Bridge State Recognition Based On Vehicle-bridge Coupling Vibration Response

In recent decades,China’s highway and high-speed railway are developing rapidly,and the construction of bridges has also experienced a period of rapid development.At the same time,Bridge collapse accidents happen frequently.Bridges are subject to various static and dynamic loads,including environmental erosion,temperature effect,material aging,natural disasters and man-made damage,which is likely to occur in local locations of the structure,including deformation,cracks,corrosion,wear and so on.The state recognition of bridges,especially the nondestructive testing of bridges,has attracted great attention from all countries,and has become one of the current research hotspots.Based on the single degree of freedom vehicle model and the train space model,the vehicle response and the bridge response are taken as the analysis data,and the wavelet transform and neural network optimized by genetic algorithm are used to identify the structural damage.Taking a Simple supported beam bridge as the research object,the feasibility of the two methods in the structural state recognition based on different vehicle models and different analysis data is proved.The main work and conclusions are as follows:The triangular series method is used to generate the irregularity samples.The comparison between the numerical simulation and the expected irregularity power spectrum shows that the simulated irregularity can truly reflect the actual situation.Several damage simulation methods are introduced,then the natural frequencies of bridge finite element model under different methods are analyzed,and it proves the feasibility of each damage simulation method.When the vehicle response is used to identify the bridge damage,the vehicle dynamic response of two vehicles crossing the bridge is obtained by using vehicle single degree of freedom model and space model.Continuous wavelet transform is used to identify single damage and multiple damage conditions of two vehicles crossing the bridge,and the influence of speed,mass,stiffness and damping of single degree of freedom vehicles on the bridge damage identification effect is discussed.The conclusion is as follows: the continuous wavelet transform method can identify the structural damage state of single degree of freedom model,but can not identify the damage state of space model.When a single degree of freedom vehicle crosses the bridge,the lower speed can identify the damage well,and the increase of speed will increase the difficulty of identification;the change of vehicle mass and damping can not affect the damage identification effect;the increase of stiffness will reduce the identification effect.When using bridge response to identify bridge damage,the displacement energy damage index and the neural network optimized by genetic algorithm.A simple supported beam bridge is taken as an example to verify the damage identification effect of this method on the bridge crossing based on single degree of freedom vehicle and space model.The following conclusions are obtained: the displacement energy entropy and its curvature can directly determine the damage location,and can also qualitatively analyze the damage degree.Displacement energy index is not sensitive to vehicle speed and road surface roughness.When the single degree of freedom vehicle model crosses the bridge,the damage identification can be realized by analyzing the displacement time history signal of a single measuring point,but it should be based on the premise of low speed;when the space model crosses the bridge,the signal of a single measuring point can not meet the requirements,it needs to realize the damage identification through the signal analysis of multiple measuring points,and it can also be used at high speed.Both the displacement energy entropy and the energy entropy curvature as the sample input can identify the damage state of the structure by the method of genetic algorithm optimization neural network.