•1

General Briefing on
Basic Theory of Wind Engineering

November 2019, Ir. Law Chi Wai

Contents
•2

1. Wind Pressure vs Wind Speed

2. Measurement of Wind Speed
3. Averaging Period of Wind Speed/Pressure
4. Return Period of Wind Speed/Pressure
5. Turbulence Intensity
6. Vertical Profiles of Wind Speed/Pressure and Turbulence Intensity
7. Spectrum of Wind Energy and Dynamic Response of Structures
8. Force Coefficients – Effects due to shape of structure
9. Topographical Effect
10. Sheltering Effect
11. Wind Induced Acceleration
12. Wind Climate
13. Wind Tunnel Test
1. Wind Pressures vs Wind Speed
•3

1
Bernoulli’s equation for a fluid 𝜌𝜌𝑣𝑣 2 + 𝜌𝜌𝑔𝑔𝑔 + 𝑝𝑝 = 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
2

Velocity Elevation Pressure

head head head

For air, the elevation head 𝜌𝜌𝑔𝑔𝑔 is negligible

1 1
∴ 𝜌𝜌𝑣𝑣 2 + 𝑝𝑝 = 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 and decrease of 𝜌𝜌𝑣𝑣 2 = increase in 𝑝𝑝
2 2
Stagnation point,
velocity = 0, so
1
𝑝𝑝 = − 𝜌𝜌𝑣𝑣 2 where 𝜌𝜌
2
is the density of air
(1.2kg/m3) and 𝑣𝑣 is
Velocity 𝑣𝑣 the wind speed.
1. Wind Pressures vs Wind Speed
•4

1. However, wind speed may not be perfectly stagnant at all points when wind flows on an object. To
account for the shape and roughness of the object, different factors have to be used to multiply to 𝑝𝑝
1
so that 𝑝𝑝 = 𝐶𝐶𝑝𝑝 𝜌𝜌𝑣𝑣 2 where 𝐶𝐶𝑝𝑝 is a force coefficient. Anyway 𝒑𝒑 ∝ 𝒗𝒗𝟐𝟐 , holding other parameters
2
constant;

2. In most of the Codes, the design wind pressures are treated as static pressures arising from wind
1
speed as with the basic pressure 𝑝𝑝 = 𝜌𝜌𝑣𝑣 2 .
2
1. Wind Pressures vs Wind Speed
•5

1. Determination of Wind Loads on Structures is by treating the wind pressures as static pressures,
which is based on the basic wind pressure multiplied by a number factors to account for (i)
direction (ii) shape; (iii) topography; (iv) sheltering; (v) dynamic effects. The total lateral loads,
moments are summed by multiplying the pressure to the exposed areas.

𝑆𝑆 = � 𝑝𝑝 𝑑𝑑𝑑𝑑
p

z
𝑀𝑀 = � 𝑝𝑝𝑧𝑧 𝑑𝑑𝑑𝑑

Total shear 𝑆𝑆 Total Moment 𝑀𝑀

2. Alternatively, time history analysis to a wind load function varying with time can be carried out.
The analysis is very time consuming and it is also very difficult to have an appropriate time history
for analysis. So this is seldom carried out.
2. Measurement of Wind Speed
•6

1. Wind speed is
traditionally measured by
anemometer, the
commonest type is the
cup anemometer;

2. When wind blows on it,

the speed of rotation
(counted by number of
turns per second) is
proportional to wind
speed;

3. Now there are some other

types such as Vane
anemometer.
2. Measurement of Wind Speed
•7

1. In Hong Kong, wind speeds

are measured in many
stations, the wind speeds
measured in Waglan Island
are the highest (during
typhoon) and used to
determine wind speed for
structural design;

2. The wind speed measured

in Waglan Island is that of
“open sea” which is un-
sheltered. So design to such
wind speeds are the
highest. There are thus
rooms for reduction for
structures situated inland.
3. Averaging Period of Wind Speed/Pressure
•8

1. Wind speed varies with time – strictly speaking, not appropriate to assign constant values;

2. So, to be a value that can be handled, wind speed has to be averaged over a period of time to
obtain a mean value. The current practice is to averaged over an hour, 10 minutes, 1 minute or 3
second;

3. The hourly mean speed is the maximum value of wind speeds (taken from the past year records),
each being averaged over a period of an hour;

4. The same principle applies to 3-sec gust speed. So obviously, maximum value of a shorter duration
wind speed is greater than that of a longer one. The 3-sec gust wind speed is greater than the
hourly wind speed;

5. In Hong Kong, both hourly mean speed and 3-sec gust speed are used. Whilst the hourly mean
speed is obtained by actual measurement of wind speeds, the 3-sec gust is determined by a
statistical formula based on the hourly mean speed as:
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 𝑜𝑜𝑜𝑜 𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠
𝑣𝑣3−𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 = 𝑣𝑣ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 1 + 3.7 ×
𝑣𝑣ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜
3. Averaging Period of Wind Speed/Pressure
•9

