Nanofluid Heat Transfer Enhancement in Porous Media With Geometric and Particle Morphology Effects

ABSTRACT

Efficient thermal management in porous, nanofluid‐filled systems is essential for applications such as compact heat exchangers, microchannel coolers, and solar thermal collectors. This study investigates the flow and heat transfer of a water‐based copper (Cu) nanofluid through a converging/diverging channel embedded in a Darcy–Forchheimer porous medium, with particular attention to two often‐overlooked factors: nanoparticle shape and viscous dissipation. The nanofluid's thermal conductivity is modeled using a shape‐dependent Maxwell–Garnett formulation, allowing the effects of spherical, columnar, and lamellar particle geometries to be distinguished. The governing nonlinear partial differential equations are reduced to a system of ordinary differential equations using similarity transformations, and solved numerically via the Lobatto IIIA method. To complement this direct numerical solution, an intelligent computing framework based on a Levenberg–Marquardt backpropagation neural network (LMBNNS) is developed and trained using MATLAB's neural fitting tool, offering a fast and accurate surrogate model for predicting velocity and temperature profiles under varying physical conditions. Twelve parametric scenarios covering channel angle, porosity, inertia effects, Eckert and Prandtl numbers, nanoparticle volume fraction, and wall stretching/shrinking are analyzed. The neural network model achieves close agreement with the numerical solutions, with mean squared errors as low as and regression coefficients approaching unity () across all scenarios, confirming both the reliability of the physical model and the predictive accuracy of the proposed computing approach. The results provide practical insight into how channel geometry and nanoparticle morphology can be jointly exploited to enhance heat transfer performance in porous thermal systems.