LES-deduced acoustic modelling for jet noise

The ability to predict jet noise is crucial for the aviation industry, particularly in order to meet its commitments to reduce aviation noise by 65% according to ACARE 2050 targets. Therefore, in order to proceed with this objective, a comprehensive methodology for the prediction of noise and, most importantly, jet noise is required. As part of this need to predict jet noise is the ability to predict jet noise at high frequency, for example with the target of a Strouhal number (St) of 10. Aeroacoustic methods reliant on the coupling of an eddy-resolving method with an acoustic extrapolation, such as the Ffowcs Williams-Hawkings (FW-H) method, need extensive mesh resolution to resolve high-frequency acoustic pressure fluctuations in the near field. An alternative to the eddy-resolving method is the turbulent statistic acoustic modelling approach that is typically coupled with a solution from a Reynolds Averaged Navier-Stokes (RANS) simulation. Investigations, however, have found that RANS alone is unable to provide suitable turbulent statistics for more complex jets such as serrated jets. As jet nozzle design continues to increase in complexity it is important to be able to have a design methodology to robustly predict turbulent statistics for a variety of jets. Further to the limitation of turbulence modelling, RANS is unable to provide unsteady data necessary to correctly determine the coefficients required for acoustic modelling. Therefore a method of using statistics from LES is pursued to inform the acoustic model. Whilst it is simple to obtain most statistics from LES, the calculation of dissipation rate (ε) from LES is not straightforward. A potentially promising method whereby the ε turbulent transport equation using mean flow field data from an LES solution.

This thesis therefore pursues the use of the ε informed by an LES solution in order to correctly calculate the corresponding ε in order to use turbulent statistics from LES to inform acoustic modelling. Further to this, two methods of acoustic coefficients determination shall be pursued in order to create a self-contained methodology for acoustic modelling informed by LES.

Between experimental and LES turbulent statistics there has been good agreement which is demonstrated. Further to this, the ability of the ε turbulent transport equation is demonstrated for jets with profiles using the transport equation matching directly calculated profiles. The use of two- point velocity statistics has been successfully used to determine acoustic coefficients for noise modelling. Finally, modelling results have compared favourable to experimental far-field noise data for both round and serrated jets showing superior high-frequency noise predictive capabilities to FW-H thereby demonstrating the advantages of the acoustic modelling approach.