Rationalizing the microstructure interplay in the thermal conductivity of a metal matrix composite via 3D imaged-based finite element modeling

The thermal performance of metallic brake lining materials plays a decisive role in the safety and efficiency of high-speed railway braking systems. In this study, a combined experimental–numerical methodology is developed to rationalize the influence of microstructural constituents on the effective thermal conductivity of a sintered metal matrix composite (MMC) brake lining. Laser Flash Analysis (LFA) is first employed to determine the thermal conductivity of some individual constituents as well as that of reference composites. X-ray CT (XCT) provides three-dimensional reconstructions of the microstructure that are subsequently used to generate realistic image-based finite element meshes. The unknown thermal conductivities of the graphite particles are identified through a Finite Element Model Updating (FEMU) scheme, where numerical predictions of the effective conductivity are iteratively matched to LFA measurements. These findings highlight the strong anisotropy of graphite particles and their favored orientation after compaction, which governs heat transport pathways. Moreover, the presence of intra-, inter-, and inter-connectivity porosity within and around the graphite is shown to significantly reduce the transverse conductivity, rationalizing the discrepancy between the FEM predictions and experimental values. Overall, the proposed approach demonstrates how combining LFA, XCT and FEMU enables the identification of constituent-level conductivities and provides new insights into the microstructure/thermal-property relationships of MMC brake linings. • FEMU-tuned XCT models match LFA data to identify constituent conductivities. • G2 graphite particles possess strong intrinsic thermal anisotropy. • XCT shows graphite particles align in compaction, driving anisotropic heat transport. • Intra, inter, and interface porosity at G2 particles lowers transverse conductivity. • FEM matches axial conductivity but overestimates transverse because of omitted porosity.