Table of contents

Aeroelastic codes based on Blade Element Momentum theory are the standard used by many wind turbine designers. These codes usually include models and corrections for unsteady aerodynamics, tip and root effect, tower shadow and other effects. In general, this kind of codes does not include models to correctly simulate aerodynamic control devices. This paper presents some modifications including the unsteady contributions due to the flap motion (based on indicial models) and the spanwise (3D) effects (based on circulation theory), in order to simulate flaps in the blades. This method can be included in BEM codes in general and it could also be applied to another kind of control devices. The validation and verification show the accuracy of this method using experimental data for two-dimensional unsteady cases, and CFD for three-dimensional steady and unsteady cases.

Due to technology developments, renewable energies are becoming competitive against fossil sources and the number of wind farms is growing, which have to be integrated into power grids. Therefore, accurate power forecast is needed and often operators are charged with penalties in case of imbalance. Yet, wind is a stochastic and very local phenomenon, and therefore hard to predict. It has a high variability in space and time and wind power forecast is challenging. Statistical methods, as Artificial Neural Networks (ANN), are often employed for power forecasting, but they have some shortcomings: they require data sets over several years and are not able to capture tails of wind power distributions. In this work a pure ANN power forecast is compared against a hybrid method, based on the combination of ANN and a physical method using computational fluid dynamics (CFD). The validation case is a wind farm sited in southern Italy in a very complex terrain, with a wide spread turbine layout.

Previous studies have shown the importance of 3D effects when calculating the performance characteristics of a scaled down turbine rotor [1-4]. In this paper the results of 3D RANS (Reynolds-Averaged Navier-Stokes) computations by Make and Vaz [1] are taken to calculate 2D lift and drag coefficients. These coefficients are assigned to FAST (Blade Element Momentum Theory (BEMT) tool from NREL) as input parameters. Then, the rotor characteristics (power and thrust coefficients) are calculated using BEMT. This coupling of RANS and BEMT was previously applied by other parties and is termed here the RANS-BEMT coupled approach. Here the approach is compared to measurements carried out in a wave basin at MARIN applying Froude scaled wind, and the direct 3D RANS computation. The data of both a model and full scale wind turbine are used for the validation and verification. The flow around a turbine blade at full scale has a more 2D character than the flow properties around a turbine blade at model scale (Make and Vaz [1]). Since BEMT assumes 2D flow behaviour, the results of the RANS-BEMT coupled approach agree better with the results of the CFD (Computational Fluid Dynamics) simulation at full- than at model-scale.

The length and flexibility of wind turbine blades are increasing over time. Typically, the dynamic response of the blades is analysed using linear models of blade deflection, enhanced by various ad-hoc non-linear correction models. For blades undergoing large deflections, the small deflection assumption inherent to linear models becomes less valid. It has previously been demonstrated that linear and nonlinear blade models can show significantly different blade response, particularly for blade torsional deflection, leading to load prediction differences. There is a need to evaluate how load predictions from these two approaches compare to measurement data from the field.

In this paper, time domain simulations in turbulent wind are carried out using the aero-elastic code Bladed with linear and non-linear blade deflection models. The turbine blade load and deflection simulation results are compared to measurement data from an onshore prototype of the GE 6MW Haliade turbine, which features 73.5m long LM blades. Both linear and non-linear blade models show a good match to measurement turbine load and blade deflections. Only the blade loads differ significantly between the two models, with other turbine loads not strongly affected. The non-linear blade model gives a better match to the measured blade root flapwise damage equivalent load, suggesting that the flapwise dynamic behaviour is better captured by the non-linear blade model. Conversely, the linear blade model shows a better match to measurements in some areas such as blade edgewise damage equivalent load.