Average wind
speed is Area
divided by T

Wind Speed and Averaging Time

4. Return Period of Wind Speed
•10

1. Wind speed used in structural design contains another time (and statistical) element called
“Return Period”;

1
2. By a return period of 𝑁𝑁 years, we mean that it is the wind speed that has the probability of being
𝑁𝑁
exceeded in one year;

3. So for a wind speed of 50 years return period, the probability of being exceeded per one year is
1/50 = 0.02. Or alternatively, it has a probability of 1 − 0.02 = 0.98 of not being exceeded;

4. The determination of the wind speed of N years return period is by “Extreme Value Analysis” in
which the values chosen for analysis are the “extreme values”, i.e. the highest wind speed, say in
each year;
4. Return Period of Wind Speed
•11

Formula for calculation of return period wind speed from a set of largest yearly fastest-mile wind speeds (as
extreme values) at 10m above ground (From “Wind Effects on Structures 3rd Edition” 3.2.2 by Simiu):
𝑣𝑣� �
𝑁𝑁 = 𝑋𝑋 + 0.78 𝑙𝑙𝑙𝑙 𝑁𝑁 − 0.577 𝑠𝑠
where 𝑋𝑋� is the mean speed;
𝑁𝑁 is the year of return period
𝑠𝑠 is the standard deviation
e.g. given 34 readings of largest yearly fastest-mile wind speeds in mph as
57, 65, 62, 58, 64, 65, 59, 65, 59, 60, 64, 65, 73, 60, 67, 50, 74
60, 66, 55, 51, 60, 55, 60, 51, 51, 62, 51, 54, 52, 59, 56, 52, 49

𝑋𝑋� = 59.15mph, s = 6.41mph , by the formula above

For 𝑁𝑁 = 50; 𝑣𝑣�
50 = 75.82mph;
For 𝑁𝑁 = 475; 𝑣𝑣�
475 = 87.08mph;
For 𝑁𝑁 = 2475; 𝑣𝑣�
2475 = 95.34mph.

The 𝑣𝑣�
50 , 𝑣𝑣�
475 and 𝑣𝑣�
2475 calculated are all greater than that determined by normal distribution, e.g. 𝑣𝑣�
50 =
1
75.82mph > that of normal distribution = 72.314mph with probability being exceeded.
50
4. Return Period of Wind Speed
•12

The distribution of the set of largest yearly fastest-mile wind speeds are “extreme values”. If assumed to be falling into
the “Gambel distribution” (the “Generalized Cumulative Distribution Function Type I” or “Fisher-Tippet Type I
distribution” for extreme value analysis) as
− 𝑥𝑥−𝜇𝜇 𝜋𝜋
𝐹𝐹 𝑥𝑥 = exp −exp where S𝐷𝐷 𝑋𝑋 = 𝜎𝜎; 𝐸𝐸 𝑋𝑋 = 𝜇𝜇 + 0.5772𝜎𝜎 and S𝐷𝐷 𝑋𝑋 is the standard deviation and 𝐸𝐸 𝑋𝑋
𝜎𝜎 6
is the expected value or the mean.
𝑑𝑑𝐹𝐹 𝑥𝑥 1 − 𝑥𝑥−𝜇𝜇 − 𝑥𝑥−𝜇𝜇
The distribution will be = exp −exp exp
𝑑𝑑𝑑𝑑 𝛼𝛼 𝜎𝜎 𝜎𝜎

Plotting both the distribution and cumulative distribution functions for the sample :

Gumbel Distribution Cumulative Gumbel Distribution

0.08 1.10
1.00
0.07
0.90
0.06 0.80 0.98
0.05 0.70
Probability
Probability

0.60
0.04
0.50
0.03 0.40
0.02 0.30
0.20 75.82mph
0.01
0.10
0 0.00
0 10 20 30 40 50 60 70 80 90 100 110 120 0 10 20 30 40 50 60 70 80 90 100 110 120
Wind speed (mph) Wind speed (mph)
4. Return Period of Wind Speed
•13

− 𝑥𝑥−𝜇𝜇 2
1
The Gambel Distribution used is compared with Normal Distribution 𝑝𝑝 𝑥𝑥 = 𝑒𝑒 2𝜎𝜎2 ;
2𝜋𝜋𝜎𝜎 2

The Gambel Distribution gives higher values

Gambel Distribution vs Normal Distribution Cumulative Gambel vs Cumulative Normal Distribution

Gumbel Normal Gumbel Normal
0.08 1.10
0.07 1.00
0.90
0.06
0.80
0.05 0.70
Probability

0.60
Probability
0.04
0.50
0.03 0.40
0.02 0.30
0.20
0.01
0.10
0.00 0.00
0 10 20 30 40 50 60 70 80 90 100 110 120 0 10 20 30 40 50 60 70 80 90 100 110 120
Wind speed (mph) Wind speed (mph)
3. Return Period of Wind Speed
•14

For a structure of design life 50 years, probability that the wind speeds of 50, 475 and 2475 years
return periods will be exceeded in the design life are:

