Vortex dynamics-driven heat transfer and flow regime assessment in a turbulent impinging synthetic jet

摘要

Synthetic jets gained attention in the last decade as a thermal management solution, especially in the electronic cooling community. Under certain conditions, they can remove heat more efficiently than conventional steady jets. In this work we aim to further the knowledge on the fundamental behavior of such flows and their role in heat transfer enhancement. A numerical canonical geometry was developed to de-couple the impinging flow from possible artifacts, such as actuator and geometry resonance. The unsteady flow was assumed in turbulent regime, which was approximated via the Finite Volume Method through the software ANSYS Fluent (TM). The turbulence was modeled using the SST k-omega model, which accurately agreed with experimental data. Synthetic jets generate a train of counter-rotating vortices that, when impinged onto a stationary wall, give rise to secondary vortices that cause a colder fluid downwash into the heated zone. We proposed an alternative definition of the Reynolds number (Re-Gamma) that characterizes the strength of the generated vortices, consequently representing the strength of the jet. We found that to increase the jet thermal efficiency: (1) Having close consecutive vortices is as important as producing strong vortices, and (2) the jet-to-surface distance should be modified such that the vortex finds its peak intensity nearest to the heated wall. Compared to the classic definition of Re-L0 the Stroke Length based Reynolds number (Re-L0) appears as a more suitable definition to establish flow regimes, with the data suggesting Re-L0 approximate to 10, 000 as a threshold where the flow fully transitioned to turbulence. Silva-Llanca et al. (2015) proposed three hypotheses to explain some disagreement found in their data: Turbulent flow at large Re-L0, significant heat losses in the experiments at low Re-L0 and unaccounted three-dimensional effects. We proved that the first two explained their entire data disagreement, thus rendering the third hypothesis unnecessary. (C) 2017 Elsevier Masson SAS. All rights reserved.