An approach for identification of bridge bending stiffness distribution using improved Gaussian peak function

Abstract

Bridge damage and bending stiffness deterioration can lead to a decrease in the bearing capacity of a bridge, posing significant safety concerns. To ensure the safe operation of bridges, it is crucial to accurately identify bending stiffness deterioration and assess the bridge performance. The bridge influence line (IL), which reflects the deterioration of bending stiffness, was widely used in the assessment of bending stiffness deterioration. Existing IL-based methods mostly assume a uniform reduction in bending stiffness for identification. In reality, the deterioration of bridge bending stiffness is far more complex, which causes difficulty in accurately identifying the bending stiffness distribution using the uniform reduction model. This paper proposes a novel method for identifying bridge stiffness distribution using an improved Gaussian peak function model. The improved model incorporates additional parameters, allowing for considering a wider range of stiffness variation and thus addressing the limitations of traditional models. The parameters of the improved Gaussian peak function model are estimated by minimizing the residual between the predicted deflection influence line (DIL) and the actual DIL using the Levenberg-Marquardt (LM) algorithm. This approach enables the accurate identification of bending stiffness distribution without baseline information. Numerical simulation and laboratory experiment are used to verify the effectiveness of the proposed method in identifying bridge bending stiffness distribution. A comparison between the identified and actual bending stiffness distribution curves under various forms of damage assesses the applicability of the proposed method.