The Effect of Pin Profiles and Process Parameters on Temperature and Tensile Strength in Friction Stir Welding of AL6061 Alloy

Abstract

The main source of the heat generation during the Friction Stir Welding (FSW) is the friction force between tool and workpiece and the plastic deformation. The geometry of the tool including the pin and the shoulder highly affects the friction force. In this study, the effects of different pin profiles with different rotational and traversing speed are evaluated in order to obtain the optimum pin profile using heat generation and tensile strength. Three different rotational speed and welding speeds are applied with threaded cylindrical, conical, stepped conical and square pin profiles. Thermocouples K type have been embedded in order to record the temperature during the welding at the advancing and the retreating side. Moreover, tensile test and microstructure analysis are performed in order to study the microstructure. The results of experimental process and design of experiments are correlated well. The better joint produced with threaded cylindrical tool pin profile with rotation speed of 1600 rpm and welding speed of 40 mm/min.

1 Introduction

Friction Stir Welding (FSW) was invented and patented by Thomas et al. at The Welding Institute (TWI) [1, 2]. There are three stages in the process plunging; welding stage and plunging out step. In the plunge stage, FSW tool which is made up of a pin and a shoulder, penetrates the plates. In some cases, there is a dwell time in which the rotating tool remains at the plate without forward motion. In the welding stage, the rotating tool moves forward along the welding seam in order to form a weld bead [3, 4]. The schematic of FSW is shown in Fig. 1 [5].

Fig. 1

Schematic of friction stir welding [3]

Although all of the welding stages are significant, the significance of the plunging stage is more than other stages, because the main part of the heat will be generated at this stage and the workpiece will be affected extremely by high temperature and stresses [6]. It should be noted that most tool wear occurs during plunge stage due to the high load and flow stress [7, 8]. In FSW, the parameters that influence quality of joint and reduce tool wear are rotational speed and traverse speed while it was reported that the influence of the geometry of the tool is more than the abovementioned process parameters [9, 10]. The geometry of the tool is separated into two different parts of shoulder and pin whereby both have a significant effect on the material flow and the welding temperature. Among all shapes that scholars considered, some of them are not compared with each other in the literatures [11,12,13,14,15,16,17,18,19,20,21]. To illustrate, Patil and Soman [13] only utilized Tri-flute and taper screw during different welding speeds. In the same way, [16] investigated different pin profiles in which the square profile was absent. [22] studied the influence of the pin profiles and the shoulder diameter on the formation of FSW stirring zone. They found that the square pin profile with 18 mm shoulder diameter have better weld quality in comparison with other pin profiles. In other research, they investigated the influence of the different pin profiles and the welding speed during the FSW. Their results indicated that the square pin profile produced defect free weld compared to other pin profiles [12].

[23], studied the influence of different pin profiles on the metallurgical and mechanical properties of Al-Metal Matrix Composite. They concluded that the square pin profile has better tensile strength in comparison with other pin profiles. However, the threaded cylindrical was not involved in their study. [24], investigated the effect of the different pin profiles on the mechanical properties of FSW of pure copper. Their results showed that the square pin profile have higher mechanical properties and better grain structure in comparison with the threaded cylindrical.

There are many considerable experimental and numerical studies on FSW of different alloys. [25] comprehensively reviewed the friction stir processes. [26] reviewed the structure and the properties of FSW. [27] gave a broad review of numerical analysis of FSW. [28] reviewed the different pin profiles. [29] studied finite element modeling approach for Friction Stir Spot Welding (FSSW) on Al6061. They applied adaptive mesh to reduce the high distortion during simulation. Besides, [30] considered the plunge stage using numerical modeling and experimental. They utilized Al2024 alloy for the experiment.

The critical part of FSW is the pin profile, because it affects the welding quality. Therefore, regarding the review of literature, in this paper, four FSW tool pin profile have been selected including; square (s), threaded cylindrical (TC), stepped conical (SC), and conical (C). Figure 2 shows the schematic of profiles.

Fig. 2

Different pin profiles. a Conical. b Square. c Threaded cylindrical. d Stepped conical

An appropriate design of the tool (especially the pin profiles) is able to generate proper heat and mixing the plasticized materials. Another significant factor which affects the heat generation is the process parameters such as rotating speed, traveling speed [25]. In this paper, in order to optimize the pin profile of FSW tool and process parameter during FSW, the thermomechanical behavior of the welded samples has been studied in detail.

