Author: Dr. Nisith R. Mandal, FIE, FRINA, Professor, Dept. of Ocean Engg & Naval Architecture, Indian Institute of Technology, Kharagpur
Friction Stir Welding (FSW) is a solid state process where consolidation of the thermally plasticized material takes place through the retreating side behind the tool that forms the welded joint. Temperature distribution in the transverse direction was found to be non-symmetric with higher temperatures located on the advancing side of the weld. The plasticized material in FSW remains not in liquid state but in highly viscous state, hence the effect of convective heat transfer is not significant. The material in the nugget zone undergoes dynamic recrystallization. The resulting microstructure and thereby the mechanical properties of the weld joint depends on the temperature fields, post-welding cooling rates and the straining rate. A correlation between the temperature attained, cooling rate as well as the straining rate of the heated weld nugget with the microstructure obtained needs to be worked out. It was generally observed that in the present work higher tool rotational speeds deterioration of weld quality took place. Purely by trial and error it was found that near consistent results were achieved in the range of 500 to 700 rpm. The dynamics of heat generation in FSW as developed in this work corroborates well with experimental measurements. This model can be used to study the effect of peak temperatures and cooling rate on the resulting microstructure of the welded joints. The aspect of dynamic recrystallization due to the thermal cycle and the straining rate and its effect on the resulting joint strength needs to be investigated to achieve consistency in weld quality.
Friction Stir Welding (FSW) is a solid state process where the peak temperature generally attend about 80% that of the liquids’ temperature of the material being welded. The joint produced in this process is asymmetric about the weld line as the material in a highly plastic state flows differently at the two sides of the welded joint. On one side of the tool, the rotational direction is the same as that of the tool travel direction. Such type of process is referred to as ‘advancing side’ and the other side is referred to as the ‘retreating side’.
There are Four distinct micro structural zones observed in FSW;moving towards the weld line from the parent metal, the four zones are, (i) parent metal zone, (ii) heat affected zone, HAZ, (iii) thermo-mechanically affected zone, TMAZ, and (iv) nugget zone. A complex interaction of process parameters, heat generation and dissipation and material flow due to sever shearing action leads to the development of the microstructure of the nugget zone, which primarily controls the weld quality and process efficiency. The material flow that takes place can be divided in 2 distinct components, (i) material in direct contact with the tool is forced to flow around the rotating tool due to friction and shearing action, (ii) threaded tool tends to push the material in downward direction closer to the tool whereas due to material-incompressibility condition, an upward motion of the material is also produced at a distance further away from the tool axis (Jata et.al 2000 and Nandan et.al. 2008).
Consolidation of the thermally plasticized material takes place at the region behind the tool that forms the welded joint. This metal transfer takes place through the retreating side. Worm hole defect is often found at the advancing side near the nugget/TMAZ interface. The following are some of the features of the FSW temperature fields and cooling rates as reported in [Nandan et.al.2008, Russel et.al1999, Sato et.al. 2002, Peel et.al. 2006).
Temperature distribution in the transverse direction is generally non-symmetric with higher temperatures often located on the advancing side of the weld. Temperature distribution in the longitudinal direction shows substantially higher temperature gradients in front of the traveling tool which happens due to the tool pin experiencing higher friction.
If the plasticized material is considered as a fluid, then the Peclet number, quantifying the relative heat transfer by convection to that of conduction, works out to ten to few hundreds., Moreover, one may conclude that convection plays an important role in heat transfer in FSW. However, the plasticized material in FSW remains in highly viscous state and not in liquid state, hence the effect of convective heat transfer is not expected to be significant.
The material in the nugget zone undergoes heavy plastic deformation as well as a condition of annealing. This leads to dynamic recrystallization, as a result of which the material strength is lowered relative to that in the base (H131 temper condition) material. The resulting microstructure and thereby the mechanical properties of the weld joint depends on the temperature fields and post-welding cooling rates.
HEAT GENERATION
In friction, stir welding heat is generated due to plastic deformation occurring in the shear layer and friction at the tool-work piece interface. The heat generation due to friction is the product of frictional force and the velocity difference between the tool and the interface layer. Whereas, the heat generated due to plastic deformation at the tool-work piece interface, is the product of shear stress and the velocity of the interface shear layer sticking to the tool. This velocity is actually the tangential speed of the tool.
Two distinct tool-work piece interface surfaces are tool shoulder and tool pin side. Whereas, the contribution due to the tool pin tip surface is negligible (Daiela et.al. 2010) towards the total heat generation. Q1 and Q2 are the components of the respective heat generated from these interfaces as shown in Fig.1.
