Explanatory Notes Wind Effects 2019 e
Explanatory Notes Wind Effects 2019 e
on
2019
FOREWORD
The Explanatory Notes (EN) give a summary of background information and
considerations reviewed in the formulation of the Code of Practice on Wind
Effects in Hong Kong 2019 (Code), and should be read in conjunction with the
Code.
As the Code has been prepared in a simple format for ease of application, the EN
is set out to explain in depth the major updates and features in the Code and to
address on situations where application of the Code may require special attention.
Buildings Department
First Issue : September 2019
Page I
Contents
Page
FOREWORDS I
Contents II
1 General 1
1.1 Scope 1
1.2 Symbols 2
Page II
Appendix A
Appendix B
Appendix C
Appendix D
Page III
Appendix E
Appendix F
References
Page IV
1 General
1.1 Scope
The Code of Practice on Wind Effects in Hong Kong 2019 (Code) has
been written to widen the scope of guidance in the Code of Practice on
Wind Effects in Hong Kong 2004 (2004 Code) and to provide more
comprehensive guidelines for design of typical buildings in Hong
Kong.
Methods of calculation have been adjusted to be more in line with
other international standards, in particular the Australian/New Zealand
Standard AS/NZS 1170.2, the American ASCE 7-16 and the Eurocode
BS EN 1991-1-4. This enables shared use of future development
work on these standards, and more direct use of data from these
standards, where needed to supplement the Code. It may also assist
with future convergence. Additional work was also carried out to
resolve differences between these standards and to obtain better
loading information for the tall and relatively slender buildings of
Hong Kong.
Both ASCE 7-16 and AS/NZS 1170.2 directly derive ultimate limit
state design loads, and intend to correspond to a storm of about 2,000
year return for important buildings, which is also the intention,
together with the load factor, of BS and BS EN standards.
In anticipation of a possible future change in this direction, the wind
load factor in current use in Hong Kong is included in the
formulations. In principle the code reference pressures could be
multiplied by the existing γw of 1.4 to enable direct calculation of ULS
wind loads (with a future γw of 1.0) and this would result in no change
to the ULS wind loads calculated with the methods in the Code.
Reference wind pressures of the 2004 Code have been retained, but
adjustments have been made to avoid unnecessary systematic
increases to the existing design loads due to use of updated pressure
coefficients and to bring the resulting loads more in line with wind
tunnel testing. The new rules will better identify buildings that are
more vulnerable to wind damage, but loads on buildings of lesser
vulnerability may be reduced.
The use of wind tunnel testing is relatively common in Hong Kong,
but rules for carrying out the testing require improvement, particularly
to ensure consistent calibration against the reference wind pressures.
Improved rules are therefore given for testing of buildings and for
investigating the effects of topography. In addition, minimum
information requirements to allow independent assessment of the wind
tunnel measurements are suggested.
Page 1
An issue of significant difficulty is to establish a reliable threshold of
when wind tunnel testing is required, beyond a simple height criterion.
Properly calibrated wind tunnel testing is the basis of the code wind
loads but clearly the numbers have been simplified greatly for
codification, to mostly envelope the maxima of results measured in a
variety of situations. The situations tested for code purposes are
however not exhaustive and often only include relatively simple
rectangular block shapes. To make the code rules more useful, they
have been extended to provide methods to cover a number of shapes
that may be treated as rectangular. Similar ideas to those presented
can also be used judicially for slightly trapezoidal shapes, although the
critical wind directions may not be exactly orthogonal in this case.
Additional guidance or wind tunnel testing should be used for
irregular shapes where the rules cannot be clearly applied. Similarly,
‘complex topography’ is where the code rules cannot be clearly
applied, particularly for highly 3-dimensional forms of hill.
Where more detailed understanding of the wind loads or building
movements is desirable, wind tunnel testing is recommended.
A summary list of the more significant changes to the Code is given in
the Foreword of the Code. These changes and the reasons for them
are described in more detail in the following sections.
1.2 Symbols
Additional symbols and descriptions are needed, compared to the
2004 Code. The nomenclature has been developed following ideas on
notation introduced by NJ Cook for BS 6399-2, but broadly follows
previous practice. Where practical, upper case letters are used with
lower-case subscripts. This enables the nomenclature to be written in
an immediately intelligible form without need for subscripting, for
example in spreadsheets and internal reports, which offers
considerable time saving compared to some other standards. A
minimum number of subscripts is used, consistent with clarity.
As usual, it is proved to be insufficient Roman letters, and Greek ones
are also used, consistent where possible with existing wind codes.
Most of the nomenclature has been defined adequately without the
need for further commentary, but definition of ‘height’ of buildings in
areas of complex topography of Hong Kong, and where there are roof
features, may need more guidance.
The height, 𝐻𝐻 , of a building above ground, may vary because the
ground level is different on different sides of a building. It is
sufficient to define 𝐻𝐻 as height of the building above the average
ground level on each face in turn. For calculation of the reference
wind pressure, 𝑄𝑄ℎ , it is easier and not particularly conservative to use
only the greatest height. There may be more value in varying 𝐻𝐻 with
Page 2
direction when considering effects of surroundings or for calculation
of force coefficients.
The height, 𝐻𝐻𝑏𝑏 , used for calculation of building accelerations and
across-wind base moments may be similarly defined above ground
level but is intended to exclude, e.g. sloping roofs and irregular roof
plant, forming a small part of the total height and which do not
continue the prismatic form of the building below. This relates to the
aerodynamic behaviour associated with dynamic across-wind forces,
which is strong for prismatic forms but weakens greatly for non-
prismatic forms.
Page 3
2 Calculation of Wind Actions
Page 4
2.2.2 Torsional Forces
Torsions arise from static gust pressures, particularly due to skew
wind directions or effect of surroundings, and from dynamic effects,
when the masses over upper levels are offset on plan.
More recently updated codes of practice such as ASCE 7-16,
BS 6399-2, AS/NZS 1170.2 and BS EN 1991-1-4 do give guidance
and this was considered in deriving a method for the Code.
Several codes provide torsional loads by offsetting lateral loads by a
constant eccentricity. However the origin of these values is obscure.
Studies of wind tunnel test data, carried out as part of this Code
development, show a trend of increased torsion with building
elongation, associated with a trapped vortex behind the windward end
when the wind is diagonal. Therefore, the eccentricity for torsion is
defined to increase linearly with the plan ratio, 𝐵𝐵/𝐷𝐷 in the Code, from
a minimum of ±0.05𝐵𝐵, for building which is square on plan to ±0.2𝐵𝐵
when 𝐵𝐵/𝐷𝐷 is 6.0. Since there is limited information available for
buildings with 𝐵𝐵/𝐷𝐷 ratio larger than 6.0, data from wind tunnel testing
is required in this condition.
In some codes of practice, the torsion is applied simultaneously with
the full lateral loads. In other cases, additional load combination
factors are used to account for non-simultaneous occurrence of the
maximum lateral and torsional loads. The latter approach is adopted
in the Code as described in the load combination section below.
Page 5
tunnel testing. It was originally calibrated from a large series of tests
of building with alternative surroundings. It does not necessarily
capture the full effects of enhanced responses that sometimes occur
due to buffeting by other buildings, but other codified methods do not
provide more reliable prediction.
As the across-wind base moment calculated using the NBCC
methodology is proportional to wind speed to the power of 3.3 rather
than 2.0 for wind pressure, the ultimate across-wind base moment is
directly calculated from the ultimate mean wind speed at the top of the
�2𝛾𝛾𝑤𝑤 𝑄𝑄ℎ ⁄𝜌𝜌𝑎𝑎
building given by and the resulting ultimate across-wind
1+3.7𝐼𝐼𝑣𝑣,ℎ
base moment is then divided by 𝛾𝛾𝑤𝑤 so that the values can be used by
the relevant structural design codes in the usual way.
The NBCC model is calibrated for:
(a) a rectangular prismatic form,
(b) mode deflections that are approximately linear with height,
(c) relatively uniform building mass with height over the top half of
the building, and
(d) 0.5 < 𝐵𝐵/𝐷𝐷 < 2.
Based on more recent experience of wind tunnel testing, the use of the
NBCC rule beyond the limit 0.5 < 𝐵𝐵/𝐷𝐷 < 2 is likely to be of similar
reliability up to about 1:4 (or 4:1) of the B/D ratio. Beyond this limit,
wind tunnel testing may be appropriate depending also on the
significance of the torsion loads which become more significant for
larger plan ratios.
