Journal of Materials Processing Technology 234 (2016) 72 83 Contents lists available at ScienceDirect Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec Friction crush welding of aluminium, copper and steel sheetmetals with flanged edges Florian A. Besler a,b,, Paul Schindele b, Richard J. Grant a,c, Michael J.R. Stegmüller b a Glyndwr University, Department of Engineering and Applied Physics, UK b Kempten University of Applied Sciences, Department of Engineering, Kempten, Germany c University College Bergen HiB, Department of Mechanical and Marine Engineering, Bergen, Norway a r t i c l e i n f o Article history: Received 25 November 2015 Received in revised form 26 February 2016 Accepted 14 March 2016 Available online 15 March 2016 Keywords: Welding Friction welding Friction crush welding Friction stir welding Solid state welding Sheetmetal Dissimilar material Aluminium copper welding a b s t r a c t The implementation of friction crush welding (FCW) offers a versatile application in the welding of sheet metals. Three materials (EN AW 5754H22, DC01 and Cu-DHP) were analysed by applying the method with flanged edges. The additional material required to form the weld is provided by the flanged edges of the parent sheet metal. The joint is formed by the relative motion between a rotating disc, which is applied with a crushing force, and two sheet metal parts. The fundamental process variables and the requirements of the welding preparation are shown. Bond strengths, as a percentage of the yield strength of the parent material, of around 95% (DC01) 90% (EN AW 5754H22) and 62% (Cu-DHP) are achieved. Microstructural investigations reveal that a dynamic solid-state deformation and recrystallization of the additional flanged material results in a fine grain microstructure in the weld region. Reduced metallurgical changes along with minimized distortion and residual stresses in the parent material indicate low heat input. By creating a fine grain microstructure in the welding line, the friction crush welding method reveals great potential, especially for welding steel. 2016 Elsevier B.V. All rights reserved. 1. Introduction 1.1. Friction welding: classification of methods Friction welding is generally categorised as pressure welding (DIN EN 14610) subordinated and then described according to the relative movement between the interfaces and according to their types of energy supply (EN ISO 15620). Considering the relative velocity and the friction motion between the interfaces, friction welding processes can be classified in two groups. The first group represents the various means where the heat is generated by the relative movement between the surfaces. The classical processes with rotary, linear and orbital motion are representatives of this group. Maalekian (Maalekian et al., 2008) compares the relative velocity between these three welding methods and the influence of the heat generation at their interfaces. This article describes disadvantages of such methods, for instance the non-uniform relative motion over the interfaces experienced in rotary friction welding. The directional and unidirectional relative velocities of the differ- Corresponding author. E-mail address: florian.besler@hs-kempten.de (F.A. Besler). ent friction welding techniques are illustrated in Fig. 1 (unlabelled arrows). Processes generating the friction due to a relative motion between a non-consumable tool and the workpiece surface can be classified as the second group. Friction Stir Welding FSW was developed by W.M. Thomas (Thomas et al., 1991) as a solid-state joining process (Fig. 1d) which operates by way of a non-consumable rotating tool with a specially designed pin and shoulder. When the pin is inserted into the interfaces of metal plates that to be joined and traversed along their mating edges the movement of the tool creates a material flow which produces the joint (Mishra and Ma, 2005). The effect of the tool wear offers a challenge in FSW and may be correlated with the rotational speed of the pin and the welding speed. Fernandez and Murr (Fernandez and Murr, 2004), detected a reduction in tool wear at decrease in rotational speed and increasing welding speed. The influence of a variation in the relative motion between the tool and the workpiece surface in FSW was also investigated by Sato and Kokawa (2001). The article describes the differences between the intrinsic strength and ductility shown by the advancing and the retreating sides; with the retreating side showing a lower strength. Cavaliere (Cavaliere et al., 2008) describes and compare the grain size as a function of the welding speed in friction stir welded aluminium (AA6083). Significant grain fining was found at welding speeds between 100 mm/min http://dx.doi.org/10.1016/j.jmatprotec.2016.03.012 0924-0136/ 2016 Elsevier B.V. All rights reserved.
