Optimum influence of tensile functions on welded parts of AA 2024-T3 produced from friction stir mechanism utilizing air and water

Friction Stir Welding (FSW) is a modern technology employed in welding aluminum alloys in various industrial fields such as airplane industries. Therefore, AA 2024-T3 alloy can be one of the active materials used in this field. From here, numerous studies have been conducted on the natural flow of air and water to cool the welded joints of this alloy. Nonetheless, the forced air and water were not utilized with this alloy. Hence, the purpose of this study is to apply these forced convection mechanisms to specify the optimal outputs of tensile functions using multi-objective optimization by the General Full Factorial (GFF) technique. In addition, the rotational speeds were 800, 1000, 1200 rpm in this experimental medium. Therefore, the optimal parameters resulting from this work were 1200 rpm and forced water to achieve the best performance. Furthermore, the morphology of the fracture region is induced by these parameters leading to appearing big dimples. Finally, these dimples contribute to enhancing the plastic fracture without brittleness.


Introduction
It is no mystery the vital role of AA 2024-T3 alloy in the fields of airplane, aerospace, and aviation aircraft [1], [2]. The main reasons for employing this alloy are represented by high performance, good fatigue, lightweight, and corrosion resistance besides low cost. Accordingly, these properties have prompted the researchers to study the Friction Stir Welding (FSW) operation on this aluminum alloy as a promising mechanism [3], [4].
The strength and rotational speed are interesting points obtained from the prior studies of AA 2024-T3 alloy under the operation FSW. Therefore, these points have contributed to the feasibility of FSW by enhancing the life of welded specimens [5]. These outputs led to focusing on dealing with the tensile case to recognize the strength behavior. Consequently, this behavior has observably grown during the increase in rotational speed. Unfortunately, it is accompanied by this behavior dropping in the ductility of this specimen [6]. From here, these outcomes have proven the brittle of welded joints. Likewise, the elongation was boosted to 8.5%, while the ultimate tensile is 93.9% of un-welded AA 2024-T3 [7]. Thus, the side effects are associated with enhancing the strength property of welded joint [8]. The primary contributors to these effects were precarious grain size depending on the increase in the rotation of tool speed. As a result, this speed caused the unsymmetrical distribution of residual stresses of the welded parts at the maximum applied tensile [9], [10]. Hence, tensile studies were unfortunate under the natural flow of air.
The authors strived to use other cooling options instead of natural airflow to enhance FSW operation. Therefore, there were trials by using types of materials instead of AA 2024-T3 alloy. As examples of these premises, Al5083 alloy under this operation has been upgraded through applying water cooling. This upgrade led to refining grain size and reducing the distributed temperature. Furthermore, both tensile ability and ductility were improved in this alloy [11]. These behaviors also appeared on AA 7075-T651 alloy, besides the increasing hardness and yield point concerning both forced air and water [12]. Therefore, these outcomes have stimulated researchers in the FSW field to develop the properties of the welded region of AA 2024-T3 alloy. Accordingly, the microhardness of AA 2024-T3 alloy has advanced by employing a natural flow of water compared to the microhardness of FSW operating with airflow. This feature was due to refining microstructure using this unique hybrid technique of this flow type with FSW [13]. Hence, the immersing of this alloy underwater raised the microhardness from 138.7 HV to 139.3 HV. In contrast, the peak temperature of this immersed alloy was dropped to 453.3 ºC [14].
Depending on this brief survey, it turned out that the natural flow of water contributed to developing the welded region properties of AA2024-T3 samples, as compared to the un-forced air. In addition, the role of forced air and water in progressing the tensile, ductility, and yield point of AA 7075-T651 has been demonstrated. Accordingly, both forced air and water were not employed with the welded parts of AA2024-T3 in the FSW environment. Consequently, the present study aims to assign the tensile functions of welded components of AA 2024-T3 resulting from these forced flows. Moreover, the optimum output of parameters is selected using the Multi-Objective optimization based on the General-Full-Factorial (GFF) technique.
According to this context, the rotational speeds utilized in the current FSW mechanism are 800, 1000, and 1200 rpm. Furthermore, the forced flow is applied in this mechanism for both air and water, separately, besides the natural flow of air. Hence, these parameters' levels and this mechanism have been adopted to test a tensile of welded parts of AA 2024-T3. Therefore, it is essential to comprehend the functions of tensile results in the current study. Therefore, in the succeeding chapters: work materials, experimental conditions, designed trials, and an optimization technique are illustrated for exhibiting a concept of the accredited methodology to specify these functions.

