AWES-QAM-1-2019

QUALITY ASSURANCE MANUAL: WIND

ENGINEERING STUDIES OF BUILDINGS

AUSTRALASIAN WIND ENGINEERING SOCIETY, 2019

 ACKNOWLEDGEMENTS

The first AWES Quality Assurance Manual for Cladding and Environmental Wind Studies
was published in 1994, with a second edition published in 2001. This third edition is largely
a revision of the 2nd edition, benefiting from comments from many within the wind
engineering community. Current revisions were made by a sub-committee comprising
A.W.Rofail, G.S. Wood and M. Eaddy and approved by the AWES Committee. Contributors
to previous versions of the manual have included (in alphabetical order), P. Carpenter, R.
Denoon, J.C.K. Cheung, J.D. Holmes, K.C.S. Kwok, C.W. Letchford.

1
 PREFACE

This Quality Assurance Manual is intended to provide guidance to the practising

construction industry professional on the conduct of wind tunnel testing and alternative
studies for buildings and structures. This Manual should assist users in specifying wind
tunnel testing appropriately and ensuring that basic testing requirements are met. Guidance
is given regarding the applicability of alternative techniques such as Computational Fluid
Dynamics (CFD).

The opening italicised paragraphs of each section are minimum normative requirements for
wind tunnel testing. The subsequent paragraphs provide informative commentary to allow
the user to understand the basis of the requirements and assess whether the minimum
requirements are sufficient for a particular project.

SCOPE

This Quality Assurance Manual sets out minimum requirements for the wind tunnel testing
of buildings and structures in simulated boundary layers. Although primarily intended for
use in Australasia, the requirements and commentaries are internationally applicable.

The Manual does not cover very specialised testing techniques (e.g. long-span bridge test
techniques), but the general principles in Section A still apply in these cases. Neither does
the manual cover testing in thermally stratified, non-fully developed boundary layers, or
highly transient flows.

2
 PART A. GENERAL

A1. Methodology. Design information should be provided for whatever wind type is
appropriate for the site in question, i.e. gales, tropical cyclones, and thunderstorm
downdrafts/outflows (tornadoes are excluded).

The use of a boundary-layer wind tunnel with acceptable methods of simulation (see
Section A2), is currently regarded as the only acceptable method for generating design
information, irrespective of wind type (i.e. gales, tropical cyclones or thunderstorm
downdrafts/outflows). Computational fluid dynamics techniques are not yet capable of
providing acceptable calculations of the fluctuating and peak pressures on a structure,
although they may be used to investigate mean environmental wind speeds depending on
the turbulence scheme and transient nature of the model used (Section C1).

A general reference for boundary-layer wind tunnel techniques is Reinhold (ed.) (1982).
Information on wind tunnels and modelling techniques is also given in the textbooks by
Aynsley, Melbourne and Vickery (1977), Holmes (2015), and papers like Cermak (1975).
In addition, the American Society of Civil Engineers (1999, ASCE 49-12, ASCE 7-16),
and the Council on Tall Buildings and Urban Habitat (2013) have published guides to wind
tunnel testing techniques for buildings and structures.

A2. Boundary-layer Simulation. The boundary-layer occurring during strong winds

produced by the wind storms characteristic of the extreme winds at the site in question
should be adequately simulated in the wind tunnel.

Wind-tunnel simulation of stationary atmospheric boundary layers requires the basic

characteristics of the natural wind to be modelled at reduced scale. The natural wind at a
site possesses the characteristics of the approach flow modified by adjacent natural
(topographical) and man-made features, giving rise to a “near-field” flow.

Apart from atmospheric dispersion studies, strong wind events will govern engineering
design through strength and serviceability (human comfort) criteria. Under these
circumstances thermal stratification effects, with the possible exception of thunderstorms,
will be insignificant and the approach flow may be considered neutrally stable. For
thunderstorm downburst winds, which have highly-transient characteristics, an acceptable
wind structure is still being developed and hence these flows are modelled with the same
characteristics as gales or tropical cyclones for the present.

These types of flows are usually modelled in boundary layer wind tunnels (BLWT) which
have the common characteristics of large cross sections (2 m or more in width and height)
and long lengths (10 m to 30 m) for flow development.

3
A3. Approach Flow. The minimum requirements for an acceptable simulation of a
neutrally stable atmospheric boundary layer are the modelling of:

(a) the variation of mean wind speed with height,

(b) the variation of longitudinal component of turbulence with height,
(c) the integral scale of turbulence, and
(d) a zero longitudinal pressure gradient.

The mean speed and turbulence intensity in the approach flow shall be modelled to within
10% of their target values. The integral scale shall be within a factor of 3 of the value
determined from the chosen geometric scaling ratio (Section A5).

Suitable models of the atmospheric boundary layer for evaluation of the wind tunnel
simulation are those due to Deaves and Harris (1978), which is incorporated in the
Australia/New Zealand Wind Loading Standard (Standards Australia, 2011), ISO 4354
(2009), and Engineering Science Data Unit (1985 and 1986). These models use a
logarithmic law to describe the mean wind speed profile with the roughness length, z 0,
being the main parameter.

The specific requirements (a-d) can be satisfied by reproducing the full depth of the
atmospheric boundary layer, but this will usually result in a small geometric scale. For
medium to low buildings, partial depth simulations to a height of 100 m are sufficient,
although constant shear stress requirements should be observed (Cermak and Cochran,
1992).

Standard techniques, for producing mean wind speed profiles and turbulence
characteristics in both full and partial depth simulations include the use of spires, fences,
grids and surface roughness elements as described by Reinhold (1982), Melbourne (1977a),
Cook (1978), Holmes and Osonphasop (1983), Barlow et al. (1999), and Cermak and
Cochran (1992).

The approach flow is usually modelled as an equilibrium boundary layer. Thus the
characteristics of the approach flow should be that of the terrain extending upwind of the
periphery of the near field model (Section A4) a distance of at least 60 times the reference
height of the study building taking into account terrain category changes. The zero
longitudinal pressure gradient is usually achieved through use of features such as adjustable
ceilings, divergent walls, or slotted or vented wind tunnel boundaries. An alternative to this
requirement is the application of blockage corrections (Section A7).

A4. Near-field Flow. Adjacent physical features such as significant buildings,

structures or topography, will influence the near field flow and must be included as part of
the local wind flow simulation. In general, all major structures and topographical features
within a radius of at least 300 m from the building site should be modelled to the correct
scale, to an accuracy of 10% or better.

4
The extent to which upwind buildings should be included in the modelling depends on
whether the site is immersed in a city centre, or is a new high-rise development in which
the surroundings are generally low-rise, but with a few isolated high-rise buildings upwind
for some directions.

The accuracy of model detail can be reduced with distance from the site. Section A9
describes the accuracy requirements for the building model under test, and for the
surrounding buildings. New buildings under construction, approved, or planned within the
next five years should be considered, and may require testing of the subject structure in
more than one surrounds configuration.

A5. Geometric scaling ratio (wind model). The geometric scaling ratio of the
atmospheric boundary layer model shall be determined from ratios of the roughness
lengths (z0) and integral scales of longitudinal turbulence (Lu). i.e.:

(𝑧𝑜 )𝑚 (𝐿𝑢 )𝑚
𝐿= =
(𝑧𝑜 )𝑝 (𝐿𝑢 )𝑝

where the subscripts indicate: m for model and p for prototype.

This length scale sets the geometric scale of building models to be used in the wind tunnel.

In practice it is possible to have a distorted integral scale of turbulence ratio, up to a factor

of 2 or 3, (see Section A3), without significantly affecting measurements, Surry (1982).

A6. Velocity scale. The wind tunnel reference mean velocity shall be chosen to
maximise the sensitivity of the measurement instrumentation (e.g. pressure transducers,
force balances, and anemometers). The velocity scale for the simulation is given by:

(𝑉̅𝑟𝑒𝑓 )
𝑉̅ = 𝑚
(𝑉̅𝑟𝑒𝑓 )𝑝

where the prototype full-scale mean velocity (𝑉̅𝑟𝑒𝑓 ) is the design mean wind speed at the
𝑝
reference height at the site in question and taken at the wind speed relevant to the design.

