Search

Peak wind gust (Wp) is a crucial meteorological variable for wind farm planning and operations. However, for many wind farm sites, there is a dearth of on-site measurements of (Wp). In this paper, we propose a machine-learning approach (called INTRIGUE) that utilizes numerous inputs from a public-domain reanalysis dataset, and in turn, generates long-term, site-specific (Wp) series. Through a systematic feature importance study, we also identify the most relevant meteorological variables for (Wp) estimation. Even though the proposed INTRIGUE approach performs very well for nominal conditions compared to specific baselines, its performance for extreme conditions is less than satisfactory.

Mesoscale meteorological phenomena, including Atmospheric Gravity Waves, or Trapped Lee Waves (TLWs) can result from flow over topography or coastal transition in the presence of stable atmospheric stratification, particularly with strong capping inversions. Satellite images show that topographically forced TLWs frequently occur around near-coastal offshore wind farms. Yet current understanding of how they interact with individual turbines and whole farm energy output is limited. This parametric study investigates the potential impact of TLWs on a UK near-coastal offshore wind farm, Westermost Rough (WMR) resulting from westerly – south-westerly flow over topography in the Southeast of England. Computational fluid dynamics (CFD) modelling (using ANSYS-CFX) of TLW situations based on real atmospheric conditions at WMR was used to better understand turbine level and whole wind farm performance in this parametric study based on real inflow conditions. These simulations indicated that TLWs have the potential to significantly alter the windspeeds experienced by and the resultant power output of individual turbines and the whole wind farm. The location of the wind farm in the TLW wave cycle was an important factor in determining the magnitude of TLW impacts, given the expected wavelength of the TLW. Where the TLW trough was coincident with the wind farm, the turbine windspeeds and power outputs were more substantially reduced compared with when the TLW peak was coincident with the location of the wind farm. These reductions were mediated by turbine windspeeds and wake losses being superimposed on the TLW. However, the same initial flow conditions interacting with topography under different atmospheric stability settings produce differing near wind farm flow. Factors influencing the flow within the wind farm under the different stability conditions include differing: hill and coastal transition recovery, windfarm blockage effects and wake recovery. Determining how much of the differences in windspeed and power output in the wind farm resulted from the TLW is an area for future development.

Anthropogenic disturbance has resulted in a global reduction in the abundance of mature, hollow‐bearing trees. Nest boxes have long been used to provide supplementary shelter sites in revegetated and regenerating landscapes, but limitations in their effectiveness when offsetting the loss of mature trees has led to increased interest in novel designs of artificial hollows. For example, mechanically excavating cavities into the trunk or branches of trees. However, the effectiveness of artificial hollows in attracting wildlife to visit small‐ or medium‐sized, growing trees in human‐disturbed landscapes has received little attention. In this study, we installed chainsaw hollows that were designed for small, hollow‐dependent mammals and birds into the trunks of live medium‐sized trees. We conducted a before‐after control‐impact experiment using passive camera traps to monitor changes in visitations by wildlife to (1) mature hollow‐bearing trees, (2) developing trees without hollows (i.e. control trees), and (3) developing trees with newly installed chainsaw hollows. We found that, compared to large hollow‐bearing trees and control trees, the developing trees that were selected for chainsaw hollow construction showed the greatest visitation rates by hollow‐dependent wildlife (i.e. number of visits) during the “post‐impact” surveys. Our results suggest that chainsaw hollows designed to replicate the external physical characteristics of natural tree hollows could be effective in attracting target hollow‐dependent fauna to developing trees in regenerating and revegetated landscapes. Further studies are required to compare the effectiveness of natural hollows, chainsaw hollows, and nest boxes when deployed in a range of human‐disturbed landscapes.

This paper presents a novel prognostic method to estimate the remaining useful life (RUL) of generators using the SCADA (Supervisory Control And Data Acquisition) systems installed in wind turbines. A data-driven wind turbine anomaly classification method is developed. The anomalies are quantified into a health indicator to measure the component degradation over time. An Autoregressive Integrated Moving Average (ARIMA) time series forecasting technique is then applied to predict the RUL of the wind turbine generator. The proposed method has been validated using industry field data showing accurate predictions of RUL with a 21 day lead time for maintenance of the turbine.

