Conventional Deep Learning (DL) methods for bearing health indicator (HI) adopt supervised approaches, requiring expert knowledge of the component degradation trend. Since bearings experience various failure modes, assuming a particular degradation trend for HI is suboptimal. Unsupervised DL methods are scarce in this domain. They generally maximise the HI monotonicity built in the middle layer of an Autoencoder (AE) trained to reconstruct the run-to-failure signals. The backpropagation (BP) training algorithm is unable to perform this maximisation since the monotonicity of HI subsections corresponding to input sample batches does not guarantee the monotonicity of the whole HI. Therefore, existing methods achieve this by searching AE hyperparameters so that its BP training to minimise the reconstruction error also leads to a highly monotonic HI in its middle layer. This is done using expensive search algorithms where the AE is trained numerous times using various hyperparameter settings, rendering them impractical for large datasets. To address this limitation, a small Convolutional Autoencoder (CAE) architecture and a hybrid training algorithm combining Particle Swarm Optimisation and BP are proposed in this work to enable simultaneous maximisation of the HI monotonicity and minimisation of the reconstruction error. As a result, the HI is built by training the CAE only once. The results from three case studies demonstrate this method’s lower computational burden compared to other unsupervised DL methods. Furthermore, the CAE-based HIs outperform the indicators built by equivalent and significantly larger models trained with a BP-based supervised approach, leading to 85% lower remaining useful life prediction errors.@en

Wind turbine blades carry the risk of impact damage during transportation, installation, and operation. Such impacts can cause levels of damage that can propagate throughout the structure compromising performance and safety. In this study, the effect of impact damage on fatigue damage propagation in test specimens representative of a spar cap-shear web adhesively-bonded connection of a wind turbine blade was investigated. In addition, the effectiveness of using acoustic emissions to detect early impact-induced fatigue damage was studied. Three impact tests with increasing levels of energy were investigated. The results showed that for an impact test with an average energy of 16.32 J, the fatigue damage accumulation process was not influenced by the size and location of the impact damage. But for impact tests with an average energy of 23.68 J and 32.13 J, greater crack density and accelerated de-lamination and de-bonding of the adhesive from the laminate could be seen in the impact zone. Acoustic emission was shown to identify the position of the damage zone for the higher energy impact tests. It was also effective in showing the progressive accumulation of fatigue damage in this zone during the fatigue test.

@en

State-of-the-art Deep Learning (DL) methods based on Supervisory Control and Data Acquisition (SCADA) system data for the detection and prognosis of wind turbine faults require large amounts of failure data for successful training and generalisation, which are generally not available. This limitation prevents benefiting from the superior performance of these methods, especially in SCADA-based failure prognosis. Data augmentation approaches have been proposed in the literature for generating failure data instances within a SCADA sequence to reduce the imbalance between healthy and faulty state data points, which is relevant to fault detection tasks. However, the successful implementation of DL-based failure prognosis methods requires the availability of multiple run-to-failure SCADA sequences. This paper proposes a data-driven method for generating synthetic run-to-failure SCADA sequences with custom operational and environmental conditions and progression of degradation. An Artificial Neural Network (ANN) is trained with signals that represent these factors to reconstruct the SCADA signals. Then, it is used to generate synthetic SCADA datasets based on data available from a wind turbine that experienced a gearbox failure. Synthetic data sets generated are evaluated on the basis of the similarity of their signal distributions, the temporal dynamics within each signal, and the temporal dynamics among different SCADA signals with those in similar field datasets. The results show that the generated synthetic datasets are consistent with their field counterparts, with a comparatively lower diversity in their dynamic behaviour in time.

