Wind Shear and Uncertainties in Power Curve Measurement and Wind Resources

Abstract

Power curve measurements are encumbered with large uncertainty as wind measurements take place only at hub height. The wind profile over the turbine rotor is an expression of the kinetic energy available to the wind turbine and the evolution of large rotors prohibits the assumption that the hub height wind speed is representative of the wind speed over the whole rotor. Even in the case where measurements cover the lower half of the turbine rotor and extrapolations are attempted, the uncertainties remain considerable. We argue for that the measurement of the wind speed over the whole rotor height should be the future preferred approach. Such a measurement will minimize the uncertainty in estimating the wind potential of a site and the uncertainty in the power curve measurement method and the AEP calculation of wind turbines. To document this, we present wind speed and power curve results from wind and power measurement campaigns, one in flat terrain suffering an energy deficit and one in complex terrain presenting a surplus. Common for both is the inadequacy of the hub height wind speed measurement to describe the energy contents of the flow. In the flat terrain campaign we use one-year period of data in order to study the wind shear profiles at heights which correspond to an assumed rotor area of a modern multi-MW turbine. The energy flux through the “turbine rotor” is seen to be subject to seasonal variations caused by differences in atmospheric stability which influence the shear profile shape. Considerable deviations occur relative to the flux measured when only the cup anemometer at hub height is used. In the complex terrain campaign, a wind turbine power curve has been measured for a period of eight months in a Midwest (US) site following a site calibration. Wind shear measurements over the lower rotor part were taken throughout this period at three heights (hub, lower tip and midway between the two). Considerable wind shear during nights and well-mixed profiles during days were observed. Large differences in the power curve and the AEP between day and night periods were observed, the power curve and the AEP being better during the night. The data analysis was combined with in-house aeroelastic simulations, over a wide range of wind shear and turbulence intensities values, in order to verify the analysis findings. The combined simulations and data analysis results made it clear that the upper turbine rotor part was influenced by the presence of a low level jet during nighttime. This caused considerable deviations from the expected power curve and AEP, which were not detected either by the site calibration or the lower rotor part rotor speed measurements. We conclude the paper by presenting results from combined cup and LIDAR power curve measurements, and suggest a method which compensates for the wind shear influence on the power curve.