Estimating Power Outputs of Polycrystalline Silicon PV Modules Using Neuronal Appraoch: A Case Study in Arid Environment

Precise power output estimates are crucial to ensuring the system operates at its highest efficiency, maximizing energy production, and optimizing return on investment. Predicting PV module power involves determining the amount of power a PV module will generate under various weather and operational circumstances. This paper's primary objective is the development of a highly accurate Artificial Neural Network (ANN) model for estimating the power output of photovoltaic (PV) modules. To attain this objective, data specific to polycrystalline silicon PV module technology was employed. The dataset was subsequently partitioned into two sub-databases: one tailored for sunny days and the other for cloudy days, taking into account the diverse environmental conditions. The output of the ANN model was validated using a subset of selected days for each sub-database category. The statistical scores show an nRMSE of approximately 5.06% for cloudy days and 1.95% for sunny days. Additionally, the correlation coefficient consistently surpasses 99% for the tested module data under both weather conditions. These results underscore a strong and reliable agreement between the predicted and actual power outputs.