▎ 摘　　要

Many studies have shown that low concentration of graphene can appreciably increase the thermal conductivity of nanofluids that is difficult to explain by conventional thermal conductivity model. In this paper, a new model considering the effects of graphene's length and thickness, interface layer and anisotropic characteristics is established. The model is further modified by considering the possibility of absence of graphene in the actual thermal conduction path, that is only bulk liquid along the heat conduction path in some cases. It is found that the thermal conductivity predicted by the present model is positively correlated with the temperature, concentration and length of nanoparticles, while is slightly negatively correlated with thickness. This model is applied to predict thermal conductivities of graphene nanofluids in different cases compared to other models. The results indicate that the present model has a strong ability to predict graphene nanofluids especially at relatively low concentrations (<0.1 vol%). (C) 2021 Elsevier B.V. All rights reserved.