vanes are made of sheet metal and contoured in a 90o circular-
arc. The gap-to-chord ratios of the vanes in the four corners are approximately
equal, i.e., 1:2.
Estimation of the Total Pressure Loss
An estimation of the pressure losses due to the wind tunnel components at

a flow speed of 30 m/s in the primary test section, called V0 hereafter, is
summarized in Table 1.
Firstly, it is noteworthy that the total pressure loss for the flow in the
entire circuit is about 550 Pa, which is equivalent to a total pressure loss
coefficient of 1.051.
Assuming that the fan efficiency is 80% at this flow speed, the fan power
required is 250 KW. Using the information in this table, the specifications of
the fan were determined as a maximum power of 500 KW and a pressure rise
of at least 1000 Pa.
Table 1 shows that the most significant pressure losses are due to the
flows in the two test sections, because they are long, to allow the development
of boundary layers. Very pronounced pressure losses also result from flows
through the screens. According to Rae and Pope (1984), the energy ratio of an
aerodynamic wind tunnel, i.e., the inverse of the total pressure loss coefficient,
is normally larger than unity.
However, the energy ratio of this wind tunnel is 0.951 only. The energy
ratio can be improved by removing the screens upstream of a test section, if it
is not in use.
@seismicisolation
@seismicisolation
Table 1. Estimation of the pressure loss for the ABRI wind tunnel
cross sectional area of main test section: 12.0m2

velocity in main test section: 30.0m/s
volume flow rate: 360.0m3/s
operation temperature: 30.0 0C
Elements pressure loss (Pa) loss coefficient Percentage (%)
(1) primary test section 98.55 0.1880 17.88
(2) diffuser #1 1.24 0.0024 0.22
(3) safety screen #1 59.23 0.1130 10.74
(4) corner #1 26.80 0.0511 4.86
(5) connection #1 1.44 0.0028 0.26
(6) corner #2 26.80 0.0511 4.86
(7)flow straightener #2
screen-1 33.88 0.0646 6.14
screen-2 56.59 0.1080 10.27
screen-3 71.95 0.1373 13.05
(8) contraction #2 0.07 0.0001 0.01
(9) second test section 31.52 0.0601 5.72
(10) diffuser #2 0.97 0.0019 0.18
breather 3.19 0.0061 0.58
(11) safety screen #2 23.46 0.0448 4.26
(12) transition #1 0.78 0.0015 0.14
(13) power section
fan housing 2.51 0.0048 0.45
fan 18.79 0.0358 3.41
(14) transition #2 1.51 0.0029 0.27
(15) diffuser #3 0.24 0.0005 0.04
(16) connection #2 0.20 0.0004 0.04
(17) corner #3 6.43 0.0123 1.17
(18) corner #4 6.43 0.0123 1.17
(19)flow straightener #1
honeycomb 8.57 0.0164 1.55
screen-1 6.88 0.0131 1.25
screen-2 14.54 0.0277 2.64
screen-3 19.99 0.0381 3.63
screen-4 26.77 0.0511 4.86
(20) contraction #1 1.99 0.0038 0.36
Total pressure loss: 551.31Pa
Sum of loss coefficients: 1.0517
Energy ratio: 0.951
Energy consumption: 248.1kW (fan efficiency is assumed to be 80%)
@seismicisolation
@seismicisolation
3. FABRICATION AND INTEGRATION

OF THE WIND TUNNEL
Table 2 shows the schedule for 2002-2004, during which the construction
of the building, the fabrication of the wind tunnel components, wind tunnel
integration and instrumentation setup were in progress. It should be mentioned
that these items were well coordinated.
After a review of the design of the wind tunnel and approval by ABRI in
2001, an architect was assigned by ABRI to work with the committee to
design the building, so that the building could be completed in time for the
integration of the wind tunnel components. The building was completed in
May, 2003. Later, the license for the building was issued by the local
government in October 2003, as noted in Table 2, after which time the
installation and integration of the wind tunnel components could be performed
within the building.
Concerning the body of the wind tunnel, the components were made of
steel, which was machined and checked at the factory, China Steel
Cooperation, Kaohsiung, Taiwan, before shipping to the Kuei-Ren Campus of
NCKU.
The fan was supplied by Voith Howden, of Germany, with a 3300 Volt
AC motor from the Tatung Company, Taiwan. The frequency inverter for fan
speed control was made by ROBICON, U. S. A. and was supplied and
integrated with the fan by a local company, Shi-Ta Cooperation.
The wind tunnel integration was contracted to China Steel Cooperation.
This work commenced on October 3, 2003 and ended with a shakedown test
on April 2, 2004. A photograph of the wind tunnel with the building is shown
in Figure 5.
Table 2. An overview of the schedule for the ABRI wind tunnel project
@seismicisolation
@seismicisolation
Figure 5. The ABRI wind tunnel with the building.
4. WIND-TUNNEL CALIBRATION
A performance validation program was initiated by the committee
members immediately after the completion of the integration of the wind
tunnel. This program included the calibration of the wind tunnel, in order to
establish a database for different operating conditions and to perform a series
of different experiments to demonstrate the capabilities of this wind tunnel.
The calibration of the wind tunnel was carried out by Professor J. J. Miau and
the other experiments were conducted by other committee members.
Specifically, Professor R. C. Chu conducted an experiment in the primary test
section on the boundary layer characteristics subjected to artificial boundary
roughness elements, to simulate atmospheric turbulent boundary layers and
Professor C. M. Cheng conducted an experiment in the secondary test section
on the effect of wind on the dynamic behavior of a bridge model. Two
additional experiments were performed in the primary test section by
Professors J. H. Chou and S. K. Ren to measure the aerodynamic forces on a
high-rise building model and by Professor K. C. Woo to determine the
diffusion of flow over buildings in a cluster. In the following, only the results
of the calibration of the wind tunnel are described. For more information
regarding the other experiments mentioned, a report was submitted to ABRI
on behalf of this committee (Miau et al., 2004).
@seismicisolation
@seismicisolation
Results of the Wind Tunnel Calibration
Table 3 shows the data for the fan, obtained on the morning of April 29,
2004, a very early test with the breathers at the downstream end of the primary
test section and the inlet of the fan open for the exchange of air with the
exterior.
The flow speed was monitored using a Pitot tube situated at the inlet of the
primary test section, with a temperature sensor positioned nearby to monitor
the air temperature with respect to time. In addition, two static pressure taps
were located upstream and downstream of the fan, in order to obtain static
pressure differential measurements. As seen in the table, the rise in the value
of the static pressure denotes the static pressure differentials obtained at a fan
speed. For each fan speed, the total pressure rise due to the fan was calculated,
according to the static pressure rise and the difference in the dynamic
pressures at the streamwise locations of the two static pressure taps.
Specifically, the dynamic pressure corresponding to a static pressure tap was
calculated, based on the known volume flow rate and the cross sectional area
where the static pressure tap was situated. The volume flow rate was
calculated in accordance with V0 and the cross sectional area of the primary
test section. Note that when the flow speed in the wind tunnel is steady over
time, the total pressure rise due to the fan is equivalent to the total pressure
loss of the flow in the tunnel circuit.
The data shown in Table 3 verifies that the fan meets the design
specifications. At the highest fan speed of 390 rpm, recommended by the fan
supplier, the total pressure rise is more than 1000 Pa, while the fan power
measured is only 412 KW, which is lower than the maximum power of
500KW. In this situation the flow velocity at the inlet of the primary test
section is higher than the 30 m/s specified in the requirement.
With regard to the efficiency of the wind tunnel, Table 4 shows the energy
ratio values obtained from a test with the breathers at the downstream end of
the primary test section and the inlet of the fan closed, which demonstrates
greater efficiency than that in Table 3. In this table, the energy ratio value was
calculated as the ratio of the dynamic pressure, using V0, to the total pressure
rise. As seen in the table, the energy ratios for fan speeds of 300rpm and above
reach a constant value of about 0.94, which is an indication that the
aerodynamic characteristics of the wind tunnel do not vary with the fan speed.
Remarkably, these energy ratio values almost coincide with the energy ratio
value of 0.951 in Table 1, reduced from the design data at V0 =30 m/s. It
should be mentioned that in Table 1 the estimation is made using a cross-
@seismicisolation
@seismicisolation
sectional area of the primary test section of 12 m2, which is about 10% larger
than the actual size. Nevertheless, this difference does not affect the predicted
energy ratio appreciably.
The energy ratio can be calculated as the ratio of the kinetic energy of the
working fluid, which is the total pressure rise multiplied by the volume flow
rate, to the power input multiplied by the electrical-to-mechanical conversion
efficiency recommended.
Accordingly, a comparison of the energy ratios using the input power and
those in Table 4 is shown in Figure 6. As seen, the values obtained using the
two methods are not generally in agreement, except at a fan speed of 300 rpm.
Physically speaking, the energy ratio values shown in Table 4 are considered
to be more accurate than those based on the power input, since the electrical-
to-mechanical conversion efficiency is not constant, but actually varies with
the fan speed.
Conceivably, the conversion efficiency could be poorer than the
recommended value at a fan speed lower than 300 rpm; but the opposite is true
at a fan speed higher than 300 rpm.
Table 3. Data for the fan test obtained with the breathers at the
downstream end of the primary test section and upstream of the inlet
of the fan open (Miau et al., 2004)
Data of the Fan Tested at Different Speeds
Data of the Fan Tested at Different Speeds
Time Fan Pitot Tube Temperature Velocity Static Total Fan
0 Pressure Pressure
2004/4/29 Speed Dynamic ( C ) (m/s) Power
V0 (m/s) rise rise
hh:mm (RPM) (Pa) (Pa) (Pa) (KW)
@seismicisolation
@seismicisolation
Table 4. The reduced energy ratio values with the breathers closed
Fan Fan Dynamic Air Velocity Total Energy

speed power Pressure density V0 (m/s) pressure ratio
(rpm) (KW) in the rise (Pa) evaluated
primary
test section
(Pa)
50 1.9 11 1.163 4.4 16 0.688
100 6.8 60 1.163 10.2 71 0.845
150 27.3 141 1.172 15.5 163 0.865
200 65.4 256 1.168 20.9 290 0.883
250 125.7 403 1.163 1.154 452 0.892
300 222 601 1.158 1.151 641 0.938
330 294.7 727 1.156 1.148 776 0.937
350 350 815 1.151 1.145 869 0.938
1.1
IAA testratio based on the power input

Energy
Voith
EnergyHowden
ratio basedtest
on the total pressure rise
1
0.9
Energy ratio
0.8
0.7
0.6
0.5
100 150 200 250 300 350 400

Runner speed(rpm)
Figure 6. A comparison of the energy ratio values reduced based on the input power
and the total pressure rise.
Figure 7 shows the measurements obtained using an accelerometer

situated in the fan, which was installed by the fan supplier. The horizontal axis
indicates the fan speed; the vertical axes on the right and left hand sides
@seismicisolation
@seismicisolation
indicate the vibration level of the fan and the frequency set by the frequency
inverter, respectively. As seen, the vibration level is less than 0.16 mm/s at the
maximum fan speed of 390 rpm, which is one order smaller than the maximum
vibration level of the fan, at 1.8 mm/s, as specified in the requirement. This
strongly implies that the building design, with an isolated foundation for the
fan, works very well. At a fan speed below 250 rpm, the fan vibration level
remains below 0.07 mm/s. Vibrations are hardly felt on touching the wind
tunnel structure. Figure 8 shows the variation in the air temperature in the
primary test section at V0= 10, 15, 20 and 30 m/s, each of which was run for
over two hours. For V0 less than 20 m/s, the temperature variations noted are
not significant. However, at V0=30 m/s, the temperature increases significantly
with time. Therefore, it is recommended that at V0 higher than 20 m/s, the
cooling system using a water spray outside the wind tunnel building be turned
on, in order to maintain a low and steady temperature.
The steadiness of the flow in the primary test section was measured using
a Pitot tube situated at the inlet of the test section. Figure 9 shows the data
obtained at four values of V0, at each of which the velocity measurements
were performed for 2 hours. Each data point shown represents the mean
velocity reduced from the Pitot tube measurements, sampled for 20 seconds at
100 Hz. As seen, over a time period of two hours the most significant
unsteadiness occurs for V0=19.6 m/s, which is 0.33% of the mean velocity.
Therefore, the long-term stability of the fan is confirmed.
60 0.2
Frequency inverter
Fan vibration
50
0.16
40
Vibration
0.12
Frequency(Hz)
Vibration(mm)
(mm/s)
30
0.08
20
0.04
10
0 0
0 50 100 150 200 250 300 350 400

Runner speed (rpm)
Figure 7. Structural vibration data of the fan at different speeds. (Kao, 2005).
@seismicisolation
@seismicisolation
36
U=10m/s
U=15m/s
35
U=20m/s
U=30m/s
34
33
Temperature(oC)
32
31
30
29
28
0 10 20 30 40 50 60 70 80 90 100 110 120

Time(min)
Figure 8. Temperature variations of the flow in the primary test section at different
flow speeds. (Kao, 2005).
Figure 9. Steadiness of flow speeds measured at x=0, at the core in the primary test
section. (Kao, 2005).
@seismicisolation
@seismicisolation
Figure 10. The one-dimensional traversing mechanism employed for the wind-tunnel
calibration. (Kao, 2005).
The mean streamwise velocity distributions near the inlet of the primary
test section were obtained using a Pitot tube and a one-dimensional traversing
mechanism. At the beginning of this experiment, it was found that the
blockage due to the traversing mechanism shown in Figure 4 was so large that
it caused an unacceptable error in velocity measurements. (Kao, 2005)
Alternatively, a one-dimensional traversing mechanism that produced much
less blockage, as shown in Figure 10, was used. The blockage due to this
traversing mechanism was also carefully studied. Figure 11 shows two iso-
value contour plots of the streamwise velocity distributions, obtained at x=0,
for V0 = 20 m/s, with the Pitot tube extended 0.9 m and 1.6 m, upstream from
the one-dimensional traversing mechanism. (Kao, 2005) Here, x=0 denotes the
streamwise location at 2.5 m downstream of the inlet of the primary test
section and y and z in Figure 11 denote the spanwise and vertical directions of
flow in the test section, respectively, where the center of the test section is
located at (y, z)= (0, 1300 mm). As shown in Figure 11, the velocity
measurements apply to a cross-sectional region, -1000 mm<y<1000 mm, and
300 mm<z<2500 mm. A comparison of the two plots in Figure 11 shows the
effect of the downstream presence of the traversing mechanism. In Figure 11a,
the mean velocity distribution in the region of z<1000 mm shows about a 1%
change in the velocity near the bottom wall; in Figure 11b, this is only about
0.5%.
@seismicisolation
@seismicisolation
(a)
(b)
Figure 11. The iso-value contour plots for the Pitot tube (a) 0.9 m and (b) 1.6 m
upstream m from the one-dimensional traversing mechanism. (Kao, 2005).
Subsequently, based on the velocity data shown in Figure 11b, the non-
uniformity of the velocity distribution reduced, in terms of the root-mean-
square of the variations, is 0.38% of the mean velocity value. Similarly, the
data obtained at V0 = 10 and 30 m/s at the same cross-sectional plane
demonstrates a non-uniformity of 0.37% and 0.34%, respectively. (Kao, 2005)
These values are lower than the 0.5 % given in the design specifications.
Figure 12 shows the mean streamwise velocity profiles and the
corresponding turbulence intensity distributions, measured along x=0 and y=0,
at V0= 10 and 20 m/s, respectively.
@seismicisolation
@seismicisolation
U/V
U/Uo
0
0.84 0.88 0.92 0.96 1 1.04 1.08 1.12 1.16 1.2
1400
U=10m/s, T.I.
U=20m/s, T.I.
1200
U=10m/s, U/Uo
U=20m/s, U/Uo
1000
800
Z(mm)
600
400
200
0
-3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5
T.I.(%)
Figure 12. Streamwise velocity profiles and the corresponding turbulence intensity
measured at x=0 and y=0. (Kao, 2005).
The measurements were carried out using a single normal hot-wire probe.
As seen, the results for the two flow conditions are quite consistent. Namely,
the boundary layer thickness is of the order of 60 mm and the freestream
turbulence intensity is about 0.35 %.
Flow angularity in the cross-sectional plane at x=0 was determined using a
cobra probe. Prior to the measurements, the cobra probe was calibrated for
inclined angles over a range of ±40, which was deemed sufficient for the
present purpose. (Kao, 2005)
Figure 13 show the iso-value contour plots for the yaw and pitch angles
measured over a cross-sectional region of -1000 mm<y<1000 mm and 300
mm<z<2300 mm, at V0 = 20 m/s, The root-mean-square of the yaw and pitch
angle variations reduced from the raw data are 0.460 and 0.3750, respectively.
However, it is noted that both the yaw and the pitch angle data in the figure
consistently show a bias. This is attributed to the uncertainty that the reference
employed in the cobra probe calibration might not coincide with that in the
wind tunnel. (Kao, 2005). This bias was later eliminated by assuming that the
flow at the core (x, y, z) = (0, 0, 1300 mm) is perfectly aligned in the x
direction. (Kao, 2005) Consequently, a vector plot, using the corrected yaw
and pitch angle data, is shown in Figure 14.
@seismicisolation
@seismicisolation
(a)
2600
2400
-0 -0.006
9
.1
2200 0.039
.1 3
39 84
-0.1
-0
2000 -0 .1 84
1800 -0.095
1600 -04.1
84 -0.050
Z(mm)
-0.18
-0.228
1400 -0.317 -0.2 73 -0.0
95
-0.3
17
1200 -0.050
1000
- 0 -0
.2 .1
800 28 84 -0.139
-0.495
-0.184
51
600
.31
-0.4
-0
62 399550 -0.006
400 -0.406
-0.3 73
-0.2
.1 .0 .0
- 0 -0 -0
200
0
-2000 -1500 -1000 -500 0 500 1000 1500 2000
Y(mm)
(b)
Figure 13. The iso-value contour plots of (a) yaw and (b) pitch angles measured by a
cobra probe at a cross-sectional plane at x=0. (Kao, 2005).
2600
2400
2200
2000
1800
1600
Z(mm)
1400
1200
1000
800
600
400
200
0
-2000 -1500 -1000 -500 0 500 1000 1500 2000
Y(mm)
Figure 14. A vector plot of flow angularity at x=0, based on the corrected yaw and
pitch angle data. (Kao, 2005).
@seismicisolation
@seismicisolation
Using this figure, the flow angularity can be described in terms of the
pitch angle, α, within =±0.3560 and the yaw angle, β, within ±0.7650.
Similarly, the measurements obtained using the Cobra probe at V0= 10 m/s
indicate that the flow angularity is limited to within a range of α= ±0.4040 and
β= ±0.9750 and that the root-mean-square of the yaw and pitch angles
measured are 0.1640 and 0.1680, respectively.
5. EXPERIMENT USING A CIRCULAR CYLINDER

AT CRITICAL REYNOLDS NUMBERS
A wealth of data concerning uniform flow over a circular cylinder has

been accumulated in the literature for Reynolds numbers of 102 to 107.
(Roshko, 1993; Zdravkovich, 1997) As noted, the flow phenomena involved
are very complicated and vary pronouncedly with the Reynolds number. For
Reynolds numbers of the order of 105, a so-called critical regime (Achenbach,
1971; Bearman, 1969; Tani, 1964; Zdravkovich, 1997) is identified, which
produces a dramatic decrease in the drag coefficient. More specifically, in the
critical regime, a series of transitions of flow states have been observed.
Namely, as the Reynolds number increases, flow transitions occur from the
sub-critical state to the one-bubble state, then from the one-bubble to the two-
bubble state. (Zdravkovich, 1997).
Physically speaking, in the critical regime, separation bubbles develop on
the surface of the circular cylinder. The development of a separation bubble
actually occurs because of laminar separation from the cylinder surface,
followed by a transition to turbulence in the separated shear layer and
subsequently reattachment to the cylinder surface. It is conceivable that the
flow phenomena referred to above are very sensitive to Reynolds number.
(Schewe, 1983; Higuchi et al., 1989).
In order to determine the flow quality of the ABRI wind tunnel, Tsai
(2006) designed a series of experiments using circular cylinders at Reynolds
numbers in the critical regime. In the following, the experimental results
obtained are presented in two parts. The first shows the base pressure
measurements of the circular cylinders employed, which are compared with
those reported in literature. The second part uses oil-film visualization of flows
around a circular cylinder at different Reynolds numbers, which evidences the
prevalence of the flow phenomena referred to above in the critical regime.
@seismicisolation
@seismicisolation
Base Pressure Measurements Using Two Circular Cylinders
Figure 15 shows a photograph of an acrylic cylinder model installed

vertically in the test section. The cylinder was situated on the first turntable,
whose center is 2.8 m downstream of the inlet of the test section. Figure 16
shows the turntable and the support mechanism beneath the floor. Using this
support mechanism, driven by a servo motor, pressure measurements at
different circumferential angles on the cylinder model were conveniently made
by rotating the turntable under computer control.
The experiments were performed using two cylinder models constructed
using different materials: one is the acrylic cylinder shown in Figure 15, which
is 300 mm in diameter, and the other is a stainless steel cylinder, 320 mm in
diameter. Since the height of the test section is 2.6 m, the aspect ratios of the
acrylic and stainless steel cylinders are 8.6 and 8, respectively. As the width of
the test section is 4 m, the geometric blockage ratios for the acrylic and
stainless steel cylinders are 7.5% and 8%, respectively. Both cylinder surfaces
were finished with a lathe and the accuracy of the dimensions (diameters) is
about 1 μm. The averaged maxima for the roughness of the acrylic and
stainless steel cylinders measured were 0.73 μm and 12.4 μm, respectively.
Therefore, the relative roughness values with respect to the diameters of the
cylinders are 2.43×10-6 and 3.88×10-5, respectively.
Figure 15. The circular cylinder model situated on the first turntable downstream of the
inlet of the primary test section.
@seismicisolation
@seismicisolation
(a)
(b)
Figure 16. (a) The first turntable in the primary test section; (b) its support mechanism.
Each cylinder has three sections; most of the pressure taps are located in
the middle section. Pressure measurements were made using Validyne DP-103
pressure transducers of a diaphragm type, which were situated inside the
cylinder model, in order to minimize the tube length to the pressure taps. In
order to prevent the pressure transducers from vibrating during the experiment,
a cushion was actually employed to hold the pressure transducers tightly. This
device could accommodate six pressure transducers, at most. Figure 17 shows
a photo of four pressure transducers situated in the stainless steel cylinder.
(Tsai, 2006) Each pressure transducer measures the pressure difference
between the pressure on the cylinder surface, denoted as P, and the static
pressure in the free stream, denoted as P0.
@seismicisolation
@seismicisolation
Figure 17. A photo of four diaphragm-type pressure transducers situated in the

stainless steel cylinder model. (Tsai, 2006).
Figure 18. A schematic drawing of a cylinder model situated in the test section. A and
B indicate the locations of the hot-wire probe for the experiments made with the
stainless steel and acrylic cylinder models, respectively. F indicates the location of the
Pitot-tube.
The experimental setup can be described using the Cartesian coordinate

system given in Figure 18, where the origin (X, Y, Z)= (0, 0, 0) is located at
the center of the circular cylinder model. In addition, the circumferential
position on the cylinder surface is denoted by θ, where θ=0 is located at the
forward stagnation point; θ is positive, if measured from θ=0 counter-
clockwise, so the base pressure was measured at θ=180o or -180o. As shown in
the figure, A and B indicate the locations of a hot-wire probe for the
experiments with the stainless steel and acrylic cylinder models, respectively.
The hot-wire velocity signals provide further information about the unsteady
behavior of the wake flow. (Miau et al., 2011).
@seismicisolation
@seismicisolation
During the experiment, the signals measured by the pressure transducers

and the hot-wire probe were digitized and sampled simultaneously, at a rate of
2000 samples per second. It is noteworthy that the sampling frequency is much
higher than the characteristic frequency of vortex shedding, which is of the
order of 10 Hz. The experiment with the acrylic cylinder used a sample time of
30 seconds for each measurement and the experiment with the stainless steel
cylinder used a sample time of 16 seconds. The Reynolds number, denoted as
Re, is defined using D, the diameter of the circular cylinder, and Vref, the
characteristic velocity. Vref is defined as the velocity reduced from the
difference between the pressure measurement at θ=0 and P0. By this
definition, the time-averaged pressure coefficient at θ=0 is unity. During the
experiment the free stream velocity was also measured using a Pitot tube
immediately downstream of the inlet of the test section. A comparison of these
two velocity values at various Reynolds numbers shows that the Vref values are
always higher than the free stream velocities measured. (Tsai, 2006) The
differences vary from 6.42 % to 3.33 %, as Re increases from 1.85 to
5.82×105. The variations are attributed to the significance of the wake
blockage at different Reynolds numbers. (Miau et al., 2011) Below, (1) shows
the definition of the pressure coefficient, Cp, where denotes the time-averaged
pressure measured on the cylinder, P0 denotes the static pressure of the free
stream and ñ denotes the density of the working fluid.
P  P0 (1)
CP 
1  V 2
2 ref
In (1), if is replaced by P, the instantaneous pressure measured, then Cp

also represents the instantaneous pressure coefficient.
The definition of the base pressure coefficient, Cpb, is given in (2), where
Pb denotes the time-averaged base pressure measured at θ=180o.
Pb  P0 (2)
C pb 
1  V 2
2 ref
A comparison of the base pressure coefficients, Cpb, for the acrylic and
stainless steel circular cylinders, and those found in the literature (Bearman,
1969; ESDU, 1980) is given in Figure 19. (Miau et al., 2011). The data for
both cylinders consistently shows that the initial transition from the sub-
critical to the critical state takes place in the neighborhood of Re=2×105. This
@seismicisolation
@seismicisolation
is in good agreement with the values reported by Bearman (1969) and the
ESDU data (1980).
Figure 19. Base pressure coefficients versus the Reynolds numbers measured. (Miau et
al., 2011).
In the figure, the circular cylinder model used by Kao (2005) is the same
as that used in the present experiment. Kao (2005) performed his experiment
at four Reynolds numbers only during the wind tunnel calibration and the base
pressure coefficients shown in the figure are in good agreement with those of
the present experiment. Also noted in the figure is that at Reynolds numbers
beyond the pronounced transition of the base pressure coefficient, say, higher
than 4×105, there is a discrepancy between the Cpb values obtained in this
study, about -0.4, and in Bearman (1969), which reports about -0.2, and the
corresponding ESDU data (1980) falls between these values. An explanation
for this is that in this flow regime the flow around a circular cylinder is in a
two-bubble state, for which the processes of flow separation and reattachment
are strongly dependent upon the surface roughness of the cylinder model, the
free stream turbulence intensity, and the blockage and aspect ratios of the
cylinder model. (Miau et al., 2011).
Oil- Film Flow Visualization
It is challenging to conduct a flow visualization experiment at Reynolds

numbers of 105, to demonstrate these flow phenomena. Humphreys (1960)
employed a thread visualization technique, which revealed the formation of
@seismicisolation
@seismicisolation
spanwise cell patterns on the cylinder surface at Reynolds numbers around

