Stress concentration factors of cold-formed high strength steel tubular T-joints

Highlights

• SCF of cold-formed tubular T-joints of S900 and S960 steel grades were investigated.

• Traditional, brace-rotated, square and diamond bird beak tubular T-joints were included.

• Member-rotation had an evident effect on the SCF values of tubular T-joints.

• Experimental SCF values were compared with the SCF values calculated from the CIDECT and literature.

• Existing SCF parametric equations are unable to predict the SCF of cold-formed tubular T-joints of S900 and S960 steel grades.

Abstract

A test programme looking into the stress concentration factors (SCF) of cold-formed high strength steel (CFHSS) traditional (without brace and/or chord rotation) and member-rotated tubular T-joints is described in this paper. The brace members of traditional T-joints were made of square, rectangular and circular hollow sections (SHS, RHS and CHS), while the brace members of member-rotated T-joints were made of SHS and RHS. On the other hand, the chord members of traditional and member-rotated tubular T-joints were made of SHS. In this study, member-rotated T-joints, include brace-rotated (BR), square bird-beak (SBB) and diamond bird-beak (DBB) tubular T-joints. The tubular members were cold-formed from thermomechanical control processed (TMCP) S900 and S960 steel grades plates. A total of 12 tests was conducted in this experimental investigation. An axial compression load was applied through the brace member, while the chord ends were supported on rollers. The brace and chord members of tubular T-joints were connected using the robotic gas metal arc welding process. The values of normal brace width-to-chord width ratio ($\beta$) ranged from 0.33 to 0.73, effective brace width-to-chord width ratio ($\beta$’) ranged from 0.40 to 0.87, brace thickness-to-chord thickness ratio ($\tau$) ranged from 0.66 to 1.28 and chord width-to-chord thickness ratio (2$\gamma$) ranged from 20.6 to 39.0. The quadratic extrapolation method was used to estimate hot spot strains at the weld toes using the strains measured in the extrapolation regions. The experimental SCF values of CFHSS traditional and member-rotated tubular T-joints were compared with the SCF values obtained from the parametric equations given in the CIDECT (Zhao et al., 2001, [1]) and studies (Tong et al., 2013, [2]; Li et al., 2018, [3]; Tong et al., 2015, [4]). It was found that the existing SCF parametric equations given in the literature (Zhao et al., 2001, [1]; Tong et al., 2013, [2]; Li et al., 2018, [3]; Tong et al., 2015, [4]) could not accurately predict the SCF of cold-formed traditional and member-rotated tubular T-joints made of S900 and S960 steel grades.

Introduction

Tubular members are commonly used in both onshore and offshore structures, which are regularly subjected to various dynamic loads. Over a period of time, due to the cyclic nature of these loads, cracks may develop at the joints, which could result in the fatigue failure. Generally, fatigue cracks initiate from the weld toe due to high stress concentrations at the weld vicinity. As fatigue failure is a cumulative process, thus, cracks can start, even when the applied load is substantially lower than the strength of the structural element. In this study, a total of 5 tubular T-joint configurations was investigated, including two types of traditional tubular T-joints and three types of member-rotated tubular T-joints, as shown in Fig. 1. The traditional tubular T-joints included RHS–RHS (where braces and chords were made of RHS) and CHS–RHS (where braces were made of CHS and chords were made of RHS) T-joints. On the other hand, member-rotated T-joints included brace-rotated (BR), square bird-beak (SBB) and diamond bird-beak (DBB) tubular T-joints. BR T-joint is a joint where only the brace member is rotated along its longitudinal axis by an angle corresponding to its cross-sectional diagonal plane. SBB T-joint is a joint where only the chord member is rotated along its longitudinal axis by an angle corresponding to its cross-sectional diagonal plane. However, DBB T-joint is a joint where both brace and chord members are rotated along their respective longitudinal axes by angles corresponding to their cross-sectional diagonal planes. The cross-sectional diagonal plane of a tubular member is shown in Fig. 1.