In wind tunnel testing of many prismatic forms, the NBCC formula
has been found to be a reasonable predictor but it is conservative for
significantly tapered or stepped buildings and buildings with irregular
plan shapes. Changes to the width and plan shape result in less
regular vortex shedding over a range of frequencies, which reduces the
excitation at the frequency of the structure. This is also reflected in
the published rules for chimneys.
Where the across-wind base moment is larger than the along-wind
base moment in the same loading direction, the along-wind force is
scaled up to achieve the same base moment as that from the across-
wind. However, if the scaling up factor is very significant for design,
or larger than 50%, then it is necessary to accurately evaluate the
across-wind responses and therefore wind tunnel testing is required.
The across-wind loads are ignored if the base moment is less than the
along-wind moment in the same loading direction.
Page 6
A specific working example is provided as below, where the
overturning moment (unit in NMm) due to different wind directions as
shown in the figure below are:
(a) WX1+: along, 190; across, 431
(b) WX1-: along, 161; across, 141
(c) WX2+: along, 340; across, 271
(d) WX2-: along, 280; across, 207
Page 7
wind tunnel testing. Otherwise the code calculation should be
followed.
For short, stiff and less slender buildings, the across-wind loads are
smaller than the along-wind loads and the requirement for checking
the across-wind base moment can be neglected.
According to a parametric study with selected parameters, assuming
natural periods of 𝐻𝐻/46 in both directions, for buildings which satisfy
𝐻𝐻/𝑚𝑚𝑚𝑚𝑚𝑚(𝐵𝐵, 𝐷𝐷) < 5, 𝐻𝐻 < 100m and 𝑁𝑁 > 0.5Hz, the along-wind base
moment is always larger than the across-wind base moment in the
same direction, even with consideration of some level of uncertainty
on the period estimation. Therefore checking of the across-wind base
moment is not required when these conditions are met.
For rectangular forms of building, 𝐹𝐹𝑥𝑥 , 𝐹𝐹𝑦𝑦 are the maximum wind
forces (the maximum of along-wind and across-wind force calculated
in the Code) in the lateral directions; 𝑀𝑀𝑧𝑧 is the maximum peak torsion
calculated in the Code, assumed to be about the centre of area.
The 24 load combinations are reasonable for providing a
3-dimensional space containing all the possible combinations of 𝐹𝐹𝑥𝑥 , 𝐹𝐹𝑦𝑦
and 𝑀𝑀𝑧𝑧 .
Page 8
The chosen load combination factors fall close to enveloping wind
tunnel testing load cases. The simplified 2-dimensional plot below
was obtained by abstracting load combination factors Fx and Fy from
various projects from various wind tunnel laboratories, including The
Boundary Layer Wind Tunnel Laboratory, CPP Inc., The Hong Kong
University of Science and Technology, Rowan Williams Davies &
Irwin Inc. and RED Consultants Limited.
The load combination factors of the Code were compared with those
of ASCE 7-16, BS 6399-2, EN 1991-1-4 and AS/NZS 1170.2. The
factors recommended in the Code are slightly conservative compared
to ASCE 7-16 but are comparable to the methods recommended in the
other standards, and to the commentary to AS/NZS 1170.2 by John
Holmes.
Many common building forms have good torsional resistance for
which the application of torsional loads may have small effect on the
internal member forces. A rule is provided in the Code to identify
such buildings so that torsional loads may be neglected to simplify the
design process. These include:
(a) buildings of single-storey up to 10 meters high. ASCE 7-16 has
similar requirements in which the height limit is 9.1 meters (30
feet).
(b) buildings up to 70 meters with peripheral lateral load resisting
construction, e.g. masonry or concrete shear walls or braced
steelwork on all faces. Bi-directional multi-bay moment frame
structures can also be considered under this category; but in case
of column set-backs, the set-back distance (measured from the
Page 9
column centre) shall not exceed 5m or 1/10th of the building
dimension in the direction of the set-back, whichever is smaller.
(c) any building which passes the ‘torsional regularity check’, which
is similar to ASCE 7-16.
As the ‘torsional regularity check’ of ASCE 7-16 is intended for
relatively low-rise buildings where the majority of the wind deflection
is caused by shear loads rather than overturning bending strains, the
rule for tall buildings should be reinterpreted as a comparison of shear
strains caused by torsion and shear strains caused by lateral shear
forces.
Although not described in the Code, the loadcases may also be
simplified conservatively for buildings which are insensitive to
torsional loads as defined in Section D.6 of ASCE 7-16.
The full neglect of torsional loads will reduce the total number of
loadcases from 24 into 8.
A further rule related to the torsional regularity check is also provided
where torsion is only a moderate effect. This limits the number of
load combinations to 16, by omitting the primarily torsion set of loads,
case 3 in Table 2-1 of the Code.
The neglect of the torsional loads in the final load combination
discussed in this section should not be taken to imply that torsional
wind loading does not exist for buildings within the limit.
The user of the Code can always reduce the number of design load
combinations by conservatively taking the full load rather than using
the reduction factors. For example, by always taking the full torsion
(combination factor of 1.0), the number of loadcases can be reduced
from 24 into 16; by always taking the full loads for all the three
components, the number of loadcases can be reduced from 24 to 8.
This is helpful when the design implication of wind loads is small.
Page 10
wind tunnel results. The reason is that above the displacement height,
negative pressures in the separated flow regions on the side and rear
faces of the building are quite uniform. Positive pressures also vary
much less than implied by the oncoming wind profile because the
wind pressures are deflected downwards by the building. A procedure
separating external and internal pressure calculations is given in
Appendix B1 to assist with non-standard cases.
A size effect factor, 𝑆𝑆𝑠𝑠 , is adopted, to take account of the fact that the
external area-average peak pressure is related to the size of the loaded
area and its location. This effect is also reflected in BS EN 1991-1-4
by the difference between 𝐶𝐶𝑝𝑝𝑝𝑝,1 and 𝐶𝐶𝑝𝑝𝑝𝑝,10 , in AS/NZS 1170.2 by the
area reduction factor 𝐾𝐾𝑎𝑎 and local pressure factor 𝐾𝐾𝑙𝑙 (the latter
depending on the location of the loaded area) and in ASCE 7-16 by
the various external pressure coefficients, 𝐺𝐺𝐶𝐶𝑝𝑝 , which also depend on
the effective loaded area and location.
In the Code, for calculating the wind forces on cladding of an enclosed
building, the size effect factor is applied to the net pressure. This is
conservative when the size factor, 𝑆𝑆𝑠𝑠 , is greater than one because the
adjustment is applied to the internal pressure as well as the external
pressure.
The size factor is also applicable for calculating the net wind loads on
building attachments, free-standing walls and open frameworks.
2.4.1 Acceleration
The Code provides a method for assessing building motions by
adopting the NBCC method for estimating across-wind acceleration.
The formula was developed by AG Davenport and his team based on
measurement from boundary layer wind tunnel studies in varying
surroundings. While there was significant scatter, it was found
reasonable to fit a representative line through the data after
normalising the responses against a ‘reduced velocity’ parameter. The
Page 11
predictions of the NBCC method have also been compared with those
from more recent wind tunnel tests. In cases where it is applicable,
the predictions have proved to be good indicators of trends and likely
across-wind sensitivity for acceleration.
In the Code, the NBCC formula has been rearranged to use consistent
units and nomenclature consistent with the Code.
People are most likely to feel motion in slender buildings, in which
the across-wind acceleration is more critical than the along-wind
acceleration. However, as discussed in NBCC 2010, if (𝐵𝐵𝐷𝐷)𝑏𝑏 >
𝐻𝐻 2 ⁄9, the along-wind acceleration could be larger than the across-
wind acceleration. In the Code, a limitation is set so that (𝐵𝐵𝐷𝐷)𝑏𝑏 ≤
𝐻𝐻 2 ⁄9. This reduces the risk that the predicted accelerations will fall
below that expected for along-wind movements, although human
comfort is normally not a problem for a building aspect ratio in this
range.
For other background on the across-wind calculation method,
reference should be made to clause 2.2.
Page 12
1 year may not be 3.67. For very slender buildings, where the critical
windspeed ( 𝑉𝑉𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 ≈ 10𝑁𝑁𝑦𝑦 for prismatic cross section) for vortex
shedding is less than that for 10 year return, the human comfort should
additionally be checked at the critical windspeed the criteria for which
can be determined through Figure 2-6 of the Code by interpolation.
