Posters

Abstract

The Uncertainty is one of the important issues when modelling the wind conditions and energy yields of wind turbines. It is a combination of modelling errors (e.g. due to the finite model resolution), model assumptions (e.g. turbulence, stability) and the uncertainty of the input data (terrain data, land use, wind distribution, etc.). Consequently, any type of model is characterized by a certain scope of quality and validity. In terms of linear models (e.g. WAsP), results of systematic investigations regarding the uncertainty are already available, whereby the site assessment by CFD-Modelling was not investigated in detail so far. The joint research project Sunday aims to fill this gap in the uncertainty of CFD-Modelling. In particular, quantitative knowledge about the dimension of each source of error and about the individual causes will be gained. The sensitivities of CFD simulations with respect to the computational grid, land use data, wind distribution as well as thermal stratification and turbulence in the atmospheric boundary layer are to be quantified and integrated as a partial aspect into the overall uncertainty of the yield calculation. A classification system for site data with respect to their complexity for CFD applications will be developed. The focus of all studies is on the uncertainty of CFD simulations as an integrative part of the overall uncertainty of the site assessment for wind farms with respect to long-term wind conditions represented by frequency distribution. For the investigated sites, the stability classes, the grid resolution and the discretization of the wind direction distribution, were varied in the CFD simulations based on the specified values from the simulation matrix. For the complex sites, a much stronger dependence is observed with respect to the horizontal grid resolution. For three internal sites, different simulations were tested with the coupling of mesoscale grid cells at different vertical levels. Especially the upper vertical levels (300 m above ground) lead to better results in the vertical profile of wind speed. The final results of the energy yield show relatively good values for the investigated sites when coupled with mesoscale grid cells. The decisive factor here is the number of grid cells used to drive the CFD model. For this purpose, different simulations were performed and analyzed.