The velocity scale in conjunction with the geometric length scale will determine the time
(T) and frequency scales (N).
𝐿 𝑡𝑚 1 𝑛𝑚
𝑇 = 𝑉̅ = 𝑡𝑝
and 𝑁=𝑇= 𝑛𝑝

5
These scales are necessary to determine wind tunnel instrumentation sampling and
frequency response requirements.

A7. Blockage ratio. The blockage ratio, defined as the projected area of the near field
simulation and the wind tunnel cross sectional area, should be less than 10% to minimise
the requirements for blockage correction. In addition, the building model under study
should not exceed half the wind tunnel height, or be located outside the middle half of the
turntable.

The constraining effect of wind tunnel walls leads to flow patterns not truly representative
of full scale. Blockage corrections to overall mean forces can be applied as outlined in
Reinhold (1982). It is still unclear as to how blockage corrections should be applied to local
fluctuating pressures in highly turbulent boundary layer flow and thus it is recommended
to minimise blockage or adopt tolerant wind tunnel techniques as described in Section A3
to avoid longitudinal pressure gradients. Blockage ratios higher than those recommended
above may be used where blockage tolerant techniques are adopted. The constraining
effects of the side walls of the wind tunnel can be minimised by locating the test model as
centrally as possible on the wind tunnel turntable.

A8. Minimum Reynolds Number. Tests shall be carried out at a Reynolds Number
based on the minimum building width and on the mean wind speed at the top of the model,
of 5 x 104, or greater. In the cases where there is extensive shielding, the minimum Reynolds
Number should be achieved on the subject building without the shielding buildings present.

Flows around bluff bodies such as buildings with sharp corners, are generally insensitive
to Reynolds Number, above a minimum value. Below this minimum value, turbulence in
the free-stream flow, and that generated by the building itself will be deficient in high
frequencies, with consequent effects on the pressure fluctuations and peaks. For buildings
with circular cross-sections, or with corners with large radii, the points of separation may
be dependent on Reynolds Number. This is particularly the case for smooth, uniform
circular cross sections, for which the Reynolds Number should be greater than 8 x 105
(Cheung and Melbourne, 1983; Macdonald et al., 1988, Eaddy 2004). Where buildings
have radiused corners, or where curved surfaces have considerable roughness, (e.g.
exposed perimeter columns, exoskeletons, deep balconies or large sunshades) some
relaxation of these requirements may be possible, but reference to previous research (e.g.
Basu, 1983) should be sought. The use of roughness to simulate super-critical conditions
for mean drag forces is possible using the roughness Reynolds Number scaling method
(Szechenyi, 1975; Holmes and Burton 2016). However, this technique has not been proven
to accurately simulate cross-wind response (Eaddy, 2004).

Some distortion of the model may be required to satisfy Reynolds Number modelling
requirements. For example, balustrades for repetitive floors are normally excluded, cross-
sectional areas of through-site links are often modified as too are porous elements.

6
A9. Accuracy of building model. The overall dimensions of the test building model
(height, plan shape, etc.) should be accurate to within 2%. Architectural details, such as
balconies, mullions, sun shades etc. should be included on the building model if they extend
from the façade by 1 m or more.

The local landscape, including all vegetation, should be modelled in a current or, “as
installed” growth state for pedestrian wind studies (unless the developer is prepared to
install temporary protection for the trees until they reach target size). In these cases, testing
should also be carried out without the effect of vegetation. Vegetation should not be
included for structural or cladding load studies, or for pedestrian safety, but can be used
for pedestrian comfort.

Surrounding buildings should be modelled with overall dimensions accurate to within at

least 10% (Section A4). Architectural details need not be included on the surrounding
buildings.

Building surface roughness elements can be important in governing separation and

reattachment of the wind flow, and hence special care should be taken at edges. However,
the modelling of architectural details is limited by the physical size of the details, and
Reynolds number effects, as discussed in A8. Immediately surrounding buildings need to
be modelled reasonably accurately so that important features such as the sizes of gaps
between buildings are correct. Buildings further away from the test model can be modelled
by approximate block shapes.

For vegetation modelling it is important to ensure equivalence of the flow. To maintain

Reynolds number requirements, modelled vegetation tends to look sparser than the
prototype equivalent. The sensitivity to vegetation modelling can be examined through a
“no vegetation” test. Vegetation should not be modelled for structural or cladding load
studies, or pedestrian safety, as the permanence of vegetation is not guaranteed and foliage
is likely to be stripped during extreme wind events.

A10. Meteorological data. Wind speed data (and direction data if available), properly
corrected for siting, shielding and instrument response effects, for a period not less than
ten years of reliable observational data shall be used for serviceability design cases such
as environmental studies and building accelerations. For extreme design studies, such as
overall structural loading, or façade pressure studies, a minimum of 20 years of reliable
observational data should be used.

Wind speeds and directions given in AS/NZS 1170.2 (Standards Australia, 2011) for
capital cities and defined regions in Australasia can be considered to satisfy the above
criterion. For locations not covered in AS/NZS1170.2, or for other countries, the user
should be satisfied that appropriate corrections have been made and that the record length
is long enough. If insufficient record length is available at a single site, then a superstation
approach should be considered (Peterka and Shahid 1998). In tropical cyclone affected

7
areas, Monte Carlo simulation techniques (e.g. Vickery et al. 2000) may be used to
supplement meteorological data in the prediction of extreme wind speeds and directions.

Either gust or mean wind speed data may be used, provided appropriate corrections are
made to relate to the reference wind speed in the wind tunnel.

Early wind data (i.e. pre-1990s) in Australia were recorded by Dines anemometers, which
had different response characteristics from the automatic weather stations (typically three-
cup or propeller anemometers) that are currently used. The exact change-over dates should
be identified, and appropriate corrections to the recorded gusts made (Holmes and Ginger,
2012). In New Zealand, heavy Munro anemometers were used prior to about the mid-
1990s. Subsequently more responsive Vector and Vaisala anemometers are used, but the
recorded data are averaged over 3 s. Exact changeover dates are available from NIWA.

A11. Directionality. Account should be made of the varying probability of exceedance

of wind speeds with wind direction.

Directional probability calculations can be carried out in a number of ways, depending on

the availability of sufficient meteorological data. These can include:

(a) A calculation of the probability of exceedance for any wind direction, from the
conditional probabilities of wind speed given a wind direction and the coefficients
determined from the wind tunnel for each direction. Methods of doing this have been
described by a number of authors (e.g. Davenport, 1971; Melbourne, 1984; Holmes,
1990; Apperley and Vickery, 1974).

(b) Use of wind directional multipliers in Standards Australia (2011), or of data derived
in an equivalent manner. The method of deriving the directional values in AS/NZS
1170.2 was given by Melbourne (1984).

(c) When directional wind information is not available, the use of an appropriate
statistical reduction factor (Davenport, 1977; Holmes, 1981) may be applied to the
structural response computed from the all-directions wind speeds.

A12. Data Retention. The wind engineering laboratory should retain all wind-tunnel
test results, and instrument calibration data for a minimum period of seven years to assist
with peer reviews as required.

8
 PART B. WIND-TUNNEL SURFACE PRESSURE MEASUREMENTS

B1. Number of pressure measurement positions. The average pressure tap density
should not be less than 1 pressure tap per 120 m² of surface area on the test building.

The density given is a minimum average value. The local density should be higher near
wall edges and discontinuities of cross-section. Higher average densities should be used
for smaller buildings and complex shapes. In some cases, it may be practical to perform
wind tunnel tests in multiple stages, with subsequent stage positions selected to improve
the definition of the first stage results. A similar approach may be required for special
studies where the presence of exposed pressure tubing would significantly influence the
flow pattern around the structure. For large areas of low-pressure gradient, the average
pressure tap density may be reduced below 1 pressure tap per 120 m², with justification.

