Articles | Volume 6, issue 2
Research article 21 Apr 2021
Research article | 21 Apr 2021
Feature selection techniques for modelling tower fatigue loads of a wind turbine with neural networks
Artur Movsessian et al.
Related subject area
Design methods, reliability and uncertainty modellingAeroelastic loads on a 10 MW turbine exposed to extreme events selected from a year-long large-eddy simulation over the North SeaOptimal scheduling of the next preventive maintenance activity for a wind farmA method for preliminary rotor design – Part 1: Radially Independent Actuator Disc modelA method for preliminary rotor design – Part 2: Wind turbine Optimization with Radial IndependenceWind farm layout optimization using pseudo-gradientsOn the scaling of wind turbine rotorsReducing cost uncertainty in the drivetrain design decision with a focus on the operational phaseWind tunnel comparison of four VAWT configurations to test load-limiting concept and CFD validationRedesign of an upwind rotor for a downwind configuration: design changes and cost evaluationFatigue lifetime calculation of wind turbine blade bearings considering blade-dependent load distributionReliability analysis of offshore wind turbine foundations under lateral cyclic loadingOperational-based annual energy production uncertainty: are its components actually uncorrelated?Change-point detection in wind turbine SCADA data for robust condition monitoring with normal behaviour modelsAugmented Kalman filter with a reduced mechanical model to estimate tower loads on a land-based wind turbine: a step towards digital-twin simulationsA surrogate model approach for associating wind farm load variations with turbine failuresNew strategies for optimized structural monitoring of wind farms: experimental campaignDifferences in damping of edgewise whirl modes operating an upwind turbine in a downwind configurationAssessment of a rotor blade extension retrofit as a supplement to the lifetime extension of wind turbinesIs the Blade Element Momentum theory overestimating wind turbine loads? – An aeroelastic comparison between OpenFAST's AeroDyn and QBlade's Lifting-Line Free Vortex Wake methodDevelopment and feasibility study of segment blade test methodologyAnalytical model for the power–yaw sensitivity of wind turbines operating in full wakeWake steering optimization under uncertaintyRadar-derived precipitation climatology for wind turbine blade leading edge erosionWESgraph: a graph database for the wind farm domainReliability-based design optimization of offshore wind turbine support structures using analytical sensitivities and factorized uncertainty modelingOptimal relationship between power and design-driving loads for wind turbine rotors using 1-D modelsDigitalization of scanning lidar measurement campaign planningMassive simplification of the wind farm layout optimization problemSystem-level design studies for large rotorsSensitivity analysis of the effect of wind characteristics and turbine properties on wind turbine loadsPerformance of non-intrusive uncertainty quantification in the aeroservoelastic simulation of wind turbinesPolynomial chaos to efficiently compute the annual energy production in wind farm layout optimizationMultipoint high-fidelity CFD-based aerodynamic shape optimization of a 10 MW wind turbineComparison between upwind and downwind designs of a 10 MW wind turbine rotorCoupled wind turbine design and layout optimization with nonhomogeneous wind turbinesAutomatic measurement and characterization of the dynamic properties of tethered membrane wingsA comparison study on jacket substructures for offshore wind turbines based on optimizationComparison of planetary bearing load-sharing characteristics in wind turbine gearboxesReducing the number of load cases for fatigue damage assessment of offshore wind turbine support structures using a simple severity-based sampling methodFrom wind to loads: wind turbine site-specific load estimation with surrogate models trained on high-fidelity load databasesA systematic approach to offshore wind turbine jacket predesign and optimization: geometry, cost, and surrogate structural code check modelsAdaptive stratified importance sampling: hybridization of extrapolation and importance sampling Monte Carlo methods for estimation of wind turbine extreme loadsGenerating wind power scenarios for probabilistic ramp event prediction using multivariate statistical post-processingEstablishing a robust testing approach for displacement measurement on a rotating horizontal-axis wind turbineApplication of a Monte Carlo procedure for probabilistic fatigue design of floating offshore wind turbinesEffects of defects in composite wind turbine blades – Part 3: A framework for treating defects as uncertainty variables for blade analysisThe risks of extreme load extrapolationMonitoring offshore wind farm power performance with SCADA data and an advanced wake modelNacelle power curve measurement with spinner anemometer and uncertainty evaluationCombined preliminary–detailed design of wind turbines
Gerard Schepers, Pim van Dorp, Remco Verzijlbergh, Peter Baas, and Harmen Jonker
Wind Energ. Sci., 6, 983–996,Short summary
In this article the aeroelastic loads on a 10 MW turbine in response to unconventional wind conditions selected from a year-long large-eddy simulation on a site at the North Sea are evaluated. Thereto an assessment is made of the practical importance of these wind conditions within an aeroelastic context based on high-fidelity wind modelling. Moreover the accuracy of BEM-based methods for modelling such wind conditions is assessed.