3×105, which is an indication of three-dimensional flow behavior. Schewe
(2001) demonstrated an oil-film photo on one side of the cylinder at a
Reynolds number pertaining to the pre-critical state, which clearly shows two
separation lines at θ=73o and 92o, which infers the presence of a laminar
separation bubble under this flow condition.
Tsai (2006) performed an oil-film flow visualization experiment in the
ABRI wind tunnel, using the stainless steel circular cylinder mentioned. The
purpose of this experiment was to gain a better understanding of the flow
characteristics of different flow states in the critical regime. (Tsai, 2006; Tsai
et al. 2010) Prior to experiment, the oil film was prepared by mixing paraffin
oil with titanium dioxide and applied to the surface of the circular cylinder.
Subsequently, after the wind tunnel was turned on and reached a steady flow
condition for 30 seconds, the limiting streamline pattern on the cylinder
surface was revealed, and prepared for photography.
Figure 20a shows a photograph obtained at Re=3.09×105 that pertains to
the initial transition from the sub-critical to the critical regime. Figure 19
shows the base pressure coefficients obtained under this flow condition to be
about -0.9. In comparison with the coefficient measured at Re=2×105, on the
verge of the critical transition, which is about -1.1, this flow condition is
referred to as the pre-critical state. In Figure 20a, a separation line is seen at
θ=100o. This separation line is interpreted to be an indication of the farthest
downstream location associated with flow separation on the cylinder. (Tsai,
2010) In this situation, the flow features the intermittent presence of separation
bubbles on the cylinder surface. A sampled signal trace of the pressure
measurements obtained at θ=-75o and Re=3.08×105 is included in Figure 20b,
for comparison, in which Cp denotes the instantaneous pressure coefficient. A
slight difference between the Reynolds numbers for Figures 20a and b is
because the two experiments were conducted separately. As seen in Figure
20b, the Cp values fluctuate between -1 and -2. As noted, the fluctuations are
very pronounced, and are attributed to the intermittent presence of a separation
bubble that develops on the circular cylinder. Hence, the separation line
indicated in Figure 20a is explained by flow separation downstream of the
separation bubble. Achenbach (1968) conducted skin friction measurements
around a circular cylinder at Re=2.6×105 and found that zero skin friction was
measured at θ=94o, which was considered to be a point of flow separation.
Achenbach’s statement (1968) is evidenced by the flow visualization in Figure
20a.
@seismicisolation
@seismicisolation
Figure 21 shows two photographs of a one-bubble state, obtained at

Re=3.64×105, which show a remarkable difference between the flow patterns
on the two sides of the cylinder.
(a)
(b)
Figure 20. An oil-film flow visualization photo of the stainless steel cylinder at
Re=3.09×105. (b) A sampled signal trace obtained at θ=-75o, for Re=3.08×105, over a
time period of 10 seconds. (Tsai et al., 2010).
In Figure 21a, the flow pattern on the negative θ side shows that an
attached turbulent flow prevails over the region from θ=-110o to -130o,
because very little oil-film remains in the region. This appearance further
implies that a separation bubble sits stably on the cylinder surface, upstream of
θ=-110o. It is also noted that the region of the attached flow does not appear
uniformly along the spanwise direction. The appearance indicates that the
mean flow in the reattachment region behaves three-dimensionally. In Figure
21b, the flow pattern on the positive θ side shows that a separation line is
@seismicisolation
@seismicisolation
identified slightly upstream of θ=90o, which tilts at an angle to the spanwise

axis. In the region of θ=110o to 130o, although the oil-film pattern appears to
be similar to that seen on the other side, comparatively more oil substance
remains.
(a)
(b)
Figure 21. Oil-film flow visualization photographs obtained at Re=3.64×105, (a) on the
negative θ side and (b) on the positive θ side. (Tsai et al., 2010).
This is explained by the fact that during the experiment, a separation

bubble may develop on this side for a period of time, but later switches to the
other side. (Tsai, et al., 2010) Figures 22a and b show two photographs
obtained on two sides of the cylinder, respectively, at Re=3.95×105, which
indicates a two-bubble state. On either side, the flow pattern is similar to that
seen in Figure 21a, which infers that a separation bubble is situated stably
upstream of the reattached flow region from θ=110o to 130o or -110o to -130o.
@seismicisolation
@seismicisolation
Overall, the oil-film flow visualization method successfully demonstrates

the mean flow patterns around a circular cylinder at three Reynolds numbers
relevant to the pre-critical state, the one-bubble state and the two-bubble state,
respectively. The information obtained explains the flow characteristics in the
critical regime.
Figure 22. Oil-film flow visualization photographs obtained at Re=3.95×105; (a) on the
negative θ side and (b) on the positive θ side. (Tsai et al., 2010).
CONCLUSION
A systematic description of the aspects of the design and performance
validation of the ABRI wind tunnel is provided in this study. The wind tunnel
notably features two long test sections in series, so its energy ratio is
significantly lower than that normally found in an aerodynamic wind tunnel. It
is worthy of note that the energy ratio reduced from the calibration data at a
fan speed higher than 300 rpm is almost the same as that reduced from the
design data for V0=30 m/s. This finding strongly infers that the design method
employed is satisfactory and accurate. According to the calibration data for the
wind tunnel and the results for the experiment with two circular cylinders at
critical Reynolds numbers, it can be confidently stated that the flow quality of
this wind tunnel is comparable to that of the wind tunnels reported in
literature. This wind tunnel is now being operated by ABRI and provides
services to governmental agencies, academic institutions and industry in
Taiwan.
@seismicisolation
@seismicisolation
ACKNOWLEDGMENTS
The authors would like to acknowledge the support of the committee
members of the ABRI wind tunnel, throughout the design, fabrication and
integration phases of the wind tunnel, from 2002-2004. Acknowledgement is
also given to the ABRI and the National Science Council for their support in
funding the experiments using the circular cylinders.
REFERENCES
Achenbach, E. (1968); Distribution of local pressure and skin friction around a
circular cylinder in cross-flow up to Re = 5×106. J. Fluid Mech., vol. 34,
pp. 625-639.
Bearman, P. W. (1969); On vortex shedding from a circular cylinder in the
critical Reynolds number regime. J. Fluid Mech., vol. 37, pp. 577-585.
Cermak, J. E., Sandborn, V. A., Plate, E. J., Binder, G. H., Chuang, H.,
Meroney, R. N., and Ito, S. (1966); Simulation of atmospheric motion by
wind tunnel flow. Report to Army under Contract DA-AMC-28-043-G20,
Colo. State University.
Cermak, J. E. (1975); Applications of fluid mechanics to wind engineering – a
freeman scholar lecture. Trans. of the ASME, J. of Fluids Eng., vol. 97,
pp. 9-38.
Cermak, J. E. (1976); Aerodynamics of building. Annual Review of Fluid
Mechanics, vol. 8, pp. 75-106.
ESDU 80025 (1980); Mean forces, pressures and flow field velocities for
circular cylindrical structures: single cylinder with two-dimensional flow.
Issued October 1980.
Gorlin, S. M. and Slezinger, I. I. (1966); Wind tunnels and their
instrumentation. Israel program for scientific translations Ltd.
Higuchi, H., Kim, H. J., and Farell, C. (1989); On flow separation and
reattachment around a circular cylinder at critical Reynolds numbers. J
Fluid Mech., vol. 200, pp. 149-171.
Houghton, E. L., and Carruther, N. B. (1976); Wind effects of building and
structures. John Willy and Sons, New York.
Humphreys, J. S. (1960); On a circular cylinder in a steady wind at transition
Reynolds numbers. J. Fluid Mech., vol. 9, issue 4, pp. 603-612.
@seismicisolation
@seismicisolation
Kao, Y. M. (2005); Calibration of the ABRI environmental wind tunnel and

experimental study of 2-D bluff-body aerodynamic flows. M. S. Thesis,
Department of Aeronautics and Astronautics, National Cheng Kung
University, Tainan, Taiwan, ROC.
Lin, Y. L., Miau, J. J., Tu, J. K., and Tsai, H. W. (2011); Non-stationary,
three-dimensional aspects of flow around a circular cylinder at critical
Reynolds numbers, AIAA Journal, vol. 49, no. 9, pp. 1857-1870.
Miau, J. J., Chou, J. H., Cheng, C. M., Chu, C. R., Woo, K. C., Ren, S. K.,
Chen, E, L., Hu, C. C. and Chen, Z. L. (2003); Design aspects of the
ABRI wind tunnel. The International Wind Engineering Symposium,
IWES 2003, November 17-18, 2003, Tamsui, Taipei County, Taiwan.
Miau, J. J., Chou, J. H., Cheng, C. M., Chu, C. R., Woo, K. C., Ren, S. K.,
Chen, E. L., Hu, C. C. and Chen, J. L. (2004); System Integration and
Performance Validation of the ABRI wind tunnel laboratory. A project
report to ABRI.
Miau, J. J., Tsai, H. W., Lin, Y. J.,Tu, J. K., Fang, C. H., and Chen, M. (2011);
Experiment on smooth circular cylinders in cross-flow in the critical
Reynolds number regime. Experiments in Fluids, Vo. 51, pp.949-967.
Rae, W. H. and Pope, A. (1984); Low speed wind tunnel testing. Wiley and
Sons, 2nd ed.
Roshko, A. (1993); Perspectives on bluff body aerodynamics. Journal of Wind
Engineering and Industrial Aerodynamics, vol. 49, pp. 70-100.
Schewe, G. (1983); On the force fluctuations acting on a circular cylinder in
cross-flow from subcritical up to transcritical Reynolds numbers. J. Fluid
Mech., vol. 133, pp. 265-285.
Schewe, G. (2001); Reynolds number effects in flow around more-or-less bluff
bodies. Journal of Wind Engineering and Industrial Aerodynamics, vol.
89, pp. 1267-1289.
Tsai, H. W. (2006). Experimental investigations of flows around circular
cylinders in the critical regime. M. S. thesis, Department of Aeronautics
and Astronautics, National Cheng Kung University, Tainan, Taiwan.
Tsai, H. W., Miau, J. J., Lin, Y. J., and Tu, J. K. (2010); Oil-film visualization
on flow over a circular cylinder in the critical regime. ISFV14, 14th
International Symposium on Flow Visualization June 21-24, 2010, Daegu,
Korea.
Zdravkovich, M. M. (1997); Flow around circular cylinders. Vol.1:
fundamentals. Oxford University Press.
@seismicisolation
@seismicisolation
@seismicisolation
@seismicisolation
Chapter 2
ANALYSIS OF WIND ENVIRONMENT AND AIR

QUALITY IN DENSELY POPULATED AREAS
USING WIND TUNNEL EXPERIMENTS
Mahmoud Bady*
Assistant Professor, Department of Environmental Engineering,
Egypt-Japan University of Science and Technology (E-JUST),
New Borg El-Arab City, Alexandria, Egypt
ABSTRACT
This chapter presents an experimental investigation of wind flow
characteristics and air quality along a street canyon located within a dense
urban area. Four typical models of a highly populated urban area are
studied and wind tunnel experiments are carried out over an extended
range of the applied wind directions. The building patterns are
represented by 1:100 scale models, where wind velocity, wind pressure
and tracer gas concentrations are measured along the two sides of the
street.
A serious problem associated with wind tunnel tests on the flow
around buildings is that of blockage. To overcome such problem, an
experiment for determining the building blocking effect was carried out
and a correction factor was estimated and considered in the measured
pressures. Details of such experiment are given at the end of this chapter.
*
E-mail: [email protected].
@seismicisolation
@seismicisolation
36 Mahmoud Bady
The study results provide evidence that building configurations and

wind directions are very important factors in determining both wind flow
and pollutant dispersion characteristics within urban domains. Also, the
results demonstrate that gaps between buildings are a very important
factor to be considered by urban planners and designers, because, for a
given building height, larger gaps induce more wind in urban canyons,
thus improving the ventilation process.
1. INTRODUCTION
In recent years, numerous investigations concerning wind flow fields in
urban areas have been carried out. Mfula et al. [1] carried out wind tunnel
experiments to investigate the effect of building array area density on the
magnitude and spatial distribution of concentrations on a test building located
within the array, under normal wind direction and different pollutant source
positions. Macdonald et al. [2] examined the effect of building aspect ratio on
plume dispersion in an obstacle array in scaled wind tunnel experiments and
compared the results with these of field measurements. Chang at al. [3] used
wind tunnel tests and CFD simulations to study the dispersion of pollutants on
a model of an urban street canyon, located within a group of buildings, with
various street canyon aspect ratios. Hoydyth et al. [4] performed wind tunnel
tests to study pollutant concentrations at an urban intersection and the nearby
environment of a regular array of uniform, low-rise rectangular urban blocks
using quantitative tracer gas techniques. However, there are only a few studies
of wind flow characteristics and air quality on high density urban areas. This
study appears to be the first in which detailed measurements of both wind
velocities and pollutant concentrations were made at numerous locations
inside different models of high density built-up areas.
The present chapter was primarily conducted in order to set out proposals
for the design of future high density areas in a way that improved urban
ventilation. This goal was attained through assessment of the effects of
different building patterns on wind environment and air quality, inside an
urban street located within a dense built-up area, over an extended range of
incident wind directions. Four models of typical built-up areas were created,
and wind tunnel experiments have been performed to measure wind velocities
and pollutant concentrations along the two sides of the study domain. The
effectiveness of each pattern in changing wind flow characteristics (and hence
the dispersion of the pollutants) inside the domain under consideration were
assessed over a wide range of applied wind directions. The obtained data was
@seismicisolation
@seismicisolation
Analysis of Wind Environment and Air Quality … 37
analyzed and the effects of the above parameters on wind flow characteristics
and air quality inside the study domain evaluated through detailed discussions.
In addition, the experimental data was used to evaluate the CFD simulation
results.
Figure 1. Simplified diagrams for four typical models of a dense built-up area.
2. GEOMETRY OF BUILDING PATTERNS

Four typical models of a high density built-up area are illustrated in a
simplified diagram for the focused study area in Figure 1, and the geometry of
the central part of model (I) is shown in Figure 2. In addition, photographs of
the wind tunnel test models as used in the experiments are shown in Figure 3
together with the central part of each model. In these models, the study domain
is the pedestrian level through a street surrounded by type-A and type-B
blocks (shown in Figure 2) and a fence of 1.5 m height around the blocks. The
study domain has a length L = 40 m, and a width D = 4 m. The white blocks in
Figure 1 express detached houses, while the grey color represents traffic roads
and void spaces between adjacent buildings.
Models (I), (II), and (III) have the same building arrangements and
dimensions, but they differ in the geometry of the central part which surrounds
the street (marked with black dashed lines). In model (I), the central part
consists of eight type-A and two type-B blocks. Narrow gaps of 1 m width
exist between the type-A buildings, and also between the type-B buildings. In
model (II), the central part forms a solid U-shape. With respect to model (III),
the street buildings are the same as those of model (I), while the outer blocks
which surround them form a solid U-shape. In model (IV), the symmetry is
@seismicisolation
@seismicisolation
38 Mahmoud Bady
interrupted by introducing other blocks of different dimensions in addition to

shifting the upper and the lower two-thirds of the array along the x-axis.
Many factors were born in mind during the design of the four building
patterns. First, these models suit the nature of existing small-lot residential
areas that are so notable in Japan [5]. Second, the four models were nominated
to examine the optimum design for densely inhabited areas which will induce
more wind inside urban domains and will improve their air quality and
comfort. Third, model (II) represents an unfavorable choice for dense urban
areas, due to the blocking effect of its geometry at some wind directions. The
U-shape of the central part of such model decelerates the wind motion and
hence traps pollutants within the pedestrian domain of the street. Fourth,
model (I) represents a spatially uniform-building array. Fifth, model (III)
exhibits a combination between models (I) and (II), as a result, it is expected
that it occupies an in-between behavior between these two models. Sixth,
model (IV) represents a model of what exists in reality, since the buildings in
this model form a staggered array.
(Release points Measurement points)
Figure 2. Street configurations for the case of model (I).
@seismicisolation
@seismicisolation
Figure 3. Test models inside the wind tunnel together with the central part of each
model.
@seismicisolation
@seismicisolation
40 Mahmoud Bady
3. NATURAL VENTILATION
In building industry, natural ventilation is one of the most important
parameters to be considered by planners and architects during the design of
new built–up areas. This importance arises from the fact that natural
ventilation is a very effective tool in enhancing air quality inside indoor
domains by dissolution of pollutants. On the other hand, the wind pressure is
known to be an essential source of power for natural ventilation. By creating
high and low pressures on the different faces of buildings, wind motivates the
air flow inside buildings and these flows are strongly dependent on wind
pressure gradients. At the same time, the arrangements of adjacent buildings
and wind directions are very important parameters which control the wind
flow characteristics (and hence natural ventilation performance of the local
wind) inside urban domains [6]. Accordingly, these two parameters will be
studied here in relation to their influence in enhancing more wind to the study
domain which in turn improves the natural ventilation process.
Generally, ventilation can be undertaken by two methods:
1) Mechanical or forced ventilation;

2) Natural ventilation.
Mechanical or forced ventilation requires the use of a system that employs

fans in some way to drive a movement of air throughout a building. This
research project however, is related to systems that utilize natural ventilation,
where fans and mechanical systems are entirely absent.
3.1. Definition of Natural Ventilation
Natural ventilation can be defined as ‘the movement of air through

openings in a buildings fabric, due to wind or to static pressures created by the
differences in temperature between the interior and exterior of the building
(generally known as the stack effect), or to a combination of these acting
together.
Natural ventilation is subjected to the variability of wind speed, wind
direction, air temperature and opening configuration. Not only do these factors
affect the rate of fresh air supply but also determine whether openings will act
as an inlet or outlet for the air in any space within a building.
@seismicisolation
@seismicisolation
3.2. Natural Ventilation Mechanisms
There are three forms of natural ventilation that must be considered. The
three forms are:
1) Wind induced only;

2) Temperature difference only;
3) Wind and temperature difference.
The following descriptions of the different processes refer to a simple

cubic structure with ventilation openings located on opposing faces as shown
in Figure 4.
Figure 4. Natural ventilation of a building by the three dominant mechanisms [7].
3.3. Wind-Induced Ventilation
Ventilation can be induced solely by the action of the wind on the surface
of the building, around the location of any openings in the building envelope.
@seismicisolation
@seismicisolation
42 Mahmoud Bady
The nature of the pressure distribution and the momentum of the incoming
wind around the structure drive this form of ventilation.
The internal and external aerodynamic pressures at an opening may be
written as a pressure coefficient, Cpi and Cpe respectively, pressure difference
normalized by the dynamic pressure at the reference height which is fixed at
the building roof level:
p p
i ref
Cp  (1)
i 1
 U ref
2
p p
e ref
Cp  (2)
e 1
 U ref
2
where:
Uref is the reference velocity (usually taken at building roof height), (m/s).
pref is a reference pressure (usually atmospheric), (Pa).
ρ is the density of air (kg/m3).
pi and pe are the internal and external pressures respectively, (Pa).
Although external pressure coefficients will vary with position around a

building, internal pressure coefficients are considered to be constant
throughout the volume of the building in many cases as mentioned by Harris
[8].
If conservation of mass is introduced (i.e. total inflow equals total
outflow), it is possible for the ventilation to be found from the pressure data
for a building. The pressure difference across any opening is found from
Equation (2).
e i 2
1
p  p  p   U ref Cp  Cp
e
2

i
 (3)
where Δp is the difference in pressure across the opening.
@seismicisolation
@seismicisolation
4. WIND TUNNEL EXPERIMENTS

The experiments were carried out under neutral atmospheric conditions in
the boundary layer wind tunnel of the Institute of Industrial Science, the
University of Tokyo, Japan. It has a working section of 2.2 m width, 1.8 m
height and an upwind length of 13 m. A schematic diagram for the test section
of the wind tunnel is illustrated in Figure 5. Also, Figure 6 illustrates a
diagrammatic layout of the wind tunnel showing the main parts of the air flow
system. In addition, Figure 7 shows a photo of the wind tunnel from inside.
The photo shows the turn table, the roughness elements, and the 3-D auto
traverse system. Moreover, a summary for the specifications of the boundary
layer wind tunnel, air-flow system devices and their operating characteristics,
and the temperature stratification system is presented in Table 1.
Figure 5. Schematic diagram of the test section of the wind tunnel.
Figure 6. A diagrammatic layout of the boundary layer wind tunnel shows the main
parts of the air flow system.
@seismicisolation
@seismicisolation
44 Mahmoud Bady
Figure 7. A view of the wind tunnel from inside shows the turn table, roughness
elements, and the auto traverse system.
Table 1. Specifications of the boundary layer wind tunnel
Thermally-stratified boundary layer wind

Name
tunnel
Type Closed circuit
Dimensions of the Test
2200 mm (W)×1800 mm (H)
Section
Length of Test Section 16470 mm
Length of Boundary Layer 12850 mm
Wind Speed Range 0.2～20 m/s
Fan Type Manual changeable low noise fan
Flow Rate, Static Pressure 80 m3/s, 80 mm Aq
External compulsion cooling type direct-current
Engine Type electric motor
Capacity, Rotational Speed DC 440V, 160 KW, 0～1,150 rpm (Thyristor
control)
Air Cooling System Type Water cooling radiator panel
Control Temperature Range 12～35 oC
@seismicisolation
@seismicisolation
Thermally-stratified boundary layer wind

Name
tunnel
Stratification System
Multi-layer electrical heater (44 layers×2)
Controlled Temperature
12～52 oC
Range
Floor Temperature Setting Electrical heater and water cooled radiator
Controlled Temperature panel.
Range 12～72 oC
Four in-house units of remote control automatic

positioning.
Main flow direction (X-axis: 0:12000 mm).
Traverse System
Normal to flow direction (Y-axis:± 820 mm).
Vertical (Z-axis 0:1150 mm).
Vertical turning capacity (θ axis: ± 360°).
Turn Table Diameter = 2000 mm, Mobility = ± 360°.
4.1. Experimental Set-Up
The building models were made of wood with a geometric scale of 1:100.
During wind tunnel tests, the model under study was mounted on the turntable
inside the wind tunnel to allow investigation into the effects of different wind
directions. Six directions have been considered and applied to each model.
These directions are: 0°, 22.5°, 45°, 90°, 270°, and 315°. In each case, wind
velocities, and tracer gas concentrations were recorded along the sides of the
street, while wind pressures were measured along the faces of each block of
the street.
4.2. Simulated Boundary Layer
The simulated boundary layer was generated with the aid of

homogeneously distributed roughness elements mounted on the tunnel floor
and by vortex generators installed at the test section inlet [9]. The spires as
well as the size and arrangement of the roughness elements controlled the
properties of the induced boundary layer.
@seismicisolation
@seismicisolation
46 Mahmoud Bady
Figure 8 shows the undisturbed stream wise velocity and the turbulent
intensity profiles of the incident flow measured 30 cm upwind of the leading
edge of the turntable. The best fit for the experimental data was estimated as:
u  z 0.27 (4)
where u is the velocity (m/s) and z is the height (m).
I (%)
0 5 10 15 20 25 30 35
0.7
0.6
0.5
0.4
z (m)
0.3
Velocity
0.2 Turbulent Intensity
0.1
0.0
2 3 4 5 6 7 8 9
u (m/s)
Figure 8. Mean velocity and turbulent intensity profiles of the incident flow.
4.3. Instrumentation and Test Conditions
Measurements of inlet wind velocity and turbulent intensity of the incident

flow have been performed using a 1-D hot wire probe constant temperature
anemometer (CTA, Dantec®) using a sampling frequency of 1.0 kHz with a
sampling time of 65 seconds at each measurement location.
@seismicisolation
@seismicisolation
To normalize the measured velocities, a suitable reference velocity is

needed. Usually in urban area studies, the reference velocity is specified at the
building height level [10]. Therefore, the reference velocity was set to the
velocity at the roof level H above the wind tunnel floor level in the free
stream. Two reference velocities were considered in the experiments. For wind
velocity measurements, the reference velocity was 4.6 m/s, while for
concentration measurements, a velocity of 0.78 m/s was applied.
4.3.1. Wind Velocity

Wind velocities at the desired sampling locations were measured using
thermal-type purified Germanium probes (GeZ-200M, Tohnic®). Figures 9-10
show a photo for the Germanium probes and the distributions of these probes
within the study domain during the experiments.
A total of 22 positions along the sides of the street at 1.5-m height were
considered to perform the measurements, as illustrated in Figure 2. Therefore,
a multi-channel anemo-thermometer system was used to measure wind
velocities at these locations simultaneously. A unique calibration for each
probe (using a standard nozzle calibration system) was carried out. A sampling
time of 90 seconds with a frequency of 10 Hz was considered sufficient to
achieve a stable average for the wind velocity at each measurement location.
Figure 9. The thermal-type purified Germanium probes (GeZ-200M, Tohnic®).
@seismicisolation
@seismicisolation
48 Mahmoud Bady
Figure 10. Distribution of the Germanium probes within the study domain.
4.3.2. Gas Concentrations

Measurements of pollutant concentrations were performed with a high
response Flame Ionization Detector (FID) (THC-2A, TECHNICA®) equipped
with a needle 11 mm long and 0.15 mm in diameter. A mixture of ethylene
(C2H4) and synthetic air was used as a tracer gas (ethylene is essentially
neutrally buoyant in air). The concentration of C2H4 in the tracer gas was
12,100 ppm. The FID device was calibrated before each experiment with fresh
air (0 ppm) and four certified calibration gases of different concentrations of
100, 200, 1,000 and 5,000 ppm. A mass flow controller (Ohkura-Riken®) set
the flow rate of the ethylene-synthetic air mixture released from the release
points (shown in Figure 2) and so controlled the vertical exhaust speed and the
source generation rate. The tracer gas was discharged continuously from four
small ground level dispersing porous plugs of 5 mm diameter at a negligible
momentum (volume flux = 1 cm3/s). A data acquisition system (NR-2000,
KEYENCE along with a PC were used to record the concentration data. A
sampling frequency of 10 Hz for 45 seconds was considered sufficient to
average the concentrations at each measurement location. The measurement
points of concentration are also shown in Figure 2 (height of 1.5 m). The
measured concentrations are presented here in terms of the dimensionless
ratio:
@seismicisolation
@seismicisolation
cH 2U ref
K (5)
Qs
where:
c is the gas concentration (kg/m3).