Page 13
3 Design Wind Pressures
3.1 General
Wind pressures of the Code are numerically closely similar to the
tabulated gust pressure in the 2004 Code. For ease of calculation, a
simplified power-law formula as shown in Equation 3-2 of the Code
has been fitted. The turbulence intensity formula (Equation 3-3 of the
Code) is the same as that in the 2004 Code. Together, these
correspond to open-sea exposure during a storm with a mean speed of
59.5 m/s at reference height of 500m.
The pressures are intended to be used with the 1.4 wind load factor of
current structural design codes and the direction factors of the Code.
When used in this way, the ULS wind loads have been calibrated to be
of at least 1,000-1,500 year return which is consistent with the level of
reliability of significant buildings implied in other international
standards and previous practice in Hong Kong.
The possible effects of climate change have been considered, while
the detection of long term trend of tropical cyclone activities in
western North Pacific based on existing data has been subject to
uncertainties due to significant natural decadal variations and the
discrepancies between various historical datasets in the basin,
available theoretical studies and climate model simulations from
different research groups suggest that tropical cyclone intensity
(maximum wind speed) will likely increase as the climate warms in
response to increase in human induced greenhouse gas emission.
Because of this, although long-term data from Waglan Island shows
windspeeds which are lower than from current Monte-Carlo extreme
wind statistics, it has been decided to retain the current level of
conservatism of the derived loads pending improved understanding.
Wind pressures are based on a gust wind speed with a peak factor of
3.7 times the standard deviation of wind turbulence plus the hourly-
mean speed, as in the 2004 Code. The latest study shows that 3.7
peak factor actually corresponds to about 0.35s-average. However,
the discerptions do not affect the calculation of structural or cladding
wind loads.
The changed description of the peak factor does not directly affect the
pressures of the Code. For example the implied mean and fluctuating
pressures remain the same as that in the 2004 Code. However, there
are implications for calculation of size factors and for the
measurement of wind pressures in wind tunnel testing.
Although the open sea turbulence intensity formula of the 2004 Code
is kept, turbulence is increased when a building is largely sheltered by
surrounding buildings. A modification (Equation 3-4 of the Code),
Page 14
depending on effective height (see below), is therefore provided for
use with the across-wind response calculation.
Page 15
3.4 Topography Effects
Large-scale topography is an obvious characteristic in the landscape of
Hong Kong, and topographic effects are likely to dominate over
terrain effects in the determination of wind loads for Hong Kong
buildings.
Topography factors in the Code are expressed in equations that can be
easily formulated in a spreadsheet. The formulation is based on the
slightly updated EN 1991-1-4 rules but has been changed to make it
easier to use in the complex topography of Hong Kong. These
changes are discussed in more detail in Appendix A3.
Page 16
4 Force and Pressure Coefficients
4.1 General
The force and pressure coefficients of the 2004 Code have been
updated and extended to take account of developing consensus in
international standards and availability of values with improved
background of quality and consistency. The original values in the
2004 Code are similar to those in CP3 (1972). The relevant building
geometries were extensively re-measured by the Building Research
Establishment (BRE) and published in BS 6399-2 (1995) and most of
them have also been adopted with some minor adjustments by the BS
EN.
A hidden weakness of the studies behind these coefficients is that they
were mostly carried out for low-rise buildings. The most serious
deficiency identified was for building with a large slenderness. The
BS EN values were limited to 𝐻𝐻/𝐷𝐷 = 5. While many of the local
pressure values in the BS EN data remain usable, it was desirable to
find another source of overall building force coefficients for buildings
of higher slenderness as described below.
Page 17
The ESDU data item also described effects of turbulence which were
apparently contradictory, sometimes increasing and sometimes
reducing the coefficients. This is a result of the way the turbulence
affects reattachment of the flow. Data from a limited number of
isolated buildings tested in shear flow was used to calibrate an
adjustment of the effective plan ratio in Equation 4-1 of the Code.
The resulting coefficients provide the correct trends of the data with
increasing slenderness and effect of plan ratio.
4.2.4 Buildings with Wings (e.g. U-, X-, Y-, Z- and L-shaped)
There is some wind tunnel data on non-rectangular shaped buildings,
but again there is no comprehensive data set which can be used for full
calibration for each particular irregular configuration listed in the
Code. However, the largest wind loads for buildings with
(approximately) rectangular corners are associated with simple
rectangular forms. Therefore the rules in the Code, which are
intended to be helpful but slightly conservative, are based on the
rectangular data.
For shapes not covered in the Code, similar rules may be applied as
below:
(a) The principle of the enclosing rectangle may be used to assess
the effective breadth, 𝐵𝐵, for any wind direction.
(b) Re-entrant surfaces should be ignored for buildings with
𝐻𝐻/𝐵𝐵 ≥ 1. Assume a straight line (as the blue line in the figures
below) on plan between extremes of each cross-section.
(c) Clause 4.2.3 may be used to assess the effect of windward faces
not normal to the wind.
(d) The effective depth, 𝐷𝐷 , should be based on the minimum
dimension in the direction of the wind, near the extremes of the
Page 18
cross-sectional breadth. Where surfaces are not parallel to the
wind, 𝐷𝐷 should be taken as the depth of the enclosed rectangle
or 𝐵𝐵/1.8, whichever is the smaller in determining the rough B/D
ratio where the force coefficients in Figure 4-2 reach their peaks.
(e) In general it is necessary to calculate the above for at least four
wind directions, each approximately at right-angles.
Selected examples to illustrate the rule 𝐷𝐷 = 𝑚𝑚𝑚𝑚𝑚𝑚(𝐷𝐷𝐸𝐸𝐸𝐸𝑐𝑐𝑙𝑙𝐸𝐸𝑠𝑠𝑝𝑝𝐸𝐸−𝑅𝑅𝑝𝑝𝑐𝑐𝑐𝑐 ,
𝐵𝐵/1.8) are given below:
Page 19
4.3 Pressure Coefficients for Building Elements
Page 20
Below the displacement height there is typically some reduction in the
peak pressures. A reduction of 20% is given for heights less than half
of the displacement height.
If supplementary information on external pressures is obtained from
international standards, e.g. for low-rise buildings, then the methods
given in Appendix B of the Code should be used to include an
appropriate internal pressure. The net pressure on building envelope
is the sum of the pressures acting on external and internal faces.
Where separate internal and external pressures are calculated, these
should be chosen to give the most critical net positive and net negative
(suction) pressure on the envelope.
Difference of internal pressures between two nearby rooms will lead
to wind loads on the internal walls. For buildings without dominant
openings, BS 6399-2 gives 0.5 as the maximum of net pressure
coefficient of internal walls; AS/NZS takes 0.4 as the worst net
pressure coefficient for internal walls which form a permanent seal
and 0.3 for internal walls which do not form a permanent seal. In this
Code, the Registered Structural Engineer/Authorized Person shall
make a judgement when an internal wall shall be designed and what
the appropriate value for the net pressure coefficient is.
In case of accidental dominant openings, the net pressure coefficient
on internal walls will be larger. This should be considered for the
internal walls between different occupied units. However, the
accidental dominant opening scenario is out of the scope of this Code,
as discussed in Appendix B1.3.
Page 21
4.3.3 Pressure Coefficient for Open Frameworks
Pressure coefficients for open planar frameworks in Table 4-2 of the
Code have been updated, assuming sharp section members using data
in ‘The designer’s guide to wind loading of building structures: Part 2:
Static Structures’, NJ Cook (the Guide). For more complex open
frameworks, force coefficients from lattice mast and tower standards
could be used. There is also further useful guidance in the Guide.
Page 22
5 Size Factor and Size and Dynamic Factor
Page 23
except for the Australian/New Zealand code 1, also apply a constant
factor with height. But this is applied to the gust-pressures which are
more uniform over height than the mean pressures. This results in
significantly non-conservative pressures towards the upper levels of
buildings, as compared to both the original Davenport factor on mean
pressures and also to wind tunnel testing, but also increased wind
shear at base level. This becomes more significant with code based
design of taller buildings.
The Code overcomes this through a size and dynamic factor, 𝑆𝑆𝑞𝑞,𝑧𝑧 , on
the peak gust pressures which varies linearly with height as shown in
Equation 5-2. This directly results in an appropriate shear and
moment at base level and a reasonable distribution of loading with
height.
The theory used for calculation of the responses is a first principle
approach based on the ESDU wind engineering data items, with the
results fitted by an empirical formula assuming a linear variation of
the size and dynamic factor with height. Alternative forms of
variation with height did not significantly increase the quality of the fit.
The factor was derived for standard exposure of the Code but gave
good results also for different cases of displacement height.
