Bulletin of the American Physical Society
70th Annual Meeting of the APS Division of Fluid Dynamics
Volume 62, Number 14
Sunday–Tuesday, November 19–21, 2017; Denver, Colorado
Session D17: Aerodynamics: Wind Energy IAerodynamics Energy
|
Hide Abstracts |
Chair: Giacomo Valerio Iungo, University of Texas at Dallas Room: 605 |
Sunday, November 19, 2017 2:15PM - 2:28PM |
D17.00001: Weather Research and Forecasting model simulation of an onshore wind farm: assessment against LiDAR and SCADA data Christian Santoni, Edgardo J. Garcia-Cartagena, Lu Zhan, Giacomo Valerio Iungo, Stefano Leonardi The integration of wind farm parameterizations into numerical weather prediction models is essential to study power production under realistic conditions. Nevertheless, recent models are unable to capture turbine wake interactions and, consequently, the mean kinetic energy entrainment, which are essential for the development of power optimization models. To address the study of wind turbine wake interaction, one-way nested mesoscale to large-eddy simulation (LES) were performed using the Weather Research and Forecasting model (WRF). The simulation contains five nested domains modeling the mesoscale wind on the entire North Texas Panhandle region to the microscale wind fluctuations and turbine wakes of a wind farm located at Panhandle, Texas. The wind speed, direction and boundary layer profile obtained from WRF were compared against measurements obtained with a sonic anemometer and light detection and ranging system located within the wind farm. Additionally, the power production were assessed against measurements obtained from the supervisory control and data acquisition system located in each turbine. Furthermore, to incorporate the turbines into very coarse LES, a modification to the implementation of the wind farm parameterization by Fitch et al. (2012) is proposed. [Preview Abstract] |
Sunday, November 19, 2017 2:28PM - 2:41PM |
D17.00002: A simple and complete model for wind turbine wakes over complex terrain Nick Rommelfanger, Mai Rajborirug, Paolo Luzzatto-Fegiz Simple models for turbine wakes have been used extensively in the wind energy community, both as independent tools, as well as to complement more refined and computationally-intensive techniques. These models typically prescribe empirical relations for how the wake radius grows with downstream distance $x$ and obtain the wake velocity at each $x$ through the application of either mass conservation, or of both mass and momentum conservation (e.g. Kati\'{c} et al. 1986; Frandsen et al. 2006; Bastankhah {\&} Port\'{e}-Agel 2014). Since these models assume a global behavior of the wake (for example, linear spreading with $x)$ they cannot respond to local changes in background flow, as may occur over complex terrain. Instead of assuming a global wake shape, we develop a model by relying on a local assumption for the growth of the turbulent interface. To this end, we introduce to wind turbine wakes the use of the entrainment hypothesis, which has been used extensively in other areas of geophysical fluid dynamics. We obtain two coupled ordinary differential equations for mass and momentum conservation, which can be readily solved with a prescribed background pressure gradient. Our model is in good agreement with published data for the development of wakes over complex terrain. [Preview Abstract] |
Sunday, November 19, 2017 2:41PM - 2:54PM |
D17.00003: ABSTRACT WITHDRAWN |
Sunday, November 19, 2017 2:54PM - 3:07PM |
D17.00004: Dispersive stresses in wind farms Antonio Segalini, Robert Braunbehrens, Ann Hyvarinen One of the most famous models of wind farms is provided by the assumption that the farm can be approximated as a horizontally-homogeneous forest canopy with vertically-varying force intensity. By means of this approximation, the flow-motion equations become drastically simpler, as many of the three-dimensional effects are gone. However, the application of the horizontal average operator to the RANS equations leads to the appearance of new transport terms (called dispersive stresses) originating from the horizontal (small-scale) variation of the mean velocity field. Since these terms are related to the individual turbine signature, they are expected to vanish outside the roughness sublayer, providing a definition for the latter. In the present work, an assessment of the dispersive stresses is performed by means of a wake-model approach and through the linearised code ORFEUS developed at KTH. Both approaches are very fast and enable the characterization of a large number of wind-farm layouts. The dispersive stress tensor and its effect on the turbulence closure models are investigated, providing guidelines for those simulations where it is impossible to resolve the farm at a turbine scale due to grid requirements (as, for instance, mesoscale simulations). [Preview Abstract] |
