Numerical investigation of scale resolved turbulence models (LES, ELES and DDES) in the assessment of wind effects on supertall structures

Abstract

Scale resolved turbulence models, such as large eddy simulation (LES), have been increasingly utilised in applications of computational wind engineering (CWE) for tall buildings, due to their proven numerical accuracy. The demand for computational time and resources in LES is still a concern in practical applications of CWE, where it is necessary to explore alternative modelling techniques. The main objective of this study is to critically investigate the possibility of using a computationally efficient hybrid RANS/LES approach, specifically embedded-LES (ELES), to predict wind effects on a supertall 406 m slender tower of non-standard geometry. Numerical predictions of LES and ELES are compared and critically analysed with experimental results obtained by performing boundary layer wind tunnel (BLWT) tests. It was shown that the choice of sub-grid scale (SGS) modelling in LES has an impact on the wind effects predicted on the structure. Out of the five SGS models tested, the wall adapting local eddy viscosity (WALE) model was the most suitable due to its treatment of the near wall region. An optimum domain configuration based on a parametric study was proposed for the application of ELES. The importance of overall mesh quality and its impact on fluctuating pressures and dynamic properties of wind is presented. Wind flow behaviour on the structure, such as Von-Karman vortex shedding and its dynamic implications, is also discussed. Finally, using the proposed modelling techniques, ELES can generate similar numerical accuracy to that of an LES model at a lower computational cost.