Operating horizontal axis wind turbines create large-scale turbulent wake structures that affect the power output of downwind turbines considerably. The computational prediction of this phenomenon is challenging as efficient low dissipation schemes are necessary that represent the vorticity production by the moving structures accurately and that are able to transport wakes without significant artificial decay over distances of several rotor diameters. We have developed a parallel adaptive lattice Boltzmann method for large eddy simulation of turbulent weakly compressible flows with embedded moving structures that considers these requirements rather naturally and enables first principle simulations of wake-turbine interaction phenomena at reasonable computational costs. The paper describes the employed computational techniques and presents validation simulations for the Mexnext benchmark experiments as well as simulations of the wake propagation in the Scaled Wind Farm Technology (SWIFT) array consisting of three Vestas V27 turbines in triangular arrangement.

The contribution of wake meandering and shape asymmetry to load and power estimates is quantified by comparing aeroelastic simulations initialized with different inflow conditions: an axisymmetric base wake, an unsteady stochastic shape wake, and a large-eddy simulation with rotating actuator-line turbine representation. Time series of blade-root and tower base bending moments are analyzed. We find that meandering has a large contribution to the fluctuation of the loads. Moreover, considering the wake edge intermittence via the stochastic shape model improves the simulation of load and power fluctuations and of the fatigue damage equivalent loads. These results indicate that the stochastic shape wake simulator is a valuable addition to simplified wake models when seeking to obtain higher-fidelity computationally inexpensive predictions of loads and power.

Wind energy has become cost competitive in recent years for several reasons. Among them, wind turbines have become more efficient, increasing its size, both rotor diameter and tower height. This growth in size makes the prediction of the wind flow through wind turbines more challenging. To avoid the computational cost related to resolve the blade boundary layer as well as the atmospheric boundary layer, actuator models have been proposed in the past few years. Among them, the Actuator Line Model (ALM) has shown to reproduce with reasonable accuracy the wind flow in the wake of a wind turbine with moderately computational cost. However, its use to simulate the flow through wind farms requires a spatial resolution and a time step that makes it unaffordable in some cases. The present paper aims to assess the ALM with coarser resolution and larger time step than what is generally recommended, taking into account an atmospheric sheared and turbulent inflow condition and comparing the results with the Actuator Disk Model with Rotation (ADM-R) and experimental data. To accomplish this, a well known wind tunnel campaign is considered as validation case.

Accurate modeling of complex terrain, especially steep terrain, in the simulation of wind fields remains a challenge. It is well known that the terrain-following coordinate transformation method (TFCT) generally used in atmospheric flow simulations is restricted to non-steep terrain with slope angles less than 45 degrees. Due to the advantage of keeping the basic computational grids and numerical schemes unchanged, the immersed boundary method (IBM) has been widely implemented in various numerical codes to handle arbitrary domain geometry including steep terrain. However, IBM could introduce considerable implementation errors in wall modeling through various interpolations because an immersed boundary is generally not co-located with a grid line. In this paper, we perform an intercomparison of TFCT and IBM in large-eddy simulation of a turbulent wind field over a three-dimensional (3D) hill for the purpose of evaluating the implementation errors in IBM. The slopes of the three-dimensional hill are not steep and, therefore, TFCT can be applied. Since TFCT is free from interpolation-induced implementation errors in wall modeling, its results can serve as a reference for the evaluation so that the influence of errors from wall models themselves can be excluded. For TFCT, a new algorithm for solving the pressure Poisson equation in the transformed coordinate system is proposed and first validated for a laminar flow over periodic two-dimensional hills by comparing with a benchmark solution. For the turbulent flow over the 3D hill, the wind-tunnel measurements used for validation contain both vertical and horizontal profiles of mean velocities and variances, thus allowing an in-depth comparison of the numerical models. In this case, TFCT is expected to be preferable to IBM. This is confirmed by the presented results of comparison. It is shown that the implementation errors in IBM lead to large discrepancies between the results obtained by TFCT and IBM near the surface. The effects of different schemes used to implement wall boundary conditions in IBM are studied. The source of errors and possible ways to improve the IBM implementation are discussed.

