Implementing and testing a panel-based method for modeling acoustic scattering from CFD input

S. Hales Swift, Purdue University


Exposure of sailors to high levels of noise in the aircraft carrier deck environment is a problem that has serious human and economic consequences. A variety of approaches to quieting exhausting jets from high-performance aircraft are undergoing development. However, testing of noise abatement solutions at full-scale may be prohibitively costly when many possible nozzle treatments are under consideration. A relatively efficient and accurate means of predicting the noise levels resulting from engine-quieting technologies at personnel locations is needed. This is complicated by the need to model both the direct and the scattered sound field in order to determine the resultant spectrum and levels. While the direct sound field may be obtained using CFD plus surface integral methods such as the Ffowcs-Williams Hawkings method, the scattered sound field is complicated by its dependence on the geometry of the scattering surface---the aircraft carrier deck, aircraft control surfaces and other nearby structures. In this work, a time-domain boundary element method, or TD-BEM, (sometimes referred to in terms of source panels) is proposed and developed that takes advantage of and offers beneficial effects for the substantial planar components of the aircraft carrier deck environment and uses pressure gradients as its input. This method is applied to and compared with analytical results for planar surfaces, corners and spherical surfaces using an analytic point source as input. The method can also accept input from CFD data on an acoustic data surface by using the G1A pressure gradient formulation to obtain pressure gradients on the surface from the flow variables contained on the acoustic data surface. The method is also applied to a planar scattering surface characteristic of an aircraft carrier flight deck with an acoustic data surface from a supersonic jet large eddy simulation, or LES, as input to the scattering model. In this way, the process for modeling the complete sound field (assuming the availability of an acoustic data surface from a time-realized numerical simulation of the jet flow field) is outlined for a realistic group of source location, scattering surface location and observer locations. The method was able to successfully model planar cases, corners and spheres with a level of error that is low enough for some engineering purposes. Significant benefits were realized for fully planar surfaces including high parallelizability and avoidance of interaction between portions of the paneled boundary. When the jet large eddy simulation case was considered the method was able to capture a substantial portion of the spectrum including the peak frequency region and a majority of the spectral energy with good fidelity.




Lyrintzis, Purdue University.

Subject Area

Applied Mathematics|Aerospace engineering|Acoustics

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server