Sven Schmitz

Characterization of Three-Dimensional Effects for the Rotating and Parked NREL Phase VI Wind Turbine

This paper addresses three-dimensional effects which are pertinent to wind turbine aerodynamics. Two computational models were applied to the National Renewable Energy Laboratory Phase VI Rotor under rotating and parked conditions, a vortex line method using a prescribed wake, and a parallelized coupled Navier-Stokes/vortex-panel solver (PCS). The linking of the spanwise distribution of bound circulation between both models enabled the quantification of three-dimensional effects with PCS. For the rotating turbine under fully attached flow conditions, the effects of the vortex sheet dissipation and replacement by a rolled-up vortex on the computed radial force coefficients were investigated. A quantitative analysis of both radial pumping and Coriolis effect, known as the Himmelskamp effect, was performed for viscous as well as in viscid flow. For the parked turbine, both models were applied at various pitch angles corresponding to fully attached as well as stalled flow. For partially stalled flow, computed results revealed a vortical structure trailing from the blades upper surface close to the 40% radial station. This trailing vortex was documented as a highly unsteady flow structure in an earlier detached eddy simulation by another group, however, it was not directly observed experimentally but only inferred. Computed results show very good agreement with measured wind tunnel data for the PCS model. Finally, a new method for extracting three-dimensional airfoil data is proposed that is particularly well suited for highly stalled flow conditions.

A Parallelized Coupled Navier-Stokes/Vortex-Panel Solver

A Navier-Stokes solver, CFX V5.6, is coupled with an in-house developed Vortex-Panel method for the numerical analysis of wind turbines. The Navier-Stokes zone is confined to the near-field around one wind turbine blade, the Vortex-Panel method models the entire vortex sheet of a two-bladed rotor and accounts for the far-field. This coupling methodology reduces both numerical diffusion and computational cost. The parallelized coupled solver (PCS) is applied to the NREL Phase VI rotor configuration under no-yaw conditions. Fully turbulent flow is assumed using the k-e and k-ω turbulence models. Results obtained are very encouraging for fully attached flow. For separated and partially stalled flow, results are in good agreement with experimental data. Discrepancies observed between the turbulence models are attributed to different prediction of the onset of separation. This is revealed by two-dimensional (2D) results of the S809 airfoil.

Accuracy of State-of-the-Art Actuator-Line Modeling for Wind Turbine Wakes

The current actuator line method (ALM) within an OpenFOAM computational fluid dynamics (CFD) solver was used to perform simulations of the NREL Phase VI rotor under rotating and parked conditions, two fixed-wing designs both with an elliptic spanwise loading, and the NREL 5-MW turbine. The objective of this work is to assess and improve the accuracy of the state-of-the-art ALM in predicting rotor blade loads, particularly by focusing on the method used to project the actuator forces onto the flow field as body forces. Results obtained for sectional normal and tangential force coefficients were compared to available experimental data and to the in-house performance code XTurb-PSU. It was observed that the ALM results agree well with measured data and results obtained from XTurb-PSU except in the root and tip regions if a three-dimensional Gaussian of width, e, constant along the blade span is used to project the actuator force onto the flow field. A new method is proposed where the Gaussian width, e, varies along the blade span following an elliptic distribution. A general criterion is derived that applies to any planform shape. It is found that the new criterion for e leads to improved prediction of blade tip loads for a variety of blade planforms and rotor conditions considered.

Application of a 'Parallelized Coupled Navier -Stokes/Vortex - Panel Solver' to the NREL Phase VI Rotor

A commercially available Navier -Stokes solver, CFX V5.6, is coupled with an in -house developed Vortex -Panel method for the numerical analysis of wind turbines. The Navier Stokes zone is confined to the near -field around one wind turbine blade, the Vortex -Panel method models the entire vortex s heet of a two -bladed rotor and accounts for the far -field. This coup ling methodology reduces both numerical diffusion and computational cost. The coupled solver is parallelized on a cluster of 4 processors. The parallelized coupled solver (PCS) is applied to some distinctive cases of the NREL Phase VI rotor configuration with and without flow separation under steady and no -yaw conditions. Fully turbulent flow is assumed using the k -� and k -� turbulence models. Calculations performed with the coupled solver show very good agreement with experiments for fully attached flow. For separated and partially stalled flow, the k -� model overpredicts rotor power while the k -� model still shows better agreement with experiments. Discrepancies between the two turbulence models are related to different prediction of the onset of separation. This is revealed by 2D airfoil data of the S809 profile.

Inviscid Circulatory-Pressure Field Derived from the Incompressible Navier-Stokes Equations

The static-pressure field in the steady and incompressible Navier–Stokes momentum equation is decomposed into circulatory (inviscid) and dissipative (viscous) partial-pressure fields. It is shown analytically that the circulatory-pressure integral over the surface of a lifting body of thickness recovers the lift generating Kutta–Joukowski theorem in the far field, and results in Maskell’s formula for the vortex-induced drag plus an additional pressure-loss term that tends to zero for an infinitely thin wake. A Poisson equation for the circulatory-pressure field is implemented as a transport equation into the FLUENT 13 solver. Numerical examples include a circular cylinder at Re=8.5×105, the S809 airfoil at Re=2×106, and the ONERA M6 wing at Re=1×106. It is shown that the circulatory-pressure field does indeed behave as an inviscid pressure field of a fully viscous solution, and provides insight into the nature of pressure drag and its contributions to local form and vortex-induced drag.

