Experimental investigation and analytical prediction of the rotational stiffness of apex connections in cold-formed steel portal frames

• Four configurations of cold-formed steel portal frame apex connections were experimentally investigated • Connections with flange bolts in addition to web bolts relocated failure away from the connection and showed moderately increased stiffness • Connections with top flange bolts and an additional haunch showed a 44% increase in stiffness over the connection with only web bolts • Brackets without a continuous flange exhibited premature failure in lateral-torsional buckling • A component model is presented to predict the moment-rotation behaviour of these types of connections An experimental campaign was devised to investigate the rotational stiffness and capacity of four different types of apex connections for use in cold-formed steel portal frames. Type A connections were implemented by bolting only the rafter channel webs to a central bracket, an often encountered design in practice which provided a benchmark configuration. These connections were observed to fail by instability of the cross-section containing the first line of bolts. Type B connections featured additional bolts connecting the top flanges of the rafter channels and brackets. This configuration successfully moved the location of failure away from the connection and resulted in a modest increase in the rotational stiffness during the initial bearing phase of 12%. Type C connections contained an additional haunch to accommodate bottom flange bolts. A more significant increase in rotational stiffness of 44% over the Type A connections was achieved, and the slip phase in the moment-rotation curve was mostly eliminated. Type D connections also contained a bottom haunch and benefitted from a direct connection between bracket and channel top and bottom flanges. The rotational stiffness exceeded that of the Type A connections by 73%. However, the connection became prone to lateral-torsional buckling of the connection bracket in the gap between the bracket flanges, resulting in premature failure. Based on the results of additional lap shear tests and material coupon tests, a component model was developed for the Type A, B and C connections which was eminently able to predict the experimentally observed connection behaviour.