Uref is the wind velocity at the height of the building model (m/s).
Qs is the tracer gas emission rate (m3/s).
4.3.3. Wind Pressures

Wind pressures on the building surfaces were measured using fine
pressure taps distributed on the building faces inside the street. Pressure
signals were transmitted pneumatically using 1 mm internal diameter plastic
tubes to the pressure transducers which convert the pneumatic pressure into a
voltage signal. Four transducers were used in the same time in order to
monitor pressures simultaneously at four measuring locations. The system of
changing the measured point is shown in Figure 11. Also, Figure 12 shows a
photo for the differential pressure transducers used in the pressure
measurement.
Figure 11. Pressure measured point changing system and the plastic tubes.
In addition to the surface measurements, a reference static pressure was

measured at a location of 2 m upstream of the model center at a height of 1 m
@seismicisolation
@seismicisolation
50 Mahmoud Bady
above the wind tunnel ground level. This reference pressure was measured
using a Pitot tube. It is important to note that the dynamic pressure is not
measured. It is calculated by measuring both static and total pressures and then
subtracting them. Total and static pressures are measured using a single Pitot
tube. The total pressure ptotal is measured at the tube tip, while the static
pressure pstatic is measured at the end of the diverging part of the tube.
Figure 12. Differential pressure transducer.
The time mean pressures were determined by averaging the instantaneous

pressure values sampled with a frequency of 100 Hz over a period of 45
seconds.
To measure wind pressure on the building faces, nine pressure taps were
placed on each building at three measuring heights (S) of 1.5 m, 4.0 m and 6.5
m as shown Figure 13.
factor was estimated and considered in the measured pressures. Details of such
experiment are given at the end of this chapter.
@seismicisolation
@seismicisolation
Figure 13. The street configurations of model (IV) and pressure tap distributions.
5. TRACER GAS EXPERIMENT

Tracer gas experiments were carried out during wind tunnel experiments
in order to attain more understanding for the wind flow characteristics within
and around the study domain.
Figure 14 shows one of the visualization experiments around the street
buildings. Also, the test rig of the tracer gas experiment is shown in Figure 15.
Figure 14. Tracer gas experiment.
@seismicisolation
@seismicisolation
52 Mahmoud Bady
Figure 15. Setup of the tracer gas experiment.
6. EXPERIMENTAL RESULTS
This section is divided into two categories. In the first, the in-domain wind
characteristics are discussed in terms of normalized wind velocity along the
sides of the street; while in the second section, air quality is analyzed and
evaluated in terms of a dimensionless concentration ratio.
6.1. Wind Flow Characteristics
Normalized wind velocities measured along the two sides of the study
domain (left side: X = –1 m and right side: X = 1 m) for the four models are
shown in Figures 16 - 19. The distance from the origin (y) is normalized by the
street length L (i.e. L = 40 m), where (y) and (L) are illustrated in Figure 2.
The figures reflect considerable changes with the building configurations on
wind flow characteristics inside the study domain. In all models, it is clear that
the velocity tends to increase as the distance in the road direction (+y
direction) increases. This behavior is referred to as the blocking effect (i.e.
deceleration in the velocity as the wind hits obstacles such as building walls)
of the traverse blocks (i.e. building B’s) especially in the case of model (II).
As the distance (y) increases towards the wide end of the street (y/L = 1.0), this
effect gradually disappears. Also, it can be noticed that at θ = 90°, the wind
@seismicisolation
@seismicisolation
velocity is nearly constant in all models, since the wind flows smoothly into
the street at this direction.
In the case of model (I), presented in Figure 16, low wind speeds were
established along the inner half of the street (y/L ≤ 0.5) at θ = 0° and 22.5°. In
these cases, wind flows into the street through the gaps and the top boundary
since the momentum transfer from the traffic road is small because the bulk of
the flow doesn’t enter the street. At the mid third of the street (0.3 ≤ y/L ≤ 0.7),
the maximum wind speed is observed to occur at θ = 45°, as the wind enters
the street from both the gaps and the domain’s open end (located on the road
side), which increases the circulation and the wind velocity in this part.
0.5 0.5
0.0 Deg. 0.0 Deg.
22.5 Deg. X = -1 22.5 Deg. X= 1
0.4 45.0 Deg.
0.4 45.0 Deg.
90.0 Deg.
90.0 Deg.
270.0 Deg.
270.0 Deg.
315.0 Deg.
315.0 Deg. 0.3
0.3
u / UH
u / UH
0.2 0.2
0.1 0.1
0.0 0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
y/ L y/ L
Figure 16. History of the wind velocity along the sides of the street of model (I) for
different wind directions.
0.5 0.5
0.0 Deg. 0.0 Deg.
22.5 Deg. X = -1 22.5 Deg. X= 1
45.0 Deg. 45.0 Deg.
0.4 90.0 Deg. 0.4 90.0 Deg.
270.0 Deg. 270.0 Deg.
315.0 Deg. 315.0 Deg.
0.3 0.3
u / UH
u / UH
0.2 0.2
0.1 0.1
0.0 0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
y/ L y/ L
Figure 17. History of the wind velocity along the sides of the street of model (II) for
With respect to model (II), which is presented in Figure 17, low velocity
values are detected in the cases of 0° and 22.5° inflow wind directions along
the closed half of the street (y/L ≤ 0.5). This trend can be attributed to the
@seismicisolation
@seismicisolation
54 Mahmoud Bady
decrease in the area subjected to flow because the street has one opening side
and it has no gaps between its surrounding buildings. In this case, most of the
flow doesn’t enter the domain. On the other hand, the greater the distance from
the closed side, the greater the wind velocity, as noticed in the second half of
the street (y/L > 0.5). As the wind angle increases (i.e. θ = 45° and 90°), the
area subjected to the inflow wind (to the domain) increases and the velocity
increases. Also, it is observed that the maximum velocity occurs at θ = 270°.
This can be attributed to the circulatory vortex that was established inside the
street due to the transfer of momentum across the shear layer at roof height.
0.5 0.0 Deg.

0.5
0.0 Deg.
22.5 Deg. X = -1 22.5 Deg. X= 1
45.0 Deg. 45.0 Deg.
0.4 90.0 Deg. 0.4 90.0 Deg.
270.0 Deg. 270.0 Deg.
315.0 Deg. 315.0 Deg.
0.3 0.3
u / UH
u / UH
0.2 0.2
0.1 0.1
0.0 0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
y/ L y/ L
Figure 18. History of the wind velocity along the sides of the street of model (III) for
0.5 0.5
0.0 Deg. 0.0 Deg.
22.5 Deg. X = -1 22.5 Deg. X= 1
45.0 Deg. 45 Deg.
0.4 90.0 Deg. 0.4 90.0 Deg.
270.0 Deg. 270.0 Deg.
315.0 Deg. 315.0 Deg.
0.3 0.3
u / UH
u / UH
0.2 0.2
0.1 0.1
0.0 0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
y/ L y/ L
Figure 19. History of the wind velocity along the sides of the street of model (IV) for
Figure 18 shows the wind velocity along the two sides of the street of
model (III). In this model, the buildings which form the street are surrounded
by a group of building blocks which form a solid U-shape. As a result, the
wind flow pattern inside the street is expected to be similar to that of model
@seismicisolation
@seismicisolation
(II). The two models show nearly the same trend and flow characteristics for
In the case of model (IV), shown in Figure 19, since there is an
unsymmetrical distribution of building blocks inside the wind environment, a
small variation in the flow characteristics is expected to occur for each change
in the applied wind direction. In the first half of the street (y/L ≤ 0.5), low
velocities are observed in all wind directions (except the case of 270° since it
is similar to that of model (II)). In these cases, wind has no considerable ability
to flow inside that part of the street. Although, there are gaps between street
buildings, the surrounding blocks were shifted to interrupt the direct flow
between the wind and the gaps, which resist the wind motion in its way to the
street domain and hence decrease its velocity. For the street second half (y/L >
0.5), wind flows inside the domain mainly from the opening side which is
located on the traffic road side. This raises the velocity in this part, especially
at θ = 22.5°.
6.2. Gas Concentrations
The effect of wind directions on tracer concentrations within the street

domain for the four models is presented in Figures 20-21. In these figures, it
can be shown that almost all maximum gas concentrations occur at θ = 0° or
22.5°. This can be attributed to the fact that the lowest wind velocities inside
the street were established at these wind directions. Also, it can be observed
that some concentration peaks are found at the left side of the street domain (X
= -1 m) near its opening end (i.e. y/L = 1). It is thought that a clockwise
circulatory flow occurs there. Also, as expected, peak concentrations are
detected around the gas generation points.
Normalized concentrations inside the street for model (I) are shown in
Figure 20. High concentrations were detected along the inner half of the street
(y/L ≤ 0.5) compared with the outer half. This can be thought to occur due to
low wind speeds in the inner half, which decreases the pollutant removal
efficiency. Also, tracer concentrations in the right side of the street (X = 1 m)
are clearly lower than that of the left side (X = -1 m). This behavior refers to
the increased purging capability of the domain’s local wind in the right part
compared with the left part as illustrated in Figure 13.
Figures 21-22 show tracer concentrations in models (II) and (III)
respectively. It is observed that the maximum concentration values in both
@seismicisolation
@seismicisolation
56 Mahmoud Bady
models occurred at θ = 0°, 22.5° (especially for X = –1 m), which indicates

inefficient removal of the pollutant by the street wind at these directions.
250 250
0.0 Deg. 0.0 Deg.
X = -1 22.5 Deg. X= 1 22.5 Deg.
45.0 Deg. 45.0 Deg.
200 90.0 Deg.
200 90.0 Deg.
270.0 Deg. 270.0 Deg.
315.0 Deg. 315.0 Deg.
150 150
K
K
100 100
50 50
0 0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
y/ L y/ L
Figure 20. Normalized concentrations along the two sides of the street in model (I).
250 250
0.0 Deg. 0.0 Deg.
X = -1 22.5 Deg. X= 1 22.5 Deg.
45.0 Deg. 45.0 Deg.
200 200 90.0 Deg.
90.0 Deg.
270.0 Deg. 270.0 Deg.
315.0 Deg. 315.0 Deg.
150 150
K
100 100
50 50
0 0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
y/ L y/ L
Figure 21. Normalized concentrations along the sides of the street in model (II).
Regarding models (II) and (III), it is noticed that high wind velocities and
high concentrations are detected at the measurement locations at the same
time. It can be interpreted that the measured velocities are in a scalar form,
which do not show the directions of velocity components at each measurement
location. Consequently, the high concentrations occur due to circulatory flows
inside the street domain of such models. It is thought that the geometry of the
solid U-shape in model (II) and the outer U-shape in model (III) plays the
major role in generating such circulatory flows.
In the case of model (IV), which is presented in Figure 23, high
concentrations are shown in the first half of the street where low wind speeds
were detected. On the other hand, low concentrations were detected in the
second half as clean air continuously comes into the street from the opening
side and dilutes the domain’s polluted air. In comparison with the other
@seismicisolation
@seismicisolation
models, this model shows low concentration values compared with models (II)
and (III), but higher than model (I).
250 250
0.0 Deg. 0.0 Deg.
X = -1 22.5 Deg. X= 1 22.5 Deg.
45.0 Deg. 45.0 Deg.
200 90.0 Deg. 200 90.0 Deg.
270.0 Deg. 225.0 Deg.
315.0 Deg. 315.0 Deg.
150 150
K
K
100 100
50 50
0 0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
y/ L y/ L
Figure 22. Normalized concentrations along the sides of the street in model (III).
250 250
0.0 Deg. 0.0 Deg.
X = -1 22.5 Deg. X= 1 22.5 Deg.
200 45.0 Deg. 45.0 Deg.
200
90.0 Deg. 90.0 Deg.
270.0 Deg. 270.0 Deg.
315.0 Deg. 315.0 Deg.
150 150
K
100 100
50 50
0 0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
y/ L y/ L
Figure 23. Normalized concentrations along the two sides of the street in model (IV).
According to the measured concentrations along the street sides for the
four different building patterns, model (I) appears to be a good choice for
future dense urban areas, in terms of its wind flow characteristics and air
quality.
Results of such model show the best purging capability of its street wind
(among the four models) reflected by the lowest concentration values along its
sides for different wind directions.
This may be referred to the presence of gaps between adjacent buildings,
which is a very important factor to be considered by planners and designers.
These gaps have the ability to introduce more wind to urban domains, thus
improving the ventilation process.
@seismicisolation
@seismicisolation
58 Mahmoud Bady
6.3. Wind Pressures
Figures 24 - 27 present the values of wind pressure coefficient which was

measured along the surfaces of the street buildings of the four models at
various wind directions.
Three measurement heights along each building surface (i.e. S = 1.5 m,
4.0 m and 6.5 m) were considered as illustrated before in Figure 13.
An important note is that the street of model (II) has no gaps between its
buildings but the locations of pressure taps are the same as those of the other
three building patterns.
Thus, the central part in this model was assumed to be composed of
buildings A1 to A8, B1 and B2 without spaces between them.
From the large amount of wind pressure values shown in the figures,
many observations are arising. First, a general observation is deduced from
these figures is that the values of Cp on all surfaces are negative for S = 1.5
and 4.0 m.
This note reflects that these surfaces are not directly exposed to the flow at
any of the applied wind directions. On the other hand, at the height of 6.5 m,
almost all buildings which are located along the right hand side of the street
(i.e. A5 ~ A8), have positive values for Cp especially in the cases of θ = 0o,
22.5o, and 45o.
This can be attributed to the increase in the area of the street opening end
which located along the traffic road side. Much amount of wind enters the
street domain through this area and causes a rise in the pressure. Another
important note is that almost all of the minimum values of Cp occur at θ = 270o
on the surfaces of buildings A4 and A8. At this wind direction, most of the
flow enters the street domain through the shear layer at the roof height and
thus a very low percentage of this flow touches the surfaces of these buildings.
A large difference is observed between the Cp values on the same surface
when the wind direction changes.
This can be observed clearly at S = 1.5 and 4.0 m, while a small difference
is shown in the case of S = 6.5 m. The amount of wind supplied to the street
domain is affected by the change of the applied wind direction (due to the
change in the area subjected to the wind flow), which in turn affects the values
of the pressure coefficient.
@seismicisolation
@seismicisolation
0.00 0.00 0.30
S = 1.5 m S = 4.0 m 0.20 S = 6.5 m
-0.05 -0.05
0.10
-0.10 -0.10 0.00
Cp
Cp
-0.10
Cp
-0.15 -0.15 -0.20
-0.30
-0.20 -0.20
-0.40
A4, A8 A3, A7 A2, A6 A1, A5 A4, A8 A3, A7 A2, A6 A1, A5 A4, A8 A3, A7 A2, A6 A1, A5
-0.25 -0.25 -0.50
0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00
A4, A8 A3, A7 A2, A6 A1, A4, A8 A3, A7 A2, A6 A1, A4, A8 A3, A7 A2, A6
y/L y/ L y/ L
A5 A4, A8 A3, A7 A5 A4, A8 A3, A7 A1, A5 A4, A8
0.00 A2, A6 A1, A5 0.00 A2, A6 A1, A5 0.10 A3, A7 A2, A6 A1, A5
A4, A8 A3, A7 A2, A6 A1, A5 A4, A8 A3, A7 A2, A6 A1, A5 0.05 A4, A8 A3, A7 A2, A6 A1, A5
-0.05 -0.05
S = 1.5 m S = 4.0 m 0.00
A4, A8 A3, A7 A2, A6 A1, A4, A8 A3, A7 A2, A6 A1, -0.05 A4, A8 A3, A7 A2, A6
-0.10 Cp -0.10
Cp
Cp
-0.10
-0.15 A5 A4, A8 A3, A7 -0.15 A5 A4, A8 A3, A7 -0.15 A1, A5 A4, A8
-0.20
-0.20 -0.20
A2, A6 A1, A5 A2, A6 A1, A5 -0.25 A3, A7 A2, A6 A1,SA5
= 6.5 m
-0.25 -0.25 -0.30
0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00
y/ L y/L y/L
Figure 24. Pressure coefficient values along the building surfaces of the street of model (I).
@seismicisolation
@seismicisolation
0.10 0.10 0.35
0.05 A4, A8 A3, A7 A2, A6 A1, A5 0.05 A4, A8 A3, A7 A2, A6 A1, A5 0.25 S=
0.00 0.00 0.15 6.5 m
-0.05 A4, A8 A3, A7 A2, A6 A1, -0.05 A4, A8 A3, A7 A2, A6 A1, 0.05
-0.10 -0.10
Cp
Cp
Cp
-0.05
-0.15 A5 A4, A8 A3, A7 -0.15
A5 A4, A8 A3, A7 -0.15
-0.20 -0.20
-0.25 -0.25 -0.25
-0.30
A2, A6 A1, A5 S = A2, A6 A1, A5 S=
-0.35
-0.30
1.5 m 4.0 m A4, A8 A3, A7 A2, A6 A1, A5
-0.35 -0.35 -0.45
0.00 0.25 0.50 0.75 1.00 0.0 0.3 0.5 0.8 1.0 0.00 0.25 0.50 0.75 1.00
y/ L y/ L
A4, A8 A3,y / LA7 A2, A6
A1, A5 A4, A8
0.10 0.10 0.35 A3, A7 A2, A6 A1, A5

0.05 A4, A8 A3, A7 A2, A6 A1, A5 0.05 A4, A8 A3, A7 A2, A6 A1, A5 0.25 A4, A8 A3, A7 A2, A6 A1, A5
0.00 0.00 0.15
-0.05 A4, A8 A3, A7 A2, A6 A1, -0.05 A4, A8 A3, A7 A2, A6 A1, 0.05 A4, A8 A3, A7 A2, A6
-0.10 -0.10
Cp
Cp
-0.05
Cp
-0.15 A5 A4, A8 A3, A7 -0.15

A5 A4, A8 A3, A7 -0.15 A1, A5 A4, A8
-0.20 -0.20
-0.25
-0.25 -0.25
-0.30
A2, A6 A1, A5 S=
-0.30
A2, A6 A1, A5 S= -0.35 A3, A7 A2, A6 A1, A5 S =
1.5 m 4.0 m -0.45
6.5 m
-0.35 -0.35
0.00 0.25 0.50 0.75 1.00 0.0 0.3 0.5 0.8 1.0 0.00 0.25 0.50 0.75 1.00
y/L y/ L y/L
Figure 25. Pressure coefficient values on the building surfaces of the street of model (II).
@seismicisolation
@seismicisolation
0.05 0.05 0.25
0.00 0.00 0.15 S=
-0.05 -0.05 S= 0.05
6.5 m
-0.10 -0.10 4.0 m
-0.05
Cp
Cp
Cp
-0.15 -0.15
-0.15
-0.20 -0.20
S= -0.25
-0.25 -0.25
1.5 m -0.35
-0.30 -0.30
-0.35 -0.35 -0.45
0.00 0.25 0.50 0.75 1.00 0.0 0.3 0.5 0.8 1.0 0.00 0.25 0.50 0.75 1.00
A4, A8 A3, A7 A2, A6 A1, A4, A8 A3, A7
y/L
A2, A6 A1, A4, A8 A3,y / LA7 A2, A6
y/ L
A5 A4, A8 A3, A7 A5 A4, A8 A3, A7 A1, A5 A4, A8
0.05
A2, A6 A1, A5 0.05
A2, A6 A1, A5 A3, A7 A2, A6 A1, A5
0.25
0.00 0.00 0.15
S=
-0.05 -0.05 S= 6.5 m
0.05
-0.10 -0.10 4.0 m
-0.05
Cp
Cp
Cp
-0.15 -0.15
-0.15
-0.20 -0.20
S= -0.25
-0.25 -0.25
1.5 m -0.35
-0.30 -0.30
-0.35 -0.35 -0.45
0.00 0.25 0.50 0.75 1.00 0.0 0.3 0.5 0.8 1.0 0.00 0.25 0.50 0.75 1.00
A4, A8 A3, A7
y/L
A2, A6 A1, A4, A8 A3, A7
y / L A2, A6 A1, A4, A8 A3,
y / LA7 A2, A6
A5 A4, A8 A3, A7 A5 A4, A8 A3, A7 A1, A5 A4, A8

Figure 26. Pressure coefficient values on the building surfaces of the street of model (III).
@seismicisolation
@seismicisolation
0.00 0.00 0.25
-0.05
S= S= 0.15 S=
-0.05
1.5 m 4.0 m 0.05 6.5 m
-0.10 -0.10
-0.05
Cp
Cp
Cp
-0.15 -0.15
-0.15
-0.20 -0.20 -0.25
-0.25 -0.25 -0.35
-0.30 -0.30 -0.45
0.00 0.25 0.50 0.75 1.00 0.0 0.3 0.5 0.8 1.0 0.00 0.25 0.50 0.75 1.00
A4, A8 A3, A7 A2, A6 A1, A4, A8 A3, A7 A2, A6 A1, A4, A8 A3, A7 A2, A6 A1,
y/ L y/ L y/ L
A5 A4, A8 A3, A7 A5 A4, A8 A3, A7 A5 A4, A8 A3, A7
A2, A6 A1, A5 A2, A6 A1, A5 A2, A6 A1, A5

0.05 0.05 0.25
0.00
A4, A8 A3, A7 A2, A6 A1, A5 0.00
A4, A8 A3, A7 A2, A6 A1, A5 A4, A8 A3, A7 A2, A6 A1, A5
0.15
-0.05 -0.05 0.05
A4, A8 A3, A7 A2, A6 A1, A4, A8 A3, A7 A2, A6 A1, A4, A8 A3, A7 A2, A6
-0.10 -0.10 -0.05
Cp
Cp
Cp
-0.15 -0.15 -0.15

A5 A4, A8 A3, A7 A5 A4, A8 A3, A7 A1, A5 A4, A8
-0.20 -0.20 -0.25
S= S= S=
-0.25 A2, A6 A1, A5 -0.25 A2, A6 A1, A5 -0.35
A3, A7 A2, A6 A1, A5
6.5 m
1.5 m 4.0 m
-0.30 -0.30 -0.45
0.00 0.25 0.50 0.75 1.00 0.0 0.3 0.5 0.8 1.0 0.00 0.25 0.50 0.75 1.00
y/ L y/ L y/L
Figure 27. Pressure coefficient values on the building surfaces of the street of model (IV).
@seismicisolation
@seismicisolation
From Figures 24-27, it can be concluded that; the determination of the

building model (among the four models) that can satisfy the maximum
natural ventilation on the building surfaces of a dense built-up area is not
easy. A large number of parameters can be considered in order to carry out
the comparisons such as wind direction variation, minimum and maximum
values of Cp… etc., but it is thought that the results may vary regarding each
one of these parameters. Thus, a consistent procedure for carrying out the
desired comparison to determine the optimum model is suggested as follows:
Since the area-averaged pressure is regarded as an accumulation of point
pressures [11], the value of (Cp)avg for a surface at a certain wind direction
is considered a good representation for all point pressures measured on this
surface. The (Cp)avg can be obtained by averaging the point pressure
coefficients measured on each surface.
Figure 28 shows the variation of (Cp)avg on the street buildings surfaces
of the four models for different wind directions. The wind pressure
coefficients presented in the figure are the average of all measurement points
for one building surface (nine points on each building surface facing the
street). In the cases of θ = 0o, 22.5o and 45o, the (Cp)avg increases in the
direction from A1 to A4 and decreases from A5 to A8 for all models. This
can be attributed to the increased pressure in that direction as a result of the
deceleration in the velocity when the wind hits the walls of traverse buildings
(especially in the case of model (II) where no space between such buildings).
The same trend appears clearly in the case of 90o, where the wind flows into
the street domain smoothly through its wide opening end. When the applied
wind direction increases to 270o, the pressure coefficient trend appears to be
a reverse for that of the 90o case. This behavior can be referred to the
circulatory vortex that was established inside the street due to the transfer of
momentum across the shear layer at the roof height. In this case, wind flows
smoothly towards the street exit, which decreases the pressure in the
direction from A4 (or A8) to A1 (or A5). In the same time, it can observe that
the maximum difference between (Cp)avg values are located on buildings A4
and A8 in all models. This can be referred to the location of these two
buildings within the array. When the wind blows at a skew angle of 315o, the
average pressure coefficient decreases in the region from A1 ~ A4 in all
models, holds nearly constant from A5 ~ A7 in the first three models, and
then decreases on the surface of A8 in all models. This behavior is
interpreted as follows: at θ = 315o, the bulk of the wind that enters the study
domain is through the top boundary. A large and non-symmetry circulatory
vortex is then formed inside the street. The surfaces of buildings A1 ~ A4
@seismicisolation
@seismicisolation
64 Mahmoud Bady
and A8 are located in the wake region of that vortex, giving rise to negative
pressures in the direction from A1 ~ A4 and on the surface of A8. On the
other hand buildings A5 ~ A7 are facing the wind that flows from both the
domain top boundary and the narrow spaces between buildings A1 ~ A4. In
the case of model (IV), the staggered distribution of the buildings around the
street prevents the arrival of the flowing wind from the spaces (between A1 ~
A4) to the buildings A5 ~ A8, which decreases the average (Cp)avg on their
surfaces. From the (Cp)avg values estimated at this wind angle, it can be said
that this direction is not approved for all models especially in the case of
model (II) which showed the lowest value of (Cp)avg.
0.00 0.00
-0.05 -0.05
-0.10 -0.10
-0.15 -0.15
Cp
Cp
0.0 Deg.
-0.20 0 Deg. -0.20
22.5 Deg.
22.5 Deg.
-0.25 45.0 Deg.
-0.25 45.0 Deg.
90.0 Deg.
90.0 Deg.
-0.30 Model (I) 270.0 Deg. -0.30 Model (II) 270.0 Deg.
315.0 Deg.
315.0 Deg.
-0.35 -0.35
A A1 A2 A3 A4 A5 A6 A7 A8 A9 A A1 A2 A3 A4 A5 A6 A7 A8 A9
Building (ID) Building (ID)

0.00 0.00
-0.05 -0.05
-0.10 -0.10
-0.15 -0.15
Cp
Cp
-0.20 0.0 Deg. -0.20 0.0 Deg.

