Aeroacoustic Reduced-Order Models Based on a priori/posteriori Data Analysis

Two Reduced-Order Models for the simulation of aeroacoustic phenomena are presented. One model is based on Proper Orthogonal Decomposition, relying on a posteriori analysis of a set of high-delity calculations, the snapshots in the parametric space. To partially reduce the number of necessary snapshots, a self-adaptive sampling technique is proposed. Instead of a regular and homogeneous one-step sampling, a gradual and optimised enhancement of the sample set is dened. The second model is based on Proper Generalized Decomposition, which can be viewed as a a-priori approach, able to build the reduced-order approximation without relying on the knowledge of the solution of the complete problem, assessing the model accuracy and, if necessary, to enrich the reduced approximation basis. Both methods are applied to the computation of scattering of sound by a circular cylinder.