No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Influence of small scale wind velocity fluctuations an the aerodynamic alternating loads
... It seems reasonable to expect that these critical scales will depend on every type of machine. However, as shown here (even with a superposition approach for the fluctuations u 0 ) and in [24] common models and simulation packages for generating synthetic wind fields do not reproduce these two-point statistics. As a consequence, quantitative evaluations of possible effects on WECs due to the non-Gaussian behavior of wind speed increments have not been carried out until very recently [24]. ...
... However, as shown here (even with a superposition approach for the fluctuations u 0 ) and in [24] common models and simulation packages for generating synthetic wind fields do not reproduce these two-point statistics. As a consequence, quantitative evaluations of possible effects on WECs due to the non-Gaussian behavior of wind speed increments have not been carried out until very recently [24]. Finally, future investigatons should focus on n-point statistics for the characterization of, e.g., gust clustering and the identification of critical wind gusts shapes. ...
A general hierarchical statistical framework for the characterization of wind turbulence is proposed. This framework should ease the understanding of different existing statistical approaches and enable a clear path for refined methods of characterization. Low-order statistical descriptions were extended by higher-order statistics with respect to one-point and two-point statistics. In particular, we showed that proper analysis leads to a superstatistics approach for the probability of velocity fluctuations. To demonstrate the importance of our considerations, we analysed wind time series of the research platform FINO 1. On one side, we showed how different statistical aspects can be reproduced quite accurately, whereas on the other, the necessity of more profound approaches was worked out. The analysis of the measured data provides some deep insights into the nature of wind turbulence. Most interestingly, for conditioned data subsets, we found higher-order statistical behavior well known for idealized turbulence. Finally, we give an outlook on how to achieve a general n-point statistical description. Copyright © 2011 John Wiley & Sons, Ltd.
... AeroDyn calculates the aerodynamic loads on wind turbine blade elements based on velocities and positions provided by dynamics analysis routines and wind inputs. For our application we use the blade element-momentum theory to calculate the axial induced velocities in the rotor plane (Mücke, 2009). The virtual WP 1.5MW WEC design is pitch regulated and has a rotor diameter of 70m. ...
This paper presents two complementary approaches to model and validate wind energy converter designs. First, the mechanical model can simulate the inner mechanics and the resulting power output for the numerical design of a converter placed into a wind field. As the simulation exploits a multi-point wind field measurement, the power output modeled displays realistic features like intermittency, which are usually omitted by simpler models. This makes it a good alternative between heavier CFD models and oversimplified standard models. Second, a stochastic model can be used to characterize the overall performance of an existing converter, and furthermore model through a stochastic Langevin equation its overall power conversion when placed under any wind conditions. While it reproduces the proper long-time behavior of the converter, the faster fluctuations are described in a statistical sense correctly. This allows for a fast and flexible modeling of the converter, which can be virtually placed into any new wind location. When used together, the two approaches could be used to virtually transport the numerical design into any wind location, and find the best design for the best location.
... The wind force time series were calculated in a classic manner and applied as collinear to the wave and current actions. In a similar type of work Mücke et al. (2009Mücke et al. ( , 2010 have analysed high frequency wind speed time series measured at the GROWIAN site at the German North Sea coast. In comparison with the standard models used for load analysis of structures and wind turbines their models show a higher probability of extremes. ...
The evaluation of structural responses is key element in the design of ships and offshore structures. Fundamental to this is the determination of the design loads to support the Rule requirements and for application in direct calculations. To date, the current design philosophy for the prediction of motions and wave-induced loads has been driven by empirical or first-principles calculation procedures based on well-proven applications such as ship motion prediction programs. In recent years, the software, engineering and computer technology available to predict the design loads imposed on ships and offshore structures has improved dramatically. Notwithstanding, with the stepwise increase in the size and structural complexity of ships and floating offshore installations and the advances in the framework of Rules and Standards it has become necessary to utilise the latest technologies to assess the design loads on new designs. Along the lines of the recommendations from the International Ship and Offshore Structures Committee (ISSC) I.2 on Loads this paper reviews some of the recent advances in the assessment of loads for ships and offshore structures with the aim to draw the overall technological landscape available for further understanding, validation and implementation by the academic and industrial communities. Particular emphasis is attributed on methodologies applicable for the prediction of environmental and operational loads from waves, wind, current, ice, slamming, sloshing and operational factors. Consideration is also given to deterministic and statistical load predictions based on model experiments, full-scale measurements and theoretical methods.
