Thermal and Mechanical Computational Study of Load-Bearing Cold-Formed Steel Drywall Systems Exposed to Fire

The structural fire behavior of typical load-bearing drywall systems, consisting of light cold-formed steel (CFS) studs, sheathed with single or double layers of gypsum plasterboards and air cavity or mineral wool insulation, is investigated. For that purpose, an efficient numerical modeling approach with satisfying accuracy is developed and validated against experimental and numerical results available in the literature. Heat transfer and structural analyses of the investigated drywall assemblies are carried out, employing the computational tools ANSYS CFX and ADINA, respectively. Transient heat transfer analyses provide the spatial and temporal temperature variation over the drywall assembly, which is assumed to be exposed to the standard ISO 834 fire curve from one side. Thermal simulation results are then introduced in the subsequent structural analyses, where the load-bearing capacity of the CFS studs under fire conditions is estimated, by means of Geometry and Material Nonlinear Analyses accounting for Imperfections (GMNIA). The load-bearing capacity assuming uniform cross-sectional temperature distributions is also evaluated and compared to the respective predictions under non-uniform temperature profiles. Results obtained by varying the imposed load or the applied temperature (fire duration), are compared to predictions of existing and literature-proposed fire design guidelines. EN 1993-1-2 provisions lead to average underestimation of the load-bearing capacity by approximately 20%, while the literature-proposed methods are found to underestimate the load-bearing capacity by 5% to 15%. The limitation of using a critical temperature of 350°C for the structural fire design of thin-walled sections is found to be rather conservative, since the investigated CFS studs are able to exhibit satisfying fire performance also under higher temperature conditions. © 2016, Springer Science+Business Media New York.