Study the effect of shielded metal arc welding process parameters, cryo-treatment and preheating on welding characteristics and modelling by an artificial neural network.

Depth of heat affected zone (HAZ), depth of penetration and bead geometry (bead height and bead width) are most important characteristics of a weldment. In addition to weld process parameters welding conditions like preheating and cryo-treatment greatly affect the weldment characteristics. It is possible to get good bead geometry with cryo-treatment and preheating for some range of current, voltage and arc travel rate for which bead formation may not be possible in normal welding condition. In SMAW (shielded metal arc welding) process, selecting appropriate values for process variables is essential in order to control weldment characteristics. Also, conditions must be selected that will ensure a predictable and reproducible weld bead. In this investigation the effects of cryo-treatment and preheating on various metallurgical aspect, namely, the depth of HAZ, weld interface, grain growth and grain refinement regions have been studied. While cryo-treatment before welding decreased the grain size in HAZ and preheating increased the grain size and cryo-treatment after welding did not significantly increase the grain size. The process was modelled using an artificial neural network. Supervised mode of learning was used while training the neural networks. The test cases showed good convergence with the measured experimental values. [ABSTRACT FROM AUTHOR]