A Cross-Diagnosis Method for Determining the Structural Condition of a Long-Span Suspension Bridge Based on the Spatial Windows of Distributed Strain Data

Using structural health monitoring (SHM) techniques, Brillouin optical time-domain analysis (BOTDA) sensors can be mounted along the main box girder entire length of a long-span suspension bridge, and the high-density measured points strain monitoring data can be obtained. However, insufficient research has been conducted on accurately diagnosing the structural condition of a long-span suspension bridge by using the abovementioned strain monitoring data. To address this issue, a cross-diagnosis method that determines the structural condition of long-span suspension bridges based on the distributed strain data spatial window is proposed in this study. First, the distributed strain data spatial window based on a long-span suspension bridge structural symmetry is defined. Then, a method that divides the distributed strain data of the bridge main box girder into different spatial windows using mutual information between the strain data from BOTDA sensors is presented. The special symmetry of the environmental temperature effect on the spatial window structural performance is carried out to separate the temperature effect from the strain monitoring data; this process can effectively reduce the interference of ambient temperature on the results of the structural condition diagnosis. Second, using a convolutional neural network, a diagnosis index of the structural condition is generated by using the correlation model between the high-density measured points and the distributed strain data belonging to one whole spatial window. Regarding one spatial window, the proposed diagnosis index can effectively reflect the variation in the distributed strain correlation model caused by the damaged condition of the long-span suspension bridge to achieve cross-diagnosis of the structural condition of the bridge. Finally, the effectiveness of the proposed method is demonstrated through a numerical simulation using strain monitoring data obtained from a real bridge.