The formulation is nominally based on calculation of the appropriate
gust factor at the top of the building, 𝑆𝑆𝑞𝑞,ℎ , with a reduced factor of less
than 1.0 near the ground. The expression for 𝑆𝑆𝑞𝑞,ℎ follows the usual
practice with terms for mean, quasi-static (‘background’) and resonant
components. The results have been found to be relatively insensitive
to the proportion of mean loads and the mean has therefore been taken
as a constant 0.5 times the peak pressure. The size factor of the Code
is used to assess the maximum equivalent quasi-static pressure at the
top. The last term in the equation is a simplified expression for the
resonant response calibrated at the Code wind speeds.
The theory also assumed that the mass distribution with height is
approximately uniform in the upper levels of the tower. Where the
mass distribution is clearly not uniform, there may be significant
differences of shear and moment distribution compared to wind tunnel
testing, but generally only in the upper parts of the building.
The estimation of dynamic responses depends on good estimation of
natural frequency but this becomes less critical for shorter buildings.
In order to avoid calculating natural frequencies of shorter buildings,
which are often of complex form in Hong Kong, a simplified constant
𝑆𝑆𝑞𝑞 factor is provided in Equation 5-3 of the Code for buildings shorter
than 50m.
1
AS/NZS 1170.2 has an overturning moment factor which increases with height. This has to be
applied to internal forces post-analysis however.
Page 24
6 Requirements for Wind Tunnel Testing
Page 25
period of vibration. The cladding pressure should typically be
measured at density of not less than 120 m2 surface area per tap.
Page 26
discrepancies between target and achieved wind conditions. Generally
the turbulence intensities averaged within 15% of the reference height
should be within 10% of the target value for all important wind
directions and the turbulence length-scales should be within a factor of
two for structural loading studies.
A factor of 3 may be acceptable for cladding pressure studies with size
of cladding panels not larger than 15m, and this may be advantageous
where the larger scale of model enables the surface detail to be better
reproduced and better instrumented.
It should be noted that practical limitations of wind tunnel testing
typically result in turbulence length-scales which are shorter than in
the atmosphere.
Page 27
6.3 Additional Requirements for Cladding
There are no fully agreed procedures for considering the influence of
cladding panel sizes on the peak cladding pressure due to the
difficulties of evaluating area averaged pressures at model scales of
typically 1:400 scale. Typically pressure taps are located with the
equivalent of several metres apart at this scale.
Although the relationship between size and time/frequency filtering
methods is subject to ongoing research, the current practice of
applying a 0.5-1s time average filter to obtain pressures relevant for
typical cladding panels remains appropriate. For large loaded areas,
area- or time-averaging may be used under guidance of experienced
wind engineers.
Page 28
In all cases 100% of the loads with existing and likely future
surroundings should also be considered for design.
6.5.2 Directionality
Although typhoon wind come from all wind directions, the direction
of travel of the storms and the effects of landfall results in reduced
wind speeds for some directions, particularly north-west.
The Code rule on direction factors is intended only for typhoon winds.
For calculation of building movements under frequent winds, data
from, e.g. Waglan Island may be used, taking account of topographic
steering, e.g. through wind tunnel topography testing. Alternatively
the typhoon direction factors may be used.
If Monte-Carlo typhoon models are used in Up-crossing or Storm
Passage methods to determine loads, the ULS wind loads should be
directly calculated by targeting the return period corresponding to the
wind pressures of the Code multiplied by 𝛾𝛾𝑤𝑤 , with 𝑆𝑆𝜃𝜃 = 1.0. The loads
can then be obtained by dividing by 𝛾𝛾𝑤𝑤 as described in clause 6.5.1 of
the Code. For residential comfort, the acceleration should be
calculated by targeting the return period corresponding to the codified
wind speed multiplied by 𝑆𝑆𝑐𝑐 , where 𝑆𝑆𝑐𝑐 is defined in Table A1-2 of the
Code, with 𝑆𝑆𝜃𝜃 = 1.0.
Page 29
6.5.3 Wind Pressure for Acceleration Calculation
The reduction factors are based on analysis of Waglan island data
compared to the code wind pressures. These include both typhoon
and non-typhoon winds, although the typhoon wind strengths are
dominant in Hong Kong for winds of longer than about two-year
return.
Page 30
number of critical loadcases applied can be reduced by inspection,
using knowledge of the structural behaviour. A greater number of
loadcases could be used to define the surface of the ellipsoid more
precisely.
To be conservative, the planes (or lines) drawn between the vertices
represented by the loadcases should only just touch the surfaces of the
ellipsoid. This requires that the load combination themselves will lie
outside the surface of the ellipsoid. This has not always been clear in
past codified practice.
Because of the generality of the responses, it is not easy to summarise
or demonstrate the sufficiency of the chosen load combinations
through simple calculation of results using existing codified methods.
Currently the most appropriate method is to demonstrate this with a
3-dimensional plot of the chosen load combinations compared to an
envelope (for all wind directions) of the response ellipsoids. In
2-dimensional this can be demonstrated by plotting envelopes of the
base moments in pairs (e.g. 𝑀𝑀𝑥𝑥𝑥𝑥 v. 𝑀𝑀𝑦𝑦𝑦𝑦 , 𝑀𝑀𝑦𝑦𝑦𝑦 v. 𝑀𝑀𝑧𝑧𝑧𝑧 ,and 𝑀𝑀𝑧𝑧𝑧𝑧 v. 𝑀𝑀𝑥𝑥𝑥𝑥 ).
A satisfactory plot would have a series of loadcases with
surfaces/lines between them which completely enclose the response
envelopes.
Page 31
Where Up-crossing or Storm Passage methods are used to determine
critical loads in particular directions, combination factors should be
demonstrated using diagrams plotted taking 𝑆𝑆𝜃𝜃 = 1.0. Otherwise the
directional 𝑆𝑆𝜃𝜃 values should be used.
Page 32
Appendix A
Supplementary Information for
Section 3: Design Wind
Pressures
A1 Wind Climate
A1.1 Wind Directionality
Some sites are sheltered from the strongest typhoon winds by
topography or other buildings. The variation below 0.85 is potentially
useful in this case. The directional factors were obtained from
Monte-Carlo (open-sea) analyses since these were more conservative
than the Waglan Island data. For directional assessment, the latter is
affected by regional topography, and is difficult to correct without
research beyond the scope of work for this revision to the 2004 Code.
This might be usefully considered as something for future research
and revision.
The reduction of the maximum value to 0.85 compensates for two
factors:
(a) Review on the methods, and pressure and force coefficients of
the 2004 Code indicated smaller factors to be applied to the code
reference pressure to obtain design loads compared to
international standards. In correcting the methods, it was
considered undesirable (and found to be unnecessary) to
systematically increase the design loads.
(b) ‘Up-crossing’ or similar methods have been widely used with
wind tunnel testing in Hong Kong and tend to result in a
reduction in the peak wind loads. For example ASCE 7-16
suggests an omni-directional reduction factor 𝐾𝐾𝐸𝐸 of 0.85 for
buildings where standard code methods (not Up-crossing) are
used.
However there may be increases in loading for some more vulnerable
buildings, which are better identified in the Code, and reductions for
others.
New extreme value analyses of extreme winds from Waglan Island
indicate that the necessary 1,000-1,500 year return ULS wind loads
(as used internationally for significant buildings) are still achieved.
This extreme analysis is valid because the strongest winds come from
directions where there is no land to change the winds. The data was
corrected for speed-up effects over Waglan Island itself.
Page A1
A2 Exposure Adjustment for Direct Shelter
The direct shelter effect of surrounding buildings is now considered
using the concepts of displacement height and a reduced effective
height. The principles behind the displacement height effect are well
established in codified form in BS 6399-2 and BS EN 1991-1-4.
However the complexity of surroundings in Hong Kong has led to a
more detailed description of how to treat surroundings on non-uniform
height.
Changes to surroundings over the lifetime of a building are normally
inevitable in any built environment. While these often provide
additional shelter, increases in loading are also possible, and may be
temporary or permanent. Various restrictions are therefore made
within the Code to minimise significant changes to loading when
future development occurs.
In particular, the benefit due to the most significant sheltering building
is not taken into account. This makes allowance for possible removal
or change of the sheltering building. The probability of simultaneous
removal of two or more sheltering buildings in one approaching wind
direction is typically smaller, so benefit is taken from the second most
sheltering building.
Buildings in the same building lot shall be treated as a whole, when
evaluating the direct sheltering effect for the determination of the most
beneficial building/buildings to be removed.