Experimental medium and work materials
AA 2024-T3 alloy was employed as the work material's plates in the current study. Here, these plates were cut into two rectangular parts with dimensions (200×100×5) mm 3 , as demonstrated in Figure 1(C). The Friction-Stir-Welding (FSW) mechanism was performed by Semi-Automatic Vertical Milling Machine, India. As illustrated in Figure 1(a), the ejector nozzle has been applied on the FSW line at pressure 0.55 MPa for water and air. Besides, the machine's parameters utilized were 800, 1000, 1200 rpm, with 28 mm/min for rotational speed and feed rate, respectively. Meanwhile, the probe welding depicted in Figure 1(b) was installed as a tool in this machine to implement the mechanism of stir-welding by friction with these rectangular parts. This probe was made from water hardening tool steel, and the dimensions of this probe can be observed in Figure 1 On the other hand, Scanning Electron Microscopy (SEM)/Energy Dispersive X-Ray Spectroscopy (EDS) model (Hitachi-SU1510, Japan) has been employed to specify the chemical composition of work materials: workpiece and Probe-Tool of welding. In addition, SEM/EDS was wielded to capture the micrographs of the fracture zone resulting from the tensile test. In Figure 2(a) and (b), the observed chemical composition was demonstrated as a royal bar for Workpiece: AA 2024-T3 and Welding Probe: Water Hardening Tool Steel-AISI-W1, respectively, depending on the EDS test. In contrast, the yellow bar indicated the standard of chemical composition to the over-mentioned work materials. According to the SEM's job, the FIJI program has been employed here, as a twin-integrated, to convert 2D-micrograph to 3D-topography [15]. This topography, as planned in results, contributes to clarifying the morphology of the fracture zone. After tensile samples were being prepared by the Wire-EDM machine (Model: DK7732-SUNDA, China), each specimen was fixed in the testing machine (Model: WDW Series-Beijing United Test, China). In Figure  3(a), the fixing status of the specimen was shown in this testing machine. Moreover, the ASTM-B557M standard was used with each sample, as depicted in Figure 3(b). Therefore, this testing machine is responsible for specifying ultimate tensile strength (TS) and the ratio of tensile strength efficiency (%TSE), as functions in the present study. Accordingly, this study has reported the influence of the natural flow of air besides ejected air and water on these tensile functions.
As aforementioned, these functions are considered as the responses to determine the optimum performance of tensile samples. Here, this performance has wholly resulted from these flow mechanisms of air and water.

Experimental design and optimization
The experimental conditions of the FSW medium are very significant to control the performance of rectangular parts welding of AA 2024-T3, as clarified in Figure 1(a) and (c). Hence, these conditions can be divided into two kinds: fixed as feeding rate and nozzle pressure. Meanwhile, the other conditions are variable behaviors on the experimental side, as demonstrated in Table 1. Therefore, flow mechanism (FM) and rotational speed (RS) were considered variable conditions. In other words, these conditions were applied as variable parameters in this medium.
Depending on Table 1, the influential parameters' levels were observably adopted to design the experimental runs. These runs were devised employing General Full Factorial (GFF) technique in the present study. Hence, the models in Eqs. (1) and (2) = + 1 −1 + 2 0 + 3 1 + 4 −1 + 5 0 + 6 1 + 7 −1 −1 + 8 −1 0 + 9 −1 1 + 10 0 −1 + 11 0 0 + 12 0 1 + 13 1 −1 + 14 1 0 + 15 1 −1 , where Er indicates the total of experimental runs, L is the levels of each parameter, n is the replicates number of corner points, and P is the number of variable parameters. On the other hand, R refers to the response, A is the coefficient of regression, while X and Y are RS and FM, respectively. Thus, the whole number of the experimental runs applied in the present work is 18. In Table 2, the number of runs and the values of these responses at each run were demonstrated. The criterion adopted in this study for responses TS and %TSE were larger-the-better (LB) of desirability function for optimizing the multi-objective. Therefore, the statistical medium Minitab 18 was employed to implement the analysis and optimization of these responses. Hence, the desirability function (D) as composite value for this multi-objective can be seen in Eqs. (3) and (4) [16], [17], [18], [19], [20]:  where dk(R) is a sub-desirability function based on LB-Criterion of desired response Rn, while Ln and Tn are the lower limit of desired response and desired target for the response, sequentially. In addition, w and k are the weight factor and responses number, individually. Currently, the weight w applied to specify D in this multiobjective optimization equals one. This concise description has presented a whole concept of the design and optimization of the existing experimental environment. Therefore, the scenario of this environment contributes to analyzing, discussing, and interpreting the outcomes in Table 2. Moreover, these outcomes lead to selecting the optimal run for TS and %TSE.

Results and discussion
The statistical analysis is considered a pointer to specify the significant parameters of TS and %TSE responses depending on P-Values and F-Values. Accordingly, these values as outputs to Analysis of Variance (ANOVA) are computed based on the values of these responses exhibited in Table 2. In turn, the values of FIT and RES were observed as a preliminary evaluation of outcomes. Thus, the RES values are (1.5-3) and (1-2.5) in Table 2 for TS and %TSE, respectively.
In Table 2, the concrete results of FIT and RES led to boosting the experimental responses to be beneficial to conduct the ANOVA. Thus, this means identifying the vital parameters of TS and %TSE in the present research [21], [22], [23]. The outcomes of ANOVA of TS were significant to the linear states: RS and FM, besides the interacted case: RS×FM since the P-Values of these cases were <5%, as illustrated in Table 3 [24]. Furthermore, the F-Value and chart of Pareto's effect in Table 3 as an index reflect the parameters rank of this study. According to this index, both FM and RS were first and second, respectively. In addition, the interacted case RS×FM was third depending on the ultimate TS outcomes. Thus, the flow mechanism has boosted the ultimate tensile strength, although the rotational speed contributed to this enhancement. Hence, the outputs of ANOVA for TS of specimen produced by FSW of AA 2024-T3 in the current work corresponded to this behavior [13], [25]. Moreover, the regression model in Eq. (5), as a polynomial form of TS, fitted and correlated with experimental outcomes, where the correlation factors of this model were 99.5%, 99.05%, and 97.99% for R-sq, R-sq(adj), and R-sq(pred), correspondingly.

