A framework characterizing the degradation of wind turbines for use in multiple-input damage-aware farm control is suggested. The focus is on the fatigue damage of the powertrain (drivetrain + generator) system, but the methodology may be extended to other components. A database of steady-state damage analyses for different operating conditions (average wind speeds, turbulence levels, power demands, and yaw misalignment angles) using aero-hydro-servo-elastic simulations is first generated. Then, a weighted damage index based on probabilistic long-term fatigue damage analysis of the powertrain system components is suggested and used to represent degradation at the farm level for control purposes. The focus is on curtailed conditions where the farm controller dispatches power commands to individual turbines in order to track a demanded power reference (rather than seeking to maximize power) at the farm level. As a secondary objective, the controller seeks to mitigate degradation through a smart combination of power commands and yaw offset angles, making use of the weighted degradation index. The potential of the proposed approach is demonstrated through a case study on the TotalControl Reference Wind Power Plant in a FLORIS-based simulation framework. The proposed farm controller is compared with the conventional one without damage mitigation feature and with damage mitigation but without yaw angle as the control input. It is found that combining yawing and downregulation effectively slows down degradation on the main bearing and powertrain as a whole.

Abstract Turbulent wind fields are known to be a major driver for structural loads and power fluctuations on offshore wind turbines. At the single‐turbine scale, there exist well‐established design standards based on wind spectra and coherence functions calibrated from years of measurements, which are used to generate multiple 10‐min wind field realisations known as synthetic turbulence boxes, themselves used as input to turbine‐scale aero‐hydro‐servo elastic codes. These methods are however not directly applicable at farm scale. When analysing the dynamics of large offshore wind farms, measurements reveal the importance of large, low‐frequency turbulent vortices for power fluctuations and hence for wind farm control and grid integration. Also, farm‐scale wind fields are needed as input to farm‐scale aero‐servo‐elastic codes for the modelling of wake dynamics, affecting structural loads. These new concerns motivate an upgrade in the original turbine‐scale wind field representation: (1) spectral models need to be based on farm‐scale measurements, (2) the frozen‐turbulence assumption merging temporal and along‐wind coherence must be lifted, (3) simplifications are needed to reduce the number of degrees of freedom as the domain becomes excessively large. This paper suggests models and algorithms for aggregated farm‐wide corrrelated synthetic turbulence generation—lumping the wind field into space‐averaged quantities—adapted to the aero‐hydro‐servo elastic modelling of large offshore wind farms. Starting from the work of Sørensen et al. in the early 2000s for grid integration purposes, methods for structural load modelling (through wake meandering and high‐resolution wind field reconstruction) are introduced. Implementation and efficiency matters involving mathematical subtleties are then presented. Finally, numerical experiments are carried out to (1) verify the approach and implementation against a state‐of‐the‐art point‐based—as opposite to aggregated—synthetic turbulence generation code and (2) illustrate the benefit of turbulence aggregation for the modelling of large offshore wind farms.

AbstractWind farm flow control (WFFC) is an emerging technology involving coordinated operation of wind turbines within a wind farm to achieve collective goals. To design and evaluate controllers, wind farm flow models are used that capture the key aerodynamics of the system whilst remaining computationally efficient for iterative controller design. This review article reveals considerable heterogeneity in the potential wind farm flow models to study WFFC. Lack of consensus is problematic as differences in results from separate studies are attributable to both controller and model effects, making it hard to draw comparative conclusions. Hence, an expert elicitation is completed surveying WFFC practitioners. Two key contributions are presented. First, a guide to available software for WFFC, which, combined with results from an expert elicitation on flow model requirements, facilitates selection of suitable software for investigating WFFC problems. Secondly, critical future research areas are identified. Research into high fidelity wind direction models (particularly transient effects) and wake meandering models for fatigue load investigations are identified as critical to the field. A lack of consensus regarding the importance of atmospheric boundary layer effects, wake induced turbulence, and lateral wind correlation identifies the requirement of sensitivity studies in these areas.

Increasing the efficiency of wind farms is important for speeding up the transition from fossil fuels to renewable energy sources. Current wind farm control relies on maximization of power generation of individual turbines. However, research has demonstrated that plant-wide wind farm control could optimize the performance of a wind farm. Wind farm simulation tools are crucial in designing, testing, and validating wind farm controls. FAST.Farm is a recently developed multi-physics engineering tool for modeling wind farm performance by solving the aero-hydro-servoelastic dynamics of each individual turbine. The capabilities of FAST.Farm for control design purposes can be extended through a co-simulation with MATLAB/Simulink. Therefore, a MATLAB/Simulink interface with FAST.Farm has been developed. The creation and operation of this interface are explained in this paper. This interface supports developing and testing advanced closed-loop control at the wind turbine and wind farm levels.

