Session T27: Aerodynamics: Wind Energy

Large eddy simulation of canonical urban geometries

Forecasts of wind energy potential at the hub height of wind turbines are important for optimizing the siting and layout of wind turbines in a distributed wind scenario where the generated electrical power is fed directly to buildings/farms in an urban/semi-urban setting. However, the effect that an array of buildings would have on the turbulence characteristics near the terrain surface, and at hub height, continues to challenge attempts to model the flow. Therefore, information about the mean flow velocity and turbulence structures of canonical urban layouts are important for the efficient design and layout of wind turbines. In our work, high-fidelity, high-order spectral element (Nek5000) simulations of flow in the atmospheric surface layer, at the lower hub heights, were performed to understand the influence of urban geometry on the turbulence characteristics. A comparison of our results, with the limited experimental data available from open literature, shows good agreement with our LES predictions, and sets the stage for the development of low-order (macro-scale) models.  We present the results of our simulations and also discuss their importance in parametrizing the effects of the urban canopy in meso-scale and macro scale models.   

Double-Gaussian model for predictions of the streamwise mean velocity and turbulence intensity in wind-turbine wakes

Wind turbine wakes pose unique challenges for predictions of mean velocity and turbulence intensity due to the high Reynolds numbers and the complex flow physics involved. In this work, the streamwise mean velocity and turbulence intensity obtained from LiDAR measurements of wakes generated by full-scale wind turbines are interrogated to assess the accuracy of existing wake models, such as the Gaussian wake model, and identify opportunities for improvements of wind turbine wake modeling. It is found that a double-Gaussian model provides a higher degree of self-similarity throughout the wake region, specifically in the near-wake, enabling more accurate predictions of the mean velocity than those obtainable with the Gaussian model. The enhanced prediction capabilities of the double-Gaussian model are also confirmed from the analysis of the wake-added turbulence intensity. The predictions of the diagonal streamwise component of the strain-rate tensor obtained with the double-Gaussian model show better agreement with the experimental data, while the Gaussian model generally overestimates the rotor-averaged wake-added turbulence intensity and locates the turbulence-intensity peaks more inwards with respect to their actual positions.   

Large eddy simulation of offshore wind turbines interacting with waves using the spectral element method

Offshore wind is attracting increasing attention as a renewable energy source to meet the decarbonization goal. Winds offshore typically feature higher speeds than on-land and thus possess a higher power generating potential. However, offshore wind turbine aerodynamics is usually complicated by ocean waves and the coupling of ocean waves and surface winds. Computational fluid dynamics (CFD) simulations provide a unique tool to directly probe complex flow structure and dynamics in offshore wind turbine systems, thus improving the physical understanding and designs of turbines with higher stability and efficiency. In this study, we perform large eddy simulations (LES) using a spectral element CFD code Nek5000 for an offshore wind turbine interacting with prescribed ocean waves. The use of the high-order spectral element method ensures capturing the highly transient vortex structures. An actuate line model (ALM) is used to reproduce the aerodynamic forces from the turbine blades without the need to actually resolve the boundary layers around the blades. A log-law based wall model is used for the atmospheric boundary layers (ABL) over the ocean surface. The spectral element LES-ALM modeling framework is first validated on the NREL-5MW wind turbine. The spectral element simulation shows good agreement with the reference data even with a coarser grid size. The validated code is then applied to simulations with flat terrain and moving wave, respectively. Effects of both ABL and moving wave are investigated in detail, focusing on aerodynamic properties along the actuate line, turbine efficiency, and wake dynamics. The results from this study will shed lights on wind/wave/turbine coupling in offshore wind turbine systems. The high-fidelity data generated can further be used to inform low-order models for design optimization.   

