Evolutionary strategy algorithm for a complete characterization of porous materials using a standing wave tube

Abstract

A widely used model for describing the attenuation of an acoustical wave propagating in a rigid open-cell porous material is the Johnson–Champoux–Allard (JCA) model. This model is based on five macroscopic parameters describing the porous medium: flow resistivity, porosity, tortuosity, viscous and thermal characteristic lengths. Simultaneously with the development of direct methods for measuring these five parameters, defining and solving inverse problems based on an artificial intelligence approach appears to overcome some of the limitations of the direct methods. In this work, an application of the evolutionary strategy (ES) algorithm for the estimation of the five parameters of a porous material from simple acoustical measurements is presented. First, a number of numerical tests are performed and the results are used like a priori knowledge of how to set up an evolutionary algorithm to solve such a difficult problem in the shortest time. In the second step, the final setup of the evolutionary algorithm is applied for evaluation of the five parameters from experimental measurements. The method seems practical and promising since it is based on simple acoustical measurements and avoids using complicated and unreliable measurement setups.