New analytical wake models based on artificial intelligence and rivalling the benchmark full-rotor CFD predictions under both uniform and ABL inflows

Abstract

New analytical wake models are derived from the soft computing technique, called Genetic Programming (GP) to predict wake velocities and turbulence intensity. The design of the wind farm’s appropriate layout is essential for minimizing cost and maximizing the wind farm power generation. This needs a precise wake velocity model to simulate the wake effect of the wind farm within a limited time duration. Furthermore, prediction of turbulence in the wake due to ambient flow and rotor-generated is extremely crucial owing to its contribution to fatigue loads and structural failures of the downstream wind turbines. This article discusses the classical to the recent analytical wake velocity and turbulence intensity models derived based on hard computing techniques in detail and their limitations. The significant constraints are the consideration of uniform inflow without integrating Atmospheric Boundary Layer (ABL) impacts for the forecast of wake velocity and estimation of single value of turbulence intensity while it radially varies at distinct downstream distances of the wind turbine. Eventually, these constraints are tackled and new analytical models for wake velocity and turbulence intensity profiles are formulated for both uniform and ABL inflows. The existing and proposed models are compared with the previous NREL Phase VI wind turbine CFD study for uniform and ABL inflows and it was observed that the proposed models are precise.