Regression modeling and Neural Computing for predicting the Ultimate Tensile Strength of Friction Stir Welded aerospace Aluminium Alloy Mishra Akshansh1,*, Vance Jonathan Ve2, Dasgupta Anish3, Saravanan M4 1Founder and Project Scientific Officer, Stir Research Technologies, Uttar Pradesh, India 2Department of Engineering Design, IIT, Madras, India 3Technical Officer, Stir Research Technologies, Uttar Pradesh, India 4Head Design Engineer, Stir Research Technologies, Uttar Pradesh, India *Corresponding Author E-mail: akshansh.frictionwelding@gmail.com
Abstract AA7075 is an aluminum alloy that's almost as strong as steel, yet it weighs just one third as much. Unfortunately its use has been limited, due to the fact that pieces of it couldn't be securely welded together by the traditional welding process. Friction Stir Welding (FSW) process overcomes the limitations of conventional welding process. The aim of our present is to compare the predicted results of the Ultimate Tensile Strength (UTS) of Friction Stir welded similar joints through Regression modeling and Artificial Neural Network (ANN) modeling. It was observed that the linear regression algorithm is able to make more accurate predictions compared to neural network algorithm for small dataset. Top Keywords Artificial Neural Network, Regression Model, Friction Stir Welding. Top |