Publishers version - DTU Orbit

More documents

Recommendations

Info

15 SAR for wind energy Charlotte B. Hasager and Merete Badger DTU Wind Energy, Risø Campus, Roskilde, Denmark 15.1 Introduction Satellite Synthetic Aperture Radar (SAR) can provide a wide range of information about the surface of the Earth. SAR instruments have been flown on several satellite platforms since the early1990s.OneofthepurposesofthesatelliteSARsisresearch.Thesuccessfulinvestigations have led to a variety of (near-) operational products including sea ice mapping, oil spill detection, ship movement, mapping of flooded land surfaces, earth quake, land slide and subsidence mapping, digital elevation mapping, urban development, vegetation and biomass changes, glacier coverage, ocean wave and ocean current information, and last but not least ocean surface wind mapping. One advantage of SAR compared to optical remote sensing is that the radar carries its own illumination source, and is thus independent of daylight. This is particularly useful near the Arctic and Antarctic where daylight is limited for several months per year. SAR operates in the microwave bands. Microwave radiation is able to penetrate clouds and precipitation. SARs are all-weatherinstruments and so not limited by cloud cover. This is particularly useful in cloudy and rainy scenarios including hurricanes. Ocean surface wind mapping from SAR has been described in numerous articles. A recent state-of-the-art white-paper on ‘Wind retrieval from Synthetic Aperture Radar - an overview’ from the SEASAR 2012 workshop ‘Advances in SAR Oceanography’ by the European Space Agency(ESA) (Dagestadet al.,2013)summarizesthetechnicalfundamentalsofsatelliteSAR ocean surface wind retrieval. A wide range of applications are also presented including ocean wind mapping for weather prediction, wind farming, tropical cyclones, polar lows, katabatic winds, gap winds, vortex streets, boundary layer rolls and atmospheric gravity waves. See this paper for references (157 in total). SAR measurements are high-resolutionobservations of the Earth surface. Although no SAR sensor has been designed specifically for wind mapping, it has become clear that SAR data is very suitable for high-resolutionwind retrievals over the ocean includingnear-shore areas. The spatial resolution of SAR makes it particularly useful for resolving mesoscale wind variability. Planning of offshore wind farms has emphasized the need for reliable ocean wind observations. Ocean wind observations are generally costly to obtain from meteorological masts or ground-based remote sensing instruments. Furthermore, such data is only valid near the local point at which it is measured. In contrast, satellite SAR can provide spatially resolved ocean wind information. Most potential offshore wind farm sites are covered by archived SAR data. Wind resource mapping can thus be performed without any delay whereas it takes time to plan and conduct a ground based observational campaign. 15.2 SAR technical description SAR is an active microwave sensorwhich transmits coherent microwaves.The imagesshowing the normalized radar cross section (NRCS) are made from advanced signal processing of the originalobservations.NRCSistherecordedbackscatteredsignalperunitarea.Themicrowaves transmitted and received are either vertically (V) or horizontally (H) polarized. Co-polarized (VV or HH) images are made if the same polarization is used for both transmitting and receiving. Cross-polarized (VH or HV) images are made otherwise. The co-polarized NRCS has traditionally been used for ocean surface wind retrieval at spatial pixel scales finer than 1 km. Cross-polarized NRCS has been tested for wind retrieval. 276 DTU Wind Energy-E-Report-0029(EN)
Also the Doppler Centroid anomaly can be used for wind retrieval. This method is particularly of interest for investigation of wind direction. This chapter focuses on wind retrieval from co-polarized SAR data (hereinafter co-pol). The microwave wavelength in SAR ranges from around 2.5 to 30.0 cm. According to the standard radar frequency letter-band nomenclature (IEEE standard 521-1984), the X band wavelength band is 2.5−3.8 cm (8−12 GHz), C band 3.8−7.5 cm (4−8 GHz), S band 7.5−15.0 cm (2−4GHz) and L band 15.0−30.0 cm (1−2 GHz). In brackets is given the frequency bands. Recent satellite SARs are listed in Table 1. Updated information is available at http: //database.eohandbook.com.Longprior to the SARs listed in Table 1 the L-band SEASAT SAR was flown for three months in 1978. All SAR sensors provide co-pol NRCS data either VV and/or HH. Several SAR can provide selected cross-pol data or quad-pol (co-and cross pol together) data. SAR Agency / country Operational Radar band Swath width (max) No. ERS-1 ESA 1991-2000 C 100 km 1 JERS-1 JAXA, Japan 1992-1998 L 75 km 1 ERS-2 ESA 1995-2011 