Quanjiang Yu, Michael Patriksson, and Serik Sagitov
Wind Energ. Sci., 6, 949–959,Short summary
There are two ways to maintain a multi-component system: corrective maintenance, when a broken component is replaced with a new one, and preventive maintenance (PM), when some components are replaced in a planned manner before they break down. This article proposes a mathematical model for finding an optimal time to perform the next PM activity and selecting the components which should be replaced. The model is fast to solve, and it can be used as a key module in a maintenance scheduling app.
Kenneth Loenbaek, Christian Bak, Jens I. Madsen, and Michael McWilliam
Wind Energ. Sci., 6, 903–915,Short summary
We present a model for assessing the aerodynamic performance of a wind turbine rotor through a different parametrization of the classical blade element momentum model. The model establishes an analytical relationship between the loading in the flow direction and the power along the rotor span. The main benefit of the model is the ease with which it can be applied for rotor optimization and especially load constraint power optimization.
Kenneth Loenbaek, Christian Bak, and Michael McWilliam
Wind Energ. Sci., 6, 917–933,Short summary
A novel wind turbine rotor optimization methodology is presented. Using an assumption of radial independence it is possible to obtain the Pareto-optimal relationship between power and loads through the use of KKT multipliers, leaving an optimization problem that can be solved at each radial station independently. Combining it with a simple cost function it is possible to analytically solve for the optimal power per cost with given inputs for the aerodynamics and the cost function.
Erik Quaeghebeur, René Bos, and Michiel B. Zaaijer
Wind Energ. Sci., 6, 815–839,Short summary
We present a technique to support the optimal layout (placement) of wind turbines in a wind farm. It efficiently determines good directions and distances for moving turbines. An improved layout reduces production losses and so makes the farm project economically more attractive. Compared to most existing techniques, our approach requires less time. This allows wind farm designers to explore more alternatives and provides the flexibility to adapt the layout to site-specific requirements.
Helena Canet, Pietro Bortolotti, and Carlo L. Bottasso
Wind Energ. Sci., 6, 601–626,Short summary
The paper analyzes in detail the problem of scaling, considering both the steady-state and transient response cases, including the effects of aerodynamics, elasticity, inertia, gravity, and actuation. After a general theoretical analysis of the problem, the article considers two alternative ways of designing a scaled rotor. The two methods are then applied to the scaling of a 10 MW turbine of 180 m in diameter down to three different sizes (54, 27, and 2.8 m).
Freia Harzendorf, Ralf Schelenz, and Georg Jacobs
Wind Energ. Sci., 6, 571–584,Short summary
Making wind turbines more reliable over their lifetime is an important goal for improving wind turbine technology. The wind turbine drivetrain has a major influence on turbine reliability. This paper presents an approach that will help to identify holistically better drivetrain concepts in an early product design phase from an operational perspective as it is able to estimate and assess drivetrain-concept-specific inherent risks in the operational phase.
Jan Wiśniewski, Krzysztof Rogowski, Konrad Gumowski, and Jacek Szumbarski
Wind Energ. Sci., 6, 287–294,Short summary
The article describes results of experimental wind tunnel and CFD testing of four different straight-bladed vertical axis wind turbine model configurations. The experiment tested a novel concept of vertically dividing and azimuthally shifting a turbine rotor into two parts with a specific uneven height division in order to limit cycle amplitudes and average cycle values of bending moments at the bottom of the turbine shaft to increase product lifetime, especially for industrial-scale turbines.
Gesine Wanke, Leonardo Bergami, Frederik Zahle, and David Robert Verelst
Wind Energ. Sci., 6, 203–220,Short summary
This article regards a rotor redesign for a wind turbine in upwind and in downwind rotor configurations. A simple optimization tool is used to estimate the aerodynamic planform, as well as the structural mass distribution of the rotor blade. The designs are evaluated in full load base calculations according to the IEC standard with the aeroelastic tool HAWC2. A scaling model is used to scale turbine and energy costs from the design loads and compare the costs for the turbine configurations.