F.A. Besler et al. / Journal of Materials Processing Technology 234 (2016) 72 83 73 Fig. 1. Classification of friction welding. and 200 mm/min. Dissimilar FSW of thin sheets (0.8 mm) of aluminium (AA5052) and pure copper (99,999%) is shown by Yusof et al., 2013. The author shows the possibility to weld different materials and describes problems like surface void defects and insufficient material flow relating to the thin metal welding. Developing a joining approach, which eliminates unequal relative stirring velocities and offers the possibility of welding thin sheet metals at high speeds, is the motivation for this work. The use of friction crush welding, FCW, offers such possibilities and is the subject of this article. 1.2. Principal of friction crush welding (FCW) FCW is based on an invention (patent pending) for joining metal workpieces using frictional heat and pressure with the possibility to produce joints between similar and dissimilar materials (Schindele, 2010). The process operates, similarly to FSW applications, with a friction contact between a non-consumable rotating tool and the workpiece surfaces. The relative motion, in contrast to FSW (Fig. 1d), is accomplished by means of a rotating disc (Fig. 1e) and with a unidirectional uniform relative speed between the disc and sheet metal edges (Besler, 2014). In this implementation of the technique the edges of the sheet metal parts to be joined are prepared with flanged edges and are then placed against each other (adjacent, in the same plain and contacting). The friction disc traverses with a constant feed rate along the edges of the workpiece which leads to a plastification of the material in the region near the FCW disc. The disc surface, with a specific circumferential profile, shapes the welding line whereby the joint is modelled by the action of crushing a certain amount of additional flanged material into the gap formed by the contacting material (see crushing zone, Fig. 2). Because of these two main processes, the friction based heating and secondly the crushing mechanism of material intermixture, the name of the process was previously defined as Friction Squeeze Welding (Schindele, 2012); however, in this work it is described as Friction Crush Welding (University Kempten, 2016). 2. Preliminary considerations 2.1. Process parameters Fig. 2. Principle and microstructural zones of Friction Crush Welding. FCW involves a complex plastic deformation and material modelling process to form the joint. The preparation of the mating sheet metal parts and the geometry of the rotating disc offer a number of parametric variables which may affect the welding process. Also the effect of the disc rotational speed, its position relative to the work piece and the speed at which it is traversed along the weld line (welding speed) need to be considered. The welding speed and the disc rotational speed could be expected to cause a change in the resistance when crushing the additional material, leading to an associated pressure distribution at the contact zone (Fig. 3). This pressure distribution will result in a crushing force F C that can be resolved into a horizontal component F H which opposes the welding direction of travel and a vertical component, or vertically downward force, F v. The geometry of the disc may also be important in this regard.
74 F.A. Besler et al. / Journal of Materials Processing Technology 234 (2016) 72 83 2.2. Disc geometry Fig. 3. Pressure distribution and forces in FCW. The friction disc geometry potentially offers an important influence on the material flow and in the forming of the welding line. The basic dimensions of the disc can be specified by three parameters (see Fig. 4): the outside disc diameter (D), the disc width (W d ) and the superficial curvature of the disc (d c ). Derived from the alignment and the basic dimensions of the two flanged sheet metal parts, a geometric consideration was conceived to determine the shape of the friction disc. The friction disc was, as an initial concept, designed with a simple semi-circular groove at the outside circumferential surface and which ran around at its mid-section. The intention of this concave profile was to produce a confinement of the material, concentrated towards the weld centre. It was anticipated that this effect would lead to a better horizontal intermixture of the material in the crushing zone so providing a better bonding strength of the joint. Another desired effect was to create a convex welding line with a reinforcing effect on the joint geometry. In the profile of the disc the groove diameter (d c ) (Fig. 4) can be described in conjunction with the material thickness (t) and gap length (G) as: dc = 2 (4t+G) The curvature of the concave disc profile in the contacting edge of the disc could also be varied and it may be that other groove geometries prove beneficial; however, the