In this article an artificial neural network based system to predict weld bead geometry using features derived from the infrared thermal video of a welding process is proposed. The multilayer perceptron and radial basis function networks are used in the prediction model and an online feature selection technique prioritises the features used in the prediction model. The efficacy of the system is demonstrated with a number of welding experiments and using the leave one out cross-validation experiments.
A novel variant of tungsten inert gas (TIG) welding called activated-TIG (A-TIG) welding, which uses a thin layer of activated flux coating applied on the joint area prior to welding, is known to enhance the depth of penetration during autogenous TIG welding and overcomes the limitation associated with TIG welding of modified 9Cr-1Mo steels. Therefore, it is necessary to develop a specific activated flux for enhancing the depth of penetration during autogeneous TIG welding of modified 9Cr-1Mo steel. In the current work, activated flux composition is optimized to achieve 6 mm depth of penetration in single-pass TIG welding at minimum heat input possible. Then square butt weld joints are made for 6-mm-thick and 10-mm-thick plates using the optimized flux. The effect of flux on the microstructure, mechanical properties, and residual stresses of the A-TIG weld joint is studied by comparing it with that of the weld joints made by conventional multipass TIG welding process using matching filler wire. Welded microstructure in the A-TIG weld joint is coarser because of the higher peak temperature in A-TIG welding process compared with that of multipass TIG weld joint made by a conventional TIG welding process. Transverse strength properties of the modified 9Cr-1Mo steel weld produced by A-TIG welding exceeded the minimum specified strength values of the base materials. The average toughness values of A-TIG weld joints are lower compared with that of the base metal and multipass weld joints due to the presence of d-ferrite and inclusions in the weld metal caused by the flux. Compressive residual stresses are observed in the fusion zone of A-TIG weld joint, whereas tensile residual stresses are observed in the multipass TIG weld joint.
It is essential to set up the activated tungsten inert gas (A-TIG) welding process parameters to produce the desired weld bead geometry and heat affected zone (HAZ) width in modified 9Cr-1Mo steel weld joints. Therefore, it becomes necessary to develop a tool for optimisation of A-TIG welding process. Genetic algorithm (GA) based model has been developed to determine the optimum process parameters. In this methodology, first independent ANN models correlating depth of penetration, weld bead width and HAZ width with current, voltage and torch speed respectively were developed. Then, GA code was developed in which the objective function was evaluated using the ANN models. There was good agreement between the target and actual values of bead geometry and HAZ width obtained using the GA optimised process parameters. Thus, a methodology using GA has been developed for optimising the A-TIG process parameters for modified 9Cr-1Mo steel.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.