Sunday, November 19, 2017 3:07PM - 3:20PM |
D17.00005: Wind Farm LES Simulations Using an Overset Methodology Shreyas Ananthan, Shashank Yellapantula Accurate simulation of wind farm wakes under realistic atmospheric inflow conditions and complex terrain requires modeling a wide range of length and time scales. The computational domain can span several kilometers while requiring mesh resolutions in $O(10^{-6})$ to adequately resolve the boundary layer on the blade surface. Overset mesh methodology offers an attractive option to address the disparate range of length scales; it allows embedding body-confirming meshes around turbine geomtries within nested wake capturing meshes of varying resolutions necessary to accurately model the inflow turbulence and the resulting wake structures. Dynamic overset hole-cutting algorithms permit relative mesh motion that allow this nested mesh structure to track unsteady inflow direction changes, turbine control changes (yaw and pitch), and wake propagation. An LES model with overset mesh for localized mesh refinement is used to analyze wind farm wakes and performance and compared with local mesh refinements using non-conformal (hanging node) unstructured meshes. Turbine structures will be modeled using both actuator line approaches and fully-resolved structures to test the efficacy of overset methods for wind farm applications. [Preview Abstract] |
(Author Not Attending)
|
D17.00006: Wind Plant Power Optimization and Control under Uncertainty. Pankaj Jha, Demet Ulker, Kyle Hutchings, Gregory Oxley The development of optimized cooperative wind plant control involves the coordinated operation of individual turbines co-located within a wind plant to improve the overall power production. This is typically achieved by manipulating the trajectory and intensity of wake interactions between nearby turbines, thereby reducing wake losses. However, there are various types of uncertainties involved, such as turbulent inflow and microscale and turbine model input parameters. In a recent NREL-Envision collaboration, a controller that performs wake steering was designed and implemented for the Longyuan Rudong offshore wind plant in Jiangsu, China. The Rudong site contains 25 Envision EN136-4 MW turbines, of which a subset was selected for the field test campaign consisting of the front two rows for the northeasterly wind direction. In the first row, a turbine was selected as the reference turbine, providing comparison power data, while another was selected as the controlled turbine. This controlled turbine wakes three different turbines in the second row depending on the wind direction. A yaw misalignment strategy was designed using Envision's GWCFD, a multi-fidelity plant-scale CFD tool based on SOWFA with a generalized actuator disc (GAD) turbine model, which, in turn, was used to tune NREL's FLORIS model used for wake steering and yaw control optimization. The presentation will account for some associated uncertainties, such as those in atmospheric turbulence and wake profile. [Preview Abstract] |
Sunday, November 19, 2017 3:33PM - 3:46PM |
D17.00007: Control strategies for wind farm power optimization: LES study Umberto Ciri, Mario Rotea, Stefano Leonardi Turbines in wind farms operate in off-design conditions as wake interactions occur for particular wind directions. Advanced wind farm control strategies aim at coordinating and adjusting turbine operations to mitigate power losses in such conditions. Coordination is achieved by controlling on upstream turbines either the wake intensity, through the blade pitch angle or the generator torque, or the wake direction, through yaw misalignment. Downstream turbines can be adapted to work in waked conditions and limit power losses, using the blade pitch angle or the generator torque. As wind conditions in wind farm operations may change significantly, it is difficult to determine and parameterize the variations of the coordinated optimal settings. An alternative is model-free control and optimization of wind farms, which does not require any parameterization and can track the optimal settings as conditions vary. In this work, we employ a model-free optimization algorithm, extremum-seeking control, to find the optimal set-points of generator torque, blade pitch and yaw angle for a three-turbine configuration. Large-Eddy Simulations are used to provide a virtual environment to evaluate the performance of the control strategies under realistic, unsteady incoming wind. [Preview Abstract] |