chosen geometry Fig. 5. Additional material volume. provides a good starting point in this work and is simple to manufacture. This shape variation will be the subject of further research as a separate parametric study. The size of the disc shoulder (S) was judged to be an important factor for the forming process of the welding line, allowing the plasticised material to be sufficiently forged beneath the disc groove. Operating as a seal between the disc and the sheet metal pieces, the disc shoulder (see Fig. 4) was designed to prevent the plasticised material from crushing-out of the welding zone. 2.3. Joint geometry The requirements of a specific additional material volume are defined by the flange height (F) of the sheet metal parts. The area between the workpieces comprising the gap (G) and the crushing zone must be filled with the additional material during the process. In order to determine this extra volume of material, a basic geometric calculation is required, see Fig. 5. The thermo-mechanical forming process of the material in FCW can be compared with a rolling process. In a similar manner, the volume of additional material (V 3 ) provided by the flanges must equate to the volume required to form the weld. The clearance beyond the disc contour and the crushing zone can be separated into the two demarcated volumes V 1 and V 2, and so it follows that V 3 = V 1 +V 2 (2-2) where the boundary of the forming process is represented by the concave disc contour. Specifying the outside bend radius of the sheet metal (R), volume V 2 can be calculated as V 2 = 1 2 [(4 t) 2 -(2 t) 2 ] (2-3) where R = 2t. Also, volume V 3 is calculated as V 3 = 2t G (2-4) This volume of the additional material (V 3 ), provided by the excess length (A) of the flanged material, can thereby be fully described by the dimensions t and G A 2t = G 2t+ 1 2 [(4 t) 2 -(2 t) 2 ] giving A = G+4 t t 3. Experimental setup 3.1. Material selection and joint preparation Fig. 4. Welding disc geometry and material concentration in the crushing zone. For the trials using aluminium, the non-hardenable wrought alloy of the 5000 series, namely EN AW-5754H22 (AlMg3), was
F.A. Besler et al. / Journal of Materials Processing Technology 234 (2016) 72 83 75 Table 1 Material choice and joint preparation for the FCW trials. Aluminium alloy Steel Copper Material designation AlMg3 DC01 Cu-DHP Material number EN AW-5754 1.0330 CW024A Specifications H22 Cold rolled R240 half hard Mechanical properties of the sheet metals Ultimate strength R m 220 270 MPa 270 410 MPa 240 300 MPa Yield strength R p0,2 130 MPa 140 280 MPa 180 MPa Young s modulus 70,500 MPa 208,000 MPa 110,000 MPa Hardness 63 HBW 105 HV 65 95 HWB Physical properties Melting Temperature 610 640 C 1250 1460 C 1083 C Thermal conductivity 140 160 W/(m K) 54 W/(m K) 305 W/(m K) Sheetmetal and joint preparation Sheetmetal length 250 mm 250 mm 250 mm Sheetmetal thickness 1 mm 1 mm 1 mm Flange height 2.5 mm 2.5 mm 2.5 mm selected due to its excellent welding behaviour for fusion, resistance and friction welding. For steel sheetmetals, a cold rolled carbon steel DC01 (1.0330) was used. This steel combines the advantages of good weldability with sufficient elongation for the FCW process. In order to prevent oxidation problems at annealing temperatures during the process, a shielding gas (argon) was injected into the friction contact zone. Copper Cu-DHP was judged to be a promising choice for FCW. In copper welding, the high thermal conductivity and high affinity for oxygen exacts particular demands on welding applications. The material properties and the joint preparation are presented in Table 1. The outcome of the work piece flange calculations (see Section 2) showed that sheet material of t = 1 mm would provide a good geometry for the trials. The preparation involved flanging the mating edges to a height F = 2.5 mm, where the bend radius is R = 2 mm and the bend angle ( ) is 90. 3.2. Welding machine concept The experimental setup for the FCW method used in this work is based on a concept of a portal welding machine with a moving Table (Fig. 6). Fig. 6. Portal welding machine for the FCW application. The machine was designed with a portal frame (1) as a rigid steel construction with an effective vibration damping system (2) provided by rubber soles at the machine base. At the centre of the machine is a ridged clamping Table with an overall length of 1500 mm, which incorporates a sheet metal clamping system (3). The clamping Table is mounted to the machine bed by way of two linear guides (4). This affords a rigid attachment so as to minimise any deformations caused by the vertical friction forces (see Fig. 3). The linear guides run the length of Table (see Fig. 6) and are driven by servo-motors along the y-axis. The y-axis motion of the machine bed provides the welding speed (v w ) with a traverse feed rate of up to 