B2. Number of wind directions. Design pressures for individual measurement

positions should be determined for at least 36 wind directions at 10° increments.

For special studies, the requirement for 36 wind directions may be relaxed with
justification, or increased for studies investigating flow phenomena that occur over narrow
bands of wind directions.

B3. Pressure coefficients and reference pressures. Wind tunnel pressure

measurements shall be provided in the form of coefficients of peak external pressure with
respect to a static (atmospheric) pressure at the building site in the absence of the building.
A suitable dynamic pressure based on a gust or mean wind speed, that can be directly
related to appropriate meteorological data, should also be used.

The static pressure reference can be obtained from the static holes of a pitot-static probe
mounted in a low-turbulence region of the wind tunnel away from the direct influence of
any near-field building. Alternatively, it could be obtained from a probe especially
designed to measure static pressure in turbulent flow conditions, or from a pressure tapping
in the wall of the wind tunnel. However, possible lateral or longitudinal static pressure
gradients in the wind tunnel should be investigated, and appropriate corrections made.

The dynamic pressure may also be obtained from a pitot-static probe, but the latter should
not be located at a position where the longitudinal turbulence exceeds about 10%, or the
reference dynamic pressure will be overestimated.

B4. Extraneous acoustic pressure fluctuations. The contribution of extraneous

acoustic pressure fluctuations to the total mean square pressure fluctuation for any
measurement point should not exceed 5%.

9
The generally incompressible pressure fluctuations generated by turbulent and separated
flows around building models should be distinguished from compressible pressure
fluctuations, generated as a result of wind-tunnel ductwork or other components and which
are specific to a particular wind-tunnel design. Acoustic pressure fluctuations in wind
tunnels can be produced by a number of sources – by vorticity generated by the fan blades,
longitudinal (organ pipe) or lateral acoustic resonances, unsteady flow in diffusers or the
vibration of turning vanes. Acoustic pressure fluctuations manifest themselves as spikes on
the spectral densities of pressure fluctuations, or as high correlations at certain frequencies.

B5. Frequency response. The amplitude frequency response of the pressure measuring
instrumentation, represented by the ratio of recorded amplitude to applied amplitude of a
sinusoidal pressure, should be shown to not depart by more than 10% from a constant
value from 0 Hz to a frequency, in model scale, equal to 2Uh/B Hz (where Uh is the mean
wind speed at the top of the building, and B is the characteristic building model width).
Low pass filtering should be applied to the pressure signal beyond the frequency range
where the frequency response to the pressure signal begins to depart from a uniform
response by more than 10%. The phase response should also not depart by more than 10%
from a linear variation over this frequency range.

The frequency response of a pressure measurement system depends on the geometry of the
pressure tubing system and on the volume and diaphragm flexibility of the pressure
transducer or sensor. A variety of techniques are available to achieve the required
frequency response criteria (Holmes, 1995). The requirement for the response limit given
has been well established by Durgin (1982), Holmes (1984), Irwin and Davies (1988), and
Letchford et al. (1992), amongst others.

B6. Digital sampling rate. The minimum rate of sampling of a pressure signal for
digital processing should be at least three times the maximum frequency within the signal
to minimise aliasing, or 6Uh/B, whichever is the larger, (Uh is the mean wind speed at the
top of the building model, and B is the characteristic model width). A low pass filter should
be less than half the sampling frequency to prevent aliasing in the data.

Experience and tests have indicated that a sampling rate of three times the highest
frequency in the pressure signal, is sufficient to resolve the peaks to an acceptable level of
accuracy. Appropriate low-pass filtering can be achieved through analogue or digital
techniques, however well-designed digital filtering may require an increased sampling rate.

B7. Determination of peak external pressure coefficients. The peak external

pressure coefficient for design shall be determined as the statistical average (maximum or
minimum) for a defined windstorm length in full-scale (not less than ten minutes or more
than three hours).

10
Various methods can be used to efficiently determine the average external extreme pressure
peaks – e.g. up-crossing counts (Melbourne, 1977b), sampling of maxima and minima
(Peterka, 1983), determination of the extremes for shorter periods than the defined
windstorm length and fitting of an extreme value distribution (Mayne and Cook, 1979).
Alternatively, if the average is taken of multiple extremes in repeated identical (ergodic)
experiments, no less than five time-series samples should be used.

For elements tested at a large length scale (>1:50), the spectral density cannot be fully
modelled in the wind tunnel and a different testing and analysis technique should be used
to estimate the peak response. Such partial turbulence matching techniques are described
in Banks et al. (2015).

B8. Internal pressures. The external pressures measured on a building model should
be combined with internal pressures that have been determined according to the porosity,
wall and roof openings and internal building geometry, in the prototype building.

Internal pressures acting on a particular cladding panel depend on many factors, including
porosity or leakage in the building envelope, whether intentional or unintentional, openings
such as air-conditioning inlets or outlets, open windows, the distribution and porosity of
internal walls and partitions, thermal (stack) effects in lift wells, etc. Usually it is not
feasible to attempt to model these effects and measure internal pressures on a building
model. However, conservative estimates for internal pressure coefficients to combine with
the measured external pressures should be made. If there exists a possibility of openings in
the building envelope occurring during wind storms (e.g. from openable windows or
accidental damage), reasonable estimates of internal pressure should be made based on the
probability of occurrence of such openings, or through simultaneous internal and external
pressure measurements. If modelling the internal pressure in the presence of a dominant
opening then proper scaling of the internal volume will need to be carried out to account
for Helmholtz resonance effects.

B9. Probability of exceedance of design pressures. The probability of exceedance for

the design pressures on cladding shall be selected to be compatible with the design criteria
of the Standard or code used to design the cladding material.

Reference should be made to the appropriate governing document such as the National
Construction Code (Australian Building Codes Board), the New Zealand Building Code,
Standards Australia (2002), or NZS 4223.4:2008 for wind loading on glass.

B10. Reporting. The design cladding pressures resulting from the study should be
presented in a way as to simplify their interpretation for the cladding designer.

The report for a cladding pressure study of a building should present the essential results
of the study as a series of diagrams of the plan and elevations of the building, showing the

11
design cladding pressures as zones of equal pressure. The zones may be irregular in shape,
representing the actual measured pressures, or they may be rectangular in order to
approximate the measured pressures in a form suitable for cladding selection and design.
The zones should typically be calculated for intervals of about 0.5 kPa, with the zone
pressure being equal to or greater than the predicted pressures within the zone. However,
in areas that experience very strong winds, it may be more convenient to use intervals of
1.0 kPa. The zoning intervals selected should be appropriate to both the intended cladding
system and the pressure gradients measured. The zone pressures should reflect the
combined effects of both positive and negative peak pressures, and both external and
internal pressures acting on the cladding. The individual design pressures used to prepare
the zonal diagrams should also be tabulated.

In addition to the results, the report should also include sections relating to the requirements
of this Manual, i.e.:

(a) A general description of the test procedure and model construction, wind tunnel
and atmospheric boundary-layer simulation. Photographs of the wind tunnel model
containing both the test and proximity models should be included.
(b) A description of the design wind speed analysis and consideration of directional
effects.
(c) Diagrams of individual pressure measurement positions, labelled with pressure tap
identification numbers.
(d) Specification of the wind directions tested, measurement duration and frequency
response.
(e) The statistical analysis of the measured external pressures.
(f) Analysis of internal pressures, and the method of incorporation into the calculated
design pressures.
The following additional information may be required by a client, and it is good practice
to make it available:

(g) A full listing of the measured mean, standard deviation, minimum and maximum
pressure coefficients for each wind direction, in either a tabulated or graphical form.

Finally, any unusual details of the study should be described. Unusual flow effects may be
recorded photographically or on video.

12
 PART C – ENVIRONMENTAL WIND STUDIES

C1. General. Sufficient information should be provided for the adequate assessment of
human comfort and safety of the intended pedestrian use areas around building
developments.

