Jet and jet–wing noise modelling based on the CABARET MILES flow solver and the Ffowcs Williams–Hawkings method

Abstract

A co-axial subsonic unheated jet with and without a swept lifting wing at free-stream conditions from a recent jet–wing TsAGI experiment is considered. For computational modelling, Monotonically Integrated Large Eddy Simulations (MILES) are conducted based on the CABARET scheme which is implemented in a modern parallel unstructured-grid compressible Navier–Stokes computational code. The computational domain of the installed configuration includes a part of the nozzle and a wing section with round flow-fences present in the experiment to preclude the downwash effects. For the isolated jet, the same size of the computational domain is applied and two grid resolutions are considered to investigate the sensitivity of the current far-field noise predictions to the computational grid. The meanflow velocity profiles predicted downstream of the nozzle are compared with the flow data available. For far-field acoustic predictions, the Ffowcs Williams-Hawkings (FW-H) integral method is used. The Ffowcs Williams-Hawkings solutions correspond to a large closed permeable control surface with multiple closing disks at the outlet side set-up in accordance with the best practice to avoid pseudo sound in the acoustic modelling. The comparison of the acoustic predictions with the far-field spectra measured in the experiment for 30° and 90° observer angles to the jet flow are presented and discussed. It is shown that the current modelling robustly captures the same relative trends of the spectra behavior as observed in the experiment: while the presence of the wing does not lead to any significant change of sound spectrum in comparison with the isolated jet for 30°, there is a [Formula: see text] sound amplification due to the jet–wing interaction at 90° angle to the jet.