Locating and Quantifying Through Circular Damage in CNT/GFRP Composite Panel Using Gaussian Fit

摘要

Previously published articles on detecting damage in electrically conductive panels mainly concentrate on electrical impedance tomography methods (EIT) which are based on using surface bounded boundary electrodes and taking advantage of an electrically conductive layer on the surface of the panel or of a conductive matrix material. In this study instead, embedded electrodes in glass fiber reinforced epoxy panels are used to locate and quantify the artificial damage inflicted on the panel. The panel was manufactured using vacuum infusion method. It consisted of 10 (S-2) glass fabric plies, where copper electrodes were embedded below the top layer and then vacuum infused with carbon nanotube (CNT) mixed epoxy. During all measurements, a constant electrical current was supplied from two outer electrodes (the source and sink) and changes in voltage from the two inner probes were recorded. In contrast to EIT methods, no complicated algorithm is used to solve the conductivity distribution of the panel but instead, a simple algorithm that fits Gaussian curves to the data obtained using a four-probe measurement technique. Using the fitted curves, we are able to detect location and magnitude of the damages within a confidence bound. This practical method reduces computational cost and also enables the use of embedded electrodes which could provide more durability for the sensors. The experimental data is in very good agreement with the finite element simulations. Comparison of relative voltage change before and after the damages is consistent and sensitive enough to detect damages down to 1/8" diameter hole inside an area of 33 in2. As expected, accuracy is higher for larger diameter holes.