Comparação do desempenho dos algoritmos RF e KNN na previsão da energia gerada por quatro tecnologias fotovoltaicas
DOI:
https://doi.org/10.21712/lajer.2025.v12.n2.p77-85Keywords:
machine learning, photovoltaics, predictive modelsAbstract
Abstract: This research aims to evaluate and compare the forecasting performance of photovoltaic energy generation using the Random Forest (RF) and K-Nearest Neighbors (KNN) algorithms, based on a dataset with limited observations. The dataset, obtained from a solarimetric station located at the Federal Technological University of Paraná, includes historical series of electricity generation from monocrystalline silicon (m-Si), polycrystalline silicon (p-Si), copper indium gallium diselenide (CIGS), and cadmium telluride (CdTe) technologies, covering the period from January 2020 to December 2023. To assess the performance of the RF and KNN models, the Mean Absolute Percentage Error (MAPE) metric was used. The results showed that, for a 4-month forecasting horizon, the Random Forest model outperformed the KNN model across all technologies.
Downloads
References
Abdelsattar M, AbdelMoety A and Eldeen AE (2025) ‘Advanced machine learning techniques for predicting power generation and fault detection in solar photovoltaic systems’, Neural Computing and Applications, vol. 37. http://doi.org/10.1007/s00521-025-11035-6.
Flores JL, Márques CR, Castro ER and Ortega JMP (2024) ‘Solar photovoltaic panel production in Mexico. A novel machine learning approach’, Environmental Research, vol. 29. http://doi.org/10.1016/j.envres.2023.118047.
Fraunhofer I (2023) ‘Photovoltaics Report’, Disponível em: www.ise.fraunhofer.de/content/dam/ise/de/documents/publications/studies/Photovoltaics-Report.pdf