II LAETA Young Researchers Meeting FEUP, Porto, 10-11 April 2012

FSW, OPTIMIZATION AND JOINT STRENGTH PREDICTION BASED ON TAGUCHI METHOD

Key words: Friction stir welding (FSW), Taguchi, mechanical properties, butt-joint Abstract. This paper is a summary of the work developed in the frame of a FCT project, FSW, optimization and joint strength prediction based on the Taguchi and artificial neural network methods, which objective is the application of the Taguchi technique design of experiments (DOE) and artificial neural networks (ANN) to optimize the Friction Stir Welding (FSW) technique parameters for butt and T joints application. For butt-joints optimization this work analyzes five different entry parameters, studying their interactions and influence on the mechanical properties. The entry parameters are: welding speed (WS); tool rotational speed (TRS); tilt angle (); probe distance from the root surface (S); and shoulder/probe ratio (D/d). The mechanical properties are: yield strength (yield); tensile strength (TS); elongation (EL); and the hardness in thermomechanically affected zone (TMAZ). A Taguchi array was chosen to define the best combinations of parameters to perform the FSW experiments. After welding 27 butt-joints, defined by the Taguchi array, it was observed that the interaction between tool rotational speed and welding speed, and the parameter probe/shoulder diameters ratio are the most influent in the joints mechanical properties.

E. R. Libnio, R. R. Rodrguez, M. A. V. de Figueiredo, P. M. G. P Moreira 1 INTRODUCTION AND STATE OF THE ART

During the past two decades, most government legislation, environmental pressures, safety regulations and consumer demand have brought concerns about the need for lighter automotive vehicles, aircraft, trains and ships that are more fuel-efficient, and have improved handling and crashworthiness. In this regard, application of light-weight materials, such as aluminum and magnesium alloys has attracted great attention to partly replace conventional steels for structural parts. Aluminum has been considered as an energy-saving structural material for advanced applications. Due to its recycling capabilities it can be considered an environmental friendly metallic material. Also, the concept of combining various sheet metals into a single welded blank has been developed to enable products to have optimal properties, . In ours days, Friction Stir Welding (FSW) is a promisingly ecologic weld method that allows lower material waste and avoids radiation and harmful gas emissions, usually associated with the fusion welding processes. This welding technique makes use of a nonconsumable welding tool to generate heat, by friction between it and the faying plates, and to induce strong plastic deformation of the workpiece material promoting its complex mixing across the joint, . FSW has become a technology of widespread interest because of its numerous advantages, most important of which are the ability to weld otherwise unweldable alloys, produces less distortion than fusion welding techniques, and also produces desirable microstructures in the weld and heat affected zone (HAZ), . The application of Taguchi quality engineering techniques is a guide and reference source for industry professionals (managers, engineers and scientists) involved in testing products and processes, . These techniques have been widely used for determination of parameters and their performance measured with minimum variation, and also applied for optimization in materials processing.

1.1 Friction Stir Welding FSW was invented in The Welding Institute, United Kingdom, in 1991 as a solid-state joining technique; it is one of the major advances in welding technology in recent years. It emerged as a commercial joining technique to weld high strength alloys that were difficult to join with the existing techniques. The process was initially developed for aluminum alloys, but since then FSW was found to be suitable for joining a large number of materials. It produces nearly defect-free welds for such demanding applications as aerospace structures, at lower cost than conventional fusion welds. A potential field of application is aircraft structures, where cost and weight can be reduced by using new joining techniques instead of riveting, .

1.1.1 Process A non-consumable rotating tool with a specially designed probe and shoulder is inserted into the abutting edges of the two parts to be joined and traversed along the line of joint. The rotating tool is forced down into the joint line under conditions where the frictional 2

E. R. Libnio, R. R. Rodrguez, M. A. V. de Figueiredo, P. M. G. P Moreira heating is sufficient to raise the temperature of the material to the range where it is plastically deformed, Figure 1.

