School Work

A study of texture patterns in friction stir welds

Description
A study of texture patterns in friction stir welds
Categories
Published
of 16
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  A study of texture patterns in friction stir welds Shaowen Xu, Xiaomin Deng * Department of Mechanical Engineering, University of South Carolina, Columbia, SC 29208, USA Received 6 July 2007; received in revised form 1 October 2007; accepted 18 November 2007Available online 4 January 2008 Abstract Texture patterns on transverse, longitudinal and horizontal cross-sections in friction stir welds (FSW) have been studied experimen-tally, and their variations with welding parameters have been analyzed. Numerical simulations of the FSW process have been carried outto understand the texture patterns. Results of this study suggest that the texture patterns are complex but a dominant theme is theappearance of bands, which occur in the advancing-side material. The banded pattern on the transverse cross-section is often in the formof onion rings. The spacing between the bands on the longitudinal and horizontal cross-sections equals the distance traveled by the weld-ing tool in one revolution. The texture patterns are found to correlate well with equivalent plastic strain contours from simulations of thecorresponding FSW process, suggesting that the texture patterns may be formed because periodically spaced material regions experiencevery different levels of plastic deformation during the FSW process.   2007 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords:  Friction stir welding; Texture; Patterns; Microstructure; Simulation 1. Introduction Since it was invented at TWI in 1991 [1], friction stirwelding (FSW) has received much attention from boththe industry and academia and has gained wide recognitionas a potentially powerful joining technology. FSW has sev-eral advantages over traditional fusion welding techniques,such as: (a) there is no melting in the process and hencethere are no melting-related joint imperfections, (b) the joint has a high mechanical efficiency, and (c) it can beapplied to many ‘‘hard-to-weld ”  materials (e.g. high-strength aluminum alloys). However, as a relatively newtechnology, a predictive scientific knowledge base forFSW has not been well established and many issues remainopen.One of the phenomena that strike researchers today isthe occurrence of banded texture in friction stir welds inthe nugget region. Such texture patterns have beenobserved on section surfaces of friction stir welds since1995 [2]. Threadgill [3] reported that the spacing between adjacent bands in the welding direction is equal to the dis-tance traveled by the forward movement of the rotatingtool in one revolution [3], which was confirmed by otherstudies [4,5].The source of the banded texture is a still a subject of current research. Lienert and Grylls [6] showed that, inwelds of aluminum alloy 6061-T651, the texture patternmay be caused by bands of fine particles of the minorphases. Sutton et al. [4] argued that, in welds of Al 2024-T3 specimens, the banded texture is the result of alternatingregions of high and low particle densities. Yang et al. [7]observed that there are about 10–25% difference in thegrain size between alternating bands in the welds of Al2024-T351 and welds of Al 2524-T351. They also pointedout that there is a strong correlation between grain sizein the nugget and welding parameter in the welds of Al2524 but not in the welds of Al 2024.The influence of the banded texture on materialproperties has been discussed by a number of researchers.Mahoney et al. [8] noticed that, in tensile tests, the fracture 1359-6454/$34.00    2007 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.actamat.2007.11.016 * Corresponding author. Tel.: +1 803 777 7144; fax: +1 803 777 0106. E-mail address:  deng@engr.sc.edu (X. Deng). www.elsevier.com/locate/actamat  Available online at www.sciencedirect.com Acta Materialia 56 (2008) 1326–1341  trajectory tends to develop along the bands. In mixed-mode I/II fracture tests on welds of Al 2024-T3 [4], itwas observed that the macroscopic fracture process isaffected by the banded structure, in that the crack pathtends to follow bands of high particle density. Moreover,micro-hardness tests reveal a periodic hardness variationacross the bands in the welds of Al 2024 and Al 2524 spec-imens [4,7].The effects of variations of welding control variables (i.e.tool rotation speed and welding speed) and tool runoff onFSW processes of Al 2024 and Al 2524 have been studiedrecently by Yan et al. [9]. The experimental results suggestthat the FSW process is a dynamic and periodic process. Inthe experiments, periodic variations of FSW process forces(both the translational resistance force and the axial forg-ing force) have been observed during the welding processes.It is found that the variations of the tool rotation and weld-ing speeds are not related to the variations of the processforces and the banded texture, and that the temporal vari-ation frequency of the process forces agrees with the toolrotational frequency. Although the periodicities of the pro-cess forces can correlate with the nature of the tool runoff,the experimental results show that the banded texture andprocess forces are independent of the magnitude of the toolrunoff.Currently, fully functional theoretical