Evaluation of Turbulence Models for the Prediction of Wind Turbine Aerodynamics

TLDR

The study aims to develop transition models and conduct grid sensitivity analyses in the tip region to better match experimental data. A third‑order spatial, second‑order temporal, implicit Navier–Stokes solver was used to simulate the NREL Phase VI wind turbine across 7–25 m/s, testing Baldwin‑Lomax, Spalart‑Allmaras, and k‑ε turbulence models with and without wall corrections. All models captured normal forces and bending moments accurately, yet struggled with chordwise forces, power output, and pitching moments; the k‑ε model with near‑wall corrections performed best overall.

Abstract

The performance of the NREL Phase VI horizontal axis wind turbine has been studied with a 3-D unsteady Navier-Stokes solver. This solver is third order accurate in space and second order accurate in time, and uses an implicit time marching scheme. Calculations were done for a range of wind conditions from 7 m/s to 25 m/s where the flow conditions ranged from attached flow to massively separated flow. A variety of turbulence models were studied: Baldwin-Lomax Model, Spalart-Allmaras one-equation model, and k-ε two equations model with and without wall corrections. It was found all the models predicted the normal forces and associated bending moments well, but most of them had difficulties in modeling the chord wise forces, power generation, and pitching moments. It was found that the k-ε model with near wall corrections did the best job of predicting most the quantities with acceptable levels of accuracy. Additional studies aimed at transition model development, and grid sensitivity studies in the tip region are deemed necessary to improve the correlation with experiments.