A method of accelerating the convergence of computational fluid dynamics for micro-siting wind mapping

Abstract

To assess wind resources, a number of simulations should be performed by wind direction, wind speed, and atmospheric stability bins to conduct micro-siting using computational fluid dynamics (CFD). This study proposes a method of accelerating CFD convergence by generating initions that are closer to the converged solution. In addition, the study proposes the 'mirrored initial condition' (IC) using the symmetry of wind direction and geography, the 'composed IC' using the vector composition principle, and the 'shifted IC' which assumes that the wind speed vectors are similar in conditions characterized by minute differences in wind direction as the well-posed initial conditions. They provided a significantly closer approximation to the converged flow field than did the conventional initial condition, which simply assumed a homogenous atmospheric boundary layer over the entire simulation domain. The results of this study show that the computation time taken for micro-siting can be shortened by around 35% when conducting CFD with 16 wind direction sectors by mixing the conventional and the proposed ICs properly.