models for pre-dicting the formation of banded texture in friction stirwelds do not seem to exist, although several studies haveoffered important insights in this regard. Biallas et al. [10]suggested that the bands are caused by a mechanism inwhich material particles flowing around the rotating toolpin are ‘‘reflected ”  at ‘‘imaginary ”  walls of the groove,leading to the formation of a ‘‘tube ”  system in the weldnugget. Krishnan [5] proposed that the bands (onion rings)are formed due to the process of friction heating generatedby the rotation of the tool and due to the process of for-ward extrusion caused by the metal on the retreating sideof the tool.As can be seen from the preceding discussions, the cur-rent literature clearly shows that the banded texture in fric-tion stir welds has close ties to the microstructure of theweld region and that the bands have a strong effect onthe mechanical behavior of the welded joints. However, afull understanding of the issues related to the banded tex-ture patterns, such as how the banded texture patternsare affected by welding process parameters and whethersuch patterns can be predicted numerically using theoreti-cal process models, is still not achieved. To this end, theauthors have carried out a combined experimental andcomputational investigation, with a goal of revealing theinfluence of some key process parameters (e.g. tool rotatingand advancing speeds) on banded texture patterns in Al6061-T6 welds and providing insights for understandingsuch patterns. This paper presents the findings of thisinvestigation.In Section 2, FSW experiments [11] on Al 6061-T6 specimens with inserted marker materials are described,and banded texture patterns on three mutually perpendic-ular weld sections are presented for several tool rotatingand advancing speeds. In Section 3, solid mechanicsbased numerical simulations of the FSW process aredescribed. These simulations have been shown to be ableto predict material flow patterns seen in friction stirwelds (see Refs [11–16]). The focus of this section is topresent evidence that there is a strong correlationbetween the banded texture and the distributions of theequivalent plastic strain in the welded joint. In Section4, the main results of the current study are summarizedwith concluding remarks. 2. Experimental results In this study, ten friction stir butt welds (see Table 1)have been made on Al 6061-T6 plate specimens using anMTS FSW Development System with the force controlfunction. For each weld, a thin pure aluminum foil (witha thickness of 0.1 mm) was used as a marker materialand was inserted at the faying surface between the twoplates (with a thickness of 8.13 mm) to be joined. The dis-tribution of the marker material after each weld enables theinvestigation of material flow patterns and the formation of banded textures in a weld.For these butt welds, the rake angle of the rotating weld-ing tool was chosen to be 2.5   degrees. All welds were madewith the same welding tool, which has a 25.4 mm shoulderdiameter and a 10.0 mm pin diameter (the pin thread spac-ing is 1.0 mm). The normal of the shoulder surface forms a7   angle with the tool axis. Other welding parameters foreach of the ten welds are given in Table 1.After the welds were made, the specimens were cut fortexture pattern analysis. Three types of cross-sections werecut to provide a three dimensional understanding of thematerial flow and texture patterns in the weld region.Fig. 1 shows a graphical definition of the three types of cross-sections, namely the transverse cross-section, the lon-gitudinal cross-section and the horizontal cross-section. Let X   be the welding direction,  Z   the plate normal direction( Z   = 0 is at the bottom surface of the plate), and  Y   thetransverse direction (which forms a right-hand coordinate Table 1Welding parametersWeldsno.Tool rotatingspeed (rpm)Tool weldingspeed (mm s  1 )Force in  Z  direction (kN)mm perrevolution1 240 1.279 22.240 0.3202 240 2.363 26.688 0.5913 240 3.316 28.912 0.8294 290 2.363 24.909 0.4895 340 2.363 24.909 0.4176 390 1.279 20.906 0.1977 390 2.363 23.574 0.3648 390 3.316 27.578 0.5109 720 2.363 22.240 0.19710 800 2.363 21.350 0.177 S. Xu, X. Deng/Acta Materialia 56 (2008) 1326–1341  1327  system with the  X   and  Z   directions. Note that the  X   –  Z  plane coincides with the faying surface before welding.Relative to the  XYZ   coordinate system, the transversecross-section is cut across the weld at a location alongthe weld where a steady state weld condition exists. Thiscross-section is parallel to the  Y   –  Z   plane, and it is viewedalong the negative  X   direction (see the arrow next to thesection in Fig. 1). The longitudinal cross-section is exactlyat the  X   –  Z   plane and it is viewed along the positive  Y   direc-tion (see the arrow next to the section in Fig. 1). The hor-izontal cross-section is parallel to the  X   –  Y   plane and is cutat a certain distance (the height  H  ) to the bottom surface of the plate. This cross-section is viewed along the negative Zdirection (i.e. from top down; also see the arrow next to thesection in Fig. 1).For each of the cross-section cut, the texture patternanalysis was performed in two steps. In the first step, thesurface was etched using a modified Keller’s etch, whichincluded 2 mL HF (48%), 3 mL HCl (conc), 20 mLHNO 3  and 175 mL H 2 O. This etching