Global mechanical tensioning for the management of residual stresses in welds

https://doi.org/10.1016/j.msea.2007.12.042Get rights and content

Abstract

The general principles behind the global mechanical tensioning technique for controlling weld residual stresses are examined using a finite element model to follow their evolution throughout the welding process. While we focus specifically on friction stir welding, the tool is represented simply as a heat source. As a result, the findings have relevance to a wide range of welding processes. For aluminium alloy friction stir welds, the maximum longitudinal weld stresses have been reported to fall approximately linearly to zero under mechanical tensioning to a level around 40% of the yield stress. Under larger tensioning levels, the weld stress becomes increasingly compressive. This behaviour is explained in terms of the reduction in compressive plastic straining ahead, and an increase in tensile plastic straining behind, the heat source as the tensioning level is increased. Finally, it is shown that tensioning during welding is much more effective than post-weld tensioning.

Introduction

Friction stir welding (FSW) is being used increasingly in a wide range of applications, particularly for joining aluminium alloys. The basic process has been described extensively elsewhere [1], but in essence involves a rotating tool consisting of a cylindrical shoulder and pin. The tool is plunged into the weld line until the shoulder is in contact with the plate surface. Once the material is sufficiently hot from frictional heating and plastic work, the tool traverses along the weld line and the hot plasticised material is extruded past the rotating pin, while constrained between the shoulder and backing bar, so as to form a joint behind the pin. Because FSW is a solid state welding method, it is particularly suited to joining high strength aluminium alloys that were previously considered unweldable using fusion techniques [1]. Although the process alleviates many of the metallurgical problems associated with fusion welding, such as liquation and solidification cracking, friction stir welds can still suffer from significant levels of residual stress, which are often similar in magnitude to those seen in fusion welds [2], [3], [4]. In general terms, the residual stresses arise from plastic misfit strains introduced as a result of the steep gradients in temperature that are generated local to the heat source as the tool advances [5]. As tensile residual stresses in welded structures produced from high strength Al-alloys can have a negative impact on service life [4], [6], [7], it is highly desirable to reduce their level as far as possible. One approach, that can be adopted to mitigate weld residual stresses, is to use weld tensioning methods to engineer the local stress state during welding by controlling the plastic misfit strains generated by the thermal field.

In welding the maximum tensile residual stresses are typically found on, or either side of, the weld line. These arise during cooling of the weld as a result of the compressive plastic misfit generated as the material expands and softens ahead of the heat source [5]. It has long been known that tensioning techniques can reduce residual stresses and the concomitant tendency for distortion (e.g. [8]). In practice, a large number of techniques have been proposed for controlling residual stresses in welding, including both thermal and mechanical tensioning methods. One of the earliest reported applications was by Greene and Holzbaur [9], who in 1946 used superimposed temperature gradients to achieve reduced residual stresses in longitudinal butt welds in ship hull structures. Local induction heating has also been investigated for residual stress improvement [10]. Michaleris and Sun [11] and Dull et al. [12] have applied thermal tensioning to reduce buckling distortion, whilst Dong et al. [13] developed an in-process thermal-stretching technique for effectively mitigating residual stresses and distortion on repair welding of aluminium panels. In addition, Barber et al. [14], van der Aa et al. [15] and Williams and co-workers [16], have applied local cooling, with either solid or liquid CO2 trailing the heat source, as a means of creating dynamically controlled low residual stress and distortion free welds. Several mechanical tensioning systems have also been proposed. Yang et al. [17], [18] have mechanically compressed the weld on cooling using a pair of rollers on both sides of the weld line, reducing both residual stress and buckling distortion. Finally, preliminary work by Williams et al. [8] has shown that the application of global, or far field, mechanical tensioning externally during the welding process can greatly reduce the tensile residual stresses in FSWs. In global mechanical tensioning a load is applied uniformly along opposite ends of the plates prior to clamping the parts for welding (see Fig. 1), so that a uniform tensile stress is maintained in the two butted plates parallel to the weld line. The clamping and tensioning loads are then released after the friction stir welding tool has traversed along the join line forming a weld. Perhaps counter-intuitively, Williams et al. found that high levels of mechanical tensioning parallel to the welding direction can actually reverse the state of stress, so that compressive longitudinal residual stresses are found in the weld region [8].

While a plethora of techniques for influencing the generation of plastic misfit strains and the resultant residual stresses during welding have emerged, at present very little work has been carried out into quantifying the effectiveness of these methods, or establishing their basic principles. This is in part because, to date, little modelling work has been carried out to study the underlying interactions between the externally imposed thermal or mechanical loads and the transient welding stresses. This paper aims to help fill this gap and thereby clarify the mechanism of mechanical tensioning. To this end we have focused on the mechanical tensioning of aluminium alloy friction stir welds. A simplified finite element model of the process has been developed and validated using experimental residual stress data available in the literature measured by X-ray synchrotron diffraction [8], [19], [20], [21], for mechanically tensioned AA7449 and AA2024 FSW's. While modelling work has been carried out on the basic FSW process previously (e.g. [22], [23], [24], [25], [26], [27], [28], [29], [30]), few studies have focused on residual stress prediction [26], [29], [30]. The study by Preston et al. [30] did look at thermal tensioning, but none have looked in detail at the effect of mechanical tensioning on FSWs. Although this paper examines friction stir welding, because the development of residual stresses have been taken to be governed solely by the heat input, the general conclusions regarding the mechanisms of stress relief by mechanical tensioning are applicable to other welding processes.