50 years return period is (𝑁𝑁 = 50)

50 50
1 49
1− 1− =1− = 0.636 = 63.6%
𝑁𝑁 50

475 years return period is (𝑁𝑁 = 475)

50 50
1 474
1− 1− =1− = 0.1 = 10%
𝑁𝑁 475

2475 years return period is (𝑁𝑁 = 2475)

50 50
1 2474
1− 1− =1− = 0.02 = 2%
𝑁𝑁 2475

These return periods and probabilities are more popularly used in earthquake engineering in
determining the level of seismicity to be designed for.
5. Turbulence Intensity
•15

1. Being a variable parameter, wind speed apart from its mean value over a number of measured
values, also has another statistical parameter which is the standard deviation, measuring its
variability from its mean value;

2. A dimensionless parameter called “turbulence intensity” is defined which is simply 𝐼𝐼𝑔𝑔 =

𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝜎𝜎
= ;
𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 𝜇𝜇

3. The equation for determining 3-sec gust from hourly mean speed is
𝑣𝑣3−𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 = 𝑣𝑣ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 1 + 3.7 × 𝐼𝐼𝑔𝑔 where 3.7 is called the gust factor by
𝑔𝑔𝑝𝑝 = 2𝑙𝑙𝑙𝑙 3600𝑓𝑓 where 𝑓𝑓 is the effective or average frequency of the structure taken in the
order from 0.2Hz to 0.3Hz, i.e. 𝑔𝑔𝑝𝑝 ranges from 3.63 to 3.74 and so take the average of 3.7.

4. The parameter is also used in determination of other design values;

6. Vertical Profile of Wind Speed and Turbulence Intensity
•16

1. Wind speed and turbulence intensity vary with height above ground;

2. The ground acts as a rough boundary providing resistance to wind speed. So wind speed at ground
level can be taken as 0. At some higher levels being freer from influence of the ground, the wind
speed increases. But it is generally assumed that the speed will achieve its maximum value at a
height, called “gradient height” and the speed becomes “gradient wind speed”;

3. As a rough boundary creating the most interfering effects to the wind, the turbulence intensity is
maximum at ground but decreases with increase in height;

4. For ease of calculation, the vertical profiles of the wind speed and turbulence intensity are
simulated by curves obeying the “power rule” or the “logarithmic rule” to match measured values.
The equations of them are respectively in the form of :
𝛼𝛼
𝑧𝑧
𝑣𝑣𝑧𝑧 = 𝑣𝑣𝑔𝑔 0 < 𝛼𝛼 < 1
𝑧𝑧𝑔𝑔
1 𝑧𝑧
𝑣𝑣𝑧𝑧 = 𝑣𝑣∗ 𝑙𝑙𝑙𝑙
𝑘𝑘 𝑧𝑧0
6. Vertical Profile of Wind Speed and Turbulence Intensity
•17

Simulation of Measured Wind Speed Simulation of Turbulence Intensity

The constants of by Mathematical Profile by Mathematical Profile
the power-rule Simulated Profile Measured Data
Simulated Profile Measured Data
and logarithm
equations are so 500 500

established so 450 450

as to simulate
400 400
the measured
data. 350 350
Height above Ground (m)

Height above Ground (m)

300 300

250 250

200 200

150 150

100 100

50 50

0 0
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 0 0.05 0.1 0.15 0.2 0.25 0.3
Speed (m/s) Turbulence Intensity
6. Vertical Profile of Wind Speed and Turbulence Intensity
•18

Gradient Gradient
wind speed height

Turbulence intensity =
𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑜𝑜𝑜𝑜 𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠
𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠
power law for
wind velocity
𝑍𝑍𝑔𝑔 profile Turbulence
𝛼𝛼 intensity profile
𝑍𝑍 𝛼𝛼 = 0.11 in
𝑉𝑉𝑧𝑧 = 𝑉𝑉𝑔𝑔 𝑍𝑍
𝛽𝛽
𝑍𝑍𝑔𝑔 2004 Code 𝐼𝐼𝑧𝑧 = 𝐼𝐼𝑔𝑔 𝛽𝛽 = −0.11 in
𝑍𝑍𝑔𝑔 2004 Code

𝑉𝑉𝑔𝑔 is the gradient wind 𝐼𝐼𝑔𝑔 is the gradient turbulence

speed intensity
6. Vertical Profile of Wind Speed and Turbulence Intensity
•19

1. The Hong Kong Wind Codes (1983, 2004 and the current 2019) adopts power rule for determination of
1
wind speed (and also wind pressure by 𝑝𝑝 = 𝜌𝜌𝑣𝑣 2 ) and turbulence intensity. The power rule equations in
2
the 2004 Code are as follows:

