Abstract
This work is to suggest a method to model the friction welded properties for dissimilar AISI 430 steel and EN 10028 P355 GH Steel joints using techniques like response surface methodology (RSM) and fuzzy logic (FL). Modelling used to predict correct parameters and their range without sacrificing weld quality is very crucial in the manufacturing sector. The techniques handled in this work used to find the welded joint properties like ultimate tensile strength (UTS), Impact toughness (IT) and axial shortening (AS) were obtained by solid state friction welding process in advance. The 27 sets of 5 input process variables like frictional pressure, forge pressure, forging time, frictional time, and chuck speed were considered. The fuzzy logic model output values are related to experimental values and the average error is found to be 2.05%, 8.9%, and 8.31% of the responses tensile stress, axial shortening, and impact toughness respectively. Similarly, RSM based model value and experimental values are related, and the average error is estimated to be 7.43%, 10.45%, and 13.78% of the responses tensile stress, axial shrinkage, and impact toughness respectively. The enhanced quality output obtained by fuzzy logic analysis is outperformed as average error yields less than response surface methodology. This work helps us to suggest the best modelling technique for the prediction of friction welded joint properties.
Keywords: Axial Shortening, Friction Welding, Fuzzy Logic, Mathematical Modelling, Response Surface Methodology