Activities per year
Abstract
Carbon fiber reinforced polymers (CFRPs) are widely employed in aerospace engineering due to their high specific strength and lightweight properties. To ensure structural integrity under cyclic fatigue loading, ultrasonic guided wave based structural health monitoring (SHM) techniques have been developed to detect internal damage, particularly delamination, which critically reduces stiffness. However, existing damage imaging algorithms are constrained by limited data utilization, reliance on manually tuned thresholds, and low imaging accuracy. To address these challenges, we propose a novel Large Language Model-enhanced Probabilistic Damage Imaging (LLM-PDI) framework for high-precision visualization of damage in composite structures. First, the proposed Time Series Large Language Model (TSLLM) extracts damage-relevant features from path-level time-series signals and specimen-level textual prompts. These features are then fused through context-based multimodal reasoning within the LLM component to generate back-projection weights for each selected path. Subsequently, an improved probabilistic damage imaging (PDI) algorithm employs the generated weights to back-project signal damage indices (DIs) into the inspection area, yielding pixel-wise damage intensity. After normalization and thresholding at 0.5, the damaged region is visualized as a binary map. The method was validated on seven CFRP laminate specimens with varying ply orientations, achieving a Dice Similarity Coefficient (DSC) of 0.8491 on the test set, significantly outperforming existing imaging algorithms.
©2026 IEEE
©2026 IEEE
| Original language | English |
|---|---|
| Title of host publication | 2026 IEEE International Instrumentation and Measurement Technology Conference (I2MTC) Proceedings |
| Publisher | IEEE |
| Number of pages | 6 |
| ISBN (Electronic) | 979-8-3315-8938-7 |
| ISBN (Print) | 979-8-3315-8939-4 |
| DOIs | |
| Publication status | Published - 25 Sept 2026 |
| Event | International Instrumentation and Measurement Technology Conference 2026: Smart Instrumentation for a Sustainable World - Nancy Congress Center - Centre Prouvé, Nancy, France Duration: 25 May 2026 → 28 May 2026 https://i2mtc2026.ieee-ims.org/ |
Publication series
| Name | |
|---|---|
| ISSN (Electronic) | 2642-2077 |
Conference
| Conference | International Instrumentation and Measurement Technology Conference 2026 |
|---|---|
| Abbreviated title | I2MTC |
| Place | France |
| City | Nancy |
| Period | 25/05/26 → 28/05/26 |
| Internet address |
Funding
This work was supported by the National Natural Science Foundation of China (Grant No. 52405164).
Research Keywords
- CFRP
- Guided Wave
- Damage Imaging
- Large Language Model
Fingerprint
Dive into the research topics of 'Probabilistic Structural Health Monitoring in Composite Structures using Large Language Models'. Together they form a unique fingerprint.Activities
- 1 Conference / Symposium
-
2026 IEEE International Instrumentation and Measurement Technology Conference (I2MTC)
XU, X. (Presenter) & LIU, C. (Presenter)
25 May 2026 → 28 May 2026Activity: Organizing or Participating in a conference / an event › Conference / Symposium
Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver