A coupled LES/high-order acoustic method for jet noise analysis

Abstract

Aircraft noise is one of the main areas of active research for the aeronautical industry due to the increasingly stringent regulations on noise emission that aviation authorities are imposing. Among the different sources that contribute to the total emitted aircraft noise, jet noise is one of the most important during take-off. Furthermore, as the by-pass ratio of turbofan engines is increased, the interaction of the jet exhaust with the high-lift devices and the wing can potentially produce new mechanisms for noise generation. On the simulation front, the rapid increase of computing power over the last decades is enabling the use of high-fidelity simulations for the study of jet noise at both industrial and academic research levels. However, most of the numerical methods used by different research groups are either too dissipative for propagating the acoustic waves or are limited to the study of simple configurations. In many cases, surface integral methods have been the preferred choice with encouraging results for isolated jet configurations. Among these methods, the Ffowcs Williams-Hawkings (FWH) formulation has been commonly applied within research communities. However, applying them in complex configurations can be challenging, which may not provide sufficient information when it comes to studying noise generation mechanisms. The work reported in this thesis is devoted to the development of a coupling framework that is suitable for complex jet noise propagation cases. In this framework, the jet noise problem is divided into two different steps. First, the acoustic sources are computed using a robust compressible Large Eddy Simulation (LES) finite volume solver, which are then transferred to a spectral/hp high-order finite element Acoustic Perturbation Equations (APE) solver that propagates the sound waves to the far-field. Two different coupling strategies are investigated. Initially, a simple methodology based on the exchange of files between the solvers is implemented with only minor modifications made to the solvers’ source code. However, the poor efficiency of data transfer meant this method is applicable only to small problems. Thus, a more efficient parallel-interface coupling technique is developed to overcome this issue. With this technique all the required data is transferred via a parallel Message Passing Interface (MPI), avoiding the bottleneck of I/O and file systems. Both coupling techniques are validated with a 2-D cylinder case demonstrating the superiority of the parallel interface method. The parallel interface coupling framework is then tested on a low Reynolds number jet, being validated against experimental and numerical results in the literature, during which a well-established FWH method is used for references. More promising results are obtained using the LES/APE method than with the FWH method. The LES/APE method is then applied to the study of a more realistic isolated jet case and is compared to the experimental data obtained at NASA. A source analysis is further carried out, in this case, to reveal the distribution and convection of sources along the jet plume at different locations. The source distribution is in good agreement with the far-field noise results. Finally, the study of a jet-flat plate installed configuration is conducted. This simplified configuration is representative of a realistic installation scenario and is particularly useful to the understanding of the installation effects. The coupling framework captures these additional flow-acoustic effects demonstrating its potential to tackle complex configurations.