Laser-induced graphene as an embedded sensor for impact damage in composite structures assisted by machine learning

Abstract

Laser-induced graphene (LIG) enables the creation of cost-effective sensing devices via one-step laser scribing on organic substrates. This work aims to demonstrate the feasibility of LIG as an embedded damage sensor in structural composite materials. LIG was produced on an inexpensive cork substrate using a low-power blue light laser engraving system. The LIG was characterised to establish the relationship between the LIG’s physical properties and the lasing parameters. Using the lasing parameters which offer the optimal LIG properties, a LIG mesh pattern lased on the cork substrate was embedded in a glass fibre composite laminate as the damage-sensing core material to assess its sensing capability for impact damage. By measuring the electrical resistance change in the LIG mesh pattern consisting of a series of horizontal and vertical channels before and after impact loading, it was able to define the location of internal damage and its damage size, validated by X-ray computed tomography results. It is demonstrated that the test data can be used to train a machine learning algorithm to develop a simple damage-sensing system with a high accuracy rate of 94.3%, which significantly reduces the manual effort for large-scale composite structural health monitoring. This simple damage sensing system can be used to monitor internal impact damage in composite structures, such as off-shore wind turbine blades, which are inaccessible for inspection by conventional non-destructive testing.