Session C13: Interact: Wind Energy: Wakes and Wake Interactions

INTERACT FLASH TALKS: Wind Energy: Wakes and Wake Interations

Each Interact Flash Talk will last around 1 minute, followed by around 30 seconds of transition time.   

A Unified Momentum Model for rotor aerodynamics and wakes across operating regimes

One-dimensional momentum theory, derived in the 19th century, is the predominant model used in engineering rotors including wind turbines, propellers, helicopters, drones, and hydrokinetic turbines. The theory represents the rotor as a porous actuator disk that imparts a thrust force on the flow, generating induced velocities and a wake. The classical theory breaks down at higher thrust coefficients and for any misalignment between inflow and rotor, which are commonly encountered in practice. Current models rely on empiricism to address these regimes. This study reports a Unified Momentum Model that predicts rotor aerodynamics and wakes across arbitrary thrust coefficients and rotor-inflow misalignments. Using conservation of mass, momentum, and energy, the limiting assumptions of the classical theory are eliminated by modeling the pressure deficit in the rotor wake, using a solution to the differential Euler equations, and by accounting for arbitrary rotor misalignment with a lifting line model. The Unified Momentum Model is validated against large eddy simulations and is also coupled with a blade element model to result in a new physics-based blade-element momentum (BEM) model without empirical corrections. Finally, the model is leveraged in applications including wind farm flow control and control co-design. The model provides a new basis for rotor modeling, design, and control tools from first-principles, rather than starting from classical theory coupled with empirical corrections.   

Wake Recovery in Downwind Wind Turbines

Creating a turbine design that allows for more economical and energy efficient farms will positively impact the wind energy industry. Downwind turbines are less susceptible to blade-tower collisions which could allow for larger, lighter, and more flexible blade designs. The potential of downwind turbines to increase power production compared to an upwind design are considered as well as the possibility for a tighter grid configuration. Unlike upwind configurations, the impact of tower wake on the overall efficiency and power production needs to be considered. These experiments were performed in the wind tunnel at Portland State University. Upwind and downwind configurations of scaled wind turbine models were compared to evaluate their wake recovery at wind speeds ranging from 3 m/s to 12m/s in an active grid with turbulent flows. Two camera stereo particle image velocimetry (PIV) were used to visualize and measure flow behavior. Mean velocities and Reynolds stresses were measured through the use of PIV; accessing three components of velocity over two planes. Power was measured through generators mounted in the hub of the turbines and power curves were developed. The measured data obtained from the experiments includes power, mean velocities, and turbulent Reynolds stresses.   

Near and far wake energy transfer mechanisms behind a wind turbine

Wake meandering, a large diffuse far wake coherent structure, and near wake strong vortical coherent structures induced by blades and the center-body nacelle exhibit regular Strouhal number characteristics in the wake of a wind turbine. The coherent structures are subject to turbulence kinetic energy transfer mechanisms as they evolve, break down, and diffuse. The interactions between coherent scales manifest as triads under a sum-zero frequency condition. To elucidate the interactions, we employ large-eddy simulation of a utility-scale wind turbine with a resolution sufficient to capture spatio-temporal evolution of dominant vortical and diffuse coherent structures and scales in the wake. A precursory inflow is created with a broad range of upwind length-scales similar to an atmospheric boundary layer. The bispectrum, dominant triadic interactions, and scale-specific coherent kinetic energy budget terms are quantified with the scale-specific energy transfer method, based on triple decomposition and dynamic mode decomposition. The triadic interactions present a "web'' or network of connections that transfer kinetic energy among upwind, turbine, and the wake meandering scales. This is observed by the modulation of upwind scales by the wind turbine, the breakdown and diffusion of coherent vortical structures, and the genesis of wake meandering.   

