Coupled FE–BE method for eigenvalue analysis of elastic structures submerged in an infinite fluid domain

Abstract

Summary For thin elastic structures submerged in heavy fluid, e.g., water, a strong interaction between the structural domain and the fluid domain occurs and significantly alters the eigenfrequencies. Therefore, the eigenanalysis of the fluid–structure interaction system is necessary. In this paper, a coupled finite element and boundary element (FE–BE) method is developed for the numerical eigenanalysis of the fluid–structure interaction problems. The structure is modeled by the finite element method. The compressibility of the fluid is taken into consideration, and hence the Helmholtz equation is employed as the governing equation and solved by the boundary element method (BEM). The resulting nonlinear eigenvalue problem is converted into a small linear one by applying a contour integral method. Adequate modifications are suggested to improve the efficiency of the contour integral method and avoid missing the eigenvalues of interest. The Burton–Miller formulation is applied to tackle the fictitious eigenfrequency problem of the BEM, and the optimal choice of its coupling parameter is investigated for the coupled FE–BE method. Numerical examples are given and discussed to demonstrate the effectiveness and accuracy of the developed FE–BE method. Copyright © 2016 John Wiley & Sons, Ltd.

Design sensitivity analysis of modal frequencies of elastic structures submerged in an infinite fluid domain

“…22,23 In spite of the advantages of solving the vibro-acoustic problems in unbounded fluid domain, the acoustic BEM makes the system of equations of the coupled FE-BE method have frequency-dependent coefficient matrices, which results in a typical nonlinear eigenvalue problem (NEP) in the modal analysis of submerged elastic structures. 24,25 Various methods have been presented to solve the NEPs (cf. Reference 26), but some of them are still troublesome and expensive to use.…”

Data-driven reduced order modeling for parametric PDE eigenvalue problems using Gaussian process regression

“…Eigenvalue problems occur naturally in many areas of science and engineering. Just to name a few applications, let us mention fluid dynamics [43,15], structural mechanics [5,46], electromagnetism [10], nuclear reactor physics [11,26,16], photonics [17,27,18], optics [32,12] for optimal control and optimal design, population dynamics [22], in acoustic problem [14] and solid-state physics [34] for calculating band structure. In nuclear physics, the eigenvalue problem represents the energetic diffusion of the neutron, where the first eigenvalue represents the criticality of the reactor and the corresponding eigenvalue represents neutron flux.…”

Divergence‐free finite elements for the numerical solution of a hydroelastic vibration problem

“…For instance, we mention [16] for a formulation in fluid displacement potential, [14,29] for an added mass formulation based on taking into account the effect of the fluid by means of a Neumann-to-Dirichlet operator and [6] for a BEM approach. We also mention [30,31,37,38] as examples of some more recent papers, which show a continuous interest on this subject.…”