Structures congress 2017 buildings and special structures

ACROSS-WIND BUILDING MOTION WITH AERODYNAMIC DAMPING The dynamic equation of motion for a tall building under wind excitation can be expressed as follows in modal domain in consideratio

Buildings and Special Structures

Structures Congress 2017

EditEd by

J G (Greg) Soules, P.E., S.E., P.Eng

Selected Papers from the Structures Congress 2017

denver, Colorado April 6–8, 2017

SPONSORED BY The Structural Engineering Institute (SEI)

of the American Society of Civil Engineers

EDITED BY

J G (Greg) Soules, P.E., S.E., P.Eng

Published by the American Society of Civil Engineers

Published by American Society of Civil Engineers

1801 Alexander Bell Drive Reston, Virginia, 20191-4382 www.asce.org/publications | ascelibrary.org

Any statements expressed in these materials are those of the individual authors and do not necessarily represent the views of ASCE, which takes no responsibility for any statement made herein No reference made in this publication to any specific method, product, process,

or service constitutes or implies an endorsement, recommendation, or warranty thereof by ASCE The materials are for general information only and do not represent a standard of ASCE, nor are they intended as a reference in purchase specifications, contracts, regulations, statutes, or any other legal document ASCE makes no representation or warranty of any kind, whether express or implied, concerning the accuracy, completeness, suitability, or utility of any information, apparatus, product, or process discussed in this publication, and assumes no liability therefor The information contained in these materials should not be used without first securing competent advice with respect to its suitability for any general or specific application Anyone utilizing such information assumes all liability arising from such use, including but not limited to infringement of any patent or patents

ASCE and American Society of Civil Engineers—Registered in U.S Patent and Trademark Office

Photocopies and permissions Permission to photocopy or reproduce material from ASCE

publications can be requested by sending an e-mail to permissions@asce.org or by locating a title in ASCE's Civil Engineering Database (http://cedb.asce.org) or ASCE Library

(http://ascelibrary.org) and using the “Permissions” link

Errata: Errata, if any, can be found at https://doi.org/10.1061/9780784480410 Copyright © 2017 by the American Society of Civil Engineers

All Rights Reserved

ISBN 978-0-7844-8041-0 (PDF) Manufactured in the United States of America

Preface

The Structures Congress has a robust technical program focusing on topics important to Structural Engineers

The papers in the proceeding are organized in 4 volumes

Volume 1 includes papers on Blast and Impact Loading and Response of Structures Volume 2 includes papers on Bridges and Transportation Structures

Volume 3 includes papers on Buildings and Nonbuilding and Special Structures Volume 4 includes papers on Other Structural Engineering Topics including; Business and Professional Practice, Natural Disasters, Nonstructural Systems and Components, Education, Research, and Forensics

Acknowledgments

Preparation for the Structures Congress required significant time and effort from the members of the National Technical Program Committee, the Local Planning Committee Much of the success of the conference reflects the dedication and hard work

by these volunteers

We would like to thank GEICO and Pearl for Sponsoring the Congress proceedings and supporting the Structures Congress in such a generous way

The Joint Program Committee would like to acknowledge the critical support of the sponsors, exhibitors, presenters, and moderators who contributed to the success of the conference through their participation

On behalf of our dedicated volunteers and staff, we would like to thank you for spending your valuable time attending the Structures Congress It is our hope that you and your colleagues will benefit greatly from the information provided, learn things you can implement and make professional connections that last for years

Eliminating the Exposure Category from Wind Design Pressure 13

Nicole Ellison and Frederick R Rutz

Wind Load Prediction on Tall Buildings in a Stochastic Framework 24

M Gibbons, J Galsworthy, M Chatten, and S Kala

Experimental Investigation of Deconstructable Steel-Concrete Shear Connections in Sustainable Composite Beams 34

Lizhong Wang, Mark D Webster, and Jerome F Hajjar

Influence of Fastener Spacing on the Slip Modulus between Cold Formed Steel and Wood Sheathing 48

Weston Loehr, Bill Zhang, Hani Melhem, and Kimberly Krammer

BRBM Frames: An Improved Approach to Seismic-Resistant Design Using Buckling-Restrained Braces 60

Leo Panian, Nick Bucci, and Steven Tipping

Implications of Modeling Assumptions on the Loss Estimation for Shear Wall Buildings 72

Kristijan Kolozvari, Vesna Terzic, and Daniel Saldana

Numerical Investigation of the Shear Buckling and Post-Buckling of Thin Steel Plates with FRP Strengthening 87

Mohamad Alipour, Alireza Rahai, and Devin K Harris

Seismic Evaluation of Incremental Seismic Retrofitting Techniques for Typical Peruvian Schools 101

Gustavo Loa, Alejandro Muñoz, and Sandra Santa-Cruz

Advanced Technical Issues Related to Wind Loading on Tall Building Structures in Consideration of Performance-Based Design 111

U Y Jeong and K Tarrant

© ASCE

ASCE 41-17 Steel Column Modeling and Acceptance Criteria 121

Daniel Bech, Jonas Houston, and Bill Tremayne

Leveraging Cloud and Parametric Workflows to Accelerate Performance Based Seismic Design 136

Kermin Chok, Pavel Tomek, Trent Clifton, and Branden Dong

Stability of Steel Columns in Steel Concentrically Braced Frames Subjected to Seismic Loading 143

Guillaume Toutant, Yasaman Balazadeh Minouei, Ali Imanpour, Sanda Koboevic, and Robert Tremblay

Classifying Cyclic Buckling Modes of Steel Wide-Flange Columns under Cyclic Loading 155

Gulen Ozkula, John Harris, and Chia-Ming Uang

Structural Behaviour of Demountable HSS Semi-Rigid Composite Joints with Precast Concrete Slabs 168

Abdolreza Ataei, Mark A Bradford, and Hamid R Valipour

Topology and Sizing Optimization of Nonlinear Viscous Dampers for the Minimum-Cost Seismic Retrofitting of 3-D Frame Structures 179

Nicolò Pollini, Oren Lavan, and Oded Amir

Structural Topology Optimization Considering Complexity 192

Saranthip Koh, May Thu Nwe Nwe, Payam Bahrami, Fodil Fadli, Cristopher D Moen, and James K Guest

Cast Steel Replaceable Modular Links for Eccentrically Braced Frames 202

J Binder, M Gray, C Christopoulos, and C de Oliveira

New Methods in Efficient Post-Tensioned Slab Design Using Topology Optimization 213

M Sarkisian, E Long, A Beghini, R Garai, D Shook, A Diaz, and R Henoch

Design and Parametric Finite Element Analysis—A Thin Lightweight Two-Way Steel Flooring System 225

Eugene Boadi-Danquah, Brian Robertson, and Matthew Fadden

Structural Form Finding of a Rope Sculpture 237

M Sarkisian, E Long, A Beghini, and N Wang

Discussion of Tubular Steel Monopole Base Connections:

The Base Weld Toe Crack Phenomenon;

Crack Identification and a Proposed Severity Classification System 248

Brian R Reese and David W Hawkins

© ASCE

Design and Theory of Passive Eddy Current Dampers in Building Structures 262

Mandy Chen and Lance Manuel

Effect of Damaged Fireproofing on the Behavior of Structural Steel Members 275

Ataollah Taghipour Anvari, Mustafa Mahamid, and Michael J McNallan

A Re-Evaluation of f’ m—Unit Strength Method, Face Shell, and Fully Bedded Mortar Joints 287

N Westin and M Mahamid

Parametric Study and Design Procedure for Skewed Extended Shear Tab Connections 301