Condition-based maintenance using routinely collected Supervisory Control and Data Acquisition (SCADA) data is a promising strategy to reduce downtime and costs associated with wind farm operations and maintenance. New approaches are continuously being developed to improve the condition monitoring for wind turbines. Development of normal behaviour models is a popular approach in studies using SCADA data. This paper first presents a data-driven framework to apply normal behaviour models using an artificial neural network approach for wind turbine gearbox prognostics. A one-class support vector machine classifier, combining different error parameters, is used to analyse the normal behaviour model error to develop a robust threshold to distinguish anomalous wind turbine operation. A detailed sensitivity study is then conducted to evaluate the potential of using high-frequency SCADA data for wind turbine gearbox prognostics. The results based on operational data from one wind turbine show that, compared to the conventionally used 10-min averaged SCADA data, the use of high-frequency data is valuable as it leads to improved prognostic predictions. High-frequency data provides more insights into the dynamics of the condition of the wind turbine components and can aid in earlier detection of faults.

Wind energy continues to make great strides in its contribution to the net CO2-zero targets of countries around the world. Both wind farms onshore and offshore are being built, but scientific challenges lie ahead if their generation is to be both reliable and economic. These challenges relate to better understanding the characteristics of the wind, how the wind inflow translates to loads and performance, and how best to build and operate the wind farms of the future so that their output can best be integrated into a 21st century energy system. Europe continues to be a hub for scientific research in wind energy and the European Academy of Wind Energy (EAWE) was created to bring together the top research institutes active in wind energy to cooperate, share knowledge and promote scientific excellence. As part of this remit, the conference, The Science of Making Torque from Wind (or TORQUE, for short) was inaugurated in 2004 in the beautiful city of Delft. The conference has gone from strength to strength and is probably the largest scientific conference devoted to wind energy in the world. History came full circle and the eighth edition, TORQUE 2020 was due to be held in Delft. Unfortunately, the global pandemic meant that this conference had to be held online, but the TU Delft Wind Energy Institute was given a further opportunity to host the ninth edition, TORQUE 2022, in person. Following the call for three-page abstracts, 435 submissions were made and, after a two-stage peer review process by over 100 reviewers, 300 full papers were accepted for publication in the proceedings. The conference consisted of three plenary sessions, 28 parallel oral sessions and two poster sessions. Oral presenters had 15 minutes to present their work followed by 5 minutes of questions. All poster presenters were allowed a recorded one-minute pitch which could be accessed in advance of the live 90-minute poster sessions.

Multiple-species (floristically diverse) plantings of trees and shrubs in former agricultural landscapes to offset (sequester) emissions of carbon dioxide are a recent component of Australian lands...

Prescribed (or “planned”) burning is used by land managers to reduce fuel-loads in order to mitigate the spread of wildfire, thereby protecting life and property, and to promote environmental heterogeneity to enhance biodiversity. Globally, many fire management agencies focus on increasing extent and frequency of prescribed burning. There is a need to assess how high levels of prescribed burning may affect the long-term, landscape-level persistence of ecological communities. We forward projected management scenarios over 21 years to explore how the operationally realistic implementation of four different prescribed burn targets, covering 5, 3, 1.5 and 0% of a large reserve per annum (p.a.) might affect provision and removal of fire-mediated habitat of 11 rare and threatened bird species. Sustained implementation of high targets (5 and 3% p.a.) homogenised the landscape toward young vegetation, substantially reducing highly suitable habitat for species requiring intermediate (20–60 years post-fire) and older (60+ years) age classes. In contrast, no prescribed burning generated insufficient habitat for species with early (

