Standard versus natural fire resistance for partially restrained composite floor beams: 2-Analysis

Abstract

A thermo-structural reduced-order finite element modeling approach is used to evaluate the results of two fire tests (presented in Part 1 of this study) on identically designed, highly composite floor beam specimens with SFRM protection, shear tab connections at each end, a lightweight concrete slab on a corrugated metal deck, and partial restraint of both axial thermal expansion and connection rotation. Model validation is performed for both an ASTM E119 standard fire and a parametric natural fire, with a constant load applied to the specimens via four-point bending to a maximum service condition. Predictions of temperature increase in the protected steel beam and the composite slab show good agreement with experimental measurements by implementing a series of weighted averaging strategies on thermal FE results for 2D cross-sections at the thick and thin portions of the composite slab. Predictions of temperature-induced deflections, which used the results of thermal FE analysis as input, also show good agreement with experimental results when little-to-no rotational stiffness is applied to the connections. The numerical models are then re-analyzed at a reduced loading level (representing a middling service condition) and with varying rotational stiffness in the shear tab connections. The results of parametric analysis are used to draw correlations between standard fire resistance and natural fire robustness as a function of applied loading and temperature increase in the steel beam. Simplified modeling strategies and recommendations for performance-based structural-fire design are discussed, and opportunities for future research are highlighted.