Predicting the broadband vibroacoustic response of systems subject to aeroacoustic loads by a Krylov subspace reduction

• Published in: Applied acoustics Vol. 74; no. 12; pp. 1394 - 1405
• Main Authors: Chronopoulos, D., Ichchou, M., Troclet, B., Bareille, O.
• Format: Journal Article
• Language: English
• Published: Kidlington Elsevier Ltd 01.12.2013; Elsevier
• Subjects: Acoustics; Aeroacoustics, atmospheric sound; Composite structures; Exact sciences and technology; FEM/SEA hybridization; Fluid–structure interaction; Fundamental areas of phenomenology (including applications); Krylov-subspace methods; Physics; Random excitations; Structural acoustics and vibration; Fluid–structure interaction; Composite structures; Random excitations; Krylov-subspace methods; FEM/SEA hybridization; Vibration; Fluid―structure interaction; Composite construction; Fluid structure interaction; Aeroacoustics; Structural acoustics
• ISSN: 0003-682X; 1872-910X
• DOI: 10.1016/j.apacoust.2013.04.006

•We reduce a fully coupled structural–acoustic system, under aeroacoustic excitations.•An SEA-like approach is applied and the acoustic response inside the cavity is computed.•A dynamic scheme for the enhancement of the method’s calculation efficiency is proposed.•The results present an excellent conformity in a broadband frequency range.•The approach radically reduces the computation time needed. The problem of the dynamic response of a structural–acoustic system in the mid-frequency range is considered in this work. The structure is a composite panel of arbitrary thickness and anisotropy. The dissipation characteristics for both, the structure and the cavity are taken into account. The system is initially modelled using finite elements, and is subsequently reduced using the Second Order ARnoldi reduction method (SOAR) which does not require inversion of large matrices for every computed frequency, thus resulting in more efficient calculation times. The fully coupled system is modelled using a Statistical Energy Analysis like (SEA-like) approach, and the energetic characteristics for each subsystem are computed and compared to the direct FEM solution. The error of the reduced model calculations for each frequency band is presented and the limits of the reliability of the reduction are explored. Different strategies concerning the reduction process parameters are investigated in order to optimize the accuracy with respect to time efficiency. The loading applied to the model comprises typical random distributed excitations, such as a ‘rain-on-the-roof’ excitation, a diffused sound field and a Turbulent Boundary Layer (TBL) excitation.