ANN-based optimization of disk-shaped microchannel heat exchanger for thermal and hydraulic performance improvement

Abstract

In thermal management systems, achieving uniform temperature distribution and minimizing pressure drop in microchannel heat exchangers remains a critical challenge. This study proposes an innovative disk-shaped microchannel heat exchanger with flow tunnel (DMHX-FT) to improve temperature uniformity and reduce pressure drop while maintaining efficient heat transfer. The DMHX-FT features a dendritic fractal microchannel layout to enhance turbulence and fluid flow equalization, along with hub-shaped flow tunnels for efficient recirculation. A feedforward backpropagation Artificial Neural Network (ANN) was employed to analyze parameter impacts and develop a predictive performance model, followed by a genetic algorithm to identify optimal solutions balancing pressure drop and temperature difference. The DMHX-FT achieves a 45 % reduction in temperature difference across various heat fluxes and a 75 % reduction in pressure drop compared to traditional designs. Experimental results align closely with numerical predictions, with discrepancies confined to a maximum of 10 %. The DMHX-FT effectively addresses key challenges in microchannel heat exchangers, offering a promising solution for advanced thermal management, supported by a robust ANN and genetic algorithm optimization framework.