Worldwide regulations concerning noise emissions of road vehicles are constantly demanding further reductions of acoustic emissions, which are considered a major environmental health concern in several countries. Among the different sources contributing to noise generation in vehicles equipped with internal combustion engines, exhaust flow noise is one of the most significant, being generated by turbulence development in the exhaust gases, and robust and reliable numerical methodologies for its prediction in early design phases are currently still needed. To this extent, Computational Aero-Acoustics (CAA) can be considered a valuable approach to characterize the physical mechanisms leading to flow noise generation and its propagation, and it could therefore be used to support exhaust system development prior to the execution of experimental testing campaigns.
This paper describes the development of a CAA methodology suitable for automotive applications that can be used to support the design of new exhaust system components in their early phases. In particular, the work focuses on the flow noise generated in a complex heavy-duty exhaust system, featuring three tailpipes located next to the ground. Firstly, the near-field acoustic field is obtained with a Direct Noise Computation (DNC) approach from an unsteady compressible 3D Computational Fluid Dynamics (3D-CFD) simulation, carried out by means of the commercially available 3D-CFD software STAR-CCM+. A Detached Eddy Simulation (DES) technique is implemented to reduce the high computational cost of the DNC approach and it has been initialized by a steady-state converged solution. The steady-state simulation has been also exploited to extract qualitative predictive indexes, as a preliminary characterization of the system in terms of mean flow and broadband noise generation.
Finally, predicted near-field noise levels are evaluated to obtain an assessment of engine exhaust system acoustic performance.