51- 64 Al-Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. Determination of Optimum Welding Parameters for FSW Samir Ali Amin Al ,**,*** Department* (Received Abstract Friction stir welding is a relatively new joining process, which involves the joining of metals without fusion or filler materials. In this study, the effect of welding T351 joints produced by FSW was investigated. Different ranges of welding parameters, as input factors, such as welding speed (6 speed (725 - 1235 rpm) were used to obtain their influences on the main responses, in terms of elongation, tensile strength, and maximum bending force. Experimental measurements of DESIGN EXPERT 8 experimental design software which was used to (RSM) models. Mathematical model of responses, as functions of used welding conditions, were obtained and analyzed by ANOVA variance to verify the adequacy of these models. The resultant quadratic models showed tha speed or welding speed increases, the tensile strength and elongation of the joint and then decrease more likely due to the occurrence of void defect. leads to increase the maximum bending force firstly to a maximum value and then decreases. However, the welding speed was found more significant than rotational speed. models and optimization with the experimental ones with confidence level of 95%. Keywords: Friction stir welding, DOE, RSM, 1. Introduction Friction stir welding (FSW) is a solid joining technology invented at the welding institute (TWI) in 1991. It has been proven to be a very successful joining technology for alloys. Compared to the conventional welding processes, FSW can produce superior mechanical properties in the weld zone. This new technique is attracting more and more research interest [1]. The FSW process appears to offer a number of advantages over conventional fusion welding techniques, such as no need for expensive consumables such as filler wire and gas shields, ease of automation on simple milling machinery, Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. 51- 64 (2015) Determination of Optimum Welding Parameters for FSW AA2024-T351 Al-Rubaie* Qasim Abbas Atiah Zuhair Altaher*** Department of Mechanical Engineering / University of Technology *E-mail: alrabiee2002@yahoo.com **E-mail: dr_qasim_uot@yahoo.com ***E-mail: zuhairsadeed@gmail.com (Received 18 June 2014; accepted 15 December 2014) Friction stir welding is a relatively new joining process, which involves the joining of metals without fusion or filler materials. In this study, the effect of welding parameters on the mechanical properties of aluminum alloys AA2024 T351 joints produced by FSW was investigated. Different ranges of welding parameters, as input factors, such as welding speed (6 - 34 mm/min) and rotational to obtain their influences on the main responses, in terms of elongation, tensile strength, and maximum bending force. Experimental measurements of main responses were taken and analyzed using DESIGN EXPERT 8 experimental design software which was used to develop the response surface methodology models. Mathematical model of responses, as functions of used welding conditions, were obtained and analyzed by ANOVA variance to verify the adequacy of these models. The resultant quadratic models showed tha speed or welding speed increases, the tensile strength and elongation of the joint firstly increase to a maximum value due to the occurrence of void defect. Increasing both welding speed and rotational speed leads to increase the maximum bending force firstly to a maximum value and then decreases. However, the welding speed was found more significant than rotational speed. A good agreement was found between the results of these experimental ones with confidence level of 95%. stir welding, DOE, RSM, Mechanical properties, Modeling and Optimization Friction stir welding (FSW) is a solid-state joining technology invented at the welding institute (TWI) in 1991. It has been proven to be a very successful joining technology for aluminum alloys. Compared to the conventional welding processes, FSW can produce superior mechanical properties in the weld zone. This new technique is attracting more and more research interest [1]. The FSW process appears to offer a number of over conventional fusion welding techniques, such as no need for expensive consumables such as filler wire and gas shields, ease of automation on simple milling machinery, good mechanical properties of the resultant joint, and low distortion [1]. During the FSW process, the tool penetrates into the workpiece, and then moves along the joint line at a constant speed (see Figure 1). The material in front of the rotating tool pin is plastically deformed and stirred back to the trail edge of the tool pin in the The tool serves three primary functions, that is, heating of the workpiece, movement of material to produce the joint, and containment of the hot metal beneath the tool shoulder. Heating is created within the workpiece both by friction between the rotating tool pin and shoulder and by severe plastic deformation of the workpiece. The localized heating softens the material around the Al-Khwarizmi Engineering Journal (2015) Determination of Optimum Welding Parameters for FSW Qasim Abbas Atiah** University of Technology Friction stir welding is a relatively new joining process, which involves the joining of metals without fusion or filler parameters on the mechanical properties of aluminum alloys AA2024- 34 mm/min) and rotational to obtain their influences on the main responses, in terms of elongation, tensile were taken and analyzed using response surface methodology models. Mathematical model of responses, as functions of used welding conditions, were obtained and analyzed by ANOVA variance to verify the adequacy of these models. The resultant quadratic models showed that as the rotation