2 Methodology

2.1 Experimental Set up

As mentioned earlier, four different pin profiles were selected for experiments. The heat treated H13 steel was used as the tool material and the workpiece AA6061-T6 with the dimension of 100 mm × 100 mm × 10 mm has been clamped. The chemical composition of a workpiece and the tool is shown in Table 1.

Table 1 Chemical composition and mechanical properties of base metal and the tool

The experiments were performed by FSW-TS-F16 friction stir welding device host machine. The single pass welding procedure has been used to fabricate the joints. The heat treatment operation is done on the FSW tool. The first step of the heat treatment is preheating cycle. According to the standard, ASTM E8 temperature for preheating is around 760 °C, held for 15 min. Then, it is soaking cycle in the austenite formation zone in which the temperature is raised up to 1010 °C and held for 30 min. After the soaking cycle, the tool is removed from the furnace and cooled to 65 °C, when the furnace has reached a temperature of 565 °C, the tools are allowed to a temper for 2 h [31]. Three rotational speeds and three traverse speeds were selected to evaluate the temperature during different speeds that are listed in Table 2. In order to record the temperature during the welding, thermocouples (K type) are embedded in the advancing side and retreating sides with the specified distance. Figure 3 shows the position of the thermocouples.

Table 2 Welding parameters and tool dimension

Fig. 3

Workpiece and thermocouples position

2.2 Microstructure Analysis

Microstructure analysis of FSW joints is performed as per ASTM to optimize the tool pin profile. In order to cut the samples, a wire cut machine has been used.

Then, the samples are grinded to remove material deformed produced from sectioning. Grinding process is followed by the polishing process. Polishing is removing the scratches from the surface of specimens to prepare them for the etching process. In order to perform etching, the specimens are immersed into Keller’s reagent for 10–20 s. The chemical composition of Keller’s reagent are listed in Table 3 [32, 33].

Table 3 Chemical composition of Keller’s reagent

2.3 Tensile Test

Tensile test is performed using a universal testing machine 50KN Amsler HA50 with standard ASTM: E8/E8M with constant crosshead speed of 0.9 mm/min. Tensile testing is a fundamental material science test that is subjected to a controlled tension force until the fracture. The results from the test are commonly used to select a material for an application, to control the quality of the weld, and also to predict the fraction of the material under different types of forces. Properties that are directly measured during a tensile test are ultimate tensile strength, maximum elongation and reduction in area. Tensile test performed using a universal testing machine with standard ASTM: E8/E8M utilized for tensile and performed with constant crosshead speed of 0.9 mm/min. Figure 4 shows the schematic of tensile samples for the test according to ASTM E8.

Fig. 4

Schematic of tensile test sample and dimension

2.4 Design of Experiment (DOE)

Design of experiments (DOE) is a systematic method in order to determine the relationship between factors affecting in a process and also to find the cause-and-effect relationships. Basically, DOE is a statistical technique for analyzing and organizing the experiments [34]. In DOE, the factors comprise of different parameters which are controlled by researchers, meanwhile the responses make up the dependent variable, which in this case, refers to productivity. To implement a DOE technique, some steps need to be followed such as: choosing the factors and their levels, choosing a response variable, choosing the experimental design, performing the experiments, analyzing the data, and promoting the best option [35].

In this research, the general factorial design has been selected for evaluating the effect of the several parameters on the heat generation and peak stress. In the first step for conducting the results. Table 4 and Table 5 show the factors and their levels for the heat generation and the stress respectively.

Table 4 Factors and levels for heat generation

Table 5 Factors and levels for peak stress

3 Results and Discussion

3.1 Heat Generation

Investigation of the heat generation in FSW is a complicated phenomenon and needs to be compared with the experimental data. In this study, heat generation of different parameters such as rotational speed, welding speed and tool pin profiles are investigated. Figure 5 illustrates the results of heat generation during FSW for advancing and retreating sides. These histograms show the maximum temperature from eight thermocouples for 36 sample welds. Highest temperature obtained from rotation speed of 1600 rpm for all pin profiles. For instance, for the threaded cylindrical tool pin profile, as the welding rotation speed increases from 800 to 1600 rpm, the temperature increases from 247.95 to 357.14 °C.

