The Paton Welding Journal, 2006, №06

2006 №06 (04)

2006 №06 (06)

---

The Paton Welding Journal, 2006, #6, 21-24 pages

On neural network application for welded joint quality control in underwater welding

I.O. Skachkov¹, A.E. Pirumov¹, S.Yu. Maksimov², E.A. Prilipko²

¹NTUU «Kiev Polytechnic Institute», Kiev, Ukraine
²E.O. Paton Electric Welding Institute of the NASU. 11 Kazymyr Malevych Str., 03150, Kyiv, Ukraine.

Abstract
Artificial neural networks have been used to asses the quality of underwater welded Joints with four types of fit-up defects: edge misalignment, change in part-to-torch distance, change in the gap between the parts and presence of tack welds. The efficiency of neural network application for the above types of defects is shown, except for a change in the edge gap.
Keywords: underwater welding, non-stationary disturbances, weld metal, quality control, electrical parameters, neural networks

References

1. Zhao, D.B., Chen, S.B., Wu, L. et al. (2001) Intelligent control for the shape of the weld pool in pulsed GTAW with filler metal. Welding Res., 11, 253–260.
2. Adolfsson, S., Bahrami, A., BolmsJo, G. et al. (1999) Online quality monitoring in short-circuit metal arc welding. Ibid., 2, 59–72.
3. Denisenko, V., Khalyavko, A. (2001) Protection from disturbances of sensors and connecting wires of industrial automation systems. Sovrem. Tekhnologii Avtomatizatsll, 1, 68–75.
4. Medvedev, V.S., Potyomkin, V.G. (2002) Neural networks: Matlab 6. Moscow: Dialog-Mifi.

---

Suggested Citation

I.O. Skachkov, A.E. Pirumov, S.Yu. Maksimov, E.A. Prilipko (2006) On neural network application for welded joint quality control in underwater welding. The Paton Welding J., 06, 21-24.

---