Assessing a wind farm’s annual energy production (AEP) involves modelling the wind resource and the wind-to-power conversion at the site. The greenfield pre-construction phase generally comprises the installation of wind measurement devices. For repowering projects, the wind data from the pre-construction phase of the existing farm can be used as wind input to assess the energy yield of the repowered wind farm. Indeed, one study demonstrates that when the modelling error correlations are known, the AEP prediction uncertainty of the repowered farm can be reduced by combining the energy production records of the existing farm with the AEP assessment for both farms. Previous studies have successfully identified the correlation structure for certain errors, especially for horizontal flow modelling, but not for vertical flow modelling. However, vertical extrapolation is essential, as the wind measurement heights are generally lower than the hub height on the repowered farm. This paper bridges this research gap and demonstrates that the correlation structure of errors in vertical profile modelling is Gaussian, with parameters dependent on shear values and heights. The distribution is validated against site data from simple to moderately complex sites in France.

This paper addresses the challenge of incorporating electricity prices into wind turbine design methods and shows how price volatility drives wind turbines towards larger rotors and lower specific power. Since wind speed and electricity prices fluctuate, current efforts to estimate a wind turbine’s revenue are based on time-series approaches. However, this paper presents a new way of accounting for price volatility based on wind distributions, which is computationally cheap and easily integrates with current wind turbine and farm design methods and tools. The new method demonstrates that a traditional wind turbine can lose more than 15% of its revenue in open energy markets like Denmark due to price volatility. Designing turbines with lower specific power can substantially increase revenue by producing more energy at low wind speeds with higher energy demand and electricity prices.

The repowering of wind farms is a rapidly developing area of research and is expected to represent 40 GW of wind projects by 2030 in the European Union. This has led to the emergence of energy yield assessment methods that incorporate operational data from existing farms with the aim of surpassing traditional methods that rely solely on physical modelling and onsite measurements. The current literature on repowering relies upon the assumption that learning from operational farm data applies to the future farm. Indeed, calibration and adjustment methods assume that physics-driven models (PDMs) have spatially and temporally correlated errors. This study investigates this assumption by analysing PDM errors for 25 pairs of nearby wind projects. A statistically significant correlation is observed. We discuss whether it is reasonable to utilise operational data from existing farms, which possess different characteristics, to improve the long-term production prediction of a repowered farm.

"Green" hydrogen produced using electrolysis and zero-carbon electricity may be a path for the EU to reduce its dependence on natural gas imports and support the green transition. The EU uses 10 Mt of natural gas yearly for hydrogen production, representing a potential 120 GW market for the wind industry. Placing the hydrogen production (electrolysers) near the energy source (wind turbines) instead of the consumer (e.g. fertiliser producers) changes the energy transportation from power to hydrogen and opens new design possibilities. This work investigates how an off-grid wind turbine designed for hydrogen production may differ from a traditional one. The paper first describes the numerical modelling framework used to estimate the hydrogen production and cost of a "hydrogen wind turbine." The main parameters of the system (generator rating, rotor diameter, electrolyser rating and battery capacity) are then optimised to reduce the levelised cost of hydrogen (LCOH). Results show that an LCOH-optimised wind turbine should have a lower specific power (i.e. a larger rotor) than an LCOE-optimised design. We find a 7% LCOH reduction compared to the reference system, opening the door for future wind turbines designed directly for hydrogen production.

We present the evolution of the methods used to create and validate the various numerical wind atlases during the past ten years of the Wind Atlas for South Africa (WASA) project. In WASA 3, we improved on the previous numerical wind atlases by: * Creating an ensemble of 2-year simulations to find the optimal set of parameterisations and surface conditions for the wind climate of South Africa. * Using a new method of generalisation and downscaling of the WRF-derived wind climate using the PyWAsP engine. * Producing the most extensive to date wind climatology for South Africa, 30 years (1990–2019) simulation covering all South Africa at 3.33 km × 3.33 km spatial resolution and 30 minutes time output. We will discuss these three areas and their improvements to the wind atlas' quality. The WASA 3 wind atlas' final error statistics show that the new WRF + PyWAsP method has a MAPE of 11.8% and 3.5% for the long-term mean power density and mean wind speed, respectively. These statistics are improved from those in WASA 1 and WASA 2. When disregarding the two masts (WM09 and WM11) located in highly complex terrain, where the methodology was never designed, the use of the WRF and WRF + PyWAsP downscaling narrows the error distributions for both long-term wind speed and power density compared to the global reanalysis, ERA5. The validated numerical wind atlas has further been used to model the wind resources of the entire land area of South Africa using the microscale WAsP model. Raster data exist with a horizontal resolution of 250 meters and three levels of 50, 100 and 150 meters a.g.l. of mean wind speed, power density, air density, Weibull A and k parameters, and ruggedness index. These data sets and the WRF dataset will be made available in the public domain at the end of the project. Data sets for other heights above the ground and offshore can easily be added later.

