Computational modelling for wind energy assessment

A brief history of computational modelling for wind resource assessment is given along with a description of the methodology within which modelling of this type is used in the wind industry. With increased computational power, models used within this methodology have moved from being linear to nonlinear and the treatment of surface boundary conditions have moved from simple roughness length parameterisations to distributed drag formulations in which momentum is absorbed through a finite depth at the lower surface. Comparison with field and wind tunnel measurements show that operation in this expanded parameter space leads to more accurate estimates of wind resource. While improvements in accuracy have resulted, a number of challenges still exist in modelling scales of turbulent motion that are near those of the main topographic features, for example, eddies shed in the lee of steep topographic features. Large eddy simulations (LES) of flows over hills with and without forest canopies show a strong coupling between intermittently separated flow in the lee of a hill with perturbations in the outer-layer flow. This points to likely difficulties in modelling the effects of these motions in Reynolds averaged Navier–Stokes (RANS) models, in which there is no mechanism for representing this type of connection between the inner and outer layers of the flow.