Development of Rotation Capacity Model for Web Cleat Connections against Progressive Collapse: Bayesian Approach

Abstract

Catenary action is recognized as one of the main resistance mechanisms for typical steel frame structures against progressive collapse. As a result of catenary action, large rotation, and axial tension demand are applied to the steel connection. Since the failure of the steel connection under catenary action is dominated by the tensile fracture of the steel component, the ultimate rotation capacity of a steel connection is controlled by the ultimate deformation capacity of the steel component. This paper proposes a calibration method for web cleat connections to evaluate the rotation capacity under column loss scenarios. The proposed method incorporates both component tests of bolted angle under tensile load conditions and structural tests of web cleat connections under column loss scenarios. First, the ultimate deformation capacity of bolted angle is developed based on the component test data. Next, the relationship between the ultimate deformation capacity of the component and the rotation capacity of the steel connection is investigated. An analytical model is established to evaluate the rotation capacity of a steel connection. Finally, based on the experimental tests of web cleat connections under simulated column loss scenarios and component tests of bolted angles, the likelihood function and the prior density function of the model parameters are developed. The component tests and structural tests are combined using Bayesian theory and the Markov chain Monte Carlo simulation method. Comparison studies show that the developed model achieves higher model accuracy against the existing equations. The proposed calibration method provides new insight into the development of the rotation capacity of steel connections, especially when the experimental test of steel connections against progressive collapse is limited.