3.1 Analyzing weld strength
Fig. 1 illustrates interaction effect of tool rotary speed and dwell time on lap shear strength of Al-Cu joints. It is seen in the figure that at 1000 RPM rotary speed, an increase in dwell time from 5 to 15s causes an 8.3% increase in weld strength. While at 1500 RPM speed, the weld strength firstly increases about 14% by increment of dwell time from 5 to 10s; but, by further increment of dwell time from 10 to 15s the weld strength decreases about 16.5%. on the other hand, when the tool rotary speed is 2000 RPM, it is seen from the fig. 1 that as dwell time increases, the weld strength decreases continuously about 27%.
When the tool rotary speed is 1000RPM, the heat input in friction stir processed (FSP) region is relatively low [16]. In such condition, by increase in dwell time the concentration of the heat increases that provides enough thermal energy for plasticization and stirring action. Thus, defects such as tunnel due to insufficient heat input is eliminated that causes increase in weld strength. Furthermore, due to concentration of more thermal at higher dwell time, the copper metal is softened and its contribution in formation of FSP region enhances. Hence, the weld shear strength increases. Fig. 2 represents macrostructure and microstructure of FSP region that produced at rotational speed of 1000 RPM and different dwell time. It is seen from macrostructure that by increase in dwell time the defect eliminated from weld macrostructure. Also, it is found from the figure that increase in dwell causes formation of further copper particles in FSP region that yields higher weld strength.
When the tool rotary speed is 1500 RPM, it is seen from the fig. 1 that by increase in dwell time from 5s to 10s the weld strength increases. As discussed, this improvement is due to providing enough heat input and sufficient material flow in FSP region that removes defect and enhances the weld strength. However, when the dwell time goes beyond a critical value of 10s, it is seen a drastic reduction occurs in lap shear strength. This decrease is because to the fact that at high dwell time, the thermal energy concentration is excessive that causes material softening. Therefore, by plunging force of the tool, the thickness of aluminum side decreases that results to a reduction in weld strength [17].
Fig. 3 illustrates macrostructure of the weld region at 1500 RPM rotary speed and different dwell time. It is found from the figure that a 5s dwell time due to less concentration of thermal energy a tunnel like defect is formed between aluminum and copper in keyhole region. Also, it is found that at 10s dwell time the defect is eliminated and sound joint is fabricated. On the other hand, it is inferred from fig. 3c that at 15s dwell time due to excessive heat input material softening occurs that causes thickness reduction in aluminum side and formation of pin hole defect in keyhole region that restricts weld strength.
However, it is seen from the fig. 1 that at 2000 RPM tool rotation, by increase in dwell time the weld strength decreases, subsequently. When the tool rotation is 2000 RPM, the heat input is relatively high; hence, at 5s dwell time the sufficient thermal energy is provided that is enough for plasticization, stirring action and material flow. However, by increase in dwell time, due to excessive heat input, the material softening occurs in both aluminum and copper that causes a reduction in the thickness of the sheets. In such condition the strength of the lap configuration drastically decreases.
The macroscopic image of weld cross section at 2000RPM rotary speed is visible in fig. 4 It is ascertained from the figure 4a that at 5s dwell time the macrostructure is free from defect and thinning in the thickness. While, at 10 and 15s dwell time due to excessive heat input and material softening, a sever thinning occurs in weld macrostructure that damages the joint strength.
3.2 Analyzing weld nugget hardness
In order to analyze the hardness of weld nugget. The cross section of the joint was prepared and the microhardness was measured in three locations of aluminum side, copper side and interface; then the average of microhardness for each weld sample was reported. Fig. 5 indicates effect of dwell time on hardness of the joint which were fabricated under different tool rotary speed.
