Elsevier

Thin-Walled Structures

Volume 102, May 2016, Pages 286-304
Thin-Walled Structures

Ultimate bending capacity of spiral-welded steel tubes – Part I: Experiments

https://doi.org/10.1016/j.tws.2015.11.024Get rights and content

Highlights

  • Investigation of material properties and spiral-welded tubes.

  • Investigation of initial geometrical imperfection of spiral-welded tubes.

  • Large scale four point bending tests on thirteen spiral-welded tubes.

  • Inclusion of girth welds and coil connection welds in the investigation.

  • Inclusion of girth welds and coil connection welds in the investigation.

  • Analysis of results from the initial measurements and bending tests.

Abstract

The present investigation refers to the bending capacity of spiral-welded steel tubes. The first part of this investigation presents the results of a full-scale experimental program, aiming to investigate the structural behavior of large-diameter spiral-welded steel tubes under bending. A companion paper (Part II) is also published, which further studies the behavior of these elements numerically, using finite element simulations.

The testing program presented in Part I consists of thirteen 42-inch-diameter spiral-welded steel tubes with D/t ranging between 65 and 120. Some of the tubular specimens contain girth welds and coil connection welds, which are shown to penalize the ultimate bending capacity of the tubes. Extensive measurements of initial imperfections and material properties are performed for each tubular specimen. The material properties of the specimens are investigated through both uniaxial tensile and compression coupon tests. A series of large-scale four-point bending tests is performed to determine the structural behavior of the tubes, resulting in local buckling failure of the tubes under consideration.

The bending behavior of the thirteen specimens is documented extensively. The study offers information with regard to the ultimate bending resistance of the specimens. In addition, the full moment–curvature equilibrium path is presented, supplemented by measurements on the development of cross sectional ovalisation and tube wall wrinkling during the bending tests.

Introduction

An economical and efficient method to manufacture relatively thin-walled large-diameter steel tubes is offered by the spiral-welded (or helical-welded) manufacturing process (HSAW). This process consists of spiral welding of a steel plate from a hot-rolled steel coil, as shown schematically in Fig. 1. Common spiral-welded tube diameters range from about 500 to 3000 mm, with wall thicknesses between 9 and 25 mm. The manufacturing process is continuous; a steel coil that runs out is connected to a new coil by means of a butt weld without interruption of the spiral-welding process. This weld, running perpendicular between two spiral welds, is denoted in this study as a coil connection weld (CCW). In addition to these coil connection welds, girth welds are also present in elements used in practice. These girth welds may be executed in the controlled environment of the production plant, but for the connection of adjacent tube segments, girth welds can also be executed on-site.

Large-diameter spiral-welded tubes are employed in onshore hydrocarbon and water pipeline applications. The diameter of hydrocarbon pipes can range from 700 mm up to about 1500 mm, with D/t ratios between 40 and 100. In addition to hydrocarbon pipelines, large-diameter steel pipelines and penstocks for water transmission are very often made of spiral-welded tubes. The diameter of large-diameter steel tubes for water transmission typically ranges from 900 to 2000 mm, with rather high diameter-to-thickness ratios; D/t ratios can range up to 240. Spiral-welded tubes are also used in structural applications, for example for tubular piling, towers, masts and other large tubular structures. An important structural application is the use of large-diameter steel tubes in combined walls as primary structural elements that resist horizontal loads from soil and water pressure. Those combined walls, often referred to as “combiwalls”, consist of a series of large diameter tubes connected by infill sheeting (see Fig. 2). For their connection, standard sheet piling slots are welded to the tube. Steel tubes used in combined walls are generally manufactured with the spiral welding process, with a diameter range up to 3000 mm, thickness up to 25 mm, specified minimum yield strength between 350 MPa and 480 MPa, and length up to 50 m, whereas typical values of the corresponding diameter-to-thickness ratio range from 65 to 120. An advantage of using the continuous spiral-welding manufacturing process for tubes in this application is the capability of producing tubes of significant length, thereby minimizing the number of girth welds in the structure.