𝛼𝛼
𝑧𝑧 𝑧𝑧 0.11
𝑣𝑣 𝑧𝑧 = 𝑣𝑣𝑔𝑔 = 59.5 ×
𝑧𝑧𝑔𝑔 500
−𝛼𝛼
𝑧𝑧 𝑧𝑧 −0.11
𝐼𝐼 𝑧𝑧 = 𝐼𝐼𝑔𝑔 = 0.087 ×
𝑧𝑧𝑔𝑔 500
𝑣𝑣 𝑧𝑧 : wind speed at height 𝑧𝑧 above ground;
𝑧𝑧𝑔𝑔 : gradient height, taken as 500m in the 2004 Code
𝑣𝑣𝑔𝑔 : wind speed at gradient height taken as 59.5m/s
𝐼𝐼𝑔𝑔 : turbulence intensity at gradient height taken as 0.087
𝛼𝛼 : power exponent

2. The power rule equation for wind speed is not explicitly given in the 2019 Code, but it can be deduced
𝑧𝑧 0.08 0.16
to be 𝑣𝑣 𝑧𝑧 = 78.528 (from the pressure equation 𝑄𝑄𝑜𝑜.𝑧𝑧 = 3.7 𝑍𝑍𝑒𝑒 /500 at (Eqn 3-2)) and that of
500
the turbulence intensity distribution same as the 2004 Code.

3. 78.5m/s is the same 3 sec gust speed of the Wind Code 2004 at the gradient height 500m.
6. Vertical Profile of Wind Speed and Turbulence Intensity
•20

4. The 2019 Code uses only 3-sec gust wind pressures (as in Table 1 of the Wind Code 2004). The
hourly mean wind pressure (as in Table 2 of the Wind Code 2004) is not used;
1
5. The 3-sec gust wind pressure is × 𝑎𝑎𝑎𝑎𝑎𝑎 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 × 3 sec 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠2
2
6. The formula in the 2019 Code for “wind reference pressure” which varies with height is given by
𝑍𝑍𝑒𝑒 0.16
𝑄𝑄𝑜𝑜,𝑧𝑧 = 3.7 in kN/m2 (for open sea exposure) which is very close to the gust wind
500
1
pressure of the wind code 2004 given by × 𝑎𝑎𝑎𝑎𝑎𝑎 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 × 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠2 = 𝑞𝑞𝑧𝑧 =
2
2
1 𝑧𝑧 0.11 𝑧𝑧 −0.11
× 1.2 × 1.05 × 56. 6 × × 1 + 3.7 × 0.087 ×
2 500 500

Density of Gradient Turning Turbulence

air Adjust
wind wind hourly wind intensity =
Add 5% to
speed speed to speed to 3 sec standard
speed
(hourly) height gust wind deviation ÷
speed mean

Density of air = 1.2kg/m3, 𝑞𝑞𝑧𝑧 so determined is in N/m2, or × 10−3 to become kN/m2

6. Vertical Profile of Wind Speed and Turbulence Intensity
•21

𝑍𝑍𝑒𝑒 0.16
7. In the formula 𝑄𝑄𝑜𝑜,𝑧𝑧 = 3.7 in the 2019 Wind Code, the symbol 𝑍𝑍𝑒𝑒 is the “effective height”
500
which takes into account the sheltering effect of the surrounding by which 𝑍𝑍𝑒𝑒 can be taken as less
than the actual height. This is a new provision in the Wind Code 2019 which is not found in the
Wind Code 2004. The use of sheltering effect is taken from the British Code BS6399-2.

𝑍𝑍𝑒𝑒 0.16
8. The formula 𝑄𝑄𝑜𝑜,𝑧𝑧 = 3.7 is for determination of 3-sec gust wind pressure which has lumped
500
the constants such as air density etc. into simple constants of 3.7 and 0.16. The profile is very close
to the Wind Code 2004.
6. Vertical Profile of Wind Speed and Turbulence Intensity
•22
6. Vertical Profile of Wind Speed and Turbulence Intensity
•23

Comparison of Basic Wind Pressures (3sec 2004 Code basic wind pressure due to 3-sec gust (in kPa)
gust) between 2004 & 2019 Wind Codes and z in m
2
1 𝑧𝑧 0.11 𝑧𝑧 −0.11
CoP2004 Gust Pressure New Code Pressure × 0.0012 × 1.05 × 56. 6 × × 1 + 3.7 × 0.087 ×
2 500 500
500

450 2019 Code basic wind pressure due to 3-sec gust (in kPa)
400
and 𝑍𝑍𝑒𝑒 in m
𝑍𝑍𝑒𝑒 0.16
350 𝑄𝑄𝑜𝑜,𝑧𝑧 = 3.7
500
Height above Ground (m)

300

250 2004 Code and 2019 Code turbulence intensity

𝑧𝑧 −0.11
200
𝐼𝐼𝑧𝑧 = 0.087 ×
500
150

100
The 2019 Code does not directly present hourly mean speed
50 or pressure for design except for simulation of wind load in
Wind Tunnel Tests. The formulae for conversion from hourly
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 mean speed to gust speed is identical to the 2004 Code.
Pressures (kPa)
6. Vertical Profile of Wind Speed and Turbulence Intensity
•24

