Machine learning insights into the strain-stress behavior of nitinol alloys at low temperatures: Utilizing the orange data mining tool

Abstract

This study entails a new technique, based on machine learning algorithms, for predicting the strain-stress behavior of Nitinol alloys. The study utilizes Orange data mining software to determine if algorithms like Linear Regression, Random Forest, k-nearest neighbors, and decision tree are related to the effectiveness. The nitinol-based alloy, which has both shape memory and superelasticity, is used in a wide range of biomedical devices, aerospace constructions, and automated apparatus. Temperature variation is a main factor affecting the Nitinol behavior. It is therefore required to carry out an in-depth analysis of strain-stress patterns. Employing machine learning algorithms along with data mining tools allows the exact prediction of the Nitinol alloy performance under any given condition. The research is purposed to establish the relationship between temperature and mechanical strength of Nitinol by analyzing strain-stress curves that were obtained from tensile tests carried out at different temperatures. Looking at overall results, kNN has the highest accuracy of the models upon comparing MSE and RMSE, and has higher R² and, therefore stands as the model of high reliability. Results show that machine learning algorithms successfully forecast the mechanical responses of Nitinol under temperature variations, which assists in determining their formability properties and hence, advancing Nitinol applications in various fields. This research demonstrates machine learning uses in developing materials science and engineering knowledge by showing the ability to predict Nitinol alloy behavior, which includes strain-stress characteristics at low temperatures. Furthermore, the research's significance lies in addressing ethical considerations and practical issues, such as the necessity of implementing an organized approach to achieve environmental benefits.