C 100 km 1 RADARSAT-1 CSA, MDA, Canada 1995-2013 C 500 km 1 Envisat ASAR ESA 2002-2012 C 420 km 1 ALOS / PALSAR JAXA, Japan 2006-2011 L 350 km 2 RADARSAT-2 CSA, MDA, Canada 2007- C 500 km 1 COSMO-SkyMed ASI, Italy 2007- X 200 km 4 TerraSAR-X TanDEM-X DLR, Germany 2007- X 100 km 1 HJ-1C CAST, China (2012-) S 100 km 1 Sentinel-1 ESA (2014-) C 400 km 2 RCM CSA (2018-) C 500 km 3 Table 22: Satellite SAR (adapted from Dagestad et al., 2013). The number of satellites is mentioned. All SAR and scatterometer are on-board sun-synchronouspolarorbitingsatellites. The SAR imageisnotprojectedtoageographicalcoordinatesystem,isnotcalibrated,andtheincidence angle dependence must be accounted for. The maximum swath width (see Table ??) is determinedfrom theradarbandandincidentangles.TheincidentanglesforsatelliteSAR vary betweensensors.Thewidestrangeisfrom 8to60degreesforALOS PALSAR. OtherSARsare within this range. The Envisat ASAR Wide Swath Mode (WSM) and Radarsat-1/2 ScanSAR mode provide the maximum swaths of 420 to 500 km, with spatial resolutions of 100 m and 75m, respectively. ThecapacityoftheSAR processingfacilitiesallowsRADARSATdatatobe made available for users in near-real-time; the same was true for Envisat. RADARSAT images can typically be downloaded via internet archives 1−3 hours after the data acquisition. This opportunityhas opened up for operational SAR-based wind mapping. Sentinel-1and Radarsat Constellation Mission (RCM) will be future operational satellites. From most SARs it is possible to order different products with different swath widths. From Envisat and RADARSAT a large suite of types of data are available. For the very high resolution products, the swath is narrow. For the wide swath mode products, the spatial resolution is lower. Thus there is a trade-off selecting either wide swath mode, i.e. more frequent coverage and large regions covered, versus high-resolution mode, i.e. finer spatial details but for smaller regions and less frequently. Before introducing wind retrieval from SAR it is adequate to introduce satellite scatterometry. Scatterometers are also radar instruments observing in microwave bands. The scatterometers ERS-1/2 (ESA) and ASCAT-1/2 on-board the METOP-A and METOP-B satellites (EUMETSAT) are C-band while the scatterometer SeaWinds on-board the QuikSCAT satellite (NASA) and the scatterometer OSCAT on-board the OCEANSAT-2 satellite (ISRO) are Ku band 1.7−2.5 cm (12−18 GHz). Scatterometers are purpose-built for operational ocean surface wind vector observations. Scatterometers have multiple antennae or disk antenna and observe each cell from different viewing angles. All scatterometers operate in VV. For wind retrieval, VV is preferred as the VV signals are stronger than HH. The spatial resolution of DTU Wind Energy-E-Report-0029(EN) 277
Page 1 and 2:
Remote Sensing for Wind Energy DTU
Page 3 and 4:
Author: Alfredo Peña, Charlotte B.
Page 5 and 6:
4 Introduction to continuous-wave D
Page 7 and 8:
8 Nacelle-based lidar systems 157 8
Page 9 and 10:
12 Complex terrain and lidars 231 1
Page 11 and 12:
1 Remote sensing of wind Torben Mik
Page 13 and 14:
Figure 2: Calibration, laboratory w
Page 15 and 16:
Figure 3: Example of scatter plots
Page 17 and 18:
1.2.3 Summary of sodars Most of tod
Page 19 and 20:
1.3.3 Wind lidars Measuring wind wi
Page 21 and 22:
Figure 6: CW wind lidars (ZephIRs)
Page 23 and 24:
Further developments Furthermore, n
Page 25 and 26:
2 The atmospheric boundary layer S
Page 27 and 28:
Figure 9: Large spatial scale varia
Page 29 and 30:
Du3 Dt Du1 Dt Du2 Dt The three mome
Page 31 and 32:
Figure 13: Consensus relations betw
Page 33 and 34:
ψ z L ∼ − 5 L . For unstable c
Page 35 and 36:
Figure 15: Behavior of the turbulen
Page 37 and 38:
Figure 17: Newly developed models t
Page 39 and 40:
The value of q0 at the surface is d
Page 41 and 42:
the spray is the source of icing on
Page 43 and 44:
u∗2 u∗1 u1(h) = u∗1 k ln h
Page 45 and 46:
Figure 26:Land-seabreeze system,whe
Page 47 and 48:
Figure28:Three dimensionalpicture o
Page 49 and 50:
sufficient information. Finally, we
Page 51 and 52:
Mann, J. (1998) Wind field simulati
Page 53 and 54:
(2010) and comparison under differe
Page 55 and 56:
To get the velocity field from the
Page 57 and 58:
τ(k) [Arbitrary units] 10 3 10 2 1
Page 59 and 60:
(Koopmans, 1974; Bendat and Piersol
Page 61 and 62:
anemometer was installed at each en
Page 63 and 64:
and fSw(f) u 2 ∗ = 1.05n 1+5.3n 5
Page 65 and 66:
and n = 0.468. This spectrum implie
Page 67 and 68:
to be calculated. We do that on a m
Page 69 and 70:
Notation A Charnock constant neutra
Page 71 and 72:
Maxey M. R. (1982) Distortion of tu
Page 73 and 74:
4.2 Basic principles of lidar opera
Page 75 and 76:
4.2.5 Wind profiling in conical sca
Page 77 and 78:
4.3.1 Behaviour of scattering parti
Page 79 and 80:
the beam radius at the output lens.
