Skip to main content

High-performance computing and wall-bounded turbulence

Time: Tue 2022-02-22 11.15

Video link: Zoom presentation

Lecturer: Ramón Pozuelo

Adverse pressure gradients (APGs) are present in many technical applications
and under some circumstances, can lead to a decrease of performance such as
detachment of the flow in airfoils. To have a better understanding of turbulent
flows at high Reynolds numbers and the effects of the APG, different experi-
mental set-ups were carried out on a simpler case, flat plate turbulent boundary
layers (TBLs) with a near-equilibrium APG. In this case, we set up a numerical
experiment, with a range of Reynolds numbers and APG intensity comparable
to wind tunnel experiments. The simulation of APG TBLs increases the compu-
tational resources needed. On one side, to solve the different variety of turbulent
scales, a very thin grid and short timesteps are needed, requiring distributed
parallel computing to accelerate the calculations and be able to accommodate
a model with a large number of grid points. The higher the Reynolds number,
the wider is the range of spatial scales and the adverse pressure gradient accel-
erates the growth of the boundary layer, together with the size of the domain
necessary to simulate the boundary layer and the free stream. We will explore
the set-p of such big simulations, the challenges and results together with the
techniques used for postprocessing.

This talk is a ACML seminar. People intrested in the thematic area Advanced Computing and Machine Leaning can contact Ricardo Vinuesa (rvinuesa@mech.kth.se) and Artem Kulachenko (artem@kth.se).

Page responsible:Ardeshir Hanifi
Belongs to: FLOW
Last changed: Feb 09, 2022