Assessing Acoustic Source Forcing Tools for Launch Vehicle Jet Noise Prediction

Jets associated with launch vehicles during the initial stages of the lift-off emit very strong Mach and acoustic waves that may interact with ground structures or with the launch vehicle structure itself. The prediction of the noise associated with these strong jets poses several challenges when compared to traditional jet noise modeling techniques: first, the acoustic and Mach waves propagating from the jet may reach high amplitudes, placing the waves in the nonlinear regime; and secondly, the jet is not embedded in free space as in traditional jet noise prediction approaches, but in the vicinity of obstructions that can be placed either in the nearor far-field. Therefore, classical acoustic analogy approaches or linearized Euler equations with source terms may be inappropriate here. The direct noise computation using either direct numerical simulations or large eddy simulations is impractical for realistic configurations, especially when the Reynolds number of the flow is high. In this work, we propose a nonlinear hybrid approach, wherein the full Euler equations are solved in the nearand far-field acoustic regions, and Navier-Stokes equations (eventually, with equations for species) are solved in the jet region to predict the acoustic source. The acoustic source on the right hand side of Euler equations are imposed using a penalization technique, where density, momentum and energy are interpolated from the Navier-Stokes flow domain (appropriate grid resolutions are utilized to discretize each flow domain). The approach is first tested on simple cases involving single-wavenumber acoustic and vorticity waves in both linear and nonlinear regimes. The effectiveness of the penalization technique for these simple cases is demonstrated. Then, a two-dimensional jet is considered in three flow regimes (high subsonic, low supersonic and high supersonic). The method is proven to be effective and accurate in terms of predicted sound pressure level spectra (direct noise computation is used to validate the results).