The parametric equations to determine the stress concentration factors (SCF) of traditional tubular joints, including the RHS–RHS configuration, are given in the CIDECT [1]. Over the last few decades, many studies were conducted to determine the SCF of various uniplanar and multiplanar tubular joints [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21]. Tong et al. [2] studied the SCF of CHS–RHS T-joints subjected to brace axial and brace in-plane bending loads. The tubular members were made of Q345B steel grade. Using the results of the comprehensive parametric study, Tong et al. [2] proposed SCF parametric equations for CHS–RHS T-joints. Li et al. [3] experimentally and numerically investigated the SCF of SBB and DBB X-joints made of Q345D steel and subjected to brace axial loads. It was found that the highest SCF occurred at the saddle locations of these X-joints. In addition, the SCF of SBB X-joints were generally found smaller than the SCF of corresponding identical DBB X-joints. Moreover, hot spot locations and their corresponding SCF parametric equations were proposed for SBB and DBB X-joints. The SCF of Q235B steel grade DBB T-joints were experimentally and numerically investigated by Tong et al. [4]. The test specimens were subjected to brace axial and brace in-plane bending loads. A comprehensive numerical parametric study was performed by duly incorporating a broad range of critical geometric parameters, and subsequently, hot spot locations and their corresponding SCF parametric equations were proposed for DBB T-joints. Noordhoek et al. [5], [6] conducted a series of tests on RHS–RHS K-, N- and KT-joints, including 134 K-joints, 67 N-joints and 16 KT-joints. These joints were made of both hot-rolled and cold-formed hollow sections, where the cold-formed steel tubular joints showed a slightly better fatigue performance compared to the hot-rolled joints. Dooren et al. [7] carried out a preliminary numerical investigation on 12 RHS–RHS K-joints with gap. The welds were modelled using the shell elements. The influence of secondary bending moment was considered in the numerical model. In addition, one experiment was also carried out for the SCF measurement and also for the validation of the numerical model. Ferreira and Branco [8] experimentally investigated the SCF of RHS–RHS T- and Y-joints. In addition, a numerical study was also conducted to investigate the influence of the weld shape as well as the 3D crack propagation law. Mang et al. [9] carried out a detailed experimental and numerical investigation on the fatigue strength of welded unstiffened RHS–RHS K-joints, with gap and overlap, and subjected to axial and in-plane bending loads. Soh and Soh [10] performed a detailed numerical parametric study to investigate the SCF of RHS–RHS T-, Y- and K-joints subjected to axial loads, in-plane and out-of-plane bending moments. A total of 21 SCF prediction equations was proposed, where the SCF equations were expressed as the function of the governing geometric parameters. Shahi [11] performed experimental and numerical investigation on the SCF of RHS–RHS multi-planar KK-gap, XX-, KY-, YY-joints, as well as uniplanar K-gap, X- and Y-joints. The welds were modelled in the numerical investigation, and the maximum SCF occurred at heel-weld and toe-weld for brace and chord members, respectively.

Kurobane et al. [12] conducted cyclic tests on cold-formed CHS–CHS (where braces and chords were made of CHS) T-joints made of mild steel grade, where the fatigue performance of five types of test specimens was studied after approximately 100,000 loading cycles. Fatigue cracks in most test specimens were initiated at the points of stress concentration near the weld toes located at the centre section of the chord and brace members. Toprac et al. [13] conducted an experimental study to investigate the fatigue strengths of different types of uniplanar CHS–CHS joints, including T-, K- and TK-joints. It was found that the point of crack initiation at the weld toe occurred at the highest stress location. Kuang et al. [14] numerically investigated the SCF of CHS–CHS T-, K- and TK-joints. The numerical analysis was performed using the shell elements, wherein the welds were not modelled. A detailed parametric study was performed for each type of CHS–CHS joint, and subsequently, empirical equations were proposed to predict the SCF at the critical locations of the joints under investigation. Wordsworth and Smedley [15] proposed empirical SCF formulae for unreinforced CHS–CHS T-, X- and Y-joints subjected to axial, in-plane and out-of-plane bending loads. Efthymiou and Durkin [16] presented a complete set of SCF equations for the design of CHS–CHS T-, Y-, and K-joints under axial, in-plane and out-of-plane bending loads. Efthymiou [17] introduced the influence function concept, which takes into account the carry-over effect in multi-planar CHS–CHS joints. Smedley and Fisher [18] developed SCF parametric equations for uniplanar CHS–CHS T-, Y-, K-, X- and KT-joints. However, the proposed equations could not accurately predict the SCF distribution along the welding path. Morgan and Lee [19] proposed a set of parametric equations to determine the SCF for CHS–CHS K-joints under axial loading. Chang and Dover [20], [21] proposed SCF prediction equations to determine the stress distribution along the intersection of CHS–CHS Y-, T-, X-, and DT-joints. The proposed equations were developed for axial, in-plane bending and out-of-plane bending loadings.

From this comprehensive literature review, it can be seen that the cold-formed high strength steel (CFHSS) tubular joints of S900 and S960 steel grades are not yet investigated for their SCF. Hence, a meticulous experimental investigation was conducted and summarised in this paper with the objectives of not only examining the applicability of the current SCF parametric equations for cold-formed S900 and S960 steel grades tubular T-joints but also to observe the effect of member-rotation(s) on the maximum SCF values in the brace and chord members of these T-joints.