Turbulence coherence in wind farms

Wind power has inherent variability across a wide range of scales due to atmospheric turbulence, and for power systems it is crucial to understand the power fluctuations of extended wind farms. Models for wind farm power fluctuations assume that turbines are passive probes of the atmospheric boundary layer, and they primarily focus on the impact of atmospheric turbulence. We employ Large Eddy Simulation (LES) to demonstrate that dynamic changes in thrust (C_T) and power (C_P) coefficients affect the coherence of velocity and power fluctuations in consecutive turbine rows. We simulated a wind farm with 28 DTU 10MW wind turbines, arranged in 7 rows and 4 columns. We consider various inflow wind speeds to examine the velocity and power coherence between turbine pairs under three scenarios: (I) a fully developed region where all turbines operate below rated power with fixed C_T and C_P, (II) the front row operates above-rated power while the downstream row operates below rated power, and (III) both rows operate above rated power.

In scenario I, the random sweeping hypothesis turbulence model by Tobin and Chamorro, JFM855, 1116-1129(2018) can effectively predict the coherence between the turbines. However, in scenario II and III, the model fails to capture the simulation results. This discrepancy arises due to the operation of above-rated turbines, with dynamically varying C_T and C_P, which have a distinctly different effect on the flow than turbines operating with fixed C_T and C_P.   

Characterizing Turbulence in the Background Flow and Wake of a Stationary Porous Disk: A Two-phase Wind Tunnel Study Under Dry Conditions

Some US offshore sites can experience intense precipitation, which warrants investigating its impact on wind turbine performance. This work is the preliminary part of a study on the influence of inertial particles on the wake dynamics of a stationary porous disk. The first part of the work focuses on the experimental design and construction of a two-phase wind tunnel, emphasizing the achievement of quasi-homogeneous isotropic turbulence (HIT) in the open tunnel configuration with inflow wind conditions without the inertial particles, also referred to as 'dry conditions.' Constant Temperature Anemometry (CTA) hotwire measurements are performed to characterize turbulence in the background flow in the tunnel, including investigating the effects of adding honeycombs, screens, and spray grids on the flow quality. In addition, 2D Computational Fluid Dynamic (CFD) studies are performed to complement this experimental work with an emphasis on enhancing the flow quality in the tunnel. The second part of the work involves studying the turbulence in the wake of a stationary porous disk under dry conditions. Turbulence statistics are calculated for the near and far wake of the porous disk and compared against the background flow in the open tunnel configuration.   

Wake Dynamics in the Generalized Quasilinear Approximation

Wind energy is an abundant renewable resource in a rapid growth period, reaching 117 GW of installed capacity in 2023: a 50% increase over the prior year. In the US, offshore wind, in particular, is experiencing dramatic growth. Because of the expense of offshore operation, wind turbines are clustered into farms. In order to maximize the efficiency of these farms, a detailed understanding of the structure of the combined wake dynamics as they interact with the marine atmospheric boundary layer is needed. We have begun a multi-year effort to use the techniques of direct statistical simulation and generalized quasilinear (GQL) approximation to understand the dynamics of offshore wind. As a first step, we report our findings on the classical Von Karman vortex street under the GQL approximation. GQL approximates the nonlinear effects by careful decimation of triadic interactions, ranging from the quasilinear approach in which fluctuations interact only with the mean all the way to a fully nonlinear direct numerical simulation. We find that very few nonlinear interactions are required to reproduce both wake structure and quantitative Strouhal-Reynolds relationships. We will conclude by suggesting generalizations of this work that are more directly applicable to offshore wind farms.   

