Identification Of Dynamic Axle Loads On Bridge Based On Piecewise Polynomial Truncated Singular Value Decomposition

Bridge moving force identification belongs to the category of structure dynamics.Most methods of bridge moving force identification translate into a linear equation combination at last,and different solving methods lead to different identification accuracy.In order to improve the moving force identification accuracy of time domain method(TDM)with measurement noise and measuring point change,based on the basic theory of the first identification method,the second identification method,frequency domain method and time domain method,combining with the characteristics of bridge moving force identification,several regularization methods are discussed in this paper,and a piecewise polynomial truncated singular value decomposition(PPTSVD)algorithm for moving force identification is developed.The numerical simulation with twelve cases of two-axle constant force and two kinds of time-varying moving force under simplified Euler beam model were analyzed.The moving dynamic axle loads were identified from bridge bending responses,acceleration responses or their combination with various noise levels based on PPTSVD.The identification accuracy of PPTSVD compares with TDM and truncated singular value decomposition(TSVD),and the comparison results of PPTSVD and TDM show that identification error increases with measurement noise in both methods.When the noise level reaches 10%,there are eleven cases of the twelve identification error more than 100% based on TDM while ten cases of the twelve identification error less than 30% based on PPTSVD.The robust noise immunity of PPTSVD is much better than TDM.The results also show that identification accuracy influenced by responses type and responses combination obviously in both methods.The identification accuracy increases with more acceleration responses included in identification case.When identify moving loads on bridge from bending responses alone,the identification results can't be accepted based on TDM while the identification results have high accuracy based on PPTSVD.PPTSVD has many good properties compared with TDM,such as better identification accuracy,robust noise immunity and less effect of response type and response combination,which is beneficial to the application of PPTSVD in the field identification of dynamic axle loads on bridge.