Machine learning-enhanced Lambert W modeling of TiO2 nanowire/Al-doped CdS QDs photovoltaic cells

Abstract

This work investigates and compares the performance of photovoltaic (PV) titanium dioxide (TiO2) nanowire/aluminum (Al)-doped cadmium sulfide (CdS) with TiO2/CdS quantum dots (QDs) non-doped based PV-cells, where the TiO2 is used as a photoanode. The parameters (power and capacitance versus voltage, and conductance) of TiO2 nanowire/Al-doped CdS were examined. The TiO2 nanowire/Al-doped CdS enhanced the current density in the PV-cells. The TiO2 nanowire promotes electron transfer to the CdS, while adding an aluminum dopant can generate more charge, increasing the electron-hole pair density by absorbing white light. The experimental relationships between current-power densities versus voltage for the PV-cells were developed using the Lambert function for modeling and improved by machine learning. This study presents the application of artificial neural networks (ANNs) for predicting the current-voltage (I-V) characteristics of photovoltaic (PV) cells. The series and shunt resistances of the TiO2 nanowire/Al-doped CdS and TiO2/CdS quantum dots (QDs) based photovoltaic cells were determined using the Lambert equation. The results indicate that the TiO2 nanowire/Al-doped CdS configuration reduced both series and shunt resistances, contributing to increased current output compared to the TiO2/CdS QDs-based cells.