Rank. 1
On the contrary, ANOVA results of %TSE clarified that the rank of RS and FM are first and second, depending on F-Value and chart of Pareto's effect, sequentially. Meanwhile, P-Values were < 5% for RS, FM, and RS×FM, in which these values were significant, as depicted in Table 4. Additionally, the regression model for %TSE in Eq. (6) reflects the sobriety of statistical analysis performance of %TSE. This sobriety performance can be seen by the correlation factors R-sq, R-sq(adj), and R-sq(pred). Moreover, the values of these factors were 97.1%, 94.52%, and 88.39%. The decrease of the rotational speed of Probe-Tool plays a notable role in dropping the ratio of tensile strength efficiency of the sample under the tensile test [26]. However, it did not supply research evidence about the relation among %TSE response and the influence of flow mechanism (FM) in FSW, like the evidence available amongst TS and FM. Therefore, with the loss of this evidence, it is difficult to recognize the dominant parameter among RS and FM in this response. Consequently, the ANOVA for %TSE in Table 4 has presented novel outcomes demonstrating RS as a dominant parameter.

Rank. 1
The outputs of ANOVA for TS and %TSE have illustrated the significant role of FM to TS and RS to %TSE, as the topped influential parameters, in the performance of FSW of AA 2024-T3. However, this analysis did not specify the impacting level of RS and FM as a pivotal job. Hence, this job can be implemented by adopting Eqs.
(3) and (4) as multi-objective optimization. Additionally, the weight factor (w) employed in Eq. (3) is equal to one. As a result, the optimal predicted parameters can be recognized in Table 5 and Figure 4. Hence, these parameters were 1200 rpm and forced water (FW) for rotational speed and flow mechanism, severally. Furthermore, the composite desirability (D) is 0.9719 for the multi-objective optimization of TS and %TSE. On the other hand, both the experimental runs No. 9 and 18 have referred to the optimal runs since the parameters in each run are consistent with the optimal parameters. Besides, the residual values between predicted and experimental outcomes for TS and %TSE were 2 and 1, sequentially.  The increasing rotational speed up to 1200 rpm contributes to growing the generated temperature by FSW operation. Thereby, the heat deeply expands to penetrate the base material zone (BMZ) and enlarge the Heat Affected Zone (HAZ) depending on the recrystallization phenomenon and growth of the grains size [25], [27], [28]. Nevertheless, the water-cool leads to efficient control in enhancing the strength of the Welding Joint Zone (WJZ) by reducing this penetration and reducing the grains' size. Besides, the water-cool diminishes the nugget region to produce minimal corrosion of WJZ [13], [29]. Thus, the multi-objective optimization outputs induced from the present work is compatible with this interpretation.
To confirm the optimal outputs in Table 5 and Figure 4, the morphology of the fracture zone in Figure 5(a)-(c) can be observed under magnification 200 μm and 20 μm for 2D and 3D-SEM-Micrograph of run No. 9. In Figure 5(b) and (c), the distribution of big dimples was an indicator of high ductile and plastic deformation of the fracture zone [30]. By contrast, Figure 5(d)-(f) refers to the brittle fracture in run No. 10 as a worst-case among 18 runs based on the minimal criterion value of the ultimate tensile strength illustrated in Table 2.
Observably, the distribution of voids as described in Figure 5(e) and (f) in the morphology of the fracture zone is the primary reason for the brittleness in this run [25]. These encouraging findings of FSW on AA 2024-T3 had revealed the character of forced water, as an active convection case, in limiting the growth of grains size at high rotational speed to boosting TS and %TSE. Hence, this study proved that the statistical analysis and multi-objective optimization techniques have contributed to understanding the role of this convection in presenting a logical performance.

Conclusions
The present work has utilized air and water ejected, besides the natural air as boundary conditions under 800, 1000, and 1200 rpm, as the rotational speed in the FSW of AA 2024-T3. Hence, it has remarkably concluded to the following outputs: 1200 rpm and forced water are considered an optimal rotational speed and flow mechanism to implement FSW of AA 2024-T3 relying on Multi-Objective Optimization using GFF technique. Consequently, the fracture zone morphology under these optimal parameters had contained large dimples distributed along this zone. Therefore, these dimples' size has reversed the efficient performance of plastic deformation. Moreover, it was enhanced the strength of the welding-joint-zone.