Proceedings
home
preface
contents
authors
keywords
copyright
reference ©2012 Civil-Comp Ltd

Paper 215

Reliability-Based Design Optimization for the Analysis of Vibro-Acoustic Problems

M. Mansouri1,2, B. Radi1 and A. El Hami2
1LM, FST Settat, Morocco
2LMR, INSA de Rouen, St Etienne de Rouvray, France

Keywords: fluid-structure interaction, vibro-acoustic, numerical simulation, finite element method, reliability based design optimization.

full paper (pdf) - reference

In this paper focuses on problems of fluid-structure interaction and specifically the vibro-acoustic coupling which is generally defined as the contact between bodies interacting according to the principles of continuum mechanics. The comprehension of the mechanisms of interactions between a fluid and an elastic solid are of prime importance in several industrial applications ( in the aerospace, automotive and civil engineering areas as well as in biomechanics). When a structure vibrates in the presence of a fluid, there is an interaction between the natural waves of each media: the fluid flow generates a structural deformation and, or the movement of a solid causes the displacement of the fluid. These applications require an effective coupling. In addition, the dynamic analysis of the industrial systems is often expensive from the numerical point of view. For the coupling in the fluid-structure finite element models, the importance of the size reduction becomes obvious because the fluid freedom degrees will be added to those of the structure. A method of condensation will be used to reduce the size of the matrices. One of the principal hypotheses in the use of component mode synthesis method is that the model is deterministic; it is to say that parameters used in the model have a defined and fixed value. In fact, all aspects of an analysis model are uncertain, and uncertainty is either ignored or accounted for using conservative assumptions. However, the fluctuations in the input parameters generate significant degradation of the quality of the deterministic solution. The volatility of the parameters and the other difficulties require consideration of variability in the formulation of the coupling problem. So it is neither financially feasible nor physically possible to eliminate the dispersion of the input parameters.

The reduction of the dispersion is generally associated with higher costs, either by production methods more efficient and accurate processes and increased efforts in quality control, hence, by accepting the existence of these uncertainties, the problem must be considered with uncertainty. Furthermore, the knowledge of the variation response of a structure involving uncertain materials, geometrical parameters, boundary conditions, tolerances of manufactures and loading conditions is essential in the global process of conception. In order to do that, the modal condensation method is extended to reliability analysis for coupled fluid-structure finite element models.

A numerical vibratory study is based on a three-dimensional structure immersed in water taking into account the acoustic aspect. In this context, the study is focused very specifically on a deterministic, stochastic and reliability analysis through numerical simulations in three-dimensional dynamic fluid-structure interaction problems. In the case of this problem, the presence of several parameters to random characters namely the Young's modulus of the structure and structure density and fluid density, which often show a great variability, which inevitably leads to a loss of precision important. Better control of these parameters is thus based on the use of stochastic methods whose main objective is to improve the quality and the reinterpretation of results from simulations. To do this, a good understanding and formulation of the main phenomena involved in the coupling problem are required.

The results of the reliability based design optimisation study shows the effectiveness of the steps followed to condense the system and to take into account the uncertain parameters. The numerical results are compared with some experimental ones. The results obtained show the potential of the proposed methodology and encourage improvement of this procedure for use in complex coupled fluid-structure systems.