Experimental convective heat transfer in a geometrically large two-dimensional impinging synthetic jet

Abstract

A jet can be synthesized by ejecting and injecting fluid from and to an orifice or channel. When actuated, a synthetic jet delivers positive momentum but no net mass flow per cycle. Small, compact synthetic jet actuators can be fabricated to operate in the subaudible acoustic range and can be packaged in orientations that allow them to deliver cooling air flow to electronic devices. For cooling, the most promising orientation is one that delivers the jet flow in a direction normal to the heated surface such that it impinges on the surface as a periodic jet. In previous studies, numerical simulations were performed by the authors, utilizing an idealized canonical geometry, with the goal of eliminating actuator artifacts from the fundamental physics that drive the problem. The present paper reports on laboratory experiments that were undertaken in order to nearly replicate the idealized synthetic jet geometry and thus allow validation of the previous numerical investigations. The experiment was designed at a geometrically large scale, with jet slot widths of 4.2 and 6 mm. The amplitude and frequency at which the jet was actuated determined the Reynolds (Re) and Womersley (Ω) numbers which are the dominant non-dimensional groups. By maintaining Re and Ω in the laboratory experiments to match those of the small scale actuators, the laboratory experiments were geometrically scaled up to allow highly resolved measurements of the unsteady velocity field and the local time-dependent Nusselt number on the target heated surface. Experiments were performed at variable jet Re, Ω, and height from the target surface. The dependence of the surface averaged Nu to jet parameters agrees with the computational results when the convective heat transfer is significantly larger than the experimental heat losses, and when the fluid remains coherent with respect to the forcing frequency. Discrepancies found between numerical and empirical local data at large spacings, high Ω and low Re, suggest flow transition to turbulence and possible emergence of three dimensional effects, which were not accounted for in the computational analyses.