Scholarlink

<p>Structural health monitoring (SHM) is the most direct and advanced method for understanding the evolution laws of structures and ensuring structural safety. The essence of SHM lies in diagnosing structural health by analyzing monitoring data. Since the introduction of machine learning paradigm for SHM, using machine learning methods to analyze the monitoring data, identify, and evaluate structural health status has become a prominent research topic in this field. For complex bridge structures, diagnosing structural health based on highly incomplete monitoring data presents an inherent high-dimensional problem. Machine learning methods are particularly well-suited for addressing these issues due to their capabilities in effective feature extraction, efficient optimization, and robust scalability. This article provides a brief review of the developments in machine learning-based structural health diagnosis, including data cleaning, structural modal parameters estimation, structural damage identification, digital twin technology, and structural reliability assessment. Additionally, the paper discusses related open questions and potential directions for future research.</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

Recent advances in structural health diagnosis: a machine learning perspective

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年3月8日

查看全文
  • Yuequan Bao,
  • Huabin Sun,
  • Yang Xu,
  • Xiaoshu Guan,
  • Qiuyue Pan,
  • Dawei Liu

摘要

Structural health monitoring (SHM) is the most direct and advanced method for understanding the evolution laws of structures and ensuring structural safety. The essence of SHM lies in diagnosing structural health by analyzing monitoring data. Since the introduction of machine learning paradigm for SHM, using machine learning methods to analyze the monitoring data, identify, and evaluate structural health status has become a prominent research topic in this field. For complex bridge structures, diagnosing structural health based on highly incomplete monitoring data presents an inherent high-dimensional problem. Machine learning methods are particularly well-suited for addressing these issues due to their capabilities in effective feature extraction, efficient optimization, and robust scalability. This article provides a brief review of the developments in machine learning-based structural health diagnosis, including data cleaning, structural modal parameters estimation, structural damage identification, digital twin technology, and structural reliability assessment. Additionally, the paper discusses related open questions and potential directions for future research.