On September 20, ICLR conducted a Friday Forum webinar titled 'Reducing the risk of extreme wind induced damage', led by Dr. Girma T. Bitsuamlak of Western University Engineering. In 2018, Canada’s insurers paid $1 billion in wind damage claims in Ontario and Quebec, a record high. Most years, wind damage is the largest cost for insurance companies in the United States. Climate studies warn that the frequency and severity of extreme hurricanes and tornadoes may increase as a result of global warming. Losses have been greatest for residential buildings, in part because they are not subject to engineering design and construction. Emerging engineering research provides a new understanding of how extreme wind can result in damage to structures. In particular, studies are beginning to focus on design and construction of individual buildings and also on neighborhoods. Wind-resilient building design and retrofit can now be supported by realistic computational modelling of urban microclimate interaction with buildings at various scales (component → building → neighbourhood → city). Alan Davenport’s “wind-loading-chain” links the modelling of extreme wind, exposure, aerodynamics, and dynamics to particular design criteria. This includes hurricane, tornado and other extreme wind events. Realizing wind-resilient communities will be presented through research projects that include modelling of the neighbourhood impact of 2010 hurricane Ivan and the 2018 tornado in Dunrobin. The models are able to evaluate risks based on wind pressure on each building at a neighbourhood scale. Modelling at neighbourhood scale is used to assess windborne debris, and wind-driven rain that contributes to large portion of overall damage. Computational modelling will play significant roles in enhancing community resiliency. Dr. Girma T. Bitsuamlak, PhD, PEng, F CSCE is Associate Professor and Canada Research Chair in Wind Engineering, Civil and Environmental Eng., Director at Boundary Layer Wind Tunnel Lab. (BLWTL) and WindEEE Research Institute, SHARCNET Site Leader, Western University.

1. Reducing the risk of extreme wind induced
damage
Girma T. Bitsuamlak, PhD, PEng, F CSCE
Canada Research Chair in Wind Engineering, and Associate Professor,
Western Site Leader for SHARCNET,
Director (Research) at Boundary Layer Wind Tunnel Laboratory, and WindEEE Research Institute,
Department of Civil and Environmental Engineering, The University of Western Ontario (UWO)
September 20, 2019
1

2. Agenda
• Introduction
• Recent advances in Computational Wind Engineering (CWE)
• Hurricane impact assessment on individual and
group of buildings
• Tornado impact assessment on individual and
group of buildings
2

3. 3
tornado
Hurricanes Source: NY Times
heat wave, energy consumption, UHI
flood
Climate stressors / consequences
“Last year, Canada’s insurers paid more than $1 billion in wind damage claims” - ICLR

4. AfterBefore
Two main tornadoes that struck Dunrobin-Gatineau and Nepean area categorized as EF2 and EF3. Loss estimate $300M
Recent extreme wind events in Canada: Dunrobin tornado
4

5. WindResearch@ Western
5
• Home of critical mass of highly experienced wind engineers,
technicians and researchers; four modern wind test
facilities; and dedicated High performance computing
center.
• Boundary Layer Wind Tunnel Laboratory
• The Insurance Research Lab for Better Homes
• The Advanced Facility for Avian Research
• WindEEE Dome
• WindEEE cluster, SHARCNET and SOSCIP
High performance computing

6. Prof. A.G. Davenport (1932–2009)
Alan G. Davenport “Wind Loading Chain”
6

7. Wind loading
Top floor acceleration
Deflection,
Pedestrian level wind
speed, etc.
AGD wind loading chain
7

8. 8
WindEEE Dome

9. Recent Advances in Computational Wind
Engineering
9

10. Computational wind engineering: Driving factors
10
• Increased frequency of natural hazards
• Multi-scale (component, building, neighborhood), multi-physics (wind,
rain, snow, debris) and the wide range of temporal and spatial
atmospheric boundary layer flow scales (e.g. turbulence) limits the
applicability of experimental approaches
• Growth of computational power and algorithm development
• The need for accurate and timely wind-induced loss prediction

