CERFACS CERFACS Scientific Activity Report Jan. 2010 – Dec. 2011

2 Simulations of aircraft wakes and rocket
launch emissions
2.1 Wake simulation in the vortex regime (R. Paoli, J. Picot, O.
Thouron, D. Kolomenskiy)
2.1.1 Interactions between jets and wing vortices (D. Kolomenskiy, R. Paoli)
Wing-tip vortices have a major influence on the dynamics of condensation trails. Numerical simulation of
their initial development requires an adequate approximation to boundary layers on the wing surfaces. On
the other hand, important hydrodynamic instabilities occur in the wake on a much larger scale. Therefore, it
seems reasonable to decompose the simulation in an upstream domain comprising the solid boundaries, and
a downstream domain that only contains the vortex wake. In the upstream domain, a RANS computation
is performed. Its outputs are then used as inflow conditions for an LES computation in the downstream
domain.
−2.2
−2.4
−2.6
RANS, full domain
RANS inflow + LES of the wake
Chow et al., experiment
Craft et al., numerical simulation
−2.8
c p
−3
−3.2
−3.4
−3.6
−3.8
−0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
x/c
FIG. 2.1: Left : minimum static pressure coefficient along the vortex centreline. Right : instantaneous
crossflow velocity field (colour) and particles location (dots).
A wind-tunnel experiment with a NACA 0012 wing (Chow et al., 1997) has been used to test these
techniques. It was found, in agreement with previous studies (Craft et al., 2006), that RANS computations
may overestimate the rate of decay of the trailing vortex. The choice of turbulence model is crucial.
Satisfactory results have been obtained with a differential Reynolds stress model in the very near field.
Downstream, an LES computation supplied with the required inflow condition can provide an accurate
prediction of the vortex wake properties. This is shown in figure 2.1(left), which shows the minimum static
pressure coefficient in the vortex core as a function of the non-dimensional distance from the trailing edge.
CERFACS ACTIVITY REPORT 181