Nonparametric Restoring Force Identification For MDOF Structures Under Partial Measurements

Identification of initiation and development of damages in engineering structures under the excitation of strong dynamic loadings plays key roles in remining load-carrying capacity and remaining life forecasting.Structural restoring force provides a direct description on the initiation and development of structural damages under dynamic loadings.Structural restoring force can be used to quantitively evaluate energy assumption during vibration and the initiation and development procedure of different structural members during vibration is helpful for the understanding of structural failure patterns.Due to the diversification of materials and forms of engineering structures,it is hard to use a general assumed parametric model to describe different nonlinear behavior of various engineering structures.Moreover,in practice,the complete structural dynamic response measurement and structural parameters such as mass are often unknown.Therefore,it is of great significance to develop nonparametric identification approaches for structural nonlinear restoring force and mass using partially available dynamic response.In this study,a nonparametric structural mass and nonlinear restoring force identification approach using only partial acceleration response and external excitation measurement is proposed and validated numerically and experimentally.This study is meaningful for post-event damage identification of engineering structures excited by strong dynamic loadings where structural nonlinearity should be considered.The main contents of this thesis are as follows:1.A structural nonlinear restoration force and mass identification approach in a nonparametric way based on Extended Kalman Filter(EKF)and Hermite Polynomial(HP)when structural parameters including mass,damping,stiffness and partial dynamic response measurements are unknown is proposed.The principle of the proposed approach and the flowchart of the algorithm are introduced in details.2.The feasibility of the proposed nonlinear restoring force and mass identification approach using HP is verified numerically with a multi-degree-of-freedom chain-like nonlinear structure equipped with a Magneto-Rheological damper(MR),which is used to mimic structural nonlinearity.The effectiveness and robustness of the proposed method for restoring force,mass and unknown dynamic response identification considering the effect of measurement noise,different masses,and nonlinearities at different locations of the structure are validated by comparing the identification results with the true values.3.In order to further investigate the feasibility of the proposed approach,experimental validation is conducted with a four-story frame structure with a MR damper on its fourth floor to mimic nonlinear behavior of the structure.Forced vibration tests on the nonlinear model structure is carried out and limited acceleration response and excitation measurement are employed for the identification.The nonparametric identification method proposed in this study is used to identify the nonlinear restoration force,structure mass and the unused dynamic response of the structure and the identification results are compared the test measurement.The results show that the proposed approach is efficient in identifying structural restoring force,unknown mass and dynamic response even using limited acceleration measurement made in the test.