Oliver Menck, Matthias Stammler, and Florian Schleich
Wind Energ. Sci., 5, 1743–1754,Short summary
Blade bearings of wind turbines experience unusual loads compared to bearings in other industrial applications, which adds some difficulty to the application of otherwise well-established calculation methods, like fatigue lifetime. As a result, different methods for such calculations can be found in the literature. This paper compares three approaches of varying complexity and comes to the conclusion that the simplest of the methods is very inaccurate compared to the more complex methods.
Gianluca Zorzi, Amol Mankar, Joey Velarde, John D. Sørensen, Patrick Arnold, and Fabian Kirsch
Wind Energ. Sci., 5, 1521–1535,Short summary
Storms, typhoons or seismic actions are likely to cause permanent rotation of offshore wind turbine foundations. Excessive rotation jeopardizes the operation of the wind turbine. In this study geotechnical, loads and probabilistic modelling are used to develop a reliability framework for predicting the rotation of the foundation under cyclic lateral loading.
Nicola Bodini and Mike Optis
Wind Energ. Sci., 5, 1435–1448,Short summary
Calculations of annual energy production (AEP) and its uncertainty are critical for wind farm financial transactions. Standard industry practice assumes that different uncertainty categories within an AEP calculation are uncorrelated and can therefore be combined through a sum of squares approach. In this project, we show the limits of this assumption by performing operational AEP estimates for over 470 wind farms in the United States and propose a more accurate way to combine uncertainties.
Wind Energ. Sci., 5, 1375–1397,Short summary
One of the major challenges when working with wind turbine sensor data in practice is the presence of systematic changes in signal behaviour induced by malfunctions or maintenance actions. We found that approximately every third signal is affected by such change points and introduce an algorithm which reliably detects them in a highly automated fashion. The algorithm enables the application of data-driven techniques to monitor wind turbine components using data from commonly installed sensors.
Emmanuel Branlard, Dylan Giardina, and Cameron S. D. Brown
Wind Energ. Sci., 5, 1155–1167,Short summary
The paper presents an application of the Kalman filtering technique to estimate loads on a wind turbine. The approach combines a mechanical model and a set of measurements to estimate signals that are not available in the measurements, such as wind speed, thrust, tower position, and tower loads. The model is severalfold faster than real time and is intended to be run online, for instance, to evaluate real-time fatigue life consumption of a field turbine using a digital twin.
Laura Schröder, Nikolay Krasimirov Dimitrov, and David Robert Verelst
Wind Energ. Sci., 5, 1007–1022,Short summary
We suggest a methodology for correlating loads with component reliability of turbines in wind farms by combining physical modeling with machine learning. The suggested approach is demonstrated on an offshore wind farm for comparing performance, loads and lifetime estimations against recorded main bearing failures from maintenance reports. It is found that turbines positioned at the border of the wind farm with a higher expected AEP are estimated to experience earlier main bearing failures.
João Pacheco, Silvina Guimarães, Carlos Moutinho, Miguel Marques, José Carlos Matos, and Filipe Magalhães
Wind Energ. Sci., 5, 983–996,Short summary
This paper introduces the Tocha wind farm, presents the different layouts adopted in the instrumentation of the wind turbines and shows initial results. At this preliminary stage, the capabilities of the very extensive monitoring layout are demonstrated. The results presented demonstrate the ability of the different monitoring components to track the modal parameters of the system, composed of tower and rotor, and to characterize the internal loads at the tower base and blade roots.
Gesine Wanke, Leonardo Bergami, and David Robert Verelst
Wind Energ. Sci., 5, 929–944,Short summary
Converting an upwind wind turbine into a downwind configuration is shown to come with higher edgewise loads due to lower edgewise damping. The study shows from modal displacements of a reduced-order turbine model that the interaction between the forces on the rotor, the rotor motion, and the tower torsion is the main reason for the observed damping decrease.
Malo Rosemeier and Matthias Saathoff
Wind Energ. Sci., 5, 897–909,Short summary
A huge number of wind turbines have reached their designated lifetime of 20 years. Most of the turbines installed were overdesigned. In practice, these turbines could potentially operate longer to increase the energy yield. For the use case turbine considered in this work, a simple lifetime extension of 8.7 years increases the energy yield by 43.5 %. When the swept rotor area is increased by means of a blade tip extension, the yield is increased by an additional 2.3 %.
