Content uploaded by Matiyas A. Bezabeh
Author content
All content in this area was uploaded by Matiyas A. Bezabeh on Oct 03, 2018
Content may be subject to copyright.
Paper ID1
Building Tomorrow’s Society
Bâtir la Société de Demain
Fredericton, Canada
June 13 – June 16, 2018/ Juin 13 – Juin 16, 2018
RISKBASED WIND DESIGN OF TALL MASSTIMBER BUILDINGS
Bezabeh, Matiyas1,4, Bitsuamlak, Girma2 and Tesfamariam, Solomon3
1 Visiting Research Student, Department of Civil Engineering, Western University and Ph.D. Candidate,
School of Engineering, The University of British Columbia, Canada
2 Associate Professor, Department of Civil Engineering, Western University, Canada
3 Professor, School of Engineering, The University of British Columbia, Canada
4 corresponding_author matiyas.bezabeh@alumni.ubc.ca
Abstract: The rapid growth of urban population and the associated environmental concerns are partly
influencing city planners and construction stakeholders to consider “Sustainable Urbanization” alternatives.
In this regard, recent urban design strategies are entertaining the use of “tall timber buildings.” Generally,
tall masstimber buildings (MTBs) utilize preengineered wood panels to form their main gravity and lateral
load resisting systems, which makes them lighter and more flexible than buildings made from concrete,
masonry or steel. As a result, frequent exposure to excessive windinduced vibrations can cause occupant
discomfort and possible inhabitability of the buildings. This paper attempts to apply a riskbased procedure
to design a 102meter tall MTB by adapting and extending the Alan G. Davenport Wind Loading Chain as
a probabilistic performancebased wind engineering framework. The structural systems of the study
building are composed of Cross Laminated Timber (CLT) shear walls, CLT floors, glulam columns, and
reinforcedconcrete link beams. Initially, aerodynamic wind tunnel tests were carried out at the Boundary
Layer Wind Tunnel Laboratory of Western University on the 1:200 scale MTB model to obtain transient
wind loads. Subsequently, using the wind tunnel data, the study MTB was structurally designed. In the risk
based performance assessment, uncertainties were incorporated at each step of the Wind Loading Chain,
i.e., local wind, exposure, aerodynamics, dynamic effects, and criteria. These uncertainties were explicitly
modeled as random variables. Dynamic structural analyses were carried out in the frequency domain to
include the amplification due to the resonance component of the excitation. Structural reliability analysis
through Monte Carlo sampling was used to propagate the uncertainties through the Wind Loading Chain to
quantify the risk of inhabitability and excessive deflection. The results of reliability analysis were used to
develop fragility curves for wind vulnerability estimations. Based on the results, the effects of various
uncertainties are discussed, and riskbased design decisions are forwarded.
Paper ID2
1 Introduction
The rapid growth of urban population and the associated environmental concerns are partly influencing city
planners and construction stakeholders to consider “Sustainable Urbanization” alternatives. Nowadays,
sustainable urbanization has emerged as a viable solution towards smart and livable cities that are more
resilient, and environmental friendly. In this regard, recent urban design strategies are entertaining the use
of “tall timber buildings.” Recent studies on the life cycle assessment of buildings indicated that wood is an
efficient construction material in terms of embodied energy and greenhouse gas emissions (FPInnovations
2010). To this end, latest design guidelines and standards in the US and Canada are considering the use
of multistory masstimber buildings (MTBs). For example, the 2015 American National Design Specification
(NDS) for Wood Construction (ANSI/AWC NDS2015) and the 2016 supplement of the Canadian National
Standard for Engineering Design in Wood (CSA O86) included chapters dedicated to the design of CLT
structural elements. In 2018, the International Building Code (IBC) Ad Hoc Committee on Tall Wood
Buildings approved the application of “heavytimber” structural elements for constructions beyond Type IV
(Baldassarra 2017). In addition to the inclusions in the building codes and standards, over the past 15
years, with the use of CLT, several tall timberbased buildings were constructed. In 2017, the University of
British Columbia (UBC) finalized the construction of 18 story (53 m) tall hybrid masstimber residential
building. Currently, the building is the tallest standing timberbased building in the world. The success of
existing timberbased buildings in meeting their design objectives enhances the confidence of architects
and developers towards pushing the height limits of timber buildings beyond 50 meters.