Assessment methods may vary from desk studies based on fundamental principles of wind
flow around isolated bluff bodies, to wind tunnel tests for more complex environments.
For the latter, the use of a boundary-layer wind tunnel with acceptable methods of
simulation (see Part A) is currently regarded as the most acceptable method. Computational
Fluid Dynamics techniques continue to have limitations when determining gust wind
speeds, or where the development is located within a suburban or urban environment
(Cochran and Derickson 2011). Nevertheless, CFD may be used to supplement a
qualitative desktop assessment. Guidance with regards to the appropriate use of CFD is
given in Franke et al. (2007), Blocken (2015), and Section E4 of this manual.

Fundamental flow patterns around building configurations, and their effect on the local
wind environment are described by Penwarden and Wise (1975), Aynsley et al. (1977).

C2. Number of wind speed measurement positions. The average density of wind
speed measurement positions should not be less than 1 position per 200 m² of the
pedestrian accessible plan area of the development site and adjacent pedestrian
thoroughfares or other outdoor areas that are accessible by occupants of neighbouring
buildings, if they are likely to be impacted by the proposed development.

From a pedestrian safety perspective, wind speed measurement positions should be

concentrated at points where wind environment problems are typically known to occur,
e.g. near corners, in front of exposed windward faces, and in open arcades linking
windward and leeward faces. For pedestrian comfort, the intended use of the space for a
specific activity (e.g. outdoor seating, swimming pool, or restaurant) should influence
measurement locations. AWES (2014) recommends minimum areas to be investigated
around developments. Where the site incorporates large open areas, relaxation of the above
measurements position density may be acceptable. Care should be taken when modelling
enclosed arcades for Reynolds number effects, see A8.

C3. Measurement of wind speed. Wind-tunnel speed measurements should be made

with an instrument with response characteristics to match the criteria for human comfort
with which the measurements are being compared. Measurements should be expressed in
the form of a ratio with a suitable reference speed that can be directly related to
meteorological data.

Wind environment assessment criteria have been expressed in terms of both gust (peak)
and mean speeds. Fluctuating and peak speed measurements may be made with hot-wire

13
or hot-film anemometers. Irwin probes may also be used (Irwin, 1981). Hot-wire or hot-
film sensors are subject to errors in the mean wind speed in regions of high turbulence
(Bottema et al., 1992). Irwin probes are also subject to errors in regions of high turbulence,
vertical flows, and with local wind-tunnel mean flows less than about 3 m/s.

Some laboratories have developed techniques for assessing the mean wind environment by
recording the erosion of a bed of material using digital photography. It has been found that
such techniques give repeatable results, particularly at windy locations on the windward
face where the erosion occurs first. When combined with climate data, images showing the
wind comfort categories based on mean wind speed can be produced.

The reference speed should be obtained from a position in the wind tunnel that is not
affected by the near-field buildings under study.

C4. Frequency response. When gust speed measurements are carried out, the
amplitude frequency response of the measuring instrumentation should be shown not to
depart by more than 10% from unity from 0 Hz to a frequency equivalent to 0.15 Hz in full
scale. The phase response should not depart by more than 10% from a linear variation
over this range.

The 0.15 Hz limit corresponds approximately to a 3 second averaging time in full scale,
this being the approximate response characteristic of meteorological anemometers used to
measure the gust speeds on which the acceptable criteria are based. This frequency
requirement can be easily met by hot-wire or hot-film anemometers and by pressure
measurement techniques. For instrumentation based on pressure measurement techniques,
the frequency response depends on the pressure tapping geometry, the dimensions of the
connecting tubing (including any restrictors etc.), the volume and diaphragm flexibility of
the pressure transducer or sensor, and on the characteristics of any pressure scanning device
(e.g. ‘Scanivalve’). A variety of techniques are available to achieve the required frequency
response, (e.g. Holmes, 1995), and digital corrections are appropriate, e.g. Halkyard et al.
(2010).

C5. Number of wind directions. Local environmental wind speeds should be

determined for at least 16 wind directions.

The actual directions chosen should depend on the sensitivity of the wind speeds at critical
locations to direction, and on availability of directional meteorological wind speed data.

C6. Assessment criteria. The assessment of the local wind environment shall be based
on the criteria required by the relevant planning authority regulations. In cases where the
planning authority has not specified criteria, the consultant should nominate suitable
criteria, which take into account the proposed activity at the location and an appropriate
level of probability of an unacceptable wind speed.

14
The most widely used criteria are those by Davenport (1972), Lawson (1975, 1990) and
Melbourne (1978b). Only the safety limit is usually adopted from the latter. AWES (2014)
provides guidance with regards to the safety criterion. Temperature and humidity may also
affect the acceptability of a given wind speed for pedestrian comfort, and these factors may
also be considered in the selection of suitable criteria for a location. Other criteria and
further background information on criteria development may be found in Hunt et al.
(1976), Isyumov and Davenport (1976), Lawson and Penwarden (1976), Penwarden and
Wise (1975), Soligo et al. (1998), Janssen et al. (2013), Ratcliff and Peterka (1990), and
Koss (2006)

Particular care should be taken for locations situated in semi-enclosed spaces, which are
treated differently to outdoor spaces from a human perception perspective. In such semi-
enclosed spaces where stationary activities are intended, the use of a criterion to achieve
wind conditions significantly better than the long duration stationary activities in outdoor
areas is recommended. An example is the ASHRAE criterion of 1.2 m/s based on a 50
percentile wind speed.

C7. Reporting. The assessment of the local wind environment for a development should
be presented in a way that simplifies the interpretation for the client.

The report should present a summary of results which indicate the suitability or otherwise
of the proposed activity for the sites investigated, e.g. thoroughfare, outdoor restaurant etc.

In addition to the results, the report should also include sections relating to the requirements
of this manual, i.e.:

(a) A general description of the test procedure and model construction, wind tunnel
and atmospheric boundary-layer simulation.

(b) A description of the analysis for the determination of the wind climate, including
analysis of directionality effects.

(c) Diagrams of measurement locations, labelled for identification.

(d) Specification of the wind directions tested, measurement sample time, and
frequency response.

(e) Effects of variation in vegetation modelling (i.e. a ‘no vegetation’ study to check
the sensitivity of the results to this effect).

Finally, any unusual details of the study should be described. Unusual flow effects may be
recorded photographically or on video.

15
 PART D - WIND-INDUCED OVERALL LOADS AND DYNAMIC RESPONSE

D1. Test objectives. Wind tunnel tests for overall loads and response of tall buildings
shall enable accurate determination of design base bending moments and shear forces,
torsional moments, effective static load distributions, and prediction of wind-induced
deflections and accelerations.

Base moments and shear forces are associated with strength design and are hence
calculated using either very long return period wind speeds associated with ultimate limit
states design approaches or a return period wind speed associated with a building lifetime
and applying a load factor to the loads predicted. Deflections and accelerations are
generally associated with serviceability design, and normally use shorter return periods.
Where wind-induced vibrations of building developments are expected to be significant,
assessment of accelerations with respect to occupant comfort criteria should be made.

D2. Building dynamic characteristics. Sufficient information regarding the structural

dynamic characteristics of the building should be provided to the wind tunnel laboratory
by the project structural engineer. These must include the mode shapes and natural
frequencies of at least the first three modes of vibration of the prototype building, mass and
mass moment of inertia distribution, and damping characteristics for the prototype
building.

The above information, except damping, would normally be obtained from a computer
model of the structure. For very tall, slender or complex-shaped buildings, the effects of
higher modes of vibration may need to be considered. Damping may be estimated from
code values or by reference to full-scale measurements of similar structures (e.g. Tamura
et al., 2000).

D3. Methods of testing. Measurements shall be made by either (a) high frequency
balance testing, (b) simultaneous multi-channel pressure measurement. (c) linear mode
aeroelastic model testing, or (d) full aeroelastic model testing. Aeroelastic testing shall be
used for cases in which reduced velocity (Uh/n0B where Uh is the mean wind speed at
building height, n0 is the fundamental natural frequency of vibration, and B is the
characteristic building width) exceeds 8 for sharp-edged building sections, 6 for
hexagonal, octagonal or chamfered corner building sections, or 4 for circular or oval
building sections.