Assessment of low-fidelity models for control of floating offshore wind farms

The U.S. Department of Energy (DOE) identifies floating offshore wind as a key technology in the transition to clean energy, looking to harness wind power over deep waters to power densely populated coastal regions. Doing so will require significant scientific and technological advancements, with the DOE targeting a 70% cost reduction (to $45 per megawatt-hour) by 2035. In support of this goal, Pacific Northwest National Lab (PNNL) is leading an Energy Earthshot Research Center (EERC) called Addressing Challenges in Energy: Floating Wind in a Changing Climate (ACE-FWICC). One of its aims is to better understand the effect of met-ocean conditions on the performance of floating offshore wind farms (FOWFs), and to develop advanced models and controllers that account for those effects to achieve farm-level objectives (e.g., maximizing energy production). In this presentation, we share preliminary results from our assessment of low-fidelity models of FOWF dynamics; such models are necessary for rapid control design and real-time control. We focus on FOWFSimDyn, a nonlinear model that incorporates two-dimensional wake effects and floating platform dynamics, comparing its predictions with those of established tools like FLORIS and FAST.Farm.   

Wind Turbine Wake Dynamics and Associated Thermal Interactions Over a Diurnal Cycle

The atmospheric boundary layer (ABL) undergoes significant changes in its structure and dynamics throughout a typical diurnal cycle. These changes can affect the behavior of wind turbine wakes in wind farms. The wakes can, in turn, alter convective heat transfer over the ground surface, leading to two-way coupling and spatial heterogeneity in surface heat fluxes. Here, we investigate the impact of spatially heterogeneous surface heat fluxes on the temperature field, as well as the wake flow structure and power time series within a wind farm during a diurnal cycle using LES. We apply a concurrent precursor inflow method, an advanced filtered actuator-line approach to represent wind, and a local 1D soil heat conduction model coupled with the LES. we incorporate Coriolis forcing that induces wind veer, causing the mean wind direction in the precursor domain to continuously vary over a 24 hour period. LES results obtained after periodic behavior is achieved and reveal that wind turbine wakes have a significant impact on the temperature field, causing increased surface temperatures behind the wind farm at night. During the morning transition, the wind farm blockage effect results in a reduction in wind speed at hub height upstream of the wind farm. Meanwhile, the formation of a low-level jet and the enhanced turbulent kinetic energy at the last row of the wind farm lead to increased velocities experienced by turbines in that area. Consequently, for a few morning hours, the first row of wind turbines generates less power compared to the last row. The generated dataset will be made available for public access through the JH turbulence database.   

Wind farm blockage and edge effects: A wind tunnel study.

An experiment was conducted inside the UTD-BLAST wind tunnel to study the flow evolving across a wind farm for different numbers of rows, N_r. Porous disks with a hub height of 8 cm and a diameter, D = 4 cm were used to set up either aligned or staggered configurations. Wind speed at different turbine locations was indirectly estimated using thrust measurements from a 1-D force sensor installed at the hub of the turbines. Results show that the increasing number of downstream rows causes a larger blockage effect i.e., more wind speed reduction for the disks at the first row. This effect is more pronounced for the staggered configuration. For both cases, the decrease in wind speed becomes asymptotic when N_r > 5. In the aligned configuration, the reduction is 2.5% compared to the single-row configuration, while in the staggered configuration, it is 3.0%. For configurations with N_r ≥ 5, the turbines situated at the edge encounter greater wind velocity in comparison to the other turbines in the same row. In an aligned configuration with N_r = 13, at row 10, the turbine at the edge encounters about 13% higher wind speed. Once a fully developed flow regime is achieved (with configurations having N_r >11), the row average wind speed at the exit row is about 1-2% higher than the previous row.   

