Study on the Damage Identification of Long-Span Arch Bridge Based on Variation Ratio of Curvature and RBF Neural Network

Half-through arch bridge is an important traffic structure, so it is extremely valuable to study the damage location questions on the condition of suspender damage. Based on the finite element modal, the efficiency of the variation ratio of curvature is researched in this paper. As it turned out, variation ratio of curvature which is a modal parameter, could locate the initial damage position, as well as the single or multiple damage detection. After data is normalized, it still has usability of detecting single damage position with 10% noise level. Convenience is provided for the subsequently accurate damage extent identification. Subsequently, radial basis function neural networks are applied to carry on the damage extent identification, and more precise results of the damage extent identification are acquired.