Studies on adverse-pressure-gradient turbulent boundary layers on wings
Time: Fri 2020-02-28 10.15
Location: Seminarierum Faxén (rum 5316), KTH Mekanik, Teknikringen 8, KTH, Stockholm (English)
Subject area: Engineering Mechanics
Doctoral student: Alvaro Tanarro , Teknisk mekanik
Opponent: Dr Omid Omidyeganeh, City University of London
Supervisor: Philipp Schlatter, Linné Flow Center, FLOW, Mekanik, SeRC - Swedish e-Science Research Centre; Ricardo Vinuesa, Linné Flow Center, FLOW, SeRC - Swedish e-Science Research Centre
The present licentiate thesis addresses the use of well-resolved simulations to simulate turbulent boundary layers (TBL) subjected to adverse pressure gradients. Within the thesis a wide variety of analyses are performed, and a method to improve the performance of the simulations is presented. The first aim of the thesis is to assess the effect of adverse pressure gradients and flow history on the development and fundamental characteristics of turbulent boundary layers. With this in mind, well-resolved large-eddy simulations (LES) of the turbulent boundary layers over two wing sections are performed using the spectral-element-method (SEM) code Nek5000. In order to assess the effects of the adverse pressure gradient on turbulent boundary layers, turbulence statistics are computed and time series are collected from the simulations. The turbulence statistics show a significant effect of the adverse pressure gradient on the mean velocity profiles, turbulent fluctuations and turbulent kinetic energy budgets. In addition, the time series are used to compute the power-spectral densities of the turbulent boundary layers and to analyse the effect of the adverse pressure gradient on the turbulent scales across the boundary layer. After having compared both wings at moderate Reynolds number Rec=400,000, the next goal is to perform high-resolution simulations of wings at higher Reynolds numbers in order to study conditions closer to those in reality, and to evaluate the effect of adverse pressure gradient with increasing Reynolds numbers. To achieve this, better and more efficient computational methods are required. In this thesis, the performance of the adaptive mesh refinement method recently implemented in Nek5000 is assessed for the first time on wing simulations. The obtained results show a large potential of this new method (which includes the use of non-conformal meshes) with respect to the previous simulations carried out with conformal meshes. Lastly, we performed a modal decomposition of the TBLs developing around both wing sections. To this end, we consider spectral proper orthogonal decomposition (SPOD), which can be used to identify the most energetic structures of the turbulent boundary layer.