An experimental study of non-sway loaded and rotationally restrained steel column assemblies under fire conditions: analysis of test results and design calculations

Abstract

Abstract This paper presents the results of an analysis of two series of experimental studies of non-sway loaded and rotationally restrained steel column assemblies under fire conditions. Although there have been a number of theoretical studies of restrained column behaviour in fire, the main objective of this study is to fill an important gap in experimental results. A separate paper has described the objectives of these experiments, their set up, instrumentation, an assessment of the testing system and a brief summary of test results. This paper gives results of a detailed analysis of the tests. The main focuses of this paper are variations in bending moment and effective lengths of the test columns. The main variables of these fire experiments are the level of axial load in a column, the initial column bending moment and the type of connection. It has been observed that bending moments in a rotationally restrained column change in a complicated manner under fire conditions. However, for fire resistant design, from comparisons between column failure temperatures of different tests and between recorded column failure temperatures and those calculated using BS 5950 Part 8 and Eurocode 3 Part 1.2, the initial bending moment and type of connection have negligible influences. In particular the following conclusions may be drawn: • The column failure temperature was mainly dependent on the axial load in a column and was not affected by the level of initially applied bending moment and also was not affected by the type of beam to column connection. • Using both BS 5950 Part 8 and Eurocode 3 Part 1.2 to calculate column failure temperatures or failure loads, better agreements between test and calculated results are reached when column bending moments are ignored and the column effective length ratio is taken as 0.7.