From Wind Code 2004

1 𝑧𝑧 0.11 𝑧𝑧 −0.11 2
1 𝑧𝑧 0.11 2
𝑞𝑞𝑧𝑧 = × 0.0012 × 1.05 × 56. 6 × × 1 + 3.7 × 0.087 × 𝑞𝑞𝑧𝑧 = × 0.0012 × 1.05 × 56. 6 ×
2 500 500 2 500
7. Spectrum of Wind Energy and Dynamic Response of
Structure
•25

1. A wind pressure, being a varying pressure with time can be considered to be made up of many wind
loads of varying frequencies;

2. In fact, it can be taken as infinite no. of wind loads of a continuously varying frequencies forming a
“spectrum”, i.e. a frequency domain of “wind energy”;

3. The use of wind energy spectrum is to facilitate analysis of dynamic response of a structure due to
wind load, which is easier than that of time history analysis;

4. The structure will respond with greater deflections and thus greater internal forces if its natural
frequency is close to that with maximum energy.
 Wind Velocity in Wind Velocity in
Wind Velocity in

•26
(m/sec) (m/sec)
(m/sec)

0
20
40
60
80
100

-15
-10
-5
0
5
10
15
0
10
20
30
40
50
60
1 1 1
14 15 14
27 27
29
40 40
43
53 53
Structure
57
66 66
71
79 79
85
92 92
99
105 105
118 113
118
131 127
131
144 141
144

Time in Second
Time in Second
157 155

Time in Second
Wind Speed Record

157

Wind Speed Component 3

170 169
170
183 183
183

Wind Speed Component 1 (Mean)

196 197 196
209 211 209
222
225 222
235
239 235
248
253 + 248
=

Wind Velocity in Wind Velocity in

(m/sec) (m/sec)
-6
-4
-2
0
2
4
6

+ …….
-20
-15
-10
-5
0
5
10
15
20
1
1
14
14

27 27

40 40

53 53

66 66

79 79
92 92
105 105
118 118
131 131
144 144
Time in Second
Time in Second

157 157
Wind Speed Component 4

170
Wind Speed Component 2

170
183 183
196
7. Spectrum of Wind Energy and Dynamic Response of

196
209
209
222
222
235
235
248
248
7. Spectrum of Wind Energy and Dynamic Response of
Structure
•27

Wind Speed Record

100
Wind Velocity in

80
Only the varying part
(m/sec)

60
40
20
0
1
14
27
40

53
66
79
92
105
118
131
144
157

170
183
196
209
222

235
248
Time in Second
Fourier Transform
⇓
Area below is the
𝑆𝑆𝑓𝑓 𝑛𝑛
variance 𝜎𝜎𝐵𝐵 2

𝑛𝑛, frequency
7. Spectrum of Wind Energy and Dynamic Response of
Structure
•28

Natural Wind spectrum

7. Spectrum of Wind Energy and Dynamic Response of
Structure
•29

Graph showing normalized auto-spectra, Von Kármán (blue), Kaimal (red) and =10m (solid),=30m
(dash)
Picture Courtesy of Milne et al
7. Spectrum of Wind Energy and Dynamic Response of
Structure
•30

1. When a single degree freedom structure of natural frequency 𝑓𝑓0 under an external excitation load of
frequency 𝑓𝑓, its motion is governed by the equation with consideration of “damping”:
𝑚𝑚𝑥𝑥̈ + 𝑘𝑘𝑥𝑥̇ + 𝑐𝑐𝑐𝑐 = 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃(𝜔𝜔𝑡𝑡)= 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 2𝜋𝜋𝑓𝑓𝑓𝑓

𝑃𝑃 𝑠𝑠𝑠𝑠𝑠𝑠 2𝜋𝜋𝑓𝑓𝑓𝑓 𝑃𝑃 𝑘𝑘
2. Solving, 𝑥𝑥 = × = × 𝐻𝐻 𝑓𝑓 𝑠𝑠𝑠𝑠𝑠𝑠 2𝜋𝜋𝑓𝑓𝑓𝑓 with 𝑓𝑓0 = 2𝜋𝜋 , the
𝑘𝑘 2 𝑘𝑘 𝑚𝑚
𝑓𝑓 2 2𝜁𝜁𝑓𝑓 2
1− +
𝑓𝑓0 𝑓𝑓0
𝑐𝑐
fundamental frequency of the system and 𝜁𝜁 = as the damping ratio.
2𝑘𝑘𝑘𝑘
1
The term = 𝐻𝐻 𝑓𝑓 is known as the “mechanical admittance” function of the system
2
𝑓𝑓 2 2𝜁𝜁𝑓𝑓 2
1− +
𝑓𝑓0 𝑓𝑓0

3. So if 𝑓𝑓 is close to 𝑓𝑓0 and 𝜁𝜁 is small, then the displacement 𝑥𝑥 will be very large, i.e. resonance.

4. A building structure is far from being a single degree of freedom. But its fundamental frequencies
usually have the greatest response and are chosen for analysis as good approximation.
7. Spectrum of Wind Energy and Dynamic Response of
Structure
•31

“Mechanical Admittance” of Structure shows the response to loads as related to its natural frequency.

