In recent years, considerable research efforts have been devoted to rectangular hollow section (RHS) joints, aiming to re-examine suitability of codified design provisions for steel grades up to S460 and to extend the validity range beyond S460. Within the research framework of drafting the next-generation international design standard ISO 14346 for hollow section joints, this paper presents recent research advances and improved design methods for RHS joints under brace axial compression or tension, failing by essential failure modes. For chord sidewall failure, the representative modified IIW method and an alternate Lan-Kuhn method together with the subsequently proposed Kim-Lee method are compared and evaluated against existing test and numerical evidence. Further, an upper limit of the codified chord sidewall bearing length is proposed to avoid overestimations for thick-walled RHS joints. For interactive chord sidewall and chord face failure, more-suitable limiting beta ranges for the occurrence of this failure mode and a linear interpolation design method are suggested. Design rules and recommendations are then proposed for chord sidewall failure and interactive failure, which have been approved to be included in the next draft of ISO 14346.

In this study, existing deformation limits are first re-examined to investigate if they can be rationally extended to high strength steel circular hollow section (CHS) joints that are subjected to in-plane bending (IPB) moment. It is pointed out that existing deformation limits, which have been developed and validated primarily for mild steel joints, need to be modified when high strength steels are involved. By noting that the ductility of IPB-loaded joints can be significantly reduced with the use of high strength steel, a new deformation limit is proposed which allows less deformation to less ductile high strength steel joints. The deformation limit, proposed in terms of joint rotation angle, is validated both numerically and experimentally. In addition, a recently proposed strain limit criterion is also discussed. To provide a guide for obtaining converged strain from finite element (FE) analysis, mesh sensitivity study is comprehensively conducted. It is shown that the element size required for the convergence of strain is substantially smaller than that required for obtaining satisfactory global response such as joint load-deformation relationship. By applying a systematic FE modeling strategy, numerical investigation is made to check the feasibility of the 5% strain limit criterion which has recently been advocated by the revised draft of ISO 14346. While the limiting principal strain of 5% is shown to be reasonable for CHS-to-CHS joints loaded by IPB, for longitudinal branch plate-to-CHS joints, a lower limiting strain appears more appropriate.

Within the research framework for updating the international design standard ISO 14346 for hollow section joints, this study examines brace failure and chord sidewall failure in full-width rectangular hollow section (RHS) X and T joints under brace axial tension, brace in-plane bending and brace out-of-plane bending. The codified design rules for full-width RHS joints are reviewed, with their limitations highlighted. Design resistance equations for brace failure, based on the modified effective width method, and also for chord sidewall failure, using the modified bearing–buckling method and the Lan–Kuhn method, are then proposed. Up-to-date experimental and numerical results are collated from the literature, which cover a wide range of geometrical parameters, steel grades ranging from S235 to S960, varying weld details, and loading cases of brace axial tension, brace in-plane bending and brace out-of-plane bending. The compiled results are adopted to evaluate the performance of full-width RHS joints and to assess suitability of the proposed design resistance equations. It is shown that weld details can significantly affect the deformation capacity and static strengths of full-width RHS joints. The proposed design resistance equations yield conservative and reliable strength predictions for full-width RHS joints. Welding guidance and user-friendly design rules, in which an extension to include Class 3 cross-sections is included, are suggested for full-width RHS X and T joints.

It is well known that the current design rules adopted by international design codes such as ISO 14346 and design guides, e.g., the CIDECT design guide No. 3, for chord sidewall failure in mild steel rectangular hollow section (RHS) joints under brace axial compression are considerably conservative, if the RHS joints are adequately supported out-of-plane. This paper presents an investigation into chord sidewall failure in RHS joints using steel grades up to S960. Representative existing design methods for chord sidewall failure in RHS joints are reviewed, and two alternative design methods, i.e., the modified bearing–buckling method and the Lan–Kuhn​ method, are proposed. Up-to-date test and numerical results reported in the literature are compiled. A wide range of geometric parameters, steel grades up to S960 and loading cases of brace axial loading, brace in-plane bending and brace out-of-plane bending are covered. The existing and proposed design methods are assessed against the collated results. The effects of brace-to-chord height ratio, brace angle, steel grade and chord stress ratio are evaluated. It is shown that the proposed design methods can provide more consistent resistance predictions for chord sidewall failure in mild steel and high-strength steel RHS joints under brace axial compression. Corresponding user-friendly design rules are suggested. The design of chord sidewall failure in RHS joints under brace axial tension, brace in-plane bending and brace out-of-plane bending is discussed. Further required research on, in particular, high-strength steel RHS joints is highlighted.

This paper deals with the evaluation of fatigue cracks under a concentrated compression (wheel) load in an I-section with full penetration welds between the web and flange. The objective is to investigate whether cracks stop or nearly stop when they have grown through the residual tensile stress field. These experimental investigations are part of a review of a crane runway girder where after 20 years of service fatigue cracks were observed in the flange at the toe of the full penetration weld. The fatigue analysis of the actual crane runway girder is described in (Wardenier et al., 2017). The fatigue tests under a concentrated wheel compression loading show that, for the specimens considered on a scale of about 1:2 with stiffeners at one side, the cracks only initiate and grow at the non-stiffened side to about 50 to 60% of the web thickness and then stop. Based only on the nominal stress range under the wheel, determined according to EN 1993-6 and neglecting the shear stress effect, an equivalent fatigue class of about 160 N/mm² was found for crack initiation in the web, whereas the minimum ratio in life between visually observed crack initiation and maximum crack length was about a factor 3. Comparison of the codes for a wheel loading in compression shows large discrepancies in effective width and fatigue classes to be used.

This paper is part of an evaluation of fatigue cracks in a crane runway girder with full penetration welds between the web and flange. The fatigue analysis of this actual crane runway girder is described in [1]. The investigation described in this paper deals with additional experimental tests on equivalent welded I sections on scale of approximately 1:2 subjected to a fluctuating line load in compression. The objective of these experimental investigations is to investigate whether cracks stop or nearly stop when they have grown through the residual tensile stress field. The test results show that, in some cases the cracks in the weld, at the weld toe with the web or with the flange initiate and grow from one side to about 50 to 60% of the web thickness and then stop. However, at the weld toe with the flange the cracks grow sometimes from both sides but with the cracks at one side having a small length and/or a small depth. The minimum ratio in life between crack initiation and maximum crack was a factor 1.2 for cracks occurring at one side only and 1.5 to 3.1 for cracks at both sides.