To make wooden skyscrapers possible, research on timber structures is getting momentum. For example,
in Canada, FPInnovations released the first comprehensive design and construction guideline of tall timber
buildings (Karacabeyli and Lum 2014). Researchers at the University British Columbia and FPInnovations
also developed both, force and displacementbased design guidelines for the timbersteel hybrid system
(Tesfamariam et al. 2015, Bezabeh et al. 2016, Bezabeh et al. 2017). In the USA, the “NHERI Tall wood
Project” was launched in 2016 to develop a resiliencebased seismic design guideline for tall timber
buildings (Pei et al. 2017). Moreover, several research projects are currently underway in different parts of
the world. While most of the studies focused on material science, fire, and seismic performance evaluations,
wind performance of highrise timber buildings has scarcely been studied. Generally, tall MTBs are
lightweight and more flexible than buildings made from steel, concrete or masonry. The increased flexibility
limits their lateral stiffness, thus making them vulnerable to excessive along and acrosswind vibrations
(Reynold et al. 2011, Popovski et al. 2014). As a result, frequent exposure to excessive windinduced
vibrations can cause occupant discomfort and possible inhabitability and unserviceability of this kind of
emerging buildings. Therefore, in this paper, a riskbased wind design procedure is applied to design a 102
meters tall MTB by adapting and extending the Alan G. Davenport Wind Loading Chain as a performance
based wind engineering (PBWE) framework.
2 Adapting the Wind Loading Chain as a probabilistic PBWE framework
Current prescriptive building design approaches consider designing for single limit state by accounting
uncertainties through safety coefficients to achieve minimum safety and acceptable serviceability levels.
Efforts to improve designs for fixed limit state started decades ago in the earthquake engineering
community after an enormous amount of monetary losses due to the 1989 Loma Prieta and 1994 Northridge
Earthquakes in California. Performancebased engineering considers a range of limit states (performances
objectives) throughout the lifetime of the structures to make riskbased design decisions. In wind
engineering, suggestions to design structures for different limit states was first introduced by Davenport
(1970). The identified limit states are ultimate strength, permanent deformation, excessive acceleration
(occupant discomfort), and integrity of the cladding and finishing materials.
Performancebased engineering requires accurate models of hazard, hazardstructure interaction,
structural properties, criteria, and consequences. In 1961, the late Alan G. Davenport introduced a
mathematical, somewhat philosophical, model to evaluate the wind load on structures similar to the current
context of performancebased engineering. He coined his thought process as interconnected chains and
named the performance evaluation framework as “Wind Loading Chain” (Davenport 1982). The weakest
link in the chain determines the final response and reliability of the system. Davenport’s Wind Loading Chain
laid the foundation for the modern wind engineering and provides a theoretical basis for many building
codes and standards. In 2011, the International Association of Wind Engineering (IAWE) recognized the
Paper ID3
Alan G Davenport Wind Loading Chain as an official wind engineering terminology (Isyumov 2012). For the
performance assessment and design of structures, the Wind Loading Chain (Figure 1) starts by modeling
the local microclimate of a target site to predict the design wind speeds. Statistical analysis of historical
wind speed data or computer simulations can be used to determine the design wind speeds (Davenport
1970). The approaching wind flow around the building site is affected by the terrain roughness and local
topography. These effects are accounted through the second element of the Wind Loading Chain. Building
aerodynamics (or simply building shape effects) is also another factor that significantly affects the wind
loads and responses of structures. Wind tunnel tests and computational fluid dynamics can be used to
study bluff body aerodynamics (Davenport 1970, Bitsuamlak 2006, and Bitsuamlak and Simiu 2010, Irwin
et al. 2013, Dagnew & Bitsuamlak 2014). Davenport (1967) developed a framework to quantify the dynamic
response of structures to wind using random vibration theory. In the framework, he suggested an approach
to linearize the dynamic wind force equation and to perform dynamic analysis under wind in the frequency
domain. Finally, the performance of the structure could be judged by comparing the peak response
demands with criteria from codes and standards.