Fig. 5

Maximum temperature for different pin profiles

The Figure shows the relation of different parameters. As the welding increases from 40 mm/min to 100 mm/min, the temperature drops from 357.14 to 247.95 °C. This happens due to the reduced heat input per unit length and dissipation of heat over a wider region of workpiece at higher welding speed. On the other hand, by increasing the rotation speed, temperature would be increased. As the welding rotation speed increases from 800 to 1600 rpm, the temperature increases from 203 to 459.6 °C for the threaded cylindrical tool pin profile in the advancing side. This happens due to friction between tool and workpiece that generate more heat. Moreover, surface area of the tool effects heat generation. In this study, square pin profile produced a higher temperature around 460 °C due to its surface shape that creates more friction during FSW.

Advancing side in FSW process is the location from where the solid material starts to transform into semi-solid one and flows around the tool pin plunged into the material. The semi-solid material retreated and cooled in the retreating side. Therefore, advancing side has more solid state nature at any point of time/location compare to retreating side during FSW process. Therefore, advancing side should generate higher friction stress (unbalanced frictional force) which ultimately generates more heat and raises the peak temperature. Moreover, advancing side produced higher temperatures in comparison with retreating side due to pushing the material at the first connection and forwarding it to retreating side.

Temperature under the shoulder is higher due to high energy density. The peak of the temperature is about 80% of the material melting point 582–652 °C. From 36 welded samples, those where generates highest temperatures compared with each other to analyze heat generation during FSW with different pin profiles. Figure 6 shows the combined graphs. As can be seen in the Figure, square pin profile produced a higher temperature around 450 °C due to its surface shape that creates more friction during FSW.

Fig. 6

Temperature graphs from different pin profiles

During the welding process, the shoulder was the same, but pin profiles changed with different speeds. We could obtain different temperatures from different speeds. However, the graphs show that differences between peak temperatures of samples welded by different pin profiles are very little and not significant. According to Eqs. (1) and (2), pin profile affects heat generation.

$$ Q_{1} = 2/3\,\pi \tau_{contact} \omega \left( {R_{shoulder}^{3} - R_{pin}^{3} } \right)\left( {1 + tan_{\alpha } } \right) $$

(1)

$$ Q_{2} = 2\,\pi \tau_{contact} \omega R_{pin}^{2} H_{pin} $$

(2)

Equations (3) and (4) give the local heat generation rate due to friction and deformation work respectively.

$$ de_{f} = \delta \left( {r\omega - U\,sin\,\theta } \right)\mu_{f} pdA $$

(3)

$$ de_{s} = \left( {1 - \delta } \right)\left( {r\omega - U\,sin\,\theta } \right)\tau_{y} dA $$

(4)

where $\delta$ is the extent of the slip, $\mu_{f}$ is the friction coefficient, $\mu_{s}$ is the shear yield stress and $p$ is the local pressure applied by the tool on the elemental area $dA$. When $\delta$ is 1, no material sticks to the tool and all the heat is generated by friction. In contrast, when $\delta = 0$, work-piece material sticks to the tool and all the heat is generated by plastic deformation [36, 37].

According to the results of the experiment, higher temperature is generated from higher rotational speed and lower temperature comes from lower traverse speed. Therefore, increasing the rotational speed will increase the temperature and increasing the welding speed will reduce temperature that proved with other researches that studied in literature review. Moreover, the highest temperature is from a square pin profile that can be demonstrated with its surface area connection with the materials.

3.2 Microstructure Analysis

The etched specimens are examined using optical microscope Leica MC170 HD and the grain size of the cross section of the joints are analyzed. The cut specimens are immersed into Keller’s reagent for 10–20 s for the etching process. Finally, Scanning Electron Microscope (SEM) is used in order to investigate the microstructure.

It is notable that all microstructures of threaded cylindrical are homogenous and there are no significant differences between advancing and retreating side. Due to utilizing threaded shapes during FSW mixing of the materials is higher and also the grain size is better. The role of the pin profile for making the microstructure is not inevitable. The results showed that, the threaded cylindrical made a homogeneous microstructure in FSW in comparing with conical and stepped conical profiles. During FSW, the temperature variation is very high, therefore, the recrystallization of samples is not complete and are broken or are in the shape of a pancake. As per discussion earlier about the heat generation by increasing the rotational speed, the temperature is also increased and thus, the microstructure should be coarse. On the other hand, by increasing the welding transverse speed, the temperature falls down and thus the microstructure is finer.