Dynamic loads play an important roll in the design of wind turbines, but establishing the life-time aerodynamic loads (e.g. extreme and fatigue loads) is a computationally expensive task. Conventional (deterministic) methods to analyze long term loads, which rely on the repeated analysis of multiple different wind samples, are usually too expensive to be included in optimization routines. We present a new stochastic approach, which solves the aerodynamic system equations (Lagrangian vortex model) in the stochastic space, and thus arrive directly at a stochastic description of the coupled loads along a turbine blade. This new approach removes the requirement of analyzing multiple different realizations. Instead, long term loads can be extracted from a single stochastic solution, a procedure that is obviously significantly faster. Despite the reduced analysis time, results obtained from the stochastic approach match deterministic result well for a simple test-case (a stationary blade). In future work, the stochastic method will be extended to rotating blades, thus opening up new avenues to include long term loads into turbine optimization.

The wind engineering community relies on multiphysics engineering software to run nonlinear time-domain simulations e.g. for design-standards-based loads analysis. Although most physics involved in wind energy are nonlinear, linearization of the underlying nonlinear system equations is often advantageous to understand the system response and exploit well- established methods and tools for analyzing linear systems. This paper presents the development and verification of the new linearization functionality of the open-source engineering tool FAST v8 for land-based wind turbines, as well as the concepts and mathematical background needed to understand and apply it correctly.

The wake structure behind a wind turbine, generated by the spectral element code Nek5000, is compared with that from the finite volume code EllipSys3D. The wind turbine blades are modeled using the actuator line method. We conduct the comparison on two different setups. One is based on an idealized rotor approximation with constant circulation imposed along the blades corresponding to Glauert's optimal operating condition, and the other is the Tjffireborg wind turbine. The focus lies on analyzing the differences in the wake structures entailed by the different codes and corresponding setups. The comparisons show good agreement for the defining parameters of the wake such as the wake expansion, helix pitch and circulation of the helical vortices. Differences can be related to the lower numerical dissipation in Nek5000 and to the domain differences at the rotor center. At comparable resolution Nek5000 yields more accurate results. It is observed that in the spectral element method the helical vortices, both at the tip and root of the actuator lines, retain their initial swirl velocity distribution for a longer distance in the near wake. This results in a lower vortex core growth and larger maximum vorticity along the wake. Additionally, it is observed that the break down process of the spiral tip vortices is significantly different between the two methods, with vortex merging occurring immediately after the onset of instability in the finite volume code, while Nek5000 simulations exhibit a 2-3 radii period of vortex pairing before merging.

The QBlade implementation of the Lifting Line Free Vortex Wake method(LLFVW) was tested in conditions analogous to floating platform motion. Comparisons against two independent test cases, using a variety of simulation methods show excellent agreement in thrust forces, rotor power, blade forces and rotor plane induction. Along with the many verifications already undertaken in literature, it seems that the code performs solidly even in these challenging cases. Further to this, the key steps are presented from a new formulation of the instantaneous aerodynamic thrust damping of a wind turbine rotor. A test case with harmonic platform motion and collective pitch is used to demonstrate how combining such tools can lead to better understanding of aeroelastic stability.

The periodic stability of an innovative load reduction system for two-bladed wind turbines is investigated using Floquet theory. The load reduction system introduces an additional cardanic degree of freedom between the hub mount and the nacelle carrier flange. A reduced rotor model is considered to analyze the time-variant behavior as a function of rotor shapes and design parameters. Aerodynamic excitation of the rotor is neglected in this study. The linearized system equations of motion of this system indicate a dependance of the rotor shape and the periodic system stability. Due to the time-variance of the asymmetric rotor the Floquet multipliers are derived to determine system stability and the dominant periodic coefficients. It can be shown that the rotor shape has a significant impact on the stability of gimbal-hinged rotors. Gyroscopic effects are the reason for a stabilization of disc-like rotors, whereas a cylindrical, two-bladed rotor assembly is in general unstable. The introduction of an additional spring-damper coupling is a possibility to stabilize such configurations, but results show that resonance phenomena have to be accounted for.

