TIG Auto Welding Process Using Adaptive Wavelet Neural Network Controller

Abstract:

The study proposed adaptive wavelet neural network controller can achieve good and precise welding control performance and use synchrotron radiation research center developed multi-gun group automatic welding system to verify the validity of the research method. Multi-gun group welding system is applied in Taiwan Photon Source (TPS). Storage ring aluminum alloy vacuum chamber of Taiwan Photon Source .In the past aluminum alloy vacuum chamber welding, it all depends on the empirical welding rule of operator to give appropriate welding current, argon flow, wire feed speed and welding speed for control. Therefore, the paper uses automatic welding skill, which takes National Instruments PXI-8180 system as basic structure, and adaptive wavelet neural network controlled four optimized parameters, I.E. welding current, wire feed speed, flow rate of argon gas and welding speed, The vacuum chamber pressure value is also up to 6.2X10-10Torr/mA. It is successfully applied to the TPS system. Therefore, it can prove the effectiveness and practicality of the method proposed in this study.

• Periodical: Applied Mechanics and Materials (Volumes 764-765)
• Pages: 634-639
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.764-765.634
• Online since: May 2015
• Authors: Yen Bin Chen, Yung Lung Lee*, Shou Jen Hsu, Chin Chun Chang, Yi Wei Chen
• Keywords: Adaptive Control, Aluminium Alloy, Wavelet Neural Network

References

[1] D. Rosenthal: Mathematical theory of heat distribution during welding and cutting, Welding J. 20 (5) (1941) 220–234.

[2] C.S. Wu, T. Pulte, D. Rehfeldt: Gas metal arc welding process monitoring and quality evaluation using neural network, Sci. Technol. Welding Joining5 (5) (2000) 324–328.

DOI: 10.1179/136217100101538380

[3] D.S. Nagesh, G.L. Datta: Prediction of weld bead geometry and penetration in shielded metal-arc welding using artificial neural networks, J. Mater. Process. Technol. 123 (2002) 303–312.

DOI: 10.1016/s0924-0136(02)00101-2

[4] Kim IS et al.: A study on relationship between process variables and bead penetration for robotic CO2 arc welding. J Mater Process Technol 2003; 136: 139–45.

DOI: 10.1016/s0924-0136(02)01126-3

[5] Lin CL, Shieh NC: Tung PC (2002) Robust wavelet neuro control for linear brushless motors. IEEE Trans Aerosp Electron Syst 38 (3): 918-932.

DOI: 10.1109/taes.2002.1039408

[6] Zhang Q: (1997) Using wavelet network in nonparametric estimation. IEEE Trans Neural Netw 8 (2): 227-236.

DOI: 10.1109/72.557660

[7] C. -C. Chang, S. -N. Hsu, G. -Y. Hsiung, and J. -R. Chen: Manufacturing and Welding of the Aluminum Ultra-high Vacuum Chambers for the Low Temperature Superconducting Magnets (in Chinese), Journal of Taiwan Vacuum Society 19(1), (2006).