Where the direct sheltering effect is calculated as shown in
Figure A2-4 of the Code, the most beneficial building which is
required to be removed in one division may be retained in another
division if it is no longer the most beneficial building.
The obstructing building height should be calculated following the
rule in the two figures below.
Page A2
Scenario for downwind slope
For the Standard Method, the user can decide the extent of surrounding
buildings around the site to be included for sheltering effect evaluation.
Generally, the influence from buildings beyond 500m or six times of the
proposed building height is negligible.
Working Example 1: Rectangular building, H=96m, nine equal
divisions
1. Consider direct sheltering effects from surroundings within a 500m
(approximately 5H) -radius circle from the proposed building (Figure
A-1).
2. Divide into 4 sectors of ±45 degrees of the considered wind direction.
3. Each sector is divided into nine equal divisions (10 degree interval) for
easy calculation (S1-S9). Normally smaller division shall give fewer
weighted average calculation because it makes it easier for a building to
cover the whole division.
4. Further divide the whole circle into five layers with radius of 100m,
200m, 300m, and 400m for easy building distance identification.
Page A3
Figure A-1 Site information, four sectors and 9 divisions within each sector
5. Demonstration for calculating HdS1 of S1(Figure A-2) in WX2-
direction
Page A4
45m. It covers the whole area of the division. With, X1/H ≈ 0.73, H1/H ≈
0.47, Hd1/H = 0.38 following Figure A2-3 in the Code.
(b) The next significant sheltering effect is from two nearby towers in the
same building lot. They are considered as a whole with windward walls
intersecting the boundary line at Points C and D. With X2=200m,
H2=165m>H=96m, X2/H2 ≈ 2.08, H2 /H = 1, so Hd2/H = 0.75.
(c) The buildings further away are not critical in this case as they are either
too far away from the site or too low to make significant sheltering
effects.
(d) Thus the second largest Hd in S1 division is HdS1/H = 0.38.
6. Demonstration for calculating HdS9 of S9 in WX2+ direction (Figure A-3)
α1
Page A5
7. Following the procedures above, the height of reduction in each division
can be calculated. The height of reduction in each of the four sectors can then be
���
𝐻𝐻𝑑𝑑� ��
𝐻𝐻����� �������
𝑑𝑑𝑑𝑑1 +𝐻𝐻 �������
𝑑𝑑𝑑𝑑2 +⋯+𝐻𝐻
calculated using = 𝑑𝑑𝑑𝑑9
. The results are summarised in the table
𝐻𝐻 9
below:
Table A-1 Summary of HdSi in each equal division (nine divisions)
S1 S2 S3 S4 S5 S6 S7 S8 S9 � 𝒅𝒅 /𝐇𝐇
𝐇𝐇
WX1+ 0.37 0.22 0.18 0.13 0.29 0.30 0.36 0.29 0.40 0.28
WX1- 0.16 0.18 0.20 0.20 0.21 0.09 0 0 0 0.12
WX2+ 0 0 0.11 0.38 0.18 0.38 0.38 0.12 0.39 0.21
WX2- 0.38 0.38 0.22 0.22 0.15 0 0 0.13 0.21 0.19
With 𝑍𝑍𝑝𝑝 = max(𝑍𝑍 − 𝐻𝐻𝐸𝐸 , 0.25𝑍𝑍), the effective building height 𝐻𝐻𝑝𝑝 is:
Table A-2 The effective building height
WX1+ 𝑯𝑯𝒆𝒆 = 𝟎𝟎. 𝟕𝟕𝟕𝟕𝑯𝑯 = 𝟔𝟔𝟔𝟔. 𝟏𝟏𝟏𝟏
WX1- 𝐻𝐻𝑝𝑝 = 0.88𝐻𝐻 = 84.5𝑚𝑚
WX2+ 𝐻𝐻𝑝𝑝 = 0.79𝐻𝐻 = 75.8𝑚𝑚
WX2- 𝐻𝐻𝑝𝑝 = 0.81𝐻𝐻 = 77.8𝑚𝑚
Page A6
Working Example 2: Rectangular building, four equal divisions
For a better understanding of the calculation procedure and the possible difference
due to the division number, the same example is repeated with four equal
divisions (S1-S4) in each 90o sector, as shown in Figure A-4.
Figure A-4 Site information, four sectors and four divisions within each sector
Page A7
Figure A-5 Detailed site map for S4 in WX2+
T6
T7
T8
1. As shown in Figure A-5, several buildings (T1, T2, T3, T4, T5 and T9) and a
building lot (highlighted in red box, including T2) potentially provide
significant sheltering effect to the proposed building.
2. For T1, HT1 = 44m, XT1 = 60m; XT1/H ≈ 0.63, HT1/H ≈ 0.47; and HdT1 /H = 0.37.
(Figure A-6)
Page A8
3. The buildings in the building lot will be treated as a whole for providing the
sheltering effect. Obviously, T2 and T6 together will provide the largest
sheltering effect in the building lot. Approximately, T2 covers about 60% of
the division while T6 covers about 40%. XT2 = 100m, HT2 = 52.4m, XT2/H ≈
1.04, HT2/H ≈ 0.55, HdT2/H = 0.44; XT6 = 158m, HT6 = 61.1m, XT6/H ≈ 1.65,
HT6/H ≈ 0.64, HdT6/H = 0.45. The combined Hd of the building lot is HT2&6 /H
= HdT2×0.6+HdT6×0.4 ≈ 0.44×0.6+0.45×0.4 = 0.44.
4. For T3, XT3 = 230m, HT3 = 51.6m, therefore XT3/H ≈ 2.40, HT3/H ≈ 0.54,
HdT3/H ≈ 0.17.
5. For T4, XT4 = 240m, HT4 = 52m, therefore XT4/H = 2.50, HT4/H ≈ 0.54, HdT4/H
≈ 0.15.
6. For T5, XT5 = 360m, HT5 = 153m > H = 96, therefore XT5/H = 3.75, HT5/H = 1,
HdT5/H ≈ 0.45.
7. For T9, XT9 = 230m, HT9 = 32m, therefore XT9/H = 3.44, HT9/H = 0.31, HdT9/H
≈ 0.
8. Although HdT5/H = 0.45 > HdT2&6/H = 0.44, T5 only covers about 5% of the
whole sector, so HdT2&6 are buildings providing the largest sheltering effect,
and therefore the whole building lot will be removed as required by the Code.
9. The height reduction for the whole division can now be calculated from that in
𝛼𝛼 𝛼𝛼 𝛼𝛼 𝛼𝛼 𝛼𝛼
T1, T3, T4 and T5 ( 𝛼𝛼1 ≈ 40%, 𝛼𝛼3 ≈ 40%, 𝛼𝛼4 ≈ 10%, 𝛼𝛼5 ≈ 5%, 𝛼𝛼9 ≈ 5%,
𝛼𝛼1 + 𝛼𝛼3 + 𝛼𝛼4 + 𝛼𝛼5 + 𝛼𝛼9 = α, Figure A-6) using the principle of weighted
average.
HdT1 H H H
𝛼𝛼1 + dT3 𝛼𝛼3 + dT4 𝛼𝛼4 + dT5 ×𝛼𝛼5
Hd in S4 division is 𝐻𝐻𝐸𝐸𝑑𝑑4 /H = 𝐻𝐻 𝐻𝐻 𝐻𝐻 𝐻𝐻
= 0.24
α
Following the procedures above, the height of reduction in each division can be
calculated. The height of reduction in each of the four sectors can be calculated
���
𝐻𝐻𝑑𝑑� ��
𝐻𝐻����� �������
𝑑𝑑𝑑𝑑1 +𝐻𝐻 �������
𝑑𝑑𝑑𝑑2 +⋯+𝐻𝐻
using = 𝑑𝑑𝑑𝑑4
and summarised below.
𝐻𝐻 4
With 𝑍𝑍𝑝𝑝 = max(𝑍𝑍 − 𝐻𝐻𝐸𝐸 , 0.25𝑍𝑍), the effective building height 𝐻𝐻𝑝𝑝 is:
Table A-4 The effective building height (4 divisions)
WX1+ 𝑯𝑯𝒆𝒆 = 𝟎𝟎. 𝟔𝟔𝟔𝟔𝑯𝑯 = 𝟔𝟔𝟔𝟔. 𝟑𝟑𝟏𝟏
WX1- 𝐻𝐻𝑝𝑝 = 0.85𝐻𝐻 = 81.6𝑚𝑚
WX2+ 𝐻𝐻𝑝𝑝 = 0.78𝐻𝐻 = 74.9𝑚𝑚
WX2- 𝐻𝐻𝑝𝑝 = 0.80𝐻𝐻 = 76.8𝑚𝑚
The two calculation cases show similar results for four divisions and nine
divisions. In comparison of the two conditions, the nine divisions case has more
Page A9
divisions in each 90o sector, but it uses less time for calculation within each
division by avoiding complicated weighted average calculation. However, it is
always allowed to simplify the weighted average calculation by conservatively
ignoring the benefit from some sheltering buildings.