Tip vortex breakdown in a high Reynolds number wind turbine wake

Wind turbines are exposed to widely varying inflow conditions that depend on the local boundary layer meteorology. These inflow conditions are characterised by varying degrees of mean velocity shear and turbulence intensity, which affect the performance and durability of the turbine as well as the downstream evolution of the wake. These effects are challenging to study in the field due to the large scales involved, and most wind tunnel experiments are conducted at low Reynolds numbers. The near wakes of wind turbines are dominated by vortices shed from the tips of the blades. As they advect downstream, the tip vortices form a helical structure with three convoluted spirals. Here, we investigate the effect of the inflow turbulence, inflow shear, and tip speed ratio on their breakdown. We present results from high Reynolds number experiments in the Variable Density Turbulence Tunnel (VDTT) at the Max Planck Institute for Dynamics and Self-Organization. This wind tunnel uses pressurized SF₆ as the working fluid to achieve a diameter-based Reynolds number of Re_D = 3 x 10⁶. Because the VDTT achieves high Reynolds numbers at low velocities, high tip speed ratios can be achieved at reasonable rotation rates. An active grid with 111 individually-controllable paddles is used to generate inflow profiles with varying degrees of velocity shear and turbulence intensity. Streamwise hot-wire measurements at multiple downstream positions quantify the breakdown of the tip vortices behind a MoWiTo 0.6 model turbine.   

Wake Dynamics of a Wind Turbine with an Oscillating Rotation Rate at High Reynolds Numbers

Understanding the wake behavior of wind turbines under unsteady conditions is critical for optimizing wind farm power generation. Dynamic induction control is a method which imposes unsteady thrust conditions on turbines, which could allow the wake to recover more quickly than the static case. Increased wake recovery can contribute to an improvement in the overall power density of a wind farm. The effect of dynamic induction control on the wake dynamics of a horizontal axis wind turbine is experimentally studied at a Reynolds number of $4\times 10^6$, and periodic oscillations of the tip-speed ratio are forced at Strouhal numbers of 0.15, 0.25, and 0.4. Force and velocity measurements are taken to characterize the turbine and its wake. Streamwise and spanwise velocity sweeps are conducted using hot-wire anemometry. While the thrust is observed to vary sinusoidally, there is no significant effect on downstream wake recovery. Instead, results show a traveling wave of velocity fluctuations propagates downstream through the wake. We examine two mechanisms for the generation of the traveling wave: the tip-speed ratio oscillation and the change in location of the breakdown of the helical tip vortices. The phase of the thrust coefficient affects the mechanism of traveling wave generation. These results help to establish the dynamic wake behavior of wind turbines under unsteady conditions at high Reynolds numbers.   

High-Fidelity Remote Sensing: Optimizing Acoustic Tomography for Wind Turbine Wake Measurements

Acoustic tomography is an emerging remote sensing methodology based on an optimal linear mapping between the observed times of flight of acoustic signals between an array of speakers and microphones and the turbulent velocity and temperature fields within. Most often, this is accomplished by defining an optimal stochastic inverse operator that relates modeled and observed velocity and temperature fluctuations, but relies heavily on an assumed distribution of their covariances, each parameterized by a characteristic length scale and standard deviation. This work details the accuracy of the estimated fields to their ground truth values, shown by large eddy simulations. Sensitivity of the method and the retrieved fluctuating fields is contained in the Jacobian of the reconstruction error, describing optimal parameter values and the tolerance of the methods to noise representative of measurement error. In the simulated ABL, the Gaussian distribution of covariances is a good approximation to first order. However, when applying acoustic tomography to wind turbine wakes and other industrial flows, many of the underlying assumptions may need to be revisited to ensure that no biases are introduced to the measured turbulent fields.   

Experimental study of a wind turbine wake in a convective boundary layer

Wind turbines are interacting with the flow in the lowest part of the atmospheric boundary layer, which is directly affected by the temperature of the Earth's surface. During daytime, the surface is mostly warmer than the air above, causing convection. The flows interacting with wind turbines are strongly turbulent and complex, making the prediction of wind turbine electricity production difficult. Laboratory studies allow us to study the effect of surface heating in a controlled environment. We use the Prandtl Wind Tunnel (PWT) at the Max Planck Institute for Dynamics and Self-Organization in Göttingen, Germany. The PWT is equipped with an active grid and heated plate, which we use to create convective, turbulent boundary layer flows. Here we present systematic measurements of surface heat flux and velocities behind a wind turbine model, over different magnitudes of surface heating, inflow velocities, and turbulence intensities. We then discuss some implications of our results on the effect of convection on the wakes of wind turbine flows.   

