Development of a new cold-formed steel roof truss system

This thesis reports on a research and development project which was aimed at developing a new cold-formed steel roof truss system for Textor Metal Industries Pty Ltd under a National Teaching Company Scheme. It presents details of analysis, design, and testing methods used in this project and their results. This Master of Engineering Degree research thesis was undertaken part-time during the project duration of two years, and includes in part some research details which are confidential. The new cold-formed steel roof truss system was designed for non-cyclonic permissible design winds of 33 mis and 41 mis. The system was developed for a maximum span of 12m, and a roof pitch of 20°-25° for a concrete tile roof, and 15°- 250 for a metal sheet roof Both the web and chord members are open thin-walled sections made from 0.8 mm - 1.0 mm G550 Zinc/Alume coated steel, and they were connected by a new fastening technology known as clinching. The web members were plain channel stud sections and the chord members were a new stiffened top plate section. In order to obtain accurate strength capacities of the thin-walled sections a rigorous material and geometric non-linear finite element analysis (FEA) was conducted on the plain channel stud web members undergoing local buckling using the Nastran package in combination with the elastic finite strip method Contour software. The method was verified by stub column tests, and considered several material grades and thicknesses. The FEA results of section capacity were compared with the predictions of the coldformed steel structures code AS 153 8 for small member lengths (ie approximating section capacity). The sections demonstrated the significant post-buckling behaviour expected of these thin-walled open sections. The comparisons of the FEA results with AS1538 predictions gave unexpected results. For some cases the FEA result was higher than the AS1538 result, but for others the AS1538 result was higher. It was clear that the difference between the FEA and AS 153 8 results was not consistent, so the new roof truss system members were designed using AS1538 with an effective length departure of 0.7L. AS 1170 .2 Section 3 was used for the determination of wind loads in preference to AS4055. AS4055 was found to determine wind loads substantially less than AS 1170.2 Section 3 due to the adoption of a revised risk philosophy which is concerned with loss to the overall community rather than loss to individual structures. This was not accepted for this research. Wind parallel to the ridge plus dead load was the critical load combination for both strength and serviceability design of the trusses. The trusses were modelled with fixed web member ends and pinned truss supports using a plane frame analysis program (SpaceGASS). Design actions from this analysis and strength capacities from AS1538 (including the 0.7L effective length departure) were used to design the thickness and grade of material required for the plain channel stud section web members and the stiffened top plate chord sections. The member connections of the new roof truss system were clinch fasteners developed by Tox Systems Pty Ltd. A single shear test program of 685 samples was conducted to determine the behaviour and capacity of these fasteners. Several multiple fastener tests were also conducted. Design strength results from the test program were used in the design of the clinch connections of the new roof truss system. A series of five full scale roof truss tests were conducted on an 8m span truss with a metal sheet roof and a 22. 5° roof pitch to validate the analysis and design assumptions used in the development of the new truss system. The single and three truss tests investigated the ultimate and serviceability behaviour of the trusses and the system effect of multiple trusses due to the presence of roof and ceiling components. A pneumatic loading system was used to simulate the critical wind load combination of wind parallel to the ridge and dead load. Both downward pressure wind and upward suction were simulated, but the upward suction wind case was taken to failure as this was critical. The test program verified the design and analysis assumptions used in this research and demonstrated the capacity of the new roof truss system to sustain the design loads. A comparison of alternative analysis assumptions was performed and showed that modelling the truss with pinned web member ends and/or pin/roller truss supports gave predictions which were not in agreement with the test results. The assumptions of fixed web member ends and pinned truss supports agreed well with the test results, and hence validated these assumptions used in the design of the trusses. The trusses in the new roof truss system were designed using the results of the analysis in combination with the member design in accordance with AS1538 (including the O. 7L effective length departure) and the fastener design in accordance with the test program. The full scale testing program confirmed that this analysis and design method was suitable for the trusses. The test truss performed adequately under both ultimate and serviceability loads. Comparison of test results with analysis and design showed that the member failure loads were in good agreement with predictions, and that the mode of failure of the members was also in good agreement with the predictions. On this basis the same process of analysis and design was used for the other trusses