Sebastian Perez-Becker, Francesco Papi, Joseph Saverin, David Marten, Alessandro Bianchini, and Christian Oliver Paschereit
Wind Energ. Sci., 5, 721–743,Short summary
Aeroelastic design load calculations play a key role in determining the design loads of the different wind turbine components. This study compares load estimations from calculations using a Blade Element Momentum aerodynamic model with estimations from calculations using a higher-order Lifting-Line Free Vortex Wake aerodynamic model. The paper finds and explains the differences in fatigue and extreme turbine loads for power production simulations that cover a wide range of turbulent wind speeds.
Kwangtae Ha, Moritz Bätge, David Melcher, and Steffen Czichon
Wind Energ. Sci., 5, 591–599,Short summary
This paper outlines a novel segment test methodology for wind turbine rotor blades. It mainly aims at improving the efficiency of the fatigue test as a future test method at Fraunhofer IWES. The numerical simulation reveals that this method has a significant time savings of up to 43 % and 52 % for 60 and 90 m blades, while improving test quality within an acceptable range of overload. This test methodology could be a technical solution for future offshore rotor blades longer than 100 m.
Jaime Liew, Albert M. Urbán, and Søren Juhl Andersen
Wind Energ. Sci., 5, 427–437,Short summary
In wind farms, the interaction between neighboring turbines can cause notable power losses. The focus of the paper is on how the combination of turbine yaw misalignment and wake effects influences the power loss in a wind turbine. The results of the paper show a more notable power loss due to turbine misalignment when turbines are closely spaced. The presented conclusions enable better predictions of a turbine's power production, which can assist the wind farm design process.
Julian Quick, Jennifer King, Ryan N. King, Peter E. Hamlington, and Katherine Dykes
Wind Energ. Sci., 5, 413–426,Short summary
We investigate the trade-offs in optimization of wake steering strategies, where upstream turbines are positioned to deflect wakes away from downstream turbines, with a probabilistic perspective. We identify inputs that are sensitive to uncertainty and demonstrate a realistic optimization under uncertainty for a wind power plant control strategy. Designing explicitly around uncertainty yielded control strategies that were generally less aggressive and more robust to the uncertain input.
Frederick Letson, Rebecca J. Barthelmie, and Sara C. Pryor
Wind Energ. Sci., 5, 331–347,Short summary
Wind turbine blade leading edge erosion (LEE) is potentially a significant source of energy loss and expense for wind farm operators. This study presents a novel approach to characterizing LEE potential from precipitation across the contiguous USA based on publicly available National Weather Service dual-polarization RADAR data. The approach is described in detail and illustrated using six locations distributed across parts of the USA that have substantial wind turbine deployments.
Erik Quaeghebeur, Sebastian Sanchez Perez-Moreno, and Michiel B. Zaaijer
Wind Energ. Sci., 5, 259–284,Short summary
The design and management of an offshore wind farm involve expertise in many disciplines. It is hard for a single person to maintain the overview needed. Therefore, we have created WESgraph, a knowledge base for the wind farm domain, implemented as a graph database. It stores descriptions of the multitude of domain concepts and their various interconnections. It allows users to explore the domain and search for relationships within and across disciplines, enabling various applications.
Lars Einar S. Stieng and Michael Muskulus
Wind Energ. Sci., 5, 171–198,Short summary
We present a framework for reducing the cost of support structures for offshore wind turbines that takes into account the many uncertainties that go into the design process. The results demonstrate how an efficient new approach, tailored for support structure design, allows the state of the art for design without uncertainties to be used within a framework that does include these uncertainties. This allows more realistic, and less conservative, design methods to be used for practical design.
Kenneth Loenbaek, Christian Bak, Jens I. Madsen, and Bjarke Dam
Wind Energ. Sci., 5, 155–170,Short summary
From the basic aerodynamic theory of wind turbine rotors, it is a well-known fact that there is a relationship between the loading of the rotor and power efficiency. It shows that there is a loading that maximizes the power efficiency, and it is common to target this maximum when designing rotors. But in this paper it is found that for rotors constrained by a load, the maximum power is found by decreasing the loading and increasing the rotor radius. Max power efficiency is therefore not optimal.
Nikola Vasiljević, Andrea Vignaroli, Andreas Bechmann, and Rozenn Wagner
Wind Energ. Sci., 5, 73–87,Short summary
A WindScanner system consisting of two synchronized scanning lidars potentially represents a cost-effective solution for multipoint measurements. However, the lidar limitations and the site limitations are detrimental to the installation of lidars and number and location of measurement positions. To simplify the process of finding suitable measurement positions and lidar installation locations, a campaign planning workflow was devised. The paper describes the workflow and how it was digitalized.
