Study On Method For Identifying Move Loads Through Data From Bridge Observation

With the development of economy and science and technology,the demand for bridges is increasing,and the development of bridges has a great influence on human travel and life.At present,the development trend of bridge engineering in China is gradually transitioning from basic construction to reliability design of bridge and the direction of bridge health monitoring.Due to the rapid increase of traffic flow,many Bridges in China have insufficient bearing capacity and other safety hazards.In recent years,many bridge collapse accident has occurred,is engaged in the statistical data,after the bridge collapse caused by overloaded vehicles is the main reason and not timely detection and maintenance of Bridges in the late take corresponding measures.Therefore,how to identify the moving load on the bridge has been closely watched.It is very important to assess the current carrying capacity of the bridge and extend the service life of the bridge.In this paper,we mainly use the simple support T beam model and ANSYS and Matlab software,which are based on the response data of the corresponding section of the bridge under the bridge,and the moving load on the bridge is identified.Based on the similarity theory,this paper studies the method of bridge moving load identification through theoretical analysis and model test.And analyze whether the error of the identification result meets the requirements.Finally,there are some problems in the process of mobile load identification.The main contents of this paper are as follows:(1)First introduced the three theories to identify the basic method of load to euler beam as the research object,the forced vibration equation under moving load,according to the principle of modal superposition,the dynamic response of the bridge.The equation of moving load is recognized by the dynamic response of the structure,and the moving load on the bridge is identified by the theoretical method.(2)Finite element model is established,according to the static and dynamic load test to collect the relevant data to the original finite element model,comparing the calculated data analysis comparison error of the two,and then to modify the related parameters in the finite element model of large sensitivity.(3)Using BP neural network method to identify moving loads,the dynamic strain under different vehicle weight and speed data were analyzed,and the selected model car speed and each section of maximum dynamic strain data as input variables,vehicle weight as output variables,to establish the BP neural network training samples.Finally,we use the trained neural network to recognize the vehicle weight of the model car.(4)According to the third and fourth chapter,the problems existing in the moving load process are optimized.Preconditioning conjugate gradient method is used to identify the high accuracy,the advantages of fast convergence speed,select the appropriate optimal matrix,which largely improve the ill-posed problem in the process of moving load in recognition.The results show that the method of using dynamic strain data to identify the moving load of the bridge is reliable and effective.