Strength and Reliability of Steel Frames with Random Properties

Abstract

This paper studies the effects of random properties on the strength and reliability of two steel frames. Monte Carlo simulation is used to analyze the behavior of a low-rise industrial frame and a grain storage bin. Both structures exhibit significant second-order effects with their strength limited by loss of frame stability. Both frames have been previously studied for the development of the direct analysis provisions of the 2005 AISC Specification for Structural Steel Buildings. The analyses consider randomness in yield strength, elastic modulus, residual stresses, and out-of-plumb (P–Delta) and out-of-straightness (P–delta) imperfections. The effects of spatial correlation of these properties are also studied. The simulation results are used to estimate probability distributions of frame strength and reliability, or probability of failure. The results provide probabilistic relationships between material and geometric properties of steel frames and the overall strength of the frames. Randomness in elastic modulus and residual stresses are shown to have a significant impact on strength and reliability of the frames by increasing lateral deflections and thereby second-order forces. The results confirm the importance of including geometric imperfections for proper assessment of frame stability.