Thesis title: Numerical tools for high-fidelity simulation of shock/boundary layer interactions
Shock-wave/turbulent-boundary-layer interactions (SBLIs) are a
characteristic phenomenon in numerous high-speed aerodynamic flows.
They are frequently encountered in external flows, including transonic and
supersonic airfoils, aerodynamic appendages, wing-body junctions,
as well as in internal flows, like supersonic engine inlets, compressors and turbines.
More broadly, SBLIs occur whenever a shock wave encounters a
turbulent boundary layer developing on a solid surface.
The impact of a shock on a boundary layer typically results
in substantial flow separation, which can lead to significant decreases in performance.
Efficiency loss of aerodynamic surfaces, structural
vibrations induced by wall pressure fluctuations,
and localized heat transfer peaks are common examples of
undesired effects stemming from SBLIs.
Consequently, the study and the understanding of this flow phenomenon
has gained a lot of interest in the aeronautic/aerospace research field.
This work concentrates on both the numerical and physical aspects of SBLIs.
In the first part, we elucidate the numerical framework utilized to
simulate these interactions using Direct Numerical Simulations (DNS),
both in Cartesian and curvilinear coordinates.
Special emphasis has been placed on the handling of convective terms,
which have been reformulated into a convenient split form ensuring
the discrete preservation of total kinetic energy.
Subsequently, various DNS of canonical flows are performed,
including turbulent boundary layers,
oblique SBLIs and compression corners.
Our primary focus is on addressing technical aspects for
state-of-the-art numerical simulations.
This encompasses the adaptation of classical synthetic
turbulence prescription at the inflow of the computational
domain, and the implementation of appropriate grid stretching
functions in the wall-normal direction.
Another significant aspect is the analysis of low-frequency
unsteadiness in SBLI configurations involving the
presence of cross-flow.
Notably, we show that, when the boundary layer meets the shock
with a non-zero sweep angle, the typical low-frequency unsteadiness can
shift to higher tones following a specific relationship
dependent on the sweep angle and on large-scale structures
at the foot of the reflected shock.
In conclusion, we present the validation results of the new code in
curvilinear coordinates for a supersonic compression corner,
along with the ongoing research projects that incorporate it.