Quality Estimation of the Resistance Spot Welding Based on Hopfield Network

Abstract:

Lots of dynamic information, which can directly or indirectly reflect the quality of welded spot, is included within the electrode displacement signal of resistance spot welding process. In this research, the displacement signal is monitored and mapped into a 15×25 element bipolarized matrix by means of some method of fuzzy theory. Some welded spots are classified into five classes according to the prototype pattern matrixes. An effective RSW quality estimation system is developed based on Hopfield network when taking the tensile shear strength of the welded spot joint as the estimation index of welded spot quality. The results of cross-validation test shows that the Hopfield network can satisfactorily accomplish the task of classification of the welded spot and has board application prospects.

• Periodical: Advanced Materials Research (Volumes 189-193)
• Pages: 3364-3369
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.189-193.3364
• Online since: February 2011
• Authors: Hong Jie Zhang, Yan Yan Hou
• Keywords: Hopfield Network, Pattern Recognition, Quality Estimation, RSW

References

[1] K.I. Johnson and J.C. Needham, Welding Journal, Vol. 51(1972), pp.122-131.

[2] D.W. Dickinson, J.E. Franklin, and A. Stanya, Welding fournal, Vol. 59(1980), p.170.

[3] M. Hao K.A. Osman and D.R. Boomer, Welding journal, Vol. 75(1996), p.1.

[4] Hongjie Zhang , Transactions of the china welding institution, vol. 30 (2009), p.97.

[5] Haitao XUE and Yongyan LI , Transactions of the china welding institution, Vol. 28(2007), p.39.

[6] Y. Cho. and S. Rhee, Welding journal, vol. 81(2002), p.104.

[7] Y. Cho and S. Rhee, Welding journal, vol. 82(2003), p.185.