Andrew P. J. Stanley and Andrew Ning
Wind Energ. Sci., 4, 663–676,Short summary
When designing a wind farm, one crucial step is finding the correct location or optimizing the location of the wind turbines to maximize power production. In the past, optimizing the turbine layout of large wind farms has been difficult because of the large number of interacting variables. In this paper, we present the boundary-grid parameterization method, which defines the layout of any wind farm with only five variables, allowing people to study and design wind farms regardless of the size.
Daniel S. Zalkind, Gavin K. Ananda, Mayank Chetan, Dana P. Martin, Christopher J. Bay, Kathryn E. Johnson, Eric Loth, D. Todd Griffith, Michael S. Selig, and Lucy Y. Pao
Wind Energ. Sci., 4, 595–618,Short summary
We present a model that both (1) reduces the computational effort involved in analyzing design trade-offs and (2) provides a qualitative understanding of the root cause of fatigue and extreme structural loads for wind turbine components from the blades to the tower base. We use this model in conjunction with design loads from high-fidelity simulations to analyze and compare the trade-offs between power capture and structural loading for large rotor concepts.
Amy N. Robertson, Kelsey Shaler, Latha Sethuraman, and Jason Jonkman
Wind Energ. Sci., 4, 479–513,Short summary
This paper identifies the most sensitive parameters for the load response of a 5 MW wind turbine. Two sets of parameters are examined: one set relating to the wind excitation characteristics and a second related to the physical properties of the wind turbine. The two sensitivity analyses are done separately, and the top most-sensitive parameters are identified for different load outputs throughout the structure. The findings will guide future validation campaigns and measurement needs.
Pietro Bortolotti, Helena Canet, Carlo L. Bottasso, and Jaikumar Loganathan
Wind Energ. Sci., 4, 397–406,Short summary
The paper studies the effects of uncertainties in aeroservoelastic wind turbine models. Uncertainties are associated with the wind inflow characteristics and the blade surface state, and they are propagated by means of two non-intrusive methods throughout the aeroservoelastic model of a large conceptual offshore wind turbine. Results are compared with a brute-force extensive Monte Carlo sampling to assess the convergence characteristics of the non-intrusive approaches.
Andrés Santiago Padrón, Jared Thomas, Andrew P. J. Stanley, Juan J. Alonso, and Andrew Ning
Wind Energ. Sci., 4, 211–231,Short summary
We propose the use of a new method to efficiently compute the annual energy production (AEP) of a wind farm by properly handling the uncertainties in the wind direction and wind speed. We apply the new ideas to the layout optimization of a large wind farm. We show significant computational savings by reducing the number of simulations required to accurately compute and optimize the AEP of different wind farms.
Mads H. Aa. Madsen, Frederik Zahle, Niels N. Sørensen, and Joaquim R. R. A. Martins
Wind Energ. Sci., 4, 163–192,Short summary
The wind energy industry relies heavily on CFD to analyze new designs. This paper investigates a way to utilize CFD further upstream the design process where lower-fidelity methods are used. We present the first comprehensive 3-D CFD adjoint-based shape optimization of a 10 MW modern offshore wind turbine. The present work shows that, with the right tools, we can model the entire geometry, including the root, and optimize modern wind turbine rotors at the cost of a few hundred CFD evaluations.
Pietro Bortolotti, Abhinav Kapila, and Carlo L. Bottasso
Wind Energ. Sci., 4, 115–125,Short summary
The paper compares upwind and downwind three-bladed configurations for a 10 MW wind turbine in terms of power and loads. For the downwind case, the study also considers a load-aligned solution with active coning. Results indicate that downwind solutions are slightly more advantageous than upwind ones, although improvements are small. Additionally, pre-alignment is difficult to achieve in practice, and the active coning solution is associated with very significant engineering challenges.
Andrew P. J. Stanley and Andrew Ning
Wind Energ. Sci., 4, 99–114,Short summary
We show that optimizing wind turbine design and wind turbine layout at the same time is superior to doing so sequentially. This coupled optimization can reduce the cost of energy by 2–5 % compared to sequential optimization. We also demonstrate that wind farms with closely spaced wind turbines can greatly benefit from different turbine designs throughout the farm. Heterogeneous turbine design can reduce the cost of energy by up to 10 % compared to farms with all identical turbines.
