Scholarlink

<p>As a core component of industrial production, the health of rolling bearings directly affects production efficiency. However, existing diagnosis methods often require manual parameter selection and involve complex models. This paper presents a novel approach that combines the whale optimization algorithm with the auto-reassigning transform to adaptively generate high-resolution time–frequency representations of vibration signals. These representations are then processed by an improved densely connected convolutional network, augmented with multi-scale convolution modules and depthwise separable convolution, thereby enhancing its capability to classification performance. Experimental results on two benchmark datasets demonstrate that the proposed method achieves high diagnostic accuracy while maintaining a favorable balance between efficiency and robustness, and outperforms several state-of-the-art baselines. This validates its practical value for real-world industrial fault diagnosis under diverse operating conditions.</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

Rolling Bearing Fault Diagnosis Based on Whale Optimization Algorithm—Auto-Reassigning Transform and Improved DenseNet

  • 期刊论文

  • 发表日期：

    2025年11月25日

查看全文
  • Siyu Chen,
  • Fangze Shang,
  • Hongliang Sun,
  • Hui Chen,
  • Jie Ma

摘要

As a core component of industrial production, the health of rolling bearings directly affects production efficiency. However, existing diagnosis methods often require manual parameter selection and involve complex models. This paper presents a novel approach that combines the whale optimization algorithm with the auto-reassigning transform to adaptively generate high-resolution time–frequency representations of vibration signals. These representations are then processed by an improved densely connected convolutional network, augmented with multi-scale convolution modules and depthwise separable convolution, thereby enhancing its capability to classification performance. Experimental results on two benchmark datasets demonstrate that the proposed method achieves high diagnostic accuracy while maintaining a favorable balance between efficiency and robustness, and outperforms several state-of-the-art baselines. This validates its practical value for real-world industrial fault diagnosis under diverse operating conditions.