Figure 1: The Alan G. Davenport Wind Loading Chain
Even though the Alan G. Davenport Wind Loading Chain framework can be used to design and assess tall
buildings for different limit states, the current wind design codes and standards only consider the linear
elastic capacity of structural systems. This is mainly due to the nonload reversal nature of wind load (uni
directional mean component), issues with damage accumulation and duration of wind storms. To accurately
assess the performance of structures under wind load, studies should aim at understanding the nonlinear
behavior under wind load and developing wind damage states, fragility and consequence curves related to
wind hazard. Towards this, collaborative research is currently underway between the University of British
Columbia (UBC), Western University, and FPInnovations to develop a unified PBWE framework by adapting
the Wind Loading Chain. The research program includes several wind tunnel tests of tall MTB models
subjected to both synoptic and nonsynoptic (tornadoes) wind loads. Different performance and damage
limit states are being developed using both linear and nonlinear aeroelastic wind tunnel tests. The first part
of the collaborative research, i.e., the application of riskbased design procedure to design tall MTBs
subjected to random wind loads is presented in this paper. For this purpose, the outcome of the probabilistic
serviceability performance assessment (probability of failure or vulnerability) is used as a risk measure.
This paper is organized as follows. Initially, we introduce the case study 102 meters (30 story) MTB. Next,
the details of aerodynamic wind tunnel tests and the obtained results are presented. Subsequently, the
structural design process of the building, carried out using wind load information from wind tunnel tests, is
presented with the design results. This is followed by uncertainty modeling, development of two
serviceability limit states using 10 and 50year return period wind speeds, and uncertainty propagation
using structural reliability analysis. The results from reliability analysis are used to develop fragility curves
for wind vulnerability estimations. Based on the results, the effects of various uncertainties are discussed,
and riskbased design decisions are forwarded.
3 Description and design details of the study masstimber building
In this paper, a 30story masstimber building is considered as a case study prototype building. The
conceptual layout of the building was introduced in 2013 by Skidmore, Owings and Merrill (SOM LLP, 2013).
The building has a typical floor plan dimensions of 30m x 42m and floortofloor height of 3.4m. Heavy
timber products such as Glulam and CLT are used to construct both the gravity and LLRS of the building.
Figures 2 (a and b) show a typical floor and 3D views of the case study building.
1 2 3 4 5
Paper ID4
a)
Figure 2: The studied 30story MTB: (a) 3D perspective view; (b)
Typical floor perspective view
b)
As shown in the figure, the shear walls and floor systems are made from CLT panels. Concrete link beams
are used to couple the movements of CLT shear walls resulting in a global “H” shaped Lateral Load
Resisting System (LLRS). As shown in Figure 2a, in addition to the two coupled CLT core walls, four single
CLT shear walls provide uplift resistance due the wind load from the wider face of the building. Glulam
columns are only used at the perimeter of the building to increase the gravity load share of the LLRS. In
this way, it is possible to increase the uplift resistance of the building. However, the absence of the interior
columns would require larger span floor systems due to increased floor deflection demands. Therefore,
endrotational restraint is provided by concrete spandrel perimeter beams. Concrete elements are extended
from the RC link beams to the end spandrel beams to act as a concrete joint. A modified HolzStahl
Komposit (HSK) system (Zhang 2017) is used to connect the CLT walls and floor systems to the concrete
joints. The modified HSK system is designed and used as shear and holddown connectors to resist story
level wind forces and net uplift forces, respectively. Additional epoxy coated vertical reinforcement bars are
also provided at the boundary elements of the CLT shear and core walls to increase their inplane bending
moment resistance.
3.1 Wind tunnel testing
Wind loads were quantified through aerodynamic wind tunnel tests at the Boundary Layer Wind Tunnel
Laboratory (BLWTL) of The University of Western Ontario (UWO). To account for the effect of the terrain
and shape of study building (second and third elements of the Wind Loading Chain), a study building was
tested using a boundary layer flow corresponding to an open country exposure condition at 1:200 geometric
scale. The aerodynamic rigid building model and test setup inside the wind tunnel are depicted in Figures
3 (a and b). Simultaneous time series of pressure fluctuations were measured using 495 taps installed on
the models. The tests were carried out for 36 wind angles of attack (AOA) in 10 degrees increment and
digitalized at the rate of 400 Hz. The time series of pressure fluctuations were converted to nondimensional
pressure coefficients (Cp) to scale the model pressures to fullscale pressure. Equation 1 was used to
calculate the time series of pressure coefficients. Average dynamic pressure (qp) at the building height is
used to normalize the measured pressure readings.
