Weld Quality Prediction for Resistance Microwelding of Fine Cu Wire Based on the Back-Propagation Neural Network

Abstract:

Article Preview

A neural network model is established to predict the joint quality in resistance microwelding (RMW) of fine Cu wire and stainless steel thin sheet. The preheat current, welding current, weld time and electrode force are selected as the input parameters and the tensile strength as the output parameter in the model. The prediction program is compiled in MATLAB through the detailed designing of hidden layer and the selection of the transfer function. The network performance is verified by experiment, and its accuracy meets the production requirements in the actual welding.

Info:

Periodical:

Edited by:

Xianghua Liu, Zhenhua Bai, Yuanhua Shuang, Cunlong Zhou and Jian Shao

Pages:

1709-1712

DOI:

10.4028/www.scientific.net/AMM.217-219.1709

Citation:

B. H. Mo and Z. N. Pan, "Weld Quality Prediction for Resistance Microwelding of Fine Cu Wire Based on the Back-Propagation Neural Network", Applied Mechanics and Materials, Vols. 217-219, pp. 1709-1712, 2012

Online since:

November 2012

Export:

Price:

$38.00

[1] L.Q. Han: Theory, Design and application of Artificial Neural Network (Chemical Industry Press, Beijing 2007).

[2] X. Wen and L. Zhou: MATLAB Simulation and Application for Neural Network (Science Press, Beijing 2003).

[3] P. Chen, Z.N. Guo and B.H. Mo: Adv. Mater. Res. Vols. 418-420 (2012), p.1448.

[4] H.T. Lee, M. Wang, R. Maev and E. Maeva: Int. J. Adv. Manuf. Technol. Vol. 22 (2003), p.727.

In order to see related information, you need to Login.