22.5 Deg. 22.5 Deg.
-0.25 45.0 Deg. -0.25 45.0 Deg.
90.0 Deg.
90.0 Deg.
-0.30 Model (III) 270.0 Deg. -0.30 Model (IV) 270.0 Deg.
315.0 Deg.
315.0 Deg.
-0.35 -0.35
A A1 A2 A3 A4 A5 A6 A7 A8 A9 A A1 A2 A3 A4 A5 A6 A7 A8 A9
Building (ID) Building (ID)
Figure 28. Area averaged pressure coefficients on the building surfaces of the street
for different building patterns.
In comparison with the other patterns, models (II) and (III) appear to be
the most undesirable choices for high-density built-up areas, regarding
ventilation performance of the street local wind. Figure 27 shows that there
are large fluctuations between the minimum and the maximum values of
(Cp)avg for different wind directions, which mean that these models are not
suitable to be constructed in sites which have variable wind directions along
the year. Also, models (I) and (IV) showed an average behavior for the
@seismicisolation
@seismicisolation
domain local wind in terms of the mean pressure coefficient compared with
the other models. If another comparison is done between these two models,
model (I) appears to be the most suitable model to be chosen for future dense
areas. This selection is based on the observation that the values of (Cp)avg in
model (I) are greater than these of model (IV) at the same wind directions.
6.4. Comparison between Numerical and Experimental Results
Some examples for comparing CFD simulation results with the

experimental data for wind velocity at three levels of z = 1.5 m, 3.0 m and 9.0
m are presented in Figures 29–31. Results of the numerical approach are
given only for the first three models, while those of model (IV) were rejected
due to the instability caused by the unstructured mesh system of such model.
Figure 29 (a) shows the wind velocity history in the region y = 0 ～ 40 m
along the two sides of the street of model (I) at 1 m apart from the street
center line. In the subplot (a), the calculated and the measured scalar
velocities along the left hand side of the street (X = - 1 m) at the three
considered heights is shown. In numerical and experimental approaches, a
general observation can be noticed that the velocity tends to increase from
the inner end of the street (y = 0 m) to its open end (y = 40 m), which is
located alongside the traffic road. Also, the figure shows satisfactory
agreements between the numerical and experimental results along the inner
half of the street (0 ≤ y ≤ 20) at z = 1.5 m and z = 3.0 m. However, a
considerable difference is observed along the second half of the street (20 < y
≤ 40). These differences could be attributed to the fact that the measurement
points near the traffic road (y = 40 m) are in locations of highly complex
circulatory flow regions. In the same time, there are peak values for the wind
velocity along this side of the street for these two wind directions near the
narrow spaces between the street buildings. This behavior reflects the
importance of the presence of such spaces between adjacent buildings to
induce more wind through urban domains. Figure 29 (b) shows the wind
velocities along the right hand side of the street (X = 1 m). Similar to the case
of X = -1 m, good agreements are observed along the inner half of the street,
while near the road side the differences between the two approaches are
larger than those of the case of (X = -1 m). In Figures 29 (a) and (b), at z = 9
m, a large difference between the calculated and the measured velocities are
shown. The velocities calculated by CFD simulation underestimate the
@seismicisolation
@seismicisolation
66 Mahmoud Bady
measured velocities. This can be attributed to the shortage of the turbulence

model in predicting the velocity gradient near building walls and roofs.
In Figures 30 and 31, the similar comparisons of numerical predictions
with the experimental results for models (II) and (III) in the two wind
directions are shown, respectively. Similar to the results of model (I), the
numerical predictions of the two models, overall, are in satisfactory
agreement with the experimental results, but the major discrepancies are
present at z = 9.0 m. In addition to the shortages of the k-ε model in
predicting the velocity gradient near walls and roofs, the computational errors
include the inaccuracies introduced by the discretization scheme, parameter
selection, and algebraic equation solutions. In general, however, the overall
agreement between numerical predictions and experimental measurements is
good.
0.8 0.8
CFD (z = 1.5) CFD (z = 3.0) CFD (z = 9.0) CFD (z = 1.5) CFD (z = 3.0) CFD (z = 9.0)
0.7 0.7 EXP (z = 1.5) EXP (z = 3.0) EXP (z = 9.0)
EXP (z = 1.5) EXP (z = 3.0) EXP (z = 9.0)
Scalar velocity (m/s)
0.6 0.6
0.5 X = -1 0.5
X=1
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0.0 0.0
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
y (m) y (m)
(A) (B)
Figure 29. Comparison between numerical and experimental wind velocities along
the sides of the street of model (I) at θ = 0º; (a) X = -1 m, and (b) X = 1 m.
0.8 0.8
0.7 0.7
EXP (z = 1.5) EXP (z = 3.0) EXP (z = 9.0) EXP (z = 1.5) EXP (z = 3.0) EXP (z = 9.0)
0.6 0.6
X = -1 X=1
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0.0 0.0
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
y (m) y (m)
(A) (B)
Figure 30. Comparison between numerical and experimental wind velocities along
the sides of the street of model (II) at θ = 0º; (a) X = -1 m, and (b) X = 1 m.
@seismicisolation
@seismicisolation
0.8 0.8
Scalar velocity (m/s) 0.7 EXP (z = 1.5) EXP (z = 3.0) EXP (z = 9.0) 0.7 EXP (z = 1.5) EXP (z = 3.0) EXP (z = 9.0)

0.6 0.6
X = -1 X=1
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0.0 0.0
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
y (m) y (m)
(A) (B)
Figure 31 Comparison between numerical and experimental wind velocities along the
sides of the street of model (III) at θ = 0º; (a) X = -1 m, and (b) X = 1 m.
7. ESTIMATION OF BLOCKING EFFECT IN WIND

TUNNEL EXPERIMENTS
factor was estimated and considered in the measured pressures.
7.1. Details of the Experiment
The experiment arrangement is shown in Figure 32.

There are three points in the figure. These points are:
 Point A: is the reference pressure measuring point (static and

dynamic).
 Point B: is the center of the turn table with the same height of A.
 Point C: is the point located on the building surface at the roof
height.
Pw is the measured wall pressure without the presence of the buildings

(hypothesis).
PW is the measured wall pressure at the building surface with the
presence of the buildings.
@seismicisolation
@seismicisolation
68 Mahmoud Bady
Figure 32. Schematic of the experimental arrangements for estimating the blocking
correction factor.
It is important to note that; the dynamic pressure is not measured. It is

calculated by measuring static and total pressures and subtracting them. Total
and static pressures are measured using a single Pitot tube. At the tube tip,
Ptotal is measured, while Pstatic is measured at the end of the diverging part of
the tube.
7.2. Methodology
It is required to estimate a correction factor for the measured pressures

due to the blocking effect of the building models. In the first stage before
adding the buildings, all quantities shown in the figure are measured.
After adding the buildings in the second stage, we can measure only PW ,
PA and VA .
Using the following guide equations:
@seismicisolation
@seismicisolation
PA  PA  PA 
V A  V A  V A 

VC / V A   
 (6)
PA  PB  P pressure loss due to blocking 
PB  PC   

PA  PB  P 
The wind pressure coefficient can be expressed as:
PW  PC  P
Cp  (7)
1
VC2
2
where P  is the pressure caused by shrinking of the stream lines due the
blocking effect.
The above set of equations can be rearranges to give:
PC  PB   

PC  PA  P    (8)
PA  PA  PA 
Then:
PC  PA  PA  P   
VC   .V A 

 (9)
V A  V A  V A 
VC   .V A  V A 
 
PW  PA  P  PA  P  

Cp  (10)
1
 2 (VA  VA ) 2
2
It can be assumed that:
@seismicisolation
@seismicisolation
70 Mahmoud Bady
P  PA  PB  0 

PA  PA  PA  0  (11)
V A  V A  V A  0
Then, the final equation becomes:
PW  PA  P  

Cp  (12)
1
 2 (VA ) 2
2
Comparing the two equations for Cp, the correction factor can be
estimated as:
C1  P   (13)
7.3. Results of the Experiment
The following are the results of the wind tunnel experiment for
measuring the blocking correction factor.
Measured quantities:
PA =-2.8855
PA =-2.8000
PB =-2.6423
PB =-2.8860
PC =-2.8793
Calculated values:
 = 0.476
 = 0.237
Correction factor is: C1= 0.1653
@seismicisolation
@seismicisolation
CONCLUSION
An investigation of wind flow characteristics and air quality over four
typical building patterns of a high density built-up area was presented, in
order to set out proposals for the design of future dense urban areas. Wind
tunnel experiments have been performed to measure wind velocities and
pollutant concentrations along the sides of a street located within this area.
The mean variables of the experiments were building array geometry and
wind direction. A comparison between CFD simulation results and the
experimental data for wind velocities and gas concentrations along the street
sides has been carried out. Some examples of such comparison were
presented in this chapter where satisfied agreements were obtained.
The study results provide evidence that building configurations and wind
direction are very important factors in determining wind flow patterns, which
controls pollutant diffusion characteristics within urban areas. Among the
building models studied, model (I) appears to be a good choice for dense
areas. Results of such model showed that the presence of gaps between
adjacent buildings is a very important factor to be considered by planners and
designers. The gaps have the ability to introduce more wind to urban
domains, which improves air quality and natural ventilation as well.
REFERENCES
[1] Mfula A., Kukadia V., Griffiths R., and Hall D., Wind Tunnel
Modeling of Urban Building Exposure to Outdoor Pollution,
Atmospheric Environment; Vol. 39 (15), pp. 2737-2745 (2005).
[2] Macdonald R., Griffiths R., and Hall D., A Comparison of Results
from Scaled Field and Wind Tunnel Modeling of Dispersion in Arrays
of Obstacles, Atmospheric Environment; Vol. 32 (22), pp. 3845-3862
(1998).
[3] Chang C., and Merony R., Concentration and Flow Distributions in
Urban Street Canyons: Wind Tunnel and Computational Data, Journal
of Wind Engineering and Industrial Aerodynamics; Vol. 91, pp. 1141-
1154 (2003).
[4] Hoydyth W., and Dabberdt W., A Fluid Modeling Study of
Concentration Distributions at Urban Intersections, Science of the Total
Environment; Vol. 146/147, pp. 425-432 (1994).
@seismicisolation
@seismicisolation
72 Mahmoud Bady
[5] Katsumata W., Strategies for Improving the Residential Environment

of Existing Suburban Small-lot Residential Areas through
Harmoniously Controlling the Rebuilding of Houses, PhD Thesis; The
University of Tokyo, Japan (2004), (in Japanese).
[6] Zhang A., Gao C., and Zhang L., Numerical Simulation of the Wind
Flow around Different Building Arrangements, Journal of Wind
Engineering and Industrial Aerodynamics, Vol. 93, pp. 891-904
(2005).
[7] Easom G., Computation and Measurement of Wind Induced
Ventilation, Ph D Thesis; University of Nottingham (2000).
[8] Harris, R.I., The Propagation of Internal Pressures in Buildings,
Journal of Wind Engineering and Industrial Aerodynamics; Vol. 34,
pp.169-184 (1990).
[9] American Society of Civil Engineering, Wind Tunnel Studies of
Buildings and Structures, ASCE Manuals and Reports on Engineering
Practice No. 67 (1990).
[10] Kim. J., and Baik J., A numerical Study of the Effects of Ambient
Wind Direction on Flow and Dispersion in Urban Street Canyon Using
the RNG k-ε Turbulent Model, Atmospheric Environment, Vol. 38, pp.
3039-3048 (2004).
[11] Uematsu Y., and Isyumov N., Wind Pressures Acting on Low-Rise
Buildings: Review, Journal of Wind Engineering and Industrial
Aerodynamics, Vol. 82, pp. 1-25 (1999).
@seismicisolation
@seismicisolation
Chapter 3
WIND TUNNEL INVESTIGATION

INTO THE DRAG CHARACTERISTICS
OF CATAMARAN FORM
I. K. A. P. Utama1 and A. Jamaluddin2

1
Department of Naval Architecture
and Shipbuilding Engineering,
ITS, Surabaya, Indonesia
2
Indonesian Hydrodynamics Laboratory,
Surabaya, Indonesia
ABSTRACT
The breakdown into the resistance or drag components of catamarans
has been widely discussed worldwide in the last 30 years. The resistance
interference (both wave and viscous parts) has been the major part among
the components. Wave resistance interference can be rather easily
estimated using tank test, whilst the viscous component is rather
complicated to determine. Tank test can be used to estimate the skin
friction, but correction should be made attributed to interference of wave
resistance on skin friction or viscous resistance. In order to isolate the
viscous resistance, hence free from wave component, wind tunnel test
was carried out. A series of tests of catamaran forms were carried out
using low speed wind tunnel. Various configurations of slender
catamaran were made in order to identify the viscous resistance (hence
the form factor) and viscous resistance interference.
@seismicisolation
@seismicisolation
74 I. K. A. P. Utama and A. Jamaluddin
The Chapter discusses the construction and experimental procedure

of the wind tunnel test. Verifications were made with tank test and CFD
(computational fluid dynamics) data in order to examine the accuracy of
using wind tunnel. Overall results demonstrate the effectiveness of using
wind tunnel test to estimate viscous resistance and its interference
component.
Keywords: Catamaran, viscous resistance, wind tunnel
NOMENCLATURE
Demihull One of the hulls which make up the catamaran
L Demihull length between perpendiculars
CF Coefficient of frictional resistance [ITTC-1957 correlation
line]
CP Pressure coefficient
CSA Cross sectional area [m2]
CV Coefficient of viscous resistance [RV/(0.5ρ(WSA)V2)]
CVP Viscous pressure resistance or normal force
ds Distance between CPs
Fr Froude number [V/√gL]
Re Reynolds number [VL/]
RF Shear force [N]
RV Viscous resistance [N]
RVP Viscous pressure or normal force [N]
S Separation between catamaran demihull centrelines [m]
V Velocity [m/s]
WSA Wetted surface area [m2]
(1 + k) Form factor, demihull
(1 +βk) Form factor, catamaran
β Viscous resistance interference factor
ø Pressure field change around the demihull
ρ Fluid density [kg/m3]
 Kinematic viscosity of fluid [m2s-1]
σ Velocity augmentation between the hulls
g Gravity acceleration [m/s2]
@seismicisolation
@seismicisolation
Wind Tunnel Investigation into the Drag Characteristics … 75
1. INTRODUCTION
Resistance characteristics are principal aspects of the catamaran design
spiral as they are strongly coupled with speed and fuel economy and,
consequently, the operating and cost efficiencies of the vessel.
In the resistance phenomena on catamaran or multihull, it restricts on
attention to the address of efficient and reliable methods for the calculation of
the wave resistance of fast displacement catamarans. It is assumed that the
frictional and the viscous-pressure part of resistance of slender hulls, as the
demihulls of catamarans are, can be successfully approached by common
semi-empirical methods (ITTC line with a form factor according to systematic
experiments and semi-empirical formulas) [ITTC, 2002]. Of course, the
problem of catamaran stern flow separation at higher speeds needs to be
considered separately, in connection with the arrangement of an optimal
propulsion system. In any case it can be assumed, due to the slenderness of the
demihulls, that the uniformity of the propeller onset flow of fast displacement
catamarans, and especially of those having SWATH-like stern sections, will
eventually contribute to a relatively high propulsive efficiency.
It is well established that the determination of the wave resistance of
symmetric slender or thin twin hull vessels can be easily achieved by
application of the classical theories of Michell-Havelock (monohulls)
[Michell, 1898] and Strettenskii-Eggers (monohull in a canal and twin-hulls)
[Eggers, 1955]. They all lead to relatively simple formulae for the wave
resistance, in terms of simple type centerline or center plane Kelvin source
distributions, the strength of which is derived immediately from the hull form
characteristics of the studied vessel. These methods consider the effects of hull
interaction on the catamaran wave resistance in an approximate way, namely
by superposition of the individual demihull's far-field wave pattern and
thereafter by employment of a modified KOCHIN function to calculate the
catamaran's wave resistance [Papanikolaou, 1997].
Eventually, there is a direct relation between the wave resistance and the
ship's hull form, defined by the hull offsets. Extending this concept also to the
frictional part of resistance, being directly proportional to the local Reynolds
number and the hull surface area, it is possible to deduce a direct functional
relationship between the sum total of the wave and frictional resistance and the
ship's hull form offsets, leading to the formulation of a systematic optimization
procedure by Lagrange's multiplier method [Papanikolaou, et al., 1989]. This
method, which proves to be very efficient and fast, was applied successfully in
the past to the design of several SWATHs and thin, but symmetric, catamarans
@seismicisolation
@seismicisolation
and the theoretical predictions have been validated, in most cases, successfully
by model experiments.
2. CATAMARAN RESISTANCE COMPONENTS

2.1. Resistance Components
The Report to the 14th I.T.T.C [1975] the Resistance Committee defined
the components of resistance. The viscous resistance is defined as the
component associated with the expenditure of energy in generating vorticity,
vortices (eddies) and turbulence.
The wave resistance is defined as the component associated with the
expenditure of energy in generating gravity waves. The resistance of a surface
vessel may be broken down into components attributed to different physical
process, which scale according to different scaling laws as written in Equation
1.
Figure 1. Breakdown of Resistance Components [Couser et al., 1997].
@seismicisolation
@seismicisolation
The resistance components can be broken down as presented in Figure 1,

which is proposed by [Couser et al., 1997].
RT ( Re ,Fr )  RV ( Re )  RW ( Fr )  1  k ( Fr ) RF ( Re )  RW ( Fr ) (1)
The total resistance coefficient can be determined with the following

equation:
RT
CT  (2)
0.5  (WSA) V 2
where WSA is the wetted surface area of both hulls in the case of catamaran.
The Froude method was based on the assumption that the total resistance
can be split into two components, one the frictional resistance and the other the
residual resistance, which is essentially the wave resistance and the resistance
due to eddies and vortices. The frictional resistance was assumed to be
sensibly equal to the resistance of a rectangular plate of the same length and
wetted surface as the hull of the ship or the model. The residual resistance, the
difference between the total resistance and frictional resistance, is then scaled
according to the Froude law.
Reference [Hughes, 1954] and [Hughes, 1966] introduced the resistance
into three components which is an important improvement over Froude's
method: skin friction (CF), form effect on skin friction (CF0), and wave making
resistances (CW). These resistance components can be written with the
following equation as:
C T  C F  C F O  CW
(3)
where C FO  kC F , hence:
CT  1  k C F  CW
(4)
It is now well known that the frictional resistance coefficient CF derived

from a flat plate is not the same as that of the hull and, furthermore, that CF is
only a part of the viscous-resistance coefficient CV. In order to improve the
Froude method, it has been suggested that the ratio is independent of the
Reynolds number Re and Froude number Fr, where k is the form factor.
@seismicisolation
@seismicisolation
1  k  CV / C F (5)
where CF is a skin friction coefficient is estimated using the ITTC 1957 ship
correlation line [ITTC, 1957].
CF = 0.075/ (log10Re-2)2 (6)
where Re = V. L/ ν is the Reynolds number.

The viscous resistance component, composed of the flat plate frictional
resistance and form drag including the earlier mentioned viscous (body)
interference, is still very difficult to estimate accurately. Whilst flat plate
resistance in isolation can be estimated with a reasonable precision using
ITTC-57 extrapolator line, the form drag (viscous pressure), caused by the
boundary layer activity and the body interference effect, and requires a series
of model test data. Reference [ITTC, 2002] recommended to determine the
form factor “k” by utilizing experimental data at very low speed or low Froude
numbers (Fr< 0.2), where Cw must become negligible. Prohaska's method
[ITTC, 2002], [Bertram, 2000] allows the form factor to be calculated by
assuming that, at low speed, the total resistance is described as:
CT  1  k  C F  a Fr n (7)
At low speed, Fr <0.2, is assumed to be a function of Fr4, the straight line

plot of CT/CF versus Fr4/ CF will intersect the ordinate (Fr=0) at (1+k),
enabling the form factor to be determined.
2.2. Catamaran Hull Resistance Interaction
The resistance interference components of a catamaran present much more

complicated phenomena than those of monohulls due to the interference
effects between the hulls. The interaction effects can be divided into two main
factors:
2.2.1. Body Interference

The flow around a symmetric demihull is asymmetric due to the influence
of other demihull. i.e. the pressure field is not symmetric relative to the
centreline of the demihulls [Insel, 1990].
@seismicisolation
@seismicisolation
This has the following outcomes:
 The perturbation velocity around the demihull increase, especially on

the inside, tunnel side, of the hull due to the venture effect. This
velocity augmentation causes an increase in skin friction resistance
and modifies the form factor. Experiments of Miyazawa [1979]
indicate an increase in perturbation velocity of up to 10% in the x-
direction compared with that of the demihull in isolation.
 A cross flow may occur under the keel which can lead into an induced
drag component which is normally neglected in monohulls. Although
this component is reported to be important in symmetric catamarans
by Pien [1976], Miyazawa [1979] has suggested that this component
is relatively small compared with in his experimental results. In these
experiments the cross flow velocity in the y-direction is about 5-7% of
the model speed. Cross flow in the entrance is outwards, while in the
run it is inwards.
 Because the wave heights at the stern inside and outside of the
demihull are different, the flow at the stern can show inwards or
outwards flow. This causes vortices and spray at the stern resulting in
an induced drag component.
 The velocity increase on the tunnel side may change the structure of
the boundary layer.
 As the waves of one demihull reach the other hull, the wetted surface,
and therefore the skin friction resistance, can change.
2.2.2. Wave Interference

As a result of two hulls running side by side, interference effects on wave
resistance may also be observed
 Due to the change in the pressure field, wave making of the demihulls
may change in other words the wave formation of a demihull may be
different than from the assumed case of the demihull in isolation.
 A favourable and unfavourable interaction between the waves of the
demihulls may occur. The transverse wave of a demihull is always
reinforced by the other hull while divergent bow wave of the one hull
can be cancelled by the divergent stern wave of the other hull or by
the reflection of the same bow wave from the other hull.
@seismicisolation
@seismicisolation
 The reflections of the divergent waves from the other demihull

complicate the interference phenomena.
 The bow wave of a demihull in the tunnel meets the bow waves of the
other demihull on the centreline, and superposition of these two waves
can become very high resulting in an unstable wave, even in breaking
waves and spray at high speeds.
 Inward or outward flows at the stern changes the wave formation at
the stern.
In the case of catamaran hull, as in [Insel and Molland, 1992] summarized

a calm-water-resistance investigation into high speed semi-displacement
catamarans, with symmetrical hull forms based on experimental work carried
out at the University of Southampton. Two interference effects contributing to
the total resistance effect were established; these are viscous and wave
interferences. The total resistance of a catamaran could be expressed by the
equation:
CT CAT  1   k   C F   CW (8)
The factor ø has been introduced to take account of the pressure-field

change around the demihulls and σ takes account of the velocity augmentation
between the hulls and would be calculated from an integration of the local
frictional resistance over the wetted surface, while (1+k) is the form factor for
the demihull in isolation.
For practical purposes, ø and σ can be combined into a viscous
interference factor β where (1+øk) σ = (1+βk). Hence
CT CAT  1   k  C F   CW (9)
 can be calculated from the equation:
CWCAT C  1   k C F CAT
   T (10)
CWDEMI CT  1  k C F DEMI
where β and τ = 1 for monohull [Sahoo, 2006 and 2007].
@seismicisolation
@seismicisolation
2.3. Viscous Resistance Interference
The flow about a catamaran is symmetric to the centerline; the flow over
each of the demihulls however is not. As far as the potential wave-making
nature of the catamaran flow, and hence the associated resistance, is concerned
this phenomena is well explored in the open literature including the well-
known wave interference phenomenon. Depending upon the slenderness of a
demihull, even the modest thin-ship theory based procedures can tackle this
resistance component within a reasonable accuracy. While the interference
viscous resistance depends on wetted hull surface area and the pressure-field
change around the demihulls as well as the velocity augmentation between the
hulls.
Basically, the viscous resistance coefficient for catamaran can be written
from the equation (9) if the wave resistance is omitted:
CV CAT  1   k  CF (11)
The viscous resistance is made up of skin friction and viscous pressure (or
form) drags. In the marine field this is treated as a flat plate friction component
together with an appropriate form factor (1+k).
Of major interest, in the field of high-speed catamarans, is the appropriate
magnitude of the form factor and its dependence on hull and operational
parameters such as: length to displacement ratio; breadth to draught ratio;
separation to length ratio and Froude number. Historically, the choice of form
factor for slender catamaran forms has often been close to unity. However,
recent research, from a number of independent researchers working in this
field, has indicated that this may not be the case and that form factors greater
than unity may be appropriate for the vessels.
Reference [Molland, et al, 1996] formula uses the demihull slenderness
ratio, L/V1/3, to determine the form factor according to the following equation:
1   k   3.03 ( L / V 1 / 3 ) 0.40 (12)
Note that this expression is for the form factor, (1+βk), of the complete
catamaran including viscous interaction effects between the demihulls.
Equation (12) is known as the Molland algorithm for catamarans, which
based on a regression analysis of random model and full scale test data of a
wide variety of hull forms, is widely used in commercial naval engineering
@seismicisolation
@seismicisolation
software. This software automatically calculates the form factor that is used
during the slender body analysis.
Reference [Insel and Molland, 1992] concluded that the form factor, for
practical purposes, is independent of speed and should thus be kept constant
over the speed range. This was a good practical solution to a complex
engineering problem at that point in time.
The viscous interference between the demihulls of catamarans also seems
to be an as yet incompletely understood effect, which can complicate the
estimation and application of simple form factors. In [Insel and Molland,
1992] state that "catamarans show substantially higher resistance than twice
that of the demihulls, even at low speeds where wave interactions are
negligible, therefore indicating viscous interactions. Additionally, flow
visualization experiments on a catamaran model indicated a change of flow
lines and pressure field; hence some form of viscous interaction”.
In order to investigate the viscous interaction due to the change of flow
speed and pressure field around the catamaran hull, it is need to conduct the
experiment in wind tunnel.
3. EXPERIMENT IN WIND TUNNEL

