Damage detection for composites using a Bayesian regularization-based electrical impedance tomography technique

Electrical impedance tomography (EIT) is a non-radiative and low-cost imaging technique that aims to estimate the interior electrical properties of an object from current-voltage measurements on its boundary. Recently, with the rapid development of nanotechnology, EIT has been applied to damage detection for composites combined with the excellent electrical properties of carbon nanotubes (CNTs). In this work, an adaptive Bayesian regularization algorithm is proposed for EIT to identify quantitatively the damage in composites. By hierarchical Bayesian modeling, the posterior probability distribution of conductivity change caused by damage is established within the context of Bayesian inference. Through a Bayesian learning algorithm, a maximum a posteriori (MAP) solution is adopted to automatically determine the optimal value of the regularization parameter, and reconstruct the unknown distribution of conductivity change, which quantitatively indicates the damage location and size. Numerical studies for synthetic data from a finite element (FE) model and experimental studies for a glass fiber reinforced polymer (GFRP) plate with a CNT sensing skin have been performed to demonstrate the applicability and effectiveness of the proposed Bayesian regularization-based EIT approach.