Full aeroelastic modelling is uncommon and is used only where higher order modes are
likely to provide a significant contribution to the structural response. For the large majority
of cases, an aeroelastic model is not required and a rigid model test (force balance or
simultaneous pressure measurement technique) can be used to obtain the wind force
spectrum. The force spectrum can be modified to obtain the response spectrum through the
application of a mechanical admittance function (for further explanation of this see

16
Holmes, 2015). However, the use of a force balance is not appropriate in cases in which
aerodynamic damping (displacement dependent effects) is expected to be significant or
approaches the level of structural damping. High aerodynamic damping in tall buildings
can occur in a low-turbulence environment (for example, facing the ocean) and where the
prototype building is exposed to design reduced velocities that are more than 80% of the
critical reduced velocity for the building section concerned. The limiting reduced velocities
represent conditions where these simple building plan shapes will typically experience
significant aerodynamic damping (Steckley, 1989).

D4. Geometric modelling. The overall dimensions of the test building model, height,
plan shape, etc., should be accurate to within 2%. Architectural details with a dimension
of greater than 1 m, or 10% of the building width, whichever is smaller, and closer than
15% of the building width to any corner should also be modelled. The geometric scaling
ratio should not be less than 1:800 and at the same time shall satisfy the requirements for
minimum Reynolds Number stipulated in Section A8.

Building surface roughness elements can be important in governing separation and

reattachment of the wind flow, and hence special care should be taken at edges. However,
the physical size, appropriateness from a flow modelling perspective, and the practicality
of incorporating small details into the building model limit the modelling of the
architectural details. A suggested lower limit to the detail of these elements would be all
features that have a maximum sectional dimension of less than 1m.

D5. Test requirements. The sample rate should be at least 5 times the scaled natural
frequency of the building at its design wind speed. Testing should be performed for a
minimum of 36 wind directions and a minimum sample period equivalent to 1 hour
duration in full-scale.

Irrespective of the sampling duration and sampling rate a sufficient data sample needs to
be acquired to provide sufficient resolution in the response spectra. Finer increments of
measurements can be used to investigate critical directional effects.

D5.A High-frequency balance testing. The model/balance frequency should be at least

30% greater than the scaled building natural frequency for the relevant loading direction.

In the high-frequency balance test, the model is mounted on a very stiff balance capable of
measuring the overturning moment on the model. Typically, a very lightweight model must
be used, with a stiff mounting method, so that the model/balance system has a very high
natural frequency (Tschanz and Davenport, 1983, Boggs, 1991).

The high-frequency balance method attempts to measure direct aerodynamic forces, but
these may be distorted by the natural dynamic characteristics of the balance and model.
Separating the model/balance frequency from the scaled building model frequency will
minimise the errors introduced by this distortion.

17
D5.B Simultaneous multi-channel pressure measurement. When multiple pressure
transducers are used to determine the fluctuating overall wind forces acting on a building
as a combination of the fluctuating local pressures on the walls of the building, the
following minimum criteria should be observed:

a) At least one pressure tap should be allocated for every 120 m2 of building surface.
b) Intermediate fluctuating wind loads per unit height should be determined from at
least five levels on the building.
c) The amplitude frequency response of the pressure measuring instrumentation,
represented by the ratio of recorded amplitude to applied amplitude of a sinusoidal
pressure, should be shown not to depart by more than 10% from a constant value
from 0 Hz to a frequency, in model scale, equal to the higher of 0.5Uh/B Hz (where
Uh is the mean wind speed at the top of the building and B is the minimum building
model width or 2.5·n0, where n0 is the lowest natural frequency of the full-scale
building). The phase response should also not depart by more than 10% from a linear
variation over this frequency range.
Using this technique, the pressures must be measured simultaneously such that the phase
lag between any pair of pressure taps is less than 30°.

Extreme care should be taken to not distort the dynamic characteristics of the pressure
tap/tubing/pressure transducer system (e.g. Halkyard et al. (2010)), and appropriate modal
and area weighting factors must be established for each tap for use in the synthesis. One
constraint of this pressure approach for tall buildings is the difficulty in getting enough
pneumatic tubing into the model for simultaneous data collection, particularly for complex
articulated facades where it would be extremely difficult to simultaneously capture the
effect of these façade articulations. However, for lower buildings, there is typically ample
room for sufficient tubes to accurately define the pressure distribution, and the roof uplift
contribution (which may be significant for a long, low building). The high-frequency
balance encompasses this addition to the base moments in a manner that may overestimate
the dynamic design loads. Additionally, actual non-linear and coupled mode shapes can
easily be accommodated directly in the simultaneous pressure approach.

D5.C Aeroelastic testing. For all types of aeroelastic modelling, frequency ratios
between modes should be modelled to an accuracy of within 5%. Where translational mode
shapes can be approximated to within 20% by a linear mode shape, a linear mode
aeroelastic model may be employed. For more complex mode shapes, a full multi-degree
of freedom aeroelastic model should be employed.

If the aeroelastic model technique is used, then the rig must be set up so as to replicate the
dynamic behaviour of the prototype building as closely as possible. This includes care in
ensuring that if coupling effects in the prototype due to similar natural frequencies in the
two translational modes occur, then the model should also have similar natural frequencies
in these modes.

18
In the case of an aeroelastic model, care should be taken to ensure similarity between the
generalised masses of the model, and those of the prototype building.

Where the centre of rotation does not match with the geometric centre of the building, care
should be taken to mount the building model from the position representing the centre of
rotation of the prototype building.

Linear translation modes can be modelled by (a) base-pivot rigid models or (b) rigid models
mounted on a cantilever at one point.

In the case where an aeroelastic rig is used, providing a straight-line approximation to the
mode shape, the position of the pivot point should be such that the straight line matches as
closely as possible to the line of best fit to the mode shape of the first two modes (with
emphasis on the upper sections of the tower).

D6. Mode-shape corrections. Results from linear mode aeroelastic or high frequency
balance tests should be corrected for mode shape where the mode shapes depart by more
than 10% from the linear approximation. The effect of coupling between different axes in
a given mode of vibration should be accounted for.

There are various correction methods available for the above, e.g. Holmes (1987), Boggs
and Peterka (1989), Xu and Kwok (1993), Holmes et al. (2003).

In the simultaneous multi-channel pressure measurement technique (Clause D5.B), non-

linearity in the mode shape can be taken account of in the weighting of the pressures
recorded at each height level on the structure.

D7. Response combination. The simultaneous effects of wind loading and response in
two orthogonal directions, as well as torsional loading and response, shall be taken
account of in design.

Direct wind forces and resonant dynamic response generate effective loads in both along-
wind and cross-wind directions on tall buildings, as well as torsional moments. These act
simultaneously and should not be treated as separate independent load cases. If the
frequencies of sway in two orthogonal directions are well separated, then the responses in
the two directions will be poorly correlated statistically. One approach for the combination
of translational load effects, when there is little or no coupling, or correlation, between
orthogonal responses, is described by Holmes (2015), (Section 9.6). Several other methods
are in use including graphical approaches.

Full aeroelastic tests (Section D5.C) have the capability of reproducing any coupling
between orthogonal modes if the correct relationship between the respective frequencies
on the prototype structure is retained. The combination of simultaneous pressure
measurement techniques (Section D5.B) and advanced structural engineering design

19
software will increasingly allow this rigid wind tunnel model technique to be used to assess
coupling effects. Dynamic coupling effects are not reproduced with high-frequency
balance testing (Section D5.A). However, the frequencies of the constructed buildings are
often different from those estimated by the structural engineer at the design stage, and
coupling effects, which usually mitigate resonant dynamic response by transferring energy
between orthogonal modes, are unpredictable and difficult to confidently reproduce in any
wind tunnel model.