Jan Hummel, Dietmar Göhlich, and Roland Schmehl
Wind Energ. Sci., 4, 41–55,Short summary
We describe a tow test setup for the reproducible measurement of aerodynamic, structural dynamic and flight dynamic properties of tethered membrane wings. The test procedure is based on repeatable automated maneuvers with the entire kite system under realistic conditions. The developed measurement method can be used to quantitatively compare different wing designs, to validate and improve simulation models, and to systematically improve kite designs.
Jan Häfele, Cristian G. Gebhardt, and Raimund Rolfes
Wind Energ. Sci., 4, 23–40,Short summary
To reduce the levelized costs of offshore wind energy, capital expenses of substructures have to be decreased significantly. Therefore, structural optimization approaches have been proposed in the recent past, mainly to improve the design of jackets. This work proposes a holistic approach to jacket optimization, which addresses some problems arising from methods that were presented in the literature.
Jonathan Keller, Yi Guo, Zhiwei Zhang, and Doug Lucas
Wind Energ. Sci., 3, 947–960,Short summary
The US Department of Energy's National Renewable Energy Laboratory (NREL) and industry partners successfully demonstrated a new gearbox design using preloaded tapered roller bearings in the planetary section. The new gearbox design demonstrated improved planetary load-sharing characteristics in the presence of rotor pitch and yaw moments, resulting in a predicted gearbox lifetime that is 3.5 times greater than the previous conventional design with cylindrical roller bearings.
Lars Einar S. Stieng and Michael Muskulus
Wind Energ. Sci., 3, 805–818,Short summary
This study was done in order to simplify the analysis needed for the assessment of safety criteria that offshore wind turbine support structures must satisfy. The work was done via simulations of the system and computations and analyses of the resulting data. The results show that the proposed methodology has great potential for simplifying the assessment procedure while retaining acceptable accuracy compared to the full analysis. There are several applications within both research and industry.
Nikolay Dimitrov, Mark C. Kelly, Andrea Vignaroli, and Jacob Berg
Wind Energ. Sci., 3, 767–790,Short summary
Wind energy site suitability assessment procedures often require estimating the loads a wind turbine will be subject to when installed. The estimation is often time-consuming and requires several iterations. We have developed a procedure for quick and accurate estimation of site-specific wind turbine loads. Our approach employs computationally efficient parametric models that are calibrated to high-fidelity load simulations. The result is a significant reduction in computation efforts.
Jan Häfele, Rick R. Damiani, Ryan N. King, Cristian G. Gebhardt, and Raimund Rolfes
Wind Energ. Sci., 3, 553–572,Short summary
The present work provides a technical basis for the design of jacket structures used as substructures for offshore wind turbines. This involves models for the geometry, costs, and structural design code checks. An example application is shown in this paper, in which three different structural designs are compared. This work may lead to improved design approaches and finally to a cost reduction of offshore substructures.
Peter Graf, Katherine Dykes, Rick Damiani, Jason Jonkman, and Paul Veers
Wind Energ. Sci., 3, 475–487,Short summary
Current approaches to wind turbine extreme load estimation are insufficient to routinely and reliably make required estimates over 50-year return periods. Our work hybridizes the two main approaches and casts the problem as stochastic optimization. However, the extreme variability in turbine response implies even an optimal sampling strategy needs unrealistic computing resources. We therefore conclude that further improvement requires better understanding of the underlying causes of loads.
Rochelle P. Worsnop, Michael Scheuerer, Thomas M. Hamill, and Julie K. Lundquist
Wind Energ. Sci., 3, 371–393,Short summary
This paper uses four statistical methods to generate probabilistic wind speed and power ramp forecasts from the High Resolution Rapid Refresh model. The results show that these methods can provide necessary uncertainty information of power ramp forecasts. These probabilistic forecasts can aid in decisions regarding power production and grid integration of wind power.
Nadia Najafi and Allan Vesth
Wind Energ. Sci., 3, 301–311,Short summary
This paper presents a well-defined procedure for measuring the displacement on a HAWT via stereo photometry. The method is demonstrated by measuring displacement during operation of a scaled down turbine model. The method is developed in (1) camera calibration and (2) tracking algorithm. We introduce an efficient, easy and practical calibration method for measurement in the large field of views that is always a challenge. Tracking algorithm also tracks the markers during rotation successfully.
