Mixing layer: Numerical and experimental control strategies

This study focuses on a numerical investigation of a mixing layer vis-a-vis an experimental study using dielectric barrier discharge (DBD) plasma actuators.¹ Upstream perturbations from the experiment^1-3 are imposed as inflow boundary condition in the direct numerical simulation (DNS) of the mixing layer using a computational methodology reported recently.^4;5 Development of the mixing layer is followed computationally through the initial stages of coalescence of vortex structures. This collaborative investigation is unique and has shed some light on the mechanism of the coalescence process of the coherent structures in the mixing layer. For example, in experiments,^1;2 it was observed that AC-DBD plasma actuation using high voltage (15kV) 30 Hz 50 per cent duty cycle pulses produced more energetic growth of the mixing layer than a pure sine mode. In the DNS carried out here, it is shown that the growth of the mixing layer corresponding to the inflow conditions provided by this experiment appears to be characterized by a new type of interaction (rather than a binomial process of pairing), where a paired and a nascent rolled-up structure coalesce. This “joint interaction” does not appear to be the “collective interaction” that has been observed experimentally6 or computationally.^4;5 This joint interaction leads to an enhanced growth of the mixing layer in comparison to a pairing^4;5 and provides a physical reason for the experimental observation^1-3 that the growth in this case is more energetic. The purpose of this effort is to enable a collaborative platform to identify various deterministic startegies for controlling the growth of a free shear layer or mixing layer, which are found in a variety of engineering applications. Vorticity thicknesss and momentum thickness evolution in the mixing layer are predicted, the latter compared with the experimental results.^1-3 DNS predictions agree reasonably well with experiments, but some differences are apparent. In essence, the effect of the wind tunnel on the mixing layer behavior is not captured in the DNS since no-slip boundary conditions, freestream turbulence and effect of the diffuser are not enforced. Comparison of Reynolds shear stress at various streamwise locations indicates local regions in which turbulence transfers energy back to the mean flow, a reverse cascade energy transfer mechanism, as observed experimentally.^1-3 This study is unique and has potential for furthering this collaborative platform for the purpose of mixing layer flow control.