Plot of Mechanical Admittance vs f/f0 Mechanical Admittance of a

ζ = 0.01 ζ = 0.02 ζ = 0.05 ζ = 0.1 ζ = 0.2 ζ = 0.5 ζ=1 structure is calculated by
100 1
𝐻𝐻 𝑓𝑓 =
2
𝑓𝑓 2 2𝜁𝜁𝑓𝑓 2
1− +
𝑓𝑓0 𝑓𝑓0

10
The parameter is obtained by
dynamic analysis of a structure
which indicates its great response
H(f)

when the external applied load has

frequency close to its natural
1
0 0.5 1 1.5 2 2.5 3 frequency.

0.1
f/f0
7. Spectrum of Wind Energy and Dynamic Response of
Structure
•32

Displacement of the 3-storey frame structures - Damping Ratio = 2%

m3 y3
1st Storey 2nd storey 3rd storey
4m k3 2.5
m2 y2
2

4m k2 1.5
m1 y1 1

Displacement (mm)
0.5
5m k1
0
0 1 2 3 4 5 6 7
-0.5
-1
6m -1.5
-2
-2.5
Time (sec)

Free oscillation of a 3-storey frame under 2% damping.

7. Spectrum of Wind Energy and Dynamic Response of
Structure
•33

m3 y3
Displacement of the 3-storey frame structures at Resonance -
Damping Ratio = 0%
4m k3
1st Storey 2nd storey 3rd storey
m2 y2
300

4m k2
200
m1 y1

Displacement (mm)
100
5m k1

0
0 1 2 3 4 5

-100
6m
-200

-300

Time (sec)

Exciting frequency = natural frequency, i.e. 𝑓𝑓 = 𝑓𝑓0 and

damping is 0
7. Spectrum of Wind Energy and Dynamic Response of
Structure
•34

Area below is the

variance 𝜎𝜎𝐵𝐵 2
𝑆𝑆𝑀𝑀𝑀𝑀 𝑓𝑓 Dynamic
resonant

Dynamic
background 𝑓𝑓, frequency
× = 𝑆𝑆𝑀𝑀 𝑓𝑓

Area below is the 𝑓𝑓, frequency

variance 𝜎𝜎𝑀𝑀 2 Dynamic
2
background
𝐻𝐻(𝑓𝑓)

𝑓𝑓, frequency
To sum up the responses, an integration over
the frequency is carried out by

� 𝑆𝑆𝑀𝑀 𝑓𝑓 𝑑𝑑𝑑𝑑
7. Spectrum of Wind Energy and Dynamic Response of
Structure
•35

1. So a structure with natural frequency (or

frequencies) close to that of wind with high
energy will have greatest dynamic responses;
2. If the Von Kármán spectrum is referred to, the Range of natural frequencies
frequency of high response are from 0.01Hz to of high-rise buildings
0.1Hz;
3. High-rise buildings will be having natural
frequencies ranging from 0.1Hz to 0.3Hz where
taller buildings generally have lower
frequencies;
4. As low frequency will be closer to wind of
higher energy, taller buildings have greater
dynamic response to wind loads;
5. Means to increase frequency is sometimes
employed to reduce dynamic response such as
increase of stiffness;
6. Another means is to increase damping, e.g.
providing damper.
8. Force Coefficients – Effects due to shape of structure
•36

1. Wind pressure acting on structures of different shapes will give different responses. The responses
are normally given by a factor called “Force coefficient”, with symbol 𝐶𝐶𝑓𝑓 ;

2. Some codes including the Hong Kong 2004 Wind Code distinguish two factors, one due to plan
dimension ratio of the building called 𝐶𝐶𝑠𝑠 and the other height factor due to height breadth ratio 𝐶𝐶ℎ
of the building so that 𝐶𝐶𝑓𝑓 =𝐶𝐶𝑠𝑠 × 𝐶𝐶ℎ ;

3. The factors are mostly found from tests. Some codes including the very old British Code CP3
Chapter V has listed many coefficients;

4. The 2004 Hong Kong Wind Code has given formulae for rectangular and circular buildings while
the 2019 Code has provided formulae but giving very different results.
8. Force Coefficients – Effects due to shape of structure
•37

2004 HK Wind Code

8. Force Coefficients – Effects due to shape of structure
•38

Force coefficient Force coefficient Cf

Cf for Rectangular for Circular
Buildings in Wind Buildings in Wind
Code 2019 Code 2019 is 0.75
(same as 2004
Code) so long
He/D < 6
9. Topographical effects
•39

1. When wind blows up a hill, the speed will be quickened or streamlines getting closer leading to
higher wind pressures;

2. The reverse will take place when wind blows downhill;

3. Different codes provide different formulae or chart for assessing the effects.
9. Topographical effects
•40

Wind Code 2004 for topographical effects, show acceleration and hence greater wind pressures
at lower heights – adoption from BS6399-2 : 1995