When representing the blade aerodynamics with rotating actuator lines, the computed forces have to be projected back to the CFD flow field as a volumetric body force. That has been done in the past with a geometrically simple uniform three-dimensional Gaussian at each point along the blade. We argue that the body force can be shaped in a way that better predicts the blade local flow field, the blade load distribution, and the formation of the tip/root vortices. In previous work, we have determined the optimal scales of circular and elliptical Gaussian kernels that best reproduce the local flow field in two-dimensions. In this work we extend the analysis and applications by considering the full three-dimensional blade to test our hypothesis in a highly resolved Large Eddy Simulation.

The current paper investigates the aeroelastic modelling of large, flexible multi- MW wind turbine blades. Most current performance prediction tools make use of the Blade Element Momentum (BEM) model, based upon a number of simplifying assumptions that hold only under steady conditions. This is why a lifting line free vortex wake (LLFVW) algorithm is used here to accurately resolve unsteady wind turbine aerodynamics. A coupling to the structural analysis tool BeamDyn, based on geometrically exact beam theory, allows for time-resolved aeroelastic simulations with highly deflected blades including bend-twist, coupling. Predictions of blade loading and deformation for rigid and flexible blades are analysed with reference to different aerodynamic and structural approaches. The emergency shutdown procedure is chosen as an examplary design load case causing large deflections to place emphasis on the influence of structural coupling and demonstrate the necessity of high fidelity structural models.

In this paper, two methods for the extraction of airfoil coefficients from 3D CFD simulations of a wind turbine rotor are investigated, and these coefficients are used to improve the load prediction of a BEM code. The coefficients are extracted from a number of steady RANS simulations, using either averaging of velocities in annular sections, or an inverse BEM approach for determination of the induction factors in the rotor plane. It is shown that these 3D rotor polars are able to capture the rotational augmentation at the inner part of the blade as well as the load reduction by 3D effects close to the blade tip. They are used as input to a simple BEM code and the results of this BEM with 3D rotor polars are compared to the predictions of BEM with 2D airfoil coefficients plus common empirical corrections for stall delay and tip loss. While BEM with 2D airfoil coefficients produces a very different radial distribution of loads than the RANS simulation, the BEM with 3D rotor polars manages to reproduce the loads from RANS very accurately for a variety of load cases, as long as the blade pitch angle is not too different from the cases from which the polars were extracted.

This paper deals with a comparison of data collected by measurements and a simulation for a complex terrain test site in southern Germany. Lidar, met mast, unmanned aerial vehicle (UAV) measurements of wind speed and direction and Computational Fluid Dynamics (CFD) data are compared to each other. The site is characterised regarding its flow features and the suitability for a wind turbine test field. A Delayed-Detached-Eddy- Simulation (DES) was employed using measurement data to generate generic turbulent inflow. A good agreement of the wind profiles between the different approaches was reached. The terrain slope leads to a speed-up, a change of turbulence intensity as well as to flow angle variations.

In this study, the performance of the mesoscale meteorological Weather Research and Forecast (WRF) model coupled with the microscale computational fluid dynamics based model WindSim is investigated and compared to the performance of WRF alone. The two model set-ups, WRF and WRF-WindSim, have been tested on three high-wind events in February, June and October, over a complex terrain at the Nygårdsfjell wind park in Norway. The wind speeds and wind directions are compared to measurements and the results are evaluated based on root mean square error, bias and standard deviation error. Both model set-ups are able to reproduce the high wind events. For the winter month February the WRF-WindSim performed better than WRF alone, with the root mean square error (RMSE) decreasing from 2.86 to 2.38 and standard deviation error (STDE) decreasing from 2.69 to 2.37. For the two other months no such improvements were found. The best model performance was found in October where the WRF had a RMSE of 1.76 and STDE of 1.68. For June, both model set-ups underestimate the wind speed. Overall, the adopted coupling method of using WRF outputs as virtual climatology for coupling WRF and WindSim did not offer a significant improvement over the complex terrain of Nygårdsfjell. However, the proposed coupling method offers high degree of simplicity when it comes to its application. Further testing is recommended over larger number of test cases to make a significant conclusion.

