Multiscale analysis on anisotropic heat conduction behaviors of multi-walled carbon nanotubes modification carbon fiber composites: random distribution and modified interface

Carbon fiber fabric composites (CFFCs) are widely used in aerospace and energy systems, but their anisotropic thermal conductivity (ATC) remains challenging to predict due to hierarchical structures and non-uniform nanofiller distributions. Existing models often oversimplify nanoscale filler randomness or ignore interfacial effects, limiting accuracy. This study addresses these gaps by developing a three-scale finite element framework integrating microscale random multi-walled carbon nanotube (MWCNT) dispersion, mesoscale interfacial layers, and macroscale fabric architecture. MWCNTs (0–0.6 % mass fraction) were experimentally incorporated into CFFCs to enhance thermal performance. Key findings include a 56.6 % in-plane thermal conductivity improvement and an 18.5 % through-thickness reduction at 0.6 % MWCNT mass fraction, driven by directional heat flux redistribution. The three-scale model predicted ATC with <9 % error compared to Hot-Disk experiments, demonstrating its reliability. This work provides critical insights into hierarchical heat transfer mechanisms, enabling tailored thermal management in advanced composites.