Page 81 and 82:
from which the value of VLOS is der
Page 83 and 84:
the atmosphere. The SNR 4 for a win
Page 85 and 86:
individual line-of-sight wind speed
Page 87 and 88:
A general approach to mitigating th
Page 89 and 90:
from ±VH sinδ (if the tilt is tow
Page 91 and 92:
as a down draught (of the same abso
Page 93 and 94:
Table 7: Combined results from 28 Z
Page 95 and 96:
in Eastern Jutland between January
Page 97 and 98:
Figure 56: Normalized power curves
Page 99 and 100:
the concept. Developments include i
Page 101 and 102:
References x horizontal position in
Page 103 and 104:
Wagner R., Mikkelsen T., and Courtn
Page 105 and 106:
5.2 End-to-end description of pulse
Page 107 and 108:
Scanner Coherent lidar measure the
Page 109 and 110:
Figure 62: Radial wind velocity ret
Page 111 and 112:
transform in order to use data obta
Page 113 and 114:
This wavelength is also the most fa
Page 115 and 116:
Eq.(132)isadaptedforcollimatedsyste
Page 117 and 118:
5.3.6 Existing systems and actual p
Page 119 and 120:
(Gottshall et al., 2010; Albers et
Page 121 and 122:
References Albers A., Janssen A. W.
Page 123 and 124:
derived from fluctuations of the wi
Page 125 and 126:
Acoustic received echo (ARE) method
Page 127 and 128:
Figure 71: Sample time-height cross
Page 129 and 130:
A gradient minimum is characterized
Page 131 and 132:
Figure73:Bragg-relatedacoustic(belo
Page 133 and 134:
stability (inversion strength) can
Page 135 and 136:
Figure 76: Combined soundingwith a
Page 137 and 138:
Figure 79: Favorite regions (shaded
Page 139 and 140:
Direct detection of MLH from acoust
Page 141 and 142:
Engelbart D.A.M.and Bange J. (2002)
Page 143 and 144:
7 What can remote sensing contribut
Page 145 and 146:
uyms 9.0 8.5 8.0 7.5 7.0 120 140 16
Page 147 and 148:
Bottom of rotor Φ rotation r w u
Page 149 and 150:
Height Height Hub 1.6 1.4 1.2 1.0 0
Page 151 and 152:
PP rated PP rated 1.0 0.8 0.6 0.4 0
Page 153 and 154:
KEprofileKEhub 1.2 1.1 1.0 0.9 0.8
Page 155 and 156:
PP rated WS Lidarms 1.0 0.8 0.6 0.4
Page 157 and 158:
8 Nacelle-based lidar systems Andre
Page 159 and 160:
• Flexibletrajectories.Dependingo
Page 161 and 162:
Figure 104: Sketch of simultaneous
Page 163 and 164:
The normal wind direction vector nw
Page 165 and 166:
Figure 109: Test site at DTU Wind E
Page 167 and 168:
Figure 112: Power curve met mast an
Page 169 and 170:
Notation C number of sent photons C
Page 171 and 172:
9 Lidars and wind turbine control -
Page 173 and 174:
for three unknowns, it is impossibl
Page 175 and 176:
model of the blade pitch actuator,
Page 177 and 178:
|GRL| [-] 1 0.8 0.6 0.4 0.2 10 k [r
Page 179 and 180:
PSD(Ωg) [(rpm) 2 /Hz] PSD(Ωg) [
Page 181 and 182:
0.04 0.03 ˆk [ rad m ] 0.02 0.63 0
Page 183 and 184:
[%] 10 0 −10 −20 −30 MyT Moop
Page 185 and 186:
PSD(θ1) [rad 2 /Hz] PSD(Moop1) [Nm
Page 187 and 188:
Pel/Pel,max [-] 1 0.98 0.96 0.94 0.
