Numerical Aeroacoustics - Towards the Prediction of Airframe Noise

The present paper discusses numerical approaches to predicting airframe noise, with emphasis on CAA (Computational AeroAcoustics) methods, which are based on the numerical solution of Euler's equations rather than on a dedicated wave equation of an acoustic analogy. First a distinction is made between turbulence noise and airframe noise, which is crucial for the understanding of the dominating source mechanisms and thus the numerical approaches to the problem. It is demonstrated that airframe noise is predominantly caused by the interaction of vorticity with geometric inhomogeneities (edges, slots, steps etc.) and partly by large structure hydrodynamic instabilities. Requirements are developed, necessary to be fulfilled in order to successfully describe airframe noise numerically. The three basically different numerical approaches to aeroacoustics (BEM/FEM based on wave eqn, DNS/LES, CAA perturbation methods) are discussed in view of their suitability, leading to the conclusion that important airframe noise phenomena are not tractable using analogy approaches, but at least require CAA methods, based on Euler's eqn. With regard to the above conclusions, DLR's cooperative CAA-research initiative "SWING" (Simulation of Wing-flow Noise Generation), involving several German universities, is briefly outlined. First results of trailing edge noise CAA-simulationsare discussed, showing very good agreement between numerical solutions and theory/experiment. Especially the Mach-number scaling and the dependence on vorticity strength of the generated sound is looked at in detail.