TAGUCHI EXPERIMENTAL DESIGN AND ARTIFICIAL NEURAL NETWORK SOLUTION OF STUD ARC WELDING PROCESS
Main Article Content
Abstract
Stud arc welding has become one of the most important unit operations in the mechanical industries. The need to reduce the time from product discovery to market introduction is inevitable. Reducing of standard deviation of tensile strength with desirable tensile strength joint as a performance character was use to illustrate the design procedure. The effects of (welding time, welding current, stud material, stud design, sheet material, sheet thickness, sheet cleaning and preheating) were studied. Design of Experiment (DOE) is a structured and organized method to determine relationships between factors affecting a process and output of the process itself. In order to design the best formulation it is of course possible to use a trial and error approach but this is not an effective way. Systematic optimization techniques are always preferable. Tensile strength quality is one of the key factors in achieving good stud welding process performance. 225 samples of stud welding was tested. Computer aided design of experiment for the stud welding process based on the neural network artificial intelligence by Matlab V6.5 software was also explain. The ANN was designed to create precise relation between process parameters and response. The proposed ANN was a supervised multi-layer feed forward one hidden layer with 8 input (control process parameters), 16 hidden and 2 output (response variables) neurons. The learning rule was based on the Levenberg-Marquardt learning algorithm.
The work of stud welding was performed at the engineering college laboratory, Baghdad University by using the DABOTEKSTUD welding machine, for 6 mm diameter stud. The sheet materials are (K14358 and K52355) according to (USN standards, and stud materials are (54NiCrMoS6 and 4OCrMnMoS8-6) according to (DIN standards).
The eight control parameters (welding time, sheet thickness, sheet coating, welding current, stud design, stud material, preheat sheet and surface condition) were studied in the mixed L16 experiments Taguchi experimental orthogonal array, to determine the optimum solution conditions.
The optimum condition was reached for the stud welding process tensile strength, where the researcher develops a special fixture for this purpose. The analysis of results contains testing sample under optimum condition, chemical composition of usage materials and micro structure of optimal condition sample.
According to that:
Practicality: the influence parameters that affect the stud welding process are welding time, which have a major effect on stud welding process, followed by sheet material and stud material.
The reduction in standard deviation was approximately (30.06 per cent) and for the range was as approximately (29.39per cent). In the other side the increase in the tensile strength mean was as approximately (30.84 per cent). The influence parameters that affect the tensile strength stud welding process are: the factor welding time has a major effect on stud welding process, followed by factor C (sheet coating) and factor F (stud material).
Article Details
How to Cite
Publication Dates
References
Jibson J.”Advance Welding”, John Wiley & Sons, 1997.
Montgomery D.C” Design and Analysis of Experiments” Second Edition John Wiley &Sons, Inc., 2001.
Montgomery D.C. "The Use of Statistical Process Control and Design of Experiments in Product and Process Improvement" JIB Transactions, Vol.24 No. 5. PP. 4-17, 1992.
Montgomery D.C “Introduction to Statistical Quality Control” John Wiley & Sons, Inc.1985.
Allen IT. and et al” Statistical Process Design for Robotic GMA We dipg of Sheet Metal” Welding Journal, PP.69-s_77-s, May 2002.
Kackar R.N.” Qft4Jng Quality Control, Parameter Decigppnd the Tagnehi „VIe/hod”. .1. Qual. Techn. , Vol. 17, No. 4, PP 176-188, 1985.
Ottoy K. N. and Antonsson E. K.” Extensions to the Taguchi Method of Product ASME Journal of Mechanical Design. January 6, 1991.
Galdmez E.V.C. and Carpinetti L.C.R. “Application as Design oLpgrimentu7 the Process of Manufhcturing of Plastic Products”, ASME Journal of Manufacturing Science and Engineering, Vol. 122, PP. 360-369, 2004.
Coit D. W. Jackson B. T. and Smith A. E.” Static Neural IVetwork Process Models: Considerations and Case Studies”. mt j. Prod. Res., Vol. 36. No. 11, 2953- 2967, 1998.
Su C. and Miao C. “Neural Network_Procedures fbr ExperpgjtilAigil.sisJ/h Censored Data”, International Journal of Quality Science, Vol.3, No.3, PP.239-253. 1998.
Hsu C.M.” Solving Multi-Response Problems through Neural Networks aJgLj‟iI7cipp7 J22gnepd‟ApglJis”. Journal of the Chinese Institute of Industrial Engineers, Vol. 18. No. 5, pp. 47-54, 2001.
Roy R., "Design of Experimental Using the Tagpchi Approach‟ Wiley, New York, 2001.
Montgomery D. C. “Design andAnalsis QjExperiments”. New York. Wiley. 1991.
Buyske S. and Trout R.” Robust Design and Taguchi Methods” .Journal of Quality Technology, vol. 22, No. 1, PP. 15 22, Jan. 2003.
Steiner 5.1-1. and MacKay R. J.” Statistical Engineering: an Algorithm far Reducing Variation in Majfpcturing Processes” American Society for Quality ASQ, 2005.
Taguchi G. and Yokoyama Y.” Taguchi Method‟s: Design of Experiments “Quad ity Engineering, Vol.4, Dearbon, MI, 1993.
Mukhopadhyay S. K. and Chakraborty D. “Optimal Process Variance under Tqggjj Loss” International Journal of Quality & Reliability Management. Vol. 12 No. 9, PP. 14-29, 1995.
Antony J. and et al “Process Optimization using Taguchi Me1fpfExperimeJ7/ul çin” Work Study, Volume 50, Number 2, PP. 51-57, 2001.
Lofthouse T. “The Taguchi Loss Function” Work Study, Volume 48, Number 6. PP. 218-222, 1999.