Figure 1 - Schematic diagram of the FSW process, .

The probe is typically slightly shorter than the thickness of the workpiece and its diameter is approximately the thickness of the workpiece. As the tool moves along the joint between the two components, the material is forced to flow from the leading edge to the trailing edge of the tool, and the material that flows around the tool undergoes extreme levels of plastic deformation. The shape of the tool promotes high hydrostatic pressure along the joint line, causing consolidation of the material plasticized due to heat generation, . The workpiece has to be firmly clamped onto a backing bar and secured against vertical, longitudinal and lateral forces, which will try to lift them and push them apart. It is established that a gap of up to 10% of the sheet thickness can be tolerated. The probe height is generally slightly smaller than the workpiece thickness, and its vertical penetration into the workpiece is halted when the shoulder makes contact with the workpiece surface. A downward forcing pressure from the shoulder helps to prevent the expulsion of softened material, in addition to providing supplementary frictional heating, . FSW joints usually consist of four different regions, as shown in Figure 2. They are (A) unaffected base metal; (B) heat affected zone (HAZ); (C) thermo-mechanically affected zone (TMAZ); and (D) friction stir processed (FSP) zone, . The formation of the above

E. R. Libnio, R. R. Rodrguez, M. A. V. de Figueiredo, P. M. G. P Moreira regions is affected by the material flow behavior under the action of rotating nonconsumable tool. However, the material flow behavior is predominantly influenced by the FSW tool profiles, FSW tool dimensions, and FSW process parameters, .

Figure 2 - Different regions of FSW joints, .

1.1.2 Parameters The nature of the FSW process is very complex and has not been studied in its entire dimension. The relationships between the fundamental control parameters and the measured process variables have never been completely explored. An Ishikawa diagram, cause and effect diagram, was constructed, as shown in Figure 3, to identify the friction stir welding process parameters that may influence the quality of joints.
Tool design Materi al Shoulder geometry Clamp design Clamp force

Tilt angle Probe geometry

Shoulder /probe diameter ratio Probe length Mechanical Properties Physical Properties Materi al

Clamp geometry Thickne ss Welding speed Chemical Composition Axial force Machine

Quality of FSW joints

Toll rotational speed

Figure 3 - Cause and effect diagram.

The tool geometry, welding parameters, and joint designs are the parameters that have significant effect on the material flow pattern and temperature distribution, thereby influencing the microstructural evolution of material. The detailed list of FSW process parameters is presented below, . 4

E. R. Libnio, R. R. Rodrguez, M. A. V. de Figueiredo, P. M. G. P Moreira Frictions stir welding process most significant parameters: i. ii. iii. iv. Tool rotational speed (TRS) [rpm] Welding speed (WS) [mm/min] Axial force (AF) [kN] Tool geometry, Figuere 4. i. D/d ratio of tool, RT ii. Probe length, L [mm] iii. Tool shoulder diameter, D [mm] iv. Probe diameter, d [mm] v. Tilt angle []

D: Shoulder diameter d: Probe diameter L: Probe length

Figure 4 - Tool geometry.

From a practical point of view, it is important to note that during numerical optimization the FSW parameter ranges should not choose at extreme conditions (e.g., 50% of the nominal/working condition), as they may correspond to failure modes in the actual weld, as observed in the literature.

1.2 Taguchi method Taguchi method is one of the most innovative methods for quality control. It was developed by Professor Genuchi Taguchi during decade of 50s. This method has two deployment areas, online and offline. Both areas are very cost sensitive to the decisions that are made with respect to the activities. Offline quality control refers to the improvement in the product and process development stages. Online

