Novel Techniques in Wind Engineering - Western University - Keynote ICWE 2015.pdf · M. Refan, H. Hangan. WindEEE: Downburst Research D. Parvu, A. Costache, H. Hangan –ICWE 14.

Novel Techniques in Wind Engineering

Horia HANGAN

INTRODUCTION

Pielke Jr. (1997):

minimize Vulnerability = f (Incidence, Exposure)

Incidence = f (Intensity, Occurrence, Frequency)

Exposure = f (Population, Property, Preparedness)

SUMMARY• New Laboratory: Novel WindEEE experiments

– Non-Synoptic Winds: Tornadoes, Downbursts

– New Flow and Structural Analysis

– Topography and Canopy Effects

– Multiscale Experiments: Wind Turbines, Solar Panels

– Measurement Techniques: Particle Tracking

• New Numerical: Multiple Space and Time Scales

– Mesoscale: WARF, Reanalysis

– Microscale: Urban wind environment

• New Full Scale: Real Space-Time Data

– Mobile Doppler Radar

– LiDAR

Climate

PDF, Vg, Vgr

ABL

V(z), Iu(z), S(f)

Aerodynamics

p(x,t)

Structural Response

x,a,M,F

LABORATORY

• WindEEE Dome : new three dimensional and time-dependent wind chamber

• can simulate various wind systems from sheared winds and gust fronts to tornadoesand downbursts

• a multi-scale, multi-purpose facility for wind research

The Wind Engineering Energy and Environment (WindEEE) Dome

www.windeee.ca

WindEEE: Preliminary Design

straight/sheared flow tornado flow downburst flow

WindEEE: Engineering Design

• 106 individually controlled fans

• 2 MW maximum power

• 5 m lift and turntable

• 1600 floor roughness elements

• 1000+ tons of steel

• 1850 m³ of concrete

• LEEDs Silver accreditation

WindEEE: Research Ready

Six Initial Design Specifications:

- Straight Mode Uniform- Straight Mode Boundary Layer- Straight Mode Shear- Tornado- Downburst- Reversed Flow Mode

+ HH 7

WindEEE: Tornado Research

M. Refan, D. Parvu, H. Hangan – ICWE14

WindEEE: Tornado Research

Data Analysis: Velocity Correlations

Fan

an

gle

10

°Fa

n a

ngl

e 2

0°

Fan

an

gle

30

°

Height 0.035 m

0.045 m 0.070 m 0.080 m 0.150 m

vu

jiji vvuunCorrelatio

,,

R. Ashton, M. Refan, H. Hangan, G. V. Iungo

Data Analysis: Independent component analysis

L. Carassale, M. Refan, H. Hangan

WindEEE: Tornado Research

(a)

(b)

(c)

(d)

(e)

(f)

• Generic Industrial and Hospital building shapes

• Tests in BLWTL and WindEEE

• Determine the loading differences

M. Refan, H. Hangan

WindEEE: Downburst Research

D. Parvu, A. Costache, H. Hangan – ICWE 14

WindEEE: Downburst Research

- Downburst TL Interaction- H/D < 1; H/D > 1- Roughness Effects- ElDamatty, Bitsuamlak,

Savory, Hangan

A. ElDamatty, A. ElAwady – ICWE14

Wind-Structure: Thunderstorm Response Spectrum

eqf

eq d 0ˆf f S n ,

eq d,eq 1ˆ S n , , f f

SDOF system

NDOF system0.002

0.01

0.05 Solari et al., W&S, 2015

Solari et al., JWEIA, 2015

Solari & De Gaetano, ICWE14

time

sp

ee

d (

m/s

)

eq,Nf

eq,2f

eq,1f

N

2

1

Wind-Structure:Gust-Front Factor (GG-F)

Modeling and analysis of thunderstorm/downburst generated gust-front wind loads effect on structures

Web-enabled module to facilitate the use of the GFF framework

http://gff.ce.nd.edu

1) Kwon, D. K., and Kareem, A., "Gust-front factor: New Framework for Wind Load Effects on Structures." Journal of Structural Engineering, ASCE, 135(6), 717-732, 2009.

2) Kwon, D. K., Kareem, A. "Generalized gust-front factor: A computational framework for wind load effects." Engineering Structures, 48, 635-644, 2013.

