Comparison of two methods simulating highly resolved atmospheric turbulence data for study of stall effects

Abstract

We compare two different methods that provide highly resolved three-dimensional turbulent wind fields for numerical investigations of stall effects. The first is computationally very expensive and explicitly simulates the turbulent wind fields using large-eddy simulation (LES). The second method generates synthetic three-dimensional turbulent wind fields from one-dimensional time series data from flights in the atmosphere. The synthetic method is comparatively fast and cheap but reproduces only statistical features of the turbulent flow.Since the focus in this study lies on the two methods by themselves, data generation is based on the same numerical simulation. The synthetic fields were generated from time series data obtained from virtual flight measurements within the LES. Different meteorological scenarios were analyzed in order to examine the influence of the different driving forces on the results.Horizontally averaged turbulence parameters of the compared fields are in good agreement. Parameters are independent of height in the synthetic flow fields since the time series used for the generation do not contain height information. In the case of a stably stratified boundary layer, the velocity fluctuations have a near-Gaussian distribution and are therefore well-reproduced by the synthetic method. Although provided with the time series, the synthetic flow fields cannot generate the non-Gaussian distribution of the vertical velocity in case of the analyzed convective boundary layers. Angles of attack of a virtual airplane calculated with the vertical velocity of wind fields generated with the two different methods show large differences. The consequences of these findings for applications will be investigated in a future study by numerical simulation of the flow around wings initialized with the different velocity fields.