Damage Detection of a Three-Dimensional, Non-Planar Carbon Fiber Composite Structure via Electrical Impedance Tomography

<p dir="ltr">Composites like carbon fiber reinforced polymers (CFRPs) are heavily leveraged in the aerospace industry, where maximizing structural strength and minimizing weight are a concern. A disadvantage of CFRPs is that they can develop barely visible impact damage (BVID) that degrades performance and require nondestructive evaluation (NDE) methods for effective identification. It would be beneficial to the design engineer and/or end-user if the damage could be detected sooner so that proactive, instead of reactive, actions can be taken to address the damage. Embedded sensing systems could fill this gap by having a method of assessing the health of the composite structure in-situ. The implementation of external sensors are likely not feasible options due to the added complexities, higher costs and increased weight. Therefore, a self-sensing material could be the solution to these obstacles.</p><p dir="ltr">Several conductive materials that posses the piezoresistive effect have been investigated for the purposes of detection damage. Piezoresistive materials have a direct electrical response due to applied strains or deformations to the material. In other words, the resistivity of the material increases as the electrical network is perturbed by an external stimuli. Irreversible changes in resistivity are therefore indicators of damage within the material, due to severed internal connections. Nanomaterials like carbon nanotubes (CNTs) and carbon black (CB) have been popular piezoresistive materials because they can electrically functionalize insulating materials, like cement mortars and glass fiber reinforced polymers (GFRPs). However, using these materials require additional processing steps to be effective sensors. The use of conductive nanomaterials is an open research area and can be a barrier for adoption by industry. Conversely, CFRPs are a piezoresistive material that is standard to the aerospace industry. CFRPs do not require modification for self-sensing capabilities and have been demonstrated in damage detecting applications.</p><p dir="ltr">Electrical impedance tomography (EIT) is a potential modality for self-sensing within CFRPs due to the ability to detect damage through a current-voltage relationship within the laminate. EIT has been demonstrated as an effective damage detection method for a variety of materials, including nanomaterials and CFRPs. The added benefit of EIT is that the damage is not just sensed, but also spatially localized. However, a lot of EIT research centers on the application to flat plates, do not use fabrics pre-impregnated with resin, and use conductively isotropic materials. However, these examples do not represent the types of materials used in the aerospace industry today.</p><p dir="ltr">Presented in this thesis is the application of EIT on a non-planar and geometrically complex prepreg CFRP laminate. A formulation that incorporates the anisotropy of the material is implemented. The conductivity of the material system was experimentally derived. Additionally, EIT reconstructions were also explored using a homogeneous best-fit conductivity, calibrated using initial experimental data and compared with the measurement approach. Two different injection schemes are proposed and evaluated in order to address the increased geometric complexities of the specimen geometry. EIT is an ill-posed and underdetermined inverse problem. This work utilizes two different minimization approaches, namely minimizing the L1-norm and L2-norm to the impact of reconstruction images. With this framework in mind, damage from a notch and two impact events were reconstructed using EIT. The notch damage was clear and distinct. A 18 (J) damage reconstruction was stymied by significant noise that prevents definitive identification of the damage. Then a 46 (J) damage detected with significant improvement, but still contained minor noise. Changing the EIT minimization to the L1-norm dramatically improved the damage reconstruction and eliminated all noise. Damage to the electrode array was likely the sources of noise, which was supported by the results. Nonetheless, the application of EIT to the CFRP specimen demonstrated damage detection capabilities, with limitations that need addressing in order to improve the quality of results. The results of this study indicate a promising approach for using self-sensing CFRPs for an embedded-sensing system.</p>