Sunday, November 19, 2017 3:46PM - 3:59PM |
D17.00008: RANS simulations of wind turbine wakes: optimal tuning of turbulence closure and aerodynamic loads from LiDAR and SCADA data. Stefano Letizia, Matteo Puccioni, Lu Zhan, Francesco Viola, Simone Camarri, Giacomo Valerio Iungo Numerical simulations of wakes produced by utility-scale wind turbines still present challenges related to the variability of the atmospheric conditions and, in the most of the cases, the lack of information about the geometry and aerodynamic performance of the wind turbine blades. In order to overcome the mentioned difficulties, we propose a RANS solver for which turbine aerodynamic forcing and turbulence closure are calibrated through LiDAR and SCADA data acquired for an onshore wind farm. The wind farm under examination is located in North Texas over a relatively flat terrain. The experimental data are leveraged to maximize accuracy of the RANS predictions in terms of wake velocity field and power capture for different atmospheric stability conditions and settings of the wind turbines. The optimization of the RANS parameters is performed through an adjoint-RANS formulation and a gradient-based procedure. The optimally-tuned aerodynamic forcing and turbulence closure are then analyzed in order to investigate effects of the atmospheric stability on the evolution of wind turbine wakes and power performance. The proposed RANS solver has low computational costs comparable to those of wake engineering models, which make it a compelling tool for wind farm control and optimization. [Preview Abstract] |
Sunday, November 19, 2017 3:59PM - 4:12PM |
D17.00009: Operation and Equivalent Loads of Wind Turbines in Large Wind Farms Soren Juhl Andersen, Jens Norkaer Sorensen, Robert Flemming Mikkelsen Wind farms continue to grow in size and as the technology matures, the design of wind farms move towards including dynamic effects besides merely annual power production estimates. The unsteady operation of wind turbines in large wind farms has been modelled with EllipSys3D(Michelsen, 1992, and S{\o}rensen, 1995) for a number of different scenarios using a fully coupled large eddy simulations(LES) and aero-elastic framework. The turbines are represented in the flow fields using the actuator line method(S{\o}rensen and Shen, 2002), where the aerodynamic forces and deflections are derived from an aero-elastic code, Flex5({\O}ye, 1996). The simulations constitute a database of full turbine operation in terms of both production and loads for various wind speeds, turbulence intensities, and turbine spacings. The operating conditions are examined in terms of averaged power production and thrust force, as well as $10$min equivalent flapwise bending, yaw, and tilt moment loads. The analyses focus on how the performance and loads change throughout a given farm as well as comparing how various input parameters affect the operation and loads of the wind turbines during different scenarios. [Preview Abstract] |
Sunday, November 19, 2017 4:12PM - 4:25PM |
D17.00010: Dynamic wake model with coordinated pitch and torque control of wind farms for power tracking Carl Shapiro, Johan Meyers, Charles Meneveau, Dennice Gayme Control of wind farm power production, where wind turbines within a wind farm coordinate to follow a time-varying power set point, is vital for increasing renewable energy participation in the power grid. Previous work developed a one-dimensional convection-diffusion equation describing the advection of the velocity deficit behind each turbine (wake) as well the turbulent mixing of the wake with the surrounding fluid. Proof-of-concept simulations demonstrated that a receding horizon controller built around this time-dependent model can effectively provide power tracking services by modulating the thrust coefficients of individual wind turbines. In this work, we extend this model-based controller to include pitch angle and generator torque control and the first-order dynamics of the drive train. Including these dynamics allows us to investigate control strategies for providing kinetic energy reserves to the grid, i.e. storing kinetic energy from the wind in the rotating mass of the wind turbine rotor for later use. [Preview Abstract] |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700