A novel numerical method for predicting the melting rate of phase change material by combining computational fluid dynamics and artificial neural network

With the increasing demand of electricity and the high pollution emissions of fossil fuels, the utilization of renewable energy has gained increased attention. The utilization of renewable energy, characterized by intermittency and variability, shows a challenge with space-time mismatches between electricity supply and demand. Energy storage technology based on phase change material (PCM) serves as an effective technical pathway to address this issue. However, due to the long time of thermal storage process for PCM, numerical simulation based on Computational Fluid Dynamics (CFD) often facing issues such as high computational consumption and long calculation times. Therefore, to address this issue, a new predictive method based on artificial neural network (ANN) is proposed in this study. A 2D numerical model of a triplex tube heat exchanger (TTHX) has been established and validated with experiment. The melting characteristics of TTHX under different working conditions were analyzed. Moreover, an ANN model was established to used to fit the numerical results with a mean squared error (MSE) of 2.75×10^-10 and R-squared (R2) of 0.990, which is of great help for the fast prediction and the optimization of PCM melting process.