Jet Noise Prediction Using Turbulent Scales from LES and RANS

This paper studies the modelling of turbulent scales used in an existing steady Reynolds averaged Navier–Stokes solution-based acoustic analogy. The turbulence in the flow has been described as a statistical model of the two-point cross-correlation of the velocity fluctuations, characterized by the turbulent length and time scales. The modelling of the turbulent length and time scales from the \(\overline{K} {-} \overline{\varepsilon }\) data used in the steady RANS-based acoustic analogy has been validated with those computed from the cross-correlation of the velocity fluctuations. This was pursued with an LES database comprising an isothermal and a heated ideally-expanded Mach 1.5 round jets. The far-field noise has been computed using the turbulent scales from both the cross-correlation data and the \(\overline{K} {-} \overline{\varepsilon }\) data. The two results agree very well, and also display reasonable match with direct predictions from the time-resolved LES data using the Ffowcs Williams-Hawkings method.