WIND ISSUES IN THE DESIGN OF TALL BUILDINGS

Peter A. Irwin RWDI

Los Angeles Tall Building Structural Design Council May 7, 2010

Wind Issues for Structural Design

Structural integrity under ultimate loads Deflections under service loads Building motions and occupant comfort Uncertainties in building structural properties (stiffness, damping) Uncertainties in wind loading Uncertainties in wind climate Codes and standards Computational Fluid Dynamics

Relationship between importance of wind and height

Importance of wind loads

Building height

Vortex shedding
Shedding frequency N is given by

U N =S b

S = Strouhal number U = wind speed b = building width

Directions of fluctuating force

wind

Peak response due to vortex excitation

Magnitude of peak response

1 density damping

Vortex excitation on a tapered spire Mode 1

Vortex excitation on a tapered spire Mode 2

Shape strategies to reduce excitation

Softened corners Tapering and setbacks Varying cross-section shape Spoilers Porosity or openings

Taper effect - Petronas towers

Taipei 101 corner softening

Original

Modified

25% REDUCTION IN BASE MOMENT

Shape effect, Taipei 101 wind tunnel tests

Burj Khalifa 828 m

Set backs, changing cross-section, orientation Completed Building Early 1:500 scale wind tunnel tests

Shanghai Center 632 m

Twisting and tapering

151 storey tower in Korea

Creation of bleed slots at edges to suppress vortex shedding Full scale rendering 1:500 wind tunnel model

Use of corner slots to bleed air through corners on a tall building

Plan view without slots Plan view with slots

Crosswind motion

Base moments reduced by 60%

Wind

Wind

With vortex excitation

10 yr
Without vortex excitation

700 yr 50 yr

(a)

(b)

With vortex excitation

50 yr 700 yr

Without vortex excitation

(c)

(d)

where where

Reliability considerations for flexible buildings

Expression for load factor

1 2 Vw W = e Kw
Vw =
Coefficient of variation of wind load Reliability index Bias factor Combination factor

=
Kw =

First order, second moment reliability analysis for rigid buildings

Code analysis method
2 2 Vw = (Vqi + Van )1 / 2 = (0.2 2 + 0.32 )1 / 2 = 0.36

1 2 Vw 1 ( 0.752 3.50.36 ) W = e = e = 1.56 Kw 1 .3

Wind tunnel method

Vw = (0.22 + 0.12 )1 / 2 = 0.22 1 ( 0.752 3.50.22 ) W = e = 1.54 1 .0

Both the coefficient of variation and bias factor are smaller in the wind tunnel method. Load factor ends up at about the same.

Reliability of flexible buildings with monotonic response

Wind load varies as power law

W U
Rigid building

2 2 Vw = (Vqi + Van )1 / 2 Flexible building 2 2 2 Vw = (Vqi + Van + Vdyn )1 / 2

2 Vdyn ndampV2 + ( n 2) 2V f2 + ( n 2) 2VV2

Damping term

Frequency term

Velocity sensitivity term

Required load factors for flexible buildings with monotonic response

Rigid building 2 2 Vw = (Vqi + Van )1 / 2 = (0.2 2 + .12 )1 / 2 = 0.22

W = e

( 0.752 3.50.22 )

= 1.54

Typical flexible building

2 2 2 Vw = (Vqi + Van + Vdyn )1 / 2 = (0.2 2 + 0.12 + 0.112 )1 / 2 = 0.249

W = e

( 0.752 3.50.249 )

= 1.63

Highly flexible building

2 2 2 Vw = (Vqi + Van + Vdyn )1 / 2 = (0.2 2 + 0.12 + 0.22 2 )1 / 2 = 0.31

W = e

( 0.752 3.50.31)

= 1.84

Note:- Wind tunnel method is assumed in all cases.

Wind loads determined directly at ultimate return period by wind tunnel method
Building type Rigid Flexible Very Flexible n 2 2.5 3 Required Actual ndamp Load Factor Load Factor Ratio 1.264n 0 1.54 1.60 1.04 0.25 1.63 1.80 1.10 0.5 1.84 2.02 1.10

Slightly conservative

Consequence of vortex shedding on reliability assessment Nakheel Tower

Low ratio of 1000 year to 50 year loads Uncertainty in loads is mostly dictated by uncertainty in this peak

Need to carefully assess uncertainty in peak vortex shedding response since the normal assumption that uncertainty in wind speed governs is no longer valid. Damping and frequency uncertainties become important.

