Hynes et al., International Journal of Advanced Engineering Technology E-ISSN 0976-3945 Research Paper MODELING OF PROCESS PARAMETERS OF FRICTION STUD USING FUZZY LOGIC SYSTEM N. Rajesh Jesudoss Hynes1, R.Kumar2, J. Angela Jennifa Sujana3

Address for Correspondence 1Associate Professor, Department of Mechanical Engineering, Mepco Schlenk Engineering College Sivakasi,Tamilnadu, India 626 005. 2 Research Scholar, Department of Mechanical Engineering, Mepco Schlenk Engineering College Sivakasi, Tamilnadu, India 626 005. 3Senior Assistant Professor, Department of Information Technology, Mepco Schlenk Engineering College Sivakasi, Tamilnadu, India 626 005. India ABSTRACT The purpose of this work is to study the influence of process parameters of friction stud welding on joining of AA 6063 and AISI 1030 . With the application of fuzzy logic, empirical models are developed for output response characteristics. Rotational speed, friction time and friction pressure are considered as the influential input process parameters. The values of Impact strength and axial shortening are predicted by fuzzy models and they are compared with the experimental data. The correlation coefficients between predicted and experimental values are 99.15% for impact strength and 98.05% for axial shortening length respectively. Moreover through response optimizer, to achieve impact strength of 206.46 KJ/m2 and axial shortening length 7.3798 mm optimal conditions are predicted as 1063 rpm rotational speed, 7 seconds friction time and 300 Kpa friction pressure. KEYWORDS Friction stud welding, Fuzzy logic, Multi-objective optimization. INTRODUCTION technique for optimizing the process parameters to Friction stud welding is a solid phase welding achieve the maximum tensile strength. Karupanan et performance involving a stud or appurtenance being al. (2014) have examined the process parameters of rotated at high speed while being forced against a of AISI 904L/ Austenitic Stainless substrate, producing heat by friction Ates et al. Steel. In that study, they conducted a set of (2007); Ellis (1972). The metal surfaces reach a experiments based on Taguchi L18 orthogonal array temperature as a result of which they flow plastically where parameters such as frictional power, burn-off under pressure, surface impurities are ejected and a length, upset power and rotational speed were forged weld is formed. The maximum temperatures involved. They used grey relational analysis and reached for the duration of welding are much lower genetic algorithm to identify the significant factor than the melting point of the metals. It is a novel and optimised level of input parameters on the variant of friction welding process. Various factors responses like fatigue life, width and welding time such as rotational speed, friction time, friction respectively. Ajith et al. (2015) studied the influence pressure, pressure etc., impact the quality of of process parameters like friction pressure, upsetting friction stud welded joints Ozdemir et al. (2007); pressure, rotational speed and burn-off length on the Mumin et al. (2008). Benefits of this welding process responses like tensile strength and hardness during are low production time, high reproducibility and low friction welding of UNS S32205 duplex stainless consumption of energy Fauzi et al. (2010); Ozdemir steel. They used artificial neural network and particle et al. (2005).The friction stud welding process is swarm optimization technique to predict the ideally suited for deep water naval applications, optimised parameters. They have developed short-term emergency repairs and for submarine regression model followed by response surface rescue Rajesh Jesudoss Hynes et al (2012 & 2014). methodology method and optimised the process Ocean engineers pursued commercial joining parameters using genetic algorithm. Manideep et al. applications for offshore platform repairs. (2012) optimised the process parameters of friction Additionally, Aluminum/ friction stud welding of AISI 321/ AISI 430 joints using Taguchi welded joints are unavoidable in pipes of tanks of L9 orthogonal array method. In that work they liquid propellants in Satellite Launch Vehicles and considered process parameters such as friction rocket fuel tanks Alves et al. (2010). Recently, there pressure, upset pressure and burn-off length (metal is a need of joining a steel ear thing pin, which is in loss). They analysed the metallurgical characteristics the shape of a stud to an aluminum car body Rajesh and micro hardness across the weld interface. Paulraj Jesudoss Hynes et al (2014 & 2015). et al. (2006) have analysed the effect of process Optimization of friction welding parameters is to be parameters of friction welding such as heating carried out in order to achieve best results. In the pressure, heating time, upsetting pressure, upsetting previous years some of the researchers had optimised time on tensile strength and metal loss using the process parameters of friction welding on multi- integrated Taguchi and ANOVA (Analysis of response characteristics. Paventhan et al. (2011 & variance) approach. The present work modelling the 2012) studied the influence of process parameters in process parameters of friction stud welding of friction welding of carbon steel/AISI 304 and AA dissimilar AA 6063/AISI 1030 joints by means of 6082/AISI 304 joints. They considered input process fuzzy logic Kovac et al. (2013) and optimized parameters as friction pressure, forging pressure, through response optimizer with the intention of friction time and forging time. They have maximizing the impact strength and minimizing axial implemented response surface methodology shortening length (metal loss). Table 1: Composition of AA 6063

