Prediction of mechanical strength of magnesium alloy AZ31 with calcium addition using a neural network based model

Abstract

The aim of the research is to develop a neural network model to predict the mechanical strength of AZ31 magnesium alloy in addition with Calcium. This would help researchers or academicians to fabricate the magnesium alloy AZ31+x wt. % Ca which finds more importance in making biodegradable implants in orthopedic surgery. As the yield strength of human bone lies between 104.9 – 114.3 MPa whereas for pure Mg it is very lesser, it is needed to make alloys suitable for biomedical applications which are to be satisfied with their mechanical properties and biocompatibility. An artificial neural network-based model is used to predict the mechanical properties of Mg alloys AZ31+x wt. % Ca with compositions of Al, Zn, Mn and Ca as countable parts with various proportions. The compositions and properties of twenty different as cast alloys were used for training and prediction of the desired mechanical properties using Artificial Neural Network and the outputs are compared with the experimental results. The results show that the Levenberg-Marquardt (LM) algorithm serves better in prediction of mechanical properties of Mg alloys AZ31 and AZ31+1 wt. % Ca. Also the results were found significant at 90%, 95% and 99% confidence limits.