Estimation of integral length scales across the neutral atmospheric boundary layer depth: A Large Eddy Simulation study

Abstract

Abstract Integral length scales of atmospheric boundary layer (ABL) flows, as well as mean velocities and turbulence intensities, play an important role in determining the response of structures subjected to wind loads. For tall building design, however, sufficient relevant data at high elevations are currently not available from experimental/field measurement data or in building standards. This study uses Large Eddy Simulation (LES) with a constant horizontal pressure gradient and streamwise periodic boundary conditions to develop horizontally homogeneous neutral ABL flow. Based on two-point velocity correlation functions ( ρ u x , ρ u y , ρ u z , ρ v y , ρ w z ), the corresponding integral length scales ( L u x , L u y , L u z , L v y , L w z ) are presented at heights across the ABL depth for better characterization of the energetic turbulent eddies. In particular, this study addresses numerical artifacts inherent in the CFD simulations, which are due to insufficient longitudinal domain lengths and streamwise periodic boundary conditions, frequently used in computational studies, that affect the estimation of longitudinal correlation function in the streamwise direction, ρ u x , and the corresponding integral length scale, L u x . An analytical expression for ρ u x is proposed and calibrated using simulations with various longitudinal domain lengths, to obtain domain-independent L u x estimates across the ABL depth. The results clearly show that the turbulent eddy structures are highly elongated (streamwise) in the near-ground region, and become shorter in the outer region of the ABL. Estimates of the integral length scales are provided at heights much greater than the upper limit of standards like ASCE 7–16. This study provides an insight into the three-dimensional coherent structures of flows throughout the ABL depth and presents data on modeling approach flows in wind tunnel tests, in numerical simulations, and in associated standard provisions.