Home > Published Issues > 2014 > Volume 3, No. 1, January 2014 >

Artificial Neural Networks in Tensile Strength and Input Parameter Prediction in Friction Stir Welding

Kudzanayi Chiteka
Production Engineering, Delhi Technological University, India

Abstract—Welding speed and rotational speed have been singled out as the most influential welding parameters which affect the tensile strength as well as the hardness in Friction Stir Welding (FSW). It is however problematic to determine the possible welding speed and rotational speed given the Ultimate Tensile Strength (UTS) since there are several combinations of welding speeds and rotational speeds that can yield the same UTS. At the same time, however, the input parameters predicted may not be available on the machine. This research is therefore aimed at using Artificial Neural Networks (ANN) in predicting the UTS given rotational speed and welding speed as well as exploring the possibility of obtaining the input parameters given the output UTS.

Index Terms—Friction stir Welding, Input parameter prediction, Tensile strength prediction, Artificial neural network

Cite: Kudzanayi Chiteka, "Artificial Neural Networks in Tensile Strength and Input Parameter Prediction in Friction Stir Welding," International Journal of Mechanical Engineering and Robotics Research, Vol. 3, No. 1, pp. 145-150, January 2014.