firstly increase to a maximum value Increasing both welding speed and rotational speed leads to increase the maximum bending force firstly to a maximum value and then decreases. However, the welding A good agreement was found between the results of these Mechanical properties, Modeling and Optimization . good mechanical properties of the resultant joint, e FSW process, the tool penetrates into the workpiece, and then moves along the joint line at a constant speed (see Figure 1). The material in front of the rotating tool pin is plastically deformed and stirred back to the trail edge of the tool pin in the welding. The tool serves three primary functions, that is, heating of the workpiece, movement of material to produce the joint, and containment of the hot metal beneath the tool shoulder. Heating is created within the workpiece both by friction e rotating tool pin and shoulder and by severe plastic deformation of the workpiece. The localized heating softens the material around the Samir Ali Amin Al-Rubaie Al-Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. 51- 64 (2015) 52 pin and, combined with the tool rotation and translation, leads to movement of material from the front to the back of the pin, thus filling the hole in the tool wake as the tool moves forward. The tool shoulder restricts the metal flow to a level equivalent to the shoulder position, that is, approximately to the initial workpiece top surface. As a result of the tool action and influence on the workpiece, when performed properly, a solid state joint is produced, that is no melting. Because of various geometrical features on the tool, material movement around the pin can be complex, with gradients in strain, temperature, and strain rate. Accordingly, the resulting nugget zone microstructure reflects these different thermomechanical histories and is not homogeneous [1]. A lot of researches [2-7] has been already done towards understanding the effect of FSW process parameters on the material flow behavior, microstructure formation, tool design and mechanical properties of friction stir welded joints, but there is a few works on studying the influence of input factors on mechanical properties using design of experiments (DOE) and response surface methodology (RSM) technique. In this investigation, an attempt has been made to determine the effect of FSW process parameters on mechanical properties (elongation, tensile strength, and maximum bending force) in Al alloy (AA2024-T351). 2. Experimental Work 2.1. Selecting the aluminum alloy and specimens preparation of plates The base material used in this investigation is 2024-T351 which was obtained from a local market with thickness of (3.2 mm). AA2024-T351 aluminum alloy is Al-Cu-Mg grade alloy of 2xxx series heat treatable of medium strength alloys. A piece of this alloy was analyzed to find its chemical composition by spectro device, as shown in Table 1 with the standard material according as ASTM B209M [8] and the standard mechanical properties of AA2024-T351 aluminum alloy, given in Table 2. The base material cut into required size is (210 mm *110 mm *3.2 mm) by a power saw cutting convenient for conducting FSW, and the plate edge was ground to ensure that there is no gap between the two plates. 2.2. Design and Manufacturing of Welding Tools The design of the tool is the key to the successful application of the process to a greater range of material and over a wider range of thickness. FSW tool of square pin profile and straight cylindrical shoulder was used. The geometry and dimensions of the tool are pin rotational diameter of 5 mm, shoulder diameter of 15 mm and pin length of 2.7 mm, see Figure 2. The friction stir welding tool was manufactured by CNC turning and milling machines, this friction stir welding tool was fabricated from tool steel labeled as X12M (density ρ = 7800 kg/m3, specific heat Cp = 500 J/kg.oC and the thermal conductivity k = 40 W/m.oC). The tool heat treatment includes heating the metal to 1020°C for 30 min and then air cooling to room temperature, which gives a hardness of 58 HRC [9], its chemical composition is tabulated in Table 3 together with the standard tool material. 2.3. Selecting the Optimum FSW Process Parameters To obtain high quality of friction stir welded joints with high mechanical properties, i.e., high welding efficiency, the main welding parameters (rotational speed and welding speed) must be selected to determine the effect of each parameter on the mechanical properties. The rotational and welding speeds were chosen, therefore the tool rotational speed (725 - 1235 rpm) and the welding speed (6-34 mm/min) were used (see Table 4). The FSW tests were carried out on a vertical milling machine with a square butt joint configuration. All the welds were produced perpendicularly to the rolling direction for aluminum alloys. 2.4. Welding Procedure A plate was fixed at a predetermined location on the backing plate (which was a 280*280*25 mm steel plate) and clamped into place. This same location was used for all plates in this study. The tool was then positioned directly over the plunge location and the pin was brought into contact with the top surface of the workpiece, see Figure 3. Each tool plunges slowly between the two sheets that are required to be welded until the Samir Ali Amin Al-Rubaie Al-Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. 51- 64 (2015) 53 shoulder of the tool touches the sheet surface. The tool was then allowed to dwell for 30-40 sec to allow the shoulder to preheat the workpiece during welding. After the dwell, the tool began to traverse along the welding line with the selected tool. When a full weld has been made, the pilot hole will be welded over and the pin was parked above the weld. When the tool was parked, it was dragged, and a park hole was left, as shown in Figure 4. 