Bayesian parameter identification of hysteretic behavior of composite walls

• The laboratory tests performed for multi-grid composite wall specimens are presented.
• A hysteretic model is proposed to model the behavior of multi-grid composite wall specimens under lateral cyclic loading.
• A Bayesian framework is presented for identification of the hysteretic model parameters with the test data.
• Identified plausible models are investigated and compared with the test data to verify the quality of identification.

A Bayesian probabilistic approach is applied for parameter identification of a hysteretic model using laboratory test data in this paper. A hysteretic model for multi-grid composite walls is proposed to model the behavior of multi-grid composite wall specimens under lateral cyclic loading. Effects of stiffness degradation, strength degradation and pinching are considered. The test data consists of observed hysteretic curves of precast composite wall specimen, composite wall specimen reinforced by light steel, retrofitted composite wall specimen and cast-in-place composite wall specimen. Using the Bayesian approach, the identification results are presented in terms of the most probable value and posterior covariance matrix of model parameters. The implied hysteretic and backbone curves with their uncertainties identified based on the test data are compared with their observed counterparts.