Properties prediction and design of self-healing epoxy resin combining molecular dynamics simulation and back propagation neural network

Abstract

It is difficult to synergistically improve the self-healing and mechanical properties of self-healing epoxy resin system based on Diels-Alder (DA) reversible covalent bond due to its different compositions. In this study, a method combining molecular dynamics (MD) simulation and back propagation (BP) neural network was employed to achieve a balance between self-healing property and mechanical property. In addition, self-healing and mechanical properties were predicted and the influence of different compositions of the system on these properties was investigated by the BP neural network. It was found that the DA bond formed by the furan group of furfuryl glycidyl ether (FGE) and the maleimide group of bismaleimide (BMI) was of great importance for self-healing property, and that diglycidyl ether of bisphenol-A (DGEBA) with a mass fraction of 25%–48% could improve the tensile property of the system. Finally, a self-healing epoxy resin system with superior comprehensive properties was designed through the proposed method under four indexes of Young’s modulus (E), self-healing efficiency (), Ultimate Tensile Strength (UTS), and glass transition temperature (Tg).