Turbulent trailing edge noise estimation using a RANS-based statistical noise model

Abstract

The prediction capability of a RANS-based Statistical Noise Model (RSNM) is assessed by applying it to a sharpedged symmetric flat strut at a Reynolds number of Re = 500,000 and zero angle of attack. The method uses a rigidhalf- plane Green's function to calculate the far-field spectra generated by the turbulence in the vicinity of the trailing edge. Because the exact form of the turbulent sources is in general not available, a model of the turbulent velocity cross-spectrum must be assumed. The cross-spectrum model must accurately reflect the frequency and phase distribution in the boundary layer close to the trailing edge, but is otherwise arbitrary. The model used in this study is Gaussian, containing velocity, time and length scales that are defined in terms of the turbulence statistics obtained from a steady RANS solution. The predicted sound spectrum shows the correct slope and levels for frequencies ranging from 300 Hz to 2 kHz.