2.5. Mechanical Tests 2.5.1. Tensile Test Tensile test was carried out on samples taken in a perpendicular direction to the welding to determine the tensile properties of the welding joints, see Figure 5. The shape and dimensions of the transverse tensile specimens according to ASTM E 8M [10] are shown in Figure 6. All tensile tests were carried out at room temperature and constant loading rate (5 mm/min) by a computerized universal testing machine (Hydraulic Tinius Olsen), which has a maximum capacity of (1000 kN). Then, the average of three specimens was taken to evaluate the tensile behavior of each welded joint. 2.5.2. Bending Test Three point bending test was carried out to determine the maximum bending force of the welded joints. Bending tests were conducted with former diameter equal to 30 mm. The shape and dimensions of the transverse bending specimens are (15.24 mm * 38.1 mm) according to ASTM E 190 [11]. The bending test was carried out at room temperature by a universal testing machine (Hydraulic LARYEE testing machine). 3. Response Surface Methodology (RSM) Response surface methodology (RSM) is a collection of mathematical and statistical techniques that are used for empirical model building and analysis of problems, in which a response of interest is influenced by several variables, and the objective is to optimize this response [12]. It has been extensively used in different engineering applications and fields. RSM is important in designing, formulating, developing, and analyzing new scientific studying and products. It is also efficient in the improvement of existing studies and products. The application of RSM to design optimization is aimed at reducing the cost of expensive analysis methods (e.g., finite element method or CFD analysis) and their associated numerical noise. By careful design of experiments, the objective is to optimize a response (output variable) which is influenced by several independent variables (input variables). An experiment is a series of tests, called runs, in which changes are made in the input variables in order to identify the reasons for changes in the output response. The advantages of design of experiments, as reviewed by Aggarwal and Singh [13], are as follows: (1) Numbers of trials are reduced. (2) Optimum values of parameters can be determined. (3) Assessment of experimental error can be made. (4) Qualitative estimation of parameters can be made. (5) Inference regarding the effect of parameters on the characteristics of the process can be made. The efficiency of RSM is significantly influenced by selecting the proper choice of experimental designs. The central composite design (CCD) is one of the most popular class of designs to fit response surfaces, building the second order (quadratic) regression model to predict the responses. Also, the most important characteristics of CCD is the spherical or rotatability property, i.e., the variance of predicted responses is the same at all points that are the same distance from the design center. Therefore, in the present research, RSM was utilized using CCD to establish predicted models for some responses (mechanical properties) as functions of input factors (welding speed and rotational speed) during FSW process of 2024- T351 Aluminum alloy using the optimum tool design. Moreover, numerical optimization was used to optimize the input parameters to obtain maximum responses. 4. Experimental Design Matrix The input parameters used in the whole experimentation were selected according to the practical experience and the limitations of the experimental measurements. These factors are given in Table 4 with two levels. The experimental design used was the response surface methodology using a central composite rotatable design for 2² factors, with 5 central points and α = ±1.414. 13 runs were performed according to the experimental design matrix (5 Samir Ali Amin Al-Rubaie Al-Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. 51- 64 (2015) 54 center points). Each parameter was used at different code levels of −1.414, −1, 0, +1, and +1.414, whereby each level used conformed to an actual value equivalent to the coded value. Thus, the input parameters studied are welding speed and rotational speed. The experimental design matrix used for input parameters in terms of actual factors with the experimental values of elongation, tensile strength and maximum bending load is given in Table 5. The software DESIGN EXPERT 8 was used to develop the model. Results of test runs are reported, as well as, the prediction models produced within a 95% confidence interval. 5. Results and Discussion 5.1. Tensile and Bending Test Results After carrying out the experiments, the welded joints were visually examined and the welds with good surface appearance were chosen and machined into the standard test specimens for the mechanical testing (according to ASTM E8M [10] for tensile test and ASTM E190 [11] for bending test). Tensile and bending tests were carried out as shown in Figures 7a, 7b, and 8. It should be noted that the testing values of tensile and bending for the base metal are (438 MPa) and (1520 N), respectively. 