Posters

Abstract

Annual energy production (AEP) assessments for individual wind farms are commonly based on a spatially homogeneous wind climate and engineering wake models. However, when extending the domain to large offshore clusters affected by external wakes, mesoscale flow variability can become a dominant source of uncertainty. This raises a practical question: how much inflow information is actually needed to obtain a reliable AEP estimate at cluster scale? We investigate a weather-regime based framework that makes use of “heterogeneous” mesoscale information (from reanalysis and the New European Wind Atlas) to drive fast engineering wake simulations. Dimensional-reduction techniques (e.g., self-organizing maps and empirical orthogonal functions) are used to identify a small set of representative weather regimes. The classification combines large-scale variables (e.g., mean sea-level pressure and pressure anomalies) with boundary-layer indicators relevant for wake development, such as atmospheric stability and shear. For each regime, representative conditions are selected and used as inflow to compute AEP for offshore wind farm clusters in the North Sea. We compare AEP estimates based on this reduced set of regimes against reference estimates derived from the full mesoscale dataset, quantifying the trade-off between computational savings and added uncertainty. The results highlight conditions under which mesoscale variability must be explicitly represented, and where simpler inflow descriptions remain sufficient. We discuss the implications and limitations of the approach for design studies and potential co-optimization of large wind farm clusters.