Structural Condition Monitoring Of Small And Medium Span Bridges Based On The Correlation Of Monitored Responses

Fast development of my country’s economy has driven the upsurge of infrastructure construction.As an important part of infrastructure,bridges have also entered a stage of rapid construction in recent years.However,after long-term use of bridges,due to material aging and environmental erosion,various damages will inevitably occur on bridges,leading to bridge accidents from time to time.Therefore,health monitoring of bridge structures has been paid more and more attention.At present,bridge monitoring systems are mostly used for bridges with large spans,and monitoring methods for small and medium span bridges are relatively lacking.Due to the low importance of small and medium span bridges,incomplete bridge monitoring information,especially the lack of vehicle load information,the structural parameters of the bridge are often unknown,s o it is urgent to study a structural health that does not depend on vehicle load and bridge struc tural parameters monitoring methods.This paper proposes a monitoring method for small and medium span bridges based on the correlation of bridge responses.Th rough the analysis of the vehicle load influence line,the mechanism that the response between different positions of the bridge does not depend on the vehicle load mapping relationship is theoretically clarified;the correlation model of different bridge response time domain data is established by using LSTM neural network,and the structural response correlation models at different locations suitable for condition monitoring of small and medium span bridges are selected.The specific contents of this article are as follows:(1)The mechanism of correlation between structural responses of different positions of small and medium span bridges under vehicle load is studied.Using the vehicle load influence lines of responses at different positions of the bridge,the correlation mapping relationship between structural responses of bridges at different positions is deduced under single and multi-car loading conditions which proves that the above-mentioned correlation mapping relationship does not depend on the vehicle load information under certain conditions;(2)The modeling method of the correlation model of different position responses of small and medium span bridges based on LSTM neural network is studied.The Ansys finite element analysis software is used to simulate the dynamic response of the simply supported beam bridge under different vehicle condi tions.The bridge deflection and acceleration response data measured at 1/4 span,mid span and 3/4 span are used as input and output of the correlation model and LSTM network technology is used to establish the time-domain correlation prediction model between the input structure responses and the output structure responses.(3)The damage state monitoring method of small and medium span bridges based on response correlation is studied.The response data of the bridge structure under different damage conditions were simulated,and the change of the prediction error of the correlation model after the bridge damage was analyzed.The sensitivity analysis of the model prediction error was used to give the most suitable monitoring model.