Development of an Artificial Intelligence-Based Algorithm for Predicting the Mechanical Properties of Weld Joints of Dissimilar S700MC-S960QC Steel Structures

Articles in Press

Authors

1 Department of Mechanical Engineering, ENSET Douala, University of Douala, Cameroon

2 Department of Mechanical Engineering, University of Douala, Douala, Cameroon

3 Laboratory of Mechanics (LM), University of Douala-Cameroon, PoBox: 1872, Douala, Cameroon

4 Kwame Nkrumah University of Science and Technology, Mechanical Engineering Department

5 Department of Engineering Science, University West, Gustava Melius Gata 2 S-461 32, Trollhättan, Sweden

Abstract

This study focuses on the development of an artificial intelligence (AI) algorithm designed to determine the mechanical properties of high-strength steels, particularly in weld joints created through dissimilar gas metal arc welding of S700MC and S960QC steels. To achieve this goal, a mathematical model based on artificial neural networks (ANNs) was employed. Sixteen experiments were conducted, generating data on the yield strength and tensile strength concerning the welding parameters and a filler wire with a similar carbon equivalent. Initially, the algorithm was set up to predict joint characteristics using only welding parameters as input variables. However, to enhance the accuracy of the predictions, the carbon equivalent of the filler metal was incorporated as an additional input variable. This adjustment resulted in improved prediction outcomes compared to those obtained without considering the filler wire. The implementation of the AI algorithm was carried out using MATLAB, specifically its R2017b version. The algorithm's ability to predict mechanical properties based on the given input variables showcases its potential utility in optimizing welding processes and ensuring the desired mechanical properties of weld joints in high strength steels.

Keywords

Civil Engineering Infrastructures Journal

Articles in Press, Accepted Manuscript
Available Online from 01 January 2025

Files

History

  • Receive Date: 08 April 2024
  • Revise Date: 03 December 2024
  • Accept Date: 01 January 2025

Share

How to cite

Statistics