An Enhanced Optimal Design of a Phase Changing Material Based Photo Voltaic System using Deep Learning

In this paper, oscillations in heat flux are smoothed out using a PCM energy storage that is controlled by artificial neural networks (ANN). The purpose of this research is to evaluate how different levels of discharging heat flow might influence the use of PCM and ANN in various settings. We compared the standard deviations of the charging and discharging heat fluxes when they were managed by ANN and when they were managed just by PID. Investigations towards testing large-scale installations as pilot projects were carried out. The TES Unit, which had a heat capacity was fuelled by a heat flux that allowed for its intensity to be adjusted. The phase transition material in the Hitec salt was comprised of KNO3, NaNO2, and NaNO3, respectively. Sigmoid function areused in order to govern the three-layer ANN. The training procedure utilised resilient backpropagation as one of its methods. To ensure the quality of the training, compare the temperatures that were predicted with those that were actually recorded. It turned out that the prognosis was right on the money. The analysis reveals that a TES unit, in conjunction with a PCM, can be used to stabilise the changing heat flux.