Presentations

Abstract

The United Kingdom has set an ambitious goal of achieving up to 50 GW of installed offshore wind capacity by 2030. The Engineering Wake Models (EWMs) that are widely used for Annual Energy Production (AEP) predictions have typically been developed based on in-farm losses, with limited analysis to-date for the much larger spacings of inter-farm wakes. To facilitate timely deployment of the large wind farms now planned it is important to assess such effects, both for their impact on planned farms, and on farms already in operation. This urgency motivates the present study which focuses on quantifying dependency of inter-farm wake losses on the EWM adopted for a range of farm design parameters. PyWake is employed to explore the effects of EWM choice, wind direction, power density (i.e., varying turbine spacing), farm capacity, and farm aspect ratio across four different idealised layout configurations. The first configuration consists of two 1 GW square arrays (Farm A and Farm B), similar to those used in a study by Frazer-Nash (2023) [1]. Additionally, three alternative configurations are analysed, where Farm B is replaced by smaller farms with capacities of 250 MW, 490 MW, and 240 MW, respectively. These smaller farms are intended to represent existing farms and allow for the assessment of inter-farm losses from the perspective of incumbent operators. The results highlight the range of predictions obtained with alternative EWMs. Of the 7 models evaluated, the TurbOGauss model consistently provides the most conservative predictions. For the two 1 GW farms, inter-farm wake losses are estimated to range from 0.5% to 1.5% at an inter-farm distance of 6 km and wake losses exceeding 1% at 10 km. These figures are highly dependent on farm configuration. Existing farms with installed capacities of 250 to 490 MW could experience inter-farm AEP losses of 2% to 3% when a 1 GW farm is located 10 km upstream (in the main wind heading) and over 1% AEP loss beyond 20 km. The study also highlights that these values are likely to vary based on the specific characteristics of both farms. For example, existing farms with a high aspect ratio are more negatively impacted if a greater proportion of turbines face the upstream farm’s wake. Additionally, higher power densities — resulting from closer turbine spacing — are likely to produce prolonged farm wakes. While internal wakes are expected to represent a larger proportion of losses than inter-farm wakes for gigawatt-scale wind farms, inter-farm wake losses may be particularly significant for farms of smaller capacity, and with a large proportion of turbines in the leading row. Finally, critical discussion is provided, drawing on comparisons to higher-fidelity modelling tools, regarding the limitations of current EWMs for inter-farm wake models, and to inform further improvements to model prediction accuracy for the gigawatt-scale wind farms now under development. [1] E. Lindsay, ‘Offshore Wind Leasing Programme - Array Layout Yield Study,’ Frazer-Nash, Tech. Rep. 55726R, Oct. 2023.