Kolja Müller and Po Wen Cheng
Wind Energ. Sci., 3, 149–162,Short summary
An efficient and accurate Monte Carlo approach is presented to assess the lifetime fatigue loading on a floating offshore wind turbine accurately. This is typically challenging in simulation effort due to the many different combinations of relevant environmental conditions which need to be considered. The applied method uses quasi-random Sobol sequences and shows promising performance with respect to convergence and accuracy.
Trey W. Riddle, Jared W. Nelson, and Douglas S. Cairns
Wind Energ. Sci., 3, 107–120,Short summary
The Department of Energy sponsored, Sandia National Laboratory led Blade Reliability Collaborative was formed to address wind turbine blade reliability. Utilizing the results of characterization and mechanical testing studies, probabilistic models were developed to assess the reliability of a wind blade with known defects. By treating defects as random variables, the results indicate that characterization of defects and reduction of design uncertainty is possible for wind turbine blades.
Stefan F. van Eijk, René Bos, and Wim A. A. M. Bierbooms
Wind Energ. Sci., 2, 377–386,Short summary
Predicting the 50-year extreme loads for wind turbines requires a tremendous computational effort. Therefore, designers often have to extrapolate from relatively small data sets and have to settle for some degree of uncertainty. We investigated the impact of this uncertainty on practical design problems by drawing subsets from a 96-year load data set and using a crude Monte Carlo method to find the 50-year load. The results show that designers have to be careful with selecting sample sizes.
Niko Mittelmeier, Tomas Blodau, and Martin Kühn
Wind Energ. Sci., 2, 175–187,Short summary
Efficient detection of wind turbines operating below their expected power output and immediate corrections help maximize asset value. The method presented estimates the environmental conditions from turbine states and uses pre-calculated power lookup tables from a numeric wake model to predict the expected power output. Deviations between the expected and the measured power output are an indication of underperformance. A demonstration of the method's ability to detect underperformance is given.
Giorgio Demurtas, Troels Friis Pedersen, and Rozenn Wagner
Wind Energ. Sci., 2, 97–114,Short summary
In this study we aimed to use the spinner anemometer calibration and nacelle transfer function determined on a reference wind turbine to assess the power performance of a second, identical turbine. An experiment was set up with a met mast and spinner anemometer on each turbine. For each wind turbine, the nacelle power curve agreed with the corresponding power curve within 0.10 % of AEP for the reference wind turbine and within 0.38 % for the second wind turbine, for a mean wind speed of 8 m s−1.
Pietro Bortolotti, Carlo L. Bottasso, and Alessandro Croce
Wind Energ. Sci., 1, 71–88,Short summary
The paper presents a new method to conduct the holistic optimization of a wind turbine. The proposed approach allows one to define the rotor radius and tower height, while simultaneously performing the detailed sizing of rotor and tower. For the rotor, the procedures perform simultaneously the design both from the aerodynamic and structural points of view. The overall optimization seeks a minimum for the cost of energy, while accounting for a wide range of user-defined design constraints.
Boslaugh, S. and Watters, P. A.: Statistics in a Nutshell by Sarah Boslaugh and Paul Andrew Watters, Copyright © 2008 Sarah Boslaugh, O'Reilly Media, Inc., Sebastopol, CA, USA, 2008.
Cosack, N.: Fatigue Load Monitoring with Standard Wind Turbine Signals, University of Stuttgart, Stuttgart, available at: https://d-nb.info/1009926721/34 (last access: 15 June 2019), 2010.
Cosack, N. and Kühn, M.: Ueberwachung von Belastungen an Windenergieanlagen durch Analyse von Standardsignalen.pdf, in: AKIDA Tagungsband, 6. Aachener Kolloquium für Instandhaltung, Diagnose und Anlagenüberwachung, 14–15 November 2006, Aachen, 277–283., 2006.
Cosack, N. and Kühn, M.: Prognose von Ermüdungslasten an Windenergieanlagen mittels Standardsignalen und neuronaler Netze.pdf, in: DMK 2007 – Dresdner Maschinenelemente Kolloquium: 5 and 6 December 2007, Dresden, 461–476, 2007.
DNV/Risø: Guidelines for Design of Wind Turbines, 2nd Edn., Jydsk Centraltrykkeri, Denmark, 2002.
Draper, N. R. and Smith, H.: Applied regression analysis, in: Wiley Series in Probability and Statistics, Wiley, 1998.
Goldberger, J., Roweis, S., Hinton, G., and Salakhutdinov, R.: Neighbourhood Components Analysis, Adv. Neural Inform. Process. Syst., 17, 513–520, 2004.