@en

Vertical-axis wind turbines (VAWTs) are considered promising solutions for urban wind energy generation due to their design, low maintenance costs, and reduced noise and visual impact compared to horizontal-axis wind turbines (HAWTs). However, deploying these turbines close to densely populated urban areas often triggers considerable local opposition to wind energy projects. Among the primary concerns raised by communities is the issue of noise emissions. Noise annoyance should be considered in the design and decision-making process to foster the social acceptance of VAWTs in urban environments. At the same time, maximising the operational efficiency of VAWTs in terms of power generation and actuation effort is equally important. This paper balances noise and aero-servo-elastic performance by formulating and solving a multi-objective optimisation problem from a controller calibration perspective. Psychoacoustic annoyance is taken as a novel indicator for the noise objective by providing a more reliable estimate of the human perception of wind turbine noise than conventional sound metrics. The computation of the psychoacoustic annoyance metric is made feasible by integrating it with an accurate and computationally efficient low-fidelity noise prediction model. For optimisation, an advanced partial-load control scheme – often used in industrial turbines – is considered, with the Kω2 controller as a baseline for comparison. Optimal solutions balancing the defined objectives are identified using a multi-criteria decision-making method (MCDM) and are subsequently assessed using a frequency-domain controller analysis framework and mid-fidelity time-domain aero-servo-elastic simulations. The MCDM results indicate the potential application of this controller in small-scale urban VAWTs to attain power gains of up to 39 % on one side and to trade off a reduction in actuation effort of up to 25 % at the cost of only a 2 % power decrease and a 6 % increase in psychoacoustic annoyance on the other side compared to the baseline. These findings confirm the flexible structure of the optimally calibrated wind speed estimator and tip-speed ratio (WSE–TSR) tracking controller, effectively balancing aero-servo-elastic performance with noise emissions and marking the first instance of integrating residential concerns into the decision-making process.@en

For scenarios of high penetration of renewable energy, it becomes increasingly relevant to improve the dispatchability of supply for wind and solar power plants. Baseload power plants, required to produce a minimum power production at all times, are discussed in this context. The baseload constraint can be satisfied with renewable sources when combined with a storage system but at a high cost. This work studies the design drivers of such a storage system when consisting of short and long-term storage. The capacities of the short-term and long-term storage components are calculated as part of a linear optimization problem with the objective of minimizing the cost of baseload, using a metric based on a net present value formulation. Our analysis, based on 10 locations in Northern Europe, highlights a high sensitivity of optimal storage sizing to storage cost assumptions. In addition, the cost of baseload is found to be correlated to the share of renewable power produced above baseload, but not to the correlation between price and wind power, suggesting arbitrage plays a minor role in the business case.@en

In this paper, the feasibility of using far-field acoustic measurements as a non-contact monitoring technique for wind turbine blade leading edge erosion is assessed. For this purpose, a DU96 W180 airfoil with several eroded leading edge configurations of different severities is experimentally investigated. The eroded leading edges are designed with pits, gouges and coating delamination scaled from a real eroded blade. To assess the feasibility of the technique in quasi-realistic configurations, experiments are carried out under clean and turbulent inflow conditions. Acoustic measurements are performed with a phased microphone array. In the absence of inflow turbulence, because of the low Reynolds number at which the experiments are carried out, the case with minor erosion severity shows similar far-field noise spectra as the clean leading-edge cases, i.e., the presence of tonal peaks caused by laminar boundary layer instability noise through a self-sustained feedback loop but with higher tonal amplitudes. Increasing the damage level (considered as moderate erosion), the spectra of the noise scattered from the suction side show that the tonal peaks shift to higher frequencies and have lower amplitudes, thus suggesting that the damage alters the flow features responsible for the acoustic feedback loop; whereas, the spectra from the pressure side show a broadband noise distribution. For heavy erosion, the far-field noise spectra show broadband features from both airfoil sides, thus suggesting that the damage has fully forced the transition to turbulent flow; in which case, an increase in the low-frequency content is observed. Conversely, in the presence of turbulent inflow, when comparing the noise scattered at the trailing edge, no difference is found. However, leading edge impingement noise decreases at medium–high frequency compared with the baseline case at a chord-length-based Strouhal number St_C~10. The experimental results also suggest that the delamination feature is the one which is the most easily detectable and the approach is valid for a wide range of angles of attack and inflow velocity.@en

In this paper, a new indicator to localize fatigue damage in a fibre glass composite structure, i.e. spar cap to shear web thick adhesive joint of a wind turbine blade, is presented. This indicator is based on the effect of damping on the phase of the mode shapes of the structure. When fatigue damage occurs, damping increases in the defective area and this leads to an increase in the local energy dissipation. This non-uniformity in the energy dissipation throughout the structure causes the structure to vibrate with mode shapes whose structural elements no longer have the same phase creating complex mode shapes. A visco-elastic finite element (FE) vibration model is developed for a thick adhesive joint of a wind turbine blade. The mass, stiffness, and damping matrix extracted from the FE model are used to determine the complex mode shapes. The results show that the damaged area is located where the spatial derivative of the phase of the components of the mode shapes is minimum. Changes in the phase of mode shapes of the structural elements are strongly dependent on the location of damage. In the locations where the strain modal energy is greater, the change in the phase is also higher.

