This paper examines the cyclic behavior of a bio-based truss structure with slotted steel plate bolted connections through a comprehensive approach involving experimental testing, FE modeling, and Bayesian model updating. Experimental tests were performed on both connection and truss levels to determine the cyclic behavior. For the joint level, a total of four distinct joint configurations were tested to determine their axial hysteretic behavior. For the truss level, cyclic tests were carried out on four hybrid planar trusses composed of the four types of truss joints. To define the axial hysteretic behavior of the joints, a robust numerical model is first defined accounting for hysteretic behavior in OpenSees. The proposed model can capture sliding, contact, pinching, strength degradation, and failure behavior. A model reduction via parameter sensitivity analysis is first performed. A Bayesian parameter identification based on Sequential Monte Carlo method was performed based on cyclic truss joints tests. A model of the hybrid truss was developed in OpenSees by integrating the joints and elastic truss elements. The developed model underwent validation using various experimental parameters, including the hysteresis curve, relative displacement at joints, and strain values at each chord. To correct the errors and the bias, a Bayesian model updating framework based on the Sequential Monte Carlo method is proposed herein to deal with model uncertainty. The proposed framework enables the acquisition of highly precise predictions by incorporating global and/or local responses while appropriately accounting for model uncertainties.
Read full abstract