Research On Comparision Of Damage Detection Indices And Damage Severity Identification For Bridge Structure

Bridge structure is the most important infra-structure form of transport, widely used inroad construction. The bridge disasters lead to huge casualties and economic losses,resulting in adverse social impact. The occurrence of bridge disasters indicate that thecollapse accident easily happen under inherent invisible injury, if bridge not be promptlydetected, repaired or reinforced. Therefore, timely damage detection, scientific assessmentand keep abreast of structural damage condition to the bridge structure, these measures canensure the safe operation of the structure and prevent disasters and accidents happen.In this paper, simply supported beam bridge and continuous girder bridge as theresearch object, carried out the preferred work based on structural dynamic characteristicsdamage index. On this basis, a method to identify the appropriate degree of injury isproposed, whose main work is as follows:(1) Based on the theoretical foundation of dynamic characteristics damageidentification, comparative analyzed indicators’ sensibility in the recognition of the damagelocation, include frequency, mode shapes, modal curvature, modal flexibility, uniform loadsurface curvature and so on Recognition results show that the modal curvature, modalflexibility and uniform load surface curvature difference indicators can achieve effectivedamage location..(2) Analyzed the noise immunity of curvature modal flexibility difference, modalcurvature difference and uniform load surface curvature difference, the results show that thecurvature modal flexibility difference have strong noise immunity, which can be used asstructural damage identification preferred indicators and can be used in the continuousgirder bridge damage identification.(3) Simply supported beam bridge and continuous girder bridge as the research object,modal flexibility difference index as the input parameters of genetic optimizatic neuralnetworks, verified the validity of the method in a single location and multi-location damage detection area. The results show that: for simply supported beam bridge, the maximumrelative error for damage identification of genetic algorithm optimization neural network fora single location is2.67%, the maximum relative error for damage identification ofmulti-position is6.83%. For continuous girder bridge, the maximum relative error fordamage identification of genetic algorithm optimization neural network for a single locationis3.83%, multi-position for maximum relative error is8.14%of damage identification.(4) Comparative analysis based on BP neural network indicates that, the maximumerror of neural network optimized for simply supported beam bridge a single locationdamage identification is2.67%, while the maximum recognition error of BP neural networkis4%, so genetic neural network optimization accuracy is better than BP network..