Previsão de potência solar utilizando métodos de simulação de Monte Carlo, K-Means e SVM: Um estudo de caso na região metropolitana de Vitória-ES

Authors

DOI:

https://doi.org/10.22279/navus.v16.2116

Abstract

The study presents an approach to forecast the power generated by photovoltaic systems, taking into account the variability of cloud cover. Using a combination of Monte Carlo Simulation, the K-Means clustering method, and predictions with the SVM tool, solar radiation data from the metropolitan region of Vitória-ES were analyzed. The models proved to be effective, producing accurate forecasts with low error rates even under varying weather conditions. The results show that the use of real data and appropriate segmentation of weather scenarios allow for a better understanding of how solar radiation variation affects energy generation. This approach can significantly contribute to solar energy planning and optimization, highlighting the need for future studies that integrate new machine learning techniques to further improve forecasts.

Downloads

Download data is not yet available.

Author Biographies

Caroline Tedesco Santos, Universidade Federal do Espírito Santo (UFES)

Mestre em Energia.

Ivaniel Fôro Maia, Instituto Capixaba de Pesquisa, Assistência Técnica e Extensão Rural (INCAPER)

Especialista em Gestão de Emergências em Saúde Pública.

Augusto César Rueda Medina, Universidade Federal do Espírito Santo (UFES)

Doutor em Engenharia Elétrica.

Jussara Farias Fardin, Universidade Federal do Espírito Santo (UFES)

Doutora em Engenharia Elétrica.

Published

2025-10-02

Issue

Section

Articles