Review report of the ISSC Technical Committee I.2 on Loads
Wind turbines are operating in the turbulent atmospheric boundary layer where they are exposed to wind speed changes in time and space on different scales [3]. This paper deals with a statistical description of the turbulent wind field on the one hand and with the interaction of turbulent wind fields with rotor blade elements with respect to dynamical changes of the lift coefficient c
l
on the other hand. In the first part we present a method to estimate the statistics of wind speed fluctuations u ′ = u – 〈u 〉 and more importantly of extreme events on different time scales based on 10-minute averaged values for the measured turbulence intensity of wind fields. In the second part we present measurement data that illustrates the effects of atmospheric like turbulence on a FX 79-W-151A airfoil under static angles of attacks.
Wind turbines are exposed to the turbulent nature of wind. The resulting dynamical loads affect the entire structure of the turbine. In order to characterize wind conditions, loads and electrical power output, standardized procedures have been developed for wind engineering practice. In most cases, those procedures are based on 10-minutes mean values and encompass the temporal dynamic on times scales of a few seconds only in a simplified manner. In this thesis, several aspects within the modelling of dynamical loads on wind turbines are analysed with respect to the effects of small-scale turbulence.
An important aspect of the conversion process is the electrical power output of a wind turbine. Power curves for wind turbines can be estimated from high-frequent data by applying the stochastic method of Langevin processes. Thereby, the drift and diffusion coefficients are describing the small-scale dynamics of the turbine on time scales within a few seconds. The dynamical conversation process of a wind turbine is taking place within this short time period on the basis of turbulent wind fluctuations. The deterministic dynamic specifies the stable fixed points of the process, which define the so-called Langevin power curve. So far, the functionality of the Langevin power curve was tested successfully by means of site measurements. The results appear reasonable, but have not been validated yet.
In this work, the Langevin power curve is estimated on the basis of numerical models. For this purpose, the electrical power output is calculated for a numerical wind turbine model with synthetic wind fields. Power curves could be quantified for laminar or turbulent flows. They could be compared to each other or with power
curves from a modified data sampling rate. We have investigated, which data basis is needed for a reliable Langevin power curve estimation. Our analyses have shown clear evidence that the quality of the stable fixed points is strongly affected by the sampling rate and the quantity of data. The turbulence intensity of the inflowing wind field has no significant influence on the stable fixed points of the Langevin power curve. However, in contradiction to classical approaches additional information can be achieved for high turbulent data.
Furthermore, it could be shown by defining the accuracy of the method that the Langevin power curve is able to detect failures during operation. For that purpose, two different failures are presented: a pitch failure during full load operation and a yaw error during partial load operation. These examples illustrate clearly that the
changing dynamics of the system lead to modified stable fixed points. Additionally, premonitions of a changing fixed point are given at an early stage.
Another important aspect is the modelling of synthetic wind fields for calculating the dynamic loads on wind turbines. The synthetic wind fields commonly used within the certification process of wind turbines obey Gaussian statistics. Actually, atmospheric wind fields are not Gaussian but highly intermittent. This implies the
existence of a higher quantity of extreme events on small time scales in contradiction to the standard wind field models. It is therefore a crucial point to identify to which extent this fundamental difference influences the dynamical loads of a wind turbine.
In our analysis, three different inflowing wind fields are used: atmospheric measurements of the GROWIAN Project, Gaussian synthetic wind fields (modelled with the Kaimal model) and intermittent synthetic wind fields (modelled with a new wind field generator based on the so-called CTRW model). The analyses have shown that the new wind field generator grasps the statistics of atmospheric velocity increments in a better way compared to the spectral Kaimal model. Thus, the CTRW model is a promising approach to model intermittent time series.
In this work, we point out that the intermittent characteristics of the atmospheric wind should also be taken into consideration for load calculations. We illustrate by means of the rotor torque that the higher quantity of extreme velocity increments leads to higher alternating loads on the drive train. Alternating loads on small time scales are crucial for a wind turbine, because an abrupt rise or fall in wind speed and thus load leads to higher strains of the drive train components as the same change over a longer time period. Our results suggest that Gaussian wind field models do not properly reproduce these higher alternating loads.
In order to calculate dynamic loads on wind turbines, the CTRW wind field generator should be capable to model the statistics of the atmospheric wind up to the two-point statistics of the second order. In this work, atmospheric wind fields (measured at the GROWIAN site) are analysed with respect to their statistical
properties. We show that the synthetic wind fields generated with the CTRW model even match the statistics of the atmospheric GROWIAN wind fields up to the two-point statistics of the forth order. In order to achieve this alignment, the parameters of the CTRW model should be identified in a specified order. The order presented here, is in general suitable to characterize arbitrary wind fields. Furthermore, our analysis provide evidence that more characteristics of the atmospheric wind should be taken into consideration for synthetic wind field modelling. Especially, sequences of increments and coherent instationarities could have a ignificant influence on both the dynamical interactions and loads of the individual components.
ResearchGate has not been able to resolve any references for this publication.