For fine ships at higher speeds such as commercial catamarans, the total
resistance is dominated by viscous resistance which can constitute between 50
and 80 per cent of the total [Couser et al., 1997]. The viscous resistance is
made up of skin friction and viscous pressure (or form) drags.
The use of wind tunnels for such marine-based investigations is described,
for example, in references [Lackenby, 1965] and [Joubert et al., 1978]. This
approach, in which the free surface is treated as a solid plane, allows the
isolation of the viscous resistance but does not take account of any influences
that surface waves may have on viscous resistance. Lackenby [1965] carried
out a number of tests on a parabolic model in both a wind tunnel and a test
tank. The overall results of this work would indicate that the influences of
surface waves on the viscous resistance are not large and that it should be
acceptable to treat the viscous resistance in isolation in this manner in order to
investigate viscous components and form effects.
The reflex model was a technique pioneered by others [Joubert et al.,
1970]) where the resistance of the hull is measured in a wind tunnel. An
implicit assumption, for the reflex model, is that the waterline is level. The
investigation has been extended to BSRA trawler series [Joubert et al., 1979].
@seismicisolation
@seismicisolation
Utama [1999] conducted an experimental investigation on a single ellipsoid

(as a reflex model) and a pair of ellipsoids in close proximity in a low-speed
wind tunnel. Ellipsoids were chosen in order to provide the facility to compare
the results for a single body with earlier experimental work on bodies of
revolution, such as in references [Lackenby, 1965] and [Joubert et al., 1978],
and the possibility of making analytical estimates of the inviscid resistance for
comparison with the numerical investigations.
3.1. Wind Tunnel Test on a Pair of Ellipsoid
A pair of Ellipsoid was tested in the 2.2 m×1.5 m low-speed closed return
wind tunnel at the University of Southampton [Davies, 1961]. The upper
ellipsoid was fitted to the overhead wind tunnel dynamometer to measure total
drag and sideforce. The lower ellipsoid (fitted with the pressure tappings)
could be adjusted vertically to alter the separation between the two bodies
(Figure 3). The models were tested in monohull form and at separation– length
ratios (S/L) of 0.27, 0.37, 0.47 and 0.57. Wind speed was set using the wind
tunnel controller and measured using a Betz manometer. Wind speeds of up to
40 m/s were used giving Reynolds numbers (based on model length) of up to
3.2×106. The experimental investigation was carried out in a wind tunnel on a
single ellipsoid and a pair of ellipsoids in close proximity (Figures 2 and 3).
The tests represented a reflex model of a multihull ship (Figure 4) and
investigated the viscous drag and the viscous interaction between the hulls.
Then, Armstrong [2003] investigated the effect of demihull separation on the
frictional resistance of catamarans.
Figure 2. Models in the wind tunnel [Utama, 1999; Molland and Utama, 2002].
@seismicisolation
@seismicisolation
Figure 3. Set-up of models in the wind tunnel [Utama, 1999; Molland and Utama,
2002].
Figure 4. Cross-section of ellipsoids in proximity, showing the equivalent waterline

[Molland and Utama, 2002].
The drag coefficient for the single ellipsoid, at a Reynolds number of

3.2×106 (Table 1), was found to be 0.004512. This compares with an estimate
of 0.0044 using the methods of reference [Young, 1939]. It is higher than the
measured value of 0.0041 for the airship [Lion, 1932], which is only
approximately similar to the shape of an ellipsoid and where some laminar
flow may have been present over the fore-end. The comparisons confirmed
that the measured drag is of the correct order of magnitude.
From Table 1 and Figure 5, at 40 m/s, it is seen that the ratio of total
viscous resistance for the twin bodies over that for the single body is about
1.10 and decreases relatively slowly to about 1.08 as the separation–length
ratio S/L is increased up to 0.47.
@seismicisolation
@seismicisolation
Figure 5. Viscous interaction (CVtwin /CVsingle ) [Molland and Utama, 2002].
Table 1. Results of wind tunnel tests on ellipsoids: wind speed= 40 m/s,

L/D= 6, equivalent L/1/3= 5.17
CF CV/CF
CV CVpress CSF
CSFpress ITTC (1+k) (1+βk) CVtwin/
S/L x x x CSF/CV
x 1000 x CVsingle
1000 1000 1000
1000
Single 4.512 - - - 3.692 1.22 - - -
0.27 4.941 1.198 4.04 3.73 3.712 - 1.33 1.095 0.818
0.37 4.824 1.202 2.09 2.02 3.716 - 1.30 1.069 0.433
0.47 4.887 1.226 1.37 1.46 3.718 - 1.31 1.083 0.280
0.57 4.985 1.190 0.70 0.89 3.678 - 1.36 1.105 0.140
At the largest separation of S/L=0.57 some interference from the lower

tunnel wall (floor) may have occurred and it is considered that the results for
S/L=0.57 should be treated with caution. The results for 20 m/s are also shown
in Figure 5 and these show similar trends to those for 40 m/s.
The pressure distributions over the ellipsoid were found to be broadly
similar to published data for single bodies, such as for bodies of revolution or
airships [Lion, 1932]. Integration of the pressures yielded the pressure drag
CVpress (see Table 1). When subtracted from the total viscous drag values, CV,
these gave values of skin friction coefficient CF, which were within 1–3 per
cent of published data such as the ITTC line [equation (6)], also shown
@seismicisolation
@seismicisolation
tabulated in Table 1. This is also in keeping with the results of Preston tube
skin friction measurements on bodies of revolution [Joubert et al., 1978] where
the levels of measured skin friction were a little lower than, but within about 3
per cent of, the ITTC line. The small changes in CF in Table 1 (for the same
nominal air speed) are due to changes in temperature during the different test
runs.
Integration of the pressures yielded sideforce values CSFpress (Table 1)
within 3–8 per cent of the directly measured sideforce, CSF, which are
considered reasonable for checking purposes.
Flow visualization studies (Figure 9) indicated the appearance of cross-
flow over the hulls when in twin- body configuration, together with the
development of a measured sideforce. The sideforce ratio, shown in Figure 6,
is seen to increase rapidly with a decrease in S/L. This change is further
supported by the development of asymmetric pressure distributions (Figure 7)
and the pressure changes shown in Figure 8, where there is a decrease in
pressure, and hence an increase in velocity, between the hulls as the S/L is
reduced. The generation of a sideforce would suggest the presence of induced
drag.
Figure 6. Sideforce as a proportion of total viscous drag (CSF /CV ) [Molland and
Utama, 2002].
@seismicisolation
@seismicisolation
Figure 7. Distribution of pressures at two longitudinal stations [Molland and Utama,

2002].
Figure 8. Distribution of pressure along a line at position 0, equivalent to a waterline

between the bodies [Molland and Utama, 2002].
@seismicisolation
@seismicisolation
Figure 9. Flow visualization study, demonstrating cross-flow [Molland and Utama,

2002].
However, while there is a rapid drop in sideforce with an increase in S/L

(Figure 6), the total drag decreases relatively slowly (Figure 5), which
suggests that the level of induced drag is not significant. Similar conclusions
were drawn in reference [Couser et al., 1998].
The tests with single and twin ellipsoids demonstrated clearly a form
effect on the single hull and a viscous interaction between the twin hulls
(Table 1). The single ellipse exhibited a form factor (1+k) of 1.22 and in the
twin-body (catamaran) mode the measured form factor (1+βk) was between
1.33 and 1.31. The results in Table 1 indicate a viscous interaction of the order
of 7–10 per cent of the single-body viscous drag and that there is little
effective change in (1+βk) with a change in hull separation S/L, a characteristic
that had been observed elsewhere such as in reference [Molland et al., 1996].
Comparison with Numerical Data

The computational fluid dynamics (CFD) investigation was also carried
out using the CFXTM RANS software package [AEA Technology, 1994 and
1997]. This approach allowed the investigation of the total drag, CV, the
pressure and velocity distributions over the single and twin bodies, together
with the distribution of skin friction and an estimate of the total CF for the
single and twin bodies. The approach also has the longer-term potential of
investigating higher Reynolds numbers and hence scale effects. In a similar
manner to the wind tunnel investigation, the formfactors can be derived as CV
/CF.
@seismicisolation
@seismicisolation
Figure 10. Comparison of numerical and experimental pressure distributions, position

0, S/L=(a) 0.37 and (b) 0.27 [Molland and Utama, 2002].
Two models were analysed and comprised ellipsoid(s) with an L/D=6.0

and an equivalent L/∆1/3 of 5.17, as tested in the wind tunnel, and further
ellipsoid(s) with an L/D=10.0 and an equivalent L/∆1/3 of 7.26. The boundaries
@seismicisolation
@seismicisolation
modelled were equivalent to the wind tunnel walls. The CFD investigations
were carried out at a Reynolds number of 2.4×106.
Examples of the numerical prediction of pressure distributions compared
with those derived from the wind tunnel are shown in Figures 10a and b. It is
seen in Figure 10a, for S/L=0.37, that a reasonable level of correlation between
the numerical and experimental results is achieved over the fore-end and mid-
body, although the correlation is not so good at the aft-end. Figure 10b shows
the results for S/L=0.27 where the correlation is again seen to be reasonable,
but not as good as that for the wider spacing of S/L=0.37.
It is seen from Table 2, for the ellipsoids with L/D= 6.0, that the numerical
total drag values are higher than the wind tunnel values in Table 1. Part of the
difference will be due to the lower Reynolds number used in the numerical
investigation, but the main differences are likely to be due to errors in the
estimate of viscous pressure drag arising from a lack of grid quality at the
ends. These were not properly in square form as required, hence failing to
capture the pressure changes with sufficient accuracy, particularly at the aft-
end. Viscous pressure drag decreases as separation of the bodies is increased.
Skin friction drag also decreases with an increase in separation, which is a
consequence of changes in velocity between the bodies. The relative changes
in resistance due to changes in separation of the bodies were again found to be
small. Form factors (Table 2) are also higher than the experimental wind
tunnel values in Table 1, but the predicted viscous interaction between the
ellipsoids, described by CVtwin /CVsingle, is smaller.
The skin friction predictions (Table 2) are within 2 per cent of the ITTC
correlation line [equation (6)], which is very satisfactory considering that the
ITTC line is a skin friction correlation line rather than a true three dimensional
friction line. It is seen in Table 2 that the change in skin friction from single to
twin hulls, equivalent to σ in equation (8), is very small.
Table 2. Results of CFD investigations on ellipsoids: L/D= 6, equivalent

L/1/3= 5.17
CV CF CVpress CSF CV/CF CVtwin/

S/L CSF/CV
x 1000 x 1000 x 1000 x 1000 (1+k) (1+βk) CVsingle
Single 5.581 4.046 1.535 - 1.379 - - -
0.27 5.740 4.105 1.635 3.078 - 1.398 1.028 0.536
0.37 5.663 4.074 1.589 1.571 - 1.390 1.015 0.277
0.47 5.623 4.063 1.560 0.910 - 1.383 1.010 0.162
0.57 5.589 4.049 1.540 0.535 - 1.380 1.000 0.096
@seismicisolation
@seismicisolation
Table 3. Results of CFD investigations on ellipsoids: L/D= 10, equivalent

L/1/3= 7.26
CV CF CVpress CSF CV/CF CVtwin/

S/L CSF/CV
x 1000 x 1000 x 1000 x 1000 (1+k) (1+βk) CVsingle
Single 4.866 4.011 0.855 - 1.213 - - -
0.203 4.921 4.046 0.875 0.559 - 1.216 1.011 0.114
0.27 4.890 4.029 0.861 0.331 - 1.214 1.005 0.068
A greater proportion of the viscous interference arises from the changes in

viscous pressure resistance, or ø in equation (8). The numerical sideforce
predictions (Table 2) are lower than the experimental results but show the
same trends.
Table 3 shows the results for the ellipsoids with L/D= 10.0. The drag and
form factors are reduced, as would be expected. Again, viscous interaction
between the ellipsoids is small. Limited tests at the higher Reynolds number of
1.0×108 were carried out for the single-body case. Decreases in CF due to
increasing the Reynolds number to 1.0×108 were as would be expected and the
form factors were found to be very similar to those derived at the lower
Reynolds number of 2.4×106.
Numerical investigations were also carried out on representative ship hulls
and these are reported in references [Utama, 1999] and [Utama and Molland,
2001]. Further discussions of the results and comparisons between the wind
tunnel and CFD results are given in references [Utama, 1999], [Molland and
Utama, 1997] and [Utama and Molland, 2001].
The overall results indicate that the form factors and viscous interference
effects predicted by the CFD investigations show broadly the same trends as
those derived from the experimental investigation. The CFD investigation also
provided a useful insight into the more detailed aspects, such as the
proportions of total skin friction drag and pressure changes when two bodies
are in proximity.
3.2. Wind Tunnel Test on a Reflex Catamaran
Recently, Jamaluddin et al., [2012] carried out an experimental work

using a wind tunnel on symmetrical catamaran for various hull clearances,
without and with turbulent stimulator at the range of Reynolds numbers
@seismicisolation
@seismicisolation
between 2.89X105 and 4.46X105. A parametric study was carried out on tests
without transition strip because of the mixing of laminar and turbulence flows.
This experimental investigations was performed in order to improve the
fundamental understanding of the viscous drag and viscous interference effects
between twin bodies, such as the hulls of catamarans, and to provide design
data.
Description of Models
The catamaran model were constructed from wooden with identical
dimensions. The catamaran model was a reflex model as shown in Figure 11.
The study objective is to determine the viscous interferences due to pressure
(ø) and flow velocity changes (σ).
The use of reflex models in wind tunnel provides an approximate means
of directly measuring the viscous resistance of the model and without the
generation of waves, as there is obviously no free surface present.
The model catamaran with two identical hulls was constructed from
wooden materials with identical dimensions. The demihulls had an overall
length of 457 mm, a wide of 475 mm and a surface area of 0.0284 m2.
Leading edge roughness (a turbulence strip) was applied to each demihull.
The turbulence stimulation comprising sand grain strips of 0.2mm diameter
and 4mm width. The strips were situated about 5 per cent aft of the leading
edge of each demihull. The static pressure is a constant value that is
determined by the flow outside the boundary layer. The velocity varies from
its free stream value to zero at the wall. The total pressure also varies from free
stream to the wall in the same way that the velocity varies.
Figure 11. Reflex model of catamaran with turbulance stimulator.
@seismicisolation
@seismicisolation
To measure the total pressure, long thin tubes connect the rake tubes to a
pressure transducer located outside of the wind tunnel model.
The blockage effects were minimized by keeping the cross-sectional area
small when compared with the tunnel cross-sectional area (<7%), and the
interference effects of the support structure were minimized by the use of a
minimum of support and by shaping the support structure for minimum drag
[Armstrong, 2003].
Tests and Apparatus

Physical model testing in wind tunnel is focused to determine the changes
in drag, pressure and flow velocity between the two hulls of catamaran in
various clearances. The test set-up model at wind tunnel was shown in Figure
12.
In the wind tunnel simulation, the model assumed stationary and the fluid
(wind) moving at a pace that is determined. Air flow in wind tunnel test
section has moved laterally homogeneous, longitudinally and vertically good
speed, static pressure and intensity turbulence.
Location of test models the front (on the wind tunnel test section) is facing
in the direction flow. Then the pressure and flow velocity between the hulls
can be measured very well. Air flow velocity can be determined by measuring
the static-pressure drop between the two parts of the wind tunnel, which
among the setting chamber and test section.
Figure 12. Test set-up model of catamaran with no turbulence stimulator in the wind
tunnel.
@seismicisolation
@seismicisolation
Figure. Open circuit wind tunnel.
The models were tested in demihull (monohull) and catamaran forms at

separation to length ratios (S/L) of 0.2, 0.3 and 0.4.
The lay-out of wind tunnel with an open circuit system as shown in the
Figure above. This wind tunnel, in the Mechanical Engineering Department of
ITS, has dimension of test section 660 x 660 mm and length of test section
1800 mm.
Test Programs
The experiments were carried out in the 1.8 m×0.66 m low-speed open
circuit wind tunnel at the Institute Technology Sepuluh Nopember, ITS. The
demihull could be adjusted laterally to alter the separation between the two
hulls. The circumferential and longitudinal positions of the pressure tappings
are given in Figure 13. One of the demihull was fitted with 57 pressure
tappings in order to measure the pressure distribution over its surface. Then, a
plot along the lateral static-pitot tube locations is shown in Figure 14. The
models were tested at separation–length ratios (S/L) of 0.2, 0.3 and 0.4.
Figure 13. Circumferential and longitudinal location of pressure tappings.
. . . . . . . . .
inner
. . . . . . . . .
L 0.3 L
Figure 14. Lateral location of pitot-static tube.

Figure 14. Lateral location of pitot-static tube.
@seismicisolation
@seismicisolation
Pressure on the surface of the hull models as measured by some position

on the pressure tap with water loaded (draft) along the hull to determine the
pressure distribution around (between) the ship’s hull. Then the flow rate is
measured with a pitot-static tube connected to a manometer, pitot-static tube
which is placed behind the stomach by a shift in the lateral direction of the Y-
axis as shown in Figure 14. Flow velocity was measured at the back of the
model with a distance of 30% of the length of the ship and the slide laterally
against the hull centerline to cover more than the width of the hulls. In this
case there are 25 and 60 points of measurements for monohull and catamaran
configurations, respectively. Wind speed was set using the wind tunnel
controller and measured using a manometer. The current research was carried
out at the range of Reynolds numbers 2.89x 105 – 4.46x 105 for models
without and with turbulence stimulators.
Data Reduction and Corrections

The acquisition of the surface pressures together with reference static and
dynamic pressures from the tunnel allowed direct calculation of the local
pressure coefficient, CP. Integration of CP was carried out to give the pressure
drag acting on the hull. Wind tunnel boundary corrections were investigated
and applied according to reference [ESDU, 1980]. The only significant
correction was that due to solid-body blockage, but this amounted to a
correction to the force coefficients of up to only 0.2 per cent for catamaran
configuration.
Results and Discussion

The technique of testing in a wind tunnel allows the pressure and flow
velocity on the demihull and catamaran models to be measured directly. The
pressure and flow velocity changes due to change of catamaran hull clearance
were determined using wind tunnel tests. In the case of the catamaran, the
pressure and flow velocities that occur on the outer and inner side along the
hull show a difference. The interference factor ratio (inner/outer) for flow
velocity (σ) shows that the larger the hull clearance the smaller the velocity
difference and vice versa for the pressure ø ratio, as shown in Figures 15 - 16
and Table 4. The pressures were integrated over the hull surface to determine
the viscous pressure drag. The total viscous drag at given wind speed could
thus be determined in which comprising of the frictional drag or shear force
(RF) and the viscous pressure or normal forces (RVP), as shown in Equation
(13).
@seismicisolation
@seismicisolation
RV  R F  RVP (13)
The frictional drag was determined by using ITTC’57 correlation line

[ITTC, 2002; Utama, 1999]. The viscous pressure drag is caused by the
development of a boundary layer along the model owing to the viscosity of the
fluid, and the consequent changing of pressure which increases in magnitude
as the boundary layer develops. This causes a drop in the pressure recovery in
the after part of the hull [Armstrong, 2003].
The tests with demihull and catamaran demonstrated clearly a form effect
on the single hull (demihull) and a viscous interaction between the two hulls of
catamaran, as shown in Table 4.
Table 4. Interference of Flow velocity (σ) and pressure (ø) from wind
tunnel result (on model with turbulence stimulator)
Reynolds S/L=0.2 S/L=0.3 S/L=0.4

number Inner/Outer
Flow Velocity (σ)
2.89 x105 1.0899 1.0571 1.0427
3.47 x105 1.0761 1.0512 1.0399
4.05 x105 1.0511 1.0369 1.0257
4.46 x105 1.0324 1.0228 1.0159
Pressure (ø)
2.89 x105 0.9939 0.9961 0.9979
3.47 x105 0.9923 0.9954 0.9965
4.05 x105 0.9930 0.9964 0.9979
4.46 x105 0.9929 0.9959 0.9967
(a)
@seismicisolation
@seismicisolation
(b)
Figure 15. Ratio of pressure changes based on Re and S/L (on model with turbulence
stimulator).
(a)
(b)
Figure 16. Ratio of flow velocity changes based on Re and S/L (on model with
turbulence stimulator).
@seismicisolation
@seismicisolation
The factor ø has been introduced to take account of the pressure-field

change around the demihulls and σ takes account of the velocity augmentation
between the hulls and would be calculated from an integration of the local
frictional resistance over the wetted surface.
In wind tunnel case, pressure resistance (along with the skin friction
resistance) contributes to the total drag, as written:
CV CAT  CVP  C F (14)
Pressure resistance may be estimated by integrating pressure coefficients

over the hull [Utama, 1999; Armstrong, 2003], as written in Equation 15.
CVP   CP .ds (15)
where CP is pressure coefficient, CVP is viscous pressure resistance coefficient

and ds is distance between two CPs.
Since CP is based on maximum cross-sectional area, CSA (whilst CVP is
based on wetted surface area, WSA), the results must be multiplied by a factor
(CSA/WSA). The new CVP’ is then:
CSA
CVP 
'
x CVP (16)
WSA
where CSA= 0.0007778 m2 and WSA= 0.028444 m2. The results obtained, in
addition to skin friction resistance, are shown in Table 5. The viscous
resistances were corrected due to blockage correction by using the formula of
Glauert [1933].
The variation in Cv with decreasing separation ration S/L is quite large,
and it is quite obvious from Figure 17 that it is almost entirely because of
variations in the value of Cvp (Figure 18). This phenomenon is also described
by Armstrong in his experimental work on NPL catamaran model [Armstrong,
2003].
At low speed regime, the viscous resistance effect is more significant than
that of at higher speed regime.
Figure 17 and 18 shows that viscous resistance and pressure resistance
catamaran with turbulence strip is higher 1,4% - 2,2% than that of without
turbulence strip.
@seismicisolation
@seismicisolation
Figure 17. Effect of hull separation on viscous resistance for models with turbulence
and without turbulence stimulators.
Figure 18. Effect of hull separation on viscous pressure resistance with turbulence and
without turbulence stimulators.
Then, the difference of viscous form factor between catamarans with

turbulence and without turbulence strip is about 1.8% average.
Figure 19 and Table 5 show the test results for the viscous form factor for
demihull (monohull) and catamaran for models both turbulence and no
turbulence stimulators. The form factor for demihull (1+k) and catamaran
(1+βk) are derived from CV/CF using Equation (4) and (9).
@seismicisolation
@seismicisolation
Table 5. Experimental viscous form factor values
1+βk
1+k
Model Catamaran Hull Clearances
Demihull 0.2 0.3 0.4
Turbulence 1.268 1.417 1.409 1.406
No Turbulence 1.254 1.385 1.378 1.394
Figure 19. Comparison viscous form factor (1 + βk) for models with turbulence and
without turbulence.
The present study indicated that the ratio of total viscous resistance for
catamaran over that demihull is obout 1.12 and decreased relatively slowly to
about 1.10 as the separation-length ratio S/L is increased up to 0.4. These
values were found to be broadly similar to published data [Utama, 1999;
Molland and Utama, 2002], which their value is about 1.10 and decreases to
about 1.08 as S/L is increased up to 0.47. Utama [1999] investigated the drag
of ellipsoids in proximity using a low speed wind tunnel with separation-
length ratio (S/L)= 0.27, 0.37, 0.47 and 0.57.
The tests with demihull and catamaran demonstrated clearly a form effect
on the demihull and a viscous interaction between the hulls. The results
indicate a viscous interaction of the order of 10-12 per cent of the demihull
viscous drag and that there is little effective change in (1+βk) with a change in
hull separation S/L.
@seismicisolation
@seismicisolation
Comparison of Form Factor Values from Wind Tunnel and Towing

