Machine Learning for Structural Health Monitoring of Aerospace Structures: A Review.

Area of Science:

• Aerospace Engineering
• Computer Science
• Materials Science

Background:

• Structural health monitoring (SHM) is crucial for aerospace safety and performance.
• Increasing system complexity necessitates advanced SHM techniques.
• Machine learning (ML) integration is transforming damage detection, localization, and prediction in aerospace.

Purpose of the Study:

• To provide a comprehensive review of recent advances in ML-based SHM for aerospace applications.
• To highlight the capabilities and challenges of various ML techniques in aerospace SHM.
• To outline a roadmap for future research and deployment of ML in aerospace SHM.

Main Methods:

• Review of supervised, unsupervised, deep, and hybrid learning techniques.
• Analysis of ML for processing high-dimensional sensor data and managing uncertainty.
• Exploration of emerging ML trends like digital twins, transfer learning, and federated learning.

Main Results:

• ML methods demonstrate significant potential for real-time diagnostics and damage prediction.
• Key challenges include data scarcity, operational variability, and interpretability in safety-critical systems.
• Current strengths and limitations of ML-based SHM are identified.

Conclusions:

• ML-based SHM offers a revolutionary approach to aerospace safety.
• Addressing challenges is essential for transitioning ML from research to operational deployment.
• Future research should focus on digital twins, transfer learning, and federated learning for robust SHM.