Application Of Moving Force Identification In Bridge Health Monitoring

With the development of urban roads and the rapid construction of expressway network in China,a large number of Bridges are put into operation.With the increasing volume of traffic and the evolution of time,the bridge in operation inevitably suffers from various diseases.The health assessment and daily maintenance of Bridges are becoming more and more important.At present,most of the bridge health monitoring systems are used to measure the dynamic response(including acceleration,strain and displacement)of Bridges under various loads(including vehicle load,wind load,temperature load and earthquake load).For the inverse analysis of bridge moving load,that is,the time-varying moving load on the bridge deck(including the speed of moving load,load position,axle load or total weight of the vehicle)is calculated from the dynamic response of the bridge,there is not yet a technology that can be scaled up,a lot of work is still in the theoretical and experimental research stage.The main research contents of this paper are as follows:1)The Midas model of the fabricated T-beam real bridge was established,and the two-axle vehicle load was applied to obtain the structural dynamic response.By changing the bridge structure parameters in the load identification algorithm,the bridge structure parameters that affect the load identification effect are compared and analyzed.It is concluded that the stiffness of the main beam has the greatest influence on the load identification effect,while the unit mass length and damping ratio of the main beam have no significant influence on the load identification effect.2)According to the data collected from the field dynamic load test,the error correction method for the stiffness of the main bridge beam and the design value is proposed;The actual vehicle speed is obtained through the preliminary analysis of the measured data,and the neural network is introduced to establish the mapping relationship between the dynamic time-history response of the bridge and the stiffness of the main beam.The l-M method is used to train the neural network,and the stiffness of the main beam is modified according to the actual data collected.3)The dynamic structural response of commonly used vehicle models was calculated to obtain the strain peak and strain variation characteristics.Then the characteristics of the measured strain data are preliminarily analyzed,the vehicle model parameters are determined and the load identification calculation is carried out.The load identification effect is good.The dynamic deflection response is obtained by applying the identified load to the main beam model with modified stiffness.After safety assessment analysis,the conclusion that the bridge is in good condition is judged.