Why the Long Phase? Disambiguating the Dynamics of a Periodically Surging Wind Turbine

Despite its relevance to floating offshore wind turbines and other non-traditional wind-energy systems, the unsteady power generation of a wind turbine translating in the streamwise direction has not been comprehensively studied in experiments. Accordingly, a horizontal-axis wind turbine was actuated in periodic surge motions in a fan-array wind tunnel at the Caltech Center for Autonomous Systems and Technologies (CAST). Experiments were conducted at a diameter-based Reynolds number of Re_D = 5.6 × 10⁵ and at tip-speed ratios between 5.7 and 9.4. Sinusoidal and trapezoidal streamwise-velocity waveforms with maximum velocities up to U/U_∞ = 0.32 were considered. Phase-averaged torque and rotation-rate measurements of the turbine in these unsteady motions showed increases in the time-averaged power in excess of 6% relative to the steady case. A first-order model was derived to capture trends in the magnitude and phase of the measured data, which allowed the aerodynamics of the turbine to be separated from other factors, such as generator characteristics and turbine inertia. These results inform the development of strategies to optimize and control the unsteady power generation of surging wind turbines.   

A Comparison of Wind Turbine and Porous Disk Wakes at High Reynolds Numbers

The present study examines the suitability of the actuator disk wind turbine model at field-relevant Reynolds numbers through an experimental comparison between the wakes of a three-bladed horizontal axis wind turbine and a matched porous disk. The experiments are conducted in the High Reynolds Test Facility at Princeton University, where high Reynolds numbers are achieved at flow velocities less than 10 m/s using air pressurized up to 238 bar as the working fluid. The rotor (18% solidity) and disk are matched in terms of diameter (20 cm) and radial solidity distribution, with the solidity of the disk varied between 30% and 60% to match rotor thrust coefficients at diameter-based Reynolds numbers of 3×10⁶ to 7×10⁶ and tip speed ratios of 3.5 to 6. Wake measurements are performed using nanoscale thermal anemometry, capturing the mean and time-varying structure of the both the near and far wakes. Particular attention is paid to the near wake, where differences are noted on account of wake dynamics driven by tip vortex and shear layer development in the rotor and disk cases, respectively. Results are compared with previous findings at much lower Reynolds numbers, leading to conclusions on the suitability of the actuator disk model at high Reynolds numbers.   

Spectral inspection of the wake and power fluctuations of wind turbines under low-level jets: an experimental study

Over time, wind turbines have grown monotonically. Current units are exposed to atmospheric phenomena that pose techno-economical challenges. One of them is the so-called Low-level jets (LLJ), characterized by a peak in the velocity profiles with typical elevation ranging from 50 m to over 200 m. These elevations are within the region covered by full-scale turbine rotors. Although positive effects have been discussed in the flow recovery in the wake under certain conditions (Doosttalab et al., JRSE, 12:053301, 2020), it is central to understand the factors contributing to the LLJ-turbine interaction. Here, we study the spectral structure of the wind turbine wake and its impact on the power fluctuations of a simple two-turbine system. Using a customized velocity profiler in a wind-tunnel environment, we replicated features of LLJs and characterized their effects using hot-wire anemometry. Our results indicate that the relative position at which the peak of the low-level jet impinges on the turbine rotor has a substantial impact on the wake and power output structures.   

Tip speed dependent vortical structures in the wake of a model wind turbine at high Reynolds numbers

The wake a of a model horizontal-axis wind turbine was studied at a Reynolds number of 4 million across a range of tip speed ratios of 4<λ<7. To achieve Reynolds numbers higher than most existing turbine wake studies, the High Reynolds number Test Facility (HRTF) was used, which uses pressurized air, up to 230 bar, as the working fluid. Hot-wire anemometry was used to acquire velocity measurements, but through the use of the nano-scale thermal anemometry probe (NSTAP), spatial and temporal filtering was avoided. Downstream distances of 0.77 

Investigation on the recovery mechanisms of wind turbine wakes through statistical analysis of LiDAR measurements and RANS simulations

The recovery of wind turbine wakes is the result of the interaction between the incoming atmospheric turbulence, the wake-generated turbulence, and the vorticity structures shed by the blades. A deeper understanding of such intertwined mechanisms is needed to improve the accuracy of wake models. In this study, LiDAR measurements of wind turbine wakes collected under a vast breadth of inflow conditions are analyzed through the LiDAR Statistical Barnes Objective Analysis (LiSBOA) tool to retrieve single-point statistics of the streamwise velocity. The results show that the blade rotation has a significant effect on the statistical properties of turbulence. The mean velocity field is used as a reference for tuning the turbulent eddy viscosity of an in-house-developed Pseudo-2D RANS model. The turbulent eddy viscosity increases in convective conditions compared to stable conditions. A secondary increasing trend of the turbulent eddy viscosity as a function of the turbine thrust coefficient is also captured for moderately turbulent inflows. Interestingly, this trend is reversed for incoming wind conditions with higher turbulence intensity. We speculate that this unexpected flow feature is linked to the action of the turbine blades on the incoming turbulent coherent structures.   