It is seen from the figure that at 1000 RPM tool rotation, the hardness value decreases about 13.8% by increase in dwell time from 5s to 15s. Also, when the tool rotary speed is 1500 RPM, as the dwell time increases, the hardness of weld nugget increases about 16% and reaches to a maximum value at 15s. Moreover, at 2000 RPM rotation speed of the tool, the hardness value firstly increases about 12.6% by increase in dwell time from 5s to 10s; but, by further increase in dwell time from 10s to 15s, the hardness decreases about 10%.
During welding of dissimilar materials. There are two important factors that significantly affect hardness of the weld nugget. One is size of microstructure and two is formation of intermetallic compounds [18]. According to Hall-Petch law, the finer microstructure results to higher hardness values. Also, intermetallic compound is type of ceramic material with high hardness and brittleness that formed under high heat input condition.
At 1000 RPM, tool rotary speed, the heat input is relatively low. Hence, sufficient heat for formation of intermetallic compound isn’t provided. In such condition increase in dwell time provides enough time and heat for recrystallization and enlarging the grains in the microstructure. Therefore, the hardness of weld nugget decreases by an increase in dwell time due to formation of coarse microstructure in weld nugget. Fig. 6 illustrates XRD pattern and microstructure of the weld nugget at 1000 RPM tool rotation and different dwell time. It is inferred from the fig. 6a that no intermetallic compound is formed at 1000 RPM tool rotation and different dwell time. Also, from fig. 6b and 6c, it is seen that by increase in dwell time the microstructure of the weld nugget becomes coarser in both the aluminum and copper sides that reduce the hardness value.
It is also found from the fig. 7 that at 1500RPM tool rotation, the hardness increases by increase in the dwell time. When the tool rotation is 1500 RPM, enough heat for formation of intermetallic compound is provided. In such condition by increase in dwell time concentration of thermal energy enhances that causes formation of more intermetallic compound in weld nugget. Therefore, weld nugget hardness increases. Fig. 7 represents XRD pattern of the samples fabricated at tool rotation of 1500RPM under different dwell time. It is seen from the figure at 5s dwell time due to low heat input, no intermetallic compound is formed in weld nugget. Also, it is seen when the dwell time is 10s, compounds such as Al2Cu and Al4Cu9 are seen in the XRD pattern. It is also seen that at 15s dwell time, number of peaks in XRD pattern is relatively higher showing more amount of intermetallic compound that exist in the weld nugget.
It is also ascertained from fig. 5 that at 2000 RPM tool rotation, the hardness value firstly increases as dwell time increases from 5 to 10s. This increase is due to formation another type of intermetallic compound like Al2O3 ceramic in weld nugget that significantly enhances the hardness value. It is further seen in the figure that as dwell time increases from 10s to 15s, the coarse and rough microstructure are formed in weld region. Nevertheless, the ceramic compound formed when dwell time is 15s, but the roughening effect of microstructure outperforms influence of intermetallic compound that causes low hardness value. Fig. 8 illustrates XRD patterns and microstructure change by dwell time at 2000 RPM rotation speed. It is seen from the fig. 8a at 10s and 15s dwell time a new type of intermetallic compound such as alumina is formed in the nugget. Therefore, the hardness increases. Also, from the fig. 8 b and c, it is observed that increase in dwell time causes coarse microstructure in both sides of aluminum and copper. Thus, a very rough structure damages the mechanical properties of weld nugget and decreases the hardness at 15s dwell time.
3.3 Statistical analysis
In order to develop empirical relationship between tool rotation sped and dwell time to lap shear strength and hardness response surface methodology is utilized. Design Expert V7 software was utilized here for regression analysis. The analysis of variances (ANOVA) was also carried out to check the validity of developed quadratic model. It was also used to identify which factor has greatest contribution on process quality characteristics. Second order polynomial model of responses including linear, interaction and quadratic terms of tensile strength and hardness are expressed in Eq. 1 and Eq. 2, respectively.