Lagrangian particle tracking in a high Reynolds number wind turbine wake

Wind turbines operate at high diameter-based Reynolds numbers which leads to wakes dominated by highly turbulent flows. As these turbulent wakes have a direct impact on the performances of the downstream wind turbines, understanding their evolution process is important. The wakes of wind turbines are characterized by the presence of coherent structures like tip vortices that affect mixing phenomena at the wake-freestream interface. Studying the wake dynamic from a Lagrangian perspective gives insight into the transport characteristics and mixing characteristics directly related to the wake recovery. 

Here, we present Lagrangian Particle Tracking (LPT) [1,2] measurements of a wind turbine wake. LPT allows for volumetric measurements of highly turbulent flows by following tracer particles within a predefined volume over an extended period of time. The LPT setup has been developed in the Variable Density Turbulence Tunnel (VDTT) [3] at the Max Planck Institute for Dynamics and Self-Organization. This wind tunnel uses pressurized SF6 to achieve full-scale Reynolds numbers on a small-scale turbine. To obtain turbulent inflow conditions, the velocity fluctuations are modified using an active grid. Illumination is provided by a 300W laser, while high temporal resolution is achieved using four Phantom high-speed cameras. As such, the experimental setup presented here allows spatial and temporal investigation of wind turbine wakes in real-world flow conditions.   

High-fidelity prediction of wind turbine wakes: Enhancing wake models using LES-trained machine learning algorithms

An innovative machine learning (ML) model is introduced to efficiently predict high-fidelity three-dimensional velocity fields in the wakes of utility-scale wind turbines. The model takes low-fidelity velocity fields from an analytical engineering wake model as input and produces high-fidelity velocity fields. Large-eddy simulations (LES) of the Sandia National Lab Scaled Wind Farm Technology (SWiFT) facility at different wind speeds, wind directions, and yaw misalignments of the turbines were performed to generate high-fidelity velocity fields for training and validation. The input to the ML model consists of the three-dimensional velocity field of the SWiFT facility obtained from the Gauss Curl Hybrid (GCH) model. When compared against LES results, the ML model reduced prediction errors of the GCH model from 20% to less than 5%. Additionally, the ML model accurately captured the non-symmetric wake deflection observed for opposing yaw angles in wake steering cases, thereby increasing the accuracy over the GCH model. The computational cost of the trained ML model is comparable to that of the GCH model while providing results nearly as accurate as the high-fidelity LES.   

Traveling waves in the wakes of dynamically controlled wind turbines

The wake dynamics of wind turbines in unsteady flow conditions may have a substantial impact on the wake losses in and overall power density of wind farms. These effects are particularly relevant for wind turbines under dynamic induction control and for floating offshore turbines undergoing platform oscillations in the streamwise direction. In this study, we investigate the effects of periodic oscillations in the rotation rate and streamwise position of wind turbines on their wake dynamics and recovery. We derive an analytical model for the near wake that describes the formation and propagation of traveling waves in the streamwise velocity and wake radius. The predictions of the model show good qualitative agreement with phase-averaged flow-field measurements from a periodically surging turbine in an optically accessible towing tank. These measurements also demonstrate that the identified unsteady flow mechanisms can lead to reductions in the streamwise extent of the wake by over 45%. We also conduct experiments in a high-pressure wind tunnel using a turbine with a periodically varying rotation rate, at Reynolds numbers closer to those of utility-scale turbines. Similar traveling-wave dynamics are evident in the wake, though no significant differences in wake recovery are observed. These results highlight unsteady flow mechanisms in wind-turbine wakes that could be leveraged to enhance the power generation of fixed-bottom and floating offshore wind farms.   

