Thermal Performance Optimization of Nano-Enhanced Phase Change Material-Based Heat Pipe Using Combined Artificial Neural Network and Genetic Algorithm Approach

Abstract

The present study focuses on the numerical investigation of nano-enhanced phase change material (Ne-PCM)-based heat pipes designed for electronic cooling applications. It uses both paraffin wax and n-eicosane as phase change materials (PCMs) that are combined with copper oxide (CuO) nanoparticles at different concentrations of 1%, 3%, 5%, and 7%. The heat input to the heat pipe ranges from 10 to 50 W in an increment of 10 W to simulate realistic operating conditions. The idea is to predict the heat pipe's thermal performance at various combinations of nanoparticles and PCMs and compare the same to the baseline case of deionized (DI) water (without PCM). The results show a constant drop in the evaporator temperature for the Ne-PCM-assisted heat pipes. Paraffin wax and n-eicosane exhibit maximum reductions of 2.86% and 1.94%, respectively, in evaporator wall temperature compared to using conventional DI water (without PCM). The thermal resistance of the heat pipe also decreases consistently with increasing the heat input for all cases, with the most significant reduction of 33.11% and 16.63% for paraffin wax–CuO- and n-eicosane–CuO-assisted heat pipes, respectively. The maximum evaporator heat transfer coefficients recorded are 257.79 W/m2K, 353 W/m2K, and 265.18 W/m2K for heat pipes using DI water (without PCM), paraffin wax–CuO, and n-eicosane–CuO, respectively. The nanoparticles act as a thermal conductivity enhancer and bring down the heat pipe's evaporator temperature with the addition of PCMs. Thus, the effective thermal conductivity of the Ne-PCM-based heat pipe is notably higher compared to heat pipes using DI water (without PCM). To understand the complex thermal behavior of the Ne-PCM-based heat pipes and to better predict their thermal performance, a predictive model has been developed using an artificial neural network (ANN). This model drives the genetic algorithm (GA) that considers the multivariable interaction of PCMs and nanoparticle concentrations to identify the optimal configuration and results in better thermal performance of the heat pipe. Thus, the combination of ANN and GA aids a useful approach for the effective prediction of the heat pipe's thermal performance. The outcomes of this study are useful for the development of sustainable solutions in electronic cooling applications.