Commercial operation and maintenance of wind farms always involves trying to find the most cost-effective solution from various possible options. In this paper, a maintenance action within a Spanish wind farm was studied, whereby a blade replacement was required to prevent catastrophic failure. The conducted replacement was accompanied by an underperformance resolved in a later blade re-pitching. We analyse the decision taken in terms of the power performance and net present value from the cash flow resulting from the energy sales. The impact of the timing of the maintenance is discussed in various what-if scenarios. The sensitivity to environmental causes of underperformance is compared by varying the duration of blade icing and comparing the performance in different wind directions. Country dynamics and subsidy impacts are hypothetically evaluated for the prevailing electricity market conditions as if the turbine were operating in either Spain, Netherlands or the UK. The findings highlight the uncertainty in power performance and the importance of maintenance accuracy. It is shown that the decision-making of operators should not only consider the seasonality of the wind resource, but also the seasonality in electricity markets.

Fire shapes ecosystems globally, including semi-arid ecosystems. In Australia, semi-arid ‘mallee’ ecosystems occur primarily across the southern part of the continent, forming an interface between the arid interior and temperate south. Mallee vegetation is characterized by short, multi-stemmed eucalypts that grow from a basal lignotuber. Fire shapes the structure and functioning of mallee ecosystems. Using the Murray Mallee region in south-eastern Australia as a case study, we examine the characteristics and role of fire, the consequences for biota, and the interaction of fire with other drivers. Wildfires in mallee ecosystems typically are large (1000s ha), burn with high severity, commonly cause top-kill of eucalypts, and create coarse-grained mosaics at a regional scale. Wildfires can occur in late spring and summer in both dry and wet years. Recovery of plant and animal communities is predictable and slow, with regeneration of eucalypts and many habitat components extending over decades. Time since the last fire strongly influences the distribution and abundance of many species and the structure of plant and animal communities. Animal species display a discrete set of generalized responses to time since fire. Systematic field studies and modeling are beginning to reveal how spatial variation in fire regimes (‘pyrodiversity’) at different scales shapes biodiversity. Pyrodiversity includes variation in the extent of post-fire habitats, the diversity of post-fire age-classes and their configuration. At regional scales, a desirable mix of fire histories for biodiversity conservation includes a combination of early, mid and late post-fire age-classes, weighted toward later seral stages that provide critical habitat for threatened species. Biodiversity is also influenced by interactions between fire and other drivers, including land clearing, rainfall, herbivory and predation. Extensive clearing for agriculture has altered the nature and impact of fire, and facilitated invasion by pest species that modify fuels, fire regimes and post-fire recovery. Given the natural and anthropogenic drivers of fire and the consequences of their interactions, we highlight opportunities for conserving mallee ecosystems. These include learning from and fostering Indigenous knowledge of fire, implementing actions that consider synergies between fire and other processes, and strategic monitoring of fire, biodiversity and other drivers to guide place-based, adaptive management under climate change.

While many operation and maintenance (O&M) decision support systems (DSS) have been already proposed, a serious research need still exists for wind farm O&M scheduling. O&M planning is a challenging task, as maintenance teams must follow specific procedures when performing their service, which requires working at height in adverse weather conditions. Here, an automated maintenance programming framework is proposed based on real case studies considering available wind speed and wind gust data. The methodology proposed consists on finding the optimal intervention time and the most effective execution order for maintenance tasks and was built on information from regular maintenance visit tasks and a corrective maintenance visit. The objective is to find possible schedules where all work orders can be performed without breaks, and to find out when to start in order to minimise revenue losses (i.e. doing maintenance when there is least wind). For the DSS, routine maintenance tasks are grouped using the findings of an agglomerative nesting analysis. Then, the task execution windows are searched within pre-planned maintenance day.