Tank Test
The model test programme was conducted at the towing tank of
Indonesian Hydrodynamic Laboratory (IHL) which is 235 m long, 11 m wide
and the water depth was maintained at a constant depth of 5.5 m. The model
test used in towing tank is higher 3 times than model used in wind tunnel. The
models were fitted with turbulence stimulation comprising sand grain strips of
0.5mm diameter and 10mm width. The strips were situated (leading edge)
about 5%LBP aft of the bow [ITTC 2002]. The model was connected to the
load cell transducer at a point located amidships and vertically at 0.45T above
base line, allowing the model to move freely in the vertical plane. Total
resistance was measured for each run over the test range of Froude numbers.
The comparison viscous form factor results between experiments in wind
tunnel and towing tank is shown in Table 6 and Figure 20. The viscous form
factor is affected by the change of hull clearance (S/L) for catamaran hull. The
interaction effects due to clearance on the demihull are quite strong.
Table 6. Experimental viscous form factor values for model with

turbulence stimulator
1+βk
1+k
Test Results Catamaran Hull Clearances
Demihull 0.2 0.3 0.4
Wind Tunnel 1.268 1.417 1.409 1.406
Towing Tank 1.277 1.426 1.415 1.410
Figure 20. Viscous form factor comparison.
@seismicisolation
@seismicisolation
Figure 20 show that the smaller the clearance (S/L), the higher the viscous
form factor becomes. This is attributed to the more intensive viscous
interaction between the hulls: primarily to modifications of the boundary layer
and velocity augment between the demihulls and to additional spray associated
with constructive interference of the two wave systems.
The viscous form factor of towing tank data is higher up to 6% than that
of wind tunnel data. Both experimental data shows a similar trend and good
agreement.
REFERENCES
AEA Technology, CFD-FLOW3D Manual, 1994.
AEA Technology, CFX4.2:Pre-processing and Postprocessing, 1997.
Armstrong, T. (2003). “The Effect Of Demihull Separation On The Frictional
Resistance of Catamarans,” Procs. FAST 2003, Ischia, Italy, 7- 10
October, pp. 22-30.
Bertram V., Practical Ship Hydrodynamics, Butterworth-Heinemann, Linacre
House, Jordan Hill, Oxford OX2 8DP, 2000, UK. pp. 74.
Bhattacharyya, G.K. and Johnson, R.A., Statistical Concepts and Methods,
John Wiley and Sons Inc, Canada, 1977.
Couser, P. R., Wellicome, J. F. and Molland, A. F. Experimental measurement
of sideforce and induced drag on catamaran demihulls. Int. Shipbuilding
Progress, September 1998, 45(443).
Couser, P. R., Molland, A. F., Armstrong, N. A. and Utama, I. K. A. P. (2003),
Calm water powering predictions for high speed catamarans, In
proceedings of Fourth International Conference on Fast Sea
Transportation, FAST’97, Sydney, July.
Davies, P. A. O. L., The new 7 ft×5 ft and 15 ft×12 ft low speed wind tunnel at
the University of Southampton. AASU Report 202, December 1961.
Eggers, K., Über Widerstandsverhältnisse von Zweikörperschiffen, Trans.
STG, Vol. 49, 1955.
ESDU (1980), Blockage Correction for Bluff Bodies in Confined Flows,
Engineering Science Data Unit, Reference No. 80024.
Glauert, H. (1933). “The Interference of a Wind Tunnel on a Symmetrical
Body,” ARC RandM, No. 1544, April.
Holtrop, J. and Mennen, G.J., An approximate power prediction method,
NSMB Paper 689, July 1982.
@seismicisolation
@seismicisolation
Hughes G., Friction and Form Resistance in Turbulent Flow, and a Proposed
Formulation for use in Model and Ship Correlation, Transactions of RINA,
Vol. 96, 1954.
Hughes G., An analysis of ship model resistance into viscous and wave
components, Transactions of RINA, Vol. 108, 1966.
Insel, M. (1990). An Investigation into the Resistance Components of High
Speed Displacement Catamarans, PhD Thesis, Faculty of Engineering and
Applied Science, University of Southampton, UK
Insel, M and Molland, A F (1991). “An Investigation into the Resistance
Components of High Speed Displacement Catamarans,” Meeting of the
Royal Institution of Naval Architects.
Insel, M and Molland, A F (1992), An Investigation into the Resistance
Components of High Speed Displacement Catamarans, Trans RINA Vol.
134.
ITTC, Testing and extrapolation methods in resistance towing tank tests,
Recommended Procedures and Guidelines, 2002.
Joubert, P.N. and Matheson, N. (1970), Wind Tunnel Tests of Two Reflex
Geosims, Journal of Ship Research, Vol.14, No.4, pp. 241–276.
Joubert, P. N., Sinclair, T. J. and Hoffmann, P. H. A further study of bodies of
revolution. J. Ship Res., 1978, 22(1).
Joubert, P.N. and Hoffmann, P.H. (1979), An Experimental Study of Viscous
Resistance of a 0.564-Cb Form, Journal of Ship Research, Vol.23, No.2,
June.
Kaklis, P., Papanikolaou, A., The Wave Resistance of a Catamaran with Non-
Symmetric Thin Demihulls, Proc. 1st Nat. Congress on Comp. Mechanics,
Athens, 1992.
Lackenby, H. (1965), An Investigation Into The Nature and Interdependence
of The Components of Ship Resistance, Trans R. Instn Naval Architects,
pp. 107.
Lyon, H. M. The eOEects of turbulence on the drag of airship models. Reports
and Memoranda 1511, Royal Aeronautical Research Council, 1932.
Michell, J.H., The wave resistance of a ship, Phil. Mag., Vol. 45, London,
1898, pp. 106–123.
Molland, A. F. and Utama, I. K. A. P. Wind tunnel investigation of a pair of
ellipsoids in close proximity. Ship Science Report 98, University of
Southampton, 1997.
Molland, A.F., Wellicome, J.F. and Couser, P.R. (1996), Resistance
experiments on a systematic series of high-speed displacement catamaran
@seismicisolation
@seismicisolation
forms: variations of length-displacement ratio and breadth-draught ratio,

Transaction RINA, 138A, 1996.
Molland, A.F. and Utama, I K A P. (2002), Experimental and Numerical
Investigations into the Drag Characteristics of a Pair of Ellipsoids in Close
Proximity, Proceedings of the Institution of Mechanical Engineers:
Engineering for the Maritime Environment, Vol. 216 No.M2.
Papanikolaou, A., Kaklis, P., Koskinas, C., Spanos, D. (1997), Hydrodynamic
Optimization of Fast-Displacement Catamarans Twenty-First Symposium
on Naval Hydrodynamics, 1997, pp. 697-714.
Papanikolaou, A., Nowacki, H., Concept Design and Optimization of a
SWATH Passenger/Car Ferry, Proc. IMAS-89 Int. Conf. on Applications
of new Technology in Shipping, Athens, 1989.
Proceedings of the 8th ITTC, Canal de Experiencias Hidrodinamicas, El
Pardo, Madrid, Spain, 1957.
Proceedings of the 14th I.T.T.C., Report of Resistance Committee, Appendix
2, Ottawa, 1975.
Sahoo, P.K, Doctor, L.J. and Pretlove, L., CFD prediction of the wave
resistance of a catamaran with staggered demihulls, International
Conference on Marine Hydrodynamics (MAHI 2006), India, 5-7 January
2006.
Sahoo P.K., M. Salas, and A. Schwetz, Practical evaluation of resistance of
high-speed catamaran hull forms – Part I, Ships and Offshore Structures,
2:4, 307 – 324, 2007.
Spanos, D., Study of the Wave Resistance of Slender Symmetric and Non-
symmetric Catamaran Configurations, Diploma Eng. Thesis, Dep. of
Naval Arch. and Marine Eng., NTUA, Athens, Feb. 1995.
Utama, I K A P (1999). Investigation of the Viscous Resistance Components
of Catamaran Forms, PhD Thesis, Department of Ship Science, University
of Southampton, UK.
Utama, I. K. A. P. and Molland, A. F. Experimental and numerical
investigations into catamaran viscous resistance. In Proceedings of Sixth
International Conference on Fast Sea Transportation, FAST’2001,
Southampton, September 2001.
Young, A. D. The calculation of the total and skin friction drags of bodies of
revolution at zero incidence. Reports and Memoranda 1874, Royal
Aeronautical Research Council, 1939.
@seismicisolation
@seismicisolation
Chapter 4
ENVIRONMENTAL WIND TUNNELS
Z. Janour and K. Jurcakova

Institute of Thermomechanics, AS CR, v.v.i., Czech Republic
ABSTRACT
A considerable variety of wind tunnels have been designed and
constructed to investigate a wide range of aerodynamic tasks. Nowadays,
environmental wind tunnels designed to simulate Atmospheric Boundary
Layer are very attractive.
They are used to determine air pollution, wind loads on buildings and
constructions, snowdrift, accident with the discharge of harmful
substances, pedestrian wind comfort, etc. Several requirements for
experiments in meteorological or environmental tasks have to be fulfilled
in order to transfer results from small scale wind tunnel experiments to
full scale: 1) proper scaling: matching the scale of the model and the
boundary layer scale; 2) matching dimensionless similarity numbers,
especially the Reynolds numbers; 3) proper simulation of air flow,
including distribution of velocity and turbulence characteristi7cs within
the boundary layer; 4) acquiring the zero pressure gradient for
equilibrium boundary layer. The similarity requirements are the starting
point for the environmental wind tunnel design. Some types of these wind
tunnels are introduced in this chapter, and their construction is briefly
described. Integral part of environmental wind tunnel laboratories is their
experimental equipment.

E-mail: [email protected].
@seismicisolation
@seismicisolation
106 Z. Janour and K. Jurcakova
Special devices are described, including systems for flow

visualization, for turbulent characteristics and concentrations
measurements. Some examples of the tasks solved in the environmental
wind tunnels are mentioned at the end.
INTRODUCTION
Although temporarily abandoned by much of the community in favour of
“not-classical” physics like particle physics or quantum mechanics, the current
popularity of fluid mechanics is now drawing the physicist back. Turbulence is
remaining the last great unsolved problem of classical physics, while the
phenomenon is omnipresent.
It must be solved in many industrial applications, taking into an account
when designing constructions, and increasingly in various environmental tasks
considering the flow in the lower atmosphere is turbulent (with a few
exceptions). Increasing needs for treatment of wind effects from engineering
point-of-view should be mention in favour of geophysics flow investigation.
Losses due to strong winds, increased demand and concern for a human
comfort, serious attempts to control air pollution, and the development and
expansion of energy-production capabilities have resulted into tasks to be
solved by engineering approach. For this, a body of knowledge has to be
accumulated.
The primary elements of this knowledge are found in the disciplines of
meteorology, fluid mechanics, aerodynamics, and structural mechanics. In the
past, fluid dynamic experts predicted that the need for experiments would
rapidly disappear as the computational fluid dynamics would in a rather short
time become capable to solve complex problems and every needed results
would be available from numerical simulations. Nowadays, it is evident that
necessary condition for further research progress is experimental data.
Verification and validation procedures of the computational models need the
excellent experimental data sets. The fundamental tool for experimental fluid
dynamics is a wind tunnel.,
The objective of this review is to bring main characteristics of such
facilities for the atmospheric flow investigation. This is accomplished through
wind tunnel flow characteristics, review of current capabilities for physical
modelling of Atmospheric Boundary Layer flow, and laboratory experimental
methods.
@seismicisolation
@seismicisolation
Environmental Wind Tunnels 107
THE ATMOSPHERIC
BOUNDARY LAYER
The concept of a boundary layer was introduced by Prandtl (1905). In the
atmospheric context it has never been easy to use the Prandtl’s definition
precisely. A useful working definition identifies the boundary layer of air
directly above the Earth’s surface in which the effects of the surface are felt
directly on time scales less than a day, and in which significant fluxes of
momentum, heat or matter are carried by turbulent motions on a scale of the
order of the depth of the boundary layer or less. Atmospheric turbulence is
becoming most conspicuous and important phenomenon there. However, it
differs from most turbulence studied by internal and external aerodynamics.
The atmospheric turbulence associated with thermal convection coexists with
mechanical turbulence (turbulence generated by wind shear), is influenced by
the rotation of the Earth. Reynolds number is a few order greaters in
comparison with the most of the laboratory flows.
The structure of the Atmospheric Boundary Layer (ABL) shows many
similarities to the turbulent layer generated in a wind tunnel (both have a
distinctive inner region and outer region). In the outer region, it shows little
dependence on the nature of the surface and the Coriolis force due to the
Earth’s rotation is important. This region is sometimes referred to as the
Ekman layer. The flow in the inner layer (sometimes called surface layer) is
mainly dependent on the surface characteristics and it is little affected by
rotation. The structure of ABL is strongly influenced by the diurnal cycle of
surface heating and cooling, and by the presence of clouds, over the land in
particular. Three main thermal regimes are defined as:
1. Neutral flow, in which buoyancy effects are absent, may be closely

approximated in the atmosphere in windy conditions;
2. The unstably stratified ABL occurs when strong surface heating
produces thermal instability or convection in the form of thermal
plums;
3. Stable stratified ABL occurs mostly at night, in response to surface
cooling by long wave emission into the space.
More details for the ABL can be find e.g. in Garatt (1992) or Stull (1988).
@seismicisolation
@seismicisolation
Figure 1. Schematic ABL structure for aerodynamic rough flow in neutrally-stratified

conditions.
ABL MODELING7
Our knowledge of the ABL is mostly related to in-situ observations of
turbulent flows. Investigation of the lower part of the ABL (up to the 200-300
m) is mainly utilized by sensors located on tower structures. For studies of the
Ekman layer balloon-borne instrumentation has been used in the past. It is
replacing by remote sensing techniques now. Unfortunately, the experiments
are extremely expensive and they provide mostly local results. That is why the
modelling techniques are frequently applied for investigation. Mathematical
and physical modelling that allows experimentation under carefully controlled
conditions. This is an advantage over real-world field experiments where no
such control is possible. Physical modelling in wind tunnels has great
flexibility for simulation of the ABL subjected to realistic surface condition.
Numerical modelling, on the other hand, has progressed to the stage where
various formulations of turbulent processes can be investigated.
@seismicisolation
@seismicisolation
The basic modelling techniques start from equations of motion.

Appropriate equations expressing the fundamental concepts of mass,
momentum and energy conservation for motion of the atmosphere may be
scale to the following forms:
  u i 
 0
t xi
U i U  u i u j
U j i 
' '
 

P     2U
  0  2i  g 3i
L  
  0 0  2 ijk  jU k
t x j x j xi U 0 L0  x j  U0  (1)
T T u i´t '  k 0    0   2T
Ui    
t xi xi   0 C p 0 0   L0U 0  x k x k
where are: xi – i-th space coordinate, t - time, P - mean pressure, Ui - i-th

component of mean velocity, u’i - i-th component of velocity fluctuation, T -
mean temperature, t’ - temperature fluctuation, ρ – mass density, g -
gravitational acceleration, Cp – specific heat at constant pressure, k0 - thermal
conductivity, ν0 – kinematic viscosity, L0 – characteristic length, U0 –
characteristic velocity, Ω0 - characteristic of angular velocity, δij - Kronecker
delta, εijk -Levi-Civit tensor.
Each of the independent and dependent variables (designated by the
asterisk) has been scaled as follows:
U i* u '* t* * * P* T*
Ui  , u i'  i , t  ,  , i  i , Pi  , T ,
U0 U0 L0 0 0  0U 02 T0
U0
t' * g*
t'  , g
T0 g0 (2)
Numerical modelling has to choose a turbulence model to close the system

of equations (1), to define initial and boundary conditions and to apply method
for numerical solution of the equations. Verification and validation procedures
have to be applied finally.
The basis of physical modelling is a boundary layer formed over the floor
of wind-tunnel. Similarity requirements must be fulfilled in order to transfer
@seismicisolation
@seismicisolation
results from small-scale wind tunnel experiments to the full scale. The general
requirements for geometric, dynamic and thermal similarity can be obtained
by inspectional analysis. One set of requirements for similarity is equality of
the nondimensional parameters shown in equations (1) in brackets. In
summary, the requirements may be summarised as follows:
1) Undistorted scaling of geometry,

2) Equality of the Rossby number Ro=U0/L0Ω0,
3) Equality of the gross Richardson number Ri= [ΔT0/T0](L0/U02)g0,
4) Equality of the Reynold number Re=U0L0/ν0,
5) Equality of the Prantl number Pr=ν0/(k0/(ρ0Cp0),
6) Equality of the Eckert number Ec= U02/(CpoΔT0)
This requirements must be supplemented by the similarity of the boundary

conditions (surface and approach flow conditions) for mean value of
dependent variables as well as main turbulent characteristics.
However, all of the requirements cannot be satisfied simultaneously by
existing laboratory facilities, and approximate simulation must be developed
and used. For example:
 When the surface is a simple plane, surface roughness and

temperature are uniform and roughness elements are small compared
to the surface-layer thickness, the surface layer flow statistics exhibit
planar-homogeneity and the vertical fluxes of momentum and heat are
constant, geometric similarity to Monin, Obukhov's (see Monin,
Obukhov (1965)) turbulence characteristics is applied for micro-scale
flows with horizontal dimension L0 (10-1 - 10) m,
 Reynolds number independence is used for micro-scale flow of
dimension L0  (10-2·102) m. Two flow fields with characteristic
scale H and characteristic velocity U are similar independently on the
Reynolds number for sufficiently large values of the Reynolds number
Re  H U/ (Townsend (1956)) according to this theory. The value
generally depends on the shape of the flow boundaries, on the surface
roughness, and especially on the information, which is required from
the simulation etc. Snyder (1972), e.g. states that for concentrations
above the roof of a cubic building, it is necessary for the simulation -
see Golden (1961) - to set ReH U/>1.1 104, where H is the height
of the building.
@seismicisolation
@seismicisolation
 Laminar-turbulent similarity is used for stratified meso-scale flow of

dimension L0  (2 103 -2 106) m. Than the Reynolds number equality
may be examined by study of the third term on the left hand side and
the second term of the right hand side the of the second equation in
(1). The Reynolds stress can be then approximated by Boussinesq
assumption about eddy diffusivity Km and these terms become
  K    
  m (  )V
 U0 L U0L  (3)
where Km  104  is approximated as a constant. For the laminar model flow

we can assume that Kmm 1  0 and for the turbulent atmosphere p  0.
Therefore similarity requires that
U0L  U L 
    0 
  m  K m  p
(4)
The approximation means that the turbulent shear stress at the topographic
surface is replaced by an equivalent viscous shear stress.
More details can be find e.g. in Snyder (1981).
LABORATORY FACILITIES
Environmental wind tunnels are designed to simulate the atmospheric
boundary layer. They are used to determine different broad tasks, including
flow patterns in the vicinity of group of buildings, wind loads on buildings, air
pollution dispersion patterns, snow drifts, etc. The features of the atmospheric
boundary layer are very different for various terrain characteristics so the
tunnels must have ability to adjust the flow conditions to simulate variety of
circumstances. Therefore a wind tunnel, which will meet requirements for
simulation of the atmospheric boundary layer, must be given special design
consideration. Neglecting the requirements of Rossby number equality, the
1
Index m denotes model index p denotes prototype.
@seismicisolation
@seismicisolation
essential requirements in addition to those usually required for an

aerodynamical wind tunnel are as follows:
1) Capability of developing the Richardson numbers comparable to the

atmosphere ones, it means from the interval ( -0.5 ≤ R i≤ 0.5),
2) Capability of developing thick momentum and thermal boundary
layers,
3) Capability of developing a zero pressure gradient in the direction of
the mean flow.
The first requirement is attainable by making provisions for heating

(cooling) the wind tunnel floor or an air-flow heating placed upstream of the
test section. These capabilities must be accompanied by the ability to operate
at low-wind speeds. Requirement two is best accomplished through use of a
long development section. By this approach the thick shear layer will have the
appropriate mean velocity, mean temperature and turbulent characteristics to
meet the requirements for approach flow similarity. The boundary layer
becomes fully developed at about 10 m from the beginning of the development
section according to Cermak (1975). It should be mention that wind tunnels
with relatively short development sections (less than 10 m) are used too. The
disadvantage of them is that, although grids and vortex generators generate
mean-velocity profiles reasonably similar to atmospheric boundary layer
profiles, a fully-developed layer with unchanging turbulence properties is not
achieved. The basic rule for meeting the requirement three is to keep blockage
of the wind tunnel intersection by the model bellow 5%. Pressure gradient can
be also controlled by varying the wind tunnel ceiling with both the model and
traverse system inside. The criterion for negligible axial static pressure
gradient is following
(5)
where δ is the height of the simulated boundary layer, Uδ is the mean flow
velocity at the top of the simulated boundary layer, ρ denotes the air density p
is static pressure.
@seismicisolation
@seismicisolation
There are two basic types of environmental wind tunnels: open circuit and
closed circuit. There are almost endless variations on the specific features of
various tunnels. The air flowing through an open circuit tunnel follows an
essentially straight path from the entrance to the development section, often
through a contraction, to the test section, followed by diffuser, fan section, and
an exhaust of the air. These tunnels operate in the suction mode of the fan.
There are also open circuit tunnels which operates in the pushing mode, they
are very popular in east Asia region. The fan at the beginning is followed by
honeycombs and contraction to filter out all rotational air movement. The air
flowing in a closed return wind tunnel, Prandtl, or Göttingen type, recirculates
continuously with little or no exchange of air with the exterior.
The wind tunnel of the Institute of Thermomechanics Academy of the
Sciences of the Czech Republic has been designed as an Eifel type with a
closed working section being 1.5 x 1.5 m2 and the length 2 m is an example of
an open circuit wind tunnel. The total length of the wind tunnel is 50 m, the
length of the entrance part being 25.5 m (Figure 2). This length provides a
reasonable boundary layer thickness and maximal velocity reached in the
centre of the wind tunnel intersection is 13 m/s. The honeycomb and screens
section at the entrance guaranties a high quality of the flow. For example: the
initial turbulence level in the working section is about 0.2%, however, it can
be increased up to 30% for measuring scaled-down models in the ABL (about
1:300) by using various turbulence generators and roughness elements (Figure
3). Atmospheric air first enters a space of 6 x 6 x 12 m3 through cloths filters
with a total area of 64 m2 (this space is located in the tower). The wind tunnel
entrance is placed in the middle of it, followed by a section of honeycombs
and screens. The whole system very efficiently minimizes the effect of
spurious turbulent fluctuations and outside gusts, and homogenizes the
velocity profiles in the wind tunnel. At the end of the wind tunnel duct there
are two bends with corner vanes leading to the laboratory roof where a 30 kW
driving fan is placed, generating velocity in the test section from 1m/s to 13
m/s. The great length of the wind tunnel makes possible both forms of
boundary layer development – the nature one, and one enhanced by turbulence
generators. In the latter case roughness elements of various shapes and heights
are fixed to the wind tunnel floor to produce required boundary layer
properties (Figure 3). The working section has glass walls and a built-in
remotely controlled 3-D traversing system carrying probes. The mock-ups of
landscape or urban sections are placed on circular tablets so that by rotating
the tablet different wind directions can be simulated.
@seismicisolation
@seismicisolation
Figure 2. The wind tunnel of the Institute of Thermomechanics.
Figure 3. The wind tunnel duct upstream of the test section can be furnished with
various turbulence generators to enhance the boundary layer development.
@seismicisolation
@seismicisolation
Propagation of emissions from various sources – point, linen, or area ones

can be simulated by different traser gases. Open circuit wind tunnel
“WOTAN” of the Environmental Wind Tunnel Laboratory of the
Meteorological Institute at the University of Hamburg is the biggest European
wind tunnel build especially for environmental tasks. The tunnel length is 26
m, with rectangular cross section 4 m wide and 2.75 to 3.25 m high. The
tunnel consist of an intake nozzle with honeycomb and flow straightener; up to
12 m long flow development zone and the test section with two turn tables.
The flow is driven by an axial blower and whole facility is operated in suction
mode, so that, the flow disturbance by the wind tunnel drive does not effect in
the test section.
Meteorological wind tunnel of Laboratory of the Faculty of Engineering
and Physical Sciences at the University of Surrey is equipped by surface
heating and surface cooling systems for generation of neutral, stable and
unstable boundary layers and neutral and stable free flows. Three dimensional,
computer controlled traversing gear is Fully computer–controlled and allows
operation for unmanned running.
More details it can be find in Bezpalcova (2006).
The Meteorological Wind Tunnel, Fluid dynamics and Diffusion
Laboratory, Colorado State University with a closed test section being 2 x 2
m2 and the length 12 m is an example of the closed circuit. The test section
floor of the tunnel may be independently heated or cooled for generation of
thermally stratified flows. Surface roughness, pressure gradient, and humidity
can be varied readily here. More details it can be find in Plate, Cermak (1963).
The both main types of the wind tunnels have advantages and
disadvantages. The main advantage of the open circuit type is its cost, and
there is no purging problem during application of visualization methods, or
pollution diffusion tasks in contrast to closed type. Corner turning vanes and
screens control well a quality of the air flow and less energy requirement for a
given test-section size and velocity in the closed type wind tunnel. If the open
circuit wind tunnel is located in a room, depending on the size of the tunnel to
the room size, it may require extensive screening at the inlet to get high-
quality flow.
The similar problem may be true if the inlet is open to the atmosphere and
e.g. the weather can affect quality of the flow.
@seismicisolation
@seismicisolation
FLOW INSTRUMENTATION
Integral part of environmental wind tunnel laboratories is their
instrumentation and experimental methods. There are two types of the
equipment of these laboratories, standard equipment of the aerodynamical
laboratories, and special devises or systems. For example devices for
measuring pressure, i.e. manometers, pressure and piezoelectric transducers as
well as flow visualization methods are standard aerodynamics methods. A
reasonable mental image of a flow is almost always necessary for a person to
have a useful understanding of a flow problem. Classic flow visualization,
mostly for low-speed flows, is done by putting something that is visible into
the flow at convenient locations and watching how tracer moves. It belongs to
the group of direct visualization methods that contains e.g. Schlieren methods.
We consider two broad categories of this flow visualization: surface
visualization and field flow visualization. The first one includes tufts, oil flow,
ink dot, china clay. Data-driven visualization in which information is
processed by either analogy or digital methods and produces flow images of
some sort is used too.
Figure 4. Visualisation of the flow inside the street canyon.
@seismicisolation
@seismicisolation
System for visualization used in the Institute of Thermomechanics

consists from fog generator, laser light sheet, high resolution black and white
CCD camera, suitable focusing system, frame grabber/digital camera/video
camera and powerful computer for processing. Qualitative analysis of the flow
inside the street canyonby visualisation method is shown in Figure 4 as an
example of the system results. Velocity and turbulence characteristic
measurement is the most fundamental part of the experiments. Classical
pressure probes are often used for reference velocity measurements; thermal
anemometry and Laser Doppler Anemometry are used to obtain point
measurements in turbulent flows with high temporal resolution. Particle image
velocimetry is another example of a tracer method and yields to area results of
mean velocity and turbulent characteristics. More details can be found e.g. in
Tropea et al., (2007).
Important for environmental wind tunnel laboratories are special
concentration measurements. The flame ionisation detector (FID) is the
industry standard method of measuring hydrocarbon concentration.
Hydrocarbons are used as a tracer gas for dispersion problem tasks. The main
component of the Flame Ionisation Detector is the combustion chamber
(Figure 5) where the flame is constantly maintained by the continuous
insertion of fuel (hydrogen) and air.
Figure 5. Scheme of the burning chamber of the Flame Ionisation Detector with high
temporal resolution.
@seismicisolation
@seismicisolation
There are two different designs of FIDs. The burning chamber is in the
control box and concentration samples of hydrocarbons are sucked into
chamber through a tube from the sampling location. This set-up mixes the
samples before its examination and provides temporal resolution up to 10 Hz.
The other system allows placing the burning chamber very close to the
sampling location and the sample is sucked through a thin needle. This system
offers temporal resolution up to 200Hz. The sample gas is introduced into a
hydrogen flame inside the FID. Any hydrocarbons in the sample will produce
ions when they are burnt. Ions are detected using a metal ion collector and
their numbers, i.e. an electrical current across the collector, is proportional to
the rate of ionisation, which in turn depends upon the concentration of
hydrocarbons in the sample gas. A concentration value can be obtained by the
use of an appropriate calibration curve. The big advantage of this system is
linear response in a wide range of hydrogen concentrations.
APPLICATION EXAMPLES
One of the key aims of European environmental policy is to improve air
quality in European cities and urban areas. To undertake these tasks, various
air pollution problems are investigated in the framework of EC Directive. The
air pollution from transport was studied at first. Nowadays, as consequences of
accidental or deliberate releases, dispersion of airborne hazardous materials in
built-up areas is main target. During the years, different types of numerical
models have been developed. Precisely measured and documented data-sets
are necessary for verification and mainly for validation of the models. These
procedures should improve the quality and quantity of model results
particularly for those dealing with air quality.
Database as well as the best practise validation procedure for numerical
models for the flow and dispersion through large idealise arrays of obstacles,
as a simplified urban area model, was set up in the framework of COST 732
project. As the project used a combination of field trial known as MUST
(Figure 6) – see Biltoft (2001) and atmospheric boundary layer physical
modelling (Figure 7) - see Bezpalcová (2006) to provide comprehensive
validation data set. The physical modelling started with similarity
requirements testing in order to transfer results from small-scale wind tunnel
experiments to full scale. The scale of the model was chosen according to size
of the site and the dimensions of the wind tunnel. The approach flow (i.e. the
boundary condition) was modelled according to ABL open terrain
@seismicisolation
@seismicisolation
classification in VDI Guideline (2000) rather than full scale data. The mean
velocity and intensity of turbulence of approach flow are shown in Figure 8.
Figure 6. Mock Urban Setting Trial.
Figure 7. Model in scale 1: 75 installed in the WOTAN wind tunnel.
@seismicisolation
@seismicisolation
1000
average vertical profile
points used for logarithmic fit
MUST field
800 60
600
ZMS [mm]
Z FS [m]
40
400
20
200
0 0
2 4 6 8 10
U mean [m/s]
Iu, slightly rough lower border

