An Intelligent Algorithm for Lab Joint in the Automatic GMA Welding Process

DOI: 10.5176/2251-1857_M318.10

Authors: Jian-Wei Zhang, Tae-Jong Yun, Min-Ho Park, Ill-Soo Kim, Byeong-Ju Jin and Han-Yong Yoon

Abstract: The automatic welding system is currently employed in high volume production industries where the cost of equipment is justified by the large number of pieces to be made. It also requires the detailed movement devices with predetermined sequences of process parameter and the use of timers to form the weld joints. However, an intelligent algorithm that predicts the optimal bead geometry and accomplishes the desired mechanical properties of the weldment in the automatic GMA(Gas Metal Arc) welding should be developed. The algorithm should also cover a wide range of material thicknesses and be applicable for all welding position. In addition, the proposed model for the automatic welding system must be available in the form of mathematical equations. The objective of this paper is concentrated on the development of an intelligent model, which employed the neural network algorithm, one of AI (Artificial Intelligence) technologies in order to study the effects of welding parameters on bead total area and predict the optimal bead total area for lab joint in the automatic GMA welding process. BP (Back Propagation) and LM (Levenberg-Marquardt) neural network algorithm have been used to develop the intelligent model. Not only the fitting of these models have been checked and compared by using variance test, but also the prediction on bead total area using the developed models have been verified.

Keywords: GMA(Gas Metal Arc) welding process, BP(Back Propagation)neural network, LM(Levenberg-Marquardt) algorithm, Lab joint, Bead total area