Weld metal chemistry and microhardness of CaO–SiO 2 –TiO 2 –MgO based electrode coatings for marine welding applications

Abstract

In order to keep their structural integrity, marine structures are becoming more reliant on advanced welding procedures and carefully chosen materials. Fabrication, maintenance, and repair of these complex systems must be done with the utmost precision. In this work, electrode coatings for shielded metal arc welding (SMAW) are developed using the extreme vertices design process. The present study introduces a novel approach for using multi-pass bead-on-plate experiments to analyze the microhardness, microstructure, and weld chemistry. Low-alloy steel plates have been covered with a layer of weld beads, and the composition of the beads was determined by analyzing the elements carbon, silicon, manganese, chromium, molybdenum, nickel, nitrogen, and iron. A Vickers microhardness tester was used to measure the welds’ microhardness. The correlation between weld chemistry, microhardness, and flux components was then determined using regression analysis. Models of artificial neural networks were developed and compared with regression analysis results. The study revealed that ANN predictions outperformed regression models by 24.78%, particularly in accurately predicting carbon content, which was negatively influenced by TiO₂. Conversely, CaO, SiO₂, and MgO exhibited positive effects on silicon content, manganese, chromium, molybdenum, and nitrogen levels. Ternary components significantly enhanced iron content and microhardness, with ANN predictions achieving improvements of 54.25% for iron and 83.25% for microhardness. Overall, ternary interactions played a crucial role in achieving substantial gains in prediction accuracy, highlighting the effectiveness of the ANN model.