The present work is motivated primarily by the need of determining the structural capacity of large-diameter spiral-welded tubes, employed in combiwall applications [3], [4]. In this application, the main loading condition is bending of the tube, and the corresponding dominant failure mode is local buckling of the tube wall. The linear and nonlinear bending behavior of spiral-welded tubes is also important for the response and safe design of large-diameter pipelines, subjected to severe ground-induced deformations in geo-hazard areas. In those areas, the pipeline is subjected to significant bending loading due to fault movement, liquefaction-induced lateral spreading and subsidence, as well as landslide action which may threaten its structural integrity.

Longitudinal bending deformation of a tubular member induces ovalisation of the tube cross-section, a special feature of tube bending, also known as the “Brazier effect” [5]. Ovalisation reduces the tube bending stiffness because of flattening, increases the local radius of the tube cross-section at the maximum compression location, and introduces a biaxial stress state, because of ring bending, leading to early yielding of the tube. Upon increasing bending deformation, structural instability of the compressed tube wall occurs, in the form of a localized wavy pattern, referred to as “local buckle”, “wrinkle” or “kink”. The formation of a local buckle is associated with a bending moment drop in the moment–curvature equilibrium path of the tubular member; this means structural failure of the tubular member. Based on previous experimental observations, thin-walled tubes with D/t values equal or above 100 subjected to bending develop wrinkles quite rapidly, leading to sudden collapse. On the other hand, in bent tubes with D/t values equal to about 60 or 70, the development of wrinkles is more gradual.

The pure bending response of metal tubes has been investigated experimentally numerous times in the last decades. Early experimental work was conducted by Moore and Clark [6] on specimens machined from aluminum-alloy rolled rod with D/t ratios ranging from 2 to 150. Their scaled experiments include bending, compression and torsion testing of which the former two are of interest for the current research. Specimens with D/t ratios larger than 15 were found to be prone to instability, either due to progressive ovalisation or the formation of a local buckle. The influence of axial tension on local buckling was found to be positive by Wilhoit and Merwin [7] in similar scaled bending experiments. A large number of small scale bending tests has been performed by Johns et al. [8]. The use of small scale experiments allowed the application of combined load of external pressure and bending. By validating the large number of scaled tests with a limited number of full scale four-point bending tests, understanding of the local buckling mechanism was further increased. More bending tests on small scale specimens were performed by Tugcu and Schroeder [9], whose research includes mainly tests on tube branches besides a few tests on plain tube, and Reddy [10], who tested tubing with D/t ratios up till 80 made of steel and aluminum. The latter investigation also includes the measurement of longitudinal profiles of the tubing at the onset of local buckling. Testing on slender tubes (D/t=81−102) of intermediate scale was performed by Van Douwen et al. [11], focusing on the application of pipelines in settlement areas. Small-scale specimens have been tested by Kyriakides and Shaw [12] and Kyriakides and Ju [13]. In these two publications combined, more than twenty tests on steel and aluminum specimens under monotonic or cyclic bending are reported.

Early large scale testing was performed by Jirsa et al. [14]. The work features six four-point bending tests on tubes up to 20 inch, two of which are concrete-coated. The moment–curvature diagrams that result from the tests were simulated by an analysis method published earlier by Ades [15], which allows a cross-sectional analysis of the bending behavior of the specimens, including the Brazier effect, but cannot capture the formation of a local buckle. Further large scale testing was performed by Sherman [16]. The testing program includes tubes with D/t ratios between 18 and 102 with various loading conditions, leading to a combination of bending moment and shear force at the critical cross sections of the specimens. More recently, more than fifty bending tests on plain and girth welded tubes were performed at the University of Alberta, which are documented in Dorey et al. [17], Delcol et al. [18], Dorey et al. [19], Mohareb et al. [20] and Yoosef-Godsi et al. [21]. The tests were performed on specimens with D/t ratios between 48 and 92. The results present a negative influence of the presence of a girth weld in the critical cross section.

One of the first investigations into the effect of the manufacturing technique of the tubes on their bending behavior was performed by Bouwkamp [22], who performed bending tests on longitudinal welded and seamless tubes. In another work by the same author [23], experimental investigations on the local buckling behavior of large diameter tubes (48 in.) have been reported. The work by Van Foeken and Gresnigt [24], [25] has further studied the influence of the manufacturing method of a tube on its local buckling behavior, by comparing the collapse and bending behavior of UOE manufactured tube with seamless tubes.

Experimental investigation on the bending behavior of spiral-welded tubes is relatively rare. The work by Zimmerman et al. [26] includes bending experiments on four spiral-welded tubes with D/t ratios between 48 and 82. Two of these experiments included internal pressure resulting in a hoop stress of 80% of the specified minimum yield strength. A direct experimental comparison with other manufacturing techniques has not been conducted, but a comparison is made with experimental data from the literature. The study concludes that the spiral weld seam may not be detrimental to tube performance and that spiral-welded line pipe performs as good as longitudinal welded linepipe in terms of local buckling behavior. Further experimental testing on spiral-welded tubes was performed by Salzgitter Mannesmann, reported by Zimmermann et al. [27]. In that research, four-point bending tests were performed on four spiral-welded tubes. During the bending tests in this investigation, two of the specimens were pressurized up to a level leading to a hoop stress of 66% of the specified minimum yield strength. In comparison with common design standards, the performance of the specimens is deemed to be satisfactory. Neither the work in [26] or the work in [27] includes a study into the influence of a girth weld or coil connection weld on the local buckling behavior. Further investigation of the behavior of spiral-welded tubes including a girth weld in the critical bending zone is announced in [27], but has not yet been published. Eight bending tests on spiral-welded tubes have been performed by Reinke et al. [28] as part of the same European research project the present investigation is part of. The research concludes that the experiments show that the spiral welded tubes performed well and that the influence of spiral-welding on the bearing behavior of the specimens is negligible.

The present investigation is part of a large research project with acronym COMBITUBE, funded by the European Commission in the framework of the RFCS program [29]. The investigation is aimed at examining the bending behavior of large-diameter steel tubes, focusing on the determination of their ultimate bending moment capacity and the corresponding buckling curvature, taking into account the effects of the spiral-welding manufacturing process. Furthermore, typical features of spiral-welded tubulars used in structural applications, such as a girth weld or a coil connection weld, are included. The current research includes a thorough experimental investigation on 42-inch-diameter steel tubes with extensive measurements of the initial geometry of the tubes. Aside from the large-scale experimental testing, numerical simulations with finite elements are also included in the investigation. In this paper, denoted as Part I, the experimental testing and corresponding results are presented in detail. The companion paper by Vasilikis et al. [30], referred to as Part II, reports the numerical simulation of the experimental procedure and an extensive parametric numerical study. The pair of papers are aimed at providing a comprehensive overview of the bending behavior of large-diameter spiral-welded tubes. Part I starts with the discussion of the manufacturing technique in Section 2, whereas the presentation of the 42-inch-diameter tubular specimens and the experimental setup are offered in Section 3. Measurements of initial imperfections and extensive material testing are presented in Section 4. Section 5 describes the experimental results from the full-scale tests. Finally, conclusions are stated in Section 6.

Section snippets

Brief description of the spiral-welded manufacturing method

The spiral-welding manufacturing technique is an economic solution for producing large-diameter tubes with a relatively high slenderness; i.e. with relatively small wall thickness. Commonly available diameters range from about 500 mm up to 3000 mm with wall thicknesses ranging from about 9 mm to 25 mm. The length of the produced tubulars may range from 8 m to about 50 m. Due to the continuous production, long lengths can be manufactured. The main limits on length are the available room in the plant

Outline of testing program and specimens

To investigate the local buckling behavior of spiral-welded tubes, a large-scale experimental program has been performed. The program features investigation of the initial geometry of the tubes, extensive material testing and four-point bending tests.

The program consists of thirteen 42-inch-diameter spiral-welded steel tubes. All specimens have a specified outer diameter of 1067 mm, with D/t ratios varying from 65 to 120, a length of 16,500 mm, and steel grades varying from X52 to X70. The

4.1. Tensile and compressive material properties

An overview of the results of the tensile tests on machined coupon specimens is presented in Table 2 in terms of tensile yield stress corresponding to 0.2% permanent plastic strain. The results show that, on the average, the longitudinal yield strength is 6.7% higher than in hoop direction. Furthermore, the yield strength on the inside of the tube wall is about 3.7% lower than on the outside of the tube wall. These values differ among the tubular specimens, but show similarities among tubes

Results of four-point bending tests

As a result of the bending loads exerted, all specimens have failed in the form of local buckling. The specimens with D/t equal to 120 failed suddenly and violently. In contrast, the transition between a stable and unstable tube wall has been much more gradual for the thick walled specimens. Especially the thick walled specimens with large initial imperfections, such as specimen T6 (see Fig. 22) have had a very smooth transition between the pre-buckling stage and post-buckling stage. Some of

Conclusions

This paper presents a summary of the results obtained from thirteen full-scale four-point bending tests on 42-inch-diameter spiral-welded steel tubes. Initial geometry, wall thickness and diameter have been carefully documented before testing. Furthermore, an extensive study into the material properties of the tubular specimens has been performed.

In the bending tests, curvature localizes before the maximum bending moment is reached due to variations in bending moment capacity over the length of

Acknowledgements

Funding for this work has been provided by the Research Fund for Coal and Steel (RFCS) of the European Commission, project COMBITUBE: “Bending Resistance of Steel Tubes in CombiWalls”, Grant agreement no. RFSR-CT-2011-00034.

References (35)

  • J.C.Wilhoit, J.E. Merwin, Critical plastic buckling parameters for tubing in bending under axial tension, in:...
  • T.G. Johns, R.E.W.R. Mesloh, J.E. Sorenson, Inelastic buckling of pipelines under combined loads, in: Proceedings of...
  • A.A. Van Douwen, A.M. Gresnigt, J.W.B. Stark, Plastic Design of Buried Steel Pipelines for Transport of Oil, Gas or...
  • S. Kyriakides et al.

    Inelastic buckling of tubes under cycling bending

    J. Press. Vessel Technol.

    (1987)
  • J.O. Jirsa, L. Fook-Hoy, J.C. Wilhoit, J.E. Merwin, U. Rice. Ovaling of pipelines under pure bending, in: Proceedings...
  • C.S. Ades

    Bending strength of tubing in the plastic range

    J. Aeronaut. Sci.

    (1957)
  • D.R. Sherman

    Tests of circular steel tubes in bending

    J. Struct. Div. ASCE

    (1976)
  • Cited by (47)

    • Bending behaviour of flax fabric-reinforced epoxy pipes

      2021, Composites Part A: Applied Science and Manufacturing
    • Static flexural behaviour of concrete-filled spiral-welded stainless-steel tubes

      2020, Thin-Walled Structures
      Citation Excerpt :

      Similar variability in local buckling location, with respect to the spiral weld seam, was also observed by the authors for comparable mild-steel hollow SWTs [23]. Consistent with the authors' findings, Van Es et al. [44] and Zimmerman et al. [45] also observed that local buckling at failure generally occurred away from and near to the weld seam respectively, for the carbon-steel SWTs they tested. The ‘wrinkly’ nature of the local buckles seen in the tested SWSSTs increased with increasing slenderness, consistent with that observed for the corresponding mild-steel SWTs [23].

    • Static flexural behaviour of concrete-filled spiral-welded mild-steel tubes

      2020, Journal of Constructional Steel Research
      Citation Excerpt :

      Furthermore, as a result of advances in automation and welding methods [8–10], SWTs can be produced with a comparable degree of reliability to that of LWTs [5]. However, high levels of residual stresses with more complicated distribution patterns compared to LWTs have been reported for SWTs [11–14]. In addition, it has been further found that the spiral welding process causes unique fabrication induced surface imperfections [15].

    View all citing articles on Scopus
    View full text