Page A10
(a) 0.9 ≤ 𝐻𝐻𝑐𝑐 /𝐻𝐻 < 1 (𝑏𝑏)𝐻𝐻𝑐𝑐 /𝐻𝐻 ≥ 1
Figure A-8 Results for 𝑯𝑯𝒅𝒅𝒅𝒅 /𝑯𝑯 = (𝟏𝟏 − 𝑯𝑯𝒆𝒆𝒅𝒅 /𝑯𝑯) as a function of 𝑯𝑯𝒅𝒅 /𝑯𝑯 and 𝑿𝑿𝒅𝒅 /𝑯𝑯
The simplified procedures for effective height calculation can be summarised as
follows:
1. Follow bullet points 1 to 3 in the page A3.
2. Use typical radius increment of 0.5𝐻𝐻 instead of 100m in bullet point 4 of
page A3 for easier reference.
3. The following steps apply to each of the four 90o sectors:
(a) Within radius of 2𝐻𝐻, apply 𝐻𝐻𝐸𝐸𝑐𝑐 /𝐻𝐻 = 0.7 for buildings taller than 0.9 𝐻𝐻;
(b) Within radius 𝑅𝑅 , where 2𝐻𝐻 < 𝑅𝑅 < 2.5𝐻𝐻 , apply 𝐻𝐻𝐸𝐸𝑐𝑐 /𝐻𝐻 = 0.7 for
buildings taller than 𝐻𝐻, and apply 𝐻𝐻𝐸𝐸𝑐𝑐 /𝐻𝐻 = 0.6 for other buildings taller
than 0.9𝐻𝐻.
(c) Within radius 2.5𝐻𝐻 < 𝑅𝑅 < 3𝐻𝐻 , apply 𝐻𝐻𝐸𝐸 /𝐻𝐻 = 0.6 for buildings taller
than 𝐻𝐻.
(d) If the gap between two adjacent buildings is smaller than half the average
of the two building breadths, they can be treated as one building.
(e) Any building fully occupying a division with 𝐻𝐻𝐸𝐸 /𝐻𝐻 ≥ 0.7 can be treated
as the most beneficial building and can therefore be removed.
The simplified procedures are illustrated below with an example shown in
Figure A-9, where the proposed building located in the centre of the figure is 90m
high, and nine equal divisions are adopted for the calculation. The displacement
height 𝐻𝐻𝐸𝐸 will be calculated for the four sectors respectively using the simplified
method.
Page A11
Figure A-9 Map of the studied site
Page A12
For the North (N) Section (Figure A-10(a)):
The surroundings are around 80m, which is about 0.9 of the proposed building
height.
(a) Within 2𝐻𝐻, identify towers T1 to T6 and T9 to T13; they all give 𝐻𝐻𝐸𝐸𝑐𝑐 /𝐻𝐻 =
0.7.
(b) Within 2.5H, identify towers T7&T8 and they give 𝐻𝐻𝐸𝐸𝑐𝑐 /𝐻𝐻 = 0.6.
(c) Division 5 is blank.
(d) Pick T1 to T5 as the most beneficial buildings, and they cover divisions 1 to
4 and divisions 6 to 9.
(e) Pick T9 and T10 to T13 as the second most beneficial buildings for division
3 and divisions 6 to 9. In this case 𝐻𝐻𝐸𝐸𝑁𝑁3 = 𝐻𝐻𝐸𝐸𝑁𝑁6 = 𝐻𝐻𝐸𝐸𝑁𝑁7 = 𝐻𝐻𝐸𝐸𝑁𝑁8 = 𝐻𝐻𝐸𝐸𝑁𝑁9 =
0.7𝐻𝐻.
(f) T7, T8 and T9 give 𝐻𝐻𝐸𝐸𝑁𝑁2 = 0.7𝐻𝐻 × 0.8 + 0.6𝐻𝐻 × 0.2 = 0.68𝐻𝐻for division
2.
(g) T6 and T7 give 𝐻𝐻𝐸𝐸𝑁𝑁1 = 0.7𝐻𝐻 × 0.8 + 0.6𝐻𝐻 × 0.2 = 0.68H for division 1.
𝐻𝐻𝑇𝑇4 𝑥𝑥𝑇𝑇4
(h) T14 gives 𝐻𝐻𝐸𝐸𝑁𝑁4 = 0.45 , with = 0.7 and = 2 for division 4.
𝐻𝐻 𝐻𝐻
(i) Because division 5 is blank, take 𝐻𝐻𝐸𝐸𝑁𝑁5 = 0.
(j) Finally 𝐻𝐻𝐸𝐸𝑁𝑁 /𝐻𝐻 = (0.7 × 5 + 0.68 × 2 + 0.45)/9 = 0.59.
Page A13
For the East (E) Sector (Figure A-10(b)):
The surroundings are around 80m, which is about 0.9 of the proposed building
height.
(a) Within 2 𝐻𝐻, identify towers T1 to T10 and they all give 𝐻𝐻𝐸𝐸𝑐𝑐 /𝐻𝐻 = 0.7.
(b) Within 2.5 𝐻𝐻, identify towers T11 to T14 and they give 𝐻𝐻𝐸𝐸𝑐𝑐 /𝐻𝐻 = 0.6.
(c) Pick up T1 to T5 as the most beneficial buildings, and they cover all the
nine divisions
(d) Pick up T6 to T8 as the second most beneficial buildings for divisions 1 to 3,
and it give 𝐻𝐻𝐸𝐸𝐸𝐸1 = 𝐻𝐻𝐸𝐸𝐸𝐸2 = 𝐻𝐻𝐸𝐸𝐸𝐸3 = 0.7𝐻𝐻.
(e) pick up T9 & T10 as the second most beneficial buildings for division 6,
and it give 𝐻𝐻𝐸𝐸𝐸𝐸6 = 0.7𝐻𝐻.
(f) Pick up T11 to T14 as the second most beneficial buildings for divisions 4,
5 and 7, and it give 𝐻𝐻𝐸𝐸𝐸𝐸4 = 𝐻𝐻𝐸𝐸𝐸𝐸5 = 𝐻𝐻𝐸𝐸𝐸𝐸7 = 0.6𝐻𝐻.
(g) Conservatively take 𝐻𝐻𝐸𝐸𝐸𝐸8 = 𝐻𝐻𝐸𝐸𝐸𝐸9 = 0.
(h) Finally 𝐻𝐻𝐸𝐸𝐸𝐸 /𝐻𝐻 = (0.7 × 4 + 0.6 × 3)/9 = 0.51.
Page A14
For the South (S) Sector (Figure A-10(c)):
The buildings are around 90m, which is about 1.0 of the proposed building height.
(a) Within 2.5𝐻𝐻, identify towers T1 to T8 and they all give𝐻𝐻𝐸𝐸𝑐𝑐 /𝐻𝐻 = 0.7.
(b) Within 3.0𝐻𝐻, identify tower T14 and it gives 𝐻𝐻𝐸𝐸𝑐𝑐 /𝐻𝐻 = 0.6.
(c) Pick up T1 to T4 as the most beneficial buildings, and they cover divisions 1
to 5; pick up T6 & T7 as the most beneficial buildings, and they cover
divisions 6 and 7.
(d) Pick up T5 and T8 as the second most beneficial buildings for division 5,
and it gives 𝐻𝐻𝐸𝐸𝑑𝑑5 = 0.7𝐻𝐻.
(e) Pick up T14 as the second most beneficial buildings for divisions 1 and 2,
and it gives 𝐻𝐻𝐸𝐸𝑑𝑑1 = 𝐻𝐻𝐸𝐸𝑑𝑑2 = 0.6𝐻𝐻.
(f) Pick up T8 to T12 as the second most beneficial buildings for division 4,
and it gives 𝐻𝐻𝐸𝐸𝑑𝑑4 = 0.6𝐻𝐻.
𝐻𝐻𝑇𝑇12 𝑥𝑥𝑇𝑇12 𝐻𝐻𝑑𝑑
(g) = 1 and = 3.5 and this gives = 0.45; this is the same for T13.