𝟐𝟐
𝑺𝑺𝒂𝒂 = 𝟏𝟏 + 𝟏𝟏. 𝟐𝟐𝜶𝜶𝒆𝒆 𝒔𝒔
9. Topographical effects
•41

2019 HKWind Code for topographical effects 𝜓𝜓𝑒𝑒 𝑖𝑖𝑖𝑖 𝑡𝑡𝑡𝑡𝑡 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝑜𝑜𝑜𝑜𝜓𝜓𝑢𝑢 𝑎𝑎𝑎𝑎𝑎𝑎 0.3
10. Sheltering effects
•42

1. Wind load on a building can be sheltered by buildings or other landscape features by which the
wind load is reduced;

2. The 1983 HK Wind Code has distinguished “general terrain” and “built-up terrain” (windward area
with 30% covered by buildings of which at least 5% are higher than 25m). Smaller wind pressures
can be used for built-up terrain;

3. The 2004 HK Wind Code deletes the provision of sheltering effects;

4. The 2019 HK Wind Code re-appended the effects by adopting the approach by BS6399-2:1997 to
account for sheltering effects by reduction of “effective heights”;

5. Full consideration for buildings around is adopted in the 2019 HK Wind Code.
10. Sheltering effects
•43

Building
sheltered by
other buildings
has the wind
profile raised,
or heights of
building taking
wind load
reduced by Hd.

Hd
10. Sheltering effects
•44

Wind Pressures Comparison to Wind

Wind Pressures Comparison to Wind
Code 2019 with and without sheltering
Code 2019 with and without sheltering
on a Building B=D=40m, H=120m
on a Building B=D=40m, H=120m
sheltered by a Building of height 80m
sheltered by a Building of height 80m at
at 30m in front
30m in front
Sheltered No Sheltering
120 120

110 Governed by Hd 110

= 1.2Hi – 0.2Xi
100 and 100

90 Ze = Z - Hd 90

80 80
Actual Height Z (m)

Actual Height Z (m)

70 70

60 60

50 50
Governed by Ze
40 40
= 0.25Z
30 30

20 20

10 10

0 0
0 20 40 60 0 1 2 3 4
Effective Height Ze Wind Pressure (kPa)
11. Wind Induced Acceleration
•45

1. In addition to determination of wind loads on buildings for structural design, the estimation of
“building acceleration” is often required to achieve human comfort for the occupants within
allowable limits imposed in codes of practice. The limits may be some absolute limit (0.15m/s2 for
residential building and 0.25m/s2 for commercial building in the Hong Kong Concrete Code 2013)
or more scientifically, frequency dependent values;

2. The acceleration is the highest at the topmost floor;

3. The acceleration, apart from being dependent on the magnitude of the wind, it is also dependent on
the “natural frequency” of the building;

4. In the simplest derivation, the principles of acceleration estimation is based on the theory of
“Simple Harmonic Motion” (SHM);
11. Wind Induced Acceleration
•46

A review of the simple harmonic motion is shown as follows by considering the sway motion of a point
mass 𝑚𝑚 under a lateral stiffness 𝑘𝑘 as a single degree of freedom problem:

The equation of the sway motion in terms of displacement 𝑥𝑥 can

Mass 𝑚𝑚 be described as
𝑚𝑚𝑥𝑥̈ + 𝑘𝑘𝑘𝑘 = 0
𝑘𝑘
⇒ 𝑥𝑥 = 𝐴𝐴 sin 𝑡𝑡 + 𝑡𝑡0
𝑚𝑚

𝑘𝑘
Stiffness 𝑘𝑘 The value = 𝜔𝜔 is known as the “circular frequency” of the
𝑚𝑚
mass. The frequency 𝑓𝑓 of the motion of the mass is related to 𝜔𝜔 by
𝜔𝜔 = 2𝜋𝜋𝑓𝑓
The frequency 𝑓𝑓 in the unit of hertz is the number of oscillations
per second of the motion.
𝑘𝑘
From 𝑚𝑚𝑥𝑥̈ + 𝑘𝑘𝑘𝑘 = 0, 𝑥𝑥̈ = − 𝑥𝑥 ⇒ 𝑥𝑥̈ = −𝜔𝜔2 𝑥𝑥
𝑚𝑚
So we can determine acceleration from displacement with known
natural frequency of the structure.
11. Wind Induced Acceleration
•47

1. For a building of a continuous length of mass instead of point mass, the following
simulation is often used in wind engineering by which a 1/3 mass of the building is used
and more conveniently related to the base bending moment instead of displacement.

2. However, for building, only the “dynamic resonant” part of the moment or displacement
should be used to estimate the acceleration.