On the wake dynamics, flow loading and power output of wind farms under wake steering control: a wind tunnel experiment

Wake steering control strategies have proven to be effective for increasing the power output of a wind farm. The variation of yaw angles of individual turbines within a wind farm produces distinctive impacts on the wake flow characteristics, and therefore alter the mean/fluctuating wind loads and power outputs of downstream turbines. In this work, systematic wind tunnel experiments were performed to evaluate the performance of modeled wind farms with 3 rows and 3 columns under aligned/staggered layouts and various wake steering control strategies. Results show that the change of upstream turbines yaw angles effectively increases the downstream wake flow velocities, especially within the rotor heights. The maximum wind farm power output can increase up to 8.5% with the yawing of first and second rows in the same direction for aligned wind farm, while the impacts of wake steering control on staggered wind farm is less distinctive. Measurements on the instantaneous wind loading of turbines in the first row reveal that the growth of yaw angle significantly increases the fatigue loading in the side-force direction across all frequency components.   

Experimental Study of Flow-Induced Oscillations of a Flexible Symmetric Hydrofoil

Flow-induced oscillations of a flexible symmetric hydrofoil with a NACA 0021 cross section are studied experimentally for a range of flow velocities and angles of attack. The hydrofoil is made of silicone rubber with a chord length of C = 2 cm and length of L = 34.7 cm, corresponding to an aspect ratio of AR = 17. Experiments are conducted in a recirculating water tunnel whereby displacement of the hydrofoil at discrete points along its length are recorded using a Panasonic DMC-FZ200 camera in the crossflow direction and Phantom Miro M110 high speed camera in the inline direction. Displacement is then quantified using Adobe After Effects tracking software. Vortex-induced vibration is observed to occur whereby the largest amplitudes of oscillations occur when the chord length of the hydrofoil is placed perpendicular to the direction of the flow, α = 90°. Performed simultaneously with the displacement measurements, hydrogen bubble flow visualization reveal the vortex shedding patterns observed in the wake of the hydrofoil.   

Using an Active Gurney Flap to Improve the Aerodynamic Performance of a Wind Turbine Blade

A Gurney flap - a small vertical tab installed near the trailing edge of the pressure surface of a wing - is often used to modify aerodynamic performance and increase the down force in race cars. Here we report the design, fabrication and testing of an ``Active Gurney Flap'' (AGF) which can be raised and lowered on demand, and is proposed for use on wind turbines to provide real-time control of the aerodynamic and structural loads experienced by the turbine blade. The torque required to raise and lower the flap was measured on a flat plate wind tunnel model, and a turbine blade wing section (chord = 0.25m, span = 1m) was fabricated and tested over a range of angles of attack at Reynolds numbers between 160,000 and 414,000. Lift and Pitching moments were measured using an aerodynamic sting. We find that the deployment torque is largely independent of Reynolds number and deployment speed. The AGF increases the lift and pitching moment of the wing over a wide range of angles of attack, with some dependence on Re, which affects the boundary layer transition and separation.   

Flow control over wind turbine blades using sweeping jets in a gust and shear environment from a windshaper, a multifan array wind generator

The performance of wind turbine blades instrumented with sweeping jets have been tested in spatially and temporally varying relative-wind conditions. Sweeping jets or fluidic oscillators allow Active Flow Control on an airfoil surface. A test bench was designed and built in order to channel air flow through a rotating hub and into three separate blades in a controlled manner. The blades were instrumented with a number of sweeping jets along their span. This research was enabled by the use of a windshaper, a new family of wind-generating facilities, which consists of an array of a large number of fans (wind-pixels) that may be arranged in various patterns and activated on demand. It is in some ways a digital wind facility that can be programmed to generate arbitrary winds of variable intensity and directions, such as uniform flows, gusts, and shear flows. In particular, sweeping jets were triggered independently on each blade depending on local flow conditions at the blade location. Force coefficients, torque coefficients, and efficiencies were evaluated in various flow conditions.