Experimental Validation of Empirical Models for Forecasting Photovoltaic Power Output: An Update and Case Study in Tangier, Morocco

Abstract

This study provides a comprehensive evaluation of eighteen empirical models commonly used for predicting photovoltaic (PV) power output. These models were thoroughly validated, and their accuracy was carefully assessed using experimental data from the PV installation at the Faculty of Science and Technology in Tangier, Morocco. The results demonstrate a strong correlation between the measured and predicted values, confirming the overall reliability of the models. Among them, the model developed by Sandia National Laboratories [1] proved to be the most suitable, delivering the most accurate power output predictions across various climatic conditions, with a normalized root mean square error (nRMSE) of approximately 7.112% during the six-month monitoring period. This outstanding performance emphasizes the model's robustness and applicability in different weather scenarios. The research not only highlights the effectiveness of the Sandia model but also provides valuable insights into the comparative performance of various empirical models, offering guidance in selecting the best tools for PV power forecasting.