We present the "most similar" method for selecting optimal measurement positions for wind resource assessment. Wind resource assessment is generally done by extrapolating a measured and long-term corrected wind climate at one location to a prediction location using a micro-scale flow model. If several measurement locations are available, standard industry practice is to make a weighted average of all the possible predictions using inverse-distance weighting. The most similar method challenges this practice. Instead of weighting several predictions, the method only selects the single measurement location evaluated to be most similar. We validate the new approach by comparing against measurements from 185 met masts from 40 wind farm sites and show improvements compared to inverse-distance weighting. Compared to using the closest measurement location, the error of power density predictions is reduced by 13 % using inverse-distance weighting and 34 % using the most similar method.

By using multiple wind measurements when designing wind farms, it is possible to decrease the uncertainty of wind farm energy assessments since the extrapolation distance between measurements and wind turbine locations is reduced. A WindScanner system consisting of two synchronized scanning lidars potentially represents a cost-effective solution for multipoint measurements, especially in complex terrain. However, the system limitations and limitations imposed by the wind farm site are detrimental to the installation of scanning lidars and the number and location of the measurement points. To simplify the process of finding suitable measurement positions and associated installation locations for the Wind Scanner system, we have devised a campaign planning workflow. The workflow consists of four phases. In the first phase, based on a preliminary wind farm layout, we generate optimum measurement positions using a greedy algorithm and a measurement "representative radius". In the second phase, we create several Geographical Information System (GIS) layers such as exclusion zones, line-of-sight (LOS) blockage and lidar range constraint maps. These GIS layers are then used in the third phase to find optimum positions of the WindScanner systems with respect to the measurement positions considering the WindScanner measurement uncertainty and logistical constraints. In the fourth phase, we optimize and generate a trajectory through the measurement positions by applying the traveling salesman problem (TSP) on these positions. The described workflow has been digitalized into a Python package named campaign-planning-tool, which gives users an effective way to design measurement campaigns with WindScanner systems. In this study, the Python package has been tested on three different sites characterized by different terrain complexity and wind farm dimensions and layouts. With minimal effort, the Python package can optimize measurement positions and suggest possible lidar installation locations for carrying out resource assessment campaigns.

Abstract. Multiple wind measurements is a way to reduce the uncertainty of wind farm energy yield assessments by reducing the extrapolation distance between measurements and wind turbines locations. A WindScanner system consisting of two synchronized scanning lidar potentially represents a cost-effective solution for multi-point measurements, especially in complex terrain. However, the system limitations and limitations imposed by the wind farm site are detrimental to the installation of scanning lidars and the number and location of the measurement positions. To simplify the process of finding suitable measurement positions and associated installation locations for the WindScanner system we have devised a campaign planning workflow. The workflow consists of four phases. In the first phase, based on a preliminary wind farm layout, we generate optimum measurement positions using a greedy algorithm and a measurement 'representative radius'. In the second phase, we create several Geographical Information System (GIS) layers of information such as exclusion zones, line-of-sight (LOS) blockage, and lidar range maps. These GIS layers are then used in the third phase to find optimum positions of the WindScanners with respect to the measurement positions considering the WindScanner measurement uncertainty. In the fourth phase, we optimize and generate trajectory through the measurement positions by applying the traveling salesman problem (TSP) on these positions. The above-described workflow has been digitized into the so-called Campaign Planning Tool (CPT) currently provided as a Python library which allows users an effective way to plan measurement campaigns with WindScanner systems. In this study, the CPT has been tested on three different sites characterized by different terrain complexity and wind farm dimensions and layouts. The CPT has shown instantly whether the whole site can be covered by one system or not.

The Bolund experiment is a study of the flow over Bolund, a small hill located near Risø DTU. In the present work, the flow over the Bolund hill has been computed using the EllipSys3D Navier-Stokes solver. The aim of this work is to identify the flow conditions at the hill in order to prepare the upcoming measurement campaign. The simulation results are used for estimating rational positions of measuring masts and instrumentation. A total of ten measuring masts are proposed consisting of two 16 m masts, six 9 m masts and two 5 m masts. The masts are instrumented with 12 cup anemometers for mean wind measurements and 21 sonics for 3-component turbulence measurements. The present report describes the simulation methodology and results of initial CFD computations of the wind conditions over the Bolund hill. The simulations have been conducted before an upcoming measuring campaign in order to estimate the best possible positions for measuring equipment. The work is conducted under the EFP project:

We present a preliminary evaluation of shelter models of different degrees of complexity using full-scale lidar measurements of the shelter on a vertical plane behind and orthogonal to a fence. Model results accounting for the distribution of the relative wind direction within the observed direction interval are in better agreement with the observations than those that correspond to the simulation at the center of the direction interval, particularly in the far-wake region, for six vertical levels up to two fence heights. Generally, the CFD results are in better agreement with the observations than those from two engineering-like obstacle models but the latter two follow well the behavior of the observations in the far-wake region.

