A B were conducted. The welding speed was designed to vary from 1 to 2 mm/s and the welders adjusted the current in accordance to the varying weld pool geometry. The arc length was set at 4 mm for both experiments. As can be observed in Fig. 13, the proposed ANFIS model can estimate the welder’s response with good accuracy. Conclusion and Remarks 52-s RESEARCH FEBRUARY 2014 VOL. WELDING 93 To derive the skilled human welder response model, a number of dynamic experiments were designed to examine how the welder responds to 3D weld pool geometry. The weld pool characteristic parameters, including the width, length, convexity, and previous adjustments made by the welder, were utilized as model inputs, and the human welder’s current adjustment was considered as the model output. A linear model was first constructed as an average model over the entire input range and an ANFIS model was then proposed to provide better modeling performance. Analysis suggests that the skilled welder adjusts the current dynamically based on the weld pool surface geometry with approximately a 1.5-second delay. His/her adjustment is, in general, positively correlated to the weld pool convexity and negatively to the weld pool width, while the correlation of his/her adjustment to the weld pool length is more complex and depends on the weld pool surface width and convexity. His/her adjustment on the current is also positively correlated to the last current adjustment. More accurately, his/her adjustment on the current correlates to the width, length, and convexity of the weld pool surface, and the last adjustment nonlinearly and the ANFIS model improves the modeling accuracy. In a future investigation, the developed neuro-fuzzy human intelligence model will be analyzed and compared with that of a novice welder and then be utilized as an intelligent controller to perform penetration control in an automated GTAW process. Acknowledgments This work was funded by the National Science Foundation under grant CMMI- 0927707 and IIS-1208420. YuKang Liu would like to thank WeiJie Zhang and Benny Porter for their assistance in conducting experiments. References 1. O’Brien, R., ed., Welding Handbook. 2008. 8th edition, vol. 2, welding processes. American Welding Society, Miami, Fla. 2. Uttrachi, G. D. 2007. Welder shortage requires new thinking. Welding Journal 86(1): 6. 3. Richardson, R. W., and Edwards, F. S. 1995. Controlling GT arc length from arc light emissions. Trends in Welding, Proceedings of the 4th International Conference, pp. 715–720. 4. Li, P. J., and Zhang, Y. M. 2001. Precision sensing of arc length in GTAW based on arc light spectrum. Journal of Manufacturing and Science Engineering 123: 62–65. 5. Hartman, D. A., DeLapp, D. R., Barnett, R. J., and Cook, G. E. 1998. A neural network/fuzzy logic system for weld penetration control. Trends in Welding Research, Proceedings of the 5th International Conference, pp. 1096–1101. 6. Balfour, C., Smith, J. S., and Al-Shamma’A, A. I. 2006. A novel edge feature correlation algorithm for real-time computer vision-based molten weld pool measurements. Welding Journal 85(1): 1-s to 8-s. 7. Mi, B., and Ume., C. 2006. Real time weld penetration depth monitoring with laser ultrasonic sensing system. Transactions of ASME: Journal of Manufacturing Science and Engineering 128: 280–286. 8. Fan, C., Lv, F., and Chen, S. 2009. Visual sensing and penetration control in aluminum alloy pulsed GTA welding. International Journal of Advanced Manufacturing Technology 42(1): 126–137. 9. Song, H. S., and Zhang, Y. M. 2009. Error analysis of a three-dimensional GTA weld pool surface measurement system. Welding Journal 88(7): 141-s to 148-s. 10. Ma, H.,Wei, S., and Lin, T. et al. 2010. Binocular vision system for both weld pool and root gap in robot welding process. Sensor Review 30(2): 116–123. 11. Lin, M. L., and Eagar, T. W. 1985. Influence of arc pressure on weld pool geometry. Welding Journal 64(4): 163-s to 169-s. 12. Rokhlin, S. I., and Guu, A. C. 1993. A study of arc force, pool depression, and weld penetration during gas tungsten arc welding. Welding Journal 72(8): 381-s to 390-s. 13. Kovacevic, R., and Zhang, Y. M. 1996. Monitoring of weld joint penetration based on weld pool geometrical appearance. Welding Journal 75(10): 317-s to 329-s. 14. Zhang, W. J., Liu, Y. K., Wang, X., and Zhang, Y. M. 2012. Characterization of three-dimensional weld pool surface in GTAW. Welding Journal 91(7): 195-s to 203-s. 15. Zhang, W. J., and Zhang, Y. M. 2012. Modeling of human welder response to 3D weld pool surface: Part I-principles. Welding Journal 91(11): 310-s to 318-s. 16. Zhang, W. J., and Zhang, Y. M. 2012. Modeling of human welder response to 3D weld pool surface: Part II-results and analysis. Welding Journal 91(12): 329-s to 353-s. 17. Hayashi, K., et al. 1995. Neuro fuzzy transmission control for automobile with variable loads. IEEE Transactions on Control Systems Technology 3(1): 49–53. 18. Tanaka, K., Sano, M., and Watanabe, H. 1995. Modeling and control of carbon monoxide concentration using a neuro-fuzzy technique. IEEE Transactions on Fuzzy Systems 3(3): 271–279. 19. Kovacevic, R., and Zhang, Y. M. 1997. Neurofuzzy model-based weld fusion state estimation. IEEE Control System Magazine 17(2): 30- 42. 20. Jang, J. S. R. 1993. ANFIS: Adaptive-network based fuzzy inference systems. IEEE Transactions on Systems, Man, and Cybernetics 23(3): 665–685. 21. Lih, W., et al. 2008. Adaptive neuro-fuzzy inference system modeling of MRR and WIWNU in CMP process with sparse experimental data. IEEE Transactions on Automation Science and Engineering 5(1): 71–83, 2008. 22. Zhai, J., et al. 2009. The dynamic behavioral model of RF power amplifiers with the modified ANFIS. IEEE Transactions on Microwave Theory and Techniques 57(1): 27–35. 23. Imen, M., Mansouri, M., and Shoorehdeli, M. A. 2011. Tracking control of mobile robot using ANFIS. Proceedings of the 2011 IEEE International Conference on Mechatronics and Automation, Aug. 7–10, Beijing, China. 24. Liu, Y. K., Zhang, W. J., and Zhang, Y. M. 2013. Adaptive neuro-fuzzy inference system (ANFIS) modeling of human welder’s response to 3D weld pool surface in GTAW. Journal of Manufacturing Science and Engineering-Transactions of the ASME 135: 1210101–12101011. Fig. 13 — Verification experiment results. A — Experiment 1; B — experiment 2.
Welding Journal | February 2014
To see the actual publication please follow the link above