11. Climate models
Surface roughness from Lidar measurement
Wind
induced
dynamic
excitation
Interaction of the built environment with wind
Wind loading chain based on multi-scale and multi-physics CFD simulation
11
Roughness model

12. Damage prediction of individual and group of
buildings under hurricane
12

13. 13
Hurricane impact assessment on individual and group of
buildings involves:
– Development of high-fidelity CFD models
– Validation of CFD models with wind tunnel test data
– Residential community modeling
– Modeling wind-driven rain ingress into residential buildings

14. 14
CFD simulation of a full wind tunnel
Target wind tunnel test CAD model of wind tunnel

15. CFD simulation of full wind tunnel: turbulence structure
15
Roughness blocks
Spires
Barrier
Test section

16. 16
CFD wind field comparison with wind tunnel

17. 17
WT: NIST database test
at UWO
LES : current CFD
simulation
Validation of
surface
pressure
coefficients

18. 18
UWO : NIST database test at UWO
LES : current Large Eddy Simulation
peak Cpmean Cp r.m.s Cp
Statistics of extreme
surface pressure
coefficients

19. • One of the neighborhoods monitored in Florida Coastal Monitoring Program in 2004
Neighbourhood scale simulations for a residential community in Florida
19
N

20. 20
Wind
Neighbourhood scale simulations for a residential community in Florida

21. 21
Wind
Wind pressure on a residential building with complex roof

22. 22

23. Progressive failure under wind events
23
Progressive aerodynamics
C0) Building with secure
enclosure
C1) After damages to
windward windows
C3) After damages to side
windows, clearstory
C4) After damages to side
windows, roof, and leeward
windows

24. Study building, wind flow zones and progressive aerodynamics
24

25. Damage scenario and wind-driven rain intrusion
No damage Damaged garage door Damaged windward doors, windows,
and some portion of roof
Damaged external windows,
doors, and about half of the roof
☼

26. Florida neighbourhood in Hurricane Ivan of 2004
26
Damage scenario and wind-driven rain intrusion
Damaged garage door
Damaged windward doors, windows,
and some portion of roof
Damaged all external windows,
doors, and about half of the roof
Average rain fall used is 1.4 mm/hr and V10 = 3m/sec

27. 27
Tornado impact assessment on individual and group
of buildings

28. • Numerical simulation of tornado-like wind-field
• Development of a generic numerical tornado model
• Interaction of tornado with a building
• A neighborhood scale assessment: Dunrobin tornado case study
28
Tornado impact assessment on individual and group of buildings involves:

29. 29
Tornado-like flow field
Tornado flow-field
• Geometric: 𝑎 =
#$
%$
, 𝑟 = 𝑟', 𝑣) ≈ 0
• Kinematic: 𝑆 = [
./
.0.23
]%5%$
= [
%$ 𝚪8
29)
]%5%:
• Dynamic: 𝑅𝑒% =
9
2=

30. 30
Engineering tornado modeling
VorTECH (TTU) Tornado Simulator (ISU) WindEEE Dome (UWO)

31. Vortex evolution
Simplified CFD modelFull CFD model Experiment/CFD comparison Full/simplified
CFD comparison
VorTECH(TTU)TornadoSimulator(ISU)WindEEEDome(UWO)

32. 32
Aerodynamic test results
(a) 0 ° (b) 30 °
(c) 60 ° (d) 90 °
𝐶𝑝 =
𝑃 − 𝑃'
1
2 ρ𝑣%FG
2
• 𝑃' is the reference atmospheric pressure
• 𝑣%FGis the mean tangential velocity at building height
Stationary tornado case Translating tornado case

33. 33
Validation of building surface pressure distribution
External building Cp (CFD) External building Cp (WindEEE)
0° Orientation

34. 34
Effect of topography on tornado flow-field
Synoptic flow
Tornadic flow
Steep-Hill
Flat ground