The flow solvers OpenFOAM and EllipSys3D are compared in the case of neutral atmospheric flow over terrain using the test cases of Askervein and Bolund hills. Both solvers are run using the steady-state Reynolds-averaged Navier–Stokes k–ϵ turbulence model. One of the main modeling differences between the two solvers is the wall-function approach. The OpenFOAM v.1.7.1 uses a Nikuradse's sand roughness model, while EllipSys3D uses a model based on the atmospheric roughness length. It is found that Nikuradse's model introduces an error dependent on the near-wall cell height. To mitigate this error the near-wall cells should be at least 10 times larger than the surface roughness. It is nonetheless possible to obtain very similar results between EllipSys3D and OpenFOAM v.1.7.1. The more recent OpenFOAM v.2.2.1, which includes the atmospheric roughness length wall-function approach, has also been tested and compared to the results of OpenFOAM v.1.7.1 and EllipSys3D. The numerical results obtained using the same wall-modeling approach in both EllipSys3D and OpenFOAM v.2.1.1 proved to be almost identical. Two meshing strategies are investigated using HypGrid and SnappyHexMesh. The performance of OpenFOAM on SnappyHexMesh-based low-aspect-ratio unstructured meshes is found to be almost an order of magnitude faster than on HypGrid-based structured and high-aspect-ratio meshes. However, proper control of boundary layer resolution is found to be very difficult when the SnappyHexMesh tool is utilized for grid generation purposes. The OpenFOAM is generally found to be 2–6 times slower than EllipSys3D in achieving numerical results of the same order of accuracy on similar or identical computational meshes, when utilization of EllipSys3D default grid sequencing procedures is included.

The validation of the accuracy was performed by comparing with estimates by the traditional Annual Maxima (AM) method and the Peak Over Threshold (POT) method, applied to measurements, for six sites: four sites located in Denmark, one site located in the Netherlands and one site located in the USA, comprising both on-shore and off-shore sites. The SC method was applied to 1-year measured wind data while the AM and POT methods were applied to long-term measured wind data. Further, the consistency of the SC method was validated from the variance of the extreme wind prediction when different years are used as the period of the short-term measured wind data.

The interaction of a wind turbine with the upstream flow-field in complex and flat terrain is studied using Reynolds-averaged Navier-Stokes (RANS) simulations with a two equation turbulence closure. The complex site modelled is Perdigao (Portugal), where a turbine is located on one of two parallel running ridges. Simulating various wind directions with and without rotor, the impact of the rotor on the flow-field upstream is determined. This is compared and related to simulations with sheared and uniform inflow. The induction zones forming for these two inflows agree to such an extent, that shear could be interpreted as linear perturbation to the uniform inflow solution. However, for complex terrain this is not the case, as it is highly dependant on flow features caused by the topography and their interaction with the rotor. Separation in the lee of the ridge plays a crucial role, as it dictates the wind turbine wake trajectory which in turn governs the orientation of the induction zone.

Most of our knowledge on forest-edge flows comes from numerical and wind-tunnel experiments where canopies are horizontally homogeneous. To investigate the impact of tree-scale heterogeneities (\({>}1\) m) on the edge-flow dynamics, the flow in an inhomogeneous forest edge on Falster island in Denmark is investigated using large-eddy simulation. The three-dimensional forest structure is prescribed in the model using high resolution helicopter-based lidar scans. After evaluating the simulation against wind measurements upwind and downwind of the forest leading edge, the flow dynamics are compared between the scanned forest and an equivalent homogeneous forest. The simulations reveal that forest inhomogeneities facilitate flow penetration into the canopy from the edge, inducing important dispersive fluxes in the edge region as a consequence of the flow spatial variability. Further downstream from the edge, the forest inhomogeneities accentuate the canopy-top turbulence and the skewness of the wind-velocity components while the momentum flux remains unchanged. This leads to a lower efficiency in the turbulent transport of momentum within the canopy. Dispersive fluxes are only significant in the upper canopy. Above the canopy, the mean flow is less affected by the forest inhomogeneities. The inhomogeneities induce an increase in the mean wind speed that was found to be equivalent to a decrease in the aerodynamic height of the canopy. Overall, these results highlight the importance of forest inhomogeneities when looking at canopy–atmosphere exchanges in forest-edge regions.