Wind Turbines under Atmospheric Icing Conditions - Ice Accretion Modeling, Aerodynamics, and Control Strategies for Mitigating Performance Degradation

This thesis presents a combined engineering methodology of ice accretion, airfoil data, and rotor performance analysis of wind turbines subject to moderate atmospheric icing conditions. The Turbine Icing Operation Control System (TIOCS) is based on strip theory for both ice accretion and aerodynamic modeling. The tool is valid for small amounts of accreted ice in the order of a few percent of the sectional airfoil chord. The TIOCS methodology is a fast engineering analysis tool for the wind industry and wind turbine operators that allows for nding guidelines for wind turbine operation during and post moderate icing events. In this thesis, the icing event was relatively short (less than an hour) and three control strategies were explored to determine wind turbine performance degradation. Preliminary results obtained for the NREL Phase VI rotor, the NREL 5MW rotor, and the in-house designed PSU-2.5 MW wind turbines subject to a representative icing condition indicate that performance degradation with respect to power loss can be mitigated with appropriate control strategies during and post an icing event.

Blade Load Unsteadiness and Turbulence Statistics in an Actuator-Line Computed Turbine–Turbine Interaction Problem

The objective of this study is to investigate how different volumetric projection techniques used in actuator-line modeling affect the unsteady blade loads and wake turbulence statistics. The two techniques for the body-force projection radius are based on either (i) the grid spacing or (ii) the combination of grid spacing and an equivalent elliptic blade planform. An array of two National Renewable Energy Laboratory 5-MW turbines separated by seven rotor diameters is simulated for 2000 s (about rotor 300 revolutions) within a large-eddy simulation (LES) solver of the neutral and moderately convective atmospheric boundary layer (ABL). The statistics of sectional angle of attack (AOA), blade loads, and turbine power histories are quantified. Moreover, the degree of unsteadiness of sectional blade loads in response to atmospheric and wake turbulence is computed via a reduced frequency based on the rate-of-change in sectional AOA. The goal of this work is to make the wind energy community aware of the uncertainties associated with actuator-line modeling approaches.

Higher-Harmonic Deployment of Trailing-Edge Flaps for Rotor-Performance Enhancement and Vibration Reduction

A computational method involving evolutionary-based optimization to provide an optimal set of higher-harmonic deployment schedules for a multisegment trailing-edge flap is investigated. The trailing-edge flap is added to the UH-60A’s rotor, with the flap’s span, deflection angles, and start/end deployment azimuth positions all optimized to minimize the total rotor power and the resulting hub vibratory loads. The formal optimization effort is carried out through the coupling of a comprehensive analysis code and one of two evolutionary algorithm-based optimizers. With regard to a single-segment flap, peak power savings over the flight envelope reached 9.5% (at an advance ratio of 0.30) with associated out-of-plane and in-plane hub vibration reductions of 66 and 22%, respectively. The dual-segment trailing-edge flap optimization with a span limitation yielded power savings of 8.9% at the same flight condition. Multi-objective optimizations were performed at a target flight condition with an advance ratio of ...

Modeling Wind Turbine Tower and Nacelle Effects within an Actuator Line Model

It is common practice in computing wind plant aerodynamics with computational fluid dynamics to represent the turbine rotors using actuator lines or disks in which body forces are applied to the flow field. It is less common in such work to include the effect of the tower and nacelle. Here we examine ways to include the effect of the tower and nacelle in a body-force setting without having to explicitly resolve them using complex geometryconforming meshes. We feel that including their effect is certainly important in better predicting the near wake, and may be of importance in the far wake. Recent research by others suggests that the nacelle wake causes interactions with the rotor wake that affect the meandering behavior of the rotor wake far downstream, and properly capturing meandering is important to computing both unsteady power and mechanical loads in a wind plant. We present different body force tower and nacelle representations of incremental complexity. We then apply these methods to two different wind turbine cases, one with emphasis on the wake, and the other with emphasis on blade loads. We show that these methods are relatively easy to implement and are capable of capturing the gross effects of towers and nacelles.

Methodology to Determine a Tip-Loss Factor for Highly Loaded Wind Turbines

The commonly observed overprediction of tip loads on wind-turbine blades by classical blade-element momentum theory is investigated by means of an analytical method that determines the exact tip-loss factor for a given blade flow angle. The analytical method is general and can be applied to any higher-fidelity computational method such as free-wake methods or computational fluid dynamics analyses. In this work, the higher-order free-wake method WindDVE is used to compute tip-vortex rollup and wake expansion in the near wake of a highly loaded wind-turbine rotor. The resulting spanwise distributions of the blade flow angle serve as input to the analytical method that is subsequently tested for the National Renewable Energy Laboratory phase 6 rotor by implementing a corrected tip-loss factor into the blade-element code XTurb. It is found that a simple modification can be added to the classical tip-loss factor in blade-element momentum theory that leads to improved prediction of blade tip loads at no additio...