Paper ID5
a)
b)
Figure 3: a) Aerodynamic model of a 30story MTB; b) Wind tunnel test setup at BLWTL
(1)
where P(x, y, z, t) is a pressure reading on the surface of the aerodynamic model at time (t) using a tap
located at (x, y, z) from the origin (center of the roof) , Po is a static pressure, T is test duration, qp is a
dynamic pressure measured at the height of the building using the Pitot tube.
Figure 4: Mean Cp distribution over the surface of the building for zerodegree wind angle of attack
The mean Cp contours over the surface of the building are presented in Figure 4 for zerodegree AOA, i.e.,
when the wider face of the model is orthogonal to the incoming wind. As shown in the figure, the windward
(N) wall is under positive pressure with a stagnation point around 2/3 of its height. The maximum of the
mean Cp value on the windward wall is 0.88. However, all the other sides are under suction (negative
pressure). The east (E) and west (W) walls (sidewalls) show a symmetric Cp distribution. To perform
dynamic structural analysis and design of the study building, pressure integration over the surface of the
building model was performed to compute the aerodynamic wind force time series in two sway, and
torsional directions.
Paper ID6
3.2 Structural design of the study masstimber building
Initially, to perform the design of the structural systems of the prototype MTB, the design gravity and wind
loads were quantified using NBCC 2010 (NRC 2010) and wind tunnel results, respectively. Preliminary
dynamic analyses were performed to quantify the equivalent static wind loads (ESWLs) which include the
inertial contribution (resonance). The details of the dynamic analysis procedure are provided in subsequent
sections. For fullscale wind load calculation, the building is assumed to be in the city of Chicago, USA. The
design of the study building considered the axial compression, inplaneshear, and inplane and outof
plane bending moment demands, and their interactions due to gravity and wind loads. Section capacities
were obtained from CSA O8614 (CSA 2014) and Structurlam CLT manufacturer catalog. The process to
design the gravity and main lateral load resisting system of the study building is outlined in the following
steps. Tables 1 and 2 summarize the final design sections of the study tall MTB.
Step1: Design of floor systems
Step2: Design of Glulam columns
Step3: Preliminary sizing of CLT shear and core walls for factored gravity loading
Step4: Check the factored inplane shear capacity of CLT walls against the wind load demand
Step5: Check the coupled axial and outofplane bending of CLT shear walls
Step6: Check the coupled axial and bending of CLT core walls
Step7: Design of HSK connections
Step8: Design of RC linkbeams
Step9: Design of RC spandrel beams
Table 1: Final design section of the prototype MTB
Structural System
Structural member
Design checks
Design
specification
Gravity load resisting
system
CLT floor system

7 layers245E
E1M5
Glulam column
axial
945 mm x 945 mm
Spandrel beam

400 mm x 500 mm
Lateral load
resisting system
CLT shear walls
axial
9 layers315E
E1M5
inplane shear
axial + outofplane bending
RC linkbeam
Shear stress limit
405 mm x 650 mm
CLT core walls
axial
9 layers315E
E1M5
axial + outofplane bending
Table 2. Details of the designed shear connectors for the first floor the building
Structural system
Shear connectors profile
Total number of SL
Number of
shear
connectors
Number of plates
inside CLT
SL per plate
CLT shear wall1
900
3
3
100
CLT shear wall2
900
3
3
100
CLT shear wall3
900
3
3
100
CLT shear wall4
900
3
3
100
CLT core wall1
2100
8
3
90
CLT core wall2
2100
8
3
90
Paper ID7
4 Dynamic structural analysis in the frequency domain
To quantify the wind load demands, functional relationships are needed to connect the elements of the
Wind Loading Chain. For this purpose, random vibration theory was used to quantify the response of
buildings subjected to stochastic wind load. In random vibration theory, the dynamic response of buildings
can be evaluated using linear modal analysis in the frequency domain. In the analysis, by simplifying the
case study building as uncoupled stickmass model, fullscale wind force time series were applied at each
floor in two sway and torsional directions. During the analysis, the dynamic properties were represented by
the mechanical admittance factor. Modal responses were combined by considering the intermodal cross
correlation due to statistical coupling effects between modes. Peak responses were computed using up
crossing rate concept by assuming the responses are Gaussian. Peak vector resultant responses were
calculated from the sway mode responses using mean square addition approach. The statistical correlation
between the sway modal responses was accounted by the joint action factor.