Mutaz Al Hijaj and Mustafa Mahamid

Scaffolding a Landmark: The Restoration of the Dome of the United States Capitol Building 319

Christopher P Pinto and Joelle K Nelson

Achieving Column-Free Platforms—Design and Construction of Large Span Station Mezzanines on the Second Avenue Subway Project 329

Renée Grigson and Michael Voorwinde

Evaluation of Full-Scale Adobe Brick Walls under Uniform Pressure 343

S Robert, H El-Emam, A Saucier, H Salim, and Scott Bade

Experimental Study of Externally Flange Bonded CFRP for Retrofitting Beam-Column Joints with High Concrete Compressive Strength 354

Olaniyi Arowojolu, Muhammad Kalimur Rahman, Baluch Muhammad Hussain, and Ali-Al Gadhib

Considerations in the Use of Side Load Pier Brackets 365

James Robert Harris and Kenneth Cobb

Retrofitting of Flange Notched Wood I-Joists with Glass Fiber Reinforced Polymer (GFRP) Plates 375

M Shahidul Islam and M Shahria Alam

Multiple Hazards and Social Vulnerability for the Denver Region 386

A Rein Starrett and R B Corotis

A Top Down Approach to Achieve Full System Modeling in Seismic Analysis and Design 406

F A Charney

© ASCE

Experimental and Numerical Investigation of Flexural Concrete Wall Design Details 418

A Behrouzi, T Welt, D Lehman, L Lowes, J LaFave, and D Kuchma

Seismic Response Study of Degraded Viscous Damping Systems for Tall Buildings in China 434

H Ataei, M Mamaghani, and K Kalbasi Anaraki

Topology Optimization and Performance-Based Design of Tall Buildings:

A Spatial Framework 447

Xihaier Luo, Arthriya Suksuwan, Seymour M J Spence, and Ahsan Kareem

Effects of Foundation Uplift on the Dynamic Response of Steel Frames 459

Mohammad Salehi, Amir Hossein Jafarieh, and Mohammad Ali Ghannad

Performance-Based Wind and Seismic Engineering:

Benefits of Considering Multiple Hazards 473

Kevin Aswegan, Russell Larsen, Ron Klemencic, John Hooper, and Jeremy Hasselbauer

Effect of Drift Loading History on the Collapse Capacity of Deep Steel Columns 485

T.-Y Wu, S El-Tawil, and J McCormick

Properties of and Applications with Full Locked Coil Rope Assemblies 495

K.-J Thiem and M Bechtold

U.S Bank Stadium: Transparent Roof Steel Collaboration 503

R John Aniol, Rick Torborg, and Eric Fielder

Advanced Analysis of Steel-Frame Buildings for Full Story Fires 515

Erica C Fischer and Amit H Varma

Integrated Fire-Structure Simulation Methodology for Predicting the Behavior of Structures in Realistic Fires 527

Chao Zhang

Structural Design, Approval, and Monitoring of a UBC Tall Wood Building 541

T Tannert and M Moudgil

Adaptive Reuse of the Historical Ferdinand Building, Boston, MA 548

The New Tocumen International Airport South Terminal in Panama City, Panama 570

Andrea Soligon, Jeng Neo, and Xiaonian Duan

Multi-Hazard Design of a New Emergency Communications Facility in

St Louis, Missouri 582

Nathan C Gould, Richard Hoehne, and Michael Shea

Prison Design in Haiti: Structural Challenges 592

David Dunkman, Christopher Hewitt, and Scott Hollingsworth

Underpinning Historic Structures at Grand Central Station, New York 604

Yazdan Majdi and Richard Giffen

Design of an Underground Viaduct for the Expansion of the Moscone Center 614

A Trgovcich, L Panian, and S Tipping

Nonbuilding and Special Structures

Extreme Wave Monitoring and In Situ Wave Pressure Measurement for the Cofferdam Construction of the Pingtan Strait Bridge 629

Zilong Ti, Shunquan Qin, Yongle Li, Dapeng Mei, and Kai Wei

What We Learned from the Cooling Tower Foundation Design Challenges from a Revamp Project 643

Silky Wong and Abhijeet Yesare

Design of Industrial Pipe Racks Using Modules, Pre-Assembled Units, and Stick-Built Construction 653

Xiapin Hua, Ron Mase, Khoi Ly, and Jkumar Gopalarathnam

Ship Impact and Nonlinear Dynamic Collapse Analysis of a Single Well Observation Platform 668

Ahmed Khalil, Huda Helmy, Hatem Tageldin, and Hamed Salem

Pile Cap Seismic Load Transfer to Soil 681

Eric Wey, Rollins Brown, Candice Kou, and C B Crouse

Constructability Solutions for Temporarily Supporting 200’ Flare Stacks during Construction Modifications 693

Mateusz Prusak, Nicholas Triandafilou, Mustafa Mahamid, and Tom Brindley

Custom Helical Pile Use for a Refinery Revamp: A Case Study 706

Eric Wey, Patrick Murray, Howard Perko, Malone Mondoy, and Paul Volpe

© ASCE

Structural Fatigue of Process Plant Modules during Ocean Transport 721

Alan Shive and Marco Camacho

Innovative Use of FRP in Large-Diameter Piles for Vessel Impact 735

M A McCarty, V Zanjani, E Grimnes, and J Marquis

Seismic Analysis and Design for Wine Barrel Storage Racks 745

Tauras Stockus and Tzong-Ying Hao

Seismic Analysis and Design of a 21,000-Gallon Frac Tank Considering the Fluid-Structure Interaction Effects for a FLEX Response at a Nuclear Power Station 758

Christine H Roy and Michael Mudlock

Seismic Behavior of Cylindrical Fluid-Filled Steel Tanks 772

Erica C Fischer and Judy Liu

A Comparison of Approximate Methods for Period Determination in Rack Structures 782

Andrew Hardyniec, Charles DeVore, and Jeffrey Travis

© ASCE

Nonlinear Dynamic Analysis of Multi-Sloshing Mode Tuned Liquid

Sloshing Dampers Installed in Tall Buildings

INTRODUCTION

Tall buildings often experience excessive building motions, which makes the serviceability of buildings in terms of accelerations and torsional velocities exceed the acceptable range according to the industrial guidelines (ISO, 1984; Isyumov, 1993, 1995) To mitigate the excessive motions, Tuned Liquid Sloshing Dampers (TLSDs) have been frequently used due to their low cost, simple frequency tuning, and low maintenance (Kareem, 1987, 1990; Fujino, 1995; Warnitchai, 1997; Tait, 2004, 2008;

Tait et al., 2004a, 2004b) The sloshing motion of the liquid in a tank can be expressed as a combination of infinite sloshing modes based on potential flow theory

in consideration of wall and free surface boundary conditions (Baucer, 1984;

Warnitchai, 1997; Tait, 2004, 2008) In order to dissipate building’s vibration energy,

© ASCE

TLSDs are usually equipped with porous screens (Fediw et al., 1995; Warnitchai, 1997) immersed in the liquid The damping force created by the porous screens are a nonlinear function of water velocity flowing through the screens, which makes it difficult to calculate the accurate damping force

Previous studies consider the nonlinear screen damping force as approximate linear form based on the assumption of the probabilistic distribution of the excitation

as sinusoidal or random signals (Caughey, 1963) However, since the wind-induced excitation of the motion is not the same as the sinusoidal or random signals, for more accurate modeling of the sloshing tank, nonlinear effects should be considered

Kaneko and Ishikawa (1999) considered the nonlinear screen damping force in their iterative solution of potential flow over the liquid domain based on Finite Difference Method Their method produces implicit solution of liquid sloshing motion requiring discretization of liquid domain which makes the solution process a bit complicated and time consuming

