Probabilistic Temperature-Dependent Mechanical Properties of Cold-Formed Steel

This study develops and applies probabilistic models for the temperature-dependent mechanical properties of cold-formed steel (CFS) to support structural fire performance assessment under uncertainty. The yield strength and elastic modulus of CFS decrease with increasing temperature, which affects the fire response of CFS structural members. A comprehensive experimental database was compiled to calibrate two types of probabilistic models, lognormal distributions and continuous logistic functions, for the retention factors of elastic modulus and 0.2% proof stress at elevated temperatures. These models were then incorporated into a Monte Carlo simulation framework to quantify the variability in fire resistance of eight end wall columns from a prototype metal building. Structural capacity was evaluated using the direct strength method (DSM) in accordance with ANSI/SDI AISI S100-24, and failure temperatures and times were determined through coupled thermal-structural analyses under ASTM E119 fire, as well as a design-basis fire exposure. Results show that both probabilistic models produce consistent fragility trends. Compared with deterministic assessments, the probabilistic framework offers a more comprehensive view of performance variability and provides a robust tool for risk-informed, performance-based fire design of CFS structures.