D8. Serviceability acceleration assessment. Serviceability accelerations shall be

assessed to ensure that risk of complaint resulting from wind-induced building motion is
kept to within a suitable criterion.

Complaints about wind-induced building motion may arise from occupants being alarmed
by the motion during extreme wind events, or being annoyed by frequently perceptible
motion (Denoon, 2000). The acceptability criteria for a particular project will depend on
the proposed use of the structure, whether it will be occupied during severe wind events,
the likely reaction of occupants to perceptible wind-induced motion, the wind climate and
the dynamic characteristics of the structure. Commonly used sets of guidelines are
ISO6897-1984, ISO10137:2007. The ISO6897-1984 guidelines were extended to peak
accelerations and different return periods by Melbourne & Cheung (1988) and these values
are presented in AWES (2011). Recent research such as that by Kwok et al. (2009) has
shown that the human response to building motion is frequency dependent. Criteria such
as those by the NBCC are considered limited, as they are independent of the natural
frequency of vibration of the structure, while the ISO6897-1984 and ISO10137:2007
guidelines are frequency dependent, a factor that has been found by many studies to be
important in the determination of motion perception and tolerance. Whichever guidelines
are used, it should be recognised that they are guidelines rather than strict criteria and some
flexibility in their application is acceptable as long as the implications for an altered risk
of complaint are recognised.

Torsional acceleration should be converted into an equivalent translational acceleration and

added into the combined acceleration with due consideration of the degree of correlation
between the accelerations along the three principal axes between vibration modes.

D9. Effective static load distribution. The base moments determined from an
aeroelastic or high-frequency balance study shall be distributed over the structure in a
rational and consistent way.

In general, the effective static wind loads arising from the mean, background (quasi-static
fluctuating), and resonant components of wind loading on a tall building, are distributed
differently from each other as a function of height, (e.g. Holmes and Kasperski, 1996;
Holmes, 2001). The background loading distribution depends on the type of load effect
(e.g. base shear or moment), and its location on the structure. However, if the resonant
component is dominant over the background component, it is acceptable to approximate

20
the combined fluctuating load distribution with an effective inertial loading as a function
of height. The latter is proportional to the mode shape of the structure. Conversely, if the
background response is dominant, the effective static wind loading may be taken as that
for the background base bending moment.

The distribution of the mean component with height based on the distribution of the mean
pressure coefficients with height along the windward and leeward faces for the most critical
wind direction. The distribution of the dynamic component with height based on either a
linear variation from the pivot point to roof level (adjusted to account for the effect of
variations in the width of the tower), or the inertial moment.

It is important to ensure that the loading distributions assumed produce values of mean and
fluctuating base bending moments equal to those measured on, or predicted from, the wind-
tunnel tests. The process of integrating the mean pressure loads from a simultaneous
pressure test to compare with the mean high-frequency balance loads, or aeroelastic test
results, as a function of wind direction is a useful quality assurance tool.

D10. Reporting. The assessment of wind-induced loads and vibrations for a

development should be presented in a way that simplifies the interpretation for the client.
In addition, the maximum total standard deviation and peak combined horizontal
accelerations should be presented and compared with the criteria.

The report should present a summary of results, which indicate the overall design base
moments, and corresponding accelerations for serviceability design. In addition, the
corresponding point load and shear force distribution with height should be provided. The
report should present as a minimum the assumed dynamic properties as well as the
directional mean and background standard deviation moment coefficients.

If the mode shape presents a significant departure from the straight-line mode shape, the
mode shape correction factor adopted should be discussed.

In discussing buildings accelerations, the maximum total standard deviation and peak
combined accelerations should be compared to acceleration acceptability guidelines as
described in Section D.7.

21
 PART E - OTHER STUDIES

Besides the Wind Engineering Studies outlined in the previous sections, there are other
types of studies that may be used to supplement those already discussed. Some of these are
briefly outlined below. This section is entirely informative and provides no normative
requirements as provided in Parts A – D. The reader is encouraged to seek further guidance
if undertaking any of these tests.

E1. Wind Tunnel Prototype Tests

These are normally performed at 1:1 scale and used to test building components for lift and
drag, discharge coefficients (effective area ratio), or wind noise. In the case of a wind noise
test the sample must be supplied by the client to ensure that the test accurately represents
the final product as method of connection, material, dynamic properties, precise gaps and
machining can influence results. For large models, where it is not practical to use a 1:1
scale model, 1:2: to 1:10 scale models may be used with suitable matching of modelling
parameters. However, the use of scale models should be carefully considered. Often the
fluid mechanics being investigated are not the same on geometrically scaled models
compared to full scale, so the observations obtained may not be valid. Careful consideration
should also be given to the wind tunnel blockage for prototype testing as this will
significantly influence the performance of the prototype.

It should be appreciated that when testing at full-scale, the turbulence structure created in
the wind-tunnel is markedly different in terms of frequency content from that experienced
in the real world and the impact of this on the results should be clearly stated.

E2. Wind Tunnel Section Model Tests

These are normally carried out at large scales of between 1:100 and 1:10. These tests are
performed to test aerodynamic performance of large industrial components, or to determine
wind loads on façade elements such as sunshade devices that cannot be tested at smaller
scales. Section model tests may also be required in cases involving ventilated double-skin
facades, although care should be taken to properly scale the internal volume for Helmholtz
resonance effects.

As with E1, the frequency content of the incident wind reproduced in the wind tunnel needs
to be documented and suitable appreciation of the impact of departures in the frequency
spectrum with respect to the full-scale should be clearly stated.

E3. Dispersion Studies

These are normally performed in the wind tunnel for point, line or area sources. Flow
visualization may be used to assist in the case of a preliminary qualitative assessment or to
provide general design guidance. However, flow visualization should not be used to
determine quantitative data for dispersion studies. A trace gas, coupled with measurements

22
of concentrations at various critical locations is required in the case where quantitative
results are required for comparison against exposure standards. Froude Number similarity
needs to be satisfied in these types of studies.

E4. Computational Fluid Dynamics (CFD)

Computational fluid dynamics (CFD) needs to be used with caution in external turbulent
flow environments (Franke et al. 2007, Blocken 2015). Current capability falls short in the
case where the modelling of local gust (or gust-equivalent means) wind speeds for the
assessment of pedestrian safety and comfort. The accuracy of CFD for this application has
not yet advanced to the point where it can be reliably used to check against accepted
comfort criteria. It should be noted that for a number of steady-state turbulence models, the
‘mean’ CFD output is an area averaged mean and does not model the transient nature of
turbulence in the flow, i.e. it smears the wind conditions in the shear layer. The
consequence of this is that a ‘mean’ CFD result would not necessarily agree with a time-
averaged mean from a physical model. Primarily, CFD has its application for modelling of
internal flow, buoyancy effects or air quality when coupled with wind tunnel data for the
boundary conditions.

CFD can be suitably benchmarked against the results of physical modelling. However,
extrapolation should be conducted with care. CFD can also be combined with the results
of physical modelling in a hybrid fashion. The two approaches both have strengths and
weaknesses in certain applications and one may be used to reinforce the other. For example,
the wind-tunnel data for a slender building with excessive accelerations may be used to
“calibrate” a CFD model. That computer model may then have architectural shape changes
explored many times via CFD until a viable solution is found. The new design should then
be confirmed in the wind tunnel.

As stated in Section A1, CFD is not yet capable of reliably determining peak design loads
on structures. However, notwithstanding the above discussion, any wind related CFD study
should conform with all components of the study as descried in this document in terms of
modelling the atmospheric turbulent boundary layer across the test domain, wind climate
analyses, and the appropriate number of wind directions for each particular study.

E5: Wind shear and Turbulence testing at Airports

Wind tunnel studies can be conducted to assess the impact of wind shear and turbulence
on aircraft operations. General background is provided in Nieuwpoort et al. (2010). For
airports in Australia, NASF Guideline B (NASF, 2018) defines the wind shear and
turbulence criteria and provides guidance for assessment, mitigation, and modelling.