Iu, slightly rough upper border
Iu, modertately rough upper border
field data, south tower, UV measurement
Iu [%], vertical profile, y=-2.25 m UW
50
40
30
ZFS [m]
20
10
5 10 15 20 25 30
Iu [%]
Figure 8. Boundary layer flow properties of the approach flow.
@seismicisolation
@seismicisolation
Detailed flow and dispersion measurement inside the MUST container

array and estimation of the travelling time (the time needed by a passive
pollutant to travel between two points) were carried out and the data-sets
formed the data base for the COST 732 validation procedure. A gaseous
dispersion over a complex urban area after an accident with a leakage of
chlorine from tanker inside a chemical factory and at the area of a railway
station was investigated under neutral thermal conditions, Jaňour et al., (2010).
Diffusion fields in a turbulent boundary layer were simulated using a model of
landscape surrounding the factory at the scale 1:1000 and the railway station in
the scale 1:500. The influence of buildings configuration near the source on
the surface gaseous concentrations was demonstrated. Fundamental
characteristic, including so-called coherent structures of the flow over urban
area was investigated in our laboratory, too. E.g. spatial distribution of
momentum flux and the third- and fourth- moments of the velocity
components obtained from one-point measurement at the inner part of a street
canyon and adjacent intersection was investigated by Kellnerova, Janour
(2011).
Time-averaged momentum flux was divided into particular quadrants
using Quadrant Analysis method. Comparison between flux events and
skewness provided an insight into the spatial distribution of strong intermittent
tendency and explained better the spatially local dominance of sweeps and
ejections.
CONCLUSION
Experimental information useful for solving aerodynamic and
hydrodynamic problems may be obtained in a different ways in interval from
in-situ to model experiments. The fundamental tool for experimental fluid
dynamics is a wind tunnel. The number of atmospheric research tasks is
increasing. Vast of codes connected with atmospheric tasks has been
developed and therefore number of data-sets is needed to validate them. This
is a reason why the number of wind tunnels for geophysical flow
investigations has been built up over the world. Some of them are designed for
special features of motion in the ABL:
 The class of wind tunnels, so called Climatic Wind Tunnels are

facilities that provide the capability to evaluate the drive systems, air
@seismicisolation
@seismicisolation
conditioners, door, and windows seals under simulated hot and cold-
weather environments and under arbitrary chosen operating cycles.
 The rotating flow facilities are constructed for the purpose of
investigating instabilities of the laminar Ekman boundary layer.
 Tornado-simulation facility is constructed for assessment of the main
characteristic of this phenomenon.
Other types of wind tunnels were constructed for investigation of:
 wind effects on buildings and structures investigations,

 loads on buildings,
 pedestrian-level wind conditions,
 windbreaks,
 snow drifting,
 soil erosion,
 etc.
The wind tunnels designed for the flow and dispersion in the atmospheric
boundary layer over different types of surfaces belongs to the most important
case of the environmental wind tunnels. Special requirements for this type of
the wind tunnels results from specific character of the ABL flow. Therefore
the structure of the ABL has been introduced at first. Then the similarity
requirements has been assessed by the inspectional analysis from the equations
of motion in order to transfer results from small scale wind tunnel
experiments/numerical simulations to full scale.
It has been stated that the requirements cannot be satisfied simultaneously
for the tasks to be solved and so called approximate simulation must be
developed. Integral part of environmental wind tunnel laboratories is their
experimental equipment. Special devises are described, including systems for
flow visualization, for turbulent characteristics and concentrations
measurements. Some examples of the tasks were mentioned at the end. Many
of problems were dropped in the chapter.
Problem of simulation of the stratified flow, of Earth´s rotation, details of
the wind tunnel construction could be discussed. Verification of the similarity
requirements, measurement repeatability and reliability were omitted.
@seismicisolation
@seismicisolation
REFERENCES
Bezpalcova K. (2006): Physical Modelling of Flow and Diffusion in Urban
Canopy, Doctoral Thesis, Charles University in Prague Faculty of
Mathematics and Physics.
Biltoft C. A. (2001): Customer report for mock urban setting test, Tech. Rep.,
WDTC-FR-01-121, U.S. Army Dugway Proving Ground, Dugway, Utah,
Cermak J. E. (1975): Applications of Fluid Mechanics to Wind
Engineering – A Freeman Scholar Lecture, J. Fluids Eng., Vol. 97, 9-38.
Garrat J. R. (1992): The atmospheric boundary layer, Cambridge University
Press, 316, Golden J., (1961): Scale Model Techniques, M. S. Thesis,
College of Eng., New York Univ, Janour Z. et al.(2010): Potential risks at
an industrial site: A wind tunnel study, Process Safety and Environmental
Protection 8 8, 185–190.
Kellnerova R., Janour Z. (2011): Flow instabilities within an urban
intersection, International Journal of Environment and Pollution. Vol. 47,
268-277,
Monin A. S., Jaglom A. M. (1965): Statističeskaja Gidrodinamika, Časť I
Izdavatelstvo Hauka, Moskva, 639, Plate E.J., Cermak J.E. (1963):
Micrometeorological wind tunnel facility, description and characteristics,
Final Report US Army Electronic Reserch and Development Activity, Fort
Huachuca, Arizona.
Prandtl L. (1905): Über Flüssigkeitbewegung bei sehr kleiner Reibung.
Verhandl. D. III Int. Math. Kongreß in Heidelberg 1904. Leipzig.
Snyder W. H. (1981): Guideline for fluid modeling of atmospheric diffusion,
Rep. No. 600/8-81-009, Environmental Sciences Research Laboratory,
Office of Research and Development, U.S. Environmental Protection
Agency, Reserch Triangle Park, NC 27711,
Stull R. B.(1989): An Introduction to Boundary Layer Meteorology. Kluwer
Acad. Pub.
Townsend A. A. (1976): The Structure of Turbulent Shear Flow Second
Edition, Cambridge University Press, Cambridge 429, VDI Guideline
(2000): VDI Guideline 3783/12, Physical modelling of flow and
dispersion processes in the atmospheric boundary layer – application of
wind tunnels, Beuth Verlag, Berlin.
@seismicisolation
@seismicisolation
@seismicisolation
@seismicisolation
Chapter 5
USE OF WIND-TUNNELS TO ANALYSE WIND

LOADING ON SCAFFOLD STRUCTURES
H. Irtaza1 and R. G. Beale2

1
Department of Civil Engineering, Aligarh Muslim University,
Aligarh, India
2
Faculty of Technology, Design and Environment, Oxford Brookes
University, Oxford, UK
ABSTRACT
This chapter describes the use of wind-tunnels to investigate the
wind pressures applied to sheeted scaffold structures. A brief description
of the problems of simulating wind in a tunnel is outlined. The tests
required to calibrate a wind-tunnel are described with sample results
presented. Wind-tunnel tests on a model of a clad cubical building and on
a model of the Silsoe Experimental Building are presented together with
pressure results showing the application of the procedures. The chapter
shows that providing the procedures described are carried out that good
agreement between wind-tunnel simulations and full-scale tests can be
achieved. Finally, the method of deriving coefficients required for
computational fluid dynamics calculations for permeable netting from
experimental pressure-velocity measurements obtained in a small wind-
tunnel is given.
@seismicisolation
@seismicisolation
126 H. Irtaza and R. G. Beale
INTRODUCTION
Wind loads on permanent structures have been described for many years
(Baker, 2007; Simiu and Scanlon, 1996) with wind-tunnels used to enable
predictions of the wind pressures to be made. However, when structures are
erected scaffolding is usually placed around the structure and sheeted to
provide access and support to permanent and temporary works during different
stages of construction in the UK and other parts of the world. Scaffold systems
are light in weight, easy to maintain, install, and dismantle. The majority of
previous research was directed at obtaining an understanding of the behaviour
of scaffold under load (Beale, 2007). This chapter is devoted to the use of
wind-tunnels to determine wind pressures.
Wind loads on scaffold structures are usually taken from tables of loads
for permanent structures such as BS6399-2 (BSI, 1997). However, these loads
do not take in to account the short lives of the structure and hence often
overestimate the pressures.
In 1994 the UK Health and Safety Executive (HSE, 1994) organised a
conference at Buxton on the wind loads on scaffolds. In the conference papers
were presented outlining aspects of research. However, many of the papers
described tests on full-scale scaffolds. The only paper describing wind-tunnel
tests was that by Schnabel (1994) describing results fully documented in
(Schnabel, 1993). However, this research did not change the codes of practice
and scaffolding in the UK and elsewhere continued to be designed with wind
forces produced for permanent structures.
In research undertaken by Yue et al., (2005) on the wind loads on integral-
lift high rise scaffolds used wind-tunnel tests to determine wind pressure shape
coefficients and the effects of wind induced vibration on the scaffold
structures. They reported that the Chinese code (CTPIS, 2000) underestimated
wind effects. Recently Amoroso et al., (2010) reported the results of wind-
tunnel tests on partially clad structures which complement the results reported
in this chapter and in (Irtaza et al., 2010, 2012, 2013).
Wind-tunnel tests on bare-pole and sheeted scaffold structures have not
been previously undertaken because of the scaling effect. Most wind-tunnel
models of buildings require that less than 3% of the tunnel is blocked by the
building and hence scales of 1:30 or smaller are often used. For example a
scale of 1:50 requires the diameter of a model scaffold tube to be less than 1
mm and of the sheeting to be 0.008 mm. At these scales the stiffness of the
scaled scaffold tube cannot easily be made the same as that of the full-scale
structure. Pressure taps on the scaled scaffold tubes and on netting/sheeting
@seismicisolation
@seismicisolation
Use of Wind-tunnels to Analyse Wind Loading … 127
cannot easily be fitted. The aeroelastic nature of netting and sheeting requires
wind-tunnel tests on clad scaffolds to be also aeroelastic because any question
regarding the static or dynamic stability of the fabric can only be accurately
answered by an aeroelastic wind-tunnel test. A rigid model test gives no
information regarding the possibility of divergence or flutter, but can be used
to predict fluctuating wind pressure due to buffeting.
Tieleman et al., (1996) and Hoxey et al., (2002) have made comparisons
between full-scale and model tests on building structures. In particular, Hoxey
et al., (2002), Richards and Hoxey (2008) and Richards et al., (2007) reported
on the Silsoe Experimental Building (SEB) which was a 6m x 6m x 6m cube
erected at Silsoe in order to get full-scale data on the wind pressure acting on a
building and to compare the data with that obtained from wind-tunnel tests.
This chapter describes the research undertaken on a model of the SEB at
1:30 scale and the correlations between wind-tunnel data and full-scale tests.
Models were therefore made of the SEB and of two fully sheeted or clad
scaffolds surrounding the building. In the models, the sheeted scaffold
surrounding the building was modelled with acrylic sheeting 4 mm thick (with
a tapered top) with pressure taps implanted both on the inner and outer faces of
the sheeting to measure pressure coefficients. These models were constructed
because were neither full-scale nor model-scale sheeted results available for
the SEB. Scale model results for the unsheeted SEB were available for
comparison (Richards et al., 2007). The models of sheeted scaffolds were used
to validate Computational Fluid Dynamics (CFD) models of sheeted and
netted scaffolds (Irtaza, 2009, Irtaza et al., 2009). The CFD models were used
to find pressure coefficients on sheets and nets surrounding a scaffold. The net
was modelled as a porous media and porous jump boundary conditions were
used to simulate a net. This chapter also describes wind-tunnel tests on two
types of net commonly used in construction industries to determine their
permeability which are required for CFD models.
TESTING FACILITIES REQUIRED FOR

WIND PRESSURE STUDIES
The wind-tunnel used for testing the models described in this chapter was
a 26m long atmospheric boundary layer type as shown in Figure 1. The test
section was 17.5 m long, 3.3 m wide and 2.2 m high. The roof of the test
section could be raised up to a maximum height of 0.6 m to keep the pressure
@seismicisolation
@seismicisolation
constant in longitudinal direction if required. The diameter of the turn table

was 2.80 m. The maximum wind speed in the test section was around 15 m/s.
The atmospheric boundary layer was simulated by using artificial roughness as
shown in Figure 2 with spires and baffles as shown in Figure 3. The wind
pressures were measured by a high frequency pressure scanner which could
simultaneously measure wind pressure at up to 64 locations.
Figure 1. Wind-tunnel.
Figure 2. Artificial roughness.
@seismicisolation
@seismicisolation
Figure 3. Spires and baffles in wind-tunnel.
It is necessary to check the flow characteristics and to calibrate them

before using them for measuring the pressure coefficients on a model. The
objective of checking the flow characteristics is to verify up to what level the
predicted and desired flow characteristics can be achieved in the test section.
Before any calibration test can be carried out, the wind speed in the tunnel
should be properly set. For the purpose of calibration for the wind-tunnel used
in this chapter, four static pressure sensor ports each on the upstream and
down stream of the contraction cone walls were installed. The ports were
located at the centre of each wall. The pressure from four static pressure ports
for both upstream and downstream stations, were averaged by interconnecting
them. The pressure drop in ' p ' (in mbar) across the contraction cone from
upstream and downstream ports was measured using a digital manometer. A
pitot static probe mounted at the centre of the test section was used to measure
the dynamic pressure ' q ' by using another digital manometer. The pressure
drop ' p ' across the contraction cone is plotted against ' q ' is shown in
Figure 4.
The axial static pressure gradient affects the accuracy in predicting drag
on a model in a wind-tunnel and an empirical correction is generally applied to
the measured drag values based on this parameter. However, this parameter
should be as small as possible in the test section. For the measurement of axial
static pressure gradient used in the experiments in the wind tunnel shown in
Figure 1 an aluminium pipe of 50 mm diameter 3 m long was cut from the
centre longitudinally and laid over the test section floor on either side from the
@seismicisolation
@seismicisolation
centre of the turn-table. Equally spaced (200mm) 11 pressure ports along the
length of the pipe were provided. The pressure ports were connected to the
scanner with the help of nylon tube and the scanner was kept below the test
section. Surface static pressures from the 11 ports were measured with
reference to atmospheric pressure using a 32 port electronic pressure scanner.
Figure 5 gives the plot of  p  patm  / q Vs X for a tunnel speed of 5
m/sec. Near zero static pressure gradients existed in the wind-tunnel which is
ideal.
6
Slope = 0.854
Δp (mbar)
θ = 40.497 deg
4
0
0 2 4 6 8
q∞ (mbar)
Figure 4. Speed calibration of wind-tunnel.
0.1
Wind speed = 5 m/s
(p-patm)/q∞
Error band =+/- 0.005

0
-0.1
-1 -0.5 0 0.5 1
X (m)
Figure 5. Static Pressure variations along the test section.
The total pressure variation along the wind tunnel height must be
determined. In the case described a pressure rake spanning half the tunnel
section width was made using a 3-axis traverse system. The total pressure was
measured by using a 32 port pressure scanner. Atmospheric pressure was used
@seismicisolation
@seismicisolation
as a reference. The accuracy of the differential pressure-measuring sensor was

± 0.03 mbar. Figure 6 shows the variation of p0 patm / q along the tunnel
height for different span-wise locations. It can be observed that variation of
' p0 ' with height is 0.3%.
1.03
y = 350 mm
(Po-Pa)/q∞
1.02
1.01
1
0 200 400 600 800 1000 1200 1400 1600 1800
Height (mm)
Figure 6. Variation of total pressure across the tunnel height at different span-wise
locations.
The flow angularity in pitch and yaw planes can be measured using a
Standard Dynamic Model (SDM) attached to balance. The flow angle
calibration can be obtained by the rotation of the drag polar. This method gives
the flow angle of a finite region rather than a point. For the experiments
described below the SDM model was instrumented with a tilt sensor to
measure pitch angle and an internal balance to measure forces. For measuring
flow angularity tests were carried out at model roll angles 0o, 90o, 180o, and
270o. Results of these measurements are shown in Figure 7. Figure 7a shows
the lift coefficient data for 0o and 180o roll angles. The intersection of the two
curves in this figure gives the flow angularity in the pitch plane. The pitch
plane flow angularity is observed to be -0.28 degree. Similarly, from Figure
7b, which shows results for model roll angles 90o and 270o, yaw plane flow
angularity is found to be 0.25 degree.
To obtain correct simulations in a wind-tunnel natural wind needs to be
developed for the scale model being used. In the tests described in the next
section the wind profile was obtained by using roughness blocks of different
sizes. The simulation was done on the basis of Silsoe Research Institute (SRI)
full-scale data (Richards et al., 2007). The velocity profile and the longitudinal
turbulence intensities simulated in the wind-tunnel and data obtained from the
site are shown in Figures 8 and 9. The sampling rate was kept at 100 samples
per second per channel and the duration of each run was
@seismicisolation
@seismicisolation
0.3
0.2
0.1
0
CL
0
180
-0.1
-0.2
-0.3
-5 -4 -3 -2 -1 0 1 2 3 4 5
α (degree)
(a) Model roll angles 0o and 180o
0.3
0.2
0.1
90
CL
0
270
-0.1
-0.2
-0.3
-5 -4 -3 -2 -1 0 1 2 3 4 5
α (degree)
(b) MODEL roll angles 90o and 270o
Figure 7. Variation of lift coefficient with angle of attack in the pitch plane.
kept to 249 seconds. This was equivalent to one hour data in the field. The
mean longitudinal wind speed profile measured in the wind-tunnel was in
good agreement with the SRI full-scale profile with a power-law exponent of
0.17. The longitudinal turbulence was slightly less than the SRI full-scale data.
It is difficult to obtain exactly the same turbulence in a wind tunnel as that
occurring naturally and many iterations often have to be undertaken. This is
done by changing the size and position of roughness.
@seismicisolation
@seismicisolation
Height/Cube Height (z/H)

Wind-tunnel
Silsoe
3
0
0 0.3 0.6 0.9 1.2 1.5
Velocity Ratio U(z)/U(h)
Figure 8. Mean velocity profile of full-scale and wind-tunnel.
2
Height/Cube height (z/H)
Silsoe U
Silsoe V
1.6 Silsoe W
Wind-tunnel
1.2
0.8
0.4
0
0 10 20 30 40 50
Turbulence Intensity (%)
Figure 9. Turbulent intensity profile comparison between full-scale and wind-tunnel.
The small-scale turbulence content (S) which is defined as

S   nSu (n) /     u U  10 (Saathoff and Melbourne, 1986)
2 2 6
u
evaluated at n  10U Lp where, n is frequency, Su ( n) is spectral density,
 u is the standard deviation of the longitudinal mean velocity U  and Lp is

the characteristic model dimension is evaluated and for the tests undertaken
@seismicisolation
@seismicisolation
was found to be 89. This value is in agreement with an equivalent value of 85

found by Ahamad and Kumar (2002) when producing wind tunnel models of
hipped roofs. The model eave height was taken as the characteristic
dimension. The reduced spectra plot at the model eave height is shown in
Figure 10. The integral scale in the wind tunnel was also evaluated at the
model eave height for the longitudinal wind speed and found to be 0.30 m. The
auto-correlation plot to find the integral scale is shown in Figure 11. The
integral scale is defined as area under the auto-correlation curve of the
fluctuating velocity component. Since the auto-correlation measurements are
usually temporal measurements at a fixed point, the Taylor hypothesis can be
used to convert the area under the auto-correlation function into a unit of
length as given by the equation given below:

Lux  U  R( )d
0 (1)
where Lux is the integral length scale, U is the mean wind speed,  is the

time and 
R( )d
0 is the area under the auto-correlation curve (Tieleman et
al, 1996).
1.0E+07
1.0E+06
f*S(f)/Variance
1.0E+05
1.0E+04
1.0E+03
1.0E+02
1.0E+01
1.0E+00
0.01 0.1 1 10 100 1000
f*h/Umean
Figure 10. Normalised reduced spectrum plot at eave height.
@seismicisolation
@seismicisolation
0.8
Autocorrelation
0.6
0.4
0.2
0
0 50 100 150 200 250 300
-0.2
Time in millisecond
Figure 11. Auto-correlation plot at eave height.
However, the auto-correlation function is often highly oscillatory, and the

area under the curve tends to cancel out, which will give an unrealistically
small length scale. A common method of overcoming this problem is to define
the integral scale as the area under the auto-correlation curve for the value
occurring before the first zero crossing and procedure was adopted to calculate
the integral length scale of the wind-tunnel experiments (Scruton, 1981).
For each run, wind pressures measured on the models are expressed in the
form of a non-dimensional pressure coefficient, defined as follows:
p(t )  p0
C p t  
1
U 2
2 (2)
where, p0 is the static (ambient/atmospheric) reference pressure, U the mean

longitudinal wind speed at the reference height (eave height) and  the air
density.
When using a wind tunnel instantaneous wind pressures at each location
should be measured at different wind directions if possible. In the tests
described below measurements were made at 15o wind directions. The mean,
RMS (root-mean-squares) and the maximum and minimum pressure
coefficients, with reference to the gradient wind pressure, should be derived
@seismicisolation
@seismicisolation
from the wind-tunnel data. The maximum and minimum pressures can be
determined by using the method proposed by Cook and Mayne (1981).
Example Tests
Three groups of wind-tunnel experiments were conducted to find the

pressures on scaled models:
1) A scaled cubical SEB (this test was undertaken for validation and
calibration purposes) and is shown in Figure 12 together with the
pressure tap positions (Figure 13).
2) A sheet clad scaffold surrounding the scaled cubical SEB and shown
in Figures 14 – 16.
3) An elevated sheet clad scaffold surrounding scaled cubical SEB.
Figure 12. Scaled model of the cubical SEB.
Figure 13. Pressure tap locations.
@seismicisolation
@seismicisolation
Figure 14. Scaled model of the cubical SEB surrounded by a sheet clad scaffold.
Figure 15. pressure tap locations on the outer faces of sheeting.
@seismicisolation
@seismicisolation
Figure 16. Pressure tap positions on inner faces.
Figure 17. Pressure contours on the roof of the model SEB.
@seismicisolation
@seismicisolation
Figure 18. Pressure coefficient contours on the windward face of the SEB.
In test cases (ii) and (iii) three models were created, one with taps on the
building (in similar positions to those shown in Figure 13), one with taps on
the outer face (the inner faces in this case had the tubing from the taps to the
sensors which would have affected the air flow) and one with taps on the inner
face. The only difference between the tap positions for test (ii) and for test (iii)
was that the spacing of the bottom row of taps on the elevated sheeting was
reduced with the other tap positions kept the same. The third set of tests was
undertaken to investigate the proposition that not cladding the bottom part of a
scaffold would reduce the pressures on the scaffold as it was thought that wind
would be carried behind the sheeting. This was found not to occur as the wind
merely went around the scaffold and not between the sheeting and the
building.
The results of pressures found for cases (i) and (ii) are given in Figures
17-19. The full set of results can be found in Irtaza (2009) and Irtaza et al.,
(2010).
In order to estimate the accuracy of the wind tunnel modelling in the tests
the results of the SEB model (case (i)) were compared with a wind tunnel
model of Richards et al., (2007) at Auckland university and the full scale
results from Silsoe (Hoxey et al., 2002). The authors’ results were shown to be
the most accurate (Irtaza et al., 2012a).
@seismicisolation
@seismicisolation
Figure 19. Pressure coefficient contours on the faces of the sheeted scaffold.
WIND-TUNNEL TESTS ON NETTING

