Numerical investigation of effective thermal conductivity for two-phase composites using a discrete model

Enhancing the effective thermal conductivity plays an important role in different areas, e.g. PCM systems, plastics or compound materials. For example, different approaches such as adding particles with high thermal conductivity (mostly graphite or metals) are known for enhancing the effective thermal conductivity of PCM systems. For designing PCM thermal energy storages or simulating storage processes, values for the effective thermal conductivity need to be known for example for material systems consisting of paraffin as PCM with graphite particles. In the present article, a discrete numerical model is described which allows estimating the effective thermal conductivity for three-dimensional geometries with particles distributed and orientated randomly in another material. Anisotropic thermal conductivity of the particles is considered since graphite in-plane and trough-plane thermal conductivity differ significantly. Influences of particle shape, distribution, orientation and linkage on the effective thermal conductivity are investigated. Additionally, deviations caused by model simplifications such as two-dimensional geometry or isotropic particle properties are studied. It is shown that 3D modelling is required and isotropic simplification might cause significant errors. Finally, results are compared to values obtained from correlation equations available from the literature.