Compressible Large-Eddy Simulation of Separation Control on a Wall-Mounted Hump

Abstract

Compressible large-eddy simulations of turbulent flow over a wall-mounted hump with active flow control are performed and compared with previous experiments. The flow is characterized by the unsteady separation before the steep trailing edge, which naturally reattaches downstream of the hump to form an unsteady turbulent separation bubble. The low Mach number large-eddy simulation demonstrated a good prediction of surface pressure coefficient, separation-bubble length, and velocity profiles compared with experiments. The effect of compressibility on the baseline flow is documented and analyzed and is found to increase the separation-bubble size, due to a reduced growth rate. Control is applied just before the natural separation point via steady suction and zero-net-mass-flux oscillatory forcing, and steady suction is shown to be more effective in decreasing the size of the separation bubble and pressure drag for the control parameters investigated. Controlled flow at a compressible subsonic Mach number is applied, and found to be slightly less effective than the same control parameters at low Mach numbers.

Modelling bubble clusters in compressible liquids

“…For the simulations contained in this work the model has been implemented in a base compressible flow code because it was readily available, but the model may be coupled to different types of flow solvers. In our case, the three-dimensional averaged equations (3.1)-(3.2) are discretized in space and solved for the density and velocity fields using a skew-symmetric formulation in generalized orthogonal coordinates (Honein & Moin 2004;Mattsson & Nordström 2004;Franck & Colonius 2010). The method uses high-order-accurate finite-difference methods in the streamwise and wall-normal directions, and a Fourier method for derivatives in the spanwise direction.…”

Supersonic flow control

“…Passive and active (by suction) techniques for the control of shock wave oscillations and separated boundary layer interaction in a transonic channel flow were investigated by Galli et al [23] who showed that passive control had no effect on shock unsteadiness. LES simulations of Franck and Colonius [24] confirmed a slight loss of control effectiveness when comparing the results for the same control parameters at low Mach numbers.…”

“…The research status in this field also indicates that the vane-type VG shapes are superior in performance when compared to the ramp/wedge-type VGs and, therefore, more innovative designs need to be looked into to match the effectiveness of the former. Although nothing concrete could be ascertained about control location, a rough estimate indicates that this distance is to be somewhere between (15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) x/ h (where x is the distance of the control location from the interaction).…”

Stabilized non-dissipative approximations of Euler equations in generalized curvilinear coordinates

“…Here the baseline formulation, that includes the total energy equation, is retained. It is also interesting to observe that a similar (but not equivalent) stabilization technique was proposed by Franck and Colonius [16]. Those authors started from the Cartesian split form of the equations (6), and replaced the Cartesian derivative operators with their strong conservation form in curvilinear coordinates, which leads to the following split convective form @qu @t þ 1 2…”

“…(14). Although the discretization of the split form (23) yields superior stability with respect to the unsplit convective form [16], it does does not guarantee discrete energy preservation from convection, whereas (14) does.…”