MS10.6 - Dynamics of Wind Energy Systems

From July 2021, a full-scale measurement campaign is carried out on two wind turbine towers with smooth surfaces (k/D  1.0∙10-6). Tower response (through accelerations) and wind pressure on the surface are measured. Throughout the one-year monitoring period, there were challenges in transferring full-scale data into a ready-to-use VIV data set. For example, the spatially varying wind conditions despite the small distance between the met masts and the towers. This work aims to provide a ready-to-use full-scale data set for the development of VIV prediction models. A data selection approach based on three simultaneous measurements of wind conditions and maximum moving standard deviation of oscillation is proposed. In addition, the load parameters obtained from the wind pressure measurements in the field and from wind tunnel tests are considered. An enhanced calculation is carried out by defining local aerodynamic damping for sectional heights of the tower. A comparison of the data set with the calculated response is presented to assess the calculation method and to give ideas on sensitive input parameters considering the full-scale condition.

Stability analysis for a large-scale wind turbine has been an increasing topic in recent years, but its intrinsic mechanism is still blurring. This paper aims at developing an explicit and fully linearized Aero-Hydro-Struc model to reveal the instability phenomena and giving a novel insight into understanding the instability mechanism. This paper introduced a series of evaluation procedures and methods for the stability judgment of a floating offshore wind turbine (FOWT) based on a linearized model. The linearized model contains linear aerodynamics model and hydrodynamics model. The nonlinear aerodynamics model including the widely-used steady engineering model - Blade Element Momentum theory (BEM) - for an operating wind turbine, and a simpler model for parked wind turbines is also introduced, they are both linearized in this paper. Besides, Morison’s equation and potential flow theory are adopted for the hydrodynamics model and their linear versions are introduced in this paper. Linear versions of aerodynamic load models are developed by means of the first-order Taylor’s expansion without considering the second-order items, and the stability analyses are conducted based on the linear models using the state-space method, as long as the implicit and nonlinear aerodynamic and hydrodynamic loads are successfully linearized as static forces and damping matrices. Stand-still, operating and parked (with yaw angles) wind turbines have been analyzed through the proposed method in this paper. It is found that the platform is unlikely to suffer from instability under these operational conditions as the hydrodynamic damping participates prominently, while the blade edgewise and the tower Side-Side (SS) movements are more possible to experience instability for a parked wind turbine with yaw misalignments, since the aerodynamic damping is negative and the hydrodynamic damping is not taken into account in these degree of freedoms (DOFs), and these two kinds of instabilities are caused by stall-induced vibrations. It’s also observed that the instabilities of blade and tower DOFs mainly arise from the blade root airfoil sections since the angles of attack are particularly high because of the higher twist. Besides, blade flapwise and edgewise, as well as the tower Fore-After (FA) and SS movements, are proved always highly damped for a normally operated wind turbine, which indicates the possibilities of the instability for these DOFs remain low. Moreover, the time-series and frequency-domain responses for an operating wind turbine have been studied as well, it’s found the blade vibrations in frequency-domain are dominated by the rotor rotation speed and are coupled with the platform and wave frequencies, while the tower responses are related to the rotor rotation speed, platform frequencies, wave frequencies, and the tower natural frequencies.

Operating wind turbines are complex systems due to fluctuating aerodynamic loads induced by turbulent wind fields, rotor rotation and the tower-rotor coupling dynamics. Difficulties have been claimed in structural identification of operating wind turbines using traditional operational modal techniques due to harmonics in responses caused by rotor rotation and time-invariant nature of such systems. Structural identification of full-size wind turbines through field measurement is even more difficult due to uncertainties of ambient excitations. In this study, structural identification is conducted to a carefully designed scale-down wind turbine model. Rotor rotation is realised via a motor and the influence of different wind fields is studied by wind tunnel tests. Dynamic responses of the blades and tower are measured through Digital Image Correlation (DIC), and the measured responses are processed by different system identification techniques including traditional Stochastic Subspace Identification (SSI) and specifically developed identification method based on Bayesian inference. It is found the later method can successfully identify the key modal parameters of the wind turbine model with sufficient accuracy, and these parameters are comparable to those of the prototype wind turbine.

Modern wind turbines are large and slender dynamical structures with a fatigue loading profile of complex nature. The guarantee of their structural integrity is paramount for materializing cost efficient and more reliable wind energy. The measurement of the global dynamic response and loads of wind turbines is fundamental to achieving this goal. However, an industry-wide, cost-effective direct sensing framework is yet to arise. Moreover, deploying physical sensors and measurement systems on every structural component of interest of a wind turbine induces prohibitive costs in deployment, maintenance and data management. Considering that direct fluid-structure interaction simulations on a farm level are not computationally feasible, the preferred path for structural response estimation on wind farms has been surrogate modelling. Within this landscape, new model architectures have risen in recent years which are able to take into account graph structured data (i.e. non-euclidean data). Wind turbines positioned in a farm, where there is a layout- and topology-dependant interplay of aerodynamic wake affecting the loading profile and power production, lend themselves perfectly to this paradigm. Thus, in this contribution, we introduce the use of graph neural networks (GNN) for layout-agnostic saptio-temporal joint modelling of fatigue loads effects, rotor-averaged wind speed and power production on individual turbines of wind farms. To this end, we generate random samples of inflow conditions from a Weibull distribution for wind speed, uniform wind direction and conditional normal distributions of wind shear and nacelle yaw angles. Additionally, 200 wind farm layouts are randomly generated based on different geometric shapes (rectangle, triangle, ellipse and sparse circles) with random parametrization (varying orientations, length/width ratio) for different numbers of turbines and minimal distance (based on the rotor diameter). Both the arbitrary layouts and the random inflow conditions are used as inputs for PyWake, a wind farm simulation tool capable of calculating wind farm flow fields, power and fatigue loads. In our analysis, we develop and compare the performance of Graph Isomorphism Networks (GIN), Graph Convolution Networks (GCN) and Graph Attention Networks (GAT) in their accuracy and ability to generalize their joint predictions for unseen layouts, uncertain inflow conditions and fatigue load estimation on the blade root, tower top and tower base of any wind turbine in the farm.