LAUSR

Abstract Material properties, residual stress distributions and cross-sectional behavior of cold-formed steel elliptical hollow sections are investigated in this study. Four cross-section series with the nominal section aspect ratio ranging from 1.65 to 3 were included in the experimental investigation. The material properties for each cross-section series and material properties distribution on half of the cross-section profile of a representative section were measured through tensile coupon tests. The distributions of bending and membrane residual stresses in both longitudinal and transverse directions were measured on the half-section profile of the same representative section. Initial local geometric imperfections were measured on five stub column specimens. Besides, stub column tests were conducted between fixed ends to ascertain the material properties of the complete cross-section in the cold-worked state as well as to study the structural behavior of cold-formed steel elliptical hollow section stub columns. In addition to experimental investigation, a finite element model was developed and verified against the test results, with which an extensive parametric study covering a broad range of cross-section geometries was carried out. Currently, there is no codified design rule for elliptical hollow section compression members. The stub column strengths obtained from experimental program and numerical analysis were only compared with the predicted strengths by the equivalent diameter method and equivalent rectangular hollow section approach proposed by previous researchers for design of hot-finished steel elliptical hollow sections, the existing traditional design rules originally developed for circular hollow section with equivalent diameter incorporated as well as the Direct Strength Method and the Continuous Strength Method that the equations were not calibrated for cold-formed steel elliptical hollow sections. The comparisons show that the Direct Strength Method offers the most accurate and reliable design strength predictions among the existing design methods, but further improvement remains possible. In this study, modifications on the Direct Strength Method and the Continuous Strength Method are proposed, which are shown to improve the accuracy of the design strength predictions.