Characterization of the electrical response of carbon-based fiber-reinforced composites for damage sensing
Gallo, Gerard J.
University of Delaware
The increased use of fiber-reinforced composites in large scale applications requires structural health monitoring (SHM) techniques that are able to detect and locate damage throughout the structure. Electrical measurements have been demonstrated to be sensitive to mechanical strain and damage in carbon-based composite materials. The study focuses on the ability to detect damage in various carbon-based composites with different levels of electrical anisotropy. A tomographic imaging technique is also introduced in order to utilize the electrical measurement technique in order to locate spatial damage in the material. The damage sensitivity of the electrical network in each composite is first characterized. Two different types of conductively modified glass fiber/epoxy/nanotube composites are manufactured and the anisotropy of the electrical networks is compared with a carbon fiber/epoxy laminate. Two-dimensional finite element models are generated and electrostatic simulations are performed to investigate the effect of the electrical anisotropy and the electrode configuration on the current flux through the material. The sensitivity of each electrical network in detecting micro-scale damage is also modeled, and simulation results are verified using experimental data gathered from incremental cyclic loading. Electrical impedance tomography (EIT) is then used in combination with electrical resistance measurements to locate permanent and subsurface damage in electrically anisotropic composites. Pre and post-damage electrical measurements are acquired from boundary electrodes. Four different current injection patterns are used to collect the measurements. EIT is then applied to generate a conductivity map of the domain. The accuracy of the algorithm is compared with normalized resistance measurements, and the spatial sensitivity of the two damage sensing techniques with respect to the current injection patterns is then compared. Drop-weight impact loading and tensile testing with induced defects are also performed on composite laminates to investigate the spatial sensitivity of the two electrical sensing techniques to more complex damage modes. Results indicate the damage sensitivity of the electrical network is highly dependent on the electrical anisotropy, regardless of the conductivity magnitude. The choice of electrode configuration and spacing becomes an important consideration as the anisotropy increases. The low axial/transverse conductivity ratio that results from the uniform dispersion of nanotubes in glass fiber composites provides the highest sensitivity. Carbon fiber laminates are comparatively less sensitive to matrix damage, but, by improving the transverse conductivity by modifying with carbon nanotubes, opportunity exists to increase damage sensitivity. Normalized resistance measurements acquired from an electrode network are found to be useful for quickly detecting localized damage and provide a rough approximation of the damage location. The speed of normalized resistance measurements allow for real-time monitoring. EIT is demonstrated to provide a more accurate assessment of the damage location but requires more time due to the computational requirements.