Machine Learning-Based Approach for Modeling the Nanofluid Flow in a Solar Thermal Panel in the Presence of Phase Change Materials

Abstract

Considering the importance of environmental protection and renewable energy resources, particularly solar energy, the present study investigates the temperature control of a solar panel using a nanofluid (NFD) flow with eco-friendly nanoparticles (NPs) and a phase change material (PCM). The PCM was used under the solar panel, and the NFD flowed through pipes within the PCM. A number of straight fins (three fins) were exploited on the pipes, and the output flow temperature, heat transfer (HTR) coefficient, and melted PCM volume fraction were measured for different pipe diameters (D_Pipe) from 4 mm to 8 mm at various time points (from 0 to 100 min). Additionally, with the use of artificial intelligence and machine learning, the best conditions for obtaining the lowest panel temperature and the highest output NFD temperature at the lowest pressure drop have been determined. While the porosity approach was used to model the PCM melt front, a two-phase mixture was used to simulate NFD flow. It was discovered that the solar panel temperature and output temperature both increased considerably between t = 0 and t = 10 min before beginning to rise at varying rates, depending on the D_Pipe. The HTR coefficient increased over time, showing similar behavior to the panel temperature. The entire PCM melted within a short time for D_Pipes of 4 and 6 mm, while a large fraction of the PCM remained un-melted for a long time for a D_Pipe of 8 mm. An increase in D_Pipe, particularly from 4 to 6 mm, reduced the maximum and average panel temperatures, leading to a lower output flow temperature. Furthermore, the increased D_Pipe reduced the HTR coefficient, with the PCM remaining un-melted for a longer time under the panel.