Performance Degradation Model Test And Parameter Identification Of Concrete Cable-stayed Bridges With Typical Damage

Cable-stayed bridges are widely used in bridge structures due to their strong spanning ability and elegant appearance.However,some components of cable-stayed bridges have accumulated more damage under the influence of unfavorable environment and external loads,such as cable corrosion and main girder cracking.The structural performance has changed greatly,which has seriously affected the service safety of cable-stayed bridge.Many concrete cable-stayed bridges have been built in China.It is a prerequisite for maintenance and strengthening in the later with considering the impact of damage on the service performance of structure and a reasonable assessment of the safety performance of these concrete cable-stayed bridges.For this reason,this paper takes the Second Hejiang Yangtze River Bridge in Sichuan Province as the engineering background,and investigates the performance degradation law and the method of parameter identification of concrete cable-stayed bridge under typical damage conditions through model test.The main research works are as follows:(1)Model tests of cable-stayed bridge segment under different damage conditions are conducted.A segment model of cable-stayed bridge with a scale ratio of 1:40 is designed based on the similarity ratio theory and the stiffness similarity principle using the Second Hejiang Yangtze River Bridge as a prototype.The static similarity analysis of the model bridge is executed based on the numerical calculation.The model tests of cable-stayed bridge under 8 groups of cable rupture and 12 groups of damaged main girder conditions are performed.The distribution of cable force,main girder stress and deflection under different damage conditions are revealed.The influence of cable failure and main girder damage on the performance degradation of cable-stayed bridges are quantified.(2)The finite element models of cable-stayed bridge under typical damage conditions are established.The elastic modulus of cable is reduced to simulate cable damage.The cable unit is deleted to simulate cable breaking.The thickness of bottom plate is reduced to simulate main girder damage.The finite element models of the prototype bridge and model bridge in service have been established.The results of finite element model are compared with the model test results.The influence of cable damage and main girder damage on the cable-stayed bridge’s cable force,pylon deformation,main girder internal force and deflection are investigated.The general law of performance degradation of cable-stayed bridges under damage conditions is clarified.(3)A damage identification method for cable-stayed bridges based on machine learning is developed.It is based on many data samples provided by the finite element model.The algorithm uses direct measurement indicators,such as cable force,main girder displacement,pylon deformation and main beam internal force,combining with the classification and regression functions of the BP neural network.A prediction method for identifying and predicting the position or degree of the main girder and cable damage of cable-stayed bridges is constructed.The effectiveness of the algorithm is verified by training and identifying 1360 groups of different working condition data.(4)A calculation model for the performance degradation of cable-stayed bridges under typical damage conditions is proposed.The analytical solution expressions for the change rules in the cable force,the internal force and deflection of main girder after the damage of cable and main girder under the action of uniform load are proposed.It is based on the virtual work principle of linear elastic bodies and the force method equation for solving statically indeterminate structures,and the cable-stayed bridge is equivalent to a multi-span continuous elastic support beam.Finally,the model is verified through a calculation example.