MODELLING AND PROCESS PARAMETERS OPTIMIZATION FOR THIN METAL SHEET JOINING BY TIG WELDING

Abstract

This study investigates multi-response optimization in Tungsten Inert Gas (TIG) welding for an optimal parametric combination to predict the weld characteristics for thin cold rolled steel (CRC) sheet To study the weld characteristics responses ; weld width, heat affected zone and under bead depression, autogenous butt welds were made. Full factorial method is used to obtain the combinations of experimental runs. The effect of welding process parameters; welding current, arc length and speed have been seen on weld geometry. Regression analysis is done to develop the mathematical models between the welding process parameters and weld characteristics responses. Analysis of variance (ANOVA) is used to check the adequacy of developed mathematical models. The confirmatory tests were also conducted to validate the accuracy of mathematical models. Sensitivity analysis is also done to analyze the effect of individual process parameter on the weld responses. ANN technique is also used for the prediction of weld responses parameters. By using the same experimental data, weld responses parameters are predicted using back propagation neural network. The appropriate neural network structure for predicting weld responses parameters is chosen by trial-and-error method. lv