IDENTIFICATION OF DEFLECTION INFLUENCE LINES OF CABLE-STAYED BRIDGES UPON MIXED MODAL DECOMPOSITION

In order to effectively remove the dynamic components in the vehicle-induced bridge response and to obtain more accurate quasi-static loading time course curves, an influence line identification method based on empirical variational hybrid modal decomposition is proposed, and the three methods of Empirical Mode Decomposition,Variational Mode Decomposition and Permutation Entropy-Variational Mode Decomposition are compared and verified by the deflection influence line identification test of cable-stayed bridge. The vehicle is transformed into a unit concentrated load, the vehicle information matrix is constructed, The axle effect is peeled off by the Tikhonov regularization method to identify the stable solutions of the deflection influence lines of cable-stayed bridges, The method proposed is verified by the data simulation and by the actual engineering cases. It is found that: the empirical variational hybrid modal decomposition is the more effective in removing the dynamic effects compared with the other three single processing methods, with a relative error of only 2.77% with respect to the simulation, and its identification error is no more than 4% when the vehicle speed is increased to 120 km/h.