Damage identification of long-span bridges based on the correlation of probability distribution of monitored quasi-static responses. (1st March 2023)

• Record Type: Journal Article
• Authors: Deng, Fan; Wei, Shiyin; Jin, Xiaowei; Chen, Zhicheng; Li, Hui
• Is Part Of: Mechanical systems and signal processing. Volume 186(2023)
• Publication Date: 2023-03-01
• Subjects: Structural health monitoring -- Damage identification -- Correlation -- Probability distribution -- Quasi-static response; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2022.109908
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: This paper proposes a statistical representation model for the structural damage identification of long-span bridges based on the correlation of probability distribution on monitored quasi-static response, and the principle of the proposed method is that the correlation of monitored quasi-static response of linear time invariant bridges is only related with structural model and parameters under the assumption that the external load distribution pattern to be identical. The probabilistic distribution function (PDF) of vehicle load can be modeled as a function of traffic flow parameters and the PDF of vehicle weight. It is validated that the PDF of vehicle weight usually keep invariant under normal operation conditions by the monitored vehicle weight data. The invariant PDF of vehicle weight is further termed as the same vehicle load distribution pattern. Under the same vehicle load distribution pattern, the PDFs of monitored quasi-static responses may vary with the traffic flow, the distance of PDFs of any two monitored quasi-static responses over two statistical durations is equal for undamaged structure. And the variation of distances at different sensors can be used as the indicator for structural health diagnosis in the probabilistic perspective. Abstract: Structural health diagnosis is one of the most critical issues in structural health monitoring (SHM). Vibration-based methods have been extensively investigated in the last few decades in the SHM community, but they encounter the challenges of local damage insensitivity and environmental sensitivity. However, the study of structural health diagnosis based on structural quasi-static response data, which has well-defined spatial coordinates and is sensitive to the local condition of the structure, is insufficient. It is difficult to extract structural models or parameters that directly indicate structural damage owing to the coupling effect of the damage and unknown external loads in the quasi-static response data. This study proposes a method based on the correlation of the probability distribution of the quasi-static response data for damage identification, and it is proved that the variation in the structural condition can be inferred from the variation in the correlation of the probability distribution of multiple quasi-static response data. The earth mover's distance (EMD) is introduced as a quantitative indicator of the variation in the probability distribution function. The difference in the EMD (DEMD) between two monitored quasi-static responses was adopted as the damage indicator. The monitored strain on the steel box girder of a multi-tower cable-stayed bridge and the monitored cable tension of the cable-stayed bridge were employed to validate the proposed method. The results show that the proposed method successfully identifies the damage to the steel box girders and stay cables.