Optimization of Parameters for CO2-Gas Shielding Arc Welding Based on the Double Weights Neural Network


Article Preview

It is very difficult to predict the process detection and performance parameters, because of the highly nonlinear and multivariable coupling of welding process. In this paper, we construct a new method in optimization of parameters for CO2-gas shielding arc welding. By using the new Double Weights Model, this algorithm can give the Direction Weight, also the Core Weight at the same time. The new network inherits the traditional BP network and RBF (Radial Basis Functions) network with multivariable parameter settings and so on. We apply the network to the optimization of welding parameters, experimental results show that this algorithm can use less generations to calculate and get more accurate optimization effects, also not serious about a local minimum compared with RBF while using the same environment and equal network scale. Experiment proves that it is feasible to control welding parameters by the Double Weights Neural Network.



Edited by:

Jerry Tian






L. L. Lv et al., "Optimization of Parameters for CO2-Gas Shielding Arc Welding Based on the Double Weights Neural Network", Advanced Materials Research, Vol. 304, pp. 175-179, 2011

Online since:

July 2011




[1] Y. Suga, M Narus and T Tkiwa: Application of neural network to visual sensing of welding and automatic tracking in robot welding. Submitted to Welding in the World, 1994, 34(2): 63~66.

[2] Shisheng Huang, Di Li and Yonglun Song: Neural network modeling and control study of the welding process. Submitted to Journal of Mechanical Engineering, 1994, 3: 23~25.

[3] S B Chen, L Wu and Q L Wang: Self-learning fuzzy neural networks and computer vision for control of pulsed GTAW. Submitted to Welding Journal, 1997(5), 52~54.

[4] Nan Ni: Optimization of technics parameters of welding robot CO2-gas Protecting and welding based on the neural network. Hefei: Hefei industrial university, college of electrical and automation engineering, (2005).

[5] Shoujue Wang, Zhaozhou Li, Xiangdong Chen and Bainan Wang: Discussion on the basic mathematical models of neurons in general purpose neuron computer. Submitted to Electronic Journals, 2001, 29(5): 577~580.

[6] Yu Cao, Xingtao Zhao: Data fitting based on new double weights neural network. Submitted to Electronic Journals, 2004, 32 (10): 1671~1673.

[7] Yu Cao: The research of double weights neural network using data fitting. Master Degree Theses of Institute of semiconductors, CAS, 2002(6): 32~34.

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