5.2. Modeling of the Elongation The average responses obtained for elongation, tensile strength and maximum bending load were used in calculating the models of the response surface per response using the least-squares method. For elongation prediction, a reduced quadratic model in coded terms was analyzed with backwards elimination of insignificant coefficients at an exit threshold of alpha = 0.1. Some coefficients were removed in order to obtain a formula with actual factors rather than coded ones. The term removed was AB. This means that the interaction of welding speed and rotational speed term had no significant effect on the elongation. Table 6 shows the statistical analysis of variance produced by the software for the remaining terms. The model is significant at 95% confidence. It is noted that the rotational speed (B), squared welding speed (A2), and the squared rotational speed (B2) terms are all significant, while the welding speed (A) term is not. The lack of fit test indicates a good model. This model illustrates that only the three terms (B, A2 and B2) have the highest impact on elongation. The final equation in terms of coded factors is : Elongation = +4.12 - 0.016 * A + 0.61 * B - 0.70 * A 2 - 1.25 * B2 …(1) And, the final equation in terms of actual factors is: Elongation = -39.12769 + 0.28011 * Welding speed + 0.079162 * Rotational speed - 7.04170E- 003 * Welding speed2 - 3.86475E-005 * Rotational speed2 …(2) Looking at the normal probability plot (Figure 9a) for the elongation data, the residuals generally that falling on a straight line implying errors, are normally distributed. Also, according to Figure 9b that depicts the residuals versus predicted responses for elongation data, it is seen that no obvious patterns or unusual structure, implying models are accurate. Figure 9c exhibits that the contour graph of welding speed and elongation as a response. It is seen that the increase in both welding speed and rotational speed leads to increase the elongation. The increase in welding speed led to increase in elongation. The joint exhibits poor elongation at a lower welding speed of 10 mm/min owing to the larger heat generation. As the welding speed increases from 10 to 20 mm/min, the negative effects of thermal cycles on joint properties are weakened, leading to an improvement in elongation. Between 20 and 30 mm/min, the elongation of the joints show a decrease with increasing welding speed due to the occurrence of void defect. Also, the increase in rotational speed led to increase in elongation. It can be seen that the elongation of the joint increases with increasing the rotational speed from 800 rpm to 980 rpm. The reason for this can be explained as follows: increasing the rotational speed would extend the shoulder dominated zone over the plate thickness. Since the material in this shoulder dominated zone is softer and easier to be stirred, extending this zone through the thickness enhances the stirring and consequently improves the elongation of the joints. However, when the rotational speed increases up to 1160 rpm, the elongation of the joint decreases because of the formation of void defect. Figure 10 manifests the predicted actual elongation data versus the actual ones for Samir Ali Amin Al-Rubaie Al-Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. 51- 64 (2015) 55 comparison reason. While Figure 11 shows the 3D graph of elongation as a function of welding speed and rotational speed. It can be noted that the increase of welding speed resulted in a slight increase in the elongation value, while the increase of rotational speed caused a higher increase in the elongation. This means that the rotational speed has the highest impact on the elongation value. In other words, the rotational speed is more significant than the welding speed in the elongation model. With increasing rotational speed for a fixed welding speed or increasing welding speed for a fixed rotational speed, the elongation of the joints firstly increased to a maximum value and then showed a decrease due to the occurrence of welding defects [6, 14]. 5.3. Modeling of the Tensile Strength Similarly, for tensile strength measurements, a reduced quadratic model in coded terms was analyzed with backwards elimination of insignificant coefficients at the exit threshold of alpha = 0.1. The term removed was AB for obtaining a formula with actual factors rather than coded ones. Table 7 reveals the statistical analysis of variance (ANOVA), and this model is significant at 95% confidence. The rotational speed (B), squared welding speed (A²) and the squared rotational speed (B²) terms are all significant, while the welding speed (A) term is not. This model indicates that these three terms have the highest impact on the tensile strength. The lack of fit test indicates a good model. The final equation in terms of coded factors is : Tensile strength = +230.40 + 3.14 * A -16.91 * B - 22.40 * A2 -27.35 * B2 …(3) The final equation in terms of actual factors is: Tensile strength = -584.03137 + 9.27514 * Welding speed + 1.56032 * Rotational speed - 0.22402 * Welding speed2 - 8.44003E-004 * Rotational speed2 …(4) The normal probability plot of residuals for tensile strength data shows that the residuals (errors) fall generally on a straight, and they are normally distributed. And, there are no obvious patterns or unusual structure, implying models are accurate. According to Figure 12 for the contour graph, it can be noticed that the increase of both welding speed and rotational speed generally increase the tensile strength value. Figure 13 depicts the predicted versus actual tensile strength data for comparison purpose. Whereas, Figure 14 reveals the 3D graph of tensile strength as a function of welding speed and rotational speed. It is seen that the increase in both welding speed and rotational speed leads to increase the tensile strength. The increase in welding speed led to increase in tensile strength. The joint exhibits poor tensile strength at a lower welding speed of 10 mm/min owing to the larger heat generation. As the welding speed increases from 10 to 20 mm/min, the negative effects of thermal cycles on joint properties are weakened, leading to an improvement in tensile strength. Between 20 and 30 mm/min, the tensile strength of the joints shows a decrease with increasing welding speed due to the occurrence of void defect. Also, the increase in rotational speed led to increase in tensile strength. It can be seen that the tensile strength of the joint increases with increasing the rotational speed from 800 rpm to 980 rpm. The reason for this can be explained as follows: increasing the rotational speed would extend the shoulder dominated zone over the plate thickness. Since the material in this shoulder dominated zone is softer and easier to be stirred, extending this zone through the thickness enhances the stirring and consequently improves the tensile strength of the joints. However, when the rotational speed increases up to 1160 rpm, the elongation of the joint decreases because of the formation of void defect. With increasing rotational speed for a fixed welding speed or increasing welding speed for a fixed rotational speed, the elongation of the joints firstly increased to a maximum value and then showed a decrease due to the occurrence of welding defects [6, 14]. 