5 Uncertainty modeling
Several studies showed the influence of other uncertainties in the wind performance assessment of tall
buildings (Minciarelli et al. 2001, Diniz and Simiu 2005, Bashor et al. 2005 and Bernardini et al. 2014). In
this section, the details of the final step of the Wind Loading Chain, i.e., criteria are presented in a
probabilistic framework. Unlike the current prescriptive building code approaches, PBWE requires accurate
models of uncertainties that affect the final performance of the structure. In PBWE of tall buildings, the
stochastic nature of the wind field, wind tunnel test procedures, structural properties, and human perception
of motion are the main sources of uncertainties. In this paper, 15 random variables are used to model these
uncertainties. The uncertainties in the wind field were introduced by three random variables (ζ1, ζ2, ζ3, ζ7) in
the powerlaw wind profile equation. Aerodynamic uncertainties in the wind tunnel tests were modeled as
nondimensional random variables (ζ4, ζ5, ζ6). Uncertainties in structural properties were represented by
modeling the first 3 building frequencies (ω1, ω2, ω3) and structural damping values (ξ1, ξ2, ξ3) as random
variables using lognormal probability distributions. Fundamental frequencies from Eigenvalue analysis were
used as a mean value with Coefficient of Variation (COV) equals to 0.01. Table 3 summarizes the
parameters and type of probability distributions to model the considered uncertainties.
6 Structural reliability analysis
Structural reliability analysis through sampling was used to propagate the uncertainties through the Wind
Loading Chain to quantify the probability of exceeding the habitability and deflection criteria of the NBCC
(2010). The resultant horizontal (PFA), ar and lateral displacement (ur) responses are considered as
engineering demand parameters. Two limit states (performance functions) are defined as follows for a
vector of random variables (X) as follows:
(2)
(3)
Equation 2 represents the habitability limit state as a function of the criteria and the resultant horizontal PFA
under 10year return period mean wind speed. The excessive deflection limit state is given in Equation 3
as a difference between the uncertain criteria and 50year peak resultant lateral displacement. Since the
first two modes of the structure are translation, both the horizontal PFA and peak lateral displacement occur
at the top floor of the building. The probability of failure, (Pf) in Equation 2 and 3 can be obtained from the
joint probability density function of all random variables fX(x) using the multifold integral as follows.
(4)
Equation 4 can be solved using both analytical approximations and/or simulation techniques. Simulation
techniques are attractive for problems that involve many random variables and offer relatively accurate
results at the expense of computational cost (Kareem 1988). Therefore, in this paper structural reliability
analysis uses a mean centered Monte Carlo sampling technique to quantify the exceedance probability
(Pf). Meancentered Monte Carlo simulations were carried out for two principal wind directions (0 and 90
)()(, xaHCxaxg rr
)()(, xuDCxuxg rr
}0 )(, :{ )(
xaxgxD Xf rf dxxfP
Paper ID8
degrees) and six mean critical damping ratios, ξ, (0.25%, 0.5%, 1%, 1.5%, 3%, and 5%), and mean wind
velocity at the building height in the range of 1065 m/s.
Table 3. Parameters of random variables
Source of uncertainty
Random
variable
Mean
COV
Distribution
type
Citation
Wind hazard uncertainties
Observation and
analysis of wind
speed
1
1
0.025
Truncated
normal
Minciarelli et al.
(2001)
Conversion factors
between different
averaging times
2
1
0.05
Truncated
normal
Minciarelli et al.