It is difficult to scale sheeting down for wind tunnel tests but the objective
of the tests described above was to give confidence in CFD models of wind
flow around scaffolds. Fully sheeted or clad scaffolds are often replaced in
practice by netted scaffolds. The netting is permeable and can be modelled in
CFD as a ‘porous jump’. A net can neither be tested as an aeroelastic model
nor as a scaled model in a wind tunnel. This is because a thin net cannot be
@seismicisolation
@seismicisolation
scaled further and also instrumentation is not available at the moment to be

used for wind-tunnel modelling within nets. However, the characteristics of
flow across the netting can only be determined by a different set of wind
tunnel experiments.
A commonly used debris net for cladding scaffold was tested for drop in
pressure versus velocity in the small wind-tunnel of the Faculty of
Technology, Design and Environment of Oxford Brookes University. The
tunnel has a cross-section 305 mm  305 mm.
After the working section a diffuser leads to the axial flow fan unit and the
air velocity is controlled by means of a double butterfly valve on the fan
outlet. The fan discharges by way of a silencer. The maximum air velocity is
such that pressure differences of the order of 30 cm of water are developed and
these may be read with suitable accuracy by the simple manometer provided.
A sample of the net is shown in Figure 20. It was manufactured from high
density polythene monofilaments.
Figure 20. Sample debris net.
The objective of testing nets is to determine the drop in pressures versus

velocity so as to simulate the nets as porous media and to determine the
coefficients for the porous media. The mean thickness of the net is required
and was measured with the help of digital micrometer screw gauge. Its average
approximate thickness was measured to be 0.42 mm.
In Computational Fluid Dynamics porous media are modelled by the
addition of a momentum source term to the standard fluid flow equations. The
source term is composed of two parts: a viscous loss term and an inertial loss
term (Fluent, 2006).
@seismicisolation
@seismicisolation
 3 3
1 
Si     Dij v j   Cij  vmag v j  (1)
 j 1 j 1 2 
where S i is the source term for the ith (x, y or z) momentum equation, and D
and C are prescribed matrices. This momentum sink contributes to the pressure
gradient in the porous cell, creating a pressure drop that is proportional to the
fluid velocity (or velocity squared) in the cell (Fluent, 2006). In the case of a
simple homogeneous porous media (i.e. a model of a net):
 1 
Si    vi  C2  vmag vi  (2)
  2 
where  is the permeability and C2 is the inertial resistance factor, D and C

are specified as diagonal matrices with 1/  and C2 , respectively, on the
diagonals (and zero for the other elements).
The experimental setup for the wind-tunnel test of a net is shown in Figure
22. Three pitot tubes were used in the experiment, one at the upstream end of
the working section to determine the static pressure and two moveable pitot
tubes, one each side of the test netting used to obtain the pressure drop across
the net.
Figure 21. Schematic of net within a small wind-tunnel.
The experimental data for the Net is given below in Table 1:
@seismicisolation
@seismicisolation
Table 1. Free stream wind against pressure drop
Free Stream wind Pressure Drop

speed (m/s) (Pa)
2.9 5.0
3.2 10
4.7 17.0
6.8 29.0
8.6 47.0
10.9 84.0
13.0 96.0
14.2 133.0
16.5 165.0
17.8 218.0
20.6 245
A regression curve is then calculated. Using Table 1 we get:
p = 0.524v 2  1.082v (5)
where p is the pressure drop and v is the velocity. Note that a simplified
version of the momentum equation, relating the pressure drop to the source
term, can be expressed as (Fluent, 2006):
p   Si n (6)
where n is the thickness of the media giving the porous jump.

Hence, comparing Eq. (5) and Eq. (6) with
 1 
p    v  C2  v 2  n (from Eq. (4)) yields the following curve
 2 
coefficients:
1
0.524  C2 n (7)
2
@seismicisolation
@seismicisolation
with   1.225 kg/m3, and a porous media thickness n equal to 0.42 mm.
The inertial resistance factor is therefore C2  2037 1 .

m
Likewise,

1.082  n (8)

With   1.7894 10

5
, the viscous inertial resistance factor
(1/permeability) for this net is therefore   6.946 109 m2 .
The values of µ and  are then able to be used in CFD programs. Irtaza
et al., (2009) and Irtaza et al., (2012) show that these values enable the flow
across porous media to be simulated in parametric analyses with results that
agree at the limits of fully permeable and impermeable (i.e. cladded) tests. An
example of the pressure comparison is Figures 21 and 22 taken from Irtaza et
al., (2009).
A permeability of 1.0*10-6 corresponds to a fully permeable sheeting and
gives results similar to those obtained without any sheeting. On the other hand,
a permeability of 1.0*10-10 corresponds to a fully clad scaffold and gives
results similar to that of a fully clad scaffold. As the results of calculations
agree at the two extremes ofa fully clad scaffold and an unsheeted scaffold
(i.e. the building alone) the authors believe that the results are fully
represenative of net clad scaffolds.
1.5
Pressure Coefficient (Cp)
Permeability=1.0*e-
1 06
0.5
0
-0.5
-1
-1.5
-2
-2.5
0 0.5 1 1.5 2 2.5 3
Distance ratio (x/X or y/Y)
Figure 21. Pressure Coefficients on the outer face of the net at different permeabilities.
@seismicisolation
@seismicisolation
1.25
Permeability=1.0*e-06
Pressure Coefficient (Cp) Permeability=1.0*e-07
0.75 Permeability=1.0*e-09
0.25
-0.25
-0.75
-1.25
0 0.5 1 1.5 2 2.5 3
Distance ratio (x/X or y/Y)
Figure 22. Pressure Coefficients on the inner face of the net at different permeabilities.
CONCLUSION
This chapter has described the use of wind-tunnels in simulating the wind
on a sheeted scaffold. The difficulties of obtaining the correct wind velocity
and turbulence profile is delineated and procedures described which enable
reliable results to be obtained.
In addition, a procedure to determine the coefficients required for the
numerical simulation is also described.
REFERENCES
Ahmad, S. and Kumar, K., 2002. ‘Effect of geometry on wind pressures on
low-rise hip roof buildings’, Journal of Wind Engineering and Industrial
Aerodynamics, 90, 755-779.
Amoroso, S., Hebert, K., Levitan, M., 2010. ‘Wind tunnel tests for mean wind
loads on partially clad structures’, Journal of Wind Engineering and
Industrial Aerodynamics, 98, 689-700.
Baker, C.J., 2007. ‘Wind engineering – Past, present and future’, Journal of
Wind Engineering and Industrial Aerodynamics, 95(9-11), 843-870.
@seismicisolation
@seismicisolation
Beale, R.G., 2007. ‘Review of Research into Scaffold Structures’. In: Civil
Engineering Computations: Tools and Techniques, Saxe-Coburg
Publications, Ch 12, 271-300.
BSI, 1997. ‘BS 6399-2: 1997, Loading for Buildings – Code of practice for
wind loads’, British Standards Institution, London, UK.
Cook, N.J. and Mayne, J.R., 1981. ‘A novel working approach to the
assessment of wind loads for equivalent static design’, Journal of Wind
Engineering and Industrial Aerodynamics, 8, 299-301.
CTPIS, 2000. (The Ministry of Construction of People’s Republic of China),
‘CTPIS temporary provisions for integral-lift scaffold structures’, China
Architecture and Building Press, Beijing (in Chinese).
Fluent, 2006. Fluent 6.3 User’s Guide.
Hoxey, R.P, Richards, P.J and Short, J.L., 2002. ‘A 6m cube in an atmospheric
boundary layer flow, part 1. Full scale and wind tunnel results’, Journal of
Wind and Structures, 5, 165-176.
HSE (Health and Safety Executive), 1994. ‘Proceedings of the Conference into
Wind Loading on Temporary Structures’, Buxton, HSE, UK.
Irtaza, H., 2009. ‘Experimental and computational determination of wind loads
on netted/sheeted scaffolds’. PhD Thesis, Oxford Brookes University, UK.
Irtaza, H., Beale, R.G. and Godley M.H.R, 2009. ‘Experimental and
Numerical Evaluation of Computational Fluid Dynamics of Wind Loads
on Sheeted Scaffolds’. In: Proceedings twelfth international conference
on Civil, Structural and Environmental Engineering Computing, Funchal,
Paper 261, 19p.
Irtaza, H., Beale, R.G., Godley, M.H.R, 2010. ‘Wind Tunnel Investigation of
the Pressures acting on Sheet Clad Scaffolds’, Report 397, School of the
Built Environment, Oxford Brookes University.
Irtaza, H., Beale, R.G. and Godley, M.H.R, 2012. ‘A wind tunnel investigation
into the pressure distribution around sheet-clad scaffolds’, Journal of
Wind Engineering and Industrial Aerodynamics, Vol. 103, 86-95.
Irtaza, H., Beale, R.G., Godley, M.H.R and Jameel, A., 2013. ‘Comparison of
wind pressure measurements on Silsoe experimental building from full-
scale observation, wind-tunnel experiments and various CFD techniques’,
International Journal of Engineering, Science and Technology, Vol. 5,
No. 1, 28-41.
Richards, P.J. and Hoxey, R.P., 2008. ‘Wind loads on the roof of a 6 m cube’,
Journal of Wind Engineering and Industrial Aerodynamics, 96, 984-993.
@seismicisolation
@seismicisolation
Richards, P.J., Hoxey, R.P., Connell, B.D. and Lander, D.P. 2007. ‘Wind-
tunnel modelling of Silsoe Cube’, Journal of Wind Engineering and
Industrial Aerodynamics, 95, 1384-1399.
Saathoff, P.J. and Melbourne, W.H., 1986. ‘Freestream turbulence and wind
tunnel blockage effects on streamwise surface pressures’, Journal of Wind
Engineering and Industrial Aerodynamics, 26, 353-370.
Schnabel, P. 1993. ‘Final report on the research project Fluidic model
experiments to determine wind loads on covered façade scaffolding
(Abschlußbericht zum Forschungsvorhaben Strömungstechnische
Modellversuche zur Ermittlung der Windlasten auf bekleidete
Fassadengerüste)’, Report A/18/91, LGA Bavaria, Munich.
Schnabel, P, 1994, ‘Model experiments on covered scaffolding in wind
tunnel’, In: Proceedings of the Conference into Wind Loading on
Temporary Structures, Buxton, HSE, 99-116.
Scruton, C., 1981. An introduction to wind effects on structures, Engineering
Design Guides, Vol 40, Oxford University Press.
Simiu, E. and Scanlan, R.H., 1996. ‘Wind effects on structures: An
introduction to wind engineering’, John Wiley & Sons, New York.
Tieleman, H.W., Surry, D. and Mehta, K.C., 1996. ‘Full/model scale
comparison of surface pressures on Texas Tech. experimental building’,
Journal of Wind Engineering and Industrial Aerodynamics, 61, 1-23.
Yue, F., Yuan, Y., Li, G.Q., Ye K.M., Chen Z.M. and Wang, Z.P., 2005.
‘Wind load on integral-lift scaffolds for tall building construction’, ASCE,
Journal of Structural Engineering, 131(5), 816-824.
@seismicisolation
@seismicisolation
@seismicisolation
@seismicisolation
INDEX
A C
access, 126
calibration, vii, 1, 2, 3, 9, 12, 18, 20, 27, 31,
agencies, 31
47, 48, 118, 129, 130, 131, 136
air quality, vii, viii, 35, 36, 38, 40, 52, 57,
chemical, 121
71, 118
China, 11, 146
air temperature, 13, 16, 40
chlorine, 121
algorithm, 81
circulation, 53
aluminium, 129
City(ies), 35, 118
ARC, 102
cladding, 139, 141
architect(s), 11, 40
classification, 119
Asia, 113
clean air, 56
assessment, 36, 122, 146
color, 37
atmosphere, 106, 107, 109, 111, 112, 115
combustion, 117
atmospheric pressure, 130
commercial, 81, 82
community, 106
B complement, 126
compulsion, 44
base, vii, 2, 22, 25, 26, 27, 28, 101, 121 computational fluid dynamics, viii, ix, 74,
Beijing, 146 88, 106, 125
bias, 20 computer, 23, 115, 117
breakdown, viii, 73 conductivity, 109
bridge model, 3, 7, 12 conference, 126, 146
building blocks, 54, 55 configuration, 40, 86, 95, 121
building code, 2 conflict, 3
Congress, 103
conservation, 42
construction, vii, viii, ix, 3, 11, 74, 105, 122,
126, 127, 147
@seismicisolation
@seismicisolation
150 Index
consumption, 10 energy, vii, 1, 9, 13, 14, 15, 31, 76, 106,

contour, 18, 19, 20, 21 109, 115
cooling, 5, 16, 44, 107, 112, 115 energy conservation, 109
correlation(s), 74, 78, 90, 96, 127, 134, 135 engineering, 32, 81, 82, 106, 145, 147
correlation function, 134, 135 environment(s), vii, 36, 55, 122
cost, 75, 115 environmental conditions, 5
critical regime, vii, 2, 22, 28, 30, 33 environmental policy, 118
critical state, 22, 26, 28, 30 Environmental Protection Agency, 123
CSA, 74, 98 environmental wind tunnel, vii, ix, 1, 2, 3,
CSF, 85, 86, 90, 91 32, 105, 106, 113, 116, 117, 122
CT, 78 equality, 110, 111
CTA, 46 equilibrium, ix, 105
CV, 74, 77, 85, 86, 88, 90, 91, 99 equipment, ix, 105, 116, 122
cycles, 122 ethylene, 48
Czech Republic, 113 evidence, viii, 36, 71
experimental pressure-velocity
measurements, ix, 125
D
data set, 106, 118 F

database, 12
dependent variable, 109, 110 fabrication, 11, 31
depth, 101, 107 filters, 113
designers, viii, 36, 57, 71 flame, 117, 118
diaphragm, 24, 25 flexibility, 108
diffusion, 12, 71, 115, 123 flow field, 7, 32, 36, 110
diffusivity, 111 fluctuations, 8, 28, 33, 64, 113
discharges, 141 fluid, 2, 14, 26, 32, 74, 93, 96, 106, 121,
discretization, 66 123, 141, 142
dispersion, viii, 3, 36, 111, 117, 118, 121, force, 33, 74, 95, 107
122, 123 formation, 27, 79, 80
displacement, 75, 80, 81, 103 formula, 81, 98
distribution, ix, 9, 18, 19, 36, 42, 55, 64, 88, friction, viii, 28, 32, 73, 77, 78, 79, 81, 82,
94, 95, 105, 121, 146 85, 88, 90, 91, 98, 104
divergence, 127 full-scale tests, ix, 125, 127
dominance, 121 funding, 32
draft, 95
draught, 81, 104
drawing, 3, 4, 25, 106 G
geometry, 37, 38, 56, 71, 110, 145

E Germany, 11
gravity, 76
Egypt, 35 grids, 112
emission, 49, 107
employment, 75
@seismicisolation
@seismicisolation
Index 151
H K
hazardous materials, 118 Korea, 33

height, viii, 5, 23, 36, 37, 42, 43, 46, 47, 48, Kui-Ren campus, vii, 1
49, 54, 58, 63, 67, 110, 112, 127, 130,
131, 134, 135
history, 65 L
homogeneity, 110
laminar, 22, 28, 84, 92, 111, 122
House(ing), 10, 102
landscape, 113, 121
human, 106
lead, 75, 79
humidity, 115
leakage, 121
hydrocarbons, 118
light, 2, 117, 126
hydrogen, 117, 118
linen, 115
hypothesis, 67, 134
Lion, 84, 85
local government, 11
I
ideal, 130 M
image(s), 116, 117
magnitude, 36, 81, 84, 96
incidence, 104
majority, 126
independence, 110
management, 7, 8
India, 104, 125
mass, 42, 48, 109
Indonesia, 73
materials, 23, 92
industry(ies), 31, 40, 117, 127
matter, 107
insertion, 117
measurement(s), ix, 2, 13, 15, 16, 18, 20,
institutions, 31
22, 23, 24, 26, 28, 36, 46, 47, 48, 49, 56,
integration, 2, 3, 11, 12, 32, 80, 98
58, 63, 65, 66, 86, 95, 102, 106, 117,
interaction effect(s), 78, 81, 101
122, 125, 129, 131, 134, 135, 146
interference, viii, 73, 74, 78, 79, 80, 81, 82,
media, 141, 143
85, 91, 92, 93, 95, 102
mental image, 116
ions, 118
metal ion, 118
isolation, 78, 79, 80, 82
micrometer, 141
Israel, 32
mission, 2
Italy, 102
mixing, 28, 92
modelling, 106, 108, 109, 118, 123, 139,
J 141, 147
modifications, 2, 102
Japan, 35, 38, 43, 72 momentum, 42, 48, 53, 54, 63, 107, 109,
joints, 5 110, 112, 121, 141, 142, 143
Jordan, 102 multiplier, 75
@seismicisolation
@seismicisolation
152 Index
requirements, ix, 3, 105, 109, 110, 111, 112,

N 118, 122
researchers, 81
neutral, 43, 115, 121
Residential, 72
NPL, 98
resistance, viii, 73, 74, 75, 76, 77, 78, 79,
80, 81, 82, 83, 84, 90, 91, 92, 98, 99,
O 100, 101, 103, 104, 142, 144
resolution, 117, 118
obstacles, 52, 118 response, 48, 107, 118
oil, vii, 2, 22, 27, 28, 29, 30, 116 risks, 123
optimization, 75 root, 19, 20, 22, 135
root-mean-square, 19, 20, 22, 135
roughness, 4, 12, 23, 27, 43, 44, 45, 92, 110,
P 113, 115, 128, 131, 132
particle physics, 106

permeability, 127, 142, 144 S
photographs, 6, 28, 30, 31, 37
physics, 106 safety, 10
pitch, 5, 20, 21, 22, 131, 132 scaling, ix, 76, 105, 110, 126
pollutants, 36, 38, 40 scaling law, 76
pollution, ix, 2, 105, 106, 111, 115, 118 semi-empirical method(s), 75
polythene, 141 sensors, 108, 139
porous media, 127, 141, 142, 144 services, 31
pressure gradient, ix, 6, 40, 105, 112, 115, shape, 37, 38, 54, 56, 84, 110, 126
129, 142 shear, 22, 54, 58, 63, 95, 107, 111, 112
probe, 20, 21, 22, 25, 46, 47, 129 shortage, 66
project, vii, 1, 2, 11, 33, 40, 118, 147 showing, ix, 43, 84, 125
proposition, 139 signals, 25, 49
prototype, 111 simulation(s), ix, 36, 37, 65, 71, 93, 105,
public safety, 2 106, 108, 110, 111, 122, 125, 131, 145
skewness, 121
skin, viii, 28, 32, 73, 77, 78, 79, 81, 82, 85,
Q 88, 90, 91, 98, 104
software, 82, 88
quantum mechanics, 106 soil erosion, 122
solution, 82, 109
R Spain, 104
specific heat, 109
reality, 38 specifications, 9, 13, 19, 43
recovery, 96 stability, 16, 127
regression, 81, 143 standard deviation, 133
regression analysis, 81 state(s), viii, 2, 22, 27, 28, 30, 82, 110
reliability, 122 statistics, 110
remote sensing, 108 steel, 11, 23, 24, 25, 26, 28, 29
stimulation, 92, 101
@seismicisolation
@seismicisolation
Index 153
stomach, 95
stratification, 43
V
stress, 111
validation, 3, 12, 31, 106, 109, 118, 121,
structure, 2, 16, 41, 42, 79, 93, 107, 108,
136
122, 126
valve, 141
supplier, 13, 15
variables, 71
surface area, 74, 75, 77, 81, 92, 98
variations, 16, 17, 19, 20, 26, 98, 104, 113,
surface layer, 107, 110
130
symmetry, 37, 63
vector, 20, 21
ventilation, viii, 36, 40, 41, 42, 57, 63, 64,
T 71
vessels, 75, 81
Taiwan, vii, 1, 2, 11, 31, 32, 33 vibration, 5, 16, 126
target, 118 viscosity, 74, 96, 109
technical assistance, 2 visualization, vii, ix, 2, 3, 22, 27, 28, 29, 30,
techniques, 36, 108, 109, 146 31, 33, 51, 82, 86, 88, 106, 115, 116,
temperature, 10, 13, 16, 40, 41, 43, 46, 86, 117, 122
109, 110, 112
test data, vii, 1, 78, 81
testing, 3, 6, 7, 33, 93, 95, 118, 127, 141
W
titanium, 28
water, 5, 16, 45, 80, 95, 101, 102, 141
tracer gas concentrations, viii, 35, 45
Wave resistance interference, viii, 73
trade, 5
wealth, 22
trade-off, 5
wind speed(s), 40, 53, 55, 56, 85, 95, 112,
transducer, 24, 50, 93, 101
128, 129, 132, 134, 135
transport, 118
wind tunnel laboratories, ix, 105
treatment, 106
wind tunnels, vii, ix, 31, 82, 105, 108, 111,
trial, 118
112, 115, 121, 122, 123
turbulence, vii, ix, 1, 7, 8, 9, 19, 20, 22, 27,
windows, 6, 7, 122
66, 76, 92, 93, 95, 96, 97, 98, 99, 100,
wood, 45
101, 103, 105, 107, 109, 110, 112, 113,
worldwide, viii, 73
114, 117, 119, 131, 132, 133, 145, 147
turbulent flows, 108, 117
X
U
X-axis, 45
UK, 102, 103, 104, 125, 126, 146
uniform, 7, 8, 22, 36, 38, 110 Y
universities, 2
urban, viii, 35, 36, 38, 40, 47, 57, 65, 71, Y-axis, 45, 95
113, 118, 121, 123
@seismicisolation
@seismicisolation

Wind Tunnels Design Construction, Types and Usage Limitations Chaplin

Uploaded by

Copyright:

Available Formats

Wind Tunnels Design Construction, Types and Usage Limitations Chaplin

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Wind Tunnels Design Construction, Types and Usage Limitations Chaplin

Uploaded by

Copyright:

Available Formats

@seismicisolation

Additional books in this series can be found on Nova’s website

Additional e-books in this series can be found on Nova’s website

NOTICE TO THE READER

Independent verification should be sought for any data, advice or recommendations

Library of Congress Cataloging-in-Publication Data

ISBN:  (eBook)

Published by Nova Science Publishers, Inc. † New York

performed in this experiment further provides information regarding the flow

Chapter 4 - A considerable variety of wind tunnels have been designed

J. J. Miau1,, Z. L. Chen1 and C. C. Hu2

two circular cylinders at Reynolds numbers in the critical regime. The

Keywords: Environmental wind tunnel, closed-return loop, circular cylinder,

2. DESIGN ASPECTS OF THE ABRI WIND TUNNEL

Gross Features of the Wind Tunnel

Figure 1 shows a schematic drawing of the wind tunnel, i.e., a closed

Figure 1. A schematic drawing of ABRI wind tunnel. (Miau et al., 2003).

The primary test section is 4 m by 2.6 m in cross section and 36.5 m in

Primary Test Section

The primary test section has dimensions of 4m by 2.6m in cross section

Figure 3. A frontal view of the fan.

These windows can be opened easily, using weight-balancing

Secondary Test Section

The secondary test section has dimensions of 6m by 2.6m in cross section

generally recommended to avoid the flow separation that occurs in a diffuser

Flow Management Sections

Flow management sections upstream of the two contraction sections

The components immediately downstream of the flow management

Figure 1 shows four 90o corner sections, which are indispensable to a

Estimation of the Total Pressure Loss

An estimation of the pressure losses due to the wind tunnel components at

cross sectional area of main test section: 12.0m2

3. FABRICATION AND INTEGRATION

Figure 5. The ABRI wind tunnel with the building.

Results of the Wind Tunnel Calibration

Fan Fan Dynamic Air Velocity Total Energy

100 6.8 60 1.163 10.2 71 0.845

150 27.3 141 1.172 15.5 163 0.865

200 65.4 256 1.168 20.9 290 0.883

250 125.7 403 1.163 1.154 452 0.892

300 222 601 1.158 1.151 641 0.938

330 294.7 727 1.156 1.148 776 0.937

350 350 815 1.151 1.145 869 0.938

IAA testratio based on the power input

100 150 200 250 300 350 400

Figure 7 shows the measurements obtained using an accelerometer

0 50 100 150 200 250 300 350 400

0 10 20 30 40 50 60 70 80 90 100 110 120

5. EXPERIMENT USING A CIRCULAR CYLINDER

A wealth of data concerning uniform flow over a circular cylinder has

Base Pressure Measurements Using Two Circular Cylinders

Figure 15 shows a photograph of an acrylic cylinder model installed

Figure 17. A photo of four diaphragm-type pressure transducers situated in the

The experimental setup can be described using the Cartesian coordinate

During the experiment, the signals measured by the pressure transducers

In (1), if is replaced by P, the instantaneous pressure measured, then Cp

Oil- Film Flow Visualization

It is challenging to conduct a flow visualization experiment at Reynolds

spanwise cell patterns on the cylinder surface at Reynolds numbers around

Figure 21 shows two photographs of a one-bubble state, obtained at

ISBN: (eBook)