5.4. Modeling of the Maximum Bending Force Similarly, for maximum bending force measurements, a reduced quadratic model in coded terms was analyzed with backwards elimination of insignificant coefficients at the exit threshold of alpha = 0.1. The term removed was AB for obtaining a formula with actual factors rather than coded ones. Table 8 reveals the statistical analysis of variance (ANOVA), and this model is significant at 95% confidence. The welding speed(A), squared welding speed (A²) and the squared Samir Ali Amin Al-Rubaie Al-Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. 51- 64 (2015) 56 rotational speed (B²) terms are all significant, while the rotational speed (B) is not significant. This model illustrates that these three terms have the highest impact on the tensile strength. The lack of fit test indicates a good model. The final equation in terms of coded factors is : Maximum bending force = + 1374.98 + 61.68 * A - 20.39 * B - 320.89 * A2 -322.80 * B2 …(5) The final equation in terms of actual factors is: Maximum bending force = -9489.29985 + 134.52393 * Welding speed + 19.41404 * Rotational speed - 3.20890 * Welding speed2 - 9.96292E-003 * Rotational speed2 …(5.6) The normal probability plot of residuals for maximum bending force data shows that the residuals (errors) fall generally on a straight, and they are normally distributed. And, there are no obvious patterns or unusual structure, implying models are accurate. According to Figure 15 for the contour graph showing the interaction of welding speed and rotational speed, it can be noticed that the increase of both welding speed and rotational speed generally increase the maximum bending force value. Figure 16 depicts the predicted versus actual maximum bending force data for comparison purpose. Whereas, Figure 17 reveals the 3D graph of maximum bending force as a function of welding speed and rotational speed. It is seen that the increase in welding speed leads to increase the maximum bending force. As indicated with the elongation and tensile strength models, it can be concluded that the maximum bending force is also smaller at the lowest and highest welding speed and rotational speed, whereas it is larger at (20 mm/min) welding speed and (980 rpm) rotational speed. 5.5. Numerical Optimization of Elongation, Tensile Strength and Maximum Bending Force The numerical optimization was provided by the Design of Experiment software to find out the optimum combinations of parameters in order to fulfill the requirements as desired. Therefore, this software was used for this optimization, based on the data from the predictive models for three responses, elongation, tensile strength and maximum bending force, as a function of two factors: welding speed and rotational speed. Elongation, tensile strength and maximum bending force are modeled with a quadratic model. To develop the new predicted models, a new objective function, named Desirability which allows to properly combining all the goals, was evaluated. Desirability is an objective function, to be maximized through a numerical optimization, which ranges from zero to one at the goal. Adjusting its weight or importance may alter the characteristics of a goal, and the aim of the optimization is to find a good set of conditions that will meet all the goals. Usually, the weights are used to establish an evaluation of the goal’s 3D importance when maximizing desirability function; in this work, weights are not changed since the three responses (elongation, tensile strength and maximum bending force) have the same importance and are not in conflict within each other. The ultimate goal of this optimization was to obtain the maximum response that simultaneously satisfied all the variable properties. Table 9 lists the constrains of each variable for numerical optimization of the elongation, tensile strength and maximum bending force. According to this table, one possible run fulfilled these specified constrains to obtain the optimum values for elongation, tensile strength and maximum bending force, as given in Table 10. It can be seen that this run gave desirability of 0.868. Figure 18 shows the surface plot for desiability as a function of welding speed and rotational speed. Figures 19,20 and 21 depict the optimum values of the elongation, tensile strength and maximum bending force, respectively. Thus, it can be concluded from these figures that the desirability reaches the maximum value of 0.868 when the optimum value of elongation is 4.1 % (Figure 19), the optimum value of tensile strength is 230.773 MPa (Figure 20) and the optimum value of maximum bending force is 1377.83 N, as shown in (Figure 21). 5.6. Comparison of Predicted Results with Experimental Ones A comparison between the actual (with the optimum tool design) and predicted for elongation, tensile strength, and maximum bending force is shown in Table 11. This table also exhibits the percentage error between the actual and predicted results of elongation, tensile strength, and maximum bending force. In addition, according to this table, it is shown a very Samir Ali Amin Al-Rubaie Al-Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. 