(2001)
Characterization of
mean wind speed
profile
3
1
0.05
Truncated
normal
Bashor et al.
(2005)
Wind speed from
ASCE 710
7
ASCE 710
wind speed
map
0.07
Normal
Bernardini et al.
(2014)
Aerodynamic
uncertainties
Aerodynamic errors in
during wind tunnel
testing
4
1
0.05
Truncated
normal
Minciarelli et al.
(2001)
Approximations
during pressure
integration
5
1
0.05
Truncated
normal

Accuracy of similitude
concept in the wind
tunnel testing
6
1
0.05
Truncated
normal
Minciarelli et al.
(2001)
Structural
properties
Building frequency
ω1, ω2, ω3
Eigenvalue
analysis
0.01
Lognormal
Bashor et al.
(2005)
Critical damping ratio
ξ1, ξ2, ξ3
0.25%,0.5%,
1%1.5%,
3%, 5%
0.3
Lognormal
Bashor et al.
(2005)
Criteria
Habitability criteria
HC
25millig
0.2
Lognormal
Bashor et al.
(2005) and
NBCC (2010)
Deflection criteria
DC
(Building
height)/500
0.1
Lognormal
NBCC 2010
For each analysis, 2.5E5 random samples were utilized. The results of the Monte Carlo simulations were
used to develop fragility curves, i.e., plots that show the probability of exceeding the serviceability limit
states conditioned on wind velocity. These curves are important to estimate the vulnerability and associated
financial costs of the casestudy MTB. Figures 5 (a and b) present the fragility curves for habitability limit
state for wind AOA 0 and 90 degrees, respectively. In general, the effect of the critical damping ratio is also
observed in the fragility curves where vulnerability is higher for low critical damping ratio. For example, for
mean wind speed at the building height (VH) of 30 m/s and critical damping ratio of 1.5%, the failure
probabilities of 0o and 90o wind directions are 66.7% and 45%, while for 5% critical damping ratio the failure
probabilities are 40% and 10%, respectively. According to Bashor and Kareem (2009), major complaints
might occur when failure probabilities exceed 54%. Therefore, by increasing the critical damping from 1.5%
to 5%, it is possible to avoid occupant complaints due to excessive building’s motion.
Paper ID9
a)
b)
Figure 5: Fragility curves for habitability limit state: a) wind AOA = 0o, b) wind AOA = 90o
a)
b)
Figure 6: Fragility curves for deflection limit state: a) wind AOA = 0o, b) wind AOA = 90o
Fragility curves were also developed for deflection limit state. Figures 6 (a and b) show the variation of
failure probability with critical damping ratio (ξ) and wind AOA. As expected the probability of failure is higher
for wind AOA = 0o, as the worst wind direction matches with the weak axis of the building. The fragility
curves for wind AOA = 90 degrees are flat and in general vulnerability is smaller as compared to wind AOA
= 0 degrees. The obtained results from Monte Carlo simulations consistently show the dependence of wind
vulnerability on critical damping ratio and wind direction.
Irrespective of the wind angle of attack, under the 1in50 year wind storm, the probability of exceeding the
H/500 drift limit of NBCC 2010 (NRC 2010) is very small. Therefore, the performance of the study building
can be enhanced by using damping mitigations instead of increasing the lateral stiffness. For example, it is
possible to reduce the failure chance by ~27% by increasing the critical damping ratio from 1.5% to 5%.
This kind of damping enhancements can be achieved by using both passive and active supplemental
damping systems. To reduce excessive windinduced vibrations, tunedmass dampers (TMDs) successfully
applied in tall buildings (such as Taipei101 in Taiwan, Citicorp Center in New York, and John Hancock
Tower in Boston). Similar external damping systems can be designed and optimized using the results of
this paper to reduce the human discomfort risk in MTBs. The presented riskbased design and analysis
framework can be used to make riskinformed decision during the design of tall MTBs. By considering the
Paper ID10
reliability of the building as decision criteria, different hybridization techniques and external damping
systems can be analyzed, optimized and compared using the proposed framework.
