G. Carte et al., NUMERICAL-SIMULATION OF THE MECHANISMS GOVERNING THE ONSET OF THE BENARD-VON-KARMAN INSTABILITY, International journal for numerical methods in fluids, 23(8), 1996, pp. 753-785
The onset of the Benard-von Karman instability consisting of the selec
tive amplification of the linear unstable mode and yielding finally th
e well-known saturated state has been described many times on the basi
s of both numerical and experimental results in various configurations
. However, neither the role of the harmonics and their coupling has be
en examined quantitatively, nor has the spatial structure of the insta
bility been studied in detail. A recently developed numerical method o
f simulation of quasi-periodic flows makes it possible to integrate th
e investigation of linear and non-linear characteristics within a sing
le numerical method, The simulation of the 2D afterbody wake presented
in this paper allows us to follow the amplification of the instabilit
y over many orders of magnitude. It is shown that at all stages of its
development the instability is characterized by a series of harmonics
, each of them amplified with a multiple of the fundamental amplificat
ion rate during the linear regime. The amplification of harmonics resu
lts from an energy transfer from the mean flow to harmonics of increas
ingly higher order. Ultimately the energy losses compensate this trans
fer and an equilibrium, commonly called saturation of the instability,
is reached. It is shown that the coupling between the fundamental har
monic and the mean flow is mainly responsible for the saturation. The
convergence rate of the development of the instability into harmonics
is investigated. A full description of the spatial structure of all si
gnificant harmonics both in the linear regime and at saturation is obt
ained. The results show that time and space characteristics of the ins
tability can be investigated simultaneously in an efficient way. Such
an approach might be particularly important in 3D wakes where the geom
etry has a strong influence on the behaviour of unstable flows.