Reynolds number effects

Originate from viscosity of air Can cause changes in flow patterns on a small scale model relative to full scale Not a concern on sharp edged structures Can be a concern on curved shape buildings Is lessened by high turbulence and surface roughness

Effect of Reynolds number on pressure coefficient around a circular cylinder

1.0

U
0.5

b
140

degrees
0 20 60 100

Cp

-0.5

Reynolds number
-1.0

Re =

Ub

-1.5

-2.0

Sub-critical Re < 2105 Kinematic viscosity of air

-2.5

Transcritical Re > 3106

1:50 Scale Model of Upper Portion of Burj Khalifa High Reynolds Number Tests, Re ~ 2x106

Note: Full scale Re~7x107

Comparison of Mean Pressure Coefficients at Low and High Reynolds Number - Burj Dubai Pressure Taps

High Reynolds number tests on Shanghai Center

Example of Mean Pressure Coefficients

Pressure tap count

Damping and dynamic response beyond the elastic limit

F2 F1 k2 k1

eff

1 E1/ 2 = 2 E

x 1 x2

= ( x2 x1 ) / x1
k 2 1 k2 2 1 + k1 2 k1 1 = k 1 + 2 + 2 2 k1

eff

Effect of stiffness reductions and inelastic deflections on effective viscous damping ratio

eff

1 E1/ 2 = 2 E

k2/k1=0.5

Example: For 20% deflection beyond elastic limit effective increment in damping ratio = 0.024. This would be additive to the damping ratio below the elastic limit which is typically assumed to be 0.01 to 0.02.

k2/k1=0.75

( x2 x1 ) / x1

Damping and inelastic response research

Non-linear time domain analysis of structures under realistic wind loading will allow us to evaluate whether more efficient structures can be developed. How feasible is it to load the structure beyond its elastic limit and how far can one go with this? Can we learn how to keep the structure stable and robust after going plastic. What can be learnt from earthquake engineering? More full scale monitoring needed at representative deflections.

Building motions and criteria

Problem is complex due to variability amongst people What return period should be used? What quantity best encapsulates comfort: acceleration; jerk; something in between; angular velocity; noises combined with motion, etc? Designers have to make decisions and move on. What is the actual experience using traditional approaches?

Table 1: Table 1:

Building Motion Criteria. Historical review of 19 buildings wind tunnel tested by RWDI in the 1980s and 1990s.
First Order Modes Frequencies (Hz) Building Height (m) 1* 2 3 4 5 6 7 8 9 10 11* 12* 13 14 15 16 17 18* 19* 249.4 163.4 198.1 137.2 236.2 178.0 215.0 110.9 163.0 124.4 207.8 145.2 94.0 141.2 143.3 259.4 175.4 247.8 259.4 fx 0.193 0.176 0.189 0.185 0.154 0.244 0.177 0.192 0.195 0.170 0.147 0.411 0.278 0.370 0.135 0.131 0.201 0.131 0.179 fy 0.199 0.224 0.184 0.135 0.169 0.250 0.149 0.164 0.208 0.224 0.171 0.213 0.243 0.216 0.180 0.125 0.157 0.154 0.159 ftor 0.295 0.250 0.300 0.323 0.400 0.400 0.331 0.250 0.224 0.204 0.250 0.440 0.139 0.356 0.333 0.263 0.200 0.211 0.236 Damping Ratio (% of critical) 2.0 2.0 1.5 2.0 2.0 2.0 1.5 2.0 1.5 2.0 1.5 2.0 2.0 2.0 2.0 1.25 1.5 2.0 2.0 Assumed Building Number

* For these buildings, wind tunnel studies predicted peak resultant accelerations above the 15 18 milli-g range.

Summary of computed frequencies of 19 buildings

0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 50 100 150 200 250 300
Buil di ng Hei ght , H ( m )

fx

fy

f = 33/H Fitting C urve

Summary of Predicted and Improved Acceleration Responses

30 25 20 15 10 5 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Building # 15-18 milli-g range accelerations Reduced acceleration responses using SDS Above 15-18 milli-g range accelerations

Acceleration ( milli - g )

Buildings with Supplemental Damping System

Building Name and Location Building Number SDS Type Installed

10-Year Peak Resultant Acceleration ( milli g ) Without SDS With SDS 16.6 15.4 15.2 15.5

1 11 12 18

Park Tower, Chicago, IL Random House, New York, NY Wall Centre, Vancouver, BC Bloomberg Tower, New York, NY Trump World Tower, New York, NY

TMD TLCD TLCD TMD

24.0 25.3 28.0 22.5

19

TMD

27.4

17.3

Park Tower, Random House and Wall Center with Damping Systems
Chicago New York

Vancouver

Trump world Tower and Bloomberg Center with Damping Systems

New York

New York

Chicago Spire Very Long period Higher Mode Effects

Higher modes can affect response ~ questions re frequency dependent motion criteria such as ISO comfort criteria.
700 600 500 400

2 nd harmonic

1st harmonic

Height, m

300 200 100 0 -1.5 -1 -0.5 0 0.5 1 1.5

Mode deflection shape

Chicago spire - motion and deflection control through use of damping system.

Assessing building motions.

25 Peak acceleration, milli-g

ISO Criteria for single frequency Moving room simulations of multiple frequencies

20 15 10 5 0 0.1 Frequency, Hz

Co Resid entia mmercial l

Meso-scale Modelling of June 1988 Event

Thank you