Machine learning analysis for heat transfer enhancement in nano-encapsulated phase change materials within L-shaped enclosure with heated blocks

Abstract

Phase change materials(PCMs) are crucial to energy storage systems due to their enhanced thermal properties. They significantly boost energy efficiency and promote sustainability. Nevertheless, the low thermal conductivity of PCMs presents a significant challenge, which is addressed by utilizing nano-encapsulation to enhance energy efficiency in energy storage systems. Motivated by this, the current study conducts a theoretical investigation to explore the heat transfer characteristics in a buoyancy-driven Nano-Encapsulated Phase Change Materials(NEPCM) nanofluid within an L-shaped porous enclosure with the impacts of a heated block and magnetic field. Furthermore, the fusion temperature plays a crucial role in initiating phase change in NEPCM, thereby impacting the heat transfer process. Hence, identifying the optimal fusion temperature is essential. To accomplish this, a machine learning approach was employed to identify the ideal fusion temperature. A dataset of 160 data points across four different fusion temperature values was used in this analysis. Additionally, the machine learning model analyzed how variations in fusion temperatures impact physical parameters. An in-house Matlab code is utilized to solve the dimensionless fluid transport equations employing the finite difference method. The results indicate that increasing the nanoparticle volume fraction significantly enhances the heat transfer rate across all physical parameters. Specifically, under higher thermal buoyancy force, increasing the volume fraction from 1% to 5% results in a 90.04% increase in the heat transfer rate. The numerical analysis demonstrates that heat transfer rates improve significantly when the fusion temperature is adjusted to 0.5, a result further validated by machine learning techniques. At this temperature, thermal buoyancy force increases by 0.98% and 2.68% compared to values of 0.1 and 0.9, respectively, while the Stefan number shows increases of 159.42% and 87.48% under these conditions; thereby, the heat transfer rate increases at this value. This computational study provides important insights into the significance of fusion temperature, emphasizing the need to determine its optimal value for improving heat transfer. Identifying this optimal value can enhance the efficiency of thermal energy storage systems and improve cooling performance in electronic devices.