Moving Force Identification Based On Acceleration And Strain Signal

In the research of bridge health monitoring,moving loads on bridges is an important factor to influence the life of the bridge,moving load identification belongs to the second category of inverse problems based on the dynamic response of bridge.In this paper,based on the summary of the moving load identification methods proposed by domestic and foreign scholars,the method of reconstructing the dynamic displacement is used to identify the moving load,specific research contents are as follows:(1)The achievements of the research about the identification of moving loads are summarized,four kinds of identification methods of IMI?IMII?TDM and FTDM are summarized,which are improved and optimized,based on the research of the causes of load identification error,the method of digital signal reconstruction is proposed to solve the problem of noise signal in order to reduce the recognition error.(2)For the displacement signal on the bridge caused by moving loads,which is decomposed into static load caused by part of the static displacement and inertial force caused by vibration displacement,the static displacement can be obtained by using empirical mode decomposition by strain signal,after removing the high frequency part of the strain history curve,the low-frequency part can be regard as static strain response,the least squares method is used to solve the beam,the relationship between curvature and vertical displacement of the vertical displacement caused by static displacement;For the vibration displacement part,the frequency domain integral method is used to obtain the signal,and the signal noise reduction is needed before integration.Finally,the displacement signal is reconstructed by the two parts of the static displacement and the vibration displacement.The numerical example and the experimental results show that the method can obtain the dynamic displacement of the structure accurately.(3)Comparing the time domain integral algorithm and frequency-domain integral algorithm,frequency domain integral algorithm in several times to the trend still can guarantee the integral precision,the trend signal generated in the integration process can be removed by the least squares fitting method.By using a set of superimposed signals of multiple sinusoidal digital noise reduction analysis of adaptive noise reduction effect was studied when the convergence factor of reasonable value,the algorithm can get stable and high degree of reduction of signal,the signal after de-noising can meet the requirements to achieve the integration of the target.(4)A simple test for verification,the test is divided into two parts,include a modal test of the whole structure and a moving load test,the dynamic characteristics of the structure were obtained by modal test,and the structure stiffness values were calibrated;test of moving loads were measured in different speed and load each measuring point of the dynamic response,the strain response after decomposition is used to identify the static displacement,the vibration displacement is integral noise acceleration signal after two times,compared with the measured displacement and reconstructed dynamic displacement,it is found that the curves fit very well,the displacement signal reconstruction from three points of the moving load identification.Either the IMII or TDM method can accurately identify the moving load under different working conditions,the signal denoising recovery to improve the load identification accuracy has certain help.