Causal analysis of turbulent Couette-Poiseuille flow using wavelet-based resolvent analysis

Flow separation is a ubiquitous phenomenon in many turbulent fluid flows and is well-studied in statistically-stationary, spatially-evolving settings. In this study, we investigate transient separation using turbulent Couette-Poiseuille flow subjected to a sudden strong adverse pressure gradient. By analyzing the time-varying mean shear and the change of the wall-normal velocity fluctuations, we observe the occurrence of reverse flow and wall-normal separation events in time. Employing wavelet-based resolvent analysis with the time-varying mean, we identify spatiotemporal forcing modes that are optimally amplified by the linearized Navier-Stokes operator, along with their corresponding response modes. Time windowed analysis localized around the separation event identifies forcing and response modes separated by a temporal delay, allowing for the causal mechanisms associated with this event to be isolated and characterized.