On using large-eddy simulation for the prediction of noise from cold and heated turbulent jets

The results of a series of large-eddy simulations of heated and unheated jets using approximately 106 grid points are presented. The computations were performed on jets at operating conditions originally investigated by Tanna in the late 1970s [H. K. Tanna, "An experimental study of jet noise Part I: Turbulent mixing noise," J. Sound Vib., 50, 405 (1977)]. Three acoustic Mach numbers are investigated (U_j/a_∞=0.5, 0.9, and 1.5) at cold (constant stagnation temperature) and heated conditions (T_j /T_∞=1.8, 2.7, and 2.3, respectively). The jets' initial annular shear layers are thick relative to experimental jets and are quasi-laminar with superimposed disturbances from linear instability theory. It is observed that qualitative changes in the jets' mean- and turbulent field structure with U_j and T_j are consistent with previous experimental data. However, the jets exhibit a faster centerline mean velocity decay rate relative to the existing data, with a corresponding 3-4 % over-prediction of the peak root-mean-square level. The acoustic pressure fluctuations in the far field are analyzed in detail. The accuracy of the overall sound pressure level predictions is found to be a strong function of the jet Mach number, with the lowest speed jets being the least accurate. At all conditions the peak acoustic frequency occurs at approximately St=fD_j/U_j=0.25. The limited resolution of the computations is shown to impact the radiated sound by yielding effectively low-pass filtered versions of the experimental spectra, with a maximum frequency of St ≈ 1.2.