10,000 mm/s. The horizontal motion along the x-axis is guided by means of a track, which is driven and is attached to a cross-head (5). This construct gives support for the assembly providing the vertical movement along the z-axis by way of two further linear guides (6) driven by a pneumatic cylinder (7). The rotation of the friction disc is produced by a spindle system (8) which consists of an electric spindle motor and the disc clamping mandrel. This spindle system is attached to the z-axis traverse. The rotational speed (n) of the spindle motor can be controlled using a frequency converter (9) which produces a rotational speed range between 12,000 and 18,000 revolutions per minute (rpm); although in this work a speed of 18,000 rpm was used for all tests. The pneumatic cylinder applies the vertical force (F V ). The total vertical motion of the spindle system can be controlled by an adjustable limit which maintains a given clearance between the friction disc and the sheet metal. Based on several initial tests it was observed that the position of the disc relative to the work piece surface might prove to be an important process setting and will be examined in a future publication. The horizontal position is defined as the clearance (C) which is the distance between the disc shoulder and the surface of the sheet metal parts. For all investigations, the overall disc diameter was held constant at 100 mm, whereas the disc width is correlated to the geometry of the sheet metal being joined and the size of the disc shoulder (S). 3.3. Welding evaluation and specimen preparation 3.3.1. Bond strength evaluation In this study the correlation between the welding speed, i.e. the speed that the disc is traversed along the weld line and the resulting tensile strength in the zone bonding the sheet metal parts is investigated. Each welded sheet was separated into four specimens as shown in Fig. 7; these were in accordance with DIN EN ISO 4136 standard. Reference samples were prepared of the parent sheet metal, i.e. without a weld, to evaluate comparative material properties. Care was taken to ensure that the rolling directions were consistent. 3.3.2. Microstructural analyses Tests on the welded sheets were conducted with microscopic examinations according to the European Standard EN ISO 17639. The tests on the welded sheet metal were examined using an optical microscope (M = 10x and M = 50x) to evaluate the superficial quality of the tested samples. In order to determine the surface quality the welding line was measured longitudinally according to DIN EN ISO 4288. The output of the surface profile was analysed to determine the average roughness and the absolute profile height. Microscopic examinations were made to analyse the microstructure of the welded specimens, this time in a transverse direction to the welding line. The analyses were conducted after transversely cross sectioning and etching using a magnifi-
76 F.A. Besler et al. / Journal of Materials Processing Technology 234 (2016) 72 83 Fig. 7. Sheet metal for tensile test evaluation. Fig. 8. Bond strength of welded Aluminium AW-5754 in correlation to the welding speed. Fig. 9. Bond strength of welded Steel (DC01) in correlation to the welding speed. cation factor between M = 25x and M = 500x. The procedure for analysing the average grain size was carried out according to the ASTM E112 13 method. These tests methods employ a grain size scale to enable an average grain size to be calculated; such methods are applicable when a single phase is mostly or entirely present. 4. Results and discussion 4.1. Bond strength analyses 4.1.1. Aluminium alloy In FCW of the EN AW-5754 aluminium alloy tests were conducted at welding speeds ranging from 500 mm/min to
F.A. Besler et al. / Journal of Materials Processing Technology 234 (2016) 72 83 77 Fig. 10. Bond strength of welded Cu-DHP in correlation to the welding speed. Fig. 11. Surface R a and R t qualities of FCW sheetmetal (a) aluminium EN AW-5754H22, (b) cold rolled carbon steel (DC01) and (c) copper Cu-DHP. 4000 mm/min. The results are shown in Fig. 8, where additionally the tensile strength of the parent material is indicated which was determined to be 243.5 ± 4 N/mm 2. Overall it is clear that the bond strength decreases as the weld speed increases with the best tensile strength obtained at 500 mm/min. These values reach an upper strength limit of approximately 90% of the parent material. 4.1.2. Steel The results of the FCW tensile tests from the DC01 steel specimens are compared with the average tensile strength of the parent steel; in this case 337.6 ± 3 N/mm 2 (Fig. 9). The tensile strength curves show the best results are achieved at welding speeds between 1500 mm/min and 2000 mm/min. Some of the specimens welded at 1500 mm/min and 2000 mm/min produced strength values approaching that of, and even the same as, the parent material.