The current paper presents the effort, in the EU AVATAR project, to establish the necessary requirements to obtain consistent lift over drag ratios among seven CFD codes. The flow around a 2D airfoil case is studied, for both transitional and fully turbulent conditions at Reynolds numbers of 3 × 10⁶ and 15 × 10⁶. The necessary grid resolution, domain size, and iterative convergence criteria to have consistent results are discussed, and suggestions are given for best practice. For the fully turbulent results four out of seven codes provide consistent results. For the laminar-turbulent transitional results only three out of seven provided results, and the agreement is generally lower than for the fully turbulent case.

In large wind farms, the turbulence induced by each turbine results in high overall turbulence levels that can be detrimental for downstream wind turbine components. In the current study, we scrutinize structural loads and dynamics, and their correlation to turbulent flow structures by conducting aeroelastic simulations in wind farms. To this end, a pseudospectral large-eddy simulation solver is coupled with a multibody dynamics module in a multiscale framework. The multirate approach leads us naturally to the development of an aeroelastic actuator sector model that represents the wind turbine forces on the flow. This makes it computationally feasible to simulate long time horizons of the two-way coupled aeroelastic system. Hence, it allows us to look at the interaction of the turbine structure with the turbulent boundary layer and the wakes of multiple turbine arrays, and to get estimates of damage equivalent loads and structural loading statistics, as longer time series are available. Results are shown for two typical wind farm layouts, i.e. aligned and staggered, for above-rated flow regimes.

As a variation of the pseudodynamic testing technique, the real-time hybrid simulation (RTHS) technique is executed in real time, thus allowing investigation of structural systems with rate-dependent components. In this paper, the RTHS is employed for performance evaluation of full-scale liquid sloshing dampers in multi-megawatt wind turbines, where the tuned liquid damper (TLD) is manufactured and tested as the physical substructure while the wind turbine is treated as the numerical substructure and modelled in the computer using a 13-degree-of-freedom (13-DOF) aeroelastic model. Wind turbines with 2 MW and 3 MW capacities have been considered under various turbulent wind conditions. Extensive parametric studies have been performed on the TLD, e.g., various tuning ratios by changing the water level, TLD without and with damping screens (various mesh sizes of the screen considered), and TLD with flat and sloped bottoms. The present study provides useful guidelines for employing sloshing dampers in large wind turbines, and indicates huge potentials of applying RTHS technique in the area of wind energy.

The presented work describes a detailed analysis of the dynamic interactions among mechanical and electrical drivetrain components of a modern wind turbine under the influence of parameter variations, different control mechanisms and transient excitations. For this study, a detailed model of a 2MW wind turbine with a gearbox, a permanent magnet synchronous generator and a full power converter has been developed which considers all relevant characteristics of the mechanical and electrical subsystems. This model includes an accurate representation of the aerodynamics and the mechanical properties of the rotor and the complete mechanical drivetrain. Furthermore, a detailed electrical modelling of the generator, the full scale power converter with discrete switching devices, its filters, the transformer and the grid as well as the control structure is considered. The analysis shows that, considering control measures based on active torsional damping, interactions between mechanical and electrical subsystems can significantly affect the loads and thus the individual lifetime of the components.

In the current paper, a method for deriving the analytical expressions for the velocity and vortex stretching terms as a function of the spherical multipole expansion approximation of the vector potential is presented. These terms are essential in the context of 3D Lagrangian vortex particle methods combined with fast summation techniques. The convergence and computational efficiency of this approach is assessed in the framework of an O(N log N)-type Fast Multipole Method (FMM), by using vorticity particles to simulate a system of coaxial vortex rings for which also the exact results are known. It is found that the current implementation converges rapidly to the exact solution with increasing expansion order and acceptance factor. An investigation into the computational efficiency demonstrated that the O(N log N)-type FMM is already viable for a particle size of only several thousands and that this speedup increases significantly with the number of particles. Finally, it is shown that the implementation of the FMM with the current analytical expressions is at least twice as fast as when opting for using even the simplest implementation of finite differences instead.

