Effect of Wind Shear and Turbulence on Wind Turbine Power Production, Based on LIDAR Measurements

Abstract

Denne bacheloroppgaven undersøker kraftproduksjonen til en multi-megawatt vindturbin som et utstyrt med en LIDAR. LIDAR målinger har blitt brukt til å estimere skjærvind og turbulens intensitet, for å komplementere turbin data. All data har blitt filtrert etter flere vilkår for å øke resultatenes pålitelighet.

Simuleringer for skjærvind og turbulens intensitet har blitt gjennomført i Ashes. Skjærvind hadde liten effekt på kraftproduksjon for økende vindhastigheter. Kraftkurver for profiler med høyt skjær falt under de simulert for lavere skjær, og reduserte AEP. Turbulens simuleringer resulterte i større avvik mellom kraftkurver. Høy turbulens intensitet presterte bedre for lave vindhastigheter, men falt signifikant med økende vindhastigheter. Kraftkurver estimert for høy turbulens intensitet reduserte AEP.

Turbin data ble sortert i fire serier med varierende skjæreksponenter: α < 0, 0 ≤ α < 0.1, 0.1 ≤ α < 0.25, og α ≥ 0.25. Dette resulterte i kraftkurver som overensstemte godt med simuleringer. Lave, positive skjæreksponenter hadde høyest AEP sammenlignet med α < 0, 0.1 ≤ α < 0.25, og α ≥ 0.25 som falt med 0.8%, 0.5%, and 0.8%, respektivt. Data sortert etter turbulens intensitet varierte signifikant. Kraftkurver for fire turbulens intensiteter ble estimert: 0 ≤ TI < 0.06, 0.06 ≤ TI < 0.12, 0.12 ≤ TI < 0.18, og TI ≥ 0.18. Disse kraftkurvene stemte relativt godt med simuleringer, og fulgte samme trend. Kraftkurver for høy turbulens intensitet hadde størts kraftproduksjon ved lavere vindhastigheter. For vindhatigheter over 9 m/s, derimot, sank kraftkurvene betraktelig og var bedre for profiler med lav turbulens intensitet. Sammenlignet med referansen 0 ≤ TI < 0.06, så sank AEP med 0.6%, 1.1%, og 1.4% for 0.06 ≤ TI < 0.12, 0.12 ≤ TI < 0.18, og TI ≥ 0.18, respektivt.

During the work of this bachelor's thesis, the power performance of a multi-megawatt wind turbine equipped with a continuous wave LIDAR was analyzed. LIDAR measurements were used to calculate wind shear and turbulence intensity to complement the turbine data. Numerous conditions filtered the entire data set to increase the reliability of results. Large intervals of data were dropped due to unreliable LIDAR measurements.

Simulations were run for various shear exponents and turbulence intensities. Wind shear simulations yielded minor changes in power output for increasing wind velocities. Power curves created for high wind shears dropped below the ones estimated for lower wind shears, reducing AEP. Turbulence simulations resulted in power curves with significant deviations. High turbulence intensities over-performed for low wind velocities but dropped substantially with increasing wind velocity. Power curves calculated for high turbulence intensities lead to a reduction in AEP.

Turbine data were binned into four ranges of shear exponents: α < 0, 0 α < 0.1, 0.1 α < 0.25, and α ≥ 0.25. This resulted in power curves that corresponded well with simulations. Low and positive wind shear yielded highest AEP, with α < 0, 0.1 ≤ α < 0.25, and α ≥ 0.25 being reduced with 0.8%, 0.5%, and 0.8%, respectively. Data binned by turbulence intensity deviated significantly. Power curves for four turbulence intensities were calculated: 0 ≤ TI < 0.06, 0.06 ≤ TI < 0.12, 0.12 ≤ TI < 0.18, and TI ≥ 0.18. These power curves corresponded relatively well with simulations, following the same trend. Power curves for higher turbulence intensities lead the ones for lower turbulence intensities at low wind velocities. Power output for velocities above 9 m/s were significantly better for lower turbulence intensities. Compared to the reference range 0 ≤ TI < 0.06, AEP was reduced with 0.6%, 1.1%, and 1.4% for 0.06 ≤ TI < 0.12, 0.12 ≤ TI < 0.18, and TI ≥ 0.18, respectively.