Analysis of variances of lap shear strength and hardness have been presented in Table 2 and 3, respectively. It is seen in ANOVA tables that the values of R2 (i.e. coefficient of determination) is in close agreement with adjusted R2. It describes that the developed models are completely valid to navigate design space [19]. Based on the analysis of variances, it is inferred that the dwell time and tool rotation are most influential factor for lap shear strength and hardness, respectively. Also, the contributions of the significant terms in developing mathematical models of lap shear strength and hardness are shown in Fig. 9. It is seen from the figure that in developing statistical model of the lap shear strength interaction of tool rotation and dwell time has significant influence; while for mathematical model of the hardness, the linear effect of tool rotation is the most significant term.
The developed empirical models can be also used to analyze effect of process factors on lap shear strength and hardness. Hence, the response surfaces of process factors on aforementioned responses were drawn and presented in fig. 10. It is seen from the fig. 10a that the maximum lap shear strength is achievable by selection of 2000 RPM tool rotation and 5s dwell time. Furthermore, maximum hardness could be obtained when the tool rotary speed is 2000 RPM and dwell time is 10s. it is seen that the variation of lap shear strength and hardness by variation of process factors don’t follow a similar trend. Thus, to achieve a unified result, its required to find optimal result in a multi-objective optimization problem.
Table 2 ANOVA reslts for lap shear strength
Source
|
Sum of squares
|
Degree of freedom
|
F-value
|
Prob>F
|
Model
|
246000
|
5
|
11.43
|
0.0029
|
N
|
3952.67
|
1
|
0.92
|
0.3699
|
t
|
37446
|
1
|
8.7
|
0.0214
|
N×t
|
109600
|
1
|
25.45
|
0.0015
|
N2
|
104
|
1
|
0.024
|
0.8839
|
t2
|
83486
|
1
|
19.39
|
0.0031
|
R2=0.9358
|
R2Adjusted=0.9188
|
Table 3 ANOVA results for hardness
Source
|
Sum of squares
|
Degree of freedom
|
F-value
|
Prob>F
|
Model
|
601.47
|
5
|
6.87
|
0.0125
|
N
|
368.17
|
1
|
21.03
|
0.0025
|
t
|
2.67
|
1
|
0.15
|
0.7
|
N×t
|
36
|
1
|
2.06
|
0.1947
|
N2
|
94.5
|
1
|
5.39
|
0.053
|
t2
|
30.9
|
1
|
1.77
|
0.2259
|
R2=0.9591
|
R2Adjusted=0.9349
|
Multi-characteristics optimization was carried out by desirability function. The toolbox of this method in Design expert software was adjusted based on range of process factors and criteria which were specified in Table 4. By performing the optimization in the software, the obtained results are identified and presented in Table 5. It is seen from table, illustrates the optimum solution that obtained through desirability function. It is evident from the table that setting of 1972 RPM tool rotation and 8.38s dwell time is the most optimal solution that causes desirability of 85.6% with strength value of 1894N and hardness of 84V. The obtained solution should be experimentally verified to show applicability of proposed approach in optimization of FSSW process. The results of confirmatory experiment have been also presented in table 4. It is found from this table that the error values for both of lap shear strength and hardness are less than 8% that shows the excellent agreement with actual and predicted values.
Table 4 Optimization criterion
Factors/Responses
|
Criterion
|
Importance
|
Tool rotation (RPM)
|
In range of 1000-200
|
-
|
Dwell time (s)
|
In range of 5-10
|
-
|
Lap shear strength (N)
|
Maximize
|
*****(5)
|
Hardness (V)
|
Minimize
|
***** (5)
|
Table 5 Optimum parameter setting along with validation test results
Factor
|
Lap shear strength (N)
|
Hardness (V)
|
Objective
|
N (RPM)
|
t (s)
|
Experiment
|
Model
|
Error
|
Experiment
|
Model
|
Error
|
Desirability
|
1982
|
8.3
|
1803
|
1894
|
0.05
|
78
|
84
|
0.076
|
0.856
|