Tool shoulder – Work piece interface
The heat generated at the tool shoulder-work piece interface has already been dealt with in (Pankaj et.al. 2011) hence not mentioned here . Considering pure friction, the heat generated at the tool shoulder and the plate interface is given by,
For welding of 6mm aluminum alloy plates, the shoulder concavity α = 60.
Tool pin – Work piece interface
As per the phenomenological point of view, achieving friction stir welded joints material flow should happen between the two mating plate edges and this is the reason why it points to the existence of flow in the shear layer around the tool pin side surface and this in further results to a state of sticking condition which was considered to estimate the extent of heat generation at the tool pin side surface to plate interface. The heat generation at this interface due to plastic deformation is given by Eq. (2).
At the same time, as the FSW tool traverses along the joint, the forward half of the tool pin experiences a reaction force F as shown in Fig.2. It is given by the product of the projected area of the tool pin and the yield stress of the aluminum alloy at the prevailing temperature of pin plate interface as shown in Eq. (3). The temperature is approximated about 80% of the liquid s temperature (5700C ) of the plate material i.e. 4560C in the present case.
Fig. 2: Lateral force acting on FSW tool pin during welding.
Therefore, the frictional force experienced by the tool pin surface will be given by as,
Hence, the heat generated due to friction of the tool pin side surface will be,
Therefore, the total heat generation considering both sticking and sliding at the pin side surface interface will be given by,
Since the tool traverses forward as it rotates, it is logical to consider the effects of shearing of metal as well as friction remaining limited to only front half of the pin surface. Hence, the contact parameter was taken 0.5. Therefore, Eq. (6) becomes,
Substituting from (3) and (4) in Eq. (7) one obtains,
EXPERIMENTAL RESULTS
Extensive test welds were carried out with AA5083. A typical setup is shown in Fig.3. These were welded with varying combinations of tool rotational speeds (710 to 2000rpm) and traverse speeds (56 to 225 mm/min). It was observed that rotational speed of 1000 rpm gives fairly consistent results at different welding speeds. Temperature recording at a location 20mm away from the weld line on top of plate surface and at the weld line at the bottom of base plate was done using 4700A Agilent data logger. The comparison of the recorded data with the theoretical ones shown in Figures 4 to 6 for different tool traverse speeds corroborates the theoretical basis presented in this work.
Fig.3: A typical FSW setup
Fig.4: Comparison of temperature distribution at the top of plate surface 20mm away from weld central line.
Fig.5: Comparison of temperature distribution for 1000 rpm and 112mm/min welding speed 20mm away from weld central line on top surface of AA5083 plate.
Fig.6: Comparison of temperature distribution for 1000 rpm and 112mm/min welding speed at weld central line on bottom surface of C-Mn steel base plate.
Mechanical Response
The welding parameters and the corresponding range of tensile strength achieved are shown in Table 1. A trend of improvement of tensile strength parameter with increasing welding speed up to 160mm/min could be noticed here; however, with further increasing, a drastic drop in strength was recorded at the welding speed of 224mm/min. The reason of this behavior has not been satisfactorily established yet. Nevertheless, it may appear from the results summary that a combination of 1000 rpm and 160 mm/min weld speed resulted in better tensile strength and also a better consistency of results were obtained with 1000 rpm and 80 mm/min weld speed.
Table 1: Weld parameters and tensile test results.
| Sl. No. | Weld parameters | Tensile strength (MPa) |
|
| Tool rotational speed (rpm) | Tool traverse speed (mm/min) | ||
| 1 | Unwelded AA5083 | 304 – 315 | |
| 2 | 1000 | 80 | 150 – 240 |
| 3 | 1000 | 112 | 160 – 225 |
| 4 | 1000 | 160 | 165 – 256 |
| 5 | 1000 | 224 | 108 – 153 |
Some of the representative ruptured surfaces of tensile test specimens are shown in Fig 7 through 10.
Fig.7: Fractured surfaces of unwelded AA5083 and sample 19 & 20
Fig.8: Fractured surfaces of samples 55 & 58
Fig.9: Cup and cone fractured surfaces of sample 31
Fig.10: Ductile and brittle phases at the fractured surfaces
The fractured surface of unwelded AA5083 test sample is shown in Fig.7, wherein one can also observe near similar fractured surface of test samples 19 and 20. However, one can clearly see the surface texture of unwelded AA5083 is uniform over the entire surface. On the other hand, in case of S – 19 and 20 the fractured surface orientation is near identical but the texture is not uniform over the surface. The top and the bottom half shows ductile fracture but in between the surface reveals brittleness. It is also well reflected in tensile strength. The tensile strength of unwelded AA5083 was about 310 MPa, whereas, for the S-19 & 20 about 256 MPa.