51- 64 (2015) 57 good agreement between the actual results of the elongation, tensile strength, and maximum bending force and the predicted results obtained by DOE and RSM technique. 6. Conclusions The following conclusions have been made from the present investigation: • Increasing both welding speed and rotational speed leads to increase the elongation and tensile strength up to 20 mm/min welding speed and 980 rpm rotational speed. The increase of welding speed resulted in a slight increase in the elongation and tensile strength, while the increase of rotational speed caused a higher increase in these two properties. This means that the rotational speed has the highest impact than welding speed on the elongation and tensile strength. • Increasing both welding speed and rotational speed leads to increase the maximum bending force up to 20 mm/min welding speed and 980 rpm rotational speed. However, the welding speed was found more significant than rotational speed. • Quadratic predicted models for elongation, tensile strength, and maximum bending force were obtained in terms of welding speed and rotational speed with 95% confidence level. • Among the 13 experiments of experimental work, 20 mm/min welding speed and 980 rpm rotational speed gives better tensile strength (245 MPa), elongation (4.7), maximum bending force (1450 N). • From the numerical optimization, the optimum values of elongation, tensile strength, and maximum bending force were found to be (4.1%), (230.733MPa) and (1377.83 N), respectively with a desirability 0.868 at (21 mm/min) welding speed and (977 rpm) rotational speed. Table 1, Standard and experimental chemical compositions of aluminum alloy AA2024-T351 (wt%). wt% Material Si Fe Cu Mn Mg Cr Zn Standard [8] ≤0.500 ≤0.500 3.800-4.900 0.300-0.900 1.200-1.800 ≤0.100 ≤0.250 Experimental 0.121 0.265 3.800 0.511 1.370 0.009 0.134 Table 2, Standard and experimental mechanical properties of aluminum alloy AA2024-T351. Property Material бy (Mpa) бu (Mpa) EL. (%) Standard [8] ≥290 ≥435 ≥15 Experimental 327 438 17.3 Table 3, Chemical composition of tool steel X12M [8]. Element C Si Mn P S Cr Cu Ni V Standard [9] 1.800 - 2.400 ≤0.400 ≤0.600 ≤0.030 ≤0.030 12.000 - 15.000 ≤0.250 ≤0.500 ≤0.300 Actual 1.870 0.278 0.270 0.009 0.001 12.440 0.079 0.200 0.023 Table 4, Levels of Input Parameters Used with Respective Coding. Factor Unit Low Level (-1) High Level (+1) -alpha (-1.414) +alpha (+1.414) Welding Speed mm/min 10 30 6 34 Rotational Speed rpm 800 1160 725 1235 Samir Ali Amin Al-Rubaie Al-Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. 51- 64 (2015) 58 Table 5 Experimental design matrix for both actual input factors and responses. Standard No. Run No. Type of Point Welding Speed (mm/min) Rotational Speed (rpm) Elongation (%) Tensile Strength (MPa) Maximum Bending force (N) 1 3 Factorial 10 800 1.1 206 673 2 13 Factorial 30 800 1.5 206 838 3 6 Factorial 10 1160 2.6 161 670 4 9 Factorial 30 1160 2.5 172 787 5 1 Axial 6 980 3.1 175 775 6 10 Axial 34 980 2.8 185 790 7 2 Axial 20 725 1.0 190 550 8 4 Axial 20 1235 2.7 150 715 9 7 Center 20 980 3.5 218 1330 10 8 Center 20 980 4.7 245 1450 11 5 Center 20 980 4.2 217 1350 12 11 Center 20 980 3.7 231 1316 13 12 Center 20 980 4.5 241 1435 Table 6 ANOVA Analysis for Response Surface Quadratic Model (Elongation, %). Source Sum of squares df Mean square F value p-value Prob > F Model 16.01 4 4.00 19.62 0.0003 significant A- Welding speed 1.930E-003 1 1.930E-003 9.464E-003 0.9249 B- Rotational speed 3.02 1 3.02 14.82 0.0049 A2 3.45 1 3.45 16.91 0.0034 B2 10.94 1 10.94 53.62 < 0.0001 Residual 1.63 8 0.20 Lack of Fit 0.58 4 0.15 0.56 0.7077 not significant Pure Error 1.05 4 0.26 Cor Total 17.64 12 17.64 Std. Dev. 0.45 R-Squared 0.9075 Mean 2.92 Adj R-Squared 0.8612 C.V.% 15.49 Pred R-Squared 0.7552 Press 4.32 Adeq Precision 12.044 Table 7 ANOVA Analysis for Response Surface Quadratic Model (Tensile strength, MPa). Source Sum of squares df Mean square F value p-value Prob > F Model 10100.35 4 2525.09 20.08 0.0003 significant A- Welding speed 79.02 1 79.02 0.63 0.4508 B- Rotational speed 2288.39 1 2288.39 18.20 0.0027 A2 3490.78 1 3490.78 27.76 0.0008 B2 5215.56 1 5215.56 41.48 < 0.0002 Residual 1005.96 8 125.74 Lack of Fit 346.76 4 86.69 0.53 0.7255 not significant Pure Error 659.20 4 164.80 Cor Total 11106.31 12 Std. Dev. 11.21 R-Squared 0.9094 Mean 199.77 Adj R-Squared 0.8641 C.V.% 5.61 Pred R-Squared 0.7626 Press 2636.86 Adeq Precision 11.335 Samir Ali Amin Al-Rubaie Al-Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. 51- 64 (2015) 59 Table 8 ANOVA Analysis for Response Surface Quadratic Model (Maximum bending force, N). Source Sum of squares df Mean square F value p-value Prob > F Model 1.310E+006 4 3.275E+005 110.86 < 0.0001 significant A- Welding speed 30433.08 1 30433.08 10.30 0.0124 B- Rotational speed 3329.39 1 3329.39 1.13 0.3194 A2 7.162E+005 1 7.162E+005 242.47 < 0.0001 B2 7.268E+005 1 7.268E+005 246.03 < 0.0001 Residual 23631.56 8 2953.95 Lack of Fit 6931.56 4 1732.89 0.42 0.7924 not significant Pure Error 16700.00 4 4175.00 Cor Total 1.334E+006 12 Std. Dev. 54.35 R-Squared 0.9823 Mean 978.69 Adj R-Squared 0.9734 C.V.