In this paper, nonlinear formulas are derived to solve dynamic sloshing motion of multi-sloshing mode a TLSD installed on a tall building in consideration of the nonlinear screen damping force To implement present method to the wind-induced tall building vibration problem, the nonlinear TLSD model is coupled with tall slender building immersed in strong winds Here, for more accurate analysis of wind-induced vibration of tall slender structure, aerodynamic damping and its time-domain formulas are also derived based on Rational Function Approximation (RFA)

(Fujino et al., 1995)

ACROSS-WIND BUILDING MOTION WITH AERODYNAMIC DAMPING

The dynamic equation of motion for a tall building under wind excitation can

be expressed as follows in modal domain in consideration of aerodynamic damping force, ~f ae:

)

~

~(

~

~

~2

ae s

velocity respectively of the mode; m~= modal mass; f~= modal wind load

Aerodynamic damping of tall buildings for across-wind excitation can be measured from forced- or free-vibration response of a pivoting model in the wind tunnel (Steckley, 1989; Watanabe et al, 1997; Katagiri et al., 2000) The aerodynamic damping depends on building geometries, exposures, aspect ratios, side ratios and amplitudes of vibration The self-excited force in terms of aerodynamic damping and

© ASCE

stiffness can be expressed as follows in generalized coordinate by aerodynamic force measurements obtained from forced vibration tests (Steckley, 1989):

x i m

β+αηω

i ; B= building width; H = building height

BUILDING COUPLED WITH MULTI-SLOSHING MODE TLSD

Figures 1(a) to 1(d) show schematic diagrams of a building coupled with simplified linear single-sloshing mode TLSD (Figure 1(a)) versus nonlinear multi-sloshing mode TLSD (Figure 1(c)), and their corresponding equivalent TMD (Tuned Mass Damper) modeling in Figures 1(b) and 1(d) respectively Here, it is noted that the Figure 1(d) illustrates the aerodynamic forces, ~f ae, and the non-conservative damping force, Q n

For a damper with dimensions of b, h, and L corresponding to width, water height and length, the equation of motion of the building-TLSD coupled system under wind loading can be represented as follows based on equilibrium conditions of the equivalent MTMD model in Figure 1(d):



=

=ω+ωξ

n

n r n eq w

ae s

s

m

m x

m

bhL f

f m x x x

1

, , 1

2

~

~

~)

~

~(

~

~

~2

x x

x x

xr n eq n r n r n eq2 n r,n ~

, , , ,

In the above, ρw = liquid density; the second term in (4) corresponds to the conservative damping force, Q , of sloshing mode n, illustrated in Figure 1(d); n m eq,n

non-is the effective mass of the nth sloshing mode defined below; x r,n , x and r,n xr,n

represent relative displacement between the motion of the equivalent mass and the building (see Figure 1), and its first and second-order time derivatives respectively for the nth liquid sloshing mode

Furthermore, the noted parameters are defined as follows based on explicit solution of Laplace equation defining the liquid in rectangular tank based on potential flow theory (Warnitchai, 1997; Tait, 2008):

)tanh(

)(

))cos(

1(2

3

2 2

,

L

h n n

n bL

π

π

−ρ

© ASCE

n n l n

L

h n L

n

n

Ξ Δ

π π

L

h n

L

x n

1

)tanh(

,

L

h n L

g n

n

In the above, x indicate the location of the screen in the x-coordinate; ns denotes the j

number of screen in the tank The amplitude of sloshing height of nth sloshing mode,

n

displacement, x r,n:

n r n

)tanh(

))cos(

1(2

L

h n n

n

n

ππ

TIME DOMAIN FORMULA OF AERODYNAMIC DAMPING

Formulas for the tall building-TLSD coupled problem are derived in domain to directly calculate the peak factors and to take advantage of the other benefits described before Among the governing equations for the time-domain analysis, the frequency-dependent aerodynamic damping and stiffness terms have to

time-be converted to terms solvable in the time-domain The Rational Function Approximation (RFA) method used originally in aeronautical engineering can be used

to represent the frequency-dependent aerodynamic damping and stiffness terms in equation (2) and convert them to the following equations with constant coefficients which can be solved in time-domain (Jeong, 2014):



=

j j H H

H

B

U x a B

U x a B

U x a f

m

1 2

2 1

2 3

1

2

~ 2

~ 2

~ 2

~

x B

U d i

a i y

H j

j

+ ω

where a for j=1 to m+3 represent rational function coefficients; j d are damping j

terms for j=1 to m; m denotes number of rational function terms The coefficients d j

should be selected to best fit the function

TIME DOMAIN BUILDING – NONLINEAR MULTI-SLOSHING MODE DAMPER COUPLED SYSTEM ANALYSIS

The governing equations for the nonlinear time-domain analysis building motion coupled with TLSD under wind-excitation can be represented as follows by combining the dynamic equation of motion for the building coupled with multi-sloshing mode TLSD in equations (3) and (4) as well as the auxiliary equations from RFA in (12) and (13):

;

~ 2

~

~ ) 2

(

~ ) 2

2 (

~ ) 2

~ 1 (

1 1

2 2 1

, ,

1 2

2 2

2 3

f m y B

U x

m m

x a B

U x

a B

U x

a m

bhL

m

j j H n

n

n r n eq

H s

H s

+ +

η + ω + η

+ ω ξ + η + ρ +

B

U d x a

In order to solve the coupled nonlinear dynamic equations in (14) to (16), Newmark Beta step-by-step integration method is used after deriving incremental formula including the nonlinear term in (15) for the iterative calculation The solutions are calculated until converged by the iterations for each time step

EXAMPLE - ACROSS-WIND RESPONSE OF COUPLED TALL NONLINEAR MULTI-MODE TLSD

BUILDING-A tall building with building height, H, equals to 300 m, with plan dimension

B and D of 20 m by 20 m respectively is analyzed under mean hourly wind speed of 33.55 m/s defined at the top of the building height above the grade in open exposure which corresponds to 10-year wind speed in Seattle, Washington The exposure is assumed to be open exposure with mean wind speed exponent, γ , of 0.14; turbulence intensity at the building height corresponds to approximately 11 % and 10 % based on ASCE (2010), ESDU respectively The mass density of the building is 225 kg / m3

and the mass is assumed to be distributed uniformly along the height of the building

Structural damping ratio, ξ , equals to 0.15 The mode shape exponent1 μ is 1.5 which

is a typical value for tall buildings

© ASCE

Across-wind load spectrum in AIJ (2006) has been used in this example due

to its simplicity and versatility covering different side ratios, although the applicable aspect ratios of the building in AIJ are limited below 6 Figure 2 illustrates across-wind spectrum used in the example The peak value of the spectra due to the vortex shedding occurs around 0.15 Hz which corresponds to the reduced frequency (=

Steckley’s (1989) motion-induced aerodynamic damping and stiffness under pivoting excitation are used in the analysis Since the amplitude-dependent nonlinear characteristic of the aerodynamic damping (Chen, 2013, 2014) is not considered in this study, the aerodynamic damping measured under small-amplitude is used Figure

3 illustrates the coefficients of aerodynamic impedance (α+iβ ) which represents aerodynamic damping and stiffness of a building with a square-shaped building plan with aspect ratio of 13.3 measured under nominal turbulence intensities of 17 % As shown in the figure, the aerodynamic damping in terms of β reduces as frequency reduce and falls below zero which means negative aerodynamic damping where the reduced frequency, K, is lower than 0.7 The figure also illustrates the fitted curves for αand β using rational functions for the time-domain analysis

