Bayesian Nonlinear Finite Element Model Updating of Bridge Columns considering Bond-slip and Rebar Buckling using Experimental and Simulated Data

• Author / Creator

  Zhenning Liu

• Reinforced concrete (RC) bridges are key elements in modern transportation system. Therefore, the serviceability and safety of such structures after earthquakes have become an area of great interest. Developing an accurate FE model that is capable of representing the important mechanics of RC structures subjected to earthquakes is important, for example, in post-event damage assessment. In particular, an updated FE model based on recorded response of an instrumented structure during an earthquake can be used to infer the hidden damage structures experienced. Among various finite element model updating (FEMU) strategies, Bayesian parameter inference (e.g., unscented Kalman filter approach) has been well received in the past decades.
  This study focused on the Bayesian nonlinear FEMU of a shake-table tested full-scale RC bridge column, with particular emphasis on learning advanced modeling aspects (e.g., bond-slip and steel rebar buckling). Using prior knowledge about the design information of the column, a nonlinear fiber-based finite element model was developed in the open-source FE software framework (OpenSees). Due to the unknown/uncertain modeling aspects, the initially developed FE model was inaccurate. It was found that the perfect bonding assumption was inappropriate, and bond-slip needed to be taken into account. Furthermore, the dynamic response data from the shake-table tests were utilized for the nonlinear FEMU to help achieve an accurate prediction on the structural behavior. Potential parameters for updating were selected by combining the in-site observations and relative importance of the model parameters, which were determined through a comprehensive sensitive analysis. The credibility of the updated model was validated by comparing FE predictions with the experimental measurements during each ground motion, in both local (i.e. strain) and global (i.e. drift ratio) scale. Moreover, damage evolution of the RC bridge column in a sequence of earthquake ground motions were also studied. The update results affirmed that the nonlinear FEMU was capable of identifying multiple unknown/uncertain modeling aspects (e.g., bond-slip related parameters) from experimental observations and showed that stochastic FEMU approach was capable of enhancing the modeling accuracy with measurements.
  Considering that the RC column expressed steel rebar buckling in the shake-table test, the buckling effect of steel rebar was further introduced into the FE model. Thus, this study also applied the FEMU approach to RC bridge columns considering both bond-slip and steel rebar buckling. With different steel material models available to incorporate the asymmetrical behavior due to steel rebar buckling, material coupon test data collected from the literature were used to assess their material model performance. With the most versatile buckling steel material model, the FE model considering both bond-slip and steel rebar buckling was used to simulate noisy seismic response data considering measurement error. Nonlinear FEMU was then conducted to examine the capability of the nonlinear FEMU approach in learning the buckling-related modeling parameters in addition to other unknown parameters (e.g., bond-slip parameters). The FEMU results have proven that the stochastic FEMU approach was capable of identifying buckling related parameters based on recorded seismic response data. This implies its potential applications for post-event damage assessment.

• Subjects / Keywords

  • Bridge Column
  • Bond-Slip Effect
  • Reinforcing Steel Buckling
  • Shake-Table Test
  • Unscented Kalman Filter
  • Model Updating

• Graduation date

  Spring 2021

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/r3-dffa-1x77

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.