The 33 rd International Congress and Exposition on Noise Control Engineering 1/1 Validation of a Wave Based Technique for the analysis of a multi-domain 3D acoustic Cavity with interior damping and loudspeaker excitation A. Hepberger a , B. Pluymers b , K. Jalics a , H.-H. Priebsch a and W. Desmet b a ACC Acoustic Competence Center G.m.B.H., Inffeldgasse 25, A-8010 Graz, Austria b K.U.Leuven, Department of Mechanical Engineering, Celestijnenlaan 300B, B-3001 Heverlee, Belgium a achim.hepberger@accgraz.com;karoly.jalics@accgraz.com; hans-herwig.priebsch@accgraz.com; b wim.desmet@mech.kuleuven.ac.be;bert.pluymers@mech.kuleuven.ac.be Abstract [821] The Wave Based Technique (WBT) is a novel prediction technique for steady-state acoustic analysis, which is based on the indirect Trefftz approach. Due to its computational efficiency, it is an attractive alternative for the widely accepted Finite Element Method (FEM), which is practically restricted to low-frequency applications. The recently developed WBT methodology allows to consider arbitrary 3-dimensional fluid domains with prescribed normal velocity and prescribed normal impedance boundary conditions. The methodology requires that the acoustic problem domains are convex. Non-convex domains have to be decomposed into convex sub-domains. At the arising interfaces between the sub-domains, specific coupling conditions are imposed. For validation purposes, a dedicated loudspeaker-driven car-like cavity (Sound Brick) has been constructed. The results of a WBT model of this cavity are compared with standard FE results as well as with measurements. It is demonstrated that the WBT is computationally more efficient than the FEM, so that the practical frequency limitation for the WBT can be shifted towards the mid-frequency range. 1 INTRODUCTION In order to reduce development time and costs, application of numerical prediction techniques has become common practice in the automotive industry. Among the wide range of simulation applications, prediction of the vehicle interior noise is still one of the most challenging ones. Especially since, at this moment, no prediction technique is readily available to tackle these problems in the entire audio frequency range of interest. The Finite Element Method (FEM) is well known for acoustic predictions in the low-frequency range. Automotive acoustic FE applications are common up to about 150-200 Hz and research work is on-going to provide accurate models for higher frequencies. The high-frequency limitation of the FEM is due to the fact that model sizes and subsequent computational efforts increase significantly with increasing frequency. Statistical Energy Analysis (SEA) has been introduced for car interior noise predictions at higher frequencies. The applicability of the SEA is limited to high frequencies (> 400 Hz for automotive applications) since the technique requires a high modal overlap. The novel Wave Based Technique (WBT) has been developed as an alternative method, that aims at covering a frequency range up to 1kHz for automotive interior noise simulations and, in this way, at