Equations in (14) to (16) are analyzed both in frequency-domain and domain for the verification with building frequencies varying from 0.08 Hz to 0.3 Hz

time-Since excessive building motions are expected due to the slenderness of the building, TLSDs are optimally designed with various dimensions to be tuned to different building frequencies, to provide 3% additional damping to the building The effective mass ratio of the damper (=m eq / m~1m /mI ) is approximately 1.4 %; and the TLSD

is optimally tuned to the structural frequencies The porous screens are also designed

to provide optimal damping to the damper based on random sloshing motion The TLSD-building system has been analysis using the equations (14) to (16) in consideration of aerodynamic damping for 17% Turbulence Intensity

Frequency-domain analysis and time domain analysis of the building-TLSD coupled system is performed for the structural frequencies varying 0.08 to 0.3 Hz using the generated wind load time series and the results are plotted in Figure 4 As

© ASCE

shown in the figure; around the region where the building frequency matches with the vortex-shedding frequency of 0.15 Hz, the amplitude of vibration reaches its maximum The aerodynamic damping effects are favorable for the frequencies higher than the vortex shedding frequency, therefore, the response reduced with AD, whereas the amplitude has increased for the frequencies lower than the vortex shedding frequency due to the unfavorable negative aerodynamic damping

With optimally designed dampers installed on the building, the response drastically reduces and the system becomes more stable even under the negative AD

at low building frequency Linear Time-domain analysis results show excellent agreements with those from the Frequency-domain analysis However, due to the system instability introduced by RFA, the result for 0.8Hz is deviated from that of the Frequency-domain analysis Nonlinear time-domain analysis results compared to those from the linear analysis have increased for high frequency region, due to the nonlinear damping force effect However, the nonlinear analysis results fall very close to the linear analysis results for low building frequencies

Figure 5 shows a time history of the average liquid pressure on screens which represents the non-conservatory damping force during the nonlinear time domain analysis The blue line represents the non-conservatory damping force per unit area (=

Q n/bh ), whereas the red line indicates directly calculated liquid pressure using the liquid velocity and loss coefficient, C , of the screen As shown in the figure, the l

non-conservatory damping force is accurately calculated based on the present method

Figure 6(a) and 6(b) illustrate liquid sloshing motions at two instantaneous times during the nonlinear time-domain analysis of building with building frequency

of 1.0Hz Whereas Figure 6(a) represents when the water sloshing height reaches its maximum on the left wall in the figure and the motion is governed by the first anti-symmetric sloshing mode, Figure 6(b) represents the moment when the water sloshing motion is contributed by multiple modes which is considered in this study

CONCLUSIONS

New formulas are derived for a nonlinear dynamic analysis of multi-sloshing TLSD In order to apply the method to wind design of tall buildings, required formulas for motion-induced aerodynamic damping and stiffness are also derived both in time- and frequency-domain The TLSD is modeled as multiple equivalent TMDs equipped with nonlinear damping force representing the non-conservatory damping force created by the screen immersed in sloshing liquid

Explicit solution for the sloshing motion of liquid in a rectangular tank is derived in generalized coordinates representing liquid sloshing motion based on potential flow theory The non-conservatory nonlinear damping force are derived for each mode for the time-domain analysis A verification of present method is

© ASCE

attempted through an example of a typical tall slender building From the analysis, the nonlinear screen damping force effects are investigated which have increased responses for the building with relatively high frequency

The proposed time domain approach will enable more accurate evaluation of wind response of tall buildings, more accurate design of TLSD reducing the expensive dynamic damper testing, evaluating non-Gaussian processes such as real peak factors, which will be very useful in tall building design

REFERENCES

American Society of Civil Engineers (ASCE) (2010), Minimum design loads for

buildings and other structures, ASCE Standard ASCE/SEI 7-10

Architectural Institute of Japan (AIJ) (2006), Recommendations for loads on

Chen, X (2013) “Estimation of stochastic crosswind response of wind-excited tall

buildings with nonlinear aerodynamic damping,” Engineering Structures, 56,

766-778

Chen, X (2014) “Extreme value distribution and peak factor of crosswind response of

flexible structures with nonlinear aeroelastic effect,”, J Struct Eng., ASCE,

online 0401491, 1-18

Deodatis, G (1996) “Simulation of ergodic multivariate stochastic processes,” J of

Eng Mech., ASCE, 122(8) 778-787

Engineering Standard Data Unit (ESDU) Wind Engineering Sub-Series, ESDU

International Inc

Fediw, A.A., Isyumov, N and Vickery, B.J (1995) “Performance of a tuned sloshing

water damper,” Journal of Wind Engineering and Industrial Aerodynamics, 57,

International Standards Organization (ISO) 6897-1984: “Guidelines for the evaluation

of the response of occupants of fixed structures, especially buildings and shore structures, to low frequency horizontal motion (0.063 to 1 Hertz).”

off-Isyumov, N (1993) “Criteria for acceptable wind-induced motions of tall buildings,”

Proceedings of the International Conference on Tall Buildings, Council on Tall Buildings and Urban Habitat, Rio de Janerio, Brazil

© ASCE

Kaneko, S., Ishikawa, M (1999) “Modeling of Tuned Liquid Damper with

Submerged Nets,” Transactions of ASME, 121, 334-343

Kareem, A and Sun, W.J (1987) “Stochastic response of structures with

fluid-containing appendages,” J of Sound and Vibration, 119(3)

Kareem, A (1990) “Reduction of wind induced motion utilizing a tuned sloshing

damper,” J of Wind Engineering and Industrial Aerodynamics, 36, 725-737

Katagiri, J., Ohkuma, T., Marukawa, H and Shimomura, S (2000) “Motion-induced wind forces acting on rectangular high-rise buildings with side ratio of 2,” J

Struct Constr Eng., AIJ 534, 25-32

Shinozuka, M and Jan, C.-M (1972) “Digital simulation of random processes and its

application,” J of Sound and Vibration, 25(1), 111-128

Steckly, A (1989) Ph.D Thesis, University of Western Ontario

Tait, M.J., El Damatty A.A and Isyumov, N (2004) “Testing of tuned liquid damper

with screens and development of equivalent TMD model,” Wind and

Structures, An International Journal, 7, 215-234

Tait, M.J., Isyumov, N and El Damatty A.A., (2004) “The efficiency and robustness

of a uni-directional tuned liquid damper ad modeling with an equivalent TMD,”

Wind and Structures, An International Journal, 7, 235-250

Tait, M.J (2004) The performance of 1-D and 2-D tuned liquid dampers, Ph.D

Thesis, University of Western Ontario

Tait, M.J (2008) “Modeling and preliminary design of a structure-TLD system,”

Engi Struct., 30, 2644-2655

Warnitchai, P and Pinkaew, T (1998) “Modeling of liquid sloshing in rectangular

tanks with flow-dampening devices,” Eng Struct., 20(7), 593-600

Watanabe, Y., Isyumov, N and Davenport, A.G (1997) “Empirical aerodynamic damping function for tall buildings,” J of Wind Eng and Ind Aerodynamics,

72, 313-321

© ASCE

Figure 1 Schematic Diagram of Tuned Liquid Sloshing Damper (TLSD) Installed on

top of the Building

Figure 2 Comparison of Normalized Base Moment Spectra of Target Value based on AIJ and Generated Time History

(d) Equivalent N.L.-MTMD model

c eq

Porous Screen

© ASCE

Figure 3 Aerodynamic impedance coefficients and corresponding Rational Function

approximation

Note: AD denotes Aerodynamic Damping

Figure 4 Accelerations of the building in consideration of aerodynamic damping and

