Scholarlink

<p>Carbon fiber reinforced polymer (CFRP) tubes serve as critical load-bearing components in aerospace and marine engineering; however, their structural reliability is profoundly influenced by the size effect. Consequently, it remains a significant challenge to directly extrapolate the damage evolution laws established from laboratory-scale specimens to full-scale engineering structures. Accordingly, radial compression tests were carried out on CFRP tubes with various dimensions in this paper. Elastic wave signals during damage evolution were captured using acoustic emission (AE) technology. To address the issues that existing damage identification models are difficult to adapt to dimensional variations, a deep learning model incorporating a self-attention mechanism was constructed to realize the intelligent identification of damage modes for CFRP tubes with different dimensions. The results demonstrate a pronounced size effect in the mechanical performance of CFRP tubes: upon normalization, the peak load and displacement of the large-scale specimens decreased by 3.8% and 17.4%, respectively, compared to the small-scale specimens. Although fiber/matrix debonding remains the predominant damage mode for CFRP tubes of different sizes, its evolution characteristics differ significantly with varying dimensions. The proposed Tabular Attention ResNet model achieved a superior identification accuracy of 97.10% for damage modes across different CFRP tube dimensions. This study reveals the damage evolution mechanism in CFRP tubes of various dimensions under radial compression, while offering a highly generalizable as well as lightweight methodological approach toward smart health monitoring of composite structures.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Size Effect and Damage Evolution of CFRP Tubes Under Radial Compression Using Acoustic Emission and Deep Learning

  • 期刊论文

  • 发表日期：

    2026年8月1日

查看全文
  • Wei Zhou,
  • Chuang Yang,
  • Sai-nan Wang,
  • Yong-na Shen,
  • Yan-bing Wang,
  • Hui Sun

摘要

Carbon fiber reinforced polymer (CFRP) tubes serve as critical load-bearing components in aerospace and marine engineering; however, their structural reliability is profoundly influenced by the size effect. Consequently, it remains a significant challenge to directly extrapolate the damage evolution laws established from laboratory-scale specimens to full-scale engineering structures. Accordingly, radial compression tests were carried out on CFRP tubes with various dimensions in this paper. Elastic wave signals during damage evolution were captured using acoustic emission (AE) technology. To address the issues that existing damage identification models are difficult to adapt to dimensional variations, a deep learning model incorporating a self-attention mechanism was constructed to realize the intelligent identification of damage modes for CFRP tubes with different dimensions. The results demonstrate a pronounced size effect in the mechanical performance of CFRP tubes: upon normalization, the peak load and displacement of the large-scale specimens decreased by 3.8% and 17.4%, respectively, compared to the small-scale specimens. Although fiber/matrix debonding remains the predominant damage mode for CFRP tubes of different sizes, its evolution characteristics differ significantly with varying dimensions. The proposed Tabular Attention ResNet model achieved a superior identification accuracy of 97.10% for damage modes across different CFRP tube dimensions. This study reveals the damage evolution mechanism in CFRP tubes of various dimensions under radial compression, while offering a highly generalizable as well as lightweight methodological approach toward smart health monitoring of composite structures.