Fuzzy correlation matrix method for health condition assessment of cable-stayed bridges based on long-term cable force monitoring

This study employs fuzzy correlation matrix analysis to investigate long-term cable force monitoring data from the full-span cables of the Lizhi Wujiang Bridge. The results demonstrate that the fuzzy correlation matrices obtained from monitoring data across different time periods exhibit a high degree of similarity. In the absence of significant structural damage, these matrices remain largely unchanged. Based on these findings, we propose using the eigenvalue normalized spectrum entropy of the fuzzy correlation matrix as an evaluation indicator for bridge condition assessment and develop an early warning and condition assessment model within the health monitoring system. To validate the sensitivity of this method in detecting cable damage, finite element simulations were conducted to generate data under various damage scenarios. A comparative analysis of the fuzzy correlation matrices in undamaged and damaged states reveals that the method exhibits high sensitivity in identifying damage in long cables—damage as slight as 5% can be detected. However, its sensitivity is relatively lower for short cables. Furthermore, by comparing the evaluation indicators derived from the analytic hierarchy process (AHP)—a method commonly used in health monitoring systems—with those obtained from the proposed approach, this study confirms the good applicability of the proposed method in condition assessment for cable-stayed bridges. Finally, practical recommendations are provided for the real-world application and reproducibility of the method.