TLSD from time and frequency domain analysis

-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2

α

α - RFA β

© ASCE

Figure 5 Time history of liquid pressure on TLSD screen from nonlinear dynamic

analysis

(a) (b) Figure 6 Instantaneous liquid sloshing motion based on nonlinear multi-sloshing

mode TLSD coupled with building ( f s =1.0Hz)

-15 -10 -5 0 5 10 15

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -3

-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2

x/L

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -3

-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2

Eliminating the Exposure Category from Wind Design Pressure

Nicole Ellison, P.E.1; and Frederick R Rutz, P.E., Ph.D.2

1Univ of Colorado Denver, Campus Box 113, P.O Box 173364, Denver, CO 80202 E-mail:

nicole.ellison@ucdenver.edu

2Univ of Colorado Denver, Campus Box 113, P.O Box 173364, Denver, CO 80202; J R Harris

& Company, 1775 Sherman St., Ste 2000, Denver, CO 80203 E-mail:

frederick.rutz@ucdenver.edu; fred.rutz@jrharrisandco.com

Abstract

Current practice in determining wind loads on a structure is based on the recommendations in the American Society of Civil Engineers Standard 7 (American Society of Civil Engineers (ASCE)

2010), Minimum Design Loads for Buildings and Other Structures An exposure category is

typically selected to represent the surface roughness surrounding a site, this from four discrete exposure categories that approximate all surface roughness conditions The quantification of these exposure categories are primarily based on research work that was completed in the early 1960’s This paper includes a review of surface roughness and why it is important to wind design, and a review of how findings from current research utilizing geographic information systems (GIS) mapping can address exposure from multiple directions As an alternative approach to the traditional exposure categories, GIS mapping data that contains surface roughness information for the United States can be used to calculate surface roughness surrounding a site The data is available from the United States Geological Survey (USGS) and from many local governments The paper presents a GIS-based approach to determining surface roughness using data directly, without the intermediate step of estimating exposure from the traditional categories

INTRODUCTION

The American Society of Civil Engineers Standard 7 (ASCE7-10), Minimum Design

Loads for Buildings and Other Structures identifies three primary exposure categories for

determining the surface roughness surrounding a building as it relates to wind load However,

rarely does a site fall into only one of these three categories Ellington and Tekie completed a

Delphi study that included 20 experts in the field of wind engineering including both practicing

engineers and those in academia based on the ASCE7-95 (American Society of Civil Engineers

1995)(ASCE 1995) Based on this study, Ellington and Tekie found a significant discrepancy

between the exposure classification chosen by the experts and concluded that “there is a

significant probability that the building exposure is classified incorrectly” and “that designers

© ASCE

populated

BACKG

Tsite Pro

have roug

e Exposure C

10 does nowwever, the aplarge discrephness values

15 study (LoThe purpose o

.1 Wind Prof(CPP) (CPP When abrupt egins to deveghness lengt

C, regardless

w contain phoproach has npancy betwee

s determinedombardo and

of this study

z based on geGeographic i

f the United

nd profile is

an, suburban,are shown inPeterka and

h surfaces th

file Changes2015) changes in relop An ab

th difference

s of what the

otos of each not significa

en roughnes

d using a sin

d Krupar 201

y is to develoeographic dainformation iStates

significantly, open count

e of at least a

e true exposu

exposure caantly changed

s values detegle roughne15)

op a method ata for the ac

is becoming

y influencedtry and flat teWind ProfilPP) (CPP 20eds will be lo

ss value from

for determinctual surface

g more and m

d by the surfaerrain develo

le Changes i015) Figure ower than in

rovided by C

w profile is

ss is defined hree (ESDU

face roughneoped over a

in Different 1.1 As sho

n areas of les

Cermak Pete

affected and

d by ESDU a1993) For

d Tekie 1999ntain in ASCLombardo asured wind dased photogr

face roughneconditions

le for the

ess surroundisignificantlyTerrains

9)

also data raphy ess

CE7-ing a

y wind such

d that hen

© ASCE

transitioning from open terrain with roughness length of 0.066 feet (ft), the adjacent roughness

would need to have a roughness length of 0.2 ft or greater or less than 0.022 ft or less

Transitions from open country to suburban or from open country to open sea would easily

qualify as an abrupt change The profile is dependent on the distance of the surface roughness

upwind of the site However, there is a transition period, meaning the wind speed is not

immediately changed when it reaches a change in roughness (Australia/New Zealand Standards

2012)

The area of influence varies greatly in the above noted standards from 0.46 km in ASCE7-10 to 100 km in ESDU Area of influence used by experts also varied for low rise

buildings Examples include 20 times the building height (Bill Esterday, personal

communication March 30, 2016); 2 km (Dr Forest Masters, personal communication, June 26,

2016); 10 km to 20 kilometers (Dr Jon Peterka, personal communication June 6, 2016); and

Lombardo and Krupar use a 3 km radius for the area of influence in their 2015 study (Lombardo

and Krupar 2015)

Another source of ambiguity is small changes in roughness, referred to as an “open patch” by ASCE7-10 There an “open patch” is defined as an opening greater than or equal to

approximately 50 m by 50 m (ASCE7 2010) It is intuitive that changes in roughness close to

the site have a greater impact on the wind loading on the site than small changes in roughness far

from the site However, Peterka suggests that small changes in roughness or “open patches” can

be ignored if they are located more than their width away from the building (Dr Jon Peterka,

CPP, personal communication June 6, 2016) AS/NZS 1170.2:2011 disregards small changes in

roughness directly adjacent the site due to lag distance downwind from the start of the new

terrain (AS/NZS 1170.2:2011)

Wind engineers and researchers will often divide the terrain surrounding the site into

section, the terrain surrounding the site is classified by the roughness This study will utilize

GIS data surrounding one test site to determine the Kz for 16 sections surrounding the site The

results will be compared to field studies by Masters Field studies have shown a direct

correlation between roughness and the turbulence intensity from wind speed measurements over

extended periods of time (Masters et al 2010)

As part of this study several codes and standards have been reviewed and utilized to develop the methods presented herein These codes include ASCE7-10, the Australian Standard

(AS/NZS 1170.2:2011) and the European wind code Engineering Sciences Data Unit (ESDU

1993), all of which provide methods for calculating a wind adjustment factor for wind to account

for surface roughness changes ASCE7-10 and AS/NZS 1170.2 use exposure categories to group roughness types

GIS MODELING

There are several different types of GIS data that are available for the United States that can

be used in quantifying the surface roughness surrounding a site These include vector data

typically available through local governments The vector data can consist of building footprints

shapefiles, tree canopy shapefiles and bodies of water shapefiles A drawback is that there are

many local governments that do not have nor use GIS data and therefore this vector data is not

uniformly available throughout the US Another type of data that is readily available for most of

the US is Land Use raster data available from the United States Geological Survey (USGS)

© ASCE

of influen

The disadvause data for Therefore, ments Lightavailable Lcts on the gro

n range HigLiDAR data

f this data va

of the sensor

ng elevation ficult to procands of point

.2 Screensho

TUDY

study includeonal Airportnce shown b

antage of thideveloped aareas such a

t Detecting ALiDAR is a round such as

gh resolution

is available aries depend

r and atmospdata, data oncess than vec

as airport runAnd Rangingremote sensi

s water, trees

n LiDAR dafrom the Unding on a varpheric interf

n bodies of wctor data or lerred to as a

R Data for Ta

tudy of a the 27.962 and locircle at 1 km

at the resolut

ot differentianways are de

g (LiDAR) ding method t

s, buildings ata can includnited States Griety of factoference Whwater, trees aland use rast

