Prediction of mechanical properties of Al-C nanocomposite: a machine learning approach

Nowadays, investigating the properties of nanocomposites and finding their optimal properties pave the wayfor a better use of them. In this study, the main purpose is the prediction of mechanical properties of aluminum-carbon(Al-C) nanocomposite being elastic modulus and ultimate tensile stress with a machine learning algorithm. TheRandom Forest Regression (RFR) algorithm has been utilized in this research. Furthermore, the independent variablesof T, S, and C illustrating ambient temperature, strain rate, and the content of carbon used in aluminum, respectively,and dependent variables of E and UTS indicating elastic modulus and ultimate tensile stress obtained from moleculardynamics simulation have been utilized to train the algorithm to predict the mentioned mechanical properties. Theresults represent that the prediction accuracies of ۸۱.۲ % and ۹۳.۱ % derived from RFR algorithm for elastic modulusand ultimate tensile stress have been obtained, respectively.

کلیدواژه ها:

Machine learning ، Mechanical properties ، Molecular dynamics ، Nanocomposite