Title: Predicting the tensile behaviour of friction stir welded AA2024 and AA5083 alloy based on artificial neural network and mayfly optimisation algorithm

Authors: P.M. Diaz; M. Julie Emerald Jiju

Addresses: Dedicated Juncture Researcher's Association, Kulasekharam, Kanyakumari, 629161, Tamil Nadu, India ' Department of MCA, CSI Institute of Technology, Thovalai, Kanyakumari, 629302, Tamil Nadu, India

Abstract: The necessity of aluminium based metal matrix composites are growing rapidly in various fields especially in automobiles. To predict the tensile behaviour of AA2024 and AA5083 alloys, a new approach has been proposed by integrating the artificial neural network with mayfly optimisation algorithm (MOA). To analyse the predicting efficiency of the proposed approach, it is compared with artificial neural networks and experimental test values. For predicting the ultimate tensile strength of AA2024 and AA5083 alloys, the proposed approach achieved very less absolute error and mean absolute error of 0.0147% and 0.3680% respectively. Similarly, the prediction of the tensile elongation of AA2024 and AA5083 alloys, the proposed ANN-MOA approach achieved very less absolute error and mean absolute error of 0.0017% and 0.3269% respectively. The results from the analysis indicated that the proposed approach has enhanced predicting accuracy than artificial neural networks.

Keywords: aluminium alloy; mechanical properties; ANN; artificial neural network; mayfly optimisation algorithm; inertial weights.

DOI: 10.1504/IJCMSSE.2023.135833

International Journal of Computational Materials Science and Surface Engineering, 2023 Vol.11 No.3/4, pp.163 - 186

Received: 30 Sep 2021
Accepted: 14 Aug 2022
Published online: 08 Jan 2024 *