“point cloud

ampa Site mo

located at thongitude -82

m radius and

tion is low, tate between esignated simdata is becomthat uses lighand anything

de objects suGeological Sors includinghile this data and buildingter data Thd” as shown

odeled in Qu

he wind coll2.540 as show

d yellow circ

typically at 3buildings anmilar to subuming more a

ht in from a

g that is withuch as peoplSurvey (USG

g the quality offers the ad

gs, the data f

e raw LiDAbelow in Fig

uick Terrain

lection statio

wn in Figurecle at 2 km r

30 meters, an

nd paved urban housinand more laser to meahin it’s

le observed fGS) website

of the sensodvantage of files are larg

R data consigure 1.2

Modeler

on at Tampa

e 1.3 with arradius

nd

ng asure

from The

or,

e and ists

reas

© ASCE

Figure 1.

radii

Data from City

as those usehot of the GIlding Model

arth View of

IS models w(City of Tamdies have utiDenver, Colohods to extra

us studies byModel in Ar

le Earth 201

this study; o

nd one usingshapefile da

on and Rutz ace roughnes

y Ellison andcMAP 10.2

16) The circ

one using ve

g LiDAR datata to determ2015) and L

ss based on t

d Rutz noted(ESRI 2012

les are a 1 k

ctor data on

ta from the Umine surface Lubbock Texthe building

d above Fig2) This will

km and 2 km

the buildingUSGS (USGroughness xas (Ellison shapefile wagure 1.4 show

l be referred

m

gs

GS and

Figure 1.

is 1 km r

Applwas used

ied Imagery

d in conjunctS(USGS 201

d of grouping

e with four c

ns higher tha

es based on tand single-sildings (ASCccount for chnce on the si

1 km and 2

e of 2 km Thexposure co the procedu

nt roughnessected at the scalculated adTable 1 anddetermined bnds to the hecalculated av

e 1.6

ot of Buildin

Quick Terration with ES16) This wi

g the points categories; w

an one-story the ASCE7-1story buildin

CE 2010)

hanges in rouite Roughn

km to simul

he roughnesoefficient in tures in the A

s factor fromsite

djustment fac

d Figure 1.5

by Masters ueight of the averaged com

ng Vector Da

ain Modeler

RI ArcMAPill be referre

in the point water, land, tr

in height A

10 of 0.016 f

gs and 3.3 ftughness, datess values wate a single

s adjustmentthe ASCE7-ASCE7-10, A

m the roughn

ctors, K z, froalong with tsing the ASCanemometer mposite result

ata for Tamp

(Applied ImP10.2 to proc

ed to as the Lcloud by elerees and sing

A roughness

ft for water,

t buildings b

ta was extracwere calculatchange in ro

t factor, K z, r

10 was calcuAS/NZS andess values ca

om the GIS m

he adjustmeCE7-10 meth

at the wind

ts for both th

pa Site mode

magery Quiccess the raw LiDAR Modevation Thegle-story buvalue was a0.066 for labetween one cted from theted for the 1 oughness at 1referred to aulated for th

d ESDU and alculated by

models for eent factors cohod for a hecollection s

he Building

eled in ArcM

ck Terrain MLiDAR datadel In summ

e points wereuildings and bassigned to eand and Daveand two sto

y Masters for

ach site are omputed fromeight of 10 mtation AlsoModel and L

MAP The cir

Modeler 2016

a obtained frmary, the pro

e classified ibuildings wiach of these enport’s 1.64ries and 5 ft

ls for two ar

d for the areatotal area ofalent velocityange in roughared to the

r measured w

summarized

m roughnessmeters which

o presented LiDAR Mod

rcle

6) rom ocess into a ith

4 ft for reas

a

f

y hness wind

d

s

h are dell

© ASCE

Masters

Equiv K z

0.634 0.646 0.728 0.791 0.880 0.876 0.830 0.842 1.000 1.009 0.900 0.822 0.761 0.685 0.660 0.644

ted K z Facto

e Results of

GIS ASCE BLDG

Equiv K z

0.754 0.767 0.717 0.702 0.931 0.933 0.901 0.807 0.937 1.001 1.001 0.935 0.786 0.752 0.676 0.727

rs and Maste

f K z with LiD

GIS AS/NZS BLDG

Equiv Kz

0.690 0.812 1.001 0.715 0.979 0.980 0.970 0.813 0.935 1.001 1.001 0.981 0.791 0.741 0.657 0.701

ers’ Equival

DAR model a

GIS ESDU BLDG

Equiv Kz

0.679 0.784 0.826 0.686 0.843 0.846 0.789 0.746 0.932 0.932 0.932 0.850 0.778 0.719 0.632 0.662

ent K z

and Building

U GIS ASC LIDAR

Equiv Kz

0.731 0.808 0.772 0.711 0.829 0.849 0.791 0.822 0.982 1.066 1.031 0.968 0.814 0.724 0.711 0.800

g Model Ove

CE

z

GIS AS/NZS LIDAR

Equiv Kz

0.701 0.810 0.971 0.677 0.832 0.883 0.795 0.830 0.973 1.043 1.007 0.969 0.822 0.713 0.683 0.816

erlaid on

z

GIS ES LIDAR

© ASCE

.7 K z Values

e Results of

orrespond wfor the weste

Kz values wefor both the

re conservati

e 1.8 presentASCE7-10,

for Building

f K z with LiD

ell with thoserly directionere lower thabuilding moive than the ted below shAS/NZS an

d ESDU me

ng ASCE7-1

and Building

d by Mastersere is open w

s calculated bLiDAR modvalues calcularison of theethods

10, AS/NZS

g Model Ove

s As expectedwater In mo

by the ASCEdel, suggestilated from fi

ers

e NZS odels

ta

© ASCE

s are new sin

e LiDAR daverified for from the air

NATING EX

As shown froustment factroughness a

ne exposure perform thes

TUNITIES

Developmentnal work willThe work co

is a complexect of current

re 1.5, that eterrain and th

for LiDAR

NZS and ESvariation be

he Building Mlding data docurrent Whnce the GIS data contains gthe specific rplanes at the

XPOSURE C

om the study,tors to accouand changes category to t

e calculation

FOR ADDI

t of GIS calc

l be requiredonsists of bot

x task as ther

t research Cexperts are n

he methods

Model using

SDU methodetween the suModel Bothoes not conthen compareddata was colgeneralized rarea based o

e airport as t

CATEGOR

, exposure caunt for terrain

in roughnesterrain surro

g ASCE7-10

ds show goodurface rough

h sets of dataain any roug

d to Google llected and aroughness d

on height Tthese have h

RY

ategories are

n Surface r

s surroundinounding a sit

WORK

hness adjustm

calculated z

he GIS proceple factors th

g this task isment on the aing for chang

0, AS/NZS a

d correlationhness calcula

a have advanghness from Earth (Googare thereforedata based onThere is also heights close

e not requireroughness ca

and ESDU

n However,ated in the Lntages and dtrees and thgle Earth 20 not include

n roughness roughness t

to buildings

ed in the dete

an be determ

us removing nology offer

s is not withoness) values dating the da

he surface rouation, illustraroughness vahness

, there are ariDAR modedisadvantage

he building d16), there ar

d in the builfactors that that is being

s

ermination omined directlythe need to

rs a variety o

out challengcan be usedata Validatiughness andated in Tablealue for diffe

d is

e 1 erent

© ASCE

Another debate among members of the wind engineering community is how far out surface roughness needs to be considered and how the effects of open patches should be

weighed The authors believe using GIS, the surface roughness using multiple fetch distances

may be readily calculated to aid in this determination Small changes in roughness can also be

investigated and compared to field wind data

To make these methods cost effective, the process could be automated and streamlined

