Energy Production Estimation Using ANN for Monofacial and Bifacial Modules Under Atacama Desert Conditions

This study presents a model for the estimation of energy production by analyzing the energy performance of single- and bifacial modules installed at the Plataforma Solar del Desierto de Atacama (PSDA) under two mounting configurations: fixed with a 20° tilt and single-axis horizontal tracking (HSAT). Electrical parameters of passivated emitter back contact (PERC) modules were monitored, environmental and irradiance data were collected between June 2022 and September 2023. These parameters were used to train an artificial neural network (ANN) to estimate the energy output of the modules in each configuration, considering the extreme desert conditions. The model tended to overestimate energy production, particularly in the fixed-tilt installations. For bifacial modules, the model performed well, although lower prediction accuracy was noted for the fixed-tilt installation compared to the monofacial setup.