Semi-supervised and unsupervised learning for health indicator extraction from guided waves in aerospace composite structures

2025-10-28

AI Summary

Extracting reliable health indicators (HIs) from guided-wave signals for aerospace composite structures remains challenging due to material variability, stochastic damage evolution, multimodal wave propagation, and—critically—the absence of ground-truth degradation labels. Method: This paper proposes a novel HI construction framework integrating semi-supervised and unsupervised learning. It introduces the Diversity-DeepSAD model, which leverages continuous auxiliary labels to characterize intermediate degradation states, and the monotonicity-constrained DTC-VAE model, designed to generate physically interpretable, trend-consistent HIs. The method fuses time-, frequency-, and time–frequency-domain features from multi-frequency guided waves, incorporating FFT and unsupervised ensemble learning. Contribution/Results: Experimental results demonstrate that DTC-VAE yields the most consistent HIs, achieving 92.3% accuracy while significantly reducing frequency dependency and variance—thereby enhancing robustness in structural health monitoring.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Search and Optimization: Mixed Discrete/Continuous Search
• Computer Vision: Multi-modal Vision

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Health indicators (HIs) are central to diagnosing and prognosing the condition of aerospace composite structures, enabling efficient maintenance and operational safety. However, extracting reliable HIs remains challenging due to variability in material properties, stochastic damage evolution, and diverse damage modes. Manufacturing defects (e.g., disbonds) and in-service incidents (e.g., bird strikes) further complicate this process. This study presents a comprehensive data-driven framework that learns HIs via two learning approaches integrated with multi-domain signal processing. Because ground-truth HIs are unavailable, a semi-supervised and an unsupervised approach are proposed: (i) a diversity deep semi-supervised anomaly detection (Diversity-DeepSAD) approach augmented with continuous auxiliary labels used as hypothetical damage proxies, which overcomes the limitation of prior binary labels that only distinguish healthy and failed states while neglecting intermediate degradation, and (ii) a degradation-trend-constrained variational autoencoder (DTC-VAE), in which the monotonicity criterion is embedded via an explicit trend constraint. Guided waves with multiple excitation frequencies are used to monitor single-stiffener composite structures under fatigue loading. Time, frequency, and time-frequency representations are explored, and per-frequency HIs are fused via unsupervised ensemble learning to mitigate frequency dependence and reduce variance. Using fast Fourier transform features, the augmented Diversity-DeepSAD model achieved 81.6% performance, while DTC-VAE delivered the most consistent HIs with 92.3% performance, outperforming existing baselines.

Problem

• Extracting reliable health indicators from guided waves in aerospace composites
• Overcoming limitations of binary labels in damage state classification
• Addressing variability in material properties and damage evolution modes

Innovation

• Diversity-DeepSAD uses continuous damage proxy labels
• DTC-VAE embeds monotonicity via trend constraints
• Unsupervised ensemble learning fuses multi-frequency health indicators