7 Concluding remarks
In this paper, a riskbased wind design procedure is applied to design a 102 meters tall MTB by adapting
and extending the Alan G. Davenport Wind Loading Chain as a probabilistic PBWE framework. Initially,
aerodynamic wind tunnel tests were carried out to obtain transient wind load information. Subsequently,
using the wind tunnel data, the study MTB was structurally designed. In the riskbased performance
assessment, two limit states were considered, i.e., habitability and deflection. The 1in10year resultant
horizontal (PFA) and 50–year lateral deflection responses were considered as engineering demand
parameters. These limit states were established based on the habitability criteria and deflection limit of the
2010 National Building Code of Canada. Uncertainties related to the nature of the wind field (hazard), wind
tunnel tests, structural properties, and human perception of motion were explicitly modeled as random
variables using probability distributions. Structural reliability approach was used to propagate the
considered uncertainties through the Wind Loading Chain to quantify the probability of exceeding the NBCC
2010 criteria. Meancentered Monte Carlo simulations were carried out for six levels of mean critical
damping ratios and two principal wind directions. In total, 9E06 simulations were carried out to calculate
the probability of exceeding the criteria (failure probabilities). The results revealed the dependence of the
exceedance probability on the wind direction and critical damping ratio. For both limit states, wind
vulnerability increases when the critical damping ratio decreases.
The results of the structural reliability analyses indicated the possibility of reducing the risk of
unserviceability (possible complaints of occupants) by adding supplemental structural damping to the study
tall MTB. Irrespective of the wind angle of attack, under the 1in50 year wind storm, the probability of
exceeding the H/500 drift limit of NBCC 2010 (NRC 2010) is small. Therefore, the performance of the study
building can be enhanced by using damping mitigations instead of increasing lateral stiffness. For example,
it is possible to reduce the failure chance by ~27% by increasing the critical damping ratio from 1.5% to
5%. This can be achieved by using both, passive and active supplemental damping systems such as TMDs.
It should be noted that due to the strong dependence of the wind response of structures on shape
(aerodynamics), structural properties, surroundings, and hazard level, the obtained results are case
specific. However, the presented riskbased design and analysis procedure can be adapted to make risk
informed decisions during the design of other types of mid and highrise buildings.
Acknowledgments
Funding for this research was provided through Mitacs Accelerated Ph.D. Fellowship program in
collaboration with FPInnovations. The assistance from the members of the Boundary Layer Wind Tunnel
Laboratory (BLWTL) of the University of Western Ontario during the experimental program is noted with
appreciation. The help of Mr. Anant Gairola during the wind tunnel tests acknowledged with thanks. Special
appreciation is extended to Dr Workamaw Wardiso from CPP Wind Engineering Consultants and Benton
Johnson from Skidmore, Owings & Merrill LLP for valuable discussions.
References
ANSI/AWC NDS2015. 2015. National design specification (NDS) for wood construction 2015 edition.
American National Standards Institute (ANSI). Washington, D.C., USA and American Wood Council
(AWC), Leesburg, VA, USA
Baldassarra, C.F. 2017. The ICC’s Proposed Changes for Tall Wooden Buildings. Wiss, Janney Elstner
and Associates, Northbrook, Illinois, USA.
Bashor, R., KijewskiCorrea, T., & Kareem, A. 2005. On the windinduced response of tall buildings: the
effect of uncertainties in dynamic properties and human comfort thresholds. In Proceedings of
Americas Conference on Wind Engineering, Baton Rouge, LA, USA.
Paper ID11
Bashor, R., & Kareem, A. 2009. Probabilistic assessment of occupant comfort in tall buildings. In
Structures Congress 2009: Don't Mess with Structural Engineers: Expanding Our Role. Structures
Congress, Austin, Texas, United States.
Bernardini, E., Spence, S. M., Kwon, D. K., & Kareem, A. 2014. Performancebased design of highrise
buildings for occupant comfort. Journal of Structural Engineering, 141(10), 04014244.
Bezabeh, M. A., Tesfamariam, S., Stiemer, S. F., Popovski, M., and Karacabeyli, E. 2016. Direct
displacementbased design of a novel hybrid structure: Steel momentresisting frames with cross
laminated timber infill walls. Earthquake Spectra, 32(3), 1565–1585.
