An artificial neural network for predicting corrosion rate and hardness of magnesium alloys. (15th January 2016)

• Record Type: Journal Article
• Authors: Xia, X.; Nie, J.F.; Davies, C.H.J.; Tang, W.N.; Xu, S.W.; Birbilis, N.
• Is Part Of: Materials & design. Volume 90(2016)
• Page Start: 1034
• Page End: 1043
• Publication Date: 2016-01-15
• Subjects: Magnesium -- Mg alloys -- Corrosion -- Neural network -- Hardness; Materials -- Periodicals; Engineering design -- Periodicals; Matériaux -- Périodiques; Conception technique -- Périodiques; Electronic journals; 620.11
• DOI: 10.1016/j.matdes.2015.11.040
• Languages: English
• ISSNs: 0264-1275

Abstract

Abstract: There presently exists a demand for development of magnesium (Mg) alloys for wrought applications. In this study, alloying additions of Zn, Ca, Zr, Gd and Sr to Mg were made in binary, ternary and quaternary combinations up to a maximum total alloy loading ~ 3 wt.%, and thus termed dilute. Such dilute alloys were studied for the purposes of potential sheet applications. The corrosion of a total of 53 custom alloys was studied in conjunction with microhardness. The results reveal that hardness increased with total alloy loading, whilst the corrosion rates did not show any clear relationship with alloy loading. Corrosion of the tested alloys was instead very sensitive to both the type and amount of the unique alloying addition. This indicates that the optimisation of properties requires a detailed knowledge of the electrochemical influence of unique alloying additions. The work contributes to an understanding of compositional effects on the corrosion of Mg, and can be exploited in prediction of corrosion resistance of existing and future Mg alloys. Graphical abstract: Highlights: Emerging Mg alloys for automotive sheet applications explored. Mg alloys designed to optimise hardness with minimal increase in corrosion rate. Artificial neural network (ANN) model constructed to manage complex data set. ANN model accurately predicted alloy hardness and corrosion rate.