Han, J. and Kamber, M.: Data Mining: Concepts and Techniques, 3rd Edn., The Morgan Kaufmann Series in Data Management Systems, 5, 83–124, 2011.
Hazewinkel, M.: Encyclopaedia of Mathematics (set), 1st Edn., Springer, Netherlands, 1994.
IRENA: Renewable Power Generation Costs in 2014, Bonn, Germany, 2015.
Lind, P. G.: Normal behaviour models for wind turbine vibrations: Comparison of neural networks and a stochastic approach, Energies, 10, 1944, https://doi.org/10.3390/en10121944, 2017.
MathWorks: Fit Data with a Shallow Neural Network – MATLAB & Simulink – MathWorks, available at: https://de.mathworks.com/help/deeplearning/gs/fit-data-with-a-neural-network.html, last access: 5 June 2019.
Melsheimer, M., Kamieth, R., Liebich, R., and Heilmann, Christoph, Grunwald, A.: Practical Experiences from a load measurement campaign for the assessment of the remaining service life of wind turbines, in: German Wind Energy Conference (DEWEK) Proceedings, 19–20 May 2015, Congress Centrum, Bremen, 1–4, 2015.
Noppe, N., Weijtjens, W., and Devriendt, C.: Modeling of quasi-static thrust load of wind turbines based on 1 s SCADA data, Wind Energ. Sci., 3, 139–147, https://doi.org/10.5194/wes-3-139-2018, 2018.
Parsian, M.: Data algorithms: recipes for scaling up with Hadoop and Spark, O'Reilly Media, Inc., O'Reilly Media, Inc., Sebastopol, CA, USA, 2015.
REN21: Renewables 2018 Global Status Report, Paris, France, 2018.
Schedat, M. and Faber, T.: Fatigue load reconstruction on wind turbine structures using structural health monitoring, in: Proceedings DEWEK 2017, 13th German Wind Energy Conference, 17–18 October 2017, Bremen, Germany, 1–4, 2017.
Schedat, M., Faber, T., and Sivanesan, A.: Structural health monitoring concept to predict the remaining lifetime of the wind turbine structure, in: IEEE Proceedings of the 24th Conference on the Domestic Use of Energy, DUE 2016, 30–31 March 2016, Cape Town, South Africa, 1–5, 2016.
Seifert, J., Vera-Tudela, L., and Kühn, M.: Training requirements of a neural network used for fatigue load estimation of offshore wind turbines, Energy Proced., 137, 315–322, https://doi.org/10.1016/j.egypro.2017.10.356, 2017.
Sharma, N. and Saroha, K.: Study of dimension reduction methodologies in data mining, in: IEEE International Conference on Computing, Communication and Automation, ICCCA 2015, 15–16 May 2015, Greater Noida, India, 133–137, 2015.
Smolka, U. and Cheng, P. W.: On the Design of Measurement Campaigns for Fatigue Life Monitoring of Offshore Wind Turbines, in: The Twenty-third International Offshore and Polar Engineering Conference, vol. 9, International Society of Offshore and Polar Engineers, Anchorage, Alaska, USA, 408–413, 2013.
Vera-Tudela, L. and Kühn, M.: On the Selection of Input Variables for a Wind Turbine Load Monitoring System, Procedia Technol., 15, 726–736, https://doi.org/10.1016/j.protcy.2014.09.045, 2014.
Vidal, R., Ma, Y., and Sastry, S. S.: Principal Component Analysis, Chemometr. Intell. Lab. Syst., 2, 37–52, https://doi.org/10.1016/0169-7439(87)80084-9, 1987.
Ziegler, L., Gonzalez, E., Rubert, T., Smolka, U., and Melero, J. J.: Lifetime extension of onshore wind turbines: A review covering Germany, Spain, Denmark, and the UK, Renew. Sustain. Energy Rev., 82, 1261–1271, https://doi.org/10.1016/j.rser.2017.09.100, 2018.
The assessment of the structural condition and technical lifetime extension of a wind turbine is challenging due to lack of information for the estimation of fatigue loads. This paper demonstrates the modelling of damage-equivalent loads of the fore–aft bending moments of a wind turbine tower, highlighting the advantage of using the neighbourhood component analysis. This feature selection technique is compared to correlation analysis, stepwise regression, and principal component analysis.
The assessment of the structural condition and technical lifetime extension of a wind turbine is...