Mass 𝑚𝑚
𝑀𝑀
Consider simple harmonic motion 𝑥𝑥̈ = −
1
𝑚𝑚ℎ
𝑚𝑚𝑥𝑥̈ + 𝑘𝑘𝑘𝑘 = 0; 𝐹𝐹 = 𝑘𝑘𝑘𝑘; 𝑀𝑀 = 𝐹𝐹𝐹 3
Approximation as
h 𝐹𝐹 𝑀𝑀 mass = 1/3 of the
∴ 𝑥𝑥̈ = − = −
𝑚𝑚 𝑚𝑚ℎ total mass

𝑀𝑀

𝐹𝐹
11. Wind Induced Acceleration
•48

For estimation by codes:

In the Hong Kong Wind Code 2004 (based on hourly wind speed pressure)

2 𝐵𝐵
𝑔𝑔𝑓𝑓 2 𝑆𝑆𝑆𝑆
1 + 2𝐼𝐼ℎ 𝑔𝑔𝑣𝑣 + Dynamic
𝜁𝜁
resonant part

In the Hong Kong Wind Code 2019 (based on gust wind speed pressure)
2 0.25
0.5 + 𝑆𝑆𝑠𝑠 𝐿𝐿0.5𝑝𝑝 =𝐵𝐵 +
𝐵𝐵0.5 𝑁𝑁𝑥𝑥 2 𝜉𝜉𝑥𝑥 Dynamic
resonant part
11. Wind Induced Acceleration
•49

1. So if the oncoming wind action (a time or frequency varying function) is known and be
presented in frequency domain (which can be converted from time domain by Fourier
Transform if the original data is in time domain) and be multiplied by the “mechanical
1
admittance” 𝐻𝐻 𝑓𝑓 = to obtain the dynamic resonant response of
2
𝑓𝑓 2 2𝜁𝜁𝑓𝑓 2
1− +
𝑓𝑓0 𝑓𝑓0

displacement or bending moment of the structure for a frequency 𝑓𝑓, the acceleration for
𝑀𝑀
the frequency can be obtained by multiplying 𝜔𝜔2 or 1 ;
𝑚𝑚𝑚
3

2. The total acceleration is the summation of all frequencies together with their associated
magnitudes or energies. But practically only a number of the resonancing frequencies
are used which are the lowest frequencies.
12. Wind Climate
•50

1. The Wind Climate is the probability distribution of wind speed with direction, i.e.
certain threshold wind speeds having different probabilities of being exceeded in
different directions;

2. The Wind Climate is determined by statistical approach on wind speed. It is region

related, being different in different regions;

3. In Hong Kong, stronger winds (based on typhoon) are found in the east and south east
directions;
12. Wind Climate
•51
12. Wind tunnel test
•52

1. For better simulation of the building shape and topography and buildings around, wind
tunnel test is invented;

2. In a wind tunnel, the target building, topography and buildings around are modeled with
reduced scale;

3. Artificial wind simulating the natural wind generated by fan and roughness arrays,
spires blow on the model;

4. Effects on the target buildings are measured (by sensors) and scale up to be the
estimated loads on the target building;
13. Wind tunnel test
•53

5. There are two types of wind tunnel tests for building structures, the topography model
and proximity model;

6. The topography model is a small scale model with scale about 1:4000 around the target
building. This is used to study how the local topography modify the wind from open sea.
That is, the oncoming wind blows to the model is modeled from that of the Code (open
sea condition) and the wind characteristics are measured at the location of the target
building (often with the target building removed). The wind speeds are mostly reduced;

7. The proximity model is of larger scale (scale in the order of 1:400) to study closely the
wind effects on the target building. The applied wind on the building can be taken from
the results of the topography model. If no topography model, the wind profiles from the
Code (open sea condition) should be applied to the proximity model directly.
13. Wind tunnel test
•54

Topography Model
13. Wind tunnel test
•55

Proximity Model
13. Wind tunnel test
•56

1
8. As pressure is proportional to square of velocity (𝑝𝑝 = 𝜌𝜌𝑣𝑣 2 ), force proportional square
2
of linear dimension, the force measured in the wind tunnel has to be scaled up by
1 2 1 2
for determination of the force of the prototype where 𝜆𝜆𝑣𝑣 is the velocity ratio
𝜆𝜆𝑣𝑣 𝜆𝜆𝑙𝑙
between the actual wind speed and the wind speed used in the wind tunnel 𝜆𝜆𝑙𝑙 is the
dimension ratio;

1 2 1 3
9. Likewise a moment measured in the wind tunnel has to be scaled up by for
𝜆𝜆𝑣𝑣 𝜆𝜆𝑙𝑙
determination of moment of the prototype;

10.So a moment of 0.03Nm measured in a wind tunnel of velocity ratio 𝜆𝜆𝑣𝑣 = 1: 10 = 0.1 and
𝜆𝜆𝑙𝑙 = 1: 500 = 0.002 represents a moment of
0.03 × 102 × 5003 = 375,000,000Nm = 375,000kNm in the prototype.
13. Wind tunnel test
•57

Pressure Tapes for Wind Tunnel Test High Frequency Base Balance HFBB
•58

The End
Thank you