DOI: 10.14704/NQ.2022.20.11.NQ66090

SUPPORT VECTOR MACHINE AND DECISION TREE ALGORITHM FOR SURFACE CHARACTERIZATION OF FRICTION STIR WELDING OF ALUMINIUM ALLOY 2024 PIPES

S.Deivanai and Dr. Manoj Soni

Abstract

For many engineering applications, the finish on a surface can have a big effect on the performance and durability of parts. Rough surfaces generally wear more rapidly and have greater friction coefficients than smooth surfaces. Typically, roughness is a dependable predictor of mechanical part performance, as irregularities tend to form nucleation sites for breaks or corrosion. Roughness is a measure of the fine irregularities on a surface. The surface roughness is a function of both raw material properties and manufacturing variables. Weld surface characteristics such as corrosion resistance, oxidation, and wear resistance toughness are affected by the cutting tool geometry, tool rotational speed and weld speed of Friction stir welding process.The objective of this study is to evaluate the weld surface roughness using support vector machine learning algorithm and decision tree algorithm for classification of weld by FSW process parameters and to build a mathematical model for prediction using artificial neural network in R studio software. Currently, there is very little information about classification of weld surface roughness using machine learning algorithms for friction stir welded aluminium alloy 2024 pipe joint available.

Keywords

Full Text

References