| description abstract | Exposed column base plate (ECBP) connections are commonly used in steel moment resisting frames. Current approaches for their design are well-established from a mechanistic standpoint. However, the reliability of connections designed as per these approaches is not as well understood. A detailed reliability analysis of the prevalent approach in the United States is performed in this study by using 59 design scenarios from steel moment frames subjected to combinations of dead, live, wind, and seismic loads. The analysis is conducted through Monte Carlo sampling reflecting uncertainties in the loads, material properties, component geometry, as well as demand and capacity models for the various components (e.g., base plate, footing, anchor rods) of the connection. Results indicate that the current design approach leads to unacceptable and inconsistent probabilities of failure across the various components. This is attributed to: (1) the use of a resistance factor for the footing bearing stress that artificially alters flexural demands on the base plate, and (2) the calibration of resistance factors for the plate and anchors without appropriate consideration of variability in demands. Two alternative approaches are examined as prospective refinements to the current approach. One eliminates the resistance factor for the bearing stress when used to determine flexural demands in the base plate, while the other considers overall failure of the connection rather than failure of individual components within the connection. For both approaches, new resistance factors are calibrated to provide consistent and acceptable probabilities of failure across all limit states and all types of loading. Design and cost implications of these alternative approaches are summarized. | |