Due to their shared genetic history, antibodies from the same clonotype often bind to the same epitope. This knowledge is used in immune repertoire mining, where known binders are used to search bulk sequencing repertoires to identify new binders. However, current computational methods cannot identify epitope convergence between antibodies from different clonotypes, limiting the sequence diversity of antigen-specific antibodies that can be identified. We describe how the antibody binding site, the paratope, can be used to cluster antibodies with common antigen reactivity from different clonotypes. Our method, paratyping, uses the predicted paratope to identify these novel cross clonotype matches. We experimentally validated our predictions on a pertussis toxoid dataset. Our results show that even the simplest abstraction of the antibody binding site, using only the length of the loops involved and predicted binding residues, is sufficient to group antigen-specific antibodies and provide additional information to conventional clonotype analysis. Abbreviations: BCR: B-cell receptor; CDR: complementarity-determining region; PTx: pertussis toxoid

In the near future, wind and solar generation are projected to play an increasingly important role in Europe's energy sector. With such fast-growing renewable energy development, the presence of simultaneous calm wind and overcast conditions could cause significant shortfalls in production with potentially serious consequences for system operators. Such events are sometimes dubbed “Dunkelflaute” events and have occurred several times in recent history. The capabilities of contemporary mesoscale models to reliably simulate and/or forecast a Dunkelflaute event are not known in the literature. In this paper, a Dunkelflaute event near the coast of Belgium is simulated utilizing the Weather Research and Forecasting (WRF) model. Comprehensive validation using measured power production data and diverse sets of meteorological data (e.g., floating lidars, radiosondes, and weather stations) indicates the potential of WRF to reproduce and forecast the boundary layer evolution during the event. Extensive sensitivity experiments with respect to grid-size, wind farm parameterization, and forcing datasets provide further insights on the reliability of the WRF model in capturing the Dunkelflaute event.

Large-scale weather systems have the potential to modulate offshore wind energy production. The Northern European sea areas have recently seen a rapid increase in wind power capacity and thus there is a need to understand how different weather systems affect offshore production from the perspective of energy system integration. In this study, mean sea level pressure data from a new-generation reanalysis (ERA5) are utilised to classify synoptic systems into 30 different weather patterns using a self-organising map (SOM) approach. ERA5 wind speeds are then used in conjunction with a reference 8 MW wind turbine power curve to estimate wind power values at selected offshore sites. We assess how wind power output varies for different weather patterns, specifically, the impact on power production and power ramps.

The adhesively-bonded connections in a wind turbine blade, e.g. the spar cap to shear web joins, are key elements for the structural integrity of the blade. These joins can suffer from damage at the bond-line which can propagate through the structure and compromise the operation of the blade. In this paper, we determine the damping properties of a test specimen representative of that joining a spar cap and a shear web during a period of progressive damage. In addition to the experimental damping measurement, an analytical dynamic model based on the visco-elastic properties of the material is developed to relate the damping to the loss factor. The experimental results show that when a crack is initiated in the test specimen, the damping increases by around 5-7%. This value increases with the propagation of transverse cracks in the adhesive and reaches a value of 35% when the adhesive layer experiences crack saturation and the damping reaches 45 % before failure occurs. Although a significant change in the damping is observed, there is no significant change in the natural frequency (

Wind by its very nature is a variable element. Its variation is different on different timescales and spatially its magnitude can change dramatically depending on local climatology and terrain. This has implications in a variety of sectors, not least in the wind energy sector. The accuracy of weather forecasting models has increased significantly in the last few decades and these models are able to give an insight into variability on the hourly and daily timescales. On shorter timescales, predicting chaotic turbulent fluctuations is far more challenging. Similarly, the ability to make seasonal forecasts is extremely limited. General circulation models (GCMs) can give insights into possible future decadal fluctuations, but there are still large uncertainties. Observational data can give useful information concerning variation on a variety of timescales, but data quality and spatial coverage can be variable. An understanding of local scale spatial variations in wind is extremely important in wind farm siting. In the last 40 years, there have been significant advances in predicting these variations using computer models, although there remain significant challenges in understanding the behavior of the wind in certain environments. Both the spatial and temporal variations of wind are important considerations when wind power is integrated into electricity networks, and this will become an ever more important consideration as wind generation makes an increasing contribution to our global energy needs.

