Scholarlink

<p>This paper proposes a multi-view context-aware graph fusion network (MCGFNet) for rolling bearing fault diagnosis, featuring an architecture that first extracts high-dimensional features from time, frequency, and wavelet domains, then constructs relational graph structures guided by physical principles rather than purely data-driven methods, and finally processes these graphs via a relational graph convolutional network (RGCN) to identify fault patterns under varying conditions. Experimental results demonstrate that MCGFNet significantly improves classification accuracy and robustness, particularly under time-varying rotational speeds. Furthermore, an integrated attention mechanism enhances model interpretability by highlighting critical nodes and relationships, offering a reliable solution for intelligent monitoring and predictive maintenance in industrial settings.</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

Multi-view context-aware graph fusion network for bearing fault diagnosis with multi-signal fusion

  • 期刊论文

  • 发表日期：

    2026年4月24日

查看全文
  • Houqiang Hua,
  • Huiying Gao,
  • Yawen Zhang,
  • Xiaoqiang Zhang,
  • Wenlin Yao

摘要

This paper proposes a multi-view context-aware graph fusion network (MCGFNet) for rolling bearing fault diagnosis, featuring an architecture that first extracts high-dimensional features from time, frequency, and wavelet domains, then constructs relational graph structures guided by physical principles rather than purely data-driven methods, and finally processes these graphs via a relational graph convolutional network (RGCN) to identify fault patterns under varying conditions. Experimental results demonstrate that MCGFNet significantly improves classification accuracy and robustness, particularly under time-varying rotational speeds. Furthermore, an integrated attention mechanism enhances model interpretability by highlighting critical nodes and relationships, offering a reliable solution for intelligent monitoring and predictive maintenance in industrial settings.