Structural Health Monitoring in Aviation: a Comprehensive Review and Future Directions for Machine Learning

Abstract

Aircraft structures are exposed to a variety of operational and environmental loads that can cause structural deformation and fractures. Structural Health Monitoring (SHM) has emerged as a promising solution for in-situ monitoring of structural components. This article presents a state-of-the-art review of SHM in aviation, current regulations, data acquisition sensors and equipment, and damage detection and identification methods. The article discusses in detail the regulations SHM specific to both civil and military aviation. A comprehensive review of conventional electrical resistance sensors, fiber optic, piezoelectric sensors and smart materials used for SHM monitoring in aircraft structures is then presented. The pros and cons of each data acquisition approach were discussed individually. The damage detection and identification section begins by describing the traditional knowledge-based methods that are combined with expert knowledge and theory, then focuses on the applicability in aircraft SHM systems of spectral or frequency domain models. The last part investigates the new paradigm, machine learning and deep learning methods such as CNN and LSTM on different types of aircraft structures through the existing literature. Furthermore, it covers an emerging approach called physics-informed neural networks (PINN), which combines physics and machine learning, and explore its potential for SHM applications.