Assessment of tensile behaviour of plain weave fabric CFRP composites using acoustic emission technique and deep learning

Abstract

Abstract The applications of Carbon Fibre Reinforced Plastic (CFRP) composites in aerospace applications have increased exponentially in the last decade. CFRP composites are now replacing conventional materials used in high temperature applications. Different types and strategies of reinforcements are used in these composites to improve their applicability to high temperature applications. One of the reinforcement strategies that has been used frequently in the recent years is the plain weave fabric configuration of reinforcing fibres in polymer matrix. A comprehensive damage assessment is essential to evaluate the characteristics of these composites at elevated temperatures. The Acoustic Emission (AE) signals generated during the damage evolution stages are used for studying the damage evolution stages when the plain weave fabric composites are tested at temperatures close to their glass transition state. A new information-theoretic parameter, Lempel-Ziv (LZ) complexity and the deep learning neural network is used for understanding the damage evolution stages. Furthermore, the neural network is used for validating the utilization of LZ complexity as a potential AE parameter for damage classification applications. The results are promising with the damage classification strategy using LZ complexity exhibits an accuracy of 85.1% which is validated through the neural network.