Habitat modification by fire and habitat loss via anthropogenic vegetation clearance and fragmentation both impact animal populations. Yet, there has been limited investigation as to whether animals that decline under one of these types of habitat change also decline under the other, and how their cumulative impacts affect the status of species and communities. Using a ~400-year chronosequence in the world's largest extant temperate woodland in south-western Australia, we examine how time since fire affects bird community richness, reporting rates and composition, and whether taxa grouped on the basis of responses to vegetation clearance and fragmentation in an adjoining agricultural landscape are associated with either recently-burnt or long-unburnt woodlands. Consistent with substantial changes in vegetation composition and structure after fire in obligate-seeder eucalypt woodlands, woodland bird communities were strongly affected by fire. Species richness and total reporting rates increased with time since fire, and community composition changed across the entire multi-century span of the chronosequence. Woodland birds most negatively impacted by vegetation clearance and fragmentation were strongly associated with long-unburnt woodlands. In a regional south-western Australian context, where extensive vegetation clearance has substantially reduced the range and populations of many woodland bird species, the ability of remaining unfragmented woodlands to support populations of these species will be strongly contingent on appropriate fire management. Specifically, as stand-replacement fires have affected 25–30% of extant woodland over recent decades, management to limit the extent of fire in remaining long-unburnt woodlands would appear a priority for conservation of woodland bird diversity.

Knowledge about the expected duration and intensity of wind power ramps is important when planning the integration of wind power production into an electricity network. The detection and classification of wind power ramps is not straightforward due to the large range of events that is observed and the stochastic nature of the wind. The development of an algorithm that can detect and classify wind power ramps is thus of some benefit to the wind energy community. In this study, we describe a relatively simple methodology using a wavelet transform to discriminate ramp events. We illustrate the utility of the methodology by studying distributions of ramp rates and their duration using 2 years of data from the Belgian offshore cluster. This brief study showed that there was a strong correlation between ramp rate and ramp duration, that a majority of ramp events were less than 15 h with a median duration of around 8 h, and that ramps with a duration of more than a day were rare. Also, we show how the methodology can be applied to a time series where installed capacity changes over time using Swedish onshore wind farm data. Finally, the performance of the methodology is compared with another ramp detection method and their sensitivities to parameter choice are contrasted.

The IEC standard 61400-1 edition 3 uses the so-called Frandsen model to assess levels of turbulence intensity (TI) within wind farms, specifically to determine turbine suitability associated with stress and fatigue. Since the Frandsen model was developed, turbines have significantly grown in size and the number of turbines in an individual wind farm has grown in number. It is of interest to test the accuracy of such models, especially when applied to large wind farms offshore. This work presents results from comparing the Frandsen model with measured data from the Greater Gabbard offshore wind farm. Comparisons are also made with a simplified version of the Frandsen model. In general, both models were shown to perform well when predicting values of TI. However, the ambient wind farm turbulence model utilised by the Frandsen model was shown to be less reliable than the use of an individual turbine wake-generated turbulence model, regardless of distance, as demonstrated using a simplified model. The difference between observed mean and 90th percentile (also known as representative TI) values was in general larger than that predicted. It is proposed that this is primarily due to model reliance on variance in the turbulence of the freestream flow rather than actually modelling the variance of the turbulence generated by individual turbines, although this would require further work to confirm this.

Linear strips of vegetation (e.g., hedges, roadsides) are characteristic of rural environments worldwide. Different types of linear features have distinct structure and landscape context, suggesting they each may offer unique opportunities for conservation in modified landscapes. We compared the avifauna of 76 streamside (riparian) sites and 33 sites in roadside vegetation—two distinctive types of linear features of rural landscapes in southern Australia. There was a marked difference in the composition of the avifauna between these linear features, reflecting their individual context within the landscape. For all response groups—woodland bird species, non-woodland species, waterbirds—riparian vegetation supported a greater species richness per site, and greater cumulative richness across multiple sites, than did roadside vegetation. For woodland species, the assemblage of greatest conservation concern, richness in both riparian and roadside sites increased with increasing width, and decreased with increasing abundance of an aggressive avian competitor. The ubiquity of linear features worldwide means that measures that enhance their conservation value will have widespread benefits. Our results demonstrate that: (1) linear features offer habitat for a broad range of species in rural environments; (2) by supporting distinct assemblages, different types of linear features have complementary roles in nature conservation; (3) wider linear features have a positive influence on species that require vegetated cover; and (4) the fauna of linear features are vulnerable to biotic influences, in this case a native avian competitor.