𝐻𝐻 𝐻𝐻 𝐻𝐻
Pick up T12 and T13 as the second most beneficial building for division 3,
and it gives 𝐻𝐻𝐸𝐸𝑑𝑑3 = 0.45𝐻𝐻 × 0.6 = 0.27𝐻𝐻.
(h) Pick up T4 as the second most beneficial building for division 6, and it gives
𝐻𝐻𝐸𝐸𝑑𝑑6 = 0.7𝐻𝐻 × 0.6 = 0.42𝐻𝐻.
(i) Conservatively take 𝐻𝐻𝐸𝐸𝑑𝑑7 = 𝐻𝐻𝐸𝐸𝑑𝑑8 = 𝐻𝐻𝐸𝐸𝑑𝑑9 = 0.
(j) Finally 𝐻𝐻𝐸𝐸𝑑𝑑 /𝐻𝐻 = (0.7 × 1 + 0.6 × 3 + 0.27 + 0.42)/9 = 0.35.
Page A15
For the West (W) Sector (Figure A-10d):
Surrounding towers T1 to T4, T9 and T13 to T16 are around 80m, which is about
0.9 of the proposed building height. Surrounding towers T5 to T8 and T10 to T12
are around 90m, which is about 1.0 of the proposed building height.
(a) Within 2.0𝐻𝐻, identify towers T1-T4 and T13 and they give 𝐻𝐻𝐸𝐸𝑐𝑐 /𝐻𝐻 = 0.7.
(b) Within 2.5𝐻𝐻, identify towers T5-T12 and they give 𝐻𝐻𝐸𝐸𝑐𝑐 /𝐻𝐻 = 0.7; identify
towers T14 to T17 and they give 𝐻𝐻𝐸𝐸𝑐𝑐 /𝐻𝐻 = 0.6.
(c) Most beneficial buildings: pick up T1 to T4 and they cover divisions 1 to 3.
Pick up T5 to T8, and they cover Divisions 6 to 9.
(d) Most beneficial buildings: pick up the two in the orange box, and it covers
division 4; pick up T17 and it covers division 5.
(e) Second most beneficial buildings: pick up T13 and it gives 𝐻𝐻𝐸𝐸𝑑𝑑1 /𝐻𝐻 = 0.7;
pick up T10 and T12, it gives 𝐻𝐻𝐸𝐸𝑑𝑑7 /𝐻𝐻 = 0.7; pick up T9 and it gives
𝐻𝐻𝐸𝐸𝑑𝑑9 /𝐻𝐻 = 0.7.
(f) Second most beneficial buildings: pick up T14 to T16, and it gives 𝐻𝐻𝐸𝐸𝑑𝑑2 =
𝐻𝐻𝐸𝐸𝑑𝑑3 = 0.6𝐻𝐻.
(g) Second most beneficial buildings: for division 4, 𝐻𝐻𝐸𝐸𝑑𝑑4 = 0.6𝐻𝐻 × 0.3 +
0.4𝐻𝐻 × 0.5 = 0.38𝐻𝐻 ; for division 5, 𝐻𝐻𝐸𝐸𝑑𝑑5 = 0.3𝐻𝐻 × 0.7 = 0.21𝐻𝐻 ; for
Division 6, 𝐻𝐻𝐸𝐸𝑑𝑑6 = 0.7𝐻𝐻 × 0.7 + 0.5𝐻𝐻 × 0.3 = 0.64𝐻𝐻 ; for division 8,
𝐻𝐻𝐸𝐸𝑑𝑑8 = 0.7𝐻𝐻 × 0.4 = 0.28𝐻𝐻.
(h) Finally 𝐻𝐻𝐸𝐸𝑑𝑑 /𝐻𝐻 = (0.7 × 3 + 0.6 × 2 + 0.38 + 0.21 + 0.64 + 0.28)/9 =
0.53.
Page A16
A3 Topographic Multiplier
Procedures for obtaining the parameter 𝑠𝑠 have changed to enable the
rules to be used with more irregular hill-forms.
In particular:
(a) The hill-slope definition has been changed.
(b) Except when downwind of an escarpment, position on a hill is
defined by site ground level relative to top and bottom of the hill.
(c) Rules for calculating 𝑠𝑠 downwind of the top height of an
escarpment and hill are combined to cover cases of relatively (but
not perfectly) flat downwind slopes.
The topographic method in the 2004 Code was made reference from
BS 6399-2. The closely similar method of the Eurocode, BS EN 1991-
1-4, including information in the UK National Annex and
UK PD 6688-1-4, was used as a starting place for development for the
Code. However, various changes were made to make it easier to use
in the context of real hill slopes which are not uniform in slope or
symmetrical.
The 2004 Code follows the ‘Simplified method’ of BS 6399-2 in using
a factor of
(1 + 1.2 𝑥𝑥 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑥𝑥 𝑠𝑠)2
The Code uses
2
2 × 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 × 𝑠𝑠
𝑆𝑆𝑐𝑐 = �1 + �
1 + 3.7𝐼𝐼𝑣𝑣,𝑧𝑧
These are the same only when 𝐼𝐼𝑣𝑣,𝑧𝑧 =0.18, and the 2004 Code becomes
non-conservative when the turbulence is lower. In the Eurocode the
simplified formula is thus restricted to buildings of less than 50m
height.
For ease of calculation, the formulas for calculating the topographic
multiplier, 𝑠𝑠, are provided in the Code. These are essentially the same
as those in the Eurocode, BS EN 1991-1-4, but have been changed to
reflect the new height based rules for positioning.
Page A17
Appendix B
Supplementary Information for
Section 4: Force and Pressure
Coefficients
B1 Pressure Coefficients for Building Envelope
with Dominant Openings
Page B1
Figure B-1 Definition of edge zones for irregular shape
Page B2
For openings that cannot be categorised as dominant, the internal
pressure coefficient can be estimated from the balance of flow, using
the formula below:
where
𝑃𝑃𝑝𝑝𝑗𝑗 is the external pressure at the 𝑗𝑗th opening
𝑃𝑃𝑐𝑐 is the wanted internal pressure
�2�𝑃𝑃𝑝𝑝𝑗𝑗 − 𝑃𝑃𝑐𝑐 �/𝜌𝜌𝑎𝑎 is the air flow speed at the 𝑗𝑗th opening
Page B3
B2 Pressure Coefficients for Building
Attachments
B2.2 Balconies
The wind loads on balustrades of a balcony are affected by the overall
wind flow around the building. The loads are also influenced by local
wind flow change due to the presence of blade walls or privacy
screens which subdivide the balconies. Balustrades close to the top
and the corners of the buildings are more likely to have higher net
pressure coefficient, due to the local flow acceleration in those
locations following the same concept discussed in B2.1 above. The
Page B4
current provision for net pressure values of balcony wall and
balustrades in the Code is based on “AWES-HB-001-2012”, which
gives pressure coefficients of ±1.8 for corner and top-floor balconies
and ±1.5 for balconies in other locations. For simplicity, ±1.8 is taken
in the Code. For balcony slabs, -1.8 is also used for uplift forces, while
+0.9 is taken for downward forces, similar to an attached canopy.
As the wind loads on the balustrades are affected by the overall wind
flow due to the tower, the reference height of the balcony and
balustrade shall be that of the building. For balcony within half of the
displacement height, reduction of wind loads of 20% is allowed, as
discussed in Section 4.3.1.
Page B5
B3 Pressure Coefficient for Free-standing
Walls
See Section 4.3.5.
Page B6
Appendix C
Supplementary Information for
Section 5: Size Factor and Size
and Dynamic Factor
C1 Equations for Calculation of Size Factors
Equations are provided for use with spreadsheet or other
computational methods. They are based on an empirical fit to the
theoretical model, as described Section 5.1.
Figure C-1
Re-plotting based on data in paper “Intrinsic and supplementary
damping in tall buildings”, Smith et al., Structures and Buildings,
Issue 63, SB 2
Page C1
The values given in Willford and Smith are also obtained using more
modern procedures which are able to determine systematic errors
based on the duration and quality of measurement. Random error
associated with many past measurements methods tended to
overestimate damping levels, particularly smaller values, which have
to be measured over a longer period.
One problem with using measured damping is that there may be
variations with amplitude of motion. Friction damping, e.g. due to
cladding, tends to reduce with amplitude. In other cases, slippage may
only occur under higher loads and be associated with a loss of
stiffness. Foundation damping also depends on amplitude. There are
also theories that on-going cracking of concrete will add to damping,
but clearly this would be contributing to fatigue damage of the
structure if it were relied on. It should be noted that the vertical
structure of a tall tower is associated with much of the elastic energy
of vibration and this is effectively pre-stressed under deadweight and
pre-stressed concrete has a material damping much closer to 0.5% of
critical than values often used in wind tunnel testing.
