Presentations

Abstract

Validation studies of large eddy simulation (LES) based modelling of operational wind farms often focus on a limited set of idealized flow cases, which do not account for e.g. temporal and spatial atmospheric heterogeneities over the site. Furthermore, by limiting the analysis on a narrow wind speed and wind direction range, most of the available observed wind farm production data is discarded from the analysis. These factors limit the general applicability and extendibility of the validation results to all operational regimes and weather conditions. Instead of constructing flow cases a priori, we propose a time-concurrent simulation approach, where the simulated wind farm is exposed to a best estimate of the local, time and spatially varying atmospheric conditions by using a meso-scale coupled LES-based weather model. Besides providing a physics-based representation of blockage and wakes, the actuator disk-based wind turbine parametrization in the LES provides time series of all relevant turbine parameters, including electrical power, wind speed and wind direction. A one-to-one comparison of this data can then be made with observations obtained from the Supervisory Control And Data Acquisition (SCADA) system of actual wind turbines. Specific wind speed and atmospheric stability regimes can then be selected from the simulated and observed wind farm data in a consistent way, ensuring that the underlying distributions of all atmospheric parameters is matched. Within the context of this study, multi-year simulations were performed for 8 operational wind farms. The simulations are validated in terms of bias and correlation in key metrics like average wind farm production, pattern of production and array efficiency. Furthermore, the key metrics are conditioned on specific atmospheric conditions to deepen the understanding of the validity of a real weather, LES-based modelling approach in different wind speed and atmospheric stability regimes.