Wake interactions between floating offshore wind turbines: A wind tunnel study

Floating offshore wind turbines (FOWTs) experience 6-degree-of-freedom motion impacting their performance, fatigue loading, and wakes leading to complex interactions between FOWT arrays. For this study, we tested two downscaled models of DTU 10 MW wind turbines with a rotor diameter (D) of 40 cm installed in tandem at the UT Dallas Boundary Layer and Subsonic Wind Tunnel (BLAST) using multi-hole pressure probes, 6-DOF force sensors, and electric generators. The upstream FOWT is installed on a mechatronic emulator reproducing typical turbine motion during offshore operations, specifically imposing sway motion with different non-dimensional amplitude (A/D) from 0 to 0.16, and Strouhal number (St) from 0 to 0.35. FOWT operational tip-speed ratios (TSRs) were set to sub-optimal (4.8), optimal (5.2), and super-optimal (7.5) values, with a downstream turbine located at a distance of 5D or 10D operating at the optimal TSR. Results show that high St and A/D enhance wake recovery, particularly at higher TSRs, increasing power capture and thrust loading of the downstream turbine, causing lateral wake expansion and vertical shrinkage, thus resulting in an elliptical mean wake shape. Increased St and A/D also increase Reynolds stresses and turbulent kinetic energy in the wake.

Decoupling wind shear from atmospheric boundary layer forcing to systematically unravel wind turbine wake dynamics

In atmospheric boundary layer (ABL), friction and Coriolis forces balance pressure gradient forcing to form the structure of the mean wind and turbulence. Wind turbines operate in the ABL, generating wakes through the extraction of mean kinetic energy that negatively impact turbine performance downwind. Many factors influence these wakes, including turbulence, Coriolis forces, thermal stratification, and the ABL velocities, which ubiquitously contain wind speed and direction shear. But wind speed and direction shear are themselves coupled with the balance of forces in the ABL, and further depend on many additional variables, including heating history, orography, and subsidence. Due to this coupling, the effects of each ABL phenomenon on wake dynamics have been challenging to study independently. To isolate the wake dynamics associated with each ABL phenomenon, large eddy simulations (LES) of elementary flows are systematically used to decouple wind shear from Coriolis forces, stratification, and turbulence. Through the analysis of momentum and turbulence budgets, these ABL effects on wake dynamics are parsed into forcing and transport processes. Finally, LES of wind turbine wakes in realistic, stratified ABL conditions are compared with the elementary flows. Transport mechanisms in the ABL are constructed from elementary flows, aiming toward the development of a generalized wake model that captures ABL phenomena from first principles.   

Experimental and numerical investigations on the aerodynamic lift characteristics and flow features of water-treading hovering of two-dimensional elliptic airfoil

This study investigates the aerodynamic performance of a 2D hovering airfoil through experiments and numerical simulations. Here, the water-treading type of hovering kinematics, combining the surging and pitching motions at a hovering frequency of 0.833 Hz, is considered. The airfoil has a rectangular planform with an elliptical cross-section, a chord length of 0.04 m, a thickness of 0.005 m, and an aspect ratio of 10. Three different pitching kinematics: (i) pure sinusoidal, (ii) and (iii) a smoothened trapezoidal waveform with base lengths of 0.2 and 0.4 times the duration of one hovering cycle, respectively, are followed to access the airfoil aerodynamic performance. The pitching amplitude is kept at 30°. It is observed that the time-averaged lift coefficient over a hovering cycle is 11.62% and 9.09% higher for case (iii) hovering mode compared to case (i) and case (ii), respectively. However, the time-averaged aerodynamic power coefficient for case (iii) hovering mode is 62.96% and 57.14% higher than case (i) and case (ii), respectively. The best energy efficiency hovering mode is seen in case (i) by comparing the ratio of the aerodynamic lift to power coefficients. The near-wake flow features at different hovering modes will be thoroughly discussed in the presentation.   

INTERACT DISCUSSION SESSION WITH POSTERS: Wind Energy: Wakes and Wake Interations

After each Flash Talk has concluded, the Interact session will be followed by interactive poster or e-poster presentations, with plenty of time for one-on-one and small group discussions.