The best-estimate damping values of the Code are based on the
available data above at stress levels associated with noticeability of
building motions. Following previous practice, damping values for
structural loading have been increased by 50%.
There is a small difference between steel and concrete buildings due to
the lower material damping associated with steel. Many buildings in
Hong Kong will be composite and values between the two may be
used in this case.
The effect of building taper on structural damping has been included
by using a structural aspect ratio of height (from foundation/base level)
over base dimension in the direction of vibration. In cases where a
flexible structure is mounted on a relatively rigid and wider base, it
shall also be checked with the height measured above the rigid base
and the structural dimension at the level just above it. The aspect ratio
for calculating damping ratio in Table C2-1 and C2-2 of the Code
shall be the larger one of the two methods.
In some cases it may be beneficial to build extra damping into the
structure. In these cases the best-estimate structural damping may
reasonably also be taken into account.
Page C2
Appendix D
Supplementary Information for
Section 6: Requirements for
Wind Tunnel Testing
D1 Considerations for Wind Tunnel Testing of
Unusual Structures
D1.1 General
Wind tunnel test procedures for common forms of buildings and
bridges are well established but are frequently based on certain
simplifying assumptions that are not generally valid. In the case of
unusual structures, especially those lacking symmetry, containing
slender elements or with rounded surfaces, the usual methods may not
be directly applicable. The discussion that follows identifies some
example cases where alternative methods should be followed, but this
is not intended as a comprehensive guide for unusual structures.
Page D1
designed to take account of the form of the local structure. In either
case specialist advice may be required.
Page D2
necessary multi-modal frequency domain methods, especially for
complex roof structures.
Page D3
Appendix E
Load Combination for Multi-
towers Sitting on the Same
Podium
E1 General
Multi-towers structurally integrated together with a common podium
is a typical and characteristic structural form in Hong Kong. It is a
challenging task to envelope all critical loading scenario with a limited
number of load cases. There is insufficient academic study in this area
nor ready guidance in design codes internationally.
This Appendix is to give reference on determination of critical load
cases for podium design in the multi-tower cases. The critical load
cases are categorised into two groups, respectively considering the
translational and torsional effects. The designer needs to review
whether those recommended load cases are sufficient to cover the
loading effects.
It is noted that the user of the Code can always reduce the number of
design load combinations by conservatively taking the full load rather
than using the reduction factors.
Page E1
Figure E-2 Plan of the building
Procedures to calculate the critical translational load cases are as
follows:
1. The towers and the podium follow their own local axis while the
selection of global axis can be independent from the local axis of
the towers and the podium as shown in Figure E-3.
2. The translational wind loads (Fx, Fy) for each tower are calculated
under the local axis following the Code.
3. For each tower, the translation loads (Fx, Fy) are combined using
the load combination factors of Case 1 and Case 2 in Table 2-1 of
the Code (copied as Table E-1 below). Each combined
translational load is represented as a vector. For example,
vector 1-1 as shown in Figure E-4 is the combined load of
1.0Fy+0.55Fx.
4. The checking plan of each tower is divided into eight subzones
with each representing one main checking direction. The vectors
of the resultant combined forces (eight in total) for each tower are
drawn on the global axis surface as shown in Figure E-5.
5. The vector forces on the podium should be obtained following the
same procedure as towers in steps 2 to 4 above.
6. Pick up forces from each tower (maximum one vector for a tower)
and the podium in each subzone and do the load combination
(vector summation of picked forces). Then select the load
combination which has the largest magnitude of the vector sum in
the subzone. If more than two vectors from a tower fall in a
subzone, the one contributing the largest vector sum magnitude
after combined with vectors from the podium and other towers
shall be picked up. If there is no vector from the podium or a
tower falling in a subzone, then zero vector can be assigned to the
tower/podium in the subzone. This gives totally eight load cases
with all eight subzones considered.
7. The accompanying torsional loads are torsions on individual
towers and the podium following requirement in clause 2.2.2 of
Page E2
the Code with load factor of 0.55. With consideration of opposite
directions of the torsional loads (+, -), 16 (8 × 2 = 16) load cases
are needed in total to envelope all the critical translational load
cases, similar to the 16 translational governed load cases in clause
2.2.4 of the Code.
Page E3
Figure E-5 Resultant combined forces for each tower
Page E4
torsional effect for a single building plan. The discussion above
shows cross references of the torsional issue from various perspectives.
As shown in Figure E-6, the partial load cases critical to the torsional
effect on podium can be generated by applying 55% translational
loads together with an equivalent torsion by shifting the full
translational loads with 10% eccentricity.
Figure E-6 (a) BS EN rule for torsion loads; (b) Equivalent uniform
distributed load to the BS EN rule; (c) Examples partial load application
It shall also be noted that the discussion above focuses on overall
torsion generated by partial tower loads above the podium roof (M*).
Therefore, the geometry of the podium is irrelevant to the M*
calculation at this step. The torsion loads caused by lateral load
eccentricity in individual towers and podium (M) can be added
separately afterwards.
The critical torsional load cases can be obtained by the following
procedures:
1. For each tower, the loads on individual towers are calculated
following the Code as shown in Figure E-7(a), (b) and (c).
2. Obtain the eight translational governed loads according to Section
E2. In each subzones defined in Figure E-5, the resultant force
(Flateral-i) can be calculated by vector summation. For example,
Flateral-1 is the resultant force by F1-1, F2-1 and F3-1, as shown in
Figure E-7(d)
3. For each Flateral-i, the maximum projecting diagonal breadth Bi can
be identified as shown in Figure E-7(e).
4. For each subzone, the overall torsion can be calculated as
𝑀𝑀𝑐𝑐∗ = 𝐹𝐹𝑙𝑙𝑎𝑎𝑐𝑐𝑝𝑝𝑐𝑐𝑎𝑎𝑙𝑙−𝑐𝑐 × 0.1𝐵𝐵𝑐𝑐 (i = 1 to 8).
Page E5
5. The maximum torsion is chosen as the overall torsional due to
partial tower loads as 𝑀𝑀∗ = max(𝑀𝑀𝑐𝑐∗ ) (i = 1 to 8).
6. Repeat the eight governing global lateral loads (Flateral-I, i = 1 to 8)
but with a load factor of 0.55. Local torsion of individual tower
and podium corresponding (all in the same direction) to each of
the eight global lateral load M is applied simultaneously with load
factor of 1.0.
7. Apply the torsion M* (concentrated torque) calculated in step 5
simultaneously at the top of the podium for all the eight cases
following the same sign of M, as shown in Figure E-8.
8. Consider the opposite directions (+, -) of M and M*, there are
totally another 16 torsional cases.
Page E6
Figure E-8 Critical torsional load cases
Notes:
1. Flateral represent the set of lateral forces from the tower and the
podium which give the largest vector sum in each of the eight
directions;
2. M represent the torsional forces induced by the lateral forces on
each tower and on the podium following the rule in Section 2.2.2;
3. M* represent the overall torsional forces induced by the partial
lateral forces above the podium roof.
Page E7
consideration of mirror cases by loading T2 and unloading T1, there
are another four load cases. Therefore, there are totally eight cases.
Unloaded Unloaded
(b) Max torsional effect in zones (c) Max torsional effect in zones
1&2 3&4
Unloaded Unloaded
(d) Max torsional effect in zones (e) Max torsional effect in zones
5&6 7&8
Page E8
Appendix F
References
References
1. Australian/New Zealand Standard: Structural design actions, Part 2: Wind
actions, AS/NZS 1170.2:2011
2. ASCE Standard: Minimum Design Loads for Buildings and Other Structures,
ASCE/SEI 7-16
3. British Standard: Loading for Buildings – Part 2: Code of practice for wind
loads, BS6399 -2: 1997
4. British Standard: Eurocode 1: Actions on structures – Part 1-4: General
actions – Wind actions, BS EN 1991-1-4:2005+A1:2010
5. Cook, N.J. The designer’s guide to wind loading of building structure – Part 2:
Static structures. Butterworth – Hernemann Ltd, 1990
6. John D. Holmes, Wind Loading of Structures, Second Edition, Taylor &
Francis
7. User’s Guide – NBC 2010 Structural Commentaries (Part 4 of Division B)
Page F1