78 F.A. Besler et al. / Journal of Materials Processing Technology 234 (2016) 72 83 Fig. 12. Surface roughness in correlation to the welding speed. However, beneath the lower threshold at 1000 mm/min the results show an average tensile strength of approximately 50% of the parent material. 4.1.3. Copper Fig. 10 shows the tensile strength of FCW for copper versus welding speed. The results are compared with the parent material with a tensile strength of 263.9 ± 2 N/mm 2. Considering the trial curves, the tensile strength is seen to approximately double for an increase in welding speed from 1000 to 5000 mm/min. The highest strength is delivered with a welding speed between 5000 and 6000 mm/min which is an average of 62% that of the parent material. There is a greater consistency in the tensile strength achieved for a welding speed of 5000 mm/min with little scatter in the results obtained for each trial. 4.2. Macrostructural analyses A visual indication of the surface quality of welds and the values of the average roughness R a and the absolute profile height R t when joining aluminium, copper and steel sheet are shown in Fig. 11. In order to produce the same welding conditions for each trial, the surface curvature of the disc was mechanically reworked after each trial. Curves of surface roughness, R a, versus the welding speed are shown in Fig. 12. The change of welding speed shows the most significant influence on the surface quality of the aluminium alloy, with the copper and steel results showing little sensitivity in this regard. The higher welding speeds for all three materials show similar surface qualities, with the copper presenting the best results. A correlation with the poorer surface qualities obtained by the aluminium welding line was seen to coincide with a deterioration in the quality of the friction disc surface. The result of making a 250 mm long weld (one trial) is shown in Fig. 13 where deposits of the sheet metal are evident as adhering layers around the disc circumference. In addition to the upper surface of the welding line, the weld quality of the rootside is a factor which can influence the performance of a joint. Images of this side of the weldline for all three material types can be seen in Fig. 14. The quality of the rootside joint profile of the welded sheet metals were assessed by an optical inspection employing a laser microscope that provides a magnification factor in a range between M = 5x and M = 500x. The evaluation of three characteristic rootside profiles can be made by providing three quality categories which can be seen in Fig. 15. The top images show surface profiles (magnification factor M = 20x), along with their attendant rootside view. The category representing the best welding result, a complete bonding, is shown as a closed surface in Fig. 15a). The laser microscope analysis of this area only reveals some small groves of less than a 0.5 m. This can be represented in the sketch shown in Fig. 16a). The rootside surface is completely closed which recognises a sufficient material flow during welding. Sheetmetal welded with this rootside quality also produces the highest bond strength. Rootside defects, such as a layer building phenomenon (Fig. 15b), represents an insufficiently closed joint. A material flow during welding results in a layer formation at the rootside. This less desirable weld showing the layer building is represented in Fig. 16b). The third category of rootside defects, could be evidenced by the production of root groves and insufficient bonding (Fig. 16c). Root groves of approximately 100 m to 400 m and insufficient bonding offer a predetermined braking point for the strength test evaluation (Fig. 15c). Such a defect resulted in a weld producing a very poor quality. 4.3. Microstructural analyses Based on microstructural analyses of transverse cross-sections, three characteristic zones of grains and precipitates can be classified (Fig. 17). Firstly, the crushing zone which is arranged in the middle of the joint. This zone, representing the gap between the sheetmetal parts, has to be filled with additional material during the crushing process. Due to the intensive material deformation and the presence of intermixture, this area is mainly affected by precipitates or oxidation inclusions. Secondly, the thermodynamically-affected zone (TMAZ). This zone between the parent material and the crushing zone is characterized by thermomechanical forming which results in a recrystallized fine-grained microstructure. Thirdly, the heat-affected zone (HAZ), which is located between the TMAZ and the parent material. This zone is exposed to a thermal treatment, but does not experience any plastic deformation which leads to different microstructural properties in comparison with the parent material. 