WindEEE: Solar Panels Research Full-scale: ground and roof mounted solar array at WindEEE Dome

WindEEE: Solar Panels Research

Pressure + force balance + strain gauge testing

Z. Samani, G. Bitsuamlak, H. Hangan

• located near Roskilde, Denmark

• 12 m high peninsula with steep escarpment

• shows topographical similarity to wind turbine sites in complex terrain

• provides meaningful test case for model validation

• comprehensive mast data available

WINDEEE : Topography – Bolund Experiment

• 1/25 Scale Model

• Large Scale PIV: 2 x 1.5 meters

• 4 simultaneous cameras

• Window overlapping

• Several exposures

WINDEEE : Topography – Bolund Experiment

WINDEEE : Topography – Bolund Experiment

0

5

10

15

20

25

30

35

40

0 5 10 15 20

Fu

ll S

cale

Hei

gh

t (m

)

U (m/s), TI (%)

WindEEE (Mean velocity, U)

0

2

4

6

8

10

12

14

16

18

20

0 10 20 30F

ull

Sca

le H

eig

ht

(m)

s/u*

Full Scale Bolund DataWindEEE Data

• Cameras 1 and 2

• Velocity Snapshots

WINDEEE : Topography – Bolund Experiment

WINDEEE : Canopy – PEI Experiment

WINDEEE : Canopy – PEI Experiment

Porosity based Forest Canopy Modeling

LAI (Leaf Area Index) measured indirect

Satellite data (MODIS)-> obtain LAI estimates

LAI distribution mapped over terrain

D. Parvu, H. Hangan

WINDEEE : Wake Experiment

Phase-Locked PIV measurements

8 azimuthal angles between 0 and 120

Two axial locations: x/R =1 and x/R =2

At each azimuthal angle 4 PIV tiles

P. Hashemi-Tari, K. Siddiqui, H. Hangan – Wind Energy (2015)

WindEEE: Wake Experiment

Radial profiles of axial deficit velocities at X=Rand X=2R for 8 azimuth angles of blade

WindEEE: Wake Experiment

Radial profiles of radial velocities at X=R and X=2R for 8 azimuth angles of blade

WindEEE: Wake Experiment

Radial profiles of Turbulence Intensity Streamlines of Instantaneous Velocity

WindEEE: Wake Experiment

Scaling Effects

WindEEE: Particle Tracking Techniques

D. Parvu, A. Costache, M. Refan, H. Hangan – ICWE14

NUMERICAL

WRF

Data Reanalysis

Urban Microscale CFD

Convective forcing modeling

1

2

34

Numerical Modeling: NCEP / NCAR Reanalysis

Kalnay et al (1996) data from 1948 to Dec. 2014

2.5° latitude by 2.5° longitude spatial resolution

bilinear interpolation method

4 times a day, daily, and monthly

Mean daily values for two wind components 67 years, 24473 data records

Numerical Modeling: Trend Data Analysis

Mean Annual Wind Speed per Direction

Mann-Kendall non-parametric test for trend (Mann, 1945; Kendall, 1970)

Sen’s slope estimator (Sen, 1968)

𝑆 =

𝑦1=1

𝑛−1

𝑦2=𝑦1+1

𝑛

𝑠𝑔𝑛 𝑥𝑦2 − 𝑥𝑦2 .

𝑌 = 𝑄 𝑦 − 1948 + 𝐵,

D. Romanic, H. Hangan-Sustainable Cities and Society (2015)

Numerical Modeling: Spectral Analysis

Low frequency wind spectra (blue line)

95% confidence intervals (grey line)

Welch method (Welch, 1967)

Sunspots: SILSO World Data Center, 1948

13-month moving average

applied on mean monthly wind speed data

σ.995 confidence level

Numerical Modeling: Urban Environment

WRF Nested run (4 domains)

~30 million cells

Chugach supercomputer 1024 cores

~20-30 min

Kraken supercomputer 576 cores

~ 120 min

CFD and downscaling ~3 million cells

~15 min

Total time ~< 1 hour

Numerical Modeling: Urban Environment

7.6% (9) (4) 5.8%

4.9 % (13)

12.3% (8) 9.6% (4)