Int J Adv Engg Tech/Vol. VII/Issue I/Jan.-March.,2016/413-417 Hynes et al., International Journal of Advanced Engineering Technology E-ISSN 0976-3945

Table 2: Composition of AISI 1030 steel

EXPERIMENTATION welding process is carried out in four stages. In the In this work, dissimilar metals AA 6063 and AISI first phase, axial load and rotational speed of the mild 1030 steel are joined by friction stud welding process steel rod are applied and the two rods are brought and the resulting mechanical properties are close to the joined surfaces. During the second stage, investigated. The chemical combinations of AA 6063 friction heating takes place at the interface the two and AISI 1030 steel are as shown in table1 & 2. rods. At the third phase, when the rotational speed Quality of any friction welded components depends was stopped and upsetting under increased pressure on the appropriate selection of the process occurs simultaneously. In the final phase, the parameters. Hence, the influence of the process pressure remains constant until the axial pressure is parameters such as friction time, friction pressure and released. According to orthogonal L9 array, the speed of rotation on the impact strength and metal experiment was carried out for different loss of the friction welded AA6063/AISI 1030 joints combinations of rotational speed, friction time and are studied in this present work. The friction stud friction pressure as shown in table 3 & 4.

Table 3 Input process parameters and their levels

Table 4 Experimental results

FUZZY LOGIC ANALYSIS a fuzzy variable has been described. Afterwards Generally soft computing techniques are helpful allocate the suitable membership functions for both when accurate mathematical modelling does not exist input variables and the output response and these techniques vary from traditional computing characteristics. In the final steps, fuzzy rules are because of its inexactness, ambiguity, and partial created. The membership function is a graphical reality. Fuzzy logic is the one of the soft computing illustration of the degree of contribution of each method that take part in a prominent role in order to variable. However there are several numbers of create the relationship in between input parameters membership functions that are recommend such as with their response characteristics. Fuzzy Logic has triangular (trimf), Trapezoidal (trapmf), Gaussian and been employed to find out the output parameters by bell-shaped (gbelmf), Gaussian (gaussmf), Gaussian means of Matlab software Kovac et al. (2013). It combinational (gauss2mf), Sigmoidal (sigmf), etc. involves following three process) fuzzification b), The graphical demonstration of fuzzy logic system is fuzzy inference) and defuzzification. In the first step shown in figure 1.

Fig. 1 Fuzzy logic system

Int J Adv Engg Tech/Vol. VII/Issue I/Jan.-March.,2016/413-417 Hynes et al., International Journal of Advanced Engineering Technology E-ISSN 0976-3945

Fig. 2 Membership function for (a) Rotational speed (b) Friction time (c) Friction Pressure

where p, q, r show the process parameters and y is a variable. Input process parameters are segregated into three membership functions; they are small (S), medium (M) and large (L). Similarly output parameters are separated into nine membership functions; very very small (VVS), very small (VS), small (S), medium low (ML), medium (M), medium high (MH), high (H), very high (VH), very very high (VVH) each. In order to characterize the connection between the inputs and outputs variables, fuzzy If- then rules are effectively employed. It can be specified as follows: Rule 1: if y1 is P1 and y2 is Q1, and y3 is R1then z1 is S1 and z2 is T1else Rule 2: if y1 is P2 and y2 is Q2, and y3 is R2then z1 is S2 Fig. 3 Membership function for (a) Impact strength (b) and z2 is T2else Axial shortening length Rule n: y1 is Pn and y2 is Qn, and y3 is Rn then z1 is Sn and In this work, triangular membership function used to z2 is Tn define the three input variables and two output wherey1, y2, y3 are corresponds to rotational speed, response characteristics as shown in figure 2 & 3. friction time and friction pressure respectively. The triangular membership function can be expressed Similarly z1, z2 corresponds to impact strength and by the following expression Emel et al (2013) axial shortening length respectively. P, Q, R, S, and T 0 ≤ are linguistic assessments described by fuzzy sets on ⎧ ⎫ ⎪ ≤ ≤ ⎪ the ranges y y y and z z respectively. 1, 2, 3 1, 2 (; , , ) = (1) ⎨ ≤ ≤ ⎬ ⎪ ⎪ ⎩0 ≤ ⎭ Table 4 Experimental results

