Skip to main navigation Skip to search Skip to main content

Typhoon wind and turbulence structure, and its impact on wind energy application

Research output: Book/ReportPh.D. thesis

269 Downloads (Orbit)

Abstract

More and more wind farms are being planned in areas affected by tropical cyclones. Therefore, a thorough understanding of the tropical cyclone wind field and its impact on wind turbine loads is required. With the aim of reducing the uncertainties associated with the inflow conditions in wind turbine load simulations, the tropical cyclone wind field has been analyzed through a combination of mesoscale modeling, spectral modeling, and measurements.
Mesoscale simulations with the Weather Research and Forecasting model were used to analyze wind profiles and mesoscale wind speed variability during tropical cyclone conditions. It was found that over the open ocean, the wind shear in the mesoscale simulations was moderate and in good qualitative agreement with drop-sonde measurements. However, in coastal regions, the modeled wind shear can be significantly greater than that suggested by current wind turbine design standards. The variability of the simulated wind field was found to depend on the boundary layer parameterization schemes used in the model. Regions of pronounced variability were found to be associated with boundary layer rolls and precipitation along the tropical cyclone rainbands. However, these organized meteorological features and wind field variability at even smaller scales cannot be fully resolved by mesoscale models.
Tropical cyclone turbulence is analyzed using high-frequency sonic anemometer measurements from four typhoon cases. These measurements were used for the first in-depth analysis of how well the IEC-recommended Mann uniform shear model captures the spectral characteristics of tropical cyclone turbulence. The analysis showed that the Mann model can provide a reasonable approximation of the turbulence spectra in certain regions within tropical cyclones. However, 1.) the frequency of the spectral peaks of the u and v wind components was found to be closer together during certain measurement periods than predicted by the Mann model, and 2.) for the same periods, more energy was observed in the v component within the energy-containing subrange than predicted by the Mann model. By combining the measurements with mesoscale simulations, it was proposed that these two features may be related to the presence of boundary layer rolls. Furthermore, the spectral energy was found to be mostly larger than predicted by the Mann model at larger scales of several hundred to thousands of meters, especially in the outer region of tropical cyclones.
In order to add the missing large scale variability to the variability predicted by the Mann model, a new method to generate wind fields has been developed. This method combines the mesoscale wind variability obtained from mesoscale simulations with the microscale turbulence obtained from the Mann model. The method was used to generate wind fields with cyclone-specific wind speed, wind shear, wind veer, and meso- to microscale variability. These wind fields were used as inflow conditions for wind turbine load simulations. Based on a limited number of wind turbine load simulations, it is suggested that the mesoscale variability added with the current method does not systematically affect fatigue loads, but does lead to an increase in extreme loads. However, further validation of the generated wind fields and load simulations for more tropical cyclone cases are needed to confirm this result. This approach is also expected to be applicable to the analysis of wind turbine loads and wake meandering under non-tropical cyclone conditions.
Original languageEnglish
Place of PublicationRisø, Rokilde, Denmark
PublisherDTU Wind and Energy Systems
Number of pages146
Publication statusPublished - 2025

Fingerprint

Dive into the research topics of 'Typhoon wind and turbulence structure, and its impact on wind energy application'. Together they form a unique fingerprint.

Cite this