Posters

Abstract

Wake modelling is a key step in the prediction of energy in pre-construction energy yield assessments. To date, most of the wind energy industry still consider wake losses, such as turbine availability, and other secondary losses to be independent. This research demonstrates a second order interdependence between wind farm availability and wake losses, finding that this effect can be significant for cases with high wake losses, and should be considered when undertaking energy yield assessments. This research utilises a time-based energy prediction model quantify differences in wake losses when considering an idealised case, where turbines are always operating, and a coupled model which integrates operating state into the wake modelling approach. Availability is synthesised on a per turbine level, using a Monte Carlo method to randomly sprinkle availability events to achieve a time-based availability target.  A search space was defined to investigate this second order effect for a range of wind farm types. Key parameters investigated included; wind farm size (number of turbines), layout density, availability rates and wake deficit models. High-performance computing resources were leveraged to scale up the analysis, enabling thousands of simulations to isolate the key drivers of this second order effect.  The results from this study show an increase of 0.2% to 2.5% in predicted annual energy production of a wind farm, based on the search space defined. This highlights the growing importance of considering the interdependence of losses, as wind farms increase in size and density in future build out scenarios. Furthermore, the study emphasises the need to review how this effect is included in industry validation and whether current practises introduce bias.