Measurements of Turbulence Profiles with Scanning Doppler Lidar for Wind Energy Applications

Click here to load reader

  • date post

  • Category


  • view

  • download


Embed Size (px)


Rod Frehlich University of Colorado, Boulder, CO RAL/NCAR, Boulder, CO. Measurements of Turbulence Profiles with Scanning Doppler Lidar for Wind Energy Applications. Acknowledgments. CTI (Steve Hannon, Sammy Henderson, Jerry Pelk, Mark Vercauteren, Keith Barr) - PowerPoint PPT Presentation

Transcript of Measurements of Turbulence Profiles with Scanning Doppler Lidar for Wind Energy Applications

  • Measurements of Turbulence Profiles with Scanning Doppler Lidar for Wind Energy Applications

    Rod Frehlich University of Colorado, Boulder, CORAL/NCAR, Boulder, CO

  • AcknowledgmentsCTI (Steve Hannon, Sammy Henderson, Jerry Pelk, Mark Vercauteren, Keith Barr)Larry Cornman, Kent Goodrich (Juneau)DARPA Pentagon Shield (Paul Benda)R. Sharman, T. Warner, J. Weil, S. Swerdlin (NCAR)Y. Meillier, M. Jensen, B. Balsley (CU)Army Research Office (Walter Bach)NSF (Steve Nelson)NREL (Neil Kelley, Bonnie Jonkman)CU renewable energy seed grantNOAA (Bob Banta, Yelena Pichugina)

  • Complements of Steve Hannon

  • Point Statistics of TurbulenceLongitudinal velocity (radial)Transverse velocityLongitudinal structure function

    Transverse structure function



  • Horizontal Isotropic TurbulenceUniversal description - longitudinal

    u longitudinal standard deviationL0u longitudinal outer scale(x) universal function (von Krmn)s

  • Horizontal Isotropic TurbulenceUniversal description - transverse

    v transverse standard deviationL0v transverse outer scale(x) universal function (von Karman)s

  • Doppler LIDAR Range WeightingTime of data maps to range (1 s = 150 m)Pulsed lidar velocity measurements filter the random radial velocity vr(z)Pulse width rRange gate length defined by processing interval pRange weighting W(r)

  • LIDAR Data and Range WeightingLIDAR radial velocity estimates at range R

    vwgt(R) - pulse weighted velocitye(R) estimation error

    vr(z) - random radial velocityW(r) - range weighting

    v(R) = vwgt (R) + e(R)vwgt(R)=vr(z)W(R-z) dzp=72 mr=66 m

  • Estimates of Random ErrorRadial velocity error variance e2 can be determined from dataSpectral noise floor is proportional to e2e2=0.25 m2/s2

  • Non-Scanning Lidar DataVertically pointed beamTime series of velocity for various altitudes zHigh spatial and temporal resolutionResolves turbulence

  • Lidar Structure Function (Radial)Corrected longitudinal structure function

    Draw(s) raw structure function e2(s) correction for estimation errorTheoretical relation (h

  • Turbulence StatisticsCalculate corrected structure functionDetermine best-fit to theoretical modelBest-fit parameters are estimates of turbulence statistics

  • Scanning Lidar DataHigh-resolution profiles of wind speed and turbulence statisticsHighest statistical accuracyRapid update rates compared with tower derived statisticsCan provide profiles at multiple locationsIdeal for some wind energy applications

  • LIDAR Velocity Map for 1oRadial velocity100 range-gates along beam 180 beams (=0.5 degree)18 seconds per scanh=R
  • Wind Speed and DirectionBest-fit lidar radial velocity for best-fit wind speed and directionFluctuations are turbulenceLongitudinal fluctuations along the lidar beamTransverse fluctuation in azimuth direction

    in situ prediction

  • Lidar Structure Function (Azimuth)Corrected azimuth structure function

    Draw(s) raw structure functione2(s) correction for estimation errorTheoretical relation (h

