Modelling a sound radiated by a turbulent jet

Noise standards around the world for aircrafts have become more stringent. Jet noise is a major source of noise from the aircraft particularly during takeoff and landing. Therefore it is important to investigate the nature of the jet noise and to be able to predict the noise level using numerical approaches. A substational amount of research work has been performed on numerical investigations of noise. The majority of these may be split into two main approaches: the semi-empirical source models, based on the steady RANS computations, which provides information about turbulent length and time scales which is translated by empirical relations into sound-source spectra, and the' direct' numerical simulations of the acoustic sources coupled with the analytical transport techniques, namely the Ffowcs Williams - Hawkings (FW-H) equation. The first approach is more specific and case dependent, however it is computationally fast, since it requires only 2D steady RANS simulations. The second approach is basically case independent, however it is very time-consuming since it requires unsteady numerical solutions for the flow field. The FW-H approach is well developed and widely validated when coupled with the Large Eddy Simulation (LES). However we have found that there are no investigations of the FW -H acoustic model coupled with the unsteady RANS simulations. It is widely accepted that the LES simulation results are usually more detailed than the RANS, but it is still not known to that extent the LES approach is more (or not more) reliable than the URANS for the jet noise prediction purposes. In addition we have discovered that despite the FW-H model being well developed in the commercial CFD software, such as FLUENT, there is no published data on the application of the FW - H model in FLUENT to the jet noise problems. ' This research is focused on the validation of the jet acoustic models that exist in FLUENT with the available experimental data. We employ three sets of experimental data, obtained by different research groups, to investigate the turbulence model approach for the source noise modelling and the acoustic model for the simulation of the noise level in the far field. For simplicity, we decided to concentrate our numerical investigations on the jets of relatively low Mach number, up to 0.6, and the flow issuing from the nozzle of a simple geometry, i.e. without chevrons. It should be noted that so far FLUENT includes two acoustical models which are applicable to jet noise problems, namely the FW-H model and the so-called Goldstein acoustic model, which is effectively one of the numerous variations of the semi empirical source models. The latter model has some difficulties in the current state of realization in FLUENT, and therefore we have reproduced the same approach in a separate software (MatLab) for the post-processing of the acoustics. In the future the results of this research will be useful for industrial applications to a wide range of jet noise problems.