Machine learning for predicting tensile properties in graphene- reinforced polyethylene composites

KL Luo and WL Luo and JL Zuo and JX Xiao and TL Jiang and LX Huang and WH Zhang, MATERIALS TODAY COMMUNICATIONS, 47, 113053 (2025).

DOI: 10.1016/j.mtcomm.2025.113053

Graphene-reinforced polyethylene composites(GRPC), as a representative of graphene-based composites, integrate the exceptional mechanical properties of graphene, such as strength and stiffness, with the remarkable abrasion and corrosion resistance of polyethylene. Investigating the mechanical properties of these composites holds significant research importance, as it can further refine the preparation processes and formulations, thereby enhancing the strength of the composites. This paper presents an integration of molecular dynamics simulations with machine learning techniques to predict the tensile mechanical properties of graphenepolyethylene composites. By taking into account various factors, including temperature, tensile strain rate, graphene deletion rate, polyethylene molecular chain length, and graphene volume fraction, a comprehensive dataset concerning the ultimate tensile strength and modulus of elasticity of the composites is generated through molecular dynamics simulations. This dataset is subsequently utilized for training and optimization using machine learning algorithms, resulting in the development of a predictive model capable of estimating the tensile mechanical properties effectively. Upon comparing feed-forward neural networks, support vector machines, and AdaBoost regression algorithms, the study found that the feed-forward neural network model exhibited the best performance in terms of prediction accuracy and overall performance, making it an ideal choice for predicting the tensile mechanical properties of GRPC.

Return to Publications page