  • Turbulence Estimates Zero ElevationStructure function of radial velocity in range a)Best fit produces estimates of u, u, L0uStructure function in azimuth b)Best fit produces estimates of v, v, L0v

    in situ predictiontropprof04

  • Turbulence Estimates H=80 mBest fit for noise corrected structure function (o)Raw structure functions (+)Radial velocity a) has small elevation angles (
  • CIRES Tethered Lifting System (TLS):Hi-Tech Kites or Aerodynamic Blimps

  • High Resolution Profiles from TLS DataTLS instrumentation used as truth for turbulence profilesHot-wire sensor for small scale velocityCold-wire sensor for small scale temperature

  • Velocity TurbulenceAlong-stream velocity u(t)Spectrum Su(f)Taylors frozen hypothesisEnergy dissipation rate

  • Temperature TurbulenceAlong-stream temperature T(t)Spectrum ST(f)Taylors frozen hypothesisTemperature structure constant CT2

  • LIDAR, SODAR and TLS (Blimp)

  • Lidar and TLS Profiles

  • Wind Energy ApplicationsTowers are expensive and limited in height coverageCW Doppler lidar can measure winds near an individual turbineScanning Doppler lidar can monitor large area upstream of a wind farmAutonomous lidar (CTI)AirportsHomeland security DC

  • Doppler Lidar at NRELRadial velocity mapLarge eddiesLarge velocity variationsMultiple elevation angles provide 3D sampling

  • High Turbulence Large fluctuations about the best fit wind speed 3D average required for accurate statistics

  • Azimuth Structure FunctionsBest-fit model provides robust turbulence statisticsProfiles produced from 3D volume scan by processing data in altitude bins

  • Atmospheric ProfilesAccurate profiles producedMost complete description of wind and turbulence availableIdeal for site resource assessment

  • Large Turbulent Length ScaleDifficult to separate turbulence and larger scale processesSimilar to troposphere and Boreas DataViolates Cartesian approximation of analysisWhat is optimal methodology?

  • Small Turbulent Length ScaleLarge corrections for spatial filteringRequires shorter lidar pulse and more accurate correctionsCritical for lower altitudes

  • Profiles from Two Angular SectorsDifferences in wind speed and directionTurbulence profiles similarImplications for site resource assessment

  • Rapid Evolution in 7 MinutesRapid change in wind speedTurbulence levels are not reduced with lower winds3D scanning required for short time averages

  • Directional ShearLarge shear in wind directionTypically at night with light windsMore data required for wind energy applications

  • Future WorkOptimize Lidar design, signal processing, and scanning patternsDetermine universal description of turbulence for better turbulence estimates (anisotropy?)Extend spatial filter correction to two dimensions for faster scanning and larger maximum rangeImprove algorithms for large length scale L0 (relax Cartesian approximation)More data required, especially for wind energy research

  • Lafayette Campaign

  • LIDAR and TLS (Blimp)

  • Low Turbulence DataAccurate corrections for pulse filtering requiredCorrect turbulence model for spatial statistics = 5.0 10-6 m2/s3v = 0.090 m/sL0 =140 m

  • Low Turbulence Conditionsin situ prediction

  • Convectionin situ prediction

  • Coherent Doppler Lidar Properties

    Direct measurement of Doppler shift from aerosol particlesDoppler shift 1 MHz for 1 m/s (2 m)Accurate radial velocity estimates with little bias Most sensitive detection methodImmune to background light Eye safe operation

  • Atmospheric LIDAR DataLidar signal is a narrow band Gaussian random process Simple statistical description for constant velocity and aerosol backscatter

  • Atmospheric LIDAR DataLidar signalRandom velocityPulse weightingRange gate defined by processing interval

  • Estimation of VelocityData from multiple pulses N improves performanceSpectral based estimators Maximum Likelihood has best performance

  • Multiple Pulse EstimatesSingle pulse data has random outliersPulse accumulation removes outliersTemporal resolution reduced

  • Estimates of Random Error-cont.Multiple pulse data has a smaller region for the noise floorThe atmospheric signal must have low frequency content

  • Hard Target DataVelocity bias determined from hard target dataBias is typically less than 2 cm/s

  • Verification of AccuracyVelocity random error depends on signal energyAccuracy is very goodAgrees with theoretical predictions if turbulence is included