Designing floating substructures for the next generation of 10MW and larger wind turbines has introduced new challenges in capturing relevant physical effects in dynamic simulation tools. In achieving technically and economically optimal floating substructures, structural flexibility may increase to the extent that it becomes relevant to include in addition to the standard rigid body substructure modes which are typically described through linear radiation-diffraction theory. This paper describes a method for the inclusion of substructural flexibility in aero-hydro-servo-elastic dynamic simulations for large-volume substructures, including wave-structure interactions, to form the basis of deriving sectional loads and stresses within the substructure. The method is applied to a case study to illustrate the implementation and relevance. It is found that the flexible mode is significantly excited in an extreme event, indicating an increase in predicted substructure internal loads.

Computational Fluid Dynamics (CFD) based on the Unsteady Reynolds Averaged Navier Stokes (URANS) equations have long been widely used to study vertical axis wind turbines (VAWTs). Following a comprehensive experimental survey on the wakes downwind of a troposkien-shaped rotor, a campaign of bi-dimensional simulations is presented here, with the aim of assessing its reliability in reproducing the main features of the flow, also identifying areas needing additional research.

Starting from both a well consolidated turbulence model (k-ω SST) and an unstructured grid typology, the main simulation settings are here manipulated in a convenient form to tackle rotating grids reproducing a VAWT operating in an open jet wind tunnel. The dependence of the numerical predictions from the selected grid spacing is investigated, thus establishing the less refined grid size that is still capable of capturing some relevant flow features such as integral quantities (rotor torque) and local ones (wake velocities).

The numerical simulation of flows over large-scale wind turbine blades without considering the transition from laminar to fully turbulent flow may result in incorrect estimates of the blade loads and performance. Thanks to its relative simplicity and promising results, the Local-Correlation based Transition Modelling concept represents a valid way to include transitional effects into practical CFD simulations. However, the model involves coefficients that need tuning. In this paper, the γ—equation transition model is assessed and calibrated, for a wide range of Reynolds numbers at low Mach, as needed for wind turbine applications. An aerofoil is used to evaluate the original model and calibrate it; while a large scale wind turbine blade is employed to show that the calibrated model can lead to reliable solutions for complex three-dimensional flows. The calibrated model shows promising results for both two-dimensional and three-dimensional flows, even if cross-flow instabilities are neglected.

Actuator models have been used to represent the presence of wind turbines in a simulation in the past few years. The Actuator Line Model (ALM) has shown to reproduce with reasonable accuracy the wind flow through wind turbines under different operational conditions. Nevertheless, there are not many simulations of wind farms performed with the ALM mainly because of its computational cost. The aim of the present paper is to evaluate the ALM in spatial resolutions coarser than what is generally recommended, also using larger time steps, in a simulation of a real wind farm. To accomplish this, simulations of one row of Horns Rev wind farm are performed, for different wind directions. It is concluded that the ALM is able to capture the main features of the interaction between wind turbines relaxing its resolution requirements. A sensitivity analysis is performed to assess the influence of the smearing factor and the spatial resolution.

Small horizontal-axis wind turbines are likely to operate in a broad range of operating flow conditions, often in highly turbulent flow, due, in part, to their varied site placements. This paper compares the computational simulations of the performance of a 5 kW horizontal-axis wind turbine to detailed field measurements, with a particular focus on the impact of unsteady operating conditions on the drivetrain performance and generator output. Results indicate that the current Blade Element Momentum Theory based aerodynamic models under-predict the effect of high turbine yaw on the rotor torque, leading to a difference between predicted and measured shaft speed and power production. Furthermore, the results show discrepancies between the predicted instantaneous turbine yaw performance and measurements.

