Posters

Abstract

For a modelled 3.8 GW offshore wind cluster, with similar characteristics to the Danish Energy Island concept, three main points are presented. First, the concern that ever-increasing cluster size may lead to every-increasing gravity wave induced annual energy production (AEP) losses is not supported by the results. Over a year, the net effect of the gravity-wave induced pressure gradient through the wind farm is fairly small. Surprisingly, the effect is greater for a smaller (~1 GW) array. Increasing effect with increasing power density will be demonstrated, but again with modest negative impact on AEP, in all case less than 1%. Current understanding of this behaviour based on relative horizontal and vertical scales will be presented. Second, it is observed that on an instantaneous (rather than yearly average) basis the effect of gravity waves (excited by interaction of the wind farm with the Atmospheric Boundary Layer (ABL) top) on power can be strongly positive or negative, with variations of up to 10% power between turbines at the same instant - an effect which is separate from wake effects. This strong effect on pattern of production necessitates inclusion of gravity wave modelling when designing wind farm control methods, and presents considerable opportunity to increase energy output in large wind farm clusters. Third, estimation of cluster long range wake effects using the present model will be presented. The results include effect on potential AEP at 5 to 100 km distant from the cluster, including sensitivity to cluster spacing. The present method includes sensitivity to ABL height and friction velocity / Monin-Obhukov length and is an alternative to the TurboPark [1] method thus allowing an ensemble modelling approach to cluster wakes predictions. Input data for the method has been derived from NEWA [2] and requires geostrophic wind speeds, which are not supplied by NEWA and must be reconstructed with one or other set of assumptions. This is a relative weakness of the method as the long range wake effects are fairly sensitive to these assumptions as will be demonstrated. This exposes the relatively poor ABL data available, an issue which ought to be addressed. The rapid method used is based on Smith's linearised boundary layer solver [3] combined with a standard wake effects model. The wake effects model provides the in-farm wake losses and the turbine thrust input needed by the gravity wave solver, which in turn provides both the gravity wave feedback and the long-range wake behaviour [4]. The new formulation used here which allows their simple and robust coupling will be described. This allows the flow gradients set up by the pressure field from wind farm / ABL interaction to be treated as terrain-induced 'speed-ups' are on land. The results presented are based on the application of this model to a year-long time series of 30 minute data from NEWA. [1] N G Nygaard, Wind Europe Technology Workshop 2023. [2] New European Wind Atlas, www.neweuropeanwindatlas.eu. [3] R B Smith, Wind Energy, 2010. [4] B J Gribben and N Adams, PO.089 Wind Europe Technology Workshop 2023.