This paper presents the results of a parametric study of wind turbine wake effects in a hypothetical offshore wind farm with varying turbine separation using a Computational Fluid Dynamics (CFD) model. Results are analyzed from a simulated 40 turbine farm with 60 layout options, 4 wind speeds and 10° directional bins. Results show that increasing turbine separation in one or both directions leads to greater power generation, though this effect diminishes for separations above 8 diameters. Similarly, turbulence intensity is shown to decrease with increases in turbine separation but with little variation beyond 8 diameters. For 3 out of 4 wind speeds when combined with a representative UK offshore wind rose the farm was shown to have an optimal layout orientation along an axis 350°-170°, though the difference in power produced between orientation angles was less than between changes in turbine separation.

Aim Managing fire is critical for the conservation of biodiversity in many ecosystems globally. To manage fire effectively, it is necessary to identify the temporal and spatial scales at which it affects a diverse range of species. This information is challenging to obtain for rare and threatened species for which data often are sparse, and in systems with long fire-return intervals (e.g. >100 years). We tested the effects of a century of fires on the distribution of 12 threatened bird species across a 100,000 km2 region in which “long-unburnt” vegetation has been identified as important for the diversity of common species. Location Semi-arid mallee woodlands of south-eastern Australia. Methods We developed spatially explicit models to identify the effects of fire history and climatic factors on the distribution of 12 threatened bird species, including two globally endangered species, the Mallee Emu-wren (Stipiturus mallee) and Black-eared Miner (Manorina melanotis). Results Fire was a driver of distribution for all species. Four species were common in younger vegetation ( 60 years post-fire) vegetation. Species’ distributions were further restricted to areas associated with particular vegetation types and climatic conditions. Main conclusions Comprehensive investigation of the response to fire by a range of threatened species highlights the importance of what is now recognized as mid-successional mallee vegetation (20–60 years post-fire), and that species’ preferences for previously identified “long-unburnt” vegetation extend to ≥60 years post-fire. Fire management conducted with incomplete knowledge, or which is focussed on introducing prescribed burns or suppressing fires for early/late-successional species alone, is unlikely to maximize biodiversity. Effective fire management for biodiversity requires the promotion of ecological processes that result in key successional stages at particular locations in the landscape.

Structural and nutritional plant traits influence the ability of insect herbivores to locate, consume and persist on their hosts yet it is uncommon for ecologists to consider how multiple plant traits influence insect community composition. We sampled herbivorous insects on two understorey shrub species common to eucalypt forests of south-eastern Australia, namely Cassinia arcuata (Asteraceae) and Daviesia ulicifolia (Fabaceae). Regression analyses were used to assess the relative influence of plant structure (canopy volume), nutritional quality (macronutrients and total phenolics) and plant productivity (leaf litter) on insect abundance and species richness. Total N content of D. ulicifolia was significantly higher than C. arcuata, while the concentrations of P, K, Ca and Mg were higher in C. arcuata. Total phenolics and leaf litter were significantly lower in D. ulicifolia compared to C. arcuata. Insect composition was similar between the two shrubs but C. arcuata supported greater abundances. Canopy volume and the macronutrients P and Ca were important predictors of insect abundance on C. arcuata, whereas canopy volume alone, but neither plant productivity nor macronutrients, influenced the abundance of insects on D. ulicifolia. Ca was an important predictor of insect species richness on C. arcuata and P was an important predictor on D. ulicifolia. By quantifying a range of plant traits, we have provided an understanding of factors likely to influence the composition of herbivorous insects inhabiting these two shrubs. Traits including leaf architecture, foliar morphology and volatile terpenoids may yet explain the greater number of insects on C. arcuata since they influence the availability of microhabitats and apparency of host plants.