Due to the natural variability of the wind, it is necessary to conduct thorough wind resource assessments to determine how much energy can be extracted at a given site. Lately, important advancements have been achieved in numerical methods of multiscale models used for high resolution wind simulations over steep topography. As a contribution to this effort, an enhanced numerical method was devised in the mesoscale compressible community (MC2) model of the Meteorological Service of Canada, adapting a new semi-implicit scheme with its imbedded large-eddy simulation (LES) capability for mountainous terrain. This implementation has been verified by simulating the neutrally stratified atmospheric boundary layer (ABL) over flat terrain and a Gaussian ridge. These preliminary results indicate that the enhanced MC2-LES model reproduces efficiently the results reported by other researchers who use similar models with more sophisticated sub-grid scale turbulence schemes. The proposed multiscale method also provides a new wind initialization scheme and additional utilities to improve numerical accuracy and stability. The resulting model can be used to assess the wind resource at meso- and micro-scales, reducing significantly the wind speed overestimation in mountainous areas.

Wake vortices (WVs) generated by aircraft are a source of risk to the following aircraft. The probability of WV related accidents increases in the vicinity of airport runways due to the shorter time of recovery after a WV encounter. Hence, solutions that can reduce the risk of WV encounters are needed to ensure increased flight safety. In this work we propose an interesting approach to model such wake vortices in real time using a hybrid Eulerian- Lagrangian approach. We derive an appropriate mathematical model, and show a comparison of the different types of solvers. We will conclude with a real life application of the methodology by simulating how wake vortices left behind by an aircraft at the Vffirnes airport in Norway get transported and decay under the influence of a background wind and turbulence field. Although the work demonstrates the application in an aviation context the same approach can be used in a wind energy context.

The accuracy of Computational Fluid Dynamics (CFD) models for Atmospheric Boundary Layer (ABL) flows relies largely on the placement of the domain boundaries and the quality of the imposed flow conditions, the inlet boundary in particular. Exploiting the parabolic nature of many ABL flows and of CFD modelled ABL flow in particular, a precursor simulation is used as source of flow data to improve the target domain's inlet flow description over the standard synthetic boundary conditions, one-directionally coupling the solutions to the two simulations. Using the approach, a case of flow over a two wind farm offshore cluster is modelled using two small coupled simulations, matching the results of a single simulation including the full cluster at a significant computational time saving, in the order of 70%. Further savings were shown to be possible by reducing the resolution of the precursor simulation, with negligible impact on the results at the target domain.

The optimisation of a Diffuser Augmented Wind Turbine has traditionally focused on maximising its power output. Optimising the design of the blade and the shape of the diffuser for maximum turbine power over a range of wind velocities is a complex process, as each will influence the others flow regime. In this paper we propose a method that combines the predictions of flow through a diffuser, using computational fluid dynamics, and the flow from a turbine blade using a modified blade element theory to predict the power output of a diffuser augmented wind turbine. Good agreement was found between the predictions from this new method and experimental data from the literature.

This contribution presents a model interface of a wind turbine generator to represent the reciprocal effects between the mechanical and the electromagnetic system. Therefore, a multi-body-simulation (MBS) model in Simpack is set up and coupled with a quasi-static electromagnetic (EM) model of the generator in Matlab/Simulink via co-simulation. Due to lack of data regarding the structural properties of the generator the modal properties of the MBS model are fitted with respect to results of an experimental modal analysis (EMA) on the reference generator. The used method and the results of this approach are presented in this paper. The MB S model and the interface are set up in such a way that the EM forces can be applied to the structure and the response of the structure can be fed back to the EM model. The results of this cosimulation clearly show an influence of the feedback of the mechanical response which is mainly damping in the torsional degree of freedom and effects due to eccentricity in radial direction. The accuracy of these results will be validated via test bench measurements and presented in future work. Furthermore it is suggested that the EM model should be adjusted in future works so that transient effects are represented.

