Numerical prediction of flow-induced noise in a simple cavity using random vibroacoustics and structural-acoustic reciprocity

This study presents a hybrid numerical framework for predicting flow-induced interior noise in cavity-backed panels excited by turbulent boundary layer wall-pressure fluctuations. The excitation model integrates the Goody empirical power spectral density for spectral fidelity and the Corcos spatial coherence function, calibrated using direct numerical simulation-informed parameters, to capture realistic broadband and partially coherent surface excitations. Structural responses are computed using finite element modeling, while acoustic responses are evaluated using reciprocity-based noise transfer functions, allowing efficient calculation of interior sound pressure levels (SPLs) under distributed excitations. A comparative assessment is conducted using three flow-induced noise source models: partially correlated (Corcos-based), non-correlated, and perfectly correlated across flow speeds from 5 to 60 m/s. Results show that the Corcos-based model provides the most physically consistent SPL predictions, accurately capturing resonance amplification, spectral broadening, and coherence-dependent energy transfer, in contrast to the non-correlated model (which underestimates SPL) and the perfectly correlated model (which overpredicts due to artificial phase alignment). The proposed framework is validated against benchmark trends and prior studies and offers a robust, scalable tool for vibroacoustic prediction in aerospace, automotive, and marine engineering applications.