The surfaces shown in Fig.8, compared to those in Fig.7, one can observe very brittle kind of fractured surface of the samples 55 & 58. As can be expected by observing the fractured surfaces, the tensile strength of about 150 MPa was obtained. Fig.9 shows a very interesting phenomenon of cup & cone fractured surface. It is not known why and how this type of surface formed in some cases. In Fig.10, the brittle and ductile phases at the fractured edges are very clearly visible.
Observations
In friction stir welding, the aspect of heat generation is very important as it strongly influences the entire process of welding and its resulting mechanical properties. The peak temperature depends primarily on the shoulder diameter and the tool plunging force. The tool rotational speed determines the straining rate; whereas, the weld speed determines the heat flow rate, thereby also the cooling rate. A variation in temperature distribution in the advancing and retreating side along the transverse direction was observed as shown in Fig.11.
Fig.11: Comparison of temperature distribution 20mm away of weld line at the advancing and retreating side of the weld
All these finally determine the grain structure of the weld nugget which controls the strength of the welded joint.
Based upon the above formulation development in this work, the physical measurements of temperature profile showed a good agreement with those of the calculated ones. . The temperature field created affects the metallurgy of the weld nugget as well as the thermo mechanically affected zone. It was observed that the recrystallization of the heavily stirred zone, i.e. the weld nugget determines the strength of the welded joint. A correlation between the temperature attained as well as the cooling rate of the heated weld nugget with the microstructure obtained needs to be worked out. The extent of brittle phase in the weld nugget should be minimized as can be seen in Fig.7 in samples 19 and 20 to achieve superior mechanical properties of the joint.
Increasing weld speed causes less of heat flow in the weld surrounding, i.e. lower rate of heat input. It leads to faster cooling of the heated material as well as gives lesser time for recrystallization. It was also observed that with increasing weld speed up to 160mm/min there was a general improvement of tensile strength of the welded samples, ; however, increasing the speed further to 224mm/min caused a drastic reduction in strength. The fractured surfaces are shown in Fig.8, wherein it can be observed that a significant amount of the material along the plate thickness in the weld nugget region constituted of brittle phase. Higher cooling rate causing faster solidification of the plasticized viscous material may have led to formation of the brittle phase.
As it stands, a certain minimum level of plunging force is necessary for achieving the full penetration of the joint as well as generation of required heat to produce necessary plastification of the material in a such way that adequate material stirring can be achieved. At the same time, the plunging pressure should not exceed as much as such that expulsion of material takes place due to digging in effect of the tool shoulder. Hence one finds that not much of play is possible in variation of plunging force. It was also observed that variation of tool rpm does not significantly affect heat generation, ; however, it does have an effect on straining rate. In the present work, it was observed that too high (2000) or too low tool rpm (500) led to deterioration of weld quality. Purely by trial and error it was found that near consistent results were achieved at 1000 rpm , The most critical parameter in case of FSW of AA5083 is that of weld speed because, keeping 1000 rpm and fixing the speed of welding always vary.
The thermal model developed in this work could be used to evaluate the effect of varying welding speed on the heat flow characteristics of the welding process. This will help in working out a correlation between the attained peak temperature , cooling rate and microstructure of weld nugget, in a way that consistent weld quality can be achieved.
CONCLUSION
The heat generated in FSW strongly influences the entire process of welding and its resulting mechanical properties. The peak temperature depends primarily on the shoulder diameter and the tool plunging force. The tool rotational speed influences the straining rate, whereas, the weld speed determines the heat flow rate, thereby the cooling rate as well. A variation in temperature distribution occurs in the advancing and retreating side along the transverse direction.
Higher cooling rate causing faster solidification of the plasticized viscous material may have led to formation of the brittle phase in the weld nugget. The extent of brittle phase in the weld nugget should be minimized to achieve superior mechanical properties of the joint.
A correlation between the temperature attained, cooling rate as well as the straining rate of the heated weld nugget with the microstructure obtained needs to be worked out. In the present work it was observed that too high (2000) or too low tool rpm (500) led to deterioration of weld quality. Purely by trial and error it was found that near consistent results were achieved at 1000 rpm.
The dynamics of heat generation in FSW as developed in this work corroborates well with experimental measurements. This model can be used to study the effect of peak temperatures and cooling rate on the resulting microstructure of the welded joints. The aspect of dynamic recrystallization and its effect on the resulting joint strength needs to be investigated to achieve consistency in weld quality.