A series of experiments have been conducted in a stratifiable Atmospheric Boundary Layer (ABL) wind tunnel, using neutral and stable conditions, in which a forest canopy has been represented by use of architectural model trees. These experiments have been replicated in Computation Fluid Dynamic (CFD) simulations using a previously validated methodology. Both the numerical simulations and the experimental data show that atmospheric stability has a significant effect on the development and extent of the forest wake and on the prevalence of the canopy flow features such as the sub-canopy jet. The analysis shows that it is possible to include both forestry and buoyancy effects in numerical simulations using two sets of source and sink terms and achieve satisfactory convergence. However, it is shown that the numerical simulations overestimate the effects of thermal stratification when using the standard configuration.

Airborne wind energy (AWE) systems harness energy at heights beyond the reach of tower-based wind turbines. To estimate the annual energy production (AEP), measured or modelled wind speed statistics close to the ground are commonly extrapolated to higher altitudes, introducing substantial uncertainties. This study proposes a clustering procedure for obtaining wind statistics for an extended height range from modelled datasets that include the variation in the wind speed and direction with height. K-means clustering is used to identify a set of wind profile shapes that characterise the wind resource. The methodology is demonstrated using the Dutch Offshore Wind Atlas for the locations of the met masts IJmuiden and Cabauw, 85 km off the Dutch coast in the North Sea and in the centre of the Netherlands, respectively. The cluster-mean wind profile shapes and the corresponding temporal cycles, wind properties, and atmospheric stability are in good agreement with the literature. Finally, it is demonstrated how a set of wind profile shapes is used to estimate the AEP of a small-scale pumping AWE system located at Cabauw, which requires the derivation of a separate power curve for each wind profile shape. Studying the relationship between the estimated AEP and the number of site-specific clusters used for the calculation shows that the difference in AEP relative to the converged value is less than 3 % for four or more clusters.

A series of probabilistic models were bench-marked during the European Energy Markets forecasting Competition 2020 to assess their relative accuracy in predicting aggregated Swedish wind power generation using as input historic weather forecasts from a numerical weather prediction model. In this paper, we report the results of one of these models which uses a deep learning approach integrating two architectures: (a) Convolutional Neural Network (CNN) LeNet-5 based architectrure; (b) Multi-Layer Perceptron (MLP) architecture -with two hidden layers-. These are concatenated into the Smooth Pinball Neural Network (SPNN) framework for quantile regression. Hyperparameters were optimised to produce the best model for every region. When tuned, the re-forecasts from the model performed favorably compared to other machine learning approaches and showed significant improvement on the original competition results, though failed to fully capture spatial patterns in certain cases when compared to other methods.

A hybrid neural network model, comprising of a convolutional neural network and a multilayer perceptron network, has been developed for day-ahead forecasting of regional scale wind power production. This model requires operational weather forecasts as input and also has the capability to ingest data from ensemble forecasts. Even though the training of the model requires significant computational cost, the actual forecasting can be done within a few minutes on any recent personal computer. The proposed model has demonstrated noteworthy performance at a recent international forecasting competition.

In the coming decades, both wind and solar power production will be playing increasingly important roles in Europe’s energy economy. It is absolutely essential that power grids are resilient against any unusual weather phenomena. One such meteorological phenomenon, “Dunkelflaute”, is causing serious concern to the renewable energy industry, which is primarily characterized by calm winds and overcast conditions. For example, a Dunkelflaute event happened in the Netherlands on 30th April 2018 leading to a significant shortfall in renewable energy generation requiring emergency intervention by the system operator. By analyzing this case, this paper investigates the performance of a state-of-the-art mesoscale model, Weather Research and Forecasting (WRF), in forecasting a Dunkelflaute event. Multiple WRF simulations are driven using real-time Global Forecast System (GFS) operational data over a range of prediction horizons. For comparison, a benchmark run is carried out using ERA5 reanalysis data as boundary conditions. Through validation using a variety of measured data covering onshore and offshore areas, wind speed is shown to be more predictable than cloud-cover in this particular case study.