In ArcMap (ESRI 2012)ArcGIS, Model Builder available and computer scripting code could be

used to automate the process and expedite the processing of the data

Looking at the larger picture, the surface roughness study could be expanded to include wind loading based on wind direction Methods used for urban morphology such as wind

directionality could be could be incorporated into the surface roughness calculation While

adding complexity to the calculations, this would provide additional information for designers

that could be incorporated into the layout and design of a building

CONCLUSION

Because of the variable nature of roughness, selecting one exposure category often does not represent the actual surface roughness conditions adjacent to the design structure Using

tools such as GIS ArcMAP and Quick Terrain Modeler (Applied Imagery Quick Terrain Modeler

2016), a more accurate surface roughness can be calculated removing the ambiguity and

uncertainty from use of only one classification to characterize the surface roughness surrounding

a site This would provide designers the opportunity to expand the limited roughness values that

are currently being used today This method can ultimately refine the design wind pressures on a

structure Exposure categories can be eliminated from the wind pressure determination

procedure if surface roughness is determined directly

GIS mapping technology is growing exponentially This technology is now widespread and available at our fingertips on our mobile devices Modeling the surface terrain around a

structure using GIS follows this trend

REFERENCES

American Society of Civil Engineers (1995) “Minimum Design Loads for Buildings and Other

Structures.” ASCE7-95

American Society of Civil Engineers (ASCE) (2010) Minimum design loads for buildings and

other structures ASCE standard, American Society of Civil Engineers, Reston, VA

Appied Imagery (2016) “Quick Terrain Modeler Software.”

Australia/New Zealand Standards (2012) “Structural Design Actions, Part 2 Wind actions.”

Ellingwood and Tekie (1999) “Wind Load Statistics for Probability-Based Structural Design.”

Journal of Structural Engineering

Ellison, N., and Rutz, F R (2015) “Surface Roughness and Its Effect on Wind Speed: Modeling

Using GIS.”

Ellison and Rutz (2016) “Comparison of Surface Roughness Assessment Using GIS Mapping

Technology to Field Measurements.” 4th American Association for Wind Workshop

Engineering Sciences Data Unit (ESDU) (1993) “Strong Winds In The Atmospheric Boundary

Layer.” (Item Number 82026 With Amendments to A to C)

ESRI (2012) “ArcMAP 10.1 for Desktop Advanced Student Edition Software.” Environmental

Systems Research Institute Google Earth (2016) “Google Earth Tampa International Airport.”

<https://www.google.com/earth/> (Oct 1, 2016)

Lombardo and Krupar (2015) “Aerodynamic Roughness Length: Comparison of Estimation

Methods and Uncertainty Quantification.” Proc 14th Int Conf on Wind Engineering, Porto

Alegre, Brazil, Int Association for Wind Engineering, 1–17

Masters et al (2010) “Toward Objective, Standardized Intensity Estimates from Surface Wind

Speed Observations.” Bull Amer Meteor., 91, 1665–1681

USGS (2016) “Earth Explorer.” 2016, <http://earthexplorer.usgs.gov/> (Aug 1, 2016)

© ASCE

Wind Load Prediction on Tall Buildings in a Stochastic Framework

M Gibbons1; J Galsworthy2; M Chatten3; and S Kala4

1RWDI, 600 Southgate Dr., Guelph, ON, Canada N1G 3W6 E-mail:

it reduces a complicated problem into a simple account of loads and effects A stochastic approach considers uncertainty in the design inputs, such as natural frequencies and damping ratio While this type of approach has been described previously in the literature, the current approach provides a framework in which wind loads can be predicted in practical design scenarios The current study focuses on tall, slender buildings in an effort to better understand the impact that uncertainties in extreme wind climate, damping ratio and natural frequency have on predicted wind loads, and how these uncertainties contribute to the overall reliability of the structure

A method is described that allows for the direct calculation of probability of failure based on a stochastic relationship between load and resistance

INTRODUCTION

The basis of any modern building code or structural design standard is to ensure that the probability of failure of a building or structure is sufficiently small What constitutes ‘sufficiently small’ is defined by the design standard and is determined based on the input of design professionals though a consensus based approach (Galambos et al 1982; Ellingwood and Tekie 1999)

In the 2010 edition of Minimum Design Loads for Buildings and Other Structures (ASCE 2010, henceforth ASCE 7-10), the targeted probabilities of failure are listed in Table C1.2.1a ASCE 7-10 is the basis for most building codes in the United States

The selection of an appropriate probability of failure depends on two main factors – the intended use of the structure and the nature of failure Also provided in Table C1.2.1a are reliability indices, β If one considers the probability of failure to be normally distributed as is done in ASCE 7-10, β relates to annualized probability of failure, pf, according to the relationship given in equation 1, where N is the expected lifespan of the structure

© ASCE

β = ϕ-1(1-N·pf) [1]

Probability of failure and β represent the code intent, however it is rare that they are used to directly to calculate a load This first principles approach would prove to be far too time consuming and unnecessarily complicated for the vast majority of building design conducted based on ASCE 7-10 Rather, the load and combination factors in Chapter 2 of ASCE 7-10, and the climatic loading values provided at various mean recurrence intervals (MRI) have been calibrated against the β factors for conventional buildings and structures This is described in great detail by Ellingwood

et al (1980) This process ensures that for the vast majority structures designed according to ASCE 7-10 meet or exceeded the reliability intended by the design standard

Historically, a target reliability index of 3.0 or a probability of failure of 3×10-5 is typically taken for wind loading of tall buildings For an occupancy category II, which describes the buildings considered in the current study, this corresponds to failure that is not sudden and does not lead to wide-spread progression of damage (ASCE 7-10) Up to this point, most tall building design for wind loading has been based on allowing elastic deformation of structural members, but not inelastic/plastic deformation, which would tend to prevent the wide-spread progression of damage

As most strong wind events are forecasted with reasonably accuracy (and, particularly

in the case of hurricanes, an abundance of caution/conservatism in the forecast), the potential failure of a structure would not be considered sudden

APPLICATION TO WIND LOADING OF TALL BUILDINGS

In editions of ASCE 7-05 and earlier, wind loads were calculated at MRIs of 50 years For load combinations where wind was the primary action, a load factor of 1.6 was prescribed This load factor was based on the assumption that wind load

increases proportional to the square of wind speed This assumption held for the vast majority of structures (i.e rigid structures) however, not for dynamically sensitive structures such as tall buildings In these structures, wind loads are typically proportional to wind speed to the power of 2.5 or greater Therefore, wind loads predicted for rigid structures based of off ASCE 7-05 are at a higher level of reliability than those for dynamically sensitive structures This is one of the reasons why in ASCE 7-10, MRIs were increased to the ultimate state based on building category and load factors reduced to 1.0 For a normal importance/Category II structure, in ASCE 7-05 wind loads were predicted at an MRI of 50 years and then multiplied by a wind load factor of 1.6, whereas in ASCE 7-10 wind loads are predicted at an MRI of 700 years and multiplied by a wind load factor of 1.0

The intent of the current study is to revisit this problem for tall buildings that have been tested by RWDI in our boundary layer wind tunnels The buildings investigated all exhibited strong across-wind response, where small changes in damping ratio and natural frequency can result in large differences in predicted wind load For example,

© ASCE

the predicted 700 year wind load from one of the buildings investigated in this study increased by 37% when damping is decreased from 2% to 1%

The sensitivity of wind loads to natural frequency and damping ratio in dynamically sensitive structures is significant due to the inherent uncertainty regarding these quantities in the design of tall buildings Most studies suggest appropriate coefficient

of variations of 5% for frequency and 40% for damping, although to date there has not been an exhaustive study on the error between as designed versus observed frequency and damping Natural frequencies and damping ratios are not static in a building, with a strong dependency on the deflection of the building and age/history

