Experimental verification of buckling of curved web close steel section (original) (raw)

Today, bridge girders with curved flanges and webs are becoming common in order to increase the aesthetic value and to improve the quality of the structure. Despite the use of these new shapes, not much research has been done in this field. The aim of the present research is to develop an experimental test setup, during which box girders with changing web curvature are subjected to a constant shear force. The specimens were made from plastic sheets (with thickness 0.125 mm) to create the webs and flanges, and MDF wood to create the external stiffeners. In a later phase, numerical models are created, which have the same properties and characteristics as the experimental specimens. These include a height of 70 mm, a width of 30 mm, and curvature radii of 1500, 5, 3, 5/3, 1 and 0.55 times the web height. The aim of the test is to analyse the elastic buckling behaviour of the webs and the failure condition of the specimen. The experimental results are analysed and compared with numerical results, for which no imperfections were included. Tests indicate that the elastic buckling load is increasing with increasing curvature, whereas for the failure load the opposite happens. The numerical models show identical behaviour, with higher loads. The failure load behaviour changes according to flat web or shell behaviour of the specimen. The results showed that flatter webs have much more postbuckling capacity than the most curved webs. The deformation pattern of the web is also different for both categories. Where the flatter webs had diagonal tension lines, the most curved webs had more horizontal bulges, shifting to the upper web-flange juncture. The difference between the numerical and experimental results can be blamed to the lack of imperfections implemented in the first approach. Earlier numerical research 1 with closed steel sections, show that the failure load line is increasing with growing curvature. The high web slenderness for the experimental specimen causes opposite results for the failure load. For an increasing curvature, the post-buckling capacity is decreasing, while the elastic buckling capacity is increasing. This all results in an increasing or decreasing failure load curve.