@en

The combined wind speed estimator and tip-speed ratio (WSE–TSR) tracking wind turbine control scheme has seen recent and increased traction from the wind industry. The modern control scheme provides a flexible trade-off between power and load objectives. On the other hand, the Kω2 controller is often used based on its simplicity and steady-state optimality and is taken as a baseline here. This paper investigates the potential benefits of the WSE–TSR tracking controller compared to the baseline by analysis through a frequency-domain framework and by optimal calibration through a systematic procedure. A multi-objective optimisation problem is formulated for calibration with the conflicting objectives of power maximisation and torque fluctuation minimisation. The optimisation problem is solved by approximating the Pareto front based on the set of optimal solutions found by an explorative search. The Pareto fronts were obtained by mid-fidelity simulations with the National Renewable Energy Laboratory (NREL) 5 MW turbine under turbulent wind conditions for calibration of the baseline and for increasing fidelities of the WSE–TSR tracking controller. Optimisation results show that the WSE–TSR tracking controller does not provide further benefits in energy capture compared to the baseline Kω2 controller. There is, however, a trade-off in torque control variance and power capture with control bandwidth. By lowering the bandwidth at the expense of generated power of 2 %, the torque actuation effort reduces by 80 % with respect to the optimal calibration corresponding to the highest control bandwidth.@en

Mesoscale meteorological phenomena, including atmospheric gravity waves (AGWs) and including 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-southwesterly 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 wind speeds 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 wind speeds 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 wind speeds 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, wind farm blockage effects, and wake recovery. Determining how much of the differences in wind speed and power output in the wind farm resulted from the TLW is an area for future development.

@en

Only a few studies on the overall impact of climate change on offshore wind power production and wind power ramps in the North Sea region have been published. This study focuses on the characteristics of expected wind power production and wind power ramps in the future climate aided by the classification of circulations patterns using a self-organizing map (SOM). A SOM is used to cluster high-resolution CMIP5-CORDEX sea level pressure data into 30 European area weather patterns. These patterns are used to better understand wind power production trends and any potential changes. An increased frequency of occurrence and extended persistence of high pressure systems lasting at least 24 h is projected in the future. Whereas a contrasting reducing tendency for low-pressure systems is estimated. No significant evidence is seen for a change in wind power capacity factor over the North Sea, though tentative evidence is seen for a reduction in wind power ramps. Annual energy production is seen to be dominated by a small number of weather patterns with westerly, south-westerly or north-westerly winds. Future wind power production is projected to become less from westerly winds and more from south-westerly and north-westerly flows. Ramp up events are primarily associated with strong south-westerly winds or weather patterns with a weak pressure gradient. Ramp down events have a stronger association with more north-westerly flow. In a future climate, a reduction in ramp up events associated with weak pressure gradients is projected.

@en

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, decIsioN-TRee-based wInd GUst Estimation) that utilizes numerous inputs from a public-domain reanalysis dataset and, in turn, generates multi-year, site-specific Wp series. Through a systematic feature importance study, we also identify the most relevant meteorological variables for Wp estimation. The INTRIGUE approach outperforms the baseline predictions for all wind gust conditions. However, the performance of this proposed approach and the baselines for extreme conditions (i.e., Wp>20 m s−1) is less satisfactory.@en

On-shore horizontal-axis wind turbines (HAWTs) provide a cost-effective solution for low carbon electricity generation. However, public acceptance is still a problem. A possible alternative to a HAWT is a vertical-axis wind turbine (VAWT), which is more visually appealing and less noisy. Furthermore, the inherent omni-directionality of VAWTs makes them suitable for installation in urban environments where the turbulence levels are high, and the wind direction variations are significant. However, the variation with the azimuth angle of the blade-effective wind speed and the angle of attack makes VAWT performance difficult to predict. This study proposes a wind speed estimator for a VAWT to address this challenge and to exploit knowledge of the blade-effective wind speed for load reduction control strategies. An Unscented Kalman Filter is used to solve the blade-effective wind speed estimation problem and is applied to a realistic 1.5 m two-bladed H-Darrieus VAWT model, for which the aerodynamic characteristics are determined using an actuator cylinder model. The system performance is evaluated using different wind speed variation scenarios. Overall, good agreement between the reference and estimated blade-effective wind speed is found both in terms of trend and absolute values.