Numerical Modeling: Downburst and Tornado

FULL SCALE

ROTATE Campaign

Doppler on Wheels

GBDTV Analysis

Similarity Analysis

PEIWEE Campaign

LiDAR

Topography

Canopy

Wake

Recent Campaigns: ROTATE

Data provided by CSWR

ROTATE=Radar Observations of Tornadoes And Thunderstorms Experiment – 2012Ground-Based Velocity Track Display

Single-Doppler radar data of five tornadoes: Kellerville, TX 1995 (F4), Spencer, SD 1998 (F4), Stockton, Oklaunion, TX 2000 (F1), Stratford, TX 2003 (F0), KS 2005 (F1), Clairemont, TX 2005 (F0), Happy, TX 2007 (EF0) and Goshen County, WY 2009 (EF2)

Nine tornado volumes: cover wind speeds associated with EF0 to EF3 rated tornadoes

Elie, Manitoba tornado 2007

Bennington Kansas EF-4 tornado 2013

Tornadoes: GBDTV Analysis

DOW 3

Doppler velocity (m/s) contour map of the Happy, TX 2007 tornado at 0203:20

UTC and at 0.3˚ radar beam angle

r (m)

z(m

)

0 200 400 600 8000

200

400

600

800

1000

Vtan (m/s)

4038363432302826242220181614121086420

5

Tornadoes: Full Scale Data

Volume

Clairemont,

volume1

(Clr v1)

Happy,

volume1

(Hp v1)

Happy,

volume2

(Hp v2)

Goshen ,

volume1

(GC v1)

Goshen ,

volume2

(GC v2)

Goshen,

volume3

(GC v3)

Stockton,

volume1

(Stc v1)

Spencer,

Volume1

(Sp v1)

Spencer,

Volume2

(Sp v2)

EF EF0 EF1 EF0 EF1 EF1 EF1 EF2 EF3 EF3

zmin (m) 25 71 38 97 75 30 43 51 85

Vtrans (m/s) 1.2 19.4 19.4 9.49 9.49 9.49 10.95 15 15

Vtan,max

(m/s)36.3 39 37.9 41.6 42 42.9 50.2 58.2 62

rc (m) 96 160 160 150 150 100 220 192 208

zmax (m) 200 250 50 42 160 41 40 40 40

Vertical

structureVBA 1-cell TD 2-cell VBA 2-cell After TD 2-cell 2-cell

Tornadoes: Scaling

R (km)

Vta

n(m

/s)

0 0.2 0.4 0.6 0.8 10

10

20

30

40 z= 40 mz= 80 mz= 120 mz= 160 mz= 200 mz= 240 mz= 280 mz= 320 m

Goshen County , WY 2009 (EF2)

• The overall maximum

tangential velocity

Vtan,max = Vtan(rc,max, zmax)

• Length scale

rc,max,D/rc,max,S

zmax,D/zmax,Sλl=

r (km)

λv=

• Velocity scale

Vtan,max,D/Vtan,max,S

Tornadoes: Scaling

Hp v1 GC v2 Sp v2r (m)

Vta

n(m

/s)

0 200 400 600 800 10000

5

10

15

20

25

30

z=120mz=136mz=160mz=155mz=200mz=194m

r (m)

Vta

n(m

/s)

0 200 400 600 800 10000

5

10

15

20

25

30

z=200mz=216mz=250mz=247m

r (m)

Vta

n(m

/s)

0 200 400 600 800 10000

10

20

30

40

z=80mz=77mz=120mz=120mz=160mz=137m

Tornadoes: Scaling

Swirl ratio

Len

gth

scale

,

l

0 0.2 0.4 0.6 0.8 1 1.2 1.40

2000

4000

6000

Before touch-down After touch-down

Clr v1

Hp v1

Hp v2

Stc v1

GC v1

GC v2

GC v3Sp v1

Sp v2

Swirl ratio

Velo

city

scale

,

v

0 0.2 0.4 0.6 0.8 1 1.2 1.40

1

2

3

4

5

Hp v1

Clr v1

Hp v2

GC v1GC v2

GC v3

Sp v1

Sp v2Stc v1

>F2

• Monte Carlo Simulations for 30,000 years

• Minimum Return Period = 4,000 years/sq.km

• Maximum in Oklahoma, Minimum in Nevada

>F4

• Monte Carlo Simulations for 18 million years

• Minimum Return Period = 16,700/sq.km95% F0, F1 and F2; Only 0.1% F5

F5=0.1*F4=0.02*F3=0.006*F2

Wind Incidence: Frequency

1921-1995 Data base by Grazulis (1993), Monte Carlo Simulations by Meyer et al. (2005)

• Width increases with F scale

- Median F2 = 100 m

- Median F4 = 600 m

• Length increases with F scale

- Median F2 = 10 km

- Median F4 = 60 km

Wind Incidence: Width and Length

Recent Campaigns: PEIWEE

• Cornell University:

– Two Lidars -> Zephyr and Gallion

• Western University (WindEEE RI and DTU):

– 1 Short Range Lidar

– 1 Quadropter

• Wind Energy Institute of Canada (WEICan):

– P.E.I site with 5 wind turbines

– Masts 80 m, 60 m, 17 m and 15 m

– 1 Zephyr Lidar

• York University:

– 6 masts of 10m

PEIWEE: WindScanner short-range LiDAR

PEIWEE: WindScanner short-range LiDAR

WindEEE WindScanner : short-range LiDAR

max. wind speed acquisition rate : 500 samples/s

PEIWEE : Wake / Topography

• Line and flower pattern scanning

• Multiple heights : 10, 15, 20 , 40, 80 m

• Multiple tilt angles : 0°, 30°, 60° and 90°

• Multiple wind directions : S to W

PEIWEE : Cliff Measurements

PEIWEE : Forest Edge Measurements

PEIWEE : Wake Measurements

DISCUSSION

Climate• Meso-scale Models + Full Scale: set proper boundary conditions

Terrain• Micro-scale Models + Full Scale: run simulations in the surface layer

Stats• Statistical Analysts: set incidence models

Wind3D• Wind Fields: 3D and Time-Dependent; Multiscale

Loads • Aerodynamic Loads: Loads = f(buildings/structures, exposure)

Response • Structural Analysis: Responses to Loads; Collapse modes

CONCLUSIONS

• New Tools for the Wind Engineering Chain • New Laboratory

Climate: ABL vs. Non-Synoptic Winds-> Flow Fields Topography and Roughness -> Reynolds and 3D effects Aerodynamic Loading-> Comparison of ABL vs. Non-Synoptic Analysis Techniques-> Spectral vs. Time-Domain vs. Modal Statistical Analysis

• New Full Scale– Doppler Radar + GBDTV; LiDAR

• New NumericalRe-Analysis, Meso-Micro coupling, Physical Simulations

REFERENCES

- Pielke Jr., R.A., Refraining the U.S. hurricane problem, Society &Natural Resources: 1997.- Grazulis, T.P., Significant Tornadoes, 1680-1991. Environmental Films, St. Johnsbury, VT, 1326, 1993.- Meyer, C. L, Brooks, H. E. and Kay, M. P., A hazard model for tornado occurrence in the United States, 16th Conference on Probability and Statistics in the Atmospheric Sciences, 2002.- Xu, Z. and Hangan, H., Scale, boundary and inlet condition effects on impinging jets with application to downburst simulations, J. of Wind Eng. and Ind. Aerodynamics, 96,2008.- Hangan, H. and Kim, J.D.*, Numerical characterization of impinging jets with application to downbursts. J. of Wind Eng. and Ind. Aerodynamics, 95, Issue 4,2007.- Refan, M.*, Hangan, H., Wurman, J., “Reproducing Tornadoes in Laboratory Using Proper Scaling”, J. Wind Eng. And Ind. Aerodynamics, 2014-Hashemi-Tari, P*.,Hangan, H., Siddiqui, K., “Flow characterization in the Near-wake region of a Horizontal Axis Wind Turbine”, Wind Energy (2015) - Romanic, D.*, Rasouli, A.*, Hangan, H., “Wind resource assessment in complex urban environment”, Wind Engineering, Vol. 39, Nr. 2, January 2015- Romanic, D., Hangan, H., “Wind Climatology of Toronto based on NCEP/NCAR reanalysis 1 data set and its potential relation to solar activity, Sustainable Cities and Society (2015)

-

Thank You !

www.windeee.ca