% 5.55 Pred R-Squared 0.9555 Press 59333.04 Adeq Precision 21.628 Table 9 Constrains of each varaible for numerical optimization of Elongation, Tensile strength and Maximum bending force. Types of variables Goal Lower Limit Upper Limit Lower Weight Upper Weight Importance A: Welding speed Is in range 10 30 1 1 3 B: Rotational speed Is in range 800 1160 1 1 3 Elongation maximize 1 4.7 1 1 3 Tensile Strength maximize 150 245 1 1 3 Maximum bending force maximize 600 1450 1 1 3 Table 10 Optimal conditions used to obtain the maximum Elongation, Tensile strength and Maximum bending force. No. Welding Speed (mm/min) Rotational Speed (rpm) Elongation (%) Tensile Strength (MPa) Maximum Bending force (N) Desirability 1 21 977 4.1 230.733 1377.83 0.868 selected Table 11 Comparison between the actual and predicted responses. Welding Speed (mm/min) Rotational Speed (rpm) Elongation (%) Tensile Strength (MPa) Maximum Bending force (N) Actual 20 980 4.7 245 1450 Predicted 21 977 4.1 230.733 1377.83 % Error - - 12.77 5.82 4.98 Table 12 Conversions for each of the units used in the present work. Unit with SI Unit with U.S. customary (nonmetric) 1 MPa 145.038 psi 1 m 3.2808 ft 1 mm/min 0.05468 ft/s 1 N 0.2248 Ibf Samir Ali Amin Al-Rubaie Fig. 1. Schematic diagram of FSW Fig. 2. Design drawing for FSW tool (square pin with flat surface shoulder). Fig. 3. Welding backing plate and fixtures in use for FSW. Fig. 4. Welded sample explained welding start, and end of welded joint. End of Welding Line Plunging of Tool Welding Direction Al-Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. 60 Schematic diagram of FSW. Design drawing for FSW tool (square pin Welding backing plate and fixtures in use for Fig. 4. Welded sample explained welding direction, Fig. 5. Tensile and bending test specimens locations (all dimensions in mm). Fig. 6. Tensile test specimen (all dimensions in mm) ASTM (E 8M). Fig. 7a. Tensile test specimens before testing Fig. 7b. Tensile test specimens after testing End of Welding Line Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. 51- 64 (2015) Tensile and bending test specimens locations Tensile test specimen (all dimensions in mm) Fig. 7a. Tensile test specimens before testing. 7b. Tensile test specimens after testing. Samir Ali Amin Al-Rubaie Al-Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. 51- 64 (2015) 61 Fig. 8. Bending test specimen after testing. Fig. 9. Modeling of elongation property: (a) Normal probability plot for Elongation (%) data, (b) Residual versus predicted responses Elongation (%) data, and (c) Contour graph of Elongation (%) as a function of welding speed and rotational speed. Fig. 10. Predicted versus actual Elongation (%) data for comparison. Fig. 11. 3D graph of Elongation (%) as a function of welding speed and rotational speed. Fig. 12. Contour graph of Tensile strength as a function of welding speed and rotational speed. Design-Expert® Software Elongation Color points by value of Elongation: 4.7 1 Internally Studentized Residuals N or m al % P ro ba bi lit y Normal Plot of Residuals -2.00 -1.00 0.00 1.00 2.00 1 5 10 20 30 50 70 80 90 95 99 Design-Expert® Software Elongation Color points by value of Elongation: 4.7 1 Predicted In te rn al ly S tu de nt iz ed R es id ua ls Residuals vs. Predicted -3.00 -2.00 -1.00 0.00 1.00 2.00 3.00 0.00 1.00 2.00 3.00 4.00 5.00 Design-Expert® Software Factor Coding: Actual Elongation Design Points 4.7 1 X1 = A: Welding speed X2 = B: Rotational speed 10 15 20 25 30 800 890 980 1070 1160 Elongation A: Welding speed B : R ot at io na l s pe ed 2 22.25 2.252.5 2.75 3 33 3.25 3.253.25 3.5 3.75 4 5 Design-Expert® Software Elongation Color points by value of Elongation: 4.7 1 Actual P re di ct ed Predicted vs. Actual 0.00 1.00 2.00 3.00 4.00 5.00 1.00 2.00 3.00 4.00 5.00 Design-Expert® Software Factor Coding: Actual Elongation Design points above predicted value Design points below predicted value 4.7 1 X1 = A: Welding speed X2 = B: Rotational speed 800 890 980 1070 1160 10 15 20 25 30 1 2 3 4 5 E lo ng at io n A: Welding speed B: Rotational speed Design-Expert® Software Factor Coding: Actual Tensile strength Design Points 245 150 X1 = A: Welding speed X2 = B: Rotational speed 10 15 20 25 30 800 890 980 1070 1160 Tensile strength A: Welding speed B : R ot at io na l s pe ed 175 175180 180 185 185190 195 200 200 205 205 205 210 210 215 220 225 230 5 (a) (c) (b) Samir Ali Amin Al-Rubaie Al-Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. 51- 64 (2015) 62 Fig. 13. Predicted versus actual Tensile strength data for comparison. Fig. 14. 3D graph of Tensile strength as a function of welding speed and rotational speed. Fig. 15. Contour graph of maximum bending force as a function of welding speed and rotational speed. Fig. 16. Predicted versus actual maximum bending force data for comparison. Fig. 17. 3D graph of maximum bending force as a function of welding speed and rotational speed. Fig. 18. 3D surface plot for desirability as a function of welding speed and rotational speed. Fig. 19. The optimum value of Elongation. Fig. 20. The optimum value of Tensile strength. Design-Expert® Software Tensile strength Color points by value of Tensile strength: 245 150 Actual P re di ct ed Predicted vs. Actual 140.00 160.00 180.00 200.00 220.00 240.00 260.00 140.00 160.00 180.00 200.00 220.00 240.00 260.00 Design-Expert® Software Factor Coding: Actual Tensile strength Design points above predicted value Design points below predicted value 245 150 X1 = A: Welding speed X2 = B: Rotational speed 800 890 980 1070 1160 10 15 20 25 30 160 180 200 220 240 260 T en si le s tr en gt h A: Welding speed B: Rotational speed Design-Expert® Software Factor Coding: Actual Maximum bending force Design Points 1450 600 X1 = A: Welding speed X2 = B: Rotational speed 10 15 20 25 30 800 890 980 1070 1160 Maximum bending force A: Welding speed B : R o ta tio na l s pe ed 800 800 900 900 900 900 1000 1000 1000 1100 1200 1300 5 Design-Expert® Software Maximum bending force Color points by value of Maximum bending force: 1450 600 Actual P re di ct ed Predicted vs. Actual 600.00 800.00 1000.00 1200.00 1400.00 1600.00 600.00 800.00 1000.00 1200.00 1400.00 1600.00 Design-Expert® Software Factor Coding: Actual Maximum bending force Design points above predicted value Design points below predicted