Acoustic Emission Energy Release Rate Model for Classification of Damage Development in Large Fiber Reinforced Plastic Composite Structures

Abstract

Acoustic emission identification of fracture mechanisms in composite materials is of great importance for the practical assessment of the serviceability of aerospace structures, such as unmanned aerial vehicles, rocket motors, and others. Many scientific research studies conducted in recent decades have demonstrated the unique capabilities of the Acoustic Emission (AE) method to detect and identify different failure mechanisms, including fiber breakage, matrix cracking, and delaminations in fracture tests of composite specimens. However, testing large composite structures, in many cases, can be accompanied by multiple time-overlapped fracture events activated simultaneously, resulting in considerable difficulties in the classification and assessment of combined prolonged AE signals. In this work, we propose the Energy Release Rate (ERR) model for the classification of AE activity related to fiber-reinforced plastic (FRP) composite fracture. The model identifies accurately both micro-scopic and macro-scopic fracture mechanisms in large aerospace structures having complex composition of materials. The model is practically applied using several machine learning methods and provides a simple means to identify and discriminate typical fracture mechanisms and their combinations. Specifically, we describe ERR model, the choice of most informative AE parameters, the choice of suitable machine learning classification methods, their training and testing on FRP specimens, and validating the model in practical tests of large full-size FRP composite structures.