4.3.1. Aluminium alloy The results of the microstructural analyses of aluminium alloy specimens welded with different welding speed v w (e.g. v w = 500 mm/min and 1500 mm/min) are shown in Fig. 18. One significant finding for all welding speeds is a consistently occurring layer formation at the interface of the two sheetmetal parts, whereby a thin layer of approximately 2-4 m thickness runs in a through-thickness direction in the line of the weld. 4.3.2. Steel Cross-sections were prepared in order to reveal the relationship between the welding line bond strength, the welding speed (v w ) and the resultant microstructure in the weld interface (Fig. 19). Investigations with lower welding speeds showed oxidation layers and a lack of bonding in the crushing zone (Fig. 19a); whereas higher welding speeds led to insufficient material being crushed and rootside defects such as root groves (Fig. 19d). The best results, with a homogenous grain microstructure, are shown at a welding speed of v w = 1500 mm/min (Fig. 19c). Another significant effect when considering the steel crosssections is the phenomenon of grain refining of the microstructure in the crushing zone. Determining the influence of the thermomechanical treatment during the welding process was the reason for further investigations of the resulting grain size. Consequently, the grain size of the crushing zone was measured by using a grain scale and then compared with the grain size of the parent material (Fig. 20). The grain size measurements of the parent material (Fig. 20 region A) shows an average grain size of 6400 m 2, whereas the grain of in the crushing zone results in a finer grain size of approximately 800 m 2 (see region B). The result of this study indicates
F.A. Besler et al. / Journal of Materials Processing Technology 234 (2016) 72 83 79 Fig. 13. Disc weld material deposits in relation to the welding speed. Fig. 14. Rootside qualities of FCW sheetmetals (a) aluminium EN AW-5754H22, (b) cold rolled carbon steel (DC01) and (c) copper Cu-DHP. a grain fining factor of 1:8 between the parent material and the microstructure in the welded area. 4.3.3. Copper The results from the copper cross-sections welded at various welding speeds are shown in Fig. 21. The images show the interface between the two sheetmetal parts (M = 100x). The cross-section reveals the formation of a crushing zone with several oxidation inclusions. Such oxidation layers will act to decrease the intermetallic strength and can therefore be seen as a probable reason for the reduced tensile strength in relation with the parent material.
80 F.A. Besler et al. / Journal of Materials Processing Technology 234 (2016) 72 83 Fig. 15. Rootside quality in FCW analysed with a laser microscope (a) complete bonding, (b) layer building and (c) root groves and insufficient bond. Fig. 16. Classification of the rootside profiles: quality categories. Fig. 17. Characteristic microstructural zones of FCW steel (DC01). Another interesting finding can be made when considering the cross-sections is the relationship between the size of the oxidation layers and the welding speed (v w ). A lower welding speed leads to a stronger oxidation, whereas the specimen welded at a higher speed shows less oxidation inclusions. This correlation is a probable reason for the increase in the tensile strength of copper specimens with an increase in the welding speed (Fig. 10). 4.4. Overall discussion The greatest influence on the bond strength with respect to the welding speed is shown by the aluminium alloy specimens (Fig. 8). The intensive heating and forming process at lower welding speeds (e.g. 500 1000 mm/min) seems to lead to a better material flow, so resulting in higher bond strengths. The thin oxide layer that crosses the welding zone, extending from the bottom to the top surfaces and running in the line of the weld, is probably a result of the superficial oxidation of the aluminium sheet metal which would indicate that oxides are not totally removed during the crushing process. The removal of oxides from the surfaces to be crushed by way of a grinding process, employed immediately before the friction disc, might resolve this problem and will be the subject of future tests. Steel welding offers the greatest potential (bond strength of 95% compared to the parent material) in the FCW process (Fig. 9). For weld speeds of less than 1000 mm/min the results show an average tensile strength of approximately 50% of the parent material. This phenomenon can be explained as the effect of an insufficient material intermixture combined with oxidation inclusions. Microstructural analyses evidence the presence of such inclusions at lower welding speeds, possibly caused by a higher fiction heat
