Nonlinear Structural Parameter Identification Method Based On Adaptive UKF

The parameter identification of the structure is affected by noise and modeling uncertainty.In actual engineering,facing with problems such as complex loads,parameter time-varying,initial value errors,and statistical characteristics mismatch.It is difficult to guarantee the recognition accuracy and convergence speed.The proposed adaptive method solves some problems to a certain extent.Based on the above problems,this paper introduces three adaptive UKF methods into the process of structural parameter identification of civil engineering.The main research contents are as follows:(1)The principle and process of the UKF method and three adaptive UKF methods are introduced and analyzed.The advantages and adaptive principles of the three adaptive UKF methods are analyzed.Four methods are used to identify the Bouc-Wen model and time-varying stiffness,and the accuracy and convergence speed of the recognition results were compared.The Chebyshev orthogonal decomposition is used to identify the structural loads and parameters simultaneously.The effectiveness of three adaptive UKF methods is verified.(2)The adaptive UKF method based on the update of the Open Sees model is introduced.The frame model and the grid structure model are established in the Open Sees software,and the elastic modulus of the material is identified.The UKF method and the dual adaptive UKF method are used for identification respectively.The numerical simulation results verify the effectiveness of the adaptive UKF method based on the Open Sees model update.(3)A dual adaptive UKF method for faded memory is introduced,which improves the robustness of the method.This method can reduce the influence of initial noise and far-step noise on the current result as the time step increases by adding the faded memory factor and exponential weighting in the dual adaptive UKF method.Numerical simulations confirm that the method has the effect of accelerating convergence.