While wind loads predicted at an ultimate MRI in the ASCE 7 procedures accounts for wind loading increases greater than the assumed velocity squared relationship, it does not explicitly consider the uncertainty regarding natural frequency or damping ratio Numerous studies have been described in the literature which tackle this problem The classical approach has been ‘First-Order Second Moment’ (FOSM) method (Davenport 1983; Chatten et al 2016), which allows for the calculation of an appropriate load factor based on an accounting of the uncertainties that contribute to a predicted wind load

Studies that have explicitly propagated uncertainties associated with a comprehensive range of parameters have found load factors for dynamically sensitive structures that are in some cases significantly larger than contemporary design practice (Gabbai et

al 2008; Bashor and Kareem 2009; Kwon et al 2015) Amongst these studies there are significant differences in the magnitudes of the load factors derived depending on the analysis approach, which parametric uncertainties were considered and the definition of load factor Comparison between studies is therefore difficult, particularly since later research was unable to replicate load factors recommended by Gabbai et al (2008) which were as high as 2.3 for rigid buildings and 3.5 for flexible buildings For the purposes of providing a basis of reference for this paper Bashor and Kareem (2009) recommended the load factor for a dynamically sensitive building is around 1.9 for the conversion between the Serviceability Limit State (SLS) loads to the Ultimate Limit State (ULS) as compared to 1.6 for a rigid structure This factor corresponds with a higher load than simply using a higher wind velocity as it also accounts for frequency and damping uncertainties This factor is reasonably comparable to the same case considered by Kwon et al (2015) Their study examined

a more extensive range of parameters and found that uncertainties associated with wind speed, frequency and damping contribute most to the uncertainty in the response

of a dynamically sensitive structure

The increased load factors for dynamically sensitive structures identified by these parametric studies (Gabbai et al 2008; Bashor and Kareem 2009; Kwon et al 2015) were based on the assumption that the structure’s dynamic properties are constant between SLS and ULS loading However full-scale data indicates that as response increases, frequencies tend to decrease and structural damping increases It is a common design assumption that damping will likely exceed nominal design values in

© ASCE

the extreme responses associated with the ULS event as inelastic behavior of the structure is to be expected (Bashor and Kareem 2009; Allsop 2011) Furthermore at the ULS aerodynamically damping is likely to play a more significant role as it generally increases proportional to wind speed As the above discussion highlights, the selection of reliable ULS wind loads for a tall or super-tall tower may warrant a project specific reliability analysis that is not required for more typical structures

MONTE CARLO METHOD

For the current study, a fully probabilistic Monte Carlo Method has been devised in order to assess the impact of input uncertainty on predicted base building moments (My, Mx and Mz) This is outlined in Figure 1

Figure 1 – Outline of proposed Monte Carlo method

5) A Monte Carlo simulation is conducted to determine probability of failure based

on the statistics of the 50 year base building moments determined in step 4 A load versus resistance comparison is made stochastically, based on the relationship shown in equation 2, where α is a combined load and resistance factor and σR is the

standard deviation of overall resistance of the structure

4) The 50 year base building moments are determined for each building scenario based on the FT1 fits Their distribtion was found to be lognormal, with a mean

value of μM50 and a standard deviation of σM50

3) A Fischer-Tippet Type 1 (FT1) distribution is fit to each base building moment for each building scenario, producing 1000 FT1 fits per moment In order to speed convergence, the natural logrithm of the base building moment was taken prior to

the FT1 fitting An example of this fit is shown in Figure 2

2) 500 years of simulated wind events are applied to each independent building scenario A unique simulated time history of wind events is applied to each independent building scenario This produces peak base building moments (My,

Mx, Mz) for each wind event Both Tropical Cyclone (TC) and non-TC extreme

wind climates are considered in the current study

1) 1000 independent building scenarios are created, these based on nominal, designed values of natural frequency and damping ratio and coefficients of variation of 0.05 and 0.40, respectively Natural frequency and damping ratio are

as-assumed to be lognormally distributed

© ASCE

In tnatu

al (andload

Fig pre fitt

Whbasocc

STU

Fivtheybetwgeowingeo

this method ural frequen(2015), whic

d damping ra

ds for dynam

gure 2 – Exa econditioned ing

hile the meth

se building mcupied floor

UDY BUIL

ve buildings h

y are all tall ween 650 ft ographic regi

nd climates, ometrically u

a focus has bncy and damp

ch state that atio have themically sensi

ample of an

d by taking

hod describedmoments, a sacceleration

DINGS

have been inand slender and 1800 ft

ions around they locatedunique A su

L(μM50, been placed ping ratio Tuncertaintie

e largest contitive structur

FT1 extrem the natural

d in the currimilar appro

ns or façade w

nvestigated i– all with as Differencethe world, a

d in differentummary of th

σM50) ≤ R(μM

on uncertainThis is consi

es regarding wtribution to tres

me value fit

l logarithm

rent study haoach could bwind loads, j

in the currenspect ratios o

es include thand are there

t surroundingheir characte

M50 · α, σR)nties regardistent with thwind climatthe uncertain

for a base m

of the base

as been appli

be applied to just with dif

nt study On

of 1:6 to 1:1hat they are lefore subject

g terrains aneristics is pro

ing wind clim

he findings o

te, natural frenty in predic

moment, moment pr

ied to the prethe predictifferent criter

e commonal13.7 and heigocated in dif

to different

nd they are aovided in Ta

[2]

mate,

of Kwon et equency cted wind

rior to

ediction of

on of top ria applied

lity is that ghts fferent extreme

ll able 1 As

© ASCE

the buildings investigated in this study are confidential, details beyond those provided

in Table 1 or images cannot be provided

Table 1 – Characteristics of the five study buildings, where H is the height of the

building, W is the width (smaller horizontal dimension) of the building, L is the

length (longer horizontal dimension) of the building

H/W 13.7 6.0 9.5 8.8 6.7

Average Damping

Average Natural Frequencies

0.13, 0.20, 0.30

0.18, 0.22, 0.30

0.13, 0.14, 0.30

0.10, 0.10, 0.26

0.14, 0.15, 0.40 Exposure

Mixed (open water, urban, suburban)

Suburban Urban Urban

Mixed (open, urban) Climate

TC Dominated No TC

Building Geometry

Rectangular in plan at base, stepped with increasing height

Square in plan at base, no tapering,

no twist

Square in plan at base, no tapering,

no twist

Square in plan at base,

no tapering,

no twist

Square in plan at base, tapered,

no twist

RESULTS

The main result of this study is a relationship between combined load and resistance factor and probability of failure These are plotted in Figures 3 through 7 This represents a major shift from how wind loads are conventionally considered, where the probability assigned to the wind loads represents the MRI of the design wind speed As discussed previously, the conventional approach is not risk consistent for the design of tall buildings, whether these buildings are designed for the SLS with a load factor of 1.6 or ULS with a load factor of 1.0

Some clear trends emerge from the plots in Figure 3 through 7 For a targeted probability of failure of 3×10-5, most moments require a combined load and resistance factor of between 1.5 and 2.0 Only the Mx moment from Building D had a combined factor greater than 2.0 for the targeted probability of failure

It is interesting to note that the combined factors for Mx and My are typically different, with only building B showing reasonable agreement at the targeted probability of failure, while all but building A are square in plan Also of note is that

© ASCE

the diff

F

F

current studferent combi

igure 3 - Co

igure 4 - Co

dy suggests tined factors

ombined Lo

ombined Lo

that in order are required

oad and Res

sistance Fac uilding B

a consistent loment