Ganesh Vijayakumar

Space-Time Loadings on Wind Turbine Blades Driven by Atmospheric Boundary Layer Turbulence

the interactions between the spatio-temporal loadings on wind turbine blade blades and the turbulence structure of the neutral and moderately convective atmospheric surface layer by combining the Blade Element Method incorporated in the FAST/AeroDyn codes from NREL with a dynamic stall model with large-eddy simulation (LES) of the atmospheric boundary layer (ABL). The inow conditions were obtained from high-resolution LES interpolated to the turbine blade. The central aim of our analysis is to search for and quantify direct causal relationships between specic space-time variabilities in the turbulent inow velocity eld and the spatio-temporal variability of forces on the turbine blades, and the integrations along the blade span that produce time variations in bending moment at the hub and shaft torque. A related interest is the impact of an accurate versus inaccurate predictions of shear rate by the LES. We nd that atmospheric turbulence is a major contributor to blade loadings and that the distribution of force uctuations is sensitive to the specic structure of ABL turbulence. A well designed, accurate LES model has signicant advantages for quantifying the role of atmospheric turbulence on wind turbine performance.

Comparing Unsteady Loadings on Wind Turbines using TurbSim and LES Flow Fields

Unsteady loading on wind turbine blades due to atmospheric turbulence may be a cause for higher than anticipated wind turbine downtime. The National Renewable Energy Laboratory (NREL) has produced a turbulence simulator, TurbSim, for use in wind turbine development and analysis. We compare the turbulence created by TurbSim with atmospheric turbulence created with low-dissipation Large-Eddy Simulation of the canonical moderately convective atmospheric boundary layer. This turbulence is used to create inow conditions to NREL’s FAST code to study the dierences in wind turbine loadings. Through examination of dierent wind turbine parameters, we observe dierences between the kinematic and dynamic turbulence ow elds. In particular, we nd that the predicted mean values for rotor power and lift are very similar between the two turbulence elds. However, the variance inherent in the LES turbulence is not found within the TurbSim kinematic turbulence. Matching the TurbSim mean Reynolds stresses to those of the LES ow eld does not cause the correlations between wind velocity and turbine loadings to match. We conclude that TurbSim is a reasonable tool for some wind turbine analysis applications, but it does not fully capture the variance associated with the canonical moderately convective atmospheric boundary layer.

Considerations in coupling LES of the atmosphere to CFD around wind turbines

rst part of this paper, we analyze the role of algorithm in the inaccurate predictions of the mean shear in the surface layer of high Reynolds number ows. Brasseur and Wei 1 have proposed a solution to this problem in the < ReLES parameter space. We perform the same simulation using two algorithms with dierent numerical dissipation characteristics, viz., a spectral algorithm and a nite volume algorithm. We repeat this procedure for two incompressible high Reynolds number ows: channel ow and a neutral atmospheric boundary layer. The increased dissipation in the nite volume algorithm compared to a spectral algorithm acts as an eective lter cuto at a lower wavenumber than the grid can represent. We nd that the nite volume algorithm experiences a lower resolved stress, increased variance in the streamwise velocity and lower variance in the vertical velocity in the surface layer. These combine together to aect the turbulence structure in the surface layer that then aects the whole boundary layer.

Blade Boundary Layer Response to Atmospheric Boundary Layer Turbulence on a NREL 5MW Wind Turbine Blade with Hybrid URANS-LES

We focus on the spatio-temporal changes in the blade boundary layer structure caused by the interaction of a wind turbine blade with the day-time Atmospheric Boundary Layer (ABL). Previous studies have shown that the time scales of the energy containing eddies in the ABL are of the order of multiple rotation time scales of commerical wind turbines and are directly correlated with the large temporal fluctuations in the integrated loads. We attempt to understand the details of the blade boundary layer dynamics that causes these fluctations by simulating a single blade of the NREL 5MW turbine in a realistic ABL. We use a psuedo-spectral code to perform Large Eddy Simulation of the Atmospheric Boundary Layer and generate realistic inflow conditions for the turbine. We develop and use a new hybrid URANS-LES model by combining the 1 equation model for the LES of ABL with the k − ω − SST − SAS model by Menter and Egorov near the blade. We then perform Hybrid URANS-LES computations of the flow around the single NREL 5MW blade to compute the spatio-temporal fluctations in the surface stresses due to ABL turbulence. We find that the time scales experienced by the NREL 5MW turbine in the ABL is of the order of multiple rotation time scales that could potentially be used to control the wind turbine. Initial results from the simulation of the turbine blade in an Atmospheric Boundary Layer indicate that the boundary layer separation on the blade in an ABL could span the entire blade, unlike uniform inflow where the separation is primarily restricted to the near root region.

Influences of Atmospheric Boundary Layer Turbulence Structure on the Space-time Variability in Wind Turbine Blade and Shaft Loadings

The atmospheric boundary layer (ABL) contains turbulence structure that is tied to its global stability state. The spatio-temporal structure of the wind vectors that pass through a large wind turbine rotor disk force loadings with variations associated with the turbulence structure of the ABL, and therefore related to global stability. We present here an analysis of daytime ABL structure in relationship to wind turbine loading by coupling well resolved spectral large-eddy simulation of canonical neutral boundary layer (NBL) and moderately convective boundary layers (MCBL) to the NREL FAST/Aerodyn design-level tool based on the Blade Element Method using the NREL 5 MW turbine with 126 m rotor disk. The loadings underlying power are significantly enhanced by the presence of atmospheric turbulence (relative to mean velocity alone). The horizontal scale of the ABL energydominant eddies is of order the rotor diameter. As these eddies sweep through the rotor disk, the distribution of fluctuations in rotor loadings change according to the structure of the thermal eddies with upward vs. downward motions, and horizontal eddies with high-speed vs. low-speed motions (relative to the mean). In the MCBL the up/down eddy structures are correlated with the high/low speed motions, and there is are significant differences in the high/low-speed structures between the NBL and MCBL that impact blade loadings. To better understand the structure and stability-based differences we designed a method to condition the velocity field based on strong up/downward and high/low speed motions on filtered planes of velocity data at hub height. Based on this condition, the variations in conditional mean winds between up/down and high/low large eddy structures are ~ 4 m/s and the wind turbine experiences these variations over time scales ~ 1-3 minutes, or 10 35 rotor revolutions, sufficiently long to include in control strategies.