Experimental crack identification using electrical impedance tomography

Abstract

This paper addresses the problem of practical crack identification in electrically conducting specimens using only boundary measurements. The method is commonly referred to in the literature as Electrical Impedance Tomography (EIT). Crack identification is determined from the electrical impedance distribution, which amounts to solving an inverse problem, starting from boundary measurements. Whereas this kind of inverse problem has been extensively addressed in its theoretical and numerical aspects, there is a scarcity of experimental results aimed at examining the applicability of the method for real conditions. We present new experimental results, based on a simple identification methodology. The efficiency and limitations of this method are assessed through a series of numerical simulations and laboratory experiments on two-dimensional geometries. Following a preliminary numerical validation stage, actual crack detection is carried out on a discrete network of resistors, as an approximation to Laplace's equation. Next, experiments are carried out on a continuous conductive medium, containing one and two flaws. Our results show that EIT is a promising candidate for crack identification in real life conditions with a potential for multiple crack detection.