A machine learning framework for predicting the shear strength of carbon nanotube-polymer interfaces based on molecular dynamics simulation data

Abstract Modern aerospace applications require lightweight materials with exceptionally high strength and stiffness. Carbon nanotube (CNT)-reinforced composites have great potential in addressing these requirements. However, one critical factor limiting the of CNT-reinforced is limited load transfer capability between CNTs through a polymer matrix, which arises due to low CNT-polymer interfacial shear at molecular scale. While dynamics (MD) simulations can be employed investigate interface, such are computationally expensive. It thus intractable explore sufficiently large design space for interface modifications optimization using MD alone, motivating use surrogate models efficiently map configurations strength. In this paper, we develop machine learning model trained functionalized CNT-epoxy corresponding The proposed ML consists (i) feature representation based on radial distribution functions; (ii) convolutional neural network representations metric strength; (iii) data augmentation method maximize training data. predict CNT pullout force quite well despite variation associated force.