Research On Parameter Identification Of Bridge Structure Based On Filter Technique

Over the years,the construction of infrastructure in China has become more and more perfect,especially the bridge construction has achieved remarkable achievements.The bridge structure is developing at a high speed in the direction of longer span and more complicated structure.Therefore,the public has put forward higher requirements for the safety of the structure,in which the high-precision structural damage identification is a top priority.Currently in the field of structural damage identification,many scholars have developed effective theoretical methods and computational models.With the advantage of the most intuitive reflection of structural state characteristics,the use of the extended Kalman filter(EKF)for physical parameter identification has become an important method for the study of sheartype building damage identification,but the theoretical basis for the application of bridge damage identification is still relatively weak.Therefore,based on EKF,the physical parameters of bridges are identified around three methods and the damage identification information is obtained:(1)The derivation process of the state equation and observation equation of the beam bridge structure under the bridge integral parameter identification method is carried out.One case is used to analyze that the method is not sensitive to damping,for which reason the quality and stiffness of the bridge are taken as the parameters to be identified.No matter the applied excitation is a regular harmonic load or an irregular actual seismic wave,the bridge overall parameter identification method always shows high noise immunity,stability and accuracy.However,there is a significant difference between the recognition speed and the initial value selection;The increase of the degree of freedom and the existence of damage will not cause a significant reduction of recognition accuracy,and the stiffness parameter can be used as an effective parameter to determine the damage of the structure;In addition to Gaussian white noise,superimposing non-Gaussian t-function noise will reduce the accuracy of physical recognition but the accuracy is always stable within a certain range.It is suitable for the situation where a stable identification value is required;(2)The method called two-unit physical parameters identification based on EKF is introduced.The adjacent sub-units are taken as research objects,in which way the matrix dimension and programming difficulty is effectively reduced.Compared with the bridge integral parameter identification method,it can obtain higher stiffness parameter identification accuracy,which is suitable for identifying bridges with stiffness damage.However,when the same bridge is identified,the convergence speed of two-unit physical parameters identification method is slightly slower than the convergence speed of the bridge integral parameter identification method.Since the single recognition can only obtain the physical parameter identification values of two adjacent units,the two-unit physical parameters identification method is more suitable for bridges with generally known damage locations;(3)The static condensation method based on EKF is deduced,by which way the physical parameters of the beam bridge structure considering the rotation effect can be identified and the damage can be identified only by obtaining the translational response.By comparing the recognition accuracy under the three numerical integration methods,it is found that the recognition accuracy of the fourth-order Runge-Kutta integral is better,and the stiffness identification accuracy of the damaged unit is lower than the undamped stiffness.At the same time,the noise immunity of the stiffness parameter and the damping coefficient is very good.The closer the initial value of the stiffness parameter is to the real value,the higher is the recognition accuracy.When the initial value error is too large or the order of magnitude is different,there will be a situation where the convergence cannot be converged or the identification value error is far more than 10%.The research results of this paper provide theoretical basis and data support for the application of EKF in the field of beam bridge structure damage identification.