E. R. Libnio, R. R. Rodrguez, M. A. V. de Figueiredo, P. M. G. P Moreira quality control refers to the monitoring of current manufacturing processes to verify the quality levels produced, . This method is based on the statistical analysis of data and offer a simple means of analysis and optimization of complex systems, . In the system design step, the engineer applies scientific and engineering knowledge to generate an elemental functional archetype design. To obtain high quality and low cost, the parameters design is the most important step in the Taguchi method. The objective of this step is to obtain the optimal settings of process parameters for the best performance output characteristics of the process, . For process parameter optimization design, the listed steps must be followed, : Step 1: Determine the quality characteristic to be optimized; Step 2: Identify the noise factors and test conditions; Step 3: Identify the control factors and their alternative levels; Step 4: Design the matrix experiment, orthogonal array (OA), and define the data analysis procedure; Step 5: Conduct the matrix experiment; Step 6: Analyze the data and determine optimum levels for control factors; Step 7: Predict the performance at these levels.

The Taguchi method proposes two ways for implements the complete analysis of obtained data to determine optimum levels for control factors and the influences of this in the process: analysis of variance (ANOVA) and signal to noise (S/N) ratio. 2 EXPERIMENTAL TRIALS

2.1 Design of L27 orthogonal array (OA) for the experiment To perform the experimental trials the columns 1, 2, 6, 7, and 8 were chosen from a typical L27 Taguchi orthogonal array to study the influence of the welding parameters initially defined. The influence of the interactions between columns 1 with 2, 1 with 8, and 2 with 8 can also be studied. The Taguchi columns chosen with the respective parameters levels ready for the experimental trials can be seen on Table 1.

E. R. Libnio, R. R. Rodrguez, M. A. V. de Figueiredo, P. M. G. P Moreira

Table 1 - L27 OA with parameters levels values for experimental trials

Test 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

A 735 735 735 735 735 735 735 735 735 1000 1000 1000 1000 1000 1000 1000 1000 1000 1500 1500 1500 1500 1500 1500 1500 1500 1500

B 216 216 216 290 290 290 360 360 360 216 216 216 290 290 290 360 360 360 216 216 216 290 290 290 360 360 360

C 0 1 2 0 1 2 0 1 2 1 2 0 1 2 0 1 2 0 2 0 1 2 0 1 2 0 1

D 0.10 0.15 0.20 0.10 0.15 0.20 0.10 0.15 0.20 0.20 0.10 0.15 0.20 0.10 0.15 0.20 0.10 0.15 0.15 0.20 0.10 0.15 0.20 0.10 0.15 0.20 0.10

E 2.00 2.50 3.00 2.50 3.00 2.00 3.00 2.00 2.50 2.00 2.50 3.00 2.50 3.00 2.00 3.00 2.00 2.50 2.00 2.50 3.00 2.50 3.00 2.00 3.00 2.00 2.50

Parameter A B C D E

Description Tool rotational speed [rot/min] Welding speed [mm/min] Tilt angle [] Probe distance from the root surface [mm] Shoulder/probe ratio (D/d)

With those columns on it can be assured that the interactions between welding, rotational speed and shoulder/probe ratio are studied, which are believed to be the most influent parameters on welding process. 2.2 Experimental work The experimental trials were performed joining two AA6082-T6, 380x150x3mm, aluminium plates in butt-joint configuration. From the possible definable parameters, two were maintained constant: the probe profile and probe diameter (probe/shoulder ratio was defined by variation of the shoulder diameter). Five parameters and their interactions were defined by the Taguchi L27 array, with 3 levels for each parameter.

E. R. Libnio, R. R. Rodrguez, M. A. V. de Figueiredo, P. M. G. P Moreira After welding, mechanical tests were performed to determine which parameter or parameter interaction had more influence in the mechanical properties previously defined. Additional tests and measurements, to study the temperature along the welding seam or the deformation caused by the FSW process were made. To perform the welding tests three tools were created, being the shoulder diameter the only difference between them. These tools were designed based on previous studies and optimized to simplify the construction process, taking into account the available machinery at our disposal. The construction process and tests performed are shown in the next sections.

Figure 5 - Welding tool.

Figure 6 - Welding probe.

