Share this page on:

Conference programme 

Back to the programme printer.gif Print

Poster session

Lead Session Chair:
Stephan Barth, Managing Director, ForWind - Center for Wind Energy Research, Germany
Maxime Philippe INNOSEA, France
Magdalena Maché (1) F P
(1) INNOSEA, Nantes, France

Printer friendly version: printer.gif Print

Download poster(0.62 MB)

Presenter's biography

Biographies are supplied directly by presenters at OFFSHORE 2015 and are published here unedited

Maxime Phillipe is R&D manager at INNOSEA, an independent engineering firm specialized in Marine Renewable Energies. He obtained his PhD, working on the numerical simulation of floating wind turbines.


Aero-elastic study with turbulent wind and wake effect: comparison between large-eddy simulation (les) and classical statistical method conform to standards


Recent guidelines for certification of offshore wind turbines recommend the use a modified Frandsen model (GL 2012, IEC 64100-1) to compute the turbulence intensity for a wind turbine inside a wind farm. The wind is then computed by a statistical model. To get the moments and forces exerted on the turbine an aero-elastic tool is necessary. Design studies use those statistical models that can provide an important number of instantaneous wind fields in a short time. The aim of this study is to evaluate the error that is committed compared to studies with physical turbulence input calculated using LES.


In this work we compare the statistical wind generator TurbSim with the model SOWFA (both developed by the National Renewable Energy Laboratory). SOWFA couples directly instantaneous wind, obtained by a Large-Eddy Simulation based on OpenFOAM, with an aero-elastic tool. Two turbines were placed in the simulation domain. The wake effect from the first turbine on the second one is computed with both methods and then compared. For both turbines 2 statistical wind-field time series were generated with TurbSim: Firstly the turbulence intensity given by the standards was introduced to TurbSim, secondly the local turbulence intensity from SOWFA was used.

Main body of abstract

The LES part of SOWFA provides instantaneous wind fields and turbulence. The wakes created by wind turbines are represented by using an actuator line model. The reaction force exerted on the flow is computed at each blade point and consider rotation of the blades and a flow dependent rotor speed.
The flow field time series with realistic turbulence computed by the Atmospheric Boundary Layer Solver are introduced at each time step in the aero-elastic simulator FAST that computes the turbine’s response. The deformation of the turbines blades computed by FAST is then injected again to the actuator line model of the Atmospheric Boundary Layer Solver.
The instantaneous wind field obtained with TurbSim is computed with the Kaimal model that is recommended by international standards. By means of a fast fourier transformation time series are computed from a theoretical energy spectrum depending on the mean wind speed, the ambient turbulent intensity and an integral scale parameter.
Taking into account the background of these models it is understandable that – even at the same mean wind speed and turbulence intensity – the distribution of the turbulence is different in both cases. To evaluate the influence of this different turbulence organization on the turbine’s response we compare the minima, maxima, mean value, standard deviation of the turbine loads and displacements. We also compare parameters describing the distribution of the fluctuations from the mean (skewness, flatness and energy spectra). Moreover a DEL comparison is shown.


This work shows that the turbine’s response gives differences of 30 % already for the mean values of the pitch angle when the wind is obtained by a Large-Eddy Simulation instead of a statistic tool. This is true even for the comparison between TurbSim and SOWFA at the same turbulence intensity. Moreover we show the parameters sensible to change in turbulence and the reasons for this sensibility. The spatial distribution of the turbulent wind computed from TurbSim has less physical sense than in SOWFA and is responsible for the differences in the FAST results.

Learning objectives
Wind turbines are exposed to complex phenomena of turbulence, especially in offshore conditions. The publication of this work is an important contribution to the understanding of the turbine’s sensibility on turbulence with more physical sense than that generated by statistical tools. A future study of the nature of turbulence and the correlation between velocity and the turbine’s response will put light in this organization and its influence on the turbine’s response.