An Empirical Spring Model for Simulating Bolt Fracture Incorporating Uncertainty

Bolt fracture constitutes a primary failure mode, particularly in partial-strength bolted steel connections, under extreme hazards such as collapse-level earthquakes and progressive collapse scenarios. Simulating bolt damage and fracture is key to studying connections’ ductility and the robustness of steel structures under such hazards. In computational modeling, spring models are commonly used to represent the bolt component. The spring model constitutes a practical alternative to continuum finite element simulations that require refined mesh sizes, large computation power, and lengthy procedures to calibrate ductile fracture material models. An empirical model is proposed to accurately capture carbon steel bolts’ nonlinear response up to failure. The model is built on establishing an empirical relation between the bolt’s plastic elongation and its grade and thread length. This is made feasible through an experimental dataset on bolts assemblies under pure tension. The model can be assigned to axial connectors in finite element simulations, axial springs in mechanical component-based simulations, or employed as part of design procedures. Most importantly, bolt fracture uncertainty is quantified to support reliability and performance-based engineering studies. The model validity is demonstrated through experimental validations at the component and connection scales for different bolt material grades and connection topologies.