A comparison between different microstructures at the stir zones in advancing and retreating sides under the various FSW conditions is shown in Fig. 7a–h. Figures (a) and (b) are related to advancing and retreating side of the conical profile with rotation speed of 1600 rpm and welding speed of 40 mm/min. Figures (c) and (d) are microstructure of square profiles. Threaded cylindrical shown in figures (e) and (f) and stepped conical is shown with figures (g) and (h). The results indicated that the microstructure of the sample with rotational speed of 1600 rpm with 100 mm/min welding speed is finer.

Fig. 7

Microstructure of advancing and retreating side. a Advancing side of stepped conical pin profile. b Advancing side of conical pin profile. c Advancing side of threaded cylindrical pin profile with rotation speed of 1600 rpm and welding speed of 100 mm/min. d Advancing side of threaded cylindrical pin profile with rotation speed of 800 rpm and welding speed of 40 mm/min. e Retreating side of stepped conical. f Retreating side of conical pin profile. g Retreating side of threaded cylindrical with rotation speed of 1600 rpm and welding speed of 100 mm/min. h Retreating side of threaded cylindrical with rotation speed of 800 rpm and welding speed of 40 mm/min

As per discussion earlier about heat generation by increasing rotational speed, temperature would be increased. Therefore, the microstructure should be coarse. On the other hand, by increasing the welding speed, temperature falls down and microstructure would be coarse again. Table 6 indicates the grain sizes.

Table 6 Grain size of joints in various samples

The role of pin profile for making the microstructure is not inevitable. Threaded cylindrical made a homogeneous microstructure in FSW in comparing with conical and stepped conical profiles.

In comparing with mechanical test and heat generation, sample with rotational speed of 1600 rpm with 100 mm/min welding speed produced sounds joint with finer microstructure.

3.3 Tensile Strength Analysis

After analyzing the microstructure, as can be seen in Fig. 8 some of the samples produced wormholes, therefore among 36 samples only 11 samples are Defect-free, and were selected for the tensile test. As mentioned earlier, these samples were tested using a universal testing machine with standard ASTM: E8/E8M utilized for tensile and performed with constant crosshead speed of 0.9 mm/min. Transverse tensile properties of FSW joints such as yield strength, tensile strength, and percentage of elongation have been evaluated. Table 7 illustrates the mechanical properties of selected samples. In another aspect, primary parameters such as welding speed has a significant role in tensile properties.

Fig. 8

Wormhole in FSW samples

Table 7 Tensile properties of 11 samples

As shown, maximum tensile strength (73.91% of parent material) is observed in that specimen in which the threaded cylindrical profile with rotational speed of 1600 rpm and welding speed of 100 mm/min is used. In contrast, the inferior properties belong to conical pin profile with the rotational speed of 800 rpm and the transverse speed of 40 mm/min. Therefore, lower tool rotational speeds produced insufficient intermixing. Thus, the mechanical strength will decrease at lower rotational speeds [38]. Moreover, it is observed that by increasing the welding speed, tensile strength will be increased. In another aspect, primary parameters such as welding speed have a significant role in the tensile properties.

Table 8 shows the tensile properties of those specimens in which maximum heat is generated. These results showed that increasing the rotation speed can increase temperature. On the other hand, by reducing the welding transverse speed temperature is also increased.

Table 8 Tensile properties of FSW joints with maximum heat generation

3.4 Design of Experiments Analyzes

3.4.1 Statistical Analysis for Heat Generation and Peak Stress

After choosing the factors, levels and the experimental parameters, data collection needs to be conducted by doing the experiments. As mentioned earlier by using general factorial design method, 36 conditions for the heat generation and 16 experiments for the peak stress are optimized. In order to decrease potential errors, two replicates for the heat generation and the peak stress is assumed. Therefore, as can be seen in Table 9 the numbers of the experiments for heat generation is equal to 72 (The number of experiments = 4 * 3 * 3 * (2 replicates) = 72).

Table 9 Result of Experiments for heat generation

As illustrated in Table 10, the number of the experiments for peak stress is equal to = 4 * 2 * 2 * (2 replicates) = 36.

