Comparison of Response Surface Methodology and Artificial Neural Networks in The Estimation Of Thermal Conductivity Mild Steel Tig Weld

Thermal conductivity is a measure of heat flow per second per unit area per temperature gradient. This study compared Response surface Methodology to Artificial Neural Networks in the opitimisation of the thermal conductivity of mild steel weld using the TIG welding process, The input parameters considered in this study were welding current, welding voltage and gas flow rate, while the measured parameters was thermal conductivity. Twenty sets of experiment were performed using 5 specimens for each run. The plate samples were 60 mm long with a wall thickness of 10mm. The models RSM R2 value for RSM was 92.38% and ANN 99.85%. This shows that ANN is a better predictor as compared to RSM.