Optimal positioning of a wall in an acoustic cavity using reduced models and surrogate based optimization
Abstract
In the context of transportation design or civil engineering, noise reduction is a constraint which is more and more included in the design processes. In order to easily take into account this constraint, new efficient technics have to be proposed for studying for instance the noise level in an aircraft cabin with respect with the arrangement within or for finding the optimal design for limiting the noise propagation. Thus, an approach based on the use of a finite element approach combined with surrogate based optimization is proposed.
The considered mechanical problem couples structural and fluid domain using an acoustic fluid governed by the Helmoltz’s equation, a porous material modeled by the Biot-Allard’s constitutive law and some thin walls governed by elasto-dynamic equation placed in the fluid. The air-structure problem is solved using Xfem in order to be able to consider an arbitrary structure placed in the acoustic cavity. After the building of the full coupled problem which takes into account the porous material, a reduced model is built using a Craig-Bampton’s approach. This step leads to a significant reduction of the computational time.
Global optimization based on this mechanical problem requires a large number of calls to the mechanical solver. Therefore a surrogate-based optimization is used. The approach is based on the Efficient Global Optimization composed of two phases: (1) a kriging metamodel is built using only a few sample points and associated responses and (2) an iterative scheme using the expected improvement allows us to find the global minimum by adding smartly new sample points to the initial surrogate model.
The whole strategy has been applied on some 3D cavity on which the position of a wall is determined in order to minimize the mean quadratic pressure in a control volume. Some examples will be presented for illustrating the performance of the proposed approach.