Heat Transfer Optimization of Ternary Nanofluid Flow over a Thin Needle with Joule Heating: Artificial Neural Network Approach

Abstract

Abstract This study investigates thermal transport in a ternary nanofluid flowing over a thin needle, with applications in microscale heat exchangers, biomedical devices, electronic cooling, and coating technologies. The analysis incorporates the combined influence of an externally applied transverse magnetic field and resistive heating, which are relevant in electrically conducting cooling fluids and electromagnetic thermal systems. Using suitable similarity transformations, the governing partial differential equations are reduced to nonlinear ordinary differential equations and solved numerically via the MATLAB bvp4c solver. A Levenberg–Marquardt based artificial neural network is then developed to predict the heat transfer rate and skin friction coefficient of the ternary nanofluid. The dataset is partitioned into training (70%), validation (15%), and testing (15%) sets to ensure model accuracy and stability. The ANN results closely match numerical solutions, with errors around 10⁻⁹–10⁻¹⁰ for heat transfer and 10⁻⁷–10⁻⁸ for skin friction. Findings indicate that water–ethylene glycol ternary nanofluids significantly enhance thermal performance, even at low nanoparticle concentrations (0.0025–0.03), outperforming hybrid nanofluids. The findings highlight the effectiveness of artificial intelligence as a fast and reliable alternative to repeated numerical simulations, offering significant potential for real-time thermal design and optimization in advanced energy and cooling technologies.