In terms of local wind speed measurements, the principles outlined in Part A are applicable
for these wind tunnel studies. The length scale for testing should be as large as possible. It
is important to ensure the stability of the boundary layer wind shear and turbulence profiles
across the test domain are maintained. Measurement should be taken with suitable
anemometry equipment capable of measuring turbulent length scales equivalent of 1-2 s

23
full-scale. The test measurement locations and wind directions are outlined in Guideline B
and should be taken with and without the subject building for direct comparison.
Measurements outside the assessment zone in Guideline B should be considered for longer
runways. The results should be analysed in such a way as to be readily interpreted for
aircraft operations at the specific airport, i.e. a 3 s gust measured at the airport anemometer.

The wind tunnel test report should include:

 presentation of the mean wind speed and turbulence profiles, and spectral density
in the boundary layer upstream and downstream of the test section relative to the
target values, to illustrate the stability of the modelled wind field,
 description of the test procedure and instrumentation including calibration,
 data acquisition parameters (sampling frequency, sample duration, and any details
of filtering or signal conditioning),
 description of the subject building or structure, including photographs of the wind
tunnel model,
 diagram of the test point layout and wind directions tested, and
 tables showing 3 s gust wind speed required at the airport anemometer to exceed
the various criteria as described in Guideline B.

24
 REFERENCES

L.W. Apperley, and B.J. Vickery, The prediction and evaluation of the ground level wind
environment, Proceedings 5th Australasian Fluid Mechanics Conference, Christchurch,
New Zealand, 1974.

American Society of Civil Engineers, Wind tunnel studies of buildings and structures,
ASCE Manuals and Reports on Engineering Practice No.67, ASCE, 1999.

American Society of Civil Engineers, Wind tunnel testing for buildings and other
structures, ASCE/SEI 49-12, 2012.

American Society of Civil Engineers, Minimum design loads and associated criteria for
buildings and other structures, ASCE/SEI 7-16, 2016.

Australian Building Codes Board, National Construction Code,

https://www.abcb.gov.au/ncc-online/NCC.

AWES, Australasian Wind Engineering Society, Wind loading handbook for Australia &
New Zealand: Background to AS/NZS1170.2 Wind Actions, 2011.

AWES, Australasian Wind Engineering Society Guidelines for Pedestrian Wind Effects
Criteria, 2014.

R.M. Aynsley, W.H. Melbourne, and B.J. Vickery. Architectural Aerodynamics. Applied
Science Publishers, 1977.

J.B. Barlow, W.H. Rae, and A. Pope, Low-wind speed wind tunnel testing. John Wiley &
Sons, 1999.

D. Banks, T.K. Guha, and Fewless Y.J., A hybrid method of generating realistic full-scale
time series of wind loads from large-scale wind tunnel studies: Application to solar arrays,
14th International Conference on Wind Engineering, Porto Alegre, Brazil, 2015.

R.I Basu, Across-wind response of slender structures of circular cross-section to

atmospheric turbulence, Ph.D. Dissertation, University of Western Ontario, London,
Canada, 1983.

B. Blocken, Computational Fluid Dynamics for urban physics: Importance, scales,

possibilities, limitations, and ten tips and tricks towards accurate and reliable simulations,
Building and Environment, Vol.19, pp.219-245, 2015.

D.W. Boggs, and J.A Peterka, Aerodynamic model tests of tall buildings, Journal of
Engineering Mechanics, ASCE, Vol.115, pp.618-635, 1989.

25
N. Bottema, J.A. Leene, and J.A. Wisse. Towards forecasting of wind comfort. Journal of
Wind Engineering and Industrial Aerodynamics, Vol.44, pp.2365-2376, 1992.

J.E. Cermak, Application of fluid mechanics to wind engineering – A Freeman Scholar

lecture, Journal of Fluids Engineering, ASCE, Vol.97, pp.9-38, 1975.

J.E. Cermak, and L.S. Cochran. Physical Modelling of the Atmospheric Surface Layer.
Journal of Wind Engineering and Industrial Aerodynamics, Vol.41-44, pp.935-946, 1992.

J.C.K. Cheung, and W.H. Melbourne, Turbulence effects on some aerodynamic parameters
of a circular cylinder at supercritical Reynolds Numbers, Journal of Wind Engineering and
Industrial Aerodynamics, Vol.14, pp.399-410, 1983.

L. Cochran, and R. Derickson, A physical modeler’s view of Computational Wind

Engineering, Journal of Industrial Aerodynamics, Vol.99, pp.139-153, 2011.

N.J. Cook, Wind-tunnel simulation of the adiabatic atmospheric boundary-layer, Journal

of Industrial Aerodynamics, Vol.3, pp.157-176, 1978.

Council on Tall Buildings and Urban Habitat, Wind Engineering Working Group, Wind
Tunnel Testing of High-Rise Buildings, CTBUH Technical Guidelines, 2013.

W.A. Dalgleish, Assessment of wind loads for glazing design, Symposium on practical
experience with flow-induced vibrations, Karlsruhe, 1979.

A.G. Davenport, On the statistical prediction of structural performance in the wind

environment, ASCE National Structural Engineering Meeting, Baltimore, 1971.

A.G. Davenport, The prediction of risk under wind loading, Proceedings 2nd International
Conference on Structural Safety and Reliability, Munich, 1977.

D.M. Deaves, and R.I. Harris, A mathematical model of the structure of strong winds,
Construction Industry Research and Information Association Report 76, 1978.

R.O. Denoon, Designing for serviceability accelerations in tall buildings, Ph.D. thesis, The
University of Queensland, 2000.

F.H. Durgin, Instrumentation requirements for measuring aerodynamic pressures and

forces on buildings and structures, In Wind Tunnel Modeling for Civil Engineering
Applications, Cambridge University Press, 1982.

M. J. Eaddy, Lift Forces on Smooth and Rough Circular Cylinders in Low and High
Turbulence flows, PhD Thesis, Department of Mechanical Engineering, Monash
University, Clayton, Australia, 2004.

26
Engineering Sciences Data Unit, Characteristics of atmospheric turbulence near the
ground, Part 11: single point data for strong winds (neutral atmosphere), ESDU Data Item
85020, 1985 (revised 1993).

Engineering Sciences Data Unit, Characteristics of atmospheric turbulence near the

ground, Part 111: Variations in space and time for strong winds (neutral atmosphere),
ESDU Data Item 86010, 1986 (revised 1991).

J. Franke, A. Hellsten, H. Schlunzen, and B. Carissimo, Best practice guideline for the
CFD simulation of flows in the urban environment, COST Action 732, Quality assurance
and improvement of microscale meteorological models, 2007.

C.R. Halkyard, G. Blanchard, R.G.J. Flay, and N. Velychko, Digital filter adaptation for
tubing response correction at reduced sampling frequencies. Journal of Wind Engineering
and Industrial Aerodynamics, Vol.98, No.12, pp. 833-842, 2010.

J.D. Holmes, Reduction factors for wind direction for use in codes and standards, Colloque,
‘Construire avec le vent’, Nantes, France, 1981.

J.D. Holmes, Effect of frequency response on peak pressure measurements, Journal of

Wind Engineering and Industrial Aerodynamics, Vol.17, pp.1-9, 1984.

J.D. Holmes, Wind action on glass and Brown’s Integral, Engineering Structures, Vol.7,
pp.226-230, 1985.

J.D. Holmes, Mode shape corrections for dynamic response to wind, Engineering
Structures, Vol.9, pp.210-212, 1987.

J.D. Holmes, Directional effects on extreme wind loads, Civil Engineering Transactions,
Institution of Engineers, Australia, Vol.CE 32, pp.45-50, 1990.

J.D. Holmes, Methods of fluctuating pressure measurement in wind engineering, State-of-

the-Art Volume on Wind Engineering, Wiley Eastern Ltd. New Delhi, 1995.

J.D. Holmes, Wind loading of structures, 3rd Edition, CRC Press, Boca Raton USA, 2015.

