Enhanced Coflux Strain Analysis for Multi-Point Anomaly Detection in Historical Timber Structures

作者：Juan Wang, Yuan Yao, Xiaoying Du, Lei Tan, Xiaohui Yang · 发表于：International Journal of Architectural Heritage · 年份：2025 · DOI：10.1080/15583058.2025.2597483 · 被引用次数：1 · 研究领域：Wood Treatment and Properties、Structural Health Monitoring Techniques、Wind and Air Flow Studies

In historical timber buildings, structural anomaly identification faces significant limitations: existing methods often overlook interdependencies among strain measurement points and exhibit inefficiencies in processing real-world engineering data. To address these challenges, this study proposes an enhanced Coflux strain analysis approach. The methodology employs a hybrid prediction model to isolate environmental variables from strain measurements, enabling the extraction of damage-sensitive residual fluctuation features. By analyzing correlation shifts in these fluctuation features under structural anomalies, the approach achieves multi-point anomaly localization. Furthermore, cross-correlation analysis establishes interconnections between anomalies across spatial locations. Validation using monitored data from the Feiyun Wood Pavilion and simulated anomaly scenarios demonstrates that the proposed method effectively identifies damage-sensitive fluctuation patterns, pinpoints anomalous measurement points and temporal periods with high accuracy, and reveals spatial-temporal interrelationships among anomalies.