Posters

Abstract

The modelling of wake losses inside and between wind farms is one of the most important elements of uncertainty, when modelling the energy yield of offshore wind farms. Accurate models are required for assessment, optimised design, and wind farm control. WakeBlaster SaaS has been continuously developed since 2016 and is a parabolic 3-D RANS model of the waked flow inside and between wind farms [Bradstock20]. WakeBlaster is a field model of the flow of multiple interacting turbines, which is designed to assess wind farm yield and losses as a function of turbine properties, turbine instance operation, variable wind resource, and atmospheric conditions. Here, we report on a set of recent improvements made to the model, and the validation of the new model modelling the wake of full-scale offshore wind farms. Modelling the yield of wind farms is a process which often suffers from a lack of detailed knowledge regarding all the input parameters required. For example, only one input is available for turbulence, atmospheric stability, and roughness, and the other two parameters are either assumed by default or derived. Data may also be given as statistical averages or binned as a function of wind speed and/or wind direction. The original WakeBlaster approach for input data was to assume default values for missing information, e.g., neutral stability, if no information on stability was provided. Often, turbulence is provided as a function of wind direction only, with neutral stability assumed. However, this will lead to a bias towards underestimating wake losses, due to the non-linear relationship between stability and wake strength. The new approach recognises that turbulence and stability inputs are not independent of each other. The missing turbulence and stability inputs are set by using complementary information as a baseline. The purpose behind this modification is to improve the representation of wake losses in wind farm scenarios with incomplete input data. The original WakeBlaster model, which assumed neutral stability, was validated against data from ten different offshore wind farms and ten onshore wind farms. An updated validation - with more finely categorised validation data - is required, in order to show how modified model inputs perform, relative to the original approach. We report the results of the validation against recently published data from Lidar wind speed measurements near offshore wind farms in the North Sea. [Bradstock2020] P. Bradstock and W. Schlez: Theory and Verification of a new 3D RANS Wake Model. Wind Energy Science, pp. 1425-1434, 2020. https://wes.copernicus.org/articles/5/1425/2020/