Identification of cable damage for cable-stayed bridge based on temperature deflection

Here, aiming at cable damage identification problem of cable-stayed bridge, a cable damage identification method based on univariate linear regression model and damage probability was proposed. Firstly, based on the monitored data of a certain cable-stayed bridge in Wuhan city under Beidou 3D monitoring system, correlation coefficients among temperature deflections at various measured points along the bridge's longitudinal direction were calculated. A univariate linear regression model was constructed taking temperature deflections at two adjacent measured points as input and output. The sensitivity of temperature deflections at main beam to cable damage under action of temperature was analyzed, and then the cable damage identification method based on the univariate linear regression model combined with damage probability was adopted to do damage recognition for the cable-stayed bridge's finite element model and the actual cable-stayed bridge. The analysis results showed that temperature deflections at main girder of cable-stayed bridge along its longitudinal direction can have higher similarity; the regression function constructed with temperature deflections at adjacent measured points can have higher fitness; taking model parameters of the constructed univariate linear regression model as damage indexes, cable damage situations can be effectively recognized for cable-stayed bridge's finite element model and actual cable-stayed bridges. © 2021, Editorial Office of Journal of Vibration and Shock. All right reserved.