Bayesian optimisation of wall blowing for drag reduction of a spatially evolving turbulent boundary layer
File(s)
Author(s)
Mahfoze, Omar Ahmed Reda Ali
Type
Thesis or dissertation
Abstract
It is estimated that 50% of the fuel burnt by a commercial airliner is associated with the skin-friction drag of the turbulent boundary layers on the airliner surfaces. Wall-blowing applied in a zero pressure gradient turbulent boundary layer (TBL) through a permeable surface is a simple active flow control strategy that can result in a considerable skin-friction drag reduction (DR). However, little is known about its full potential of achieving energy savings, which is reduction of the sum of the energies needed to move the main flow and to operate the control strategy. Thus, the aim of this thesis is to improve the efficiency of a wall-blowing strategy and hence achieve energy savings by reducing the skin-friction drag in a TBL.
In this work, high-order numerical methods are used to conduct direct numerical simulations (DNS) and large eddy simulations (LES) of TBL reaching up to Re = 4300 (the Reynolds number is based on the momentum thickness and free stream velocity), and controlled by wall-blowing. Different LES approaches are examined with different mesh resolutions for turbulent channel flows and TBL. It is found that an implicit LES approach based on the numerical dissipation associated with the discretisation of the diffusion terms achieves a good compromise between cost and accuracy.
The simulations are coupled with a Bayesian optimisation framework to maximise the energy savings by optimising the amplitude and the streamwise coverage of the wall-blowing strategy. The control performance is assessed by using experimental measurements of two different types of blowing devices for the estimation of the energy savings. For a moving vehicle, it is assumed that the blowing air can be provided from other systems, avoiding the additional drag associated with drawing the ambient air. It is shown that wall-blowing with blowing amplitude less than 1% of the free stream velocity can achieve a significant DR of 95% and energy savings of 7% due to the long-lasting DR downstream of the control region.
In this work, high-order numerical methods are used to conduct direct numerical simulations (DNS) and large eddy simulations (LES) of TBL reaching up to Re = 4300 (the Reynolds number is based on the momentum thickness and free stream velocity), and controlled by wall-blowing. Different LES approaches are examined with different mesh resolutions for turbulent channel flows and TBL. It is found that an implicit LES approach based on the numerical dissipation associated with the discretisation of the diffusion terms achieves a good compromise between cost and accuracy.
The simulations are coupled with a Bayesian optimisation framework to maximise the energy savings by optimising the amplitude and the streamwise coverage of the wall-blowing strategy. The control performance is assessed by using experimental measurements of two different types of blowing devices for the estimation of the energy savings. For a moving vehicle, it is assumed that the blowing air can be provided from other systems, avoiding the additional drag associated with drawing the ambient air. It is shown that wall-blowing with blowing amplitude less than 1% of the free stream velocity can achieve a significant DR of 95% and energy savings of 7% due to the long-lasting DR downstream of the control region.
Version
Open Access
Date Issued
2021-07
Online Publication Date
2022-04-29T10:55:58Z
Date Awarded
2022-02
Copyright Statement
Creative Commons Attribution NonCommercial Licence
Advisor
Laizet, Sylvain
Sponsor
Imperial College London
UK Research and Innovation
Engineering and Physical Sciences Research Council
European Union
Grant Number
EP/R029326/1
2016163847
2018184381
2019215138)
Publisher Department
Aeronautics
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)