Page 189 and 190:
fL weighting function GRL transfer
Page 191 and 192:
E. Hau, Windkraftanlagen, 4th ed. S
Page 193 and 194:
¨¦¦§©¡§ ¥§¨¦¦§£ ¡¥
Page 195 and 196:
vertical (m) 150 100 50 R d 0
Page 197 and 198:
With feedback only, on the other ha
Page 199 and 200:
Figure 136: Estimated preview requi
Page 201 and 202:
Normalized C r = C r P Q 2 W (r) b
Page 203 and 204:
Normalized C r = C r P Q 2 W (r) b
Page 205 and 206:
Coherence 1 0.8 0.6 0.4 0.2 0 10
Page 207 and 208:
Coherence 1 0.8 0.6 0.4 0.2 0 10
Page 209 and 210:
Magnitude Squared 10 8 10 7 10 6 10
Page 211 and 212:
Figure 148: During simulation, FAST
Page 213 and 214:
Figure 149: Collective flap respons
Page 215 and 216:
Magnitude (abs) blade pitch gen spe
Page 217 and 218:
• Measurement coherence, which ca
Page 219 and 220:
Jonkman, B. (2009) TurbSim user’s
Page 221 and 222:
11 Lidars and wind profiles Alfredo
Page 223 and 224:
z [m] 160 100 80 60 40 20 10 15 20
Page 225 and 226: z [−] zo 1 κ ln 40 38 36 34 32 3
Page 227 and 228: z [m] z [m] 1000 900 800 700 600 50
Page 229 and 230: the growth of the length scale, agr
Page 231 and 232: 12 Complex terrain and lidars Ferha
Page 233 and 234: Figure 158: The ZephIR models which
Page 235 and 236: Uconst wΑx l h Φ h.tanΦ Figure 1
Page 237 and 238: Figure 162: Lavrio: The scatter plo
Page 239 and 240: Figure 163: Panahaiko: The scatter
Page 241 and 242: References Albers A. and Janssen A.
Page 243 and 244: Wexler (1968), where the limitation
Page 245 and 246: 13.2.1 Systematic turbulence errors
Page 247 and 248: where x is the center of the scanni
Page 249 and 250: The theoretical systematic errors a
Page 251 and 252: Height (m) 160 140 120 100 80 60 40
Page 253 and 254: Height (m) Height (m) 160 140 120 1
Page 255 and 256: RMSPE (%) 70 60 50 40 30 20 10 0 vu
Page 257 and 258: involves interaction of all compone
Page 259 and 260: Lindelöw P. (2007) Fibre Based Coh
Page 261 and 262: plications, including meteorologica
Page 263 and 264: Figure 173: Rayleigh-Jeans’ appro
Page 265 and 266: Figure 176: Brightness temperature
Page 267 and 268: with rain are shown in Fig. 179. A
Page 269 and 270: This method gives a good accuracy (
Page 271 and 272: Figure 183: A recent maintenance in
Page 273 and 274: Figure 185: Temperature profiles in
Page 275: References s point in space Sν(s)
Page 279 and 280: Christiansen et al., 2006). The res
Page 281 and 282: educe speckle noise, a random noise
Page 283 and 284: Figure 188: Envisat ASAR wind field
Page 285 and 286: Figure 190: Envisat ASAR wind field
Page 287 and 288: Satellites in sun-synchronous polar
Page 289 and 290: 500 overlapping scenes for wind res
Page 291 and 292: (with the fewest samples). The unce
Page 293 and 294: Badger M., Hasager C. B., Thompson
Page 295 and 296: Ren Y. Z., Lehner S., Brusch S., Li
Page 297 and 298: tracking, climate studies, air-sea
Page 299 and 300: The modified logarithmic wind profi
Page 301 and 302: sea ice mask was applied and σ0 me
Page 303 and 304: QuikSCAT 25 20 15 10 5 Horns Rev Fi
Page 305 and 306: Figure 203:Example of an ASCAT coas
Page 307 and 308: References Bourassa M.A., Legler D.
Page 309: DTU Wind Energy Technical Universit
show all

Publishers version - DTU Orbit

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?