With the coming installation of hundreds of GW of offshore wind power, penetration of the inherent power fluctuations into the electricity grid will become significant. Therefore, the use of wind farms as power reserve providers to support the regulation of the grid’s voltage and frequency through delivering a desired power is expected to increase. As a result, wind turbines will not be necessarily delivering the maximum available power anymore – known as curtailed or derated operation – and will have to be able to deal with time-variant power demand. For this purpose, power setpoints from the grid are dispatched at the farm-level and then tracked at the turbine-level under the constraint of available power in the wind (known as active power control). The idea of this work is taking advantage of the additional degree of freedom lying in the power dispatch between turbines when operating in curtailed conditions. As failure of power train system components is frequent, costly and predictable, we seek to introduce power train degradation into the farm control objectives. To this end, a data-driven model of drivetrain fatigue damage as function of wind conditions and derating factor adapted to the farm active power control objective function is developed based on the pre-analysis of single-turbine simulations and degradation calculations, where the increased turbulence intensity due to wind farm wake effect is also considered. The proposed analytical power train degradation model is computationally efficient, can reflect the fatigue damage of individual gears and bearings in the overall power train life function and in contrast with high-fidelity models can be easily adjusted for different drivetrain configurations. A case study on the TotalControl reference wind power plant is demonstrated.

Abstract A novel approach for the modeling of rotor‐integrated aerodynamic loads is suggested to answer the need for a comprehensive, insightful, and analytical actuator disc model. All the six degrees of freedom (including tangential components) are considered. It is shown that loads may be written as a quadratic form of a reduced six‐component velocity vector at the hub. The individual contributions of lift and drag, azimuthal variations, as well as blade pitching and tip losses are isolated. Errors introduced by the necessary approximations are discussed, and parametric corrections are considered. Parameter identification methods are then suggested, and the performance of the resulting calibrated analytical models is assessed. Results show that the new modeling approach is able to accurately model both the mean values of the thrust and power coefficients and their derivatives with respect to tip‐speed ratio and pitch angle across the full range of operating wind speeds. Furthermore, it is able to reconstruct the general rotor behavior with a minimal amount of information available. Tangential components are also well modeled, although they require the knowledge of airfoil properties. The model's architecture leaves room for extensions to dynamic flow, skewed flow, and azimuthal load variations.

The operation of future wind power plants (WPPs) will rely on maximizing overall profit rather than power production. Factors like current electricity price, grid code compliance, and maintenance costs will affect operational decision. Therefore, it is expected that plant control strategies will trade different control objectives while respecting the industrial practice, where the different levels in the hierarchy may be provided by different equipment vendors. This work proposes a plant controller that aims to track an operator power command at the WPP level and to reject low-frequency loading due to turbulent winds at individual turbines. The plant control respects the hierarchy in which the turbine-level controller takes precedence, interacting only via power set-point commands to each turbine. The control algorithm is based on parallel control loops along with look-up tables and gain scheduling, and its simplicity and effectiveness make it well-suited for industrial applications. Preliminary simulation results with a 32-turbine WPP indicate that there is a significant synergy effect by coordinating operation of the turbines in a plant. Structural loads can be reduced on certain turbines without excessively burdening the others, while an accurate plant-wide power tracking is still provided.

A supervisory control algorithm for large wind power plants has been designed. The controller tracks a total power command specified by the plant operator, while compensating for fluctuations in rotor thrust at individual wind turbines. By taking advantage of the fact that each wind turbine is its own independently-controlled generating unit, the plant controller is able to reduce thrust fluctuations on highly-stressed wind turbines, while tightly following the commanded power, provided that this is below the total power available in the given wind conditions.

This article presents the Real-Time Hybrid Model (ReaTHM®) tests that were performed on a 10-MW semi-submersible floating wind turbine in the Ocean Basin at SINTEF Ocean in March 2018. The ReaTHM test method was used for the model tests to circumvent the limitations encountered when performing model tests with wind and waves. The physical model was subject to physical waves, while the rotor and tower loads were simulated in real-time and applied on the model by use of a cable-driven parallel robot. Recent advances in the ReaTHM test method allowed for extended testing possibilities and load application up to the 3p frequency and the first tower bending frequency.

Model testing of offshore structures in ocean basins has been accepted as a necessary step for the validation and calibration of numerical models, as well as for final design checks in extreme environments. While offshore wind power makes no exception, model testing has not shown its full potential due to inherent modeling challenges. Generating highly-controlled wind fields in ocean basins and alleviating aerodynamic modeling errors due to Reynolds number mismatch in Froude-based scaling are prominent examples. To circumvent these issues, the concept of Real-Time Hybrid Model (ReaTHM) testing has been suggested by SINTEF Ocean and NTNU, Norway. Here the wind loads are no longer physically modeled but computed from online-measured motions and a numerical wind field. They are then actuated in real time on the scale model by means of actuators, while subjected to -physical-wave and current loads. This paper aims at presenting design considerations regarding the choice of the actuator(s) and its/their interface with the scale model. The pros and cons of the chosen solution, namely cable-driven parallel robots using industrial servo drives, are presented. The focus is then directed toward the mapping between wind loads to be actuated and tension commands on each cable, called tension allocation. Two layouts corresponding to two ReaTHM testing campaigns performed in SINTEF Ocean’s ocean basin are then compared on various aspects, with emphasis on tension allocation. In addition to proving the feasibility of the chosen technical solution, results show a tradeoff between flexibility (with respect among others to wind direction) and usability on other structures versus minimization of cable tensions. The latter aspect is treated in detail, using theory first adapted from literature on cable-driven parallel robots, then illustrated through relevant examples for the current application.

