Effective deffect identifications in honeycombs

Abstract

The image reconstruction problem based on Electrical Impedance Tomography (EIT) is an ill-posed inverse problem of finding such conductivity distribution that minimizes some optimisation criterion, which can be given by a suitable primal objective function. This paper describes new algorithms for the reconstruction of the surface conductivity distribution, which are based on stochastic methods to be used for the acquirement of more accurate reconstruction results and stable solution. The proposed methods are expected to non-destructive test of materials. There are shown examples of the identification of voids or cracks in special structures called honeycombs. Instead of the experimental data we used the phantom evaluated voltage values based on the application of finite element method. The results obtained by this new approach are compared with results from the known deterministic approach to the same image reconstruction.