The aim of this project is to improve the wind farm performance in the operational phase by accurate predictions of wind turbine production and loading. Accurate wind characteristics on the turbine scale is the key challange, it will be determined by inte grating mesoscale and microscale models. Recent improvements in both mesoscale meteorological models and microscale models based on Computational Fluid Dynamics (CFD) make a tighter integration possible. The improvements includes better description of th e flow stratification and turbulence, along with better compatibility of the boundary conditions - the difference in scales between meso and micro models has simply been reduced during the years. In a forecast situation the accuracy of a CFD simulation w ill heavily depend on the input from the mesoscale. In areas with complex terrain the accuracy is expected to be of a poor quality. In such cases it is necessary to recognize the salient processes underling the forecast error and improve the accuracy of t he mesoscale model before running the microscale fluid dynamics model. To further improve on the accuracy machine learning techniques will be used. A new mesoscale-microscale coupling model is proposed: trained on historical observations, the model uses mesoscale forecast output to issue a high quality site-specific forecast which will be used to scale the CFD modelling of the flow at the site. The outcome will be wind farm simulations more representative of local terrain conditions than usual idealized problems considering a wider range of stratification regimes and their individual effects accounted for in wind farm optimization procedures.
Project leader: Catherine Meissner
Institution: WINDSIM AS