Direct numerical simulation of sound generation in turbulent shear flows

Abstract

Direct numerical simulations of turbulent shear flows and their radiated acoustic fields have provided new insights into the mechanisms of sound generation, as well as the efficacy and limitations of the acoustic analogy approach. At present, these computations are expensive, and limited to relatively low Reynolds number and canonical flows. The computational approach consists of using high-order accurate numerical methods, together with accurate and robust nonreflecting and buffer-zone boundary conditions. This talk will focus on the results of computations of sound generated in mixing layers, jets, and open cavity flows. The relationship between linear stability waves in the flow and the radiated acoustic field is examined, in the context of a forced subsonic mixing layer and a fully turbulent supersonic round jet. For the subsonic flow, the acoustic sources can be modeled as wave packets, which radiate a superdirective acoustic field. For the supersonic jet, the directly computed (i.e., nonlinear source) acoustic field is compared to predictions based on radiating stability waves. Finally, the flow/acoustic instabilities in the subsonic flow over an open cavity are examined, and control strategies for reducing the internal acoustic load are explored. [Work supported by NSF (CTS-9501349) and AFOSR (F49620-98-1-0095).]