process will resultin a high-quality image of the texture pattern, includingthe location of the marker material. In the second step,an image of the etched section surface was taken with aMicrotek ScanMake 4900 scanner at a resolution of 4800    2400 dpi.In the following, experimental findings based on a seriesof cross-sectional images of the ten welds will be presented.Observations of texture patterns will be made and thedependence of the patterns on certain key process parame-ters will be described. To interpret properly material flowpatterns in a weld, it is necessary to distinguish betweenthe materials that srcinally belong to either of the twoplates of the weld. For convenience of discussion, the mate-rial that srcinally belongs to the plate on the advancingside of the faying surface is referred to as the advancingside material, and the material that srcinally belongs tothe plate on the retreating side of the faying surface isreferred to as the retreating side material. It is noted that,before welding, the interface between the two materials(which is the faying surface) is flat and perpendicular tothe top and bottom plate surfaces, but after welding, theinterface is no longer flat – it becomes severely distorted,as described in subsequent sections.  2.1. Texture patterns on the transverse cross-section Figs. 2–4 present images of texture patterns on the trans-verse cross-section of various welds. In particular, Figs. 2and 3 show the effect of the tool welding speed on the tex-ture pattern when the tool rotating speed is, respectively,240 rpm (revolution per minute) and 390 rpm. In each case,three images are shown, which corresponds to a weldingspeed of 1.279 mm s  1 , 2.363 mm s  1 , and 3.316 mm s  1 ,respectively. On the other hand, Fig. 4 illustrates the influ-ence of the tool rotating speed on the texture pattern whenthe welding speed is 2.363 mm s  1 . In this case, six imagesare shown for tool rotating speeds of 240, 290, 340, 390,720, and 800 rpm.In all images, the advancing side is on the left and theretreating side is on the right. A dark line (e.g. see Figs. 2and 5a) can be seen in the center region of the weld, wind-ing down from the top plate surface, through the weldregion, to the bottom plate surface. This dark line is theinterface between the advancing side material and retreat-ing side material and is made visible by the marker mate-rial. The interface line is not always obvious and oftenappears to be disconnected (e.g. in Figs. 3 and 4; see alsoFig. 5b), reflecting mainly the effect of the tool rotatingspeed, as discussed below individually for each of the threefigures.In Fig. 2, the interface between the advancing andretreating side materials starts near the right side of the Fig. 1. A schematic of a friction stir weld, showing the orientation of the transverse cross-section, the horizontal cross-section, and the longitudinal cross-section relative to the tool welding and rotation directions.1328  S. Xu, X. Deng/Acta Materialia 56 (2008) 1326–1341  weld nugget on the top surface of the plate (see point A inFig. 5a) and runs down to somewhere in the middle of theweld nugget at the bottom plate surface (see point F inFig. 5a). It is a continuous and curved line, which intersectsthe weld centerline (the vertical dashed line in the centerof the welds) several times. Near the top plate surfaceand in the lower middle region of the nugget, the advancingside material passes the weld centerline and pushes the Fig. 2. Texture patterns on the transverse cross-section of three welds made at a tool rotation speed of 240 RPM and a welding speed of: (a) 1.279 mm s  1 ,(b) 2.363 mm s  1 , and (c) 3.316 mm s  1 . The dashed vertical line in the center of each picture represents the weld centerline before welding.Fig. 3. Texture patterns on the transverse cross-section of three welds made at a tool rotation speed of 390 rpm and a welding speed of: (a) 1.279 mm s  1 ,(b) 2.363 mm s  1 , and (c) 3.316 mm s  1 . The dashed vertical line in the center of each picture represents the weld centerline before welding. S. Xu, X. Deng/Acta Materialia 56 (2008) 1326–1341  1329  retreating side material to the right. In contrast, in theupper middle region of the nugget, the retreating side mate-rial passes the weld centerline and pushes the advancingside material to the left. Nevertheless, the advancing sidematerial is not mixed with the retreating side materialalong the interface. Intuitively, however, the longer inter-face and the complexity of the interface seems to suggesta better bond between the two sides than a straight verticalinterface can. A banded texture clearly has developed in theadvancing side material (see the dark region in the lowerleft part near the interface), which becomes more obviousas the welding speed is increased (see Fig. 2c). The size of the dark region with the banded texture is seen to decreaseas the welding speed is increased. Another banded texture Fig. 4. Texture patterns on the transverse cross-section of six welds made at a welding speed of 2.363 mm s  1 and a tool rotation speed of: (a) 240 rpm, (b)290 rpm, (c) 340 rpm, (d) 390 rpm, (e) 720 rpm, and (f) 800 rpm. The dashed vertical line in the center of each picture represents the weld centerline beforewelding.1330  S. Xu, X. Deng/Acta Materialia 56 (2008) 1326–1341
Search
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks