The results of large-eddy and similar simulations and acoustic propagation calculations provided by six participants for the Rudimentary Landing Gear (RLG) category of the 2012 BANC-II workshop are analyzed and compared with experimental results. The general conclusion is that industrial accuracy remains elusive, and very significant differences exist between methods, so that the field needs more maturity, in addition to computing power. We present both blind comparisons from the workshop, and revised results the participants were invited to provide after having access to the experimental data. We recall that BANC-I and other comparisons indicate good code-to-code and experimental agreement for time-averaged pressures, but not so much for unsteady pressures and forces. Possibly as a result of these differences, the agreement for radiated noise is quite poor in the blind test, with typical differences of over 5 dB for the SPL and for the spectra over the dominant frequency range, and reaching 10 dB towards both ends of the spectra. The collection of numerical results brackets the experimental result. Some differences arise from misunderstandings, which were not corrected. The post-workshop results are much closer to experiment, with differences reduced to about 3 dB for SPL, and for spectra up to a Strouhal number of 15. This is not up to industrial needs, but is very encouraging and close to the accuracy achieved for jet noise. The results are sensitive to the use of solid or permeable surfaces in the Ffowcs Williams-Hawkings (FW-H) equation, and to numerous simulation details such as upwind-biased differencing.
展开▼
