期刊论文

Cui, J

英文

Measurement

0263-2241

2026

257

118674

10.1016/j.measurement.2025.118674

CRediT authorship contribution statement Naiwei Lu: Writing – review & editing, Validation, Supervision, Resources, Project administration, acquisition, Conceptualization. Jian Cui: Writing – original draft, Visualization, Software, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Weiming Zeng: Visualization, Validation, Investigation. Xiangyuan Xiao: Visualization, Formal analysis, Data curation. Yuan Luo: Visualization, acquisition, Formal analysis, Data curation.

本校为第一且通讯机构

The complex structural systems of large bridges pose challenges for damage identification using traditional mechanical or data-driven methods. A key limitation of supervised machine learning methods is the lack of effective samples representing structural damage conditions. A novel domain adaptation transfer learning (TL) method, SE-TL-ResNet34 model, is proposed for structural damage identification, enabling the balanced training and reducing overfitting. First, convolutional neural networks (CNNs) were utilized to extract modal damage feature...