With the intention to determine the validity of the established fuzzy logic model; the predicted values are compared with the experimental values. The relationships between the experimental values and predicted values are demonstrated in figure 4 & 5. Table 5 shows the fuzzy logic data, difference and accuracy.

Fig. 4 Correlation between experimental and predicted impact strength

Int J Adv Engg Tech/Vol. VII/Issue I/Jan.-March.,2016/413-417 Hynes et al., International Journal of Advanced Engineering Technology E-ISSN 0976-3945

metals. In fuzzy logic model R2 values are determined as 99.15% for impact strength and 98.05% for axial shortening respectively. CONCLUSION Based on the experimental results on friction stud welding of AA 6063 and AISI 1030 steel, the fuzzy logic models are developed successfully. The fuzzy logic model predicts impact strength and axial shortening with 11.50% and 7.44 % error respectively. Optimum level of process parameters of Fig. 5 Correlation between experimental and predicted friction stud welding of AA 6063 and AISI 1030 axial shortening length steel is predicted as 1063 rpm rotational speed, 7 sec MULTI-OBJECTIVE OPTIMIZATION friction time and 300 Kpa friction pressure. The Response optimizer is a technique that permits for maximum impact strength and minimum axial compromise between the different responses. It is shortening length are achieved as 206.46 KJ/m2 and carried out by obtaining the individual desirability (d) 7.3798 mm respectively. for each response, and then combining the individual ACKNOWLEDGEMENTS desirability to attain the composite desirability (D) The authors gratefully acknowledge the financial consequently maximizing or minimizing the support of this work by SERB of Department of composite desirability and predicting the optimal Science & Technology, New Delhi. (Sanction Letter input process variable settings. In this work, multi- No.: SERB / F / 1452 / 2013-2014 dated 10.06.2013). response optimization of impact strength and axial REFERENCES shortening length of friction stud welding of AA 1. Ates, H., Turker, M., &Kurt, A. (2007). Effect of 6063 and AISI 1030 steel metals are carried out. The friction pressure on the properties of friction welded optimal level for achieve the maximum impact MA956 iron-based super alloy. Materials and Design, strength (IS) and minimum axial shortening length 28, 948–953. 2. Ellis, C. R. G. (1972). Continuous drive friction (AS) value are determined. The permissible limits of welding of mild steel. Welding Journal, 183-197. friction stud welding process parameters are 3. Ozdemir, N., Sarsilmaz, F.,& Hascalik, A. (2007). considered as: Effect of rotational speed on the interface properties of 800 rpm ≤ RS ≥ 1600 rpm friction-welded AISI 304L to 4340 steel. Material Design, 28, 301–307. 3 sec ≤ FT ≥ 7 sec 4. Mumin, S.,Erol, A. H., & Kaan, O. (2008). An 300 Kpa ≤ FP ≥ 500 Kpa experimental study on joining of severe plastic deformed materials with friction welding method. Materials and Design, 29, 265–274. 5. Fauzi, M. N. A., Uday, M. B., Zuhailawati, H., & Ismail, A. B. (2010). Microstructure and mechanical properties of alumina-6061 aluminum alloy joined by friction welding. Materials and Design, 31, 670–676. 6. Ozdemir, N. (2005). Investigation of the mechanical properties of friction-welded joints between AISI 304L, AISI Steel as a function rotational speed. Materials Letters59, 2504–2509. 7. Rajesh Jesudoss Hynes, N., Nagaraj, P.,& Angela Jennifa Sujana, J. (2012). Investigation on Joining of Aluminum and Mild Steel by Friction Stud Fig. 6 Optimal level of process parameters of friction Welding. Materials and Manufacturing Processes, stud welding 27(12), 1409-1413. The individual desirability for impact strength and 8. Rajesh Jesudoss Hynes, N., Nagaraj, P., axial shortening are obtained as 0.77128 and 0.96063 Palanichamy, R., Arumugham, C. A. K., & Angela respectively. The optimal composite desirability for Jennifa Sujana, J. (2014). Numerical Simulation of heat flow of Friction Stud Welding of Dissimilar this experiment is 0.86077. Figure 6 demonstrates the metals. The Arabian Journal for Science and optimum level of joining performance condition at a Engineering, 39(4), 3217-3224. particular process parametric arrangement i.e. 9. Alves, E. P., Neto, F. P.,&An, C.Y. (2010). Welding of rotational speed (RS) 1063 rpm, friction time (FT) 7 AA1050 aluminium with AISI 304 stainless steel by sec and friction pressure (FP) 300 Kpa. Impact process. Journal of Aerospace 2 Technology and Management, 2(3), 301-306. strength can be reached as high as 206.46 KJ/m ; 10. Rajesh Jesudoss Hynes, N., Nagaraj, P., &Angela axial shortening can be attained as low as 7.3798 Jennifa Sujana, J. (2014). Mechanical evaluation and mm. microstructure of friction stud welded aluminium – mild steel joints. The Arabian Journal for Science and RESULTS AND DISCUSSION Engineering, 39, 5017-5023. From the graphical plot of experimental and fuzzy 11. Rajesh Jesudoss Hynes, N., Nagaraj, P., &Angela logic predicted values are almost same and the Jennifa Sujana, J. (2014). Ultrasonic evaluation of deviations between experimental and fuzzy predicted friction stud welded AA6063/AISI 1030 steel joints. Materials and design, 62, 118-123. values are low. The absolute error for impact strength 12. Paventhan, R., Lakshminarayanan, P. R., & and axial shortening length was determined to be Balasubramanian, V. (2011). Prediction and 11.50% and 7.44% respectively. The accuracy of optimization of friction welding parameters for joining prediction of developed models for impact strength aluminium alloy and stainless steel. Transactions of Nonferrous Metals Society of China, 21, 1480-1485. and axial shortening are determined as 98.72% and 13. Paventhan, R., Lakshminarayanan, P. R., & 99.17% respectively. From the results it was revealed Balasubramanian, V. (2012). Optimization of Friction that fuzzy logic modelling can be effectively used for Welding Process Parameters for Joining Carbon Steel prediction the impact strength and axial shortening and Stainless Steel. Journal of Iron and Steel Research, parameters of friction stud welding of dissimilar 19(1), 66-71.

