Comparison of analysis with tests of cold-formed RHS portal frames

Different types of structural analyses, first order elastic, second order elastic, first order plastic, and plastic zone analysis, were used to simulate the behaviour of 3 portal frames constructed from cold-formed rectangular hollow sections. All analyses underestimated the magnitude of the deflections of the frame mainly due to the non-inclusion of flexible joint behaviour. A second order inelastic analysis, which included gradual yielding, multi-linear stress-strain curves, and member imperfections underestimated the deflections and the hence the second order effects, and therefore slightly overestimated the strength of the frame. A second order inelastic analysis that did not account for member imperfections provided the best estimates of the frame strengths, as it omitted the beneficial effects of the unsympathetic imperfections while underestimating the magnitude of the second order effects. 1 PORTAL FRAME TESTS As part of a project to investigate the suitability of cold-formed rectangular hollow sections RHS for plastic design (Wilkinson and Hancock 1997, 1998, 1999), three portal frames were tested under simulated gravity and transverse wind load. The general layout of the frames is shown in Figure 1. Each frame spanned 7 metres, with an eaves height of 3 metres, and a total height of 4 metres, constructed from 150 × 50 × 4.0 RHS in either Grade C350 or Grade C450 (DuraGal) steel to AS 1163 (Standards Australia 1991). There was a collar tie which joined the midpoint of each rafter. Each frame was pin based. There was a welded internal sleeve connection between the column and rafter as shown in Figure 3. The apex joint is shown in Figure 4. Full details of the connections are in Wilkinson and Hancock (1999). The frame was laterally braced at several critical locations. A downwards vertical load was provided via an MTS actuator connected to a gravity load simulator (which ensured the load remained vertical as the frame swayed). A horizontal point load acted on the north column, via steel cables hung over an auxiliary frame connected to lead weights. Table 1 briefly summarises the results of the portal frame tests. The value of yield stress presented in Table 1 is from a coupon cut from one web of each specimen. There was variability of yield stress around the section, and considerable strain hardening. Full details on stress-strain curves can be found in Wilkinson and Hancock (1999). Figure 2 indicates where local buckles formed in the frames. Figures 5 and 6 show the load-deflection curves. Strong floor Gravity load simulator 1000