2.2.1 Tools configuration The tool design was performed based on designs and profiles from previous works and several articles, . It was configured to easy assembly in the milling machine and also capable of using different types of probes, including adjustment of the distance between the pin and the welding root. Three tools were made, the only difference between them was just the shoulder diameter 12mm, 15mm, and 18mm. The probe has 6mm diameter, a conical tip with 4mm diameter in the end and 2mm height, and was finalized with 3 equally separated oblique cuts with 1mm depth and 3mm height, to promote the material flow around the pin. 2.2.2 Welding tests To understand the temperature response, 4 thermocouples were mounted along the welding line, in holes with 0.8 mm diameter and 1.5 mm depth, in order to know the temperature the closest possible of the welding stir zone. After performing the 27 welds, several properties were analyzed, such as temperature along the welding stir, deformation caused by the weld, and the mechanical properties, yield and tensile strengths, elongation and hardness in the weld seam.

E. R. Libnio, R. R. Rodrguez, M. A. V. de Figueiredo, P. M. G. P Moreira 2.2.2.1 Parameters influence on temperature and deformation The temperature was measured on the shoulder boundaries during welding process, Figure 7. For temperature analysis, only the highest values were considered, and for the study of parameters influence in the temperature response the maximum temperature of the 4 points was used.

With the complete information about the temperature along the welding process, an
Figure 7 - Assembly in the milling machine and welding process.

analysis of variance was performed to study which parameters showed the highest influence promoting elevation of the temperature along the weld seam. From Figure 8 it is possible to observe that the shoulder/pin ratio and the interaction of the speeds with the shoulder/pin ratio have influence in the temperature behavior during the FSW process.

Figure 8 - Contribution of factors and interactions in temperature response.

Some preliminary conclusions: 9

E. R. Libnio, R. R. Rodrguez, M. A. V. de Figueiredo, P. M. G. P Moreira Taguchis robust orthogonal array design method is suitable to analyze the preliminaries temperature results of the experiment; The percentage of contribution of each parameters to the temperature developed during the friction stir welding process was evaluated. It was found that the shoulder/probe ratio has a contribution of 38.84%, the interaction between rotational speed and shoulder/probe ratio has a contribution of 17.38%, and the interaction between welding speed and shoulder/probe ratio has a contribution of 18.5%; It can be seen a trend to obtain higher temperatures associated with the increase of the shoulder/probe ratio, this trend is clear when only the parameter influence is studied. But even when this parameter is associated with the tool rotational speed or the welding speed the same conclusion can be found, especially for ratios higher than 2.5.

2.2.2.2 Parameters influence on mechanical properties, tensile and yield strengths, and elongation The AA6082-T6 base material mechanical properties, , in comparison with the highest values obtained on the tensile tests performed are shown in the Table 2.
Table 2 - Base material vs welded material tests

FSW Base material Tensile test results

yield [MPa] 276 166 BJ15

tensile [MPa] 323 246 BJ15

Elongatio n% 17.5 7.1 BJ1

The tensile tests were performed in accordance with ASTM E 8M 04 standard, . For each welding sample, 4 tensile test specimens were machined, those specimens, after testing, allowed to know the tensile strength, yield strength and elongation for the welded material. The tests were performed in a servo-hydraulic test machine with an extensometer to measure strain. As can be seen in Table 2, the sample that achieved the similar values to those of the base material was BJ15. In Figure 9 is presented the BJ15 sample stress/strain curve. This welded sample presents very good characteristics for a welded specimen, especially when compared to previous works with the same material, .

10

E. R. Libnio, R. R. Rodrguez, M. A. V. de Figueiredo, P. M. G. P Moreira

Figure 9 - Tensile test curve for sample BJ15.