Int J Adv Engg Tech/Vol. VII/Issue I/Jan.-March.,2016/413-417 Hynes et al., International Journal of Advanced Engineering Technology E-ISSN 0976-3945

14. Karupanan, B., Mahendra, K. M., Paul, S., &Abdullah, N. S. (2014). Studies and Parameter Optimization of AISI 904LSuper Austenitic Stainless Steel Using Friction Welding. Materials Research 17(4), 908-919. 15. Ajith, P. M, Birendra, K. B., Sathiya, P., & Aravindan, S. (2015). Multi-objective optimization of friction welding of UNS S32205 duplex stainless steel. Defence Technology 11, 157-165. 16. Ajith, P. M, Birendra, K. B., Sathiya, P., & Aravindan, S. (2015). Multi-objective Optimization of Continuous Drive Friction Welding Process Parameters Using Response Surface Methodology with Intelligent Optimization Algorithm. Journal of Iron and Steel Research, 22(10), 954-960. 17. Manideep, D.,& Balachandar, K. (2012). Welding Parameters Metallurgical Properties Correlation of Friction Welding of Austenitic Stainless Steel and Ferritic Stainless Steel. Journal of applied sciences, 12, 1013-1019. 18. Paulraj, S., Aravindan, S., & Noorul, H. A. (2006). Optimization for friction welding parameters with multiple performance characteristics. International Journal of Mechanics and Materials in Design, 3(4), 309 -318. 19. Kovac, P., Rodic, D., Pucovsky, V., Savkovic, B., & Gostimirovic, M. (2013). Application of fuzzy logic and regression analysis for modeling Surface roughness in face milling. Journal of Intelligent Manufacturing, 24, 755–762. doi10.1007/s10845-012-0623-z. 20. Emel, K., & Babur, O. (2013). Fuzzy logic and regression modelling of cutting parameters in drilling using vegetable based cutting fluids. Indian journal of engineering & materials sciences, 20, 51-58. 21.

Int J Adv Engg Tech/Vol. VII/Issue I/Jan.-March.,2016/413-417