F.A. Besler et al. / Journal of Materials Processing Technology 234 (2016) 72 83 81 Fig. 18. Cross-sections of aluminium alloy (AW-5754), welded at different welding speeds, v w. Fig. 19. Cross-sections of steel (DC01), welded with different welding speeds. input and insufficient bonding at the rootside of the welding line. An optimal welding speed range is demonstrated, as opposed to a trend. A grain refinement is shown in all welded steel specimens; noting that a change in heat input and increasing crushing forces at higher welding speeds might also have a positive influence on the grain refining effect and should be investigated further. Considering Fig. 10, the tensile strength for the copper copper joint shows an approximately linear increase in specimen bond strength for an increase in weld speeds. This linear relationship can probably be attributed to the level of thermal heat input associated with a given welding speed. It would seem logical to assume that greater welding speeds lead to less heat input in the weld, although further analyses will be required to quantify temperatures and the heat distribution due to the thermal conductivity of the materials. The phenomenon of increasing bond strength with an increase in weld speed (the converse of that demonstrated by the aluminium welds) is probably based on the decrease in oxidation inclusions and oxide layers present at the copper interfaces (see Fig. 21). Trials with steel and copper sheet metals show a marked increase in bond strength with higher welding speeds; however, the opposite can be seen with the aluminium alloy. The reason for this discrepancy is not entirely clear from the available evidence; but it is possible that the plastic deformation of the sheet metal caused by the FCW disc results in a crushing force that is predominantly influenced by the welding speed, i.e. the traverse speed. This force may need to exceed a certain value, which is material dependant, to plasticise the material and create a satisfactory bond. With aluminium, it would seem that the force which is sufficient for plasticisation may have already been reached at the lowest weld-
82 F.A. Besler et al. / Journal of Materials Processing Technology 234 (2016) 72 83 Fig. 20. Grain fining in the crushing zone of steel (DC01). Fig. 21. Cross-sections of copper (Cu-DHP) specimens welded with various welding speeds. ing speeds tested. With increasing welding speeds the aluminium is plasticised to such an extent that the resulting crushing forces decrease to a level that are insufficient to provide the conditions necessary for a crush weld. Further work has to be conducted to measure the level of temperature input and the crushing forces during the process. Also tests should be made to demonstrate the effects of a change in disc rotational speed. The surface quality and at least the superficial forming of the weld are seen to be strongly dependent on the sheet metal material. Another important parameter will probably be the friction behaviour between the workpiece material and the disc; thus inferring that the disc material is also important, which will be verified in subsequent tests. An influencing factor for both the sheet metal and disc material is the ability for the disc to model or shape the weld, with the surface quality directly related to the disc defects on the contacting surface of the friction disc. Growing disc defects are directly dependent on declining surface welding quality. Analyses of the surface and transverse cross sections through the welding line reveal root grooves; also thin cracks along the crushing zone are evident. Rootside defects show the signs of being caused by inadequate material flow due to insufficient material being crushed during the welding process. One possible reason for this effect could be a lack of additional material volume due to insufficiently high flanges on the sheet metal parts, although the specimens for the tests were appropriately prepared in accordance with the calculations shown in Section 2.3. Another reason could be as a result of inconstant process crushing forces (Fig. 3), caused by contact pressure variations between the friction disc and the sheetmetal parts. It would seem logical that the cracks likely be the cause of
