Performance Evaluation of Far-Field Noise Prediction Models for High-Lift Devices Using Reynolds-Averaged Navier-Stokes Simulations

This computational aeroacoustics (CAA) study is a performance evaluation of Amiet's leading edge turbulence interaction noise and trailing edge self-noise models modified for high-lift devices (HLDs). The chosen reference geometry was a three-element McDonnell-Douglas 30P30N airfoil at a freestream Mach number of 0.14, a chord-based Reynolds number of 1,350,000, and at a geometric angle of attack of 3 degrees. The predicted measurements were compared with the surrogate experimental measurements taken from the University of Toronto Institute for Aerospace Studies (UTIAS) Hybrid Anechoic Wind Tunnel (HAWT). The far-field spectral predictions confirmed that the main element leading edge is the most prominent contributor to the broadband signal. For the two-element configuration (i.e., with the slat removed), the turbulence interaction noise stemming from the flap leading edge shows good agreement from low- to mid-frequencies and defines the overall spectral shape of the two-element prediction. A sensitivity study of the source term data extraction locations for the leading edge model was investigated, where it was found that collecting input data in the shear layers yields high uncertainty in the far-field noise spectra. Conversely, input variables extracted in the cove separation regions are much less sensitive to the extraction position. Lastly, far-field directivity maps generated from observers located 3.75 chord lengths from the airfoil mid-chord suggests that the main element leading edge contribution is most significant across all observer locations.