Parameter Identification Of Non-linear Seismic Continuous Beam Bridge Under Earthquake Action

The isolation continuous beam bridge separates the superstructure of the bridge from the motion of the ground motion or support which may cause damage to the earthquake by installing an isolation device on the pier as far as possible,thus,the traditional continuous beam bridge has been widely used in the current engineering projects because it can reduce the damage of the superstructure under earthquake action.A suitable mechanical model of isolated bridge is used to identify the physical parameters of the isolated bridge in real time and present the current performance state of the isolated bridge.It is of great significance to the safety assessment and maintenance decision of the bridge and the safety of the lifeline engineering.(1)On the basis of comprehensive comparative analysis of simplified methods for isolated continuous beam bridges and various nonlinear mechanical models of supports and piers,this paper adopts a model method of single pier isolated beam bridges simplified to two particles.The bouc-wen model with rubber bearing characteristics as the nonlinear model of bearing hysteresis and the classical bouc-wen model as the elastoplastic mechanics model of bridge pier are selected to establish the nonlinear analysis model of isolated continuous beam bridge.Based on the extended Kalman filter(EKF)algorithm suitable for on-line identification,the substructure method of the isolated continuous beam bridge is derived from the substructure method,and the single degree of freedom bouc-wen hysteresis nonlinear model of the isolated bearing and the single degree of freedom bouc-wen model of the pier are introduced step by step.From the single pier model to the two-span continuous beam model of isolated continuous beam bridge,the on-line recursive real-time parameteridentification is realized by using extended kalman filter algorithm.The adaptive extended kalman filter(EKF)algorithm is applied to identify the time-varying parameters of nonlinear isolated bridges.The single degree of freedom hysteresis bouc-wen model of isolated bearings and the single pier isolated bridge model are studied.The identification of single time-varying parameter and multi-time-varying parameter is investigated respectively.The main conclusions are as follows:(1)The single pier isolated beam bridge is simplified as a two-particle model.The bouc-wen model with rubber bearing characteristics is used as the support hysteresis nonlinear model,and the classical bouc-wen model is used as the elastoplastic mechanical model of the pier.The mechanical behavior of isolated continuous beam bridge under earthquake action can be correctly described.(2)On the elastoplastic restoring force model of all isolated continuous beam piers,compared with the broken linear restoring force model,the classical bouc-wen curve model can not only reflect the real-time variation characteristics of the RC pier restoring force model,but also be more suitable for the on-line identification of structural physical parameters in time domain.(3)Whether it is bouc-wen model of single freedom isolation support,single degree of freedom bouc-wen model of pier,single pier model of isolated continuous beam bridge or two-span continuous beam model,the extended kalman filter method can be used when the information of external excitation and acceleration response is known.The parameters of the nonlinear structure and the displacement and velocity responses of the structure are identified.The recognition speed is fast and not affected by external excitation and noise.The identification accuracy and speed of physical parameters are lower than those of model parameters,and the recognition accuracy decreases slightly with the increase of identification parameters.For the parameter whose absolute value is too small,it is easy to be affected by the parameter with cross-correlation,and the result of identifying fluctuation is large.For nonlinear isolated bridges with abrupt structural parameters,the single degree of freedom hysteresis bouc-wen model of isolated bearings and the single pier isolated bridge model are identified by adaptive extended kalman filter with single time-varying parameters and multiple time-varying parameters.It is found that the adaptive extended kalman filter algorithm is insensitive to the number of time-varying parameters to be identified and the combination of different physical variables.The ability of tracing and identifying parametersaccurately and effectively.The tracking and identification ability of AEKF is of practical significance for on-line monitoring and damage diagnosis of nonlinear isolated bridge structures.