Xin YANG

Condition and structural health monitoring (CM/SHM) is a pivotal component of predictive maintenance (PdM) strategies across diverse industrial sectors, including mechanical rotating machinery, aircraft structures, wind turbines, and civil infrastructures. Conventional deep learning models, while effective for fault diagnosis and anomaly detection through automatic feature learning from sensor data, often struggle with operational variability, imbalanced or scarce fault datasets, and multimodal sensory data from complex systems. Deep generative models (DGMs) including deep autoregressive models, variational autoencoders, generative adversarial networks, diffusion-based models, and emerging large language models, offer transformative capabilities by synthesizing high-fidelity data samples, reconstructing latent system states, and modeling complex multimodal data streams. This review systematically examines state-of-the-art DGM applications in CM/SHM across the four main industrial systems mentioned above, emphasizing their roles in addressing key challenges: data generation, domain adaptation and generalization, multimodal data fusion, and downstream fault diagnosis and anomaly detection tasks, with rigorous comparison among signal processing, conventional machine learning or deep learning models, and DGMs. Lastly, we discuss current limitations of DGMs, including challenges of explainable and trustworthy models, computational inefficiencies for edge deployment, and the need for parameter-efficient fine-tuning strategies. Future research directions can focus on zero-shot and few-shot learning, robust multimodal data generation, hybrid architectures integrating DGMs with physics knowledge, and reinforcement learning with DGMs to enhance robustness and accuracy in industrial scenarios.

Damage imaging plays a crucial role in structural health monitoring (SHM) systems for fast and efficient damage assessment. Delay-and-sum (DAS) beamforming is a widely used algorithm in non-destructive testing for damage imaging, but its effectiveness is often compromised by the use of sparse ultrasonic transducer arrays and the difficulty in detecting progressive delamination larger than the wavelength using guided wave-based methods under fatigue loading. Although X-ray imaging offers detailed assessments of progressive delamination, its application is still limited due to the need to interrupt fatigue loading cycles and its high operational cost. To this end, we propose a novel Damage Imaging framework that uses the fine-tuned Conditional Diffusion Model for SHM systems (DI-CDM). Leveraging the powerful image generation capabilities of diffusion models, the framework was fine-tuned by combining DAS beamforming images derived from ultrasonic sparse array data with X-ray images captured during fatigue loading cycles of the composite structures. The proposed approach can generate damage images that reveal the progression of delamination size in the fatigue loading process. The framework was validated through numerical simulations and experimental data from NASA datasets for composite structures, demonstrating its potential and effectiveness by applying diffusion models in SHM applications to enable fast, high-resolution damage imaging.

Structural health monitoring is designed to detect and analyze damage in mechanical, aerospace, and industrial structures to prevent potential catastrophic failures. The effectiveness of identifying structural faults and operational damage significantly relies on the sensing technology. Multisensory data fusion, as one of the promising techniques, combines data from multiple sensor modalities for reliable and accurate fault diagnosis. In this paper, we propose a novel decision-level multisensory data fusion approach. This approach combines data from an ultrasonic guided-wave sensor modality and a vibration sensor modality for the first time, aiming to achieve precise and stable damage identification. The Bayesian probabilistic distributions from each sensor modality are obtained using a multi-level Metropolis–Hastings algorithm. Subsequently, a multilayer perceptron network is trained to automatically allocate confidence scores for each sensor modality. The final fusion results are derived from the proposed decision-level fusion approach, combining the Bayesian probabilistic distributions from each sensor modality with allocated confidence scores. Simulation and experiment results indicate that the decision-level fusion approach produces more accurate and reliable outcomes for damage identification compared to relying on a single modality, especially in the case of progressive damage.

This study presents a novel approach for real-time vision-based structural health monitoring, focusing on evaluating the deformation state of lattice structures. The structures are renowned for their remarkable recovery capabilities and exhibit similar mechanical responses under compressive loads. Despite these characteristics, quickly assessing the health status of the applied structure using knowledge from another related lattice structure is usually time-consuming and impractical. To address this, we propose to combine Gaussian process classification with transfer learning, termed TL-GPC, to detect damage states under compressive loads while also achieving uncertainty quantification. By employing structural deformations captured via the optical flow algorithm as inputs, the internal transfer kernel factor in TL-GPC is tailored to model knowledge transfer between source and target domain inputs. Experimental results show that the proposed TL-GPC model can deliver higher damage detection accuracy while ensuring stable uncertainty quantification in scenarios with limited and unbalanced experimental data.