A hybrid computing scheme for forward and reverse mappings of metal inert gas welding process

A hybrid computing scheme has been developed for carrying out forward and reverse mappings of metal inert gas welding process. As the input-output relationships of this process may not be the same over the entire range of the variables, the mappings have been done using the clusters made by the data points. Optimisation has been carried out to improve the performances of both fuzzy clustering techniques as well as radial basis function neural network used for the clustering and reasoning, respectively. Two approaches of hybrid computing scheme have been proposed, in which a binary-coded genetic algorithm has been utilised to decide the optimal structure of the network (through clustering of the data using two different approaches separately) and a back-propagation algorithm has been employed to determine the optimised parameters of the network. The hybrid scheme has yielded better performance compared to that developed using the genetic algorithm only.

[1]  De-Shuang Huang,et al.  A Hybrid Forward Algorithm for RBF Neural Network Construction , 2006, IEEE Transactions on Neural Networks.

[2]  K. W. Um,et al.  A prediction of welding process parameters by prediction of back-bead geometry , 2000 .

[3]  Dilip Kumar Pratihar,et al.  Global versus cluster-wise regression analyses for prediction of bead geometry in MIG welding process , 2007 .

[4]  Manoranjan Dash,et al.  Entropy-based fuzzy clustering and fuzzy modeling , 2000, Fuzzy Sets Syst..

[5]  Prasad K. Yarlagadda,et al.  A study on prediction of bead height in robotic arc welding using a neural network , 2002 .

[6]  Ill-Soo Kim,et al.  An investigation into an intelligent system for predicting bead geometry in GMA welding process , 2005 .

[7]  Dilip Kumar Pratihar,et al.  Forward and reverse mappings of the tungsten inert gas welding process using radial basis function neural networks , 2009 .

[8]  D. S. Nagesh,et al.  Prediction of weld bead geometry and penetration in shielded metal-arc welding using artificial neural networks , 2002 .

[9]  Jacek M. Zurada,et al.  Nonlinear Blind Source Separation Using a Radial Basis Function Network , 2001 .

[10]  Dilip Kumar Pratihar,et al.  Modeling of TIG welding process using conventional regression analysis and neural network-based approaches , 2007 .

[11]  Yujun He,et al.  Short-term Load Forecasting Model Using Fuzzy C Means Based Radial Basis Function Network , 2006, Sixth International Conference on Intelligent Systems Design and Applications.

[12]  John H. Holland,et al.  Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence , 1992 .

[13]  Henry Leung,et al.  Detection of small objects in clutter using a GA-RBF neural network , 2002 .

[14]  Witold Pedrycz,et al.  Improving RBF networks performance in regression tasks by means of a supervised fuzzy clustering , 2006, Neurocomputing.

[15]  James C. Bezdek,et al.  Fuzzy mathematics in pattern classification , 1973 .

[16]  C. Butler,et al.  MODELLING AND OPTIMIZING OF A MIG WELDING PROCESS—A CASE STUDY USING EXPERIMENTAL DESIGNS AND NEURAL NETWORKS , 1997 .

[17]  Prasad K. Yarlagadda,et al.  Prediction of welding parameters for pipeline welding using an intelligent system , 2002 .