Performance-based steel building system reliability assessment by introducing criticality and failure gate theory

Conventional seismic assessment often treats components in isolation, obscuring how local damage propagates through a frame and reduces reliability. This paper proposes a performance-based system-reliability framework that ranks component criticality by coupling global reliability with connectivity and load-path redundancy. The first contribution is an explicit integration of Performance-Based Design and Load and Resistance Factor Design. Multi-state demand–capacity checks at Immediate Occupancy (IO), Life Safety (LS), and Collapse Prevention (CP) are mapped into a system reliability space, allowing design targets and acceptance criteria to be set and verified with a single, coherent set of reliability measures. The second contribution is a Failure Gate methodology for system criticality. Combinatorial limit-state functions generate a system reliability index; a Dynamic Redundancy Measure quantifies the loss of reserve capacity relative to the intact system; and an Element Reliability Matrix (EMR) captures interaction and topology, including series and parallel effects, as well as weak seams. The gate model, consistent with plastic-hinge mechanics, removes a member only when both ends lose the ability to transfer moment, thereby avoiding penalties for partially functional elements. The resulting criticality index combines reliability, redundancy, and connectivity losses to yield transparent, reproducible rankings. Applications to 3-, 9-, and 20-story steel moment frames demonstrate that the loss of lower-story columns incurs the most significant reliability penalties and that failure trajectories are essentially topology-driven rather than record-specific. The method scales with height, guides retrofit where reliability loss is most acute, and clarifies target reliability assignment by linking component damage to system performance.