Using Machine Learning Models to Predict Damage in Aerospace Composites, Subjected to Real Flight Loads and Profiles

This paper presents a novel approach to structural health monitoring (SHM) of composite aircraft structures by incorporating realistic flight loads extracted from actual airline data. The research introduces a multi-scale framework that maps high-fidelity aerodynamic and fuel consumption data to structural models, enabling accurate simulation of dynamic loading conditions on composite components. Unlike previous studies that use simplified loading conditions, this approach captures the time-varying nature of in-flight loads while monitoring multiple damage types, including delamination and ply damage. The machine learning time-series analysis demonstrates promising results in identifying and predicting progressive damage states under these realistic conditions, offering significant improvements over existing SHM techniques for aerospace applications.