The work carried out at the Department Engine Acoustics is concerned with the finale objective of reducing community noise, i.e. noise perceived by people in the vicinity of airports. The modeling of engine-noise radiation is not straightforward as many effects influence the way sound travels from the source to the far-field observer: acoustic waves are potentially reflected at duct exits, diffracted at intake and nozzle lips, refracted by mean flow, scattered by turbulence, etc. The current topics addressed by AT-TRA in this field are briefly presented below.
Numerical & analytical activities
In computational simulations one intends to divide complex physical phenomena into smaller simplified problems. For each one of these a sub-domain with limited boundaries is identified. The objective of this splitting is to save computation time as CFD continues to present a severe challenge for unsteady flow (acoustics being a small part of it). The information is passed from one sub-domain onto the next one using appropriate coupling strategies. The source region or near-field region defines the sub-domain where aerodynamic sound is generated. In this region non-linearities and viscosity effects often play an important role therefore the Navier-Stokes equations (also simplified with the use of turbulence models) are solved. Beyond to a certain distance to the source, non-linearities and viscosity effects can be neglected. A new sub-domain can be identified within which the linearised Euler equations (LEE) are solved. These equations enable to account for the effect of mean flow gradients and complex geometry on the propagation. Practically, LEE equations are solved in the geometric near-field. Finally an integral method - like Ffowcs Williams and Hawkings (FW-H) or Kirchhoff – can be employed to project the results into the far field at the observer location. Such integral methods do not require the use of a mesh in the space between the integration surface and the observer. As drawbacks they cannot account for complex flow features and interaction with body surfaces.
In the last year our attention was focused on the development of a special formulation of FW-H derived in order to best post-process TRACE output results. This formulation is valid for porous surfaces surrounded by a homogeneous flow. The convective problem is solved. The solution works in both time and frequency domains. Using an azimuthal decomposition the surface integral is transformed into a line integral which can be efficiently solved numerically. The use of FW-H will be often debated. The application of this method to open rotors has shown that far-field acoustic levels may strongly depend on the position of the FW-H surface because the convection of vorticity through the integration surface generates spurious noise.
Our work also deals with the use of ray theory and CAA simulation to calculate the sound field radiated from engine intake. Previous results have showed that flow gradients and intake contour shape are two effects which have a significant impact on far-field directivity.
Measurements with far-field microphone arrays (Arraymesstechnik) help characterize the directivity of sound generated by engines. Should sensors be installed at the source, e.g. inside the combustor, a correlation or transfer function between source and far field can be established, on which basis for instance transmission and haystacking effects can be quantified.