AI-assisted CFD energy and exergy analysis of turbulent natural convection of ternary hybrid nanofluids in a 3D open-ended enclosure with a wavy heated wall

Abstract

Passive cooling systems have attracted significant attention in recent years due to their cost-effectiveness and strong thermal performance. This study presents a detailed numerical investigation of steady-state natural convection in a three-dimensional open-ended cubic cavity featuring a wavy heated wall. The cavity is filled with ternary hybrid nanofluid composed of water ($H_{2}O$) as the base fluid and three types of nanoparticles: copper ($Cu$), copper oxide ($CuO$), and aluminum oxide ($A{l}_2{O}_3$). Buoyancy-induced fluid motion is modeled using the Boussinesq approximation. The governing equations for both laminar and turbulent flows are solved using the Reynolds-Averaged Navier–Stokes (RANS) method with the realizable $k$-$\varepsilon$ turbulence model, following an experimentally validated approach. A parametric analysis examines the effects of Rayleigh number $(10^6\le Ra\le 10^{12})$, nanoparticle volume fraction $(0\text{\%}\le \varphi \le 5\text{\%})$, and the amplitude of wall waviness $(0\text{\%}\le A\le 30\text{\%})$ on thermal performance. The results reveal that incorporating wavy wall geometries in combination with nanofluids can substantially enhance the thermal performance of the system. Under certain optimized conditions, this configuration leads to a greater enhancement in heat transfer compared to the increase in entropy generation, resulting in a system efficiency exceeding unity. These findings highlight the strong potential of geometrically engineered surfaces for improving thermal transport in energy systems. To supplement the numerical results, an artificial neural network (ANN) was trained using the Levenberg–Marquardt algorithm on 72 datasets, accurately predicting average Nusselt numbers and validating the simulation trends as a fast and reliable predictive tool.