value 1450 600 X1 = A: Welding speed X2 = B: Rotational speed 800 890 980 1070 1160 10 15 20 25 30 600 800 1000 1200 1400 1600 M ax im um b en di ng fo rc e A: Welding speed B: Rotational speed Design-Expert® Software Factor Coding: Actual Desirability 1.000 0.000 X1 = A: Welding speed X2 = B: Rotational speed 800 890 980 1070 1160 10 15 20 25 30 0.000 0.200 0.400 0.600 0.800 1.000 D es ira bi lit y A: Welding speed B: Rotational speed 0.8680.868 Design-Expert® Software Factor Coding: Actual Elongation Design Points 4.7 1.0 X1 = A: Welding speed X2 = B: Rotational speed 10 15 20 25 30 800 890 980 1070 1160 Elongation A: Welding speed B : R ot at io na l s pe ed 2.0 2.02.5 3.0 3.03.0 3.5 4.0 5 Prediction 4.1 Design-Expert® Software Factor Coding: Actual Tensile strength Design Points 245 150 X1 = A: Welding speed X2 = B: Rotational speed 10 15 20 25 30 800 890 980 1070 1160 Tensile strength A: Welding speed B : R ot at io na l s pe ed 175 175 185 185195 205 205 205 215 225 5 Prediction 230.773 Samir Ali Amin Al-Rubaie Al-Khwarizmi Engineering Journal, Vol. 11, No. 1, P.P. 51- 64 (2015) 63 Fig. 21. The optimum value of Maximum bending force. Notation A welding speed B rotational speed Cp specific heat k thermal conductivity Greek Letters ρ density α star point 7. References [1] Rajif S. Mishra and Murray W. Mahoney, “Friction Stir Welding and Processing”, Wiley Inc., 2007. [2] R. Palanivel, P. Mathews, N. Muragan, and I. Dinaharan, “Effect of Tool Rotational Speed and Pin Profile on Microstructure and Tensile Strength of Dissimilar Friction Stir Welded AA5083-H111 and AA6351-T6 Aluminum Alloys”, Journal of Materials and Design, Vol. 40, pp. 7-16, 2012. [3] I. Rani and R. Marpu, “The Effect of Variation of Tool Geometry on Friction Stir Welded Aluminum Alloys- An Experimental Investigation”, International Journal of Mechanical Engineering and Robotics Research, Vol. 1, No. 1, pp. 91-98, 2012. [4] H. Mohanty, D. Venkateswarlu, M. Mahapatra, P. Kumar and N. Mandal, “Modeling the Effects of Tool Probe Geometries and Process Parameters on Friction Stirred Aluminum Welds”, Journal of Mechanical Engineering and Automation, Vol. 2(4), pp. 74-79, 2012. [5] R. Narsu and I. Rani, “An Experimental Investigation of the Effect of Variation of Tool Geometry And Optimization of Process Parameters on Friction Stir Welded Aluminum Alloys”, International Journal of Research in Aeronautical and Mechanical Engineering, Vol. 1, Issue. 7, pp. 261-266, 2013. [6] M. Koilraj, V. Sundareswaran, S. Vijayan, and S. Rao, “Friction stir Welding of Dissimilar Aluminum Alloys AA2219 to AA5083- Optimization of Process Parameters Using Taguchi Technique”, Journal of Materials and Design, Vol. 42, pp. 1-7, 2012. [7] P. Prasanna, C. Penchalayya, and D. Rao, “Optimization and Validation of Process Parameters in Friction Stir Welding on AA6061 Aluminum Alloy Using Gray Relational Analysis”, International Journal of Engineering Research and Applications (IJERA), Vol. 3, Issue. 1, pp. 1471-1481, 2013. [8] Standard Specification for Aluminum and Aluminum Alloy ASTM Sheet and Plate, ASTM B209M. [9] Standard JIS G 4404, Alloy Tool Steel, Material Number 1.208, 1983. [10] Standard Test Method for Tension Testing of Metallic Materials, ASTM E8M, 1988. [11] Standard Method for Guided Bend Test for Ductility of Welds, ASTM E190, 1980. [12] Montgomery, D. C., “Design and Analysis of Experiments”, 5th Edition, John Wiley & Sons Inc., 2000. [13] A. Aggarwal and H. Singh, “Optimization of Machining Techniques - A Retrospective and Literature Review”, Sadhana, Vol. 30, Part 6, pp. 699-711, 2005. [14] Huijie Liu, Huijie Zhang, Qing Pan, and Lei Yu, “Effect of FSW Parameters on Microstructural Characteristics and Mechanical Properties of 2219-T6 Aluminum Alloy Joints”, International Journal of Material Forming, Vol. 5, pp. 235-241, 2012 Design-Expert® Software Factor Coding: Actual Maximum bending force Design Points 1450 600 X1 = A: Welding speed X2 = B: Rotational speed 10 15 20 25 30 800 890 980 1070 1160 Maximum bending force A: Welding speed B : R ot at io na l s pe ed 800 800 900 900 900 900 1000 1000 1000 1100 1200 1300 5 Prediction 1377.83 � ا�ر���� ��� �� ا����� ���� 1، ا���د�11ا���ارز�� ا������� ا�� ،2015( 51- 64( 64 ا������� *�ا�(� ��!م (AA2024-T351))��'� ��!�&ت ا���!م ا��$!��� �#"! � ا������م �+!,-.� ا���/ ا �� ***زھ� ا�(!ھ **3!�2 "!س (�� *ا� *��� أ��� ��� ��� ا���������� ***،**،*�� ا�'�&%� ا�$��#�#"�� / ! � ا� alrabiee2002@yahoo.com: ا�-*,� ا+��$*و�)* dr_qasim_uot@yahoo.com: ا�-*,� ا+��$*و�)** zuhairsadeed@gmail.com: ا�-*,� ا+��$*و�)*** �� ا��& �ة � -��ً، و ا�:ي , $80م �/�م ا��%�دن 7�ون ا�5$�ام ا+�4��ر او ا��#اد ا�����1ا�0/�م ,�F) ھ:ه ا��را��، D� ا�$/A05��7 . BC ا+@$��?) ھ# =���0 �/�م " .,�C �/�م ا�A05 ا+@$��?)ا��L&�%& *�M�D N&)AA2024-T351 (*O7 �&#/0ت ا�0/�م =J0 ا�5#اص ا���������� � -�GH ا+����#م D ،ت ا�0/�مL&�%�� �P0$5& ت�,��� � ا�5$�ام &PQ#7 ا�0/�م �=*� RS& ،ل�Uإد R&34-6(=#ا mm/min (* ورا��� و ا��) rpm 1235-725(=� ا� ���O$�+ا ��+�7 ،� �H*ا+�$'��7ت ا� J0= ھ�*�M�D J0= و �4/0#ل J�W%ء ا����، و !#ة ا+�/Yو&� ا��C& .D ت����! R�0/D ام و�� D� إ�5$! ��? ��-,*' �، و!#ةYو&� ا��C& ،���O$�L� ء���'�� �)DESIGN EXPERT 8 ( �7�'$�Lا�%J�W 7]�$%��ل Z&��*7 ا�$��4� ا�$'*,-( ا+�/��4/0#ل =J0 ���ذج & ��/O ا�)RSM .( ذج ا�*,�_) �^�$'��7ت#���، ا+�$D ��?)،���O� إ�$/�4ل و R�0/D ا�Yو&�ا��C& J�W%ءا���ف ا�0/�م ?�وال �`*و) و!#ةا+�/ N,�-D ا�� $%��0 ��7$%��لANOVA ذج����7�b��7 c 87,�دة �*=� ا�0/�م و ا� *=� ا��ورا���، 8D b��Fداد . ���HL& N& BC/$0 ھ:ه ا� �'D��ا����ذج ا��*7%� ا� � و ا+�$���O ا��C& J�W= ���! Jو&�Yا��$D �M �,ا��وث ا�%�#ب ا�P*اF��d) ا�-@ g- 7 h!� .L? ان ز,�دة ��? N& �=*� دي وا� *=� ا�0/�مjD ورا����ا� ��!h ا+�/��ء ا�J ز,�دة !#ة$D �M J�W= ���! Jاو+ً ا� J�W%ا&*. ا� N& N�, ���� و"� �7ن �*=� ا�0/�م �D#ن ا?�M�D *S*ا &N ا� *=�F، و&C ورا����و . ا� � ZH�$� N�7 ھ:ه ا����ذج و ا+&l& ��0S ا��$�ZH ا�%���0 �7 $#ى �" BFا#D � ).�CM)95% و"