@en

Handbook of Wind Resource Assessment provides a comprehensive description of the background theory, methods, models, applications, and analysis of the discipline of wind resource assessment, covering topics such as climate variability, measurement, wind distributions, numerical modeling, statistical modeling, reanalysis datasets, applications in different environments (onshore and offshore), wind atlases, and future climate. The text provides an up-to-date assessment of the tools available for wind resource assessment and their application in different environments. It also summarizes our present understanding of the wind climate and its variability, with a particular focus on its relevance to wind resource assessment. Written by a highly qualified professional in the fields of wind resource assessment, wind turbine condition monitoring, and wind turbine wake modeling, sample topics included in Handbook of Wind Resource Assessment are as follows: • Climate variability, covering temporal scales of variation, power spectrum, short term variation and turbulence, the spectral gap, and long-term variation • Measurement, covering history of wind speed measurement, types of measurement, terrestrial measurements, anemometers, wind vanes, lidars, sodars and remote sensing • Distributions, covering synoptic scale wind distributions, turbulent scale distributions, contrast between mean and extreme values, and extreme value statistics • Physical modeling, covering spatial scales of variability, the governing equations, models of varying complexity, mass consistent models, linearized models and semi-empirical models • Statistical modeling, covering the use of measure-correlate-predict (MCP), wind indices and spatial interpolation Handbook of Wind Resource Assessment serves as a comprehensive text that brings together the different aspects of wind resource assessment in one place. It is an essential resource for anyone who wishes to understand the underlying science, models, or applications of wind resources, including postgraduates, academics, and wind resource professionals.

@en

In recent years, with the development of the wind power industry and the increase in the number of wind turbines, the condition monitoring of blades and the detection of damage are increasingly important. In this work, a new non-contact damage-detection approach is experimentally investigated based on the measurement of airfoil aerodynamic noise. A NACA 0018 airfoil with chord of 200 mm with different trailing edge crack sizes, 0.2, 0.5, 1.0 and 2.0 mm, is investigated. Experiments are conducted at different mean flow velocities, inflow turbulence intensities and angles of attack. Far-field noise scattered from the airfoil is measured by means of a microphone array. The spectral differences of sound pressure level between the damaged cases and the baseline (without any damage) are compared. As expected, at small angles of attack, with clean or low turbulence intensities (e.g. ∼ 4% in the experiment) flow, by increasing the size of the crack, tonal noise appears at trailing-edge thickness-based Strouhal number,St_h , approximatively equal to 0.1. However, at higher angles of attack (e.g. ± 10° and ± 15°) or under conditions of high turbulence intensity (e.g. ∼ 7%), the amplitude of the tonal peak diminishes suggesting that complementary measurements or longer acquisition time to remove inflow turbulence effects are needed to monitor trailing edge cracks.@en

This paper presents the state-of-the-art technologies and development trends of wind turbine drivetrains – the system that converts kinetic energy of the wind to electrical energy – in different stages of their life cycle: design, manufacturing, installation, operation, lifetime extension, decommissioning and recycling. Offshore development and digitalization are also a focal point in this study. Drivetrain in this context includes the whole power conversion system: main bearing, shafts, gearbox, generator and power converter. The main aim of this article is to review the drivetrain technology development as well as to identify future challenges and research gaps. The main challenges in drivetrain research identified in this paper include drivetrain dynamic responses in large or floating turbines, aerodynamic and farm control effects, use of rare-earth material in generators, improving reliability through prognostics, and use of advances in digitalization. These challenges illustrate the multidisciplinary aspect of wind turbine drivetrains, which emphasizes the need for more interdisciplinary research and collaboration.
@en

Wind turbine blade spar cap to shear web adhesively bonded connections can suffer from damage at the bond-line which can propagate through the structure compromising blade integrity. This study investigates changes in the stiffness and damping of a thick adhesive joint test specimen during a fatigue test. The stiffness is calculated using an extensometer and damping is determined using vibration and thermography. Fatigue tests showed three distinct phases of damage. Firstly, transverse cracks and delamination initiate and grow with little change in stiffness observed while damping increases by 18.6% and 17.8% inferred from the vibration and thermographic methods, respectively. In the second phase, the number of transverse cracks increases reaching saturation and the stiffness reduces by 4.7%. The increase in the loss factor is 65.7% and 95.6% from the vibration and thermographic methods, respectively. In the final phase, the crack density remains constant whilst de-bonding of the joint begins and grows until failure. The loss factor increases by 111.4% and 116.9% from the vibration and thermographic methods, respectively though the stiffness shows a cumulative reduction of only 8.6%. The results show the potential for monitoring changes in damping to infer incipient damage in an adhesively bonded composite joint structure.

@en

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.@en

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.@en