Dynamique basse fréquence et structures cohérentes autour du décrochage d'un profil mince

In the last years, particular emphasis was put on understanding the mechanisms behind stall on wings working at transitional Reynolds numbers (Re≈10^5) as is the case for unmanned aired vehicles, e.g., drones. In this regime, there is a melting pot of phenomena, among which we can find multistability and low-frequency oscillations. Despite the vast literature on these subjects, there is still a gap in the understanding of the mechanisms behind and the interaction between them. The objective of this work is to bridge this gap and to provide a clearer picture of the dynamics involved. To this end, we have conducted synchronised pressure-forced and particle image velocimetry measurements in a NACA0012 on a wind tunnel. Self-sustained bistable dynamics are observed over a broad range of Re and at angles slightly higher than the angle of stall. The system jumps intermittently between two states of high and low lift. We show that this process is random and memoryless. The mean residence time follows a super-exponential evolution typical of subcritical bifurcations. Flow topology and dynamics in each state are analyzed at different Re. Above a critical Re, the high lift state features a small laminar separation bubble, low fluctuations and no sign of low-frequency unsteadiness. Below this value, a longer bubble is present, accompanied by a low-frequency broadband peak with a Strouhal number one order of magnitude lower than the one for wake vortex shedding. On the other hand, the low lift state exhibits similar features for all Re tested, featuring intermittency between long bubbles and massive separation. Broadband low-frequency unsteadiness is also observed, with a Strouhal number close to the one found for the high lift state. Employing an extension of the Multiscale proper orthogonal decomposition, we associate this low-frequency energy with a shrinking-expansion of the bubble. Two stall scenarios are proposed depending on the Re regime. A first one where the transitions are triggered by the bubble breathing and another driven by background noise. Finally, a stochastic model based on the lift coefficient is developed. The model sucessfully reproduces the large-scale dynamics and statistics of the bistability. It also predicts the existence of an unstable branch linked by two saddle-node bifurcations with consequences on the stability of the high and low lift states.