In similarity with the study for the temperature response, an analysis of variance was performed to understand which are the most influent parameters or interactions on tensile strength, yield strength and elongation. However, it is important to note that all the mechanical properties values are relative to those found for the base AA6082-T6 aluminum alloy, referred in Table 2. The most influent parameters or parameters interactions are those that are represented in Table 3.
Table 3 - Parameters inside the confidence level interval (>95%) Mechanical property Parameter / interaction A B C D E A*B A*E B*E E A B D E A*B B*E Signification level (P) 0.0001 0.0001 0.0003 0.0001 0.0001 0.0001 0.0001 0.0001 0.0190 0.0180 0.0010 0.0040 0.0300 0.0000 0.0150 % Contribution 5.36 5.20 4.29 8.48 14.26 11.47 7.89 14.77 5.06 3.55 6.75 5.25 5.70 12.24 4.96 Residual error (% Contribuition)

Tensile strength

28.25

Yield strength

85.31

Elongation

58.76

It is possible to found trends, especially for the shoulder/probe ratio, parameter E. Besides that, the same parameter E was the most influent in the temperature response study, showing in all studies a trend to give the best values at a 2.5.

11

E. R. Libnio, R. R. Rodrguez, M. A. V. de Figueiredo, P. M. G. P Moreira 3 CONCLUSIONS

As preliminary result from the tests and analysis already completed it is possible to assume that the most influent characteristics for the FSW process are the ratio between the shoulder and the probe, and the interaction between the tool rotational speed and the welding speed. As future considerations, when it is not possible to control the downward force, the tilt angle should be removed from the entry parameters, and the values for each parameter level probably should be more spaced to avoid interference between the uncontrollable characteristics that constitute the residual error. As last consideration, while continuing the project and initializing the T-joint configuration and studies, a new Taguchi array should be selected, with just 4 entry parameters, and the chosen columns should permit a triangular configuration, thereby maximizing the relevance of parameters interactions. ACKNOWLEDGEMENTS This work was supported by the FCT project PTDC/EME-TME/114703/2009. Dr. Moreira acknowledges POPH QREN-Tipologia 4.2 Promotion of scientific employment funded by the ESF and MCTES. The collaboration of Mr. Jos Almeida during the welding trials is gratefully acknowledged.

12

E. R. Libnio, R. R. Rodrguez, M. A. V. de Figueiredo, P. M. G. P Moreira REFERENCES 1. Kim, D., Badarinarayan, Harsha., Kim, Ji Hoon., Kim, Chongmin., Okamoto, Kazutaka., Wagoner, R. H. and Chung, Kwansoo, Numerical simulation of friction stir butt welding process for AA5083-H18 sheets. European Journal of Mechanics - A/Solids, 2009. 29(2): p. 204-215. Yousif. Y. K., D.K.M.a.K.B.I., Prediction of Friction Stir Welding Characteristic Using Neural Network. Jordan Journal of Mechanical and Industrial Engineering, 2008. Volume 2(Number 3): p. 151 - 155. Rodrigues, D.M., Loureiro, A., Leitao, C., Leal, R. M., Chaparro, B. M. and Vilaa, P., Influence of friction stir welding parameters on the microstructural and mechanical properties of AA 6016-T4 thin welds. Materials & Design, 2009. 30(6): p. 1913-1921. Ross, P.J., ed. Aplicaes das tcnicas Taguchi na engenharia da qualidade. MAKRON Books. Vol. 1. 1991, Makron, McGraw-Hill: So Pablo. Guimares Pires Moreira, P.M., Lightweight stiffened panels: mechanical characterization of emerging fabrication technologies, in Departamento de Engenharia Mecnica e Gesto Industrial,Faculdade de Engenharia. 2008, Universidade do Porto: Porto. ESAB AB, W.A., Friction Stir Welding Technical Handbook. 2010. Vijayan, S., Raju, R., Subbaiah, K., Sridhar, N. and Rao, S. R. K., Friction stir welding of almg alloy optimization of process parameters using Taguchi method. Experimental Techniques, 2010. 34(5): p. 37-44. K. Hans Raj., R.S.S.a.P.S., Experimental Studies of Friction Stir Welding Process. International Journal of Engineering Studies, 2010. Volume 2 (Number 3): p. 279-288. Mohamadreza, N., Abbas, S. Milani and Spiro Yannacopoulos, Taguchi Optimization of Process Parameters in Friction Stir Welding of 6061 Aluminum Alloy: A Review and Case Study. Engineering Fracture Mechanics, 2011. Vol.3( No.2): p. 144-155. Jayaraman, M., Sivasubramanian, R., Balasubramanian, V. and Lakshminarayanan, A. K. , Optimization of process parameters for friction stir welding of cast aluminium alloy A319 by Taguchi method. Journal of Scientific & Industrial Research, 2009. 68: p. 36-43. Lakshminarayanan, A.K.a.B., V., Process parameters optimization for friction stir welding of RDE-40 aluminium alloy using Taguchi technique. Transactions of Nonferrous Metals Society of China, 2008. 18(3): p. 548-554. Macleod, C., Dror, Geva. and Maxwell, Grant, Training Artificial Neural Networks Using Taguchi Methods. Artificial Intelligence Review, 1999. 13(3): p. 177-184. Davis, A.M., Interaction of the friction stir welding tool and work-piece as influenced by process parameters [M.S. dissertation]. Dissertations & Theses: Full Text [database on the Internet], 2010. Arora, K.P., Sunil. Schaper, Michael. and Kumar, Rajneesh, Effect of process parameters on friction stir welding of aluminum alloy 2219-T87. The International Journal of Advanced Manufacturing Technology, 2010. 50(9): p. 941-952. Hattingh, D.G., Blignault, C., Van Niekerk, T. I. and James, M. N., Characterization of the influences of FSW tool geometry on welding forces and weld tensile strength using an instrumented tool. Journal of Materials Processing Technology, 2008. 203(1-3): p. 46-57. 13

