GPU CABARET Flow and Noise Solutions of an Installed Jet Configuration

摘要

The GPU-CABARET method in the Wall Modelling LES (WMLES) version has been systematically tested on the NASA wall-mounted hump problem and then applied for flow and noise simulations corresponding to the SYMPHONY jet installation noise experiment. The jet installation case corresponds to a complete jet-pylon-wing-flap configuration including the fuselage body. For far-field noise predictions, the Ffowcs Williams - Hawkings method based on multiple penetrable control surfaces is used. In comparison with the conventional isolated jet noise problem, it is shown that the design of acoustic control surfaces for the installed case needs a special consideration in order to include all important jet-pylon-flap-interaction noise sources and needs additional averaging across several control surfaces to remove the pseudo sound effects. The complete LES model of the jet installation noise includes 243 Million cells and is run for 8.5 days on 4 high-end GPU cards (2 home-office-style workstations) to obtain 600 convective time units of the signal time. The obtained acoustic predictions are compared with the data for the same jet installation noise case digitized from the literature.