Bezabeh, M. A., Tesfamariam, S., Popovski, M., Goda, K., & Stiemer, S. F. 2017. Seismic base shear
modification factors for timbersteel hybrid structure: collapse risk assessment approach. Journal of
Structural Engineering, 143(10), 04017136.
Bitsuamlak, G. 2006. Application of computational wind engineering: A practical perspective. In Third
National Conference in Wind Engineering (pp. 119).
Canadian Standards Association (CSA). 2014. Engineering Design in Wood. CSA O8614. Toronto:
Canadian Standards Association.
Dagnew, A. K., & Bitsuamlak, G. T. 2014. Computational evaluation of wind loads on a standard tall
building using LES. Wind and Structures, 18(5), 567598.
Davenport, A. G. 1967. Gust loading factors. Journal of the Structural Division. 93(3), 1134.
Davenport A. G. 1970. Wind Loading on Tall Buildings. Symp. Wind Effects High Rise Bldgs., Evanston,
Ill., USA.
Davenport, A. G. 1982. The interaction of wind and structures. Engineering meteorology, 527572.
Diniz, S. M., & Simiu, E. 2005. Probabilistic descriptions of wind effects and windload factors for
databaseassisted design. Journal of structural engineering, 131(3), 507516.
FPInnovations 2010. A Synthesis of Research on Wood Products & Greenhouse Gas Impacts. Technical
Report No. TR19R.
Isyumov, N. 2012. Alan G. Davenport's mark on wind engineering. Journal of Wind Engineering and
Industrial Aerodynamics, 104, 1224.
Irwin, P., Denoon, R., & Scott, D. 2013. Wind Tunnel Testing of HighRise Buildings. Routledge.
Karacabeyli, E., & Lum, C. 2014. Technical guide for the design and construction of tall wood buildings in
Canada. PointeClaire, Quebec City, Canada.
Minciarelli, F., Gioffrè, M., Grigoriu, M., and Simiu, E. 2001. Estimates of extreme wind effects and wind
load factors: Influence of knowledge uncertainties. Prob. Eng. Mech., 16(4), 331–340.
NRCC (National Research Council of Canada). 2010. National building code of Canada, Associate
Committee on the National Building Code, Ottawa, Canada.
Pei, S., van de Lindt, J. W., Ricles, J., Sause, R., Berman, J., Ryan, K., Dolan J.D., Buchanan, A.,
Robinson, T., McDonnel E., Blomgren H., Popovski, M., Rammer, D., & Blomgren, H. 2017.
Development and FullScale Validation of ResilienceBased Seismic Design of Tall Wood Buildings:
The NHERI Tallwood Project. In Proceedings of the New Zealand Society for Earthquake Engineering
Annual Conference, April 2729, Wellington, New Zealand.
Popovski, M., Mohammad, M., Ni, C., Rezai, M., Below, K., Malczyk, R., & Sherstobitoff, J. 2014. Design
and construction of tall wood buildings: Input data, testing and advanced analysis. In Proceedings of
the 13st World Conference on Timber Engineering (WCTE 2014). Quebec, Canada.
Ramage, M., Foster, R., Smith, S., Flanagan, K., & Bakker, R. 2017. Super tall timber: design research
for the next generation of natural structure. The Journal of Architecture, 22(1), 104122.
Reynolds, T., Harris, R., & Chan, W. S. 2011. Dynamic Response of Tall Timber Buildings to Wind Load.
In 35th Annual Symposium of IABSE/52nd Annual Symposium of IASS/6th International Conference on
Space Structures, London.
Zhang, X. 2017. Seismic design of timber steel hybrid highrise buildings. PhD Dissertation, University of
British Columbia, Canada.
Skidmore, Owins and Merril (SOM). 2013. Timber tower research project.
Tesfamariam, S., Stiemer, S. F., Bezabeh, M., Goertz, C., Popovski, M., and Goda, K. 2015. Forcebased
design guideline for timbersteel hybrid structures: Steel moment resisting frames with CLT infill walls.
University of British Columbia, Okanagan, Canada
Zhang, X. 2017. Seismic design of timber steel hybrid highrise buildings. PhD Dissertation, University of
British Columbia, Canada.