"Predicting the modulus of elasticity for biocompatible titanium alloys"

Abstract

Titanium alloys have been present for decades as the main components for the production of various orthopedic and dental elements. However, modern times require titanium alloys of different composition, with lower modulus of elasticity, without the presence of toxic alloying elements such as Al and V [1]. Traditional methods used to detect dependencies between parameters, as well as alloy design, are often not particularly effective and usually require large investments of time and resources. The study introduces the machine learning technique Extra Tree Regression, which, through the analysis of data from 246 biocompatible titanium alloys, identifies factors associated with reduced elastic modulus [2]. The three most influential were: specific heat and mass fraction of titanium and mass fraction of titanium silicon. Using data on the most influential factors, four-component diagrams were designed where certain alloy compositions reach a modulus of up to 54 GPa. In addition, Monte Carlo simulations were used to demonstrate the feasibility of modeling multicomponent alloys with elastic modulus below 70 GPa.