Table 10 Result of experiment for peak stress

3.4.2 Determining of Significant Factors for Heat Generation

In this study, Expert-design software is used for conducting statistical analyses. The result of ANOVA for the heat generation and the peak stress are presented in Table 11. Basically, P-value is a significant parameter that is usually used to identify the statistically significant factors due to it has influenced the final responses. It should be noted that, whenever P-values are less than 0.05, factors should be considered as significant. In contrast, when P-values are more than 0.05, it should be assumed as an insignificant factor [39]. Based on Table 5, the Model, F-value of 6526.42 implies that the model is significant. In addition, the P-values is less than 0.0500 which indicate the model terms are significant. In this case A, B, C, AB, AC, BC, ABC are significant model terms.

Table 11 ANOVA result for heat generation

Table 12 verifies the accuracy of the model. As can be seen in the table, The “Pred R-Squared” of 0.9994 is in reasonable agreement with the “Adj R-Squared” of 0.9997. “Adeq Precision” measures the signal to noise ratio that a ratio greater than 4 is desirable. So ratio of 262.871 indicates an adequate signal.

Table 12 Model accuracy

3.4.3 Residual Analysis

The residual versus predicted value and normal probability plots are two significant graphical approaches that were used in order to check the validity of the regression model [40]. The results of the residual versus predicted value shows that the difference between the predicted values and the observed values. If the residuals have a regular pattern, it will infer that the suggested model is not adequate. Moreover, residuals in normal probability plot should be laid in a straight line [41]. As can be seen in Fig. 9, the straight line confirms that the model is adequate and correct. Moreover, the structure less pattern of the residual versus predicted value confirms that the developed model is adequate (Fig. 10).

Fig. 9

Normal plot of residual

Fig. 10

Residual versus predicted value

3.4.4 Peak Stress

3.4.4.1 Determining of Significant Factors

Table 13 shows the result of ANOVA for the peak stress. As can be seen, factors A, B, C and interactions AB and ABC are significant and have a considerable effects on the peak stress.

Table 13 ANOVA result for peak stress

Tabel 14 indicates that the “Pred R-Squared” of 0.8533 is in reasonable agreement with the “Adj R-Squared” of 0.9355. Also “Adeq Precision” measures the signal to noise ratio that a ratio greater than 4 is desirable. So ratio of 18.276 reveals an adequate signal.

Table 14 Model accuracy for peak stress

3.4.5 Residual Analysis for Peak Stress

As can be seen in Fig. 11, the straight line confirms that the model is adequate and correct. Moreover, the structure less pattern of the residual versus predicted value confirms that the developed model is adequate and has a constant error (Fig. 12).

Fig. 11

Normal plot of residuals

Fig. 12

Residual versus predicted value

3.4.6 Optimum Condition

The results of experiment and design of experiments have a good correlation. In most samples, advancing side produced higher temperature. Higher temperature obtained from square pin profile as discussed earlier due to its shape and geometry of contact area. However, from the experiments, we found that threaded cylindrical produced higher quality joints. It is observed that in some cases that temperatures are not exactly the same as DOE results because of some technical errors while getting temperature from thermocouples. Moreover, the results indicated that the highest tensile strength comes from threaded cylindrical with rotation speed of 1600 rpm and welding speed of 100 mm/min.

For the heat generation, the actual factor for the welding speed is obtained at 100 mm/min. Table 15 indicates the optimum parameters for heat generation and peak stress. The maximum temperature was 459.8 °C for the threaded cylindrical and rotation speed of 1600 rpm. Moreover, the stress of 237.68 MPa is obtained for threaded cylindrical.

Table 15 Optimum parameters from DOE

4 Conclusion

Experimental analyses and DOE are performed by different rotational and welding speeds with various pin profiles. During experimental step 11 samples are approved for further studies due to absence of weld defect.

• In regard to heat generation, the results of experimental and DOE are almost similar.

• Increasing the rotational speed will increase temperature. On the other hand, by increasing the welding speed, temperature comes down.

• Highest strength produced with higher rotational speed due to better intermixing of the materials during the welding process.

• In aspect of microstructure, With the increase of rotational speed, the welding speed grain size would be bigger. Moreover, pin profiles have an influence on the microstructure and grain size after welding.

• The results of experiments and DOE for heat generation, tensile test and microstructure analysis indicate that sounds joint produced with threaded cylindrical tool pin profile with rotation speed of 1600 rpm and welding speed of 100 mm/min.