Time-variant Structural Parameter Identification Method Based On Adaptive Forgetting Factor

With the development of civil engineering,structural health monitoring has become an important method for evaluating the safety of structures in the operation stage.Parameter identification is an important part of structural health monitoring.The external environment or human factors may cause structural parameters to change and cause damage.In actual engineering,structural damage often changes with time,and accurate tracking of changes in structural parameters has become a key problem to be solved in structural health monitoring.This paper takes the time-variant structure as the research object,and introduces the adaptive forgetting factor on the basis of the unscented Kalman filter method,so that it can track and identify the changes of the physical parameters of the structure.The main research contents of this article are:(1)The effect of the Unscented Kalman filter(UKF)method on the parameter recognition of time-variant linear and nonlinear structures is studied.The research results show that the traditional UKF method has a poorer effect on the parameter recognition of time-variant linear structures.To solve the above problems,an adaptive forgetting factor is introduced in the process of UKF parameter identification,and an improved adaptive forgetting factor UKF method is used to track time-variant parameters.The method is used to identify and simulate the time-variant structural parameters,which verifies the effectiveness of the method.(2)Considering the seismic action,perform numerical simulation for parameter identification of time-variant five-DOF linear structure and three-DOF hysteretic nonlinear structure model.The research results show that the adaptive forgetting factor UKF method has high accuracy in identifying structural mutations,linear gradual changes and nonlinear gradual parameters,with good tracking effects,strong anti-noise ability,and effective identification of structural parameters.(3)For the time-variant structure under the moving load,a simple supported beam finite element numerical model based on Euler-Bernoulli beam element is established,taking into account incomplete monitoring information and the influence of factors such as the size,speed,simulation error of the moving load.The physical parameters of the structure are expanded into the state vector,and the elastic modulus and damage identification of the simply supported beam element are numerically simulated.The research results show that the adaptive forgetting factor UKF method has good effects and applicability for simply supported beam structure parameter tracking and damage identification.