F.A. Besler et al. / Journal of Materials Processing Technology 234 (2016) 72 83 83 the reduced bond strength, and it should be additionally noted that this could significantly affect the fatigue behaviour of the welds. 5. Conclusions The work presented in this article describes a method to weld aluminium alloy (EN AW-5754), steel (DC01) and copper (Cu-DHP) by using a new process; namely, friction crush welding. The main results, analysing the important process parameters of welding speed and bond strength, show good promise along with the level of surface finish with sheetmetals of 1 mm thickness. Following conclusions can be drawn from preliminary trials: The welding speed shows the greatest influence on the bond quality produced by FCW process; however, the results were material dependent. For aluminium low welding speeds produced high bond strength values: for copper and steel high welding speeds were necessary. Steel demonstrated the greatest strength with an average bond strength approximately 95% of the parent material when processed at weld speeds from 1500 to 2000 mm/min. The aluminium alloy and copper joint reached bond strengths of 90 and 64% of the parent material, respectively. High quality welds surface finishes can be produced by this technique. Steel welding offers a grain fining effect in the welding zone. This effect was judged to increase the mechanical properties in the bonding zone. Oxidation inclusions and insufficient bonding could be investigated as main problems within samples providing low bond strength. The thermal behaviour during the process and the resulting crushing force seem to be an important factor when considering improvements in material flow and bond strength. Future work should consider the application of temperature measurement devices and measurement of the crushing forces. Acknowledgments (Laboratory for Material Sciences) at Kempten University of Applied Sciences for their support. References Besler, F.A., 2014. DVS Congress: Große Schweißtechnische Tagung; DVS-Studentenkongress; Vorträge der Veranstaltungen in Berlin am 15. und 16. September 2014, Berlin,S. 103 105. Cavaliere, P., Squillace, A., Panella, F., 2008. Effect of welding parameters on mechanical and microstructural properties of AA6082 joints produced by friction stir welding. J. Mater. Process. Technol. 200 (1-3), 364 372, http://dx. doi.org/10.1016/j.jmatprotec.2007.09.050. DIN EN. 14610. Welding and allied processes Definitions of metal welding processes. EN ISO. 15620. Friction welding of metallic materials (EN ISO 15620:2000). EN ISO. 17639. Destructive tests on welds in metallic materials Macroscopic and microscopic examination of welds (ISO 17639:2003); German version (EN ISO 17639:2013). Fernandez, G., Murr, L., 2004. Characterization of tool wear and weld optimization in the friction-stir welding of cast aluminum 359+20% SiC metal-matrix composite. Mater. Charact. 52 (1), 65 75, http://dx.doi.org/10.1016/j.matchar. 2004.03.004. Maalekian, M., Kozeschnik, E., Brantner, H.P., Cerjak, H., 2008. Comparative analysis of heat generation in friction welding of steel bars. Acta Mater. 56 (12), 2843 2855, http://dx.doi.org/10.1016/j.actamat.2008.02.016. Mishra, R.S., Ma, Z.Y., 2005. Friction stir welding and processing. Mater. Sci. Eng. R Rep. 50 (1-2), 1 78, http://dx.doi.org/10.1016/j.mser.2005.07.001. Sato, Y.S., Kokawa, H., 2001. Distribution of tensile property and microstructure in friction stir weld of 6063 aluminum. Metall. Mater. Trans. A 32 (12), 3023 3031, http://dx.doi.org/10.1007/s11661-001-0177-8. Schindele, P., 2010. Verfahren zum Fügen von Werkstücken (DE102010054453A1). Schindele, P., 2012. Method for friction squeeze welding lap joints (WO002013087176A1). Thomas, W.M., Nicholas, E.D., Needham, J.C., 1991. Improvements relating to friction welding (European Patent) (EP 0 615 480 B1). University Kempten, 2016. Research Project Friction Crush Welding (Reibquetschschweißen). http://www.hochschule-kempten.de/forschung/ forschungsschwerpunkte-und-projekte/forschungsschwerpunkt-3-fertigungsund-automatisierungstechnik/reibschweissen-und-reibbeschichten/ reibquetschweissen.html. Yusof, F., Firdaus, A., Fadzil, M., Hamdi, M., 2013. Ultra-thin friction stir welding (FSW) between aluminum alloy and copper, 219 224 10.1533/978-1-78242-164-1.219. The authors would like to thank Werner Saft (Laboratory for Joining and Welding), Prof. Dierk Hartmann and Petra Schittenhelm