35. Tornado impact on Dunrobin neighborhood (EF3 tornado)
Courtesy: Northern tornadoes project (NTP) at UWO

36. 36
Tornado impact on Dunrobin neighborhood (EF3 tornado)
AfterBefore

37. Dunrobin tornado wind field description
Thin, laminar appearing tornado in
nature (Manitoba, 2007) : single-
cell structure
Numerical simulations of laminar
appearing vortex obtained by
controlling swirl ratio
“Fuzzy”, turbulent appearing
tornado in nature (El-Reno, 2013)
: multi-cell structure
Numerical simulations of multi-
cell vortex obtained by
controlling swirl ratio
• Ideally: Doppler radar velocity measurements can be used as
target to calibrate numerical simulations
• In absence of Doppler radar: qualitative estimation of vortex
parameters (aspect ratio, swirl ratio)
• Aspect ratio: 0.5 (range in nature 0.1-0.9)
• Swirl ratio: 0.65-0.85 (qualitative inspection of available videos)
• Target core diameter at the ground level: ~250-300 m (same
order of magnitude as the damaged neighborhood)
• EF3 rating: 62.5 m/s to 73.6 m/s (3-sec gust)
Source: Environment Canada + NTP
• EF3 speed is set to 𝑣H + 𝑣J%3KLM3JNOK = 𝑣J3%PFJ
• Average translation speed
estimated 15 m/s (based on
damage length and duration)
• A representative time taken by
the vortex to travel through the
neighborhood (𝑡L = 40𝑠)
• EF scale wind speed converted
approximately from 3-s gust to a
40 s average to obtain target 𝑣H
for a stationary tornado during
calibration stage.

38. 38
Donrobin model/physics setup

39. Stationary tornado wind field calibration
• EF-3 wind speed target (40-s average):53 m/s-62.5 m/s
• 𝑣H + 𝑣J%3KLM3JNOK5 𝑣J3%PFJ
• Average near ground tangential velocity: 38m/s -47.5 m/s
• Maximum near ground tangential velocity achieved (𝑣H,V3W) ~38m/s
• Near ground core diameter (𝑑Z)~300 m (engulfing the neighborhood)
𝑣H,V3W
𝑑Z

40. 40
Translating tornado over Dunrobin neighborhood

41. 41
Wind borne debris
𝑈G =
𝜌V 𝑡𝐼𝑔
0.5𝜌3bc
Debris flight speed estimation (Wills et al.)
Debris specification Flight speed
Timber rod (d=10mm) 11 m/s
Timber sheet (100mm x 50 mm) 32 m/s
110 mm long wooden missile 30 m/s
20 mm stone missile 30 m/s
Debris classification
EF0 EF1
EF2 EF3 EF4
Tornado size picked based on U.S. Nuclear Regulatory Commission, Regulatory Guide 1.76,
Design-Basis Tornado and Tornado Missiles for Nuclear Power Plants, Revision 1, 2007.

42. • Current research team: Tibebu, Anwar, Meseret, Matiyas, Barilelo, Abiy, Kimberly, Anant,
Chris, Tsinuel, Eric, Matt, Hadil, Muna, Shea, Tewodros, Kenny, Cody, Shea, Kate,
Dagmawi
• All former graduate students, postdocs and visiting scholars
• Current collaborators: Fitsum from BCIT, Solomon from UBC, Popovski from
FPInnovation, Laxmi from McGill, Arindam from FIU, Kaoshan from Tongji, David from
McMaster, Shaker from Ryerson, Miriam from UWO, Martha from UWO and CEE faculty at
UWO
• Funding agencies - NSERC CRC, Discovery, CFI, CRD; OCE; and SOSCIP
• Industry supporters - FM Global, FPInnovations, IBM, Wausau Tile Inc., Theakston Inc.,
Kilmaat, and Stephenson Engineering
• Western’s Boundary Layer Wind Tunnel Laboratory, WindEEE Research Institute, and
SHARCNET
• Wind research group at UWO
• FIU, and UF
Acknowledgment