To achieve cost-effective and reliable structural design of floating wind turbines, efficient and accurate time domain numerical approaches are required to analyse structural responses in design conditions, e.g. wind and waves. This paper focuses on validation of a time-domain numerical approach for determining forces and moments in structural components of floaters. The approach considers floating wind turbines as a system of several structural components, e.g. blades, rotational shaft, nacelle, tower, mooring lines, columns, pontoons and braces. A finite element model is developed to represent global stiffness of the structural components. The external and inertia loads on the structural components are modelled as distributed loads. Hydrodynamic loads on each structural component are derived from the corresponding hydrodynamic coefficients obtained by solving the first order boundary value problem using WAMIT. A 1:30 scaled braceless semi-submersible model test which implements the ReaTHM® testing approach was done by SINTEF Ocean, formerly MARINTEK, in its ocean basin. Measurements of the global forces and moments at the base of a side column of the model and rigid-body motions of the model are compared to the corresponding simulations. This paper focuses on responses in moderate waves for which linear hydrodynamic loads are applicable. Differences in the corresponding simulations and measurements are found to be small, while possible reasons, e.g. synchronizations, non-linear effects, and uncertainties in the measurements and simulations, for the differences are analysed. Essential information about the model test, descriptions of the numerical models, calibrations and results and discussions of the validation are given in this paper.

Real-Time Hybrid Model (ReaTHM) tests of a braceless semi-submersible wind turbine were carried out at MARINTEK’s Ocean Basin in 2015. The tests sought to evaluate the performance of the floating wind turbine (FWT) structure in environmental conditions representative of the Northern North Sea. In order to do so, the tests employed a new hybrid testing method, wherein simulated aerodynamic loads were applied to the physical structure in the laboratory. The test method was found to work well, and is documented in [1]. The present work describes some of the experimental results. The test results showed a high level of repeatability, and permitted accurate investigation of the coupled responses of a FWT, including unique conditions such as blade pitch faults. For example, the influence of the wind turbine controller can be seen in decay tests in pitch and surge. In regular waves, aerodynamic loads due to constant wind had little influence on the structure motions (except for the mean offsets). Tests in irregular waves with and without turbulent wind are compared directly, and the influence of the wave-frequency motions on the aerodynamic damping of wind-induced low-frequency motions can be observed.

This article presents a method for performing Real-Time Hybrid Model testing (ReaTHM testing) of a floating wind turbine (FWT). The advantage of this method compared to the physical modelling of the wind in an ocean basin, is that it solves the Froude-Reynolds scaling conflict, which is a key issue in FWT testing. ReaTHM testing allows for more accurate testing also in transient conditions, or degraded conditions, which are not feasible yet with physical wind. The originality of the presented method lies in the fact that all aerodynamic load components of importance for the structure were identified and applied on the physical model, while in previous similar projects, only the aerodynamic thrust force was applied on the physical model. The way of applying the loads is also new. The article starts with a short review (mostly references) of ReaTHM testing when applied to other fields than marine technology. It then describes the design of the hybrid setup, its qualification, and discusses possible error sources and their quantification. The second part of the article [1] focuses on the performance of a braceless semi-submersible FWT, tested with the developed method. The third part [2] describes how the experimental data was used to calibrate a numerical model of the FWT.

Real-time hybrid model testing (ReaTHM) is a new approach for conducting small-scale experimental campaign [1], [2], [3]. In the case of a floating wind turbine in a wave basin, the aerodynamic loads on the wind turbine may be applied based on simultaneous simulations (coupled to the experiments), while the wave loads and floater response are physically tested. The objective of this paper is to demonstrate the effects of actuation limitation on the ReaTHM testing setup for a particular platform: numerical simulations are employed to examine the effects of not including some components of the aerodynamic loads (or of inducing error, for example in the direction of the force actuation).

Within the field of hydrodynamics, it is fairly easy to find examples of model tests whose performance is impaired by only a subpart of the whole system, which may not be the one of interest. Real-time hybrid testing (RTHT) overcomes this issue by performing scale model testing only on a subpart of the whole structure, the remainder being simulated numerically. The loads acting on the virtual substructure are calculated from online-measured motions of the physical substructure and actuated back on the latter in real-time. RTHT involves data measurement, filtering, force estimation, motion observing and force actuation. The main challenge is to fit all of those items into one time step. A simple case study is suggested. It consists in a linearized one degree of freedom floating wind turbine, whose floating substructure is physically tested while wind loads are numerically simulated and actuated. Design rules to build the corresponding RTHT set up are then presented.