J.D. Holmes, and D. Burton, Drag coefficients for rough circular cylinders revisited, 8th
International Colloquium on Bluff Body aerodynamics and Applications, Boston,
Massachusetts, USA, June 7-11, 2016.

J.D. Holmes, and J.Ginger, The gust wind speed duration in AS/NZS1170.2. Australian
Journal of structural Engineering, Vol. 13, pp. 207-218, 2012.

J.D. Holmes, and C. Osonphasop, Flow behind two-dimensional barriers on a roughened

ground plane, and applications for atmospheric boundary-layer modelling, Proceedings 8th.

27
Australasian Fluid Mechanics Conference, Newcastle, December 1983, pp.11B.13-16,
1983.

J.D. Holmes, A.W. Rofail, and L.J. Aurelius, High frequency base balance methodologies
for tall buildings with torsional and coupled resonant modes, Proceedings of 11th
International Conference on Wind Engineering, Lubbock, Texas, June 2003.

J.C.R. Hunt, E.C. Poulton, and J.C. Mumford, The effects of wind on people: new criteria
based on wind tunnel experiments, Building and Environment, Vol.2, pp.15-28, 1976.

International Organization for Standardization, Guidelines for the evaluation of the

response of occupants of fixed structures, especially buildings and off-shore structures, to
low-frequency horizontal motion (0,063 to 1 Hz) ISO 6897:1984, International
Organization for Standardization, Geneva, Switzerland, 1984.

H.P.A.H. Irwin, A simple omni-directional sensor for wind tunnel studies of pedestrian
level winds, Journal of Wind Engineering and Industrial Aerodynamics, Vol.7, pp.219-
239, 1981.

International Organisation for Standardisation, Wind Actions on Structures, ISO

4354:2009.

International Organisation for Standardisation, Bases for the design of structures –

Serviceability of buildings and walkways against vibrations, ISO10137:2007.

N. Isyumov, and A.G. Davenport, The ground level wind environment in built-up areas,
Proceedings of the 4th International Conference on Wind Effects on Buildings and
Structures, Cambridge University Press, U.K., pp.403-422, 1976.

W.D. Janssen, B. Blocken, and T. van Hooff, Pedestrian wind comfort around buildings:
comparison of wind comfort criteria based on whole-flow field data for a complex case
study, Building and Environment, Vol.59, pp.547-562, 2013.

H.H. Koss, On differences and similarities of applied wind comfort criteria, Journal of
Wind Engineering and Industrial Aerodynamics, Vol.94, pp.781-797, 2006.

K.C.S. Kwok, P.A. Hitchcock, and M.D. Burton, Perception of Vibration and occupant
comfort in wind-excited tall buildings, Journal of Wind Engineering and Industrial
Aerodynamics, 2009.

T.V. Lawson, and A.D. Penwarden, The effects of wind on people in the vicinity of
buildings, Proceedings of the 4th International Conference on Wind Effects on Buildings
and Structures, Cambridge University Press, U.K., pp.605-622, 1976.

28
C.W. Letchford, and J.D. Ginger, Wind environment assessment: A case study in the
Brisbane CBD, Proceedings 3rd Asia-Pacific Symposium on Wind Engineering, Hong
Kong, 1993.

C.W. Letchford, P. Sandri, M.L. Levitan, and K.C. Mehta, Frequency response
requirements for fluctuating wind pressure measurements, Journal of Wind Engineering
and Industrial Aerodynamics, Vol.40, pp.263-276, 1992.

J.R. Mayne, and N.J. Cook, Acquisition, analysis and application of wind loading data,
Proceedings 5th International Conference on Wind Engineering, Fort Collins, Pergamon
Press, 1979.

P.A. Macdonald, K.C.S. Kwok, and J.D. Holmes, Wind loads on circular storage bins, silos
and tanks: 1. Point pressure measurements on isolated structures, Journal of Wind
Engineering and Industrial Aerodynamics, Vol.31, pp.165-188, 1988.

W.H. Melbourne, Development of Natural Wind Models at Monash University,

Proceedings 6th Australasian Hydraulics and Fluid Mechanics Conference, Adelaide,
December, pp.190-194, 1977a.

W.H. Melbourne, Probability distributions associated with the wind loading of structures,
Civil Engineering Transactions, Institution of Engineers Australia, Vol.CE19, pp.58-67,
1977b.

W.H. Melbourne, Wind environment studies in Australia, Journal of Wind Engineering

and Industrial Aerodynamics, Vol. 3, pp.201-214, 1978a.

W.H. Melbourne, Criteria for environmental wind conditions, Journal of Wind

Engineering and Industrial Aerodynamics, Vol. 3, pp.241-249, 1978b.

W.H. Melbourne, Designing for directionality, Workshop on Wind Engineering and

Industrial Aerodynamics, Highett, Victoria, 1984.

W.H. Melbourne, and J.C.K. Cheung, Designing for serviceable accelerations in tall
buildings, Proceedings of the 4th International Conference on Tall Buildings, Hong Kong
and Shanghai, 1988.

NASF (National Airports Safeguarding Framework), Managing the risk of building

generated windshear and turbulence at airports, Guideline B Version 2.2.5, 2018

National Research Council of Canada, User’s guide – NBC 1995 Structural Commentaries
(Part 4), Ottawa, Canada, 1995.

New Zealand Building Code.

http://www.legislation.govt.nz/regulation/public/1992/0150/latest/DLM162576.html

29
New Zealand Standard NZS4223.4:2008 Glazing in buildings – Part 4: Wind, dead, snow,
and live actions.

A.M.H. Nieuwpoort, J.H.M. Gooden, and J.L. de Prins, Wind criteria due to obstacles at
and around airports, National Aerospace Laboratory, NLR-TP-2010-312, 2010.

A.D. Penwarden, and A.F.E. Wise. Wind environment around buildings. Building
Research Establishment (U.K.) Report, HMSO, 1975.

J.A. Peterka. Selection of local peak pressure coefficients for wind tunnel studies of
buildings. Journal of Wind Engineering and Industrial Aerodynamics, Vol. 13, pp 477-
488, 1983.

J.A. Peterka and S. Shahid, Design gust wind speeds in the United States. Journal of
Structural Engineering, Vol. 124, pp 2017-214.

Ratcliff, M.A. and Peterka, J.A., Comparison of Pedestrian Wind Acceptability Criteria,
Journal of Wind Engineering and Industrial Aerodynamics, Vol.36, pp.791-800, 1990.

T. Reinhold (ed). Wind tunnel modeling for civil engineering applications. Cambridge
University Press, 1982.

Standards Australia/Standards New Zealand, Structural design actions Part 0: General

Principles, Australian/ New Zealand Standard AS/NZS 1170.0, 2002.

Standards Australia/Standards New Zealand, Structural design actions Part 2: Wind

Actions, Australian/ New Zealand Standard AS/NZS 1170.2, 2011.

M.J. Soligo, P.A. Irwin, C.J. Williams, and G.D. Schuyler, A comprehensive assessment
of pedestrian comfort including thermal effects, Journal of Wind Engineering and
Industrial Aerodynamics, Vol. 77-78, pp. 753-766, 1998.

A. Steckley, Motion-induced wind forces on chimneys and tall buildings, Ph.D. thesis,
University of Western Ontario, Canada, 1989.

D. Surry, Consequences of distortions in the flow including mismatching scales and

intensities of turbulence, In Wind Tunnel Modeling for civil Engineering applications.
Cambridge University Press, 1982.

E. Szechenyi, Supercritical Reynolds Number simulation for two-dimensional flow over

circular cylinders, Journal of Fluid Mechanics, Vol.70, pp.529-542, 1975.

P.J. Vickery, and L.A. Twisdale, Wind-field and filling models for hurricane wind speed
predictions, Journal of the Structural Division, American Society of Civil Engineers,
Vol.121, pp.1700-1709, 1995.

30
Y.L. Xu, and K.C.S. Kwok, Mode shape corrections for wind tunnel tests of tall buildings,
Engineering Structures, Vol.15, pp.387-392, 1993.

31