A Multimodal Gated Recurrent Unit Neural Network Model for Damage Assessment in CFRP Composites Based on Lamb Waves and Minimal Sensing

Zhuang, Long; Luo, Kai; Yang, Zhibo

doi:10.1109/tim.2023.3348884

Cited by 6 publications

(2 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The field of defect detection has experienced a significant surge in innovation and breakthroughs, catalyzed by the fast-paced advancements in artificial intelligence and deep learning technologies. Zhuang et al [5] introduced an innovative multimodal gated recurrent unit (MGNN) model that detects and localizes damage in CFRP composites. This model assimilates inputs from time-domain Lamb wave (LW) and continuous wavelet transform (CWT) energy signals, significantly bolstering its robustness against interference and precision in damage feature extraction.…”

Section: Small Target Defect Detection Methodsmentioning

confidence: 99%

MAA-YOLOv8: enhanced steel surface defect detection through multi-head attention mechanism and lightweight feature fusion

Han,

Zhang

et al. 2024

Phys. Scr.

View full text Add to dashboard Cite

In the process of industrial production, product defects often arise due to improper operations among other reasons, rendering the detection of such flaws an indispensable procedure. However, the vast array of defect types, coupled with their complex characteristics, poses ongoing challenges for contemporary defect detection algorithms within industrial settings. To solve this problem, the present study introduces an enhanced steel surface defect detection model based on the modified YOLOv8 algorithm—termed the MAA-YOLOv8 model—to augment the accuracy and practicality of the algorithm. Initially, a multi-head attention mechanism was incorporated into the C2f to bolster the feature extraction capabilities within the backbone network and diversify the attention maps. Secondly, in the neck structure, we design a multi-channel feature fusion module (McPAN) to solve the problem of balance between computational efficiency and the ability to capture useful features. A series of experiments conducted on the NEU-DET dataset reveal that the MAA-YOLOv8 model achieves a mean Average Precision (mAP) of 94.4\%, representing an enhancement of 11.1\% over the original YOLOv8s model. The MAA-YOLOv8 model proposed in this study substantially elevates the performance of steel surface defect detection while ensuring the speed of detection.

show abstract

Section: Small Target Defect Detection Methodsmentioning

confidence: 99%

MAA-YOLOv8: enhanced steel surface defect detection through multi-head attention mechanism and lightweight feature fusion

Han,

Zhang

et al. 2024

Phys. Scr.

View full text Add to dashboard Cite

show abstract

“…Renato S M et al built supervised machine learning classifiers based on the AE technology, and finally succeeded in recognizing the damage mechanisms of composite materials [37]. Zhuang et al [38] combined feature fusion techniques with multi-modal gated recurrent unit neural networks (MGNN) and energy signals from Lamb waves (LW) and continuous wavelet transform (CWT) to accurately detect and precisely locate damage in composite materials, significantly enhancing the accuracy and reliability of damage detection. Compared to deep learning, machine learning has advantages in acoustic emission signal processing, including strong model interpretability, high data efficiency, low computational resource requirements, fast processing speed, and good stability, especially in feature engineering where it can effectively incorporate domain knowledge, making it suitable for real-time monitoring and scenarios with limited data.…”

Section: Introductionmentioning

confidence: 99%

Applicability of a new composite amusement ride safety device based on acoustic emission monitoring technology

Liu,

Liang,

Liu

et al. 2024

Mater. Res. Express

View full text Add to dashboard Cite

Under the premise of ensuring safety, it is of great significance to realize the lightweight of the non-main load-bearing parts of amusement facilities. The purpose of this study is to study the failure process of a newly designed carbon fiber bumper by using acoustic emission technology. First of all, the design of carbon fiber anti-collision bar can effectively restrain passengers, and its weight is reduced by nearly two-thirds compared to traditional metal materials. Subsequently, the load-bearing capacity of the bumper was tested and acoustic emission monitoring was carried out. The test results show that this new type of combined structure of amusement facility has high reliability, which exceeds the safety factor of 3.5 required by the steel structure of amusement facility. In addition, Renyi entropy was used to select the best window function of short-time Fourier transform, and the frequency domain characteristics of acoustic emission signals of typical damage modes were discussed through appropriately selected windows function. A classifier based on supervised machine learning is established by combining frequency features and acoustic emission feature parameters. Furthermore, the use of classifiers helps to understand the damage behaviour of composite structures.

show abstract

Automated ultrasonic testing for near-surface flaws in CFRP

Wang,

Deng,

Wang

et al. 2024

Nondestructive Testing and Evaluation

View full text Add to dashboard Cite

A Multimodal Gated Recurrent Unit Neural Network Model for Damage Assessment in CFRP Composites Based on Lamb Waves and Minimal Sensing

Cited by 6 publications

References 36 publications

MAA-YOLOv8: enhanced steel surface defect detection through multi-head attention mechanism and lightweight feature fusion

MAA-YOLOv8: enhanced steel surface defect detection through multi-head attention mechanism and lightweight feature fusion

Applicability of a new composite amusement ride safety device based on acoustic emission monitoring technology

Automated ultrasonic testing for near-surface flaws in CFRP

Contact Info

Product

Resources

About