Novel Techniques in Wind Engineering Horia HANGAN INTRODUCTION - PowerPoint PPT Presentation

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 Climate – New Flow and Structural Analysis PDF, V g , V gr – Topography and Canopy Effects – ABL Multiscale Experiments: Wind Turbines, Solar Panels – Measurement Techniques: Particle Tracking V(z), I u (z), S(f) • New Numerical: Multiple Space and Time Scales Aerodynamics – Mesoscale: WARF, Reanalysis p(x,t) – Microscale: Urban wind environment Structural • New Full Scale: Real Space-Time Data Response – Mobile Doppler Radar x,a,M,F – LiDAR

LABORATORY The Wind Engineering Energy and Environment (WindEEE) Dome • WindEEE Dome : new three dimensional and time-dependent wind chamber • can simulate various wind systems from sheared winds and gust fronts to tornadoes and downbursts • a multi-scale, multi-purpose facility for wind research 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 Height 0.035 0.045 m 0.070 m 0.080 m 0.150 m Fan angle 10° m      u u v v  i , j i , j Correlatio n   Fan angle 20° u v Fan angle 30° R. Ashton, M. Refan, H. Hangan, G. V. Iungo

Data Analysis: Independent component analysis L. Carassale, M. Refan, H. Hangan

WindEEE: Tornado Research • Generic Industrial and Hospital building shapes (c) (a) (b) • Tests in BLWTL and WindEEE • Determine the loading differences (d) (e) (f) 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 speed (m/s) f eq time SDOF system   f ˆ    N f f S n , eq,N eq d 0 NDOF system   0.002   ˆ S f     f f n , , 2 eq,2 eq d,eq 1   0.01 f 1 Solari et al., W&S, 2015 eq,1   0.05 Solari et al., JWEIA, 2015 Solari & De Gaetano, ICWE14

Wind-Structure: Gust-Front Factor ( G G-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

WINDEEE : Topography – Bolund Experiment • 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 Full Scale Bolund Data WindEEE (Mean velocity, U) WindEEE Data 20 40 18 35 16 Full Scale Height (m) 30 14 Full Scale Height (m) 25 12 10 20 8 15 6 10 4 5 2 0 0 0 5 10 15 20 0 10 20 30 U (m/s), TI (%) s/u *

WINDEEE : Topography – Bolund Experiment • Cameras 1 and 2 • Velocity Snapshots

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=R  and 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 Streamlines of Instantaneous Velocity Radial profiles of Turbulence Intensity  

WindEEE : Wake Experiment Scaling Effects 

WindEEE : Particle Tracking Techniques D. Parvu, A. Costache, M. Refan, H. Hangan – ICWE14

NUMERICAL 1 2 WRF  3 Data Reanalysis  4 Urban Microscale CFD  Convective forcing modeling 

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) 𝑜−1 𝑜 𝑇 = 𝑡𝑕𝑜 𝑦 𝑧 2 − 𝑦 𝑧 2 . 𝑧 1 =1 𝑧 2 =𝑧 1 +1 𝑍 = 𝑅 𝑧 − 1948 + 𝐶, 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 (4) 5.8% 7.6% (9) 12.3% (8) 9.6% (4) 4.9 % (13)

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=R adar O bservations of T ornadoes A nd T hunderstorms E xperiment – 2012 G round- B ased V elocity T rack D isplay Single-Doppler radar data of five tornadoes: Elie, Manitoba tornado 2007 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 Bennington Kansas EF-4 tornado 2013

Tornadoes: GBDTV Analysis V tan (m/s) 40 1000 38 36 34 32 800 30 28 26 24 22 600 z (m) 20 18 16 14 12 400 DOW 3 10 8 6 4 200 2 0 5 0 0 200 400 600 800 Doppler velocity (m/s) contour map of r (m) the Happy, TX 2007 tornado at 0203:20 UTC and at 0.3˚ radar beam angle

Tornadoes: Full Scale Data Clairemont, Happy, Happy, Goshen , Goshen , Goshen, Stockton, Spencer, Spencer, volume1 volume1 volume2 volume1 volume2 volume3 volume1 Volume1 Volume2 Volume (Clr v1) (Hp v1) (Hp v2) (GC v1) (GC v2) (GC v3) (Stc v1) (Sp v1) (Sp v2) EF EF0 EF1 EF0 EF1 EF1 EF1 EF2 EF3 EF3 z min (m) 25 71 38 97 75 30 43 51 85 V trans (m/s) 1.2 19.4 19.4 9.49 9.49 9.49 10.95 15 15 V tan,max 36.3 39 37.9 41.6 42 42.9 50.2 58.2 62 (m/s) r c (m) 96 160 160 150 150 100 220 192 208 z max (m) 200 250 50 42 160 41 40 40 40 Vertical VBA 1-cell TD 2-cell VBA 2-cell After TD 2-cell 2-cell structure

Tornadoes: Scaling • The overall maximum 40 z= 40 m tangential velocity z= 80 m z= 120 m z= 160 m V tan,max = V tan ( r c,max , z max ) z= 200 m 30 z= 240 m z= 280 m V tan (m/s) z= 320 m Velocity scale • 20 V tan,max,D / V tan,max,S λ v = Length scale • 10 r c,max,D / r c,max,S Goshen County , WY 2009 (EF2) λ l = z max,D / z max,S 0 0 0.2 0.4 0.6 0.8 1 R (km) r (km)