Impact damage identification for composite structures via laser-induced graphene-based electrical impedance tomography
Deng Zhou, Gang Yan, Zaixing Huang
This study develops an embedded impact damage identification method for composite structures using laser-induced graphene (LIG) combined with electrical impedance tomography (EIT) and deep learning. An LIG sensing area was directly fabricated on a polyimide-based flexible printed circuit (FPC) via a CO2 laser and embedded within the interlayer region of glass fiber reinforced polymer composites. This LIG-FPC sensor serves as a highly sensitive sensing layer for capturing localized conductivity change induced by impact damage. After impact, the change in the pathways of the conductive network results in measurable change in the boundary voltages. The boundary voltage change was acquired by an EIT system and processed by a modified residual network (ResNet18) to reconstruct internal conductivity change distribution. A simulation-based dataset mapping boundary voltage changes to conductivity change distributions was used to train the network. Experimental results demonstrate that the proposed method can precisely localize damage. This approach effectively overcomes limitations associated with traditional sensor-based methods, providing a robust, fast, and non-destructive solution for in-situ structural health monitoring (SHM) of composite structures.