Application of materials informatics to vapor-grown carbon nanofiber/vinyl ester nanocomposites through self-organizing maps and clustering techniques

摘要: Abstract Data mining and knowledge discovery techniques were employed herein to acquire new information on the viscoelastic, flexural, compressive, tensile properties of vapor-grown carbon nanofiber (VGCNF)/vinyl ester (VE) nanocomposites. Formulation processing factors (curing environment, presence or absence dispersing agent, mixing method, VGCNF weight fraction, type, high-shear time, sonication time) testing temperature utilized as inputs true ultimate strength, yield engineering elastic modulus, flexural storage loss tan delta selected outputs. The data algorithms used in this study include self-organizing maps (SOMs) clustering techniques. SOMs demonstrated that had most significant effects output responses followed by time. also produce optimal using certain combination(s) inputs. Fuzzy C-means algorithm (FCM) was applied discover patterns nanocomposite behavior subsequent a principal component analysis (PCA), which is dimensionality reduction technique. Utilizing these techniques, specimens separated into different clusters based (30 °C 120 °C being dominant responses), delta, Furthermore, VGCNF/VE cluster their viscoelastic (storage moduli) at same temperature. FCM results indicate that, while all framework are essential, significant. This work highlights utility context materials informatics for trends material not immediately known.