2. 3.

4. 5.

6. 7. 8. 9.

10.

11. 12. 13. 14. 15.

E. R. Libnio, R. R. Rodrguez, M. A. V. de Figueiredo, P. M. G. P Moreira 16. 17. Elangovan, K.a.B., V., Influences of pin profile and rotational speed of the tool on the formation of friction stir processing zone in AA2219 aluminium alloy. Materials Science and Engineering: A, 2007. 459(1-2): p. 7-18. Elangovan, K., V. Balasubramanian, and M. Valliappan, Influences of tool pin profile and axial force on the formation of friction stir processing zone in AA6061 aluminium alloy. The International Journal of Advanced Manufacturing Technology, 2008. 38(3): p. 285-295. Elangovan, K., lasubramanian, V. and Balasubramanian, V., Influences of tool pin profile and tool shoulder diameter on the formation of friction stir processing zone in AA6061 aluminium alloy. Materials & Design, 2008. 29(2): p. 362-373. Elangovan, K.a.B., V., Influences of tool pin profile and welding speed on the formation of friction stir processing zone in AA2219 aluminium alloy. Journal of Materials Processing Technology, 2008. 200(1-3): p. 163-175. Materials, A.S.f.T.a., Standard Test Methods for Tension Testing of Metallic Materials [Metric]. 2004, ASTM International. Moreira, P.M.G.P., Santos, T., Tavares, S. M. O., Richter-Trummer, V., Vilaa, P. and de Castro, P. M. S. T., Mechanical and metallurgical characterization of friction stir welding joints of AA6061-T6 with AA6082-T6. Materials & Design, 2009. 30(1): p. 180-187. De Giorgi, M., et al., Effect of shoulder geometry on residual stress and fatigue properties of AA6082 fsw joints. Journal of Mechanical Science and Technology, 2009. 23(1): p. 26-35. Cavaliere, P., Squillace, A. and Panella, F., Effect of welding parameters on mechanical and microstructural properties of AA6082 joints produced by friction stir welding. Journal of Materials Processing Technology, 2008. 200(1-3): p. 364372.

18. 19. 20. 21.

22. 23.

14