A new canopy model for the Reynolds-averaged Navier-Stokes (RaNS) method with the k—ɛ turbulence model was developed. To derive the effect of vegetation on the transport of turbulent quantities, it uses a Taylor series expansion of the velocity magnitude, which is part of the definition of the canopy drag force, and assumes that the turbulent kinetic energy is much smaller than the kinetic energy of the mean flow. The resultant model is composed by a sum of velocity moments of increasing order. Initially, it was expected that truncating the sum at low-order, including only terms that can be expressed using quantities available within the k—ɛ RaNS model, would provide better accuracy than traditional models, based mainly on dimensional arguments. However, the results obtained with this approach mimic those obtained with a model based on dimensional arguments and calibrated using results of large-eddy simulations, proving the validity of both approaches and showing that the accuracy in the modelling of the flows over vegetation is limited by the k—ɛ model itself and not by the modelling of vegetation effects on turbulence.

2D and 3D steady state simulations were done using the commercial CFD package Star-CCM+ with three different RANS turbulence models. Lift and drag coefficients were simulated at different angles of attack for the NREL S826 airfoil at a Reynolds number of 100 000, and compared to experimental data obtained at NTNU and at DTU. The Spalart-Allmaras and the Realizable k-epsilon turbulence models reproduced experimental results for lift well in the 2D simulations. The 3D simulations with the Realizable two-layer k-epsilon model predicted essentially the same lift coefficients as the 2D Spalart-Allmaras simulations. A comparison between 2D and 3D simulations with the Realizable k-epsilon model showed a significantly lower prediction in drag by the 2D simulations. From the conducted 3D simulations surface pressure predictions along the wing span were presented, along with volumetric renderings of vorticity. Both showed a high degree of span wise flow variation when going into the stall region, and predicted a flow field resembling that of stall cells for angles of attack above peak lift.

Vortex-induced integral loading fluctuations on a single suspended blade at various inflow angles were modeled in the presents work by means of stochastic modelling methods. The reference time series were obtained by 3D DES CFD computations carried out on the DTU 10MW reference wind turbine blade. In the reference time series, the flapwise force component, Fx, showed both higher absolute values and variation than the chordwise force component, Fz, for every inflow angle considered. For this reason, the present paper focused on modelling of the Fx and not the Fz whereas Fz would be modelled using exactly the same procedure. The reference time series were significantly different, depending on the inflow angle. This made the modelling of all the time series with a single and relatively simple engineering model challenging. In order to find model parameters, optimizations were carried out, based on the root-mean-square error between the Single-Sided Amplitude Spectra of the reference and modelled time series. In order to model well defined frequency peaks present at certain inflow angles, optimized sine functions were superposed on the stochastically modelled time series. The results showed that the modelling accuracy varied depending on the inflow angle. None the less, the modelled and reference time series showed a satisfactory general agreement in terms of their visual and frequency characteristics. This indicated that the proposed method is suitable to model loading fluctuations on suspended blades.

Wind turbine blades have been becoming longer and more slender during the last few decades. The longer lever arm results in higher stresses at the blade root. Hence, the unsteady loads induced by turbulence, gust, or wind shear increase. One promising way to control these loads is to use flexible trailing edges near the blade tip. The unsteady effects which appear during the motion of a flexible trailing edge must be considered for the load calculation during the design process because of their high influence on aeroelastic effects and hence on the fatigue loads. This is not yet possible in most of the wind turbine simulation environments. Consequently, an empirical model is developed in the present study which accounts for unsteady effects during the motion of the trailing edge. The model is based on Fourier analyses of results generated with Reynolds-Averaged Navier-Stokes (RANS) simulations of a typical thin airfoil with a deformable trailing edge. The validation showed that the model fits Computational Fluid Dynamics (CFD) results simulated with a random time series of the deflection angle.