Comparison of LES and Stochastic Source Generation Methods for Aero- and Hydro-Acoustic Design Guidance

Abstract

This paper presents a comparison of two hybrid Computational Aero-Acoustics (CAA) / Computational Hydro-Acoustics (CHA) approaches for the prediction of noise in air/water. The first method uses the Finite Element Method (FEM) to solve Lighthill’s equation in the frequency domain where acoustic sources (Lighthill tensor) are supplied from Large Eddy Simulations (LES). The second method uses the Discontinuous Galerkin Method (DG) to solve acoustic perturbation equations in the time domain where acoustic source terms are generated using the Fast Random Particle Mesh (FRPM) method, with turbulent source information from Reynolds-Averaged Navier-Stokes (RANS) simulations. The FRPM-DG method in the time domain can suffer from computational efficiency issues for very low Mach number cases which are seen in maritime applications. Simple algorithmic modifications for the time-domain FRPM-DG method are suggested to make it more attractive for these very low Mach number water flows. The current work demonstrates that the FRPM-DG method can provide satisfactory noise results at a fraction of the cost of LESFEM based methods. This is especially attractive for design studies and providing noise minimisation design guidance.