Section snippets

Global mechanical tensioning experiments

The examination of the mechanical tensioning method is based on friction stir welding trials undertaken by BAE Systems and Airbus UK and associated residual stress measurements described in detail elsewhere [8], [19], [20], [21]. In essence, a hydraulic tensioning rig was used to apply a tensioning load uniformly along the ends of pairs of plates parallel to the weld line during welding. The plates were drilled at each end to allow rigid clamping in the tensioning rig, as shown in Fig. 1. Two

Model validation

The FE model was validated against residual stress data available in the literature obtained by synchrotron diffraction for AA2024-T3 [19], [20], and AA7449-W51 [21] friction stir welds. Examples of comparisons between the modelling predictions and diffraction measurements are shown in Fig. 4, Fig. 5, for welds produced using the conditions given in Table 1. In all cases the diffraction data were corrected for instrument and solute related unstrained lattice parameter (d0) effects [41]. The use

Principles of mechanical tensioning

In the following discussion, the validated model for the 3 mm 2024-T6 plate is used to examine the evolution of temperature and stress during welding as a function of mechanical tensioning level. From the previous section it is clear that tensioning has a significant influence on the final residual stress state after welding. To give an initial overview of the effect of the heat source without tensioning, Fig. 6 shows 2D maps of the thermal field and the longitudinal and transverse stress fields

Conclusions

The principles behind the global mechanical tensioning technique for controlling residual stresses in welding have been investigated using a relatively simple FE model, applied to friction stir welding, based on representing the process purely in terms of a moving heat source. When coupled to a temperature and kinetically dependent material softening model, this approach has been shown to be very successful for obtaining reliable residual stress predictions and for exploring the effectiveness

Acknowledgements

The authors would like to thank A. Wescott, S. Morgan, D. Price of BAE Systems, M. Poad of Airbus UK, and J. Altenkirch, and A. Steuwer for the provision of samples and residual stress diffraction data. This work is supported through the University of Manchester EPSRC Light Alloys Portfolio Partnership (EP/D029201/1) in collaboration with Airbus UK.

References (43)

  • W.M. Thomas et al.

    Mater. Des.

    (1997)
  • M. Peel et al.

    Acta Mater.

    (2003)
  • P. Staron et al.

    Phys. B: Condens. Matter

    (2004)
  • M. Ericsson et al.

    Int. J. Fatigue

    (2003)
  • G. Bussu et al.

    Int. J. Fatigue

    (2003)
  • P.S. Pao et al.

    Scripta Mater.

    (2001)
  • J. Altenkirch et al.

    Mat. Sci. Eng. A

    (2008)
  • M. Song et al.

    Int. J. Mach. Tools Manuf.

    (2003)
  • P. Ulysse

    Int. J. Mach. Tools Manuf.

    (2002)
  • C.M. Chen et al.

    Int. J. Mach. Tools Manuf.

    (2003)
  • R.V. Preston et al.

    Acta Mater.

    (2004)
  • S. Benavides et al.

    Scripta Mater.

    (1999)
  • P.A. Colegrove et al.

    J. Mater. Process. Technol.

    (2005)
  • O.R. Myhr et al.

    Acta Metall.

    (1991)
  • R.S. Mishra et al.

    Mater. Sci. Eng: Res. Rep.

    (2005)
  • K. Masubuchi

    Analysis of Welded Structures: Residual Stresses, Distortion and Their Consequences

    (1980)
  • S.W. Williams et al.
  • T.W. Greene et al.

    Weld. J. Res. Suppl.

    (1946)
  • P.A. McGuire et al.

    Computational Analysis and Experimental Evaluation for Residual Stresses From Induction Heating

    (1979)
  • P. Michaleris et al.

    Weld. J.

    (1997)
  • R.M. Dull et al.
  • Cited by (91)

    • Predicting residual stress in a 316L electron beam weld joint incorporating plastic properties derived from a crystal plasticity finite element model

      2023, International Journal of Pressure Vessels and Piping
      Citation Excerpt :

      It is also a characteristic of friction stir welding in weldable age hardening aluminium alloys such as AA7050 and AA2199 [9]. While precipitation dissolution [10] and local tensile plastic straining [11] in the fusion zone (FZ) are believed to be the causes of forming M-shape residual stress profile in friction stir welding of aluminium alloys, no clear explanation can be found in the literation about formation of M-shape stress profile in austenitic stainless steels. Earlier findings by Horne et al. [12] show considerable deviation between the residual stresses estimated across an EB weld joint using finite element analysis (FEA) and the actual residual stresses which were measured using the contour method and X-ray diffraction measurements.

    • Characterization and modeling of the hardening and softening behaviors for 7XXX aluminum alloy subjected to welding thermal cycle

      2022, Mechanics of Materials
      Citation Excerpt :

      As the temperature increases and reaches the peak temperature, the relative volume fraction decreases rapidly. In the cooling stage, the relative volume fraction gradually decreases until no longer change can be found, which is consistent with the results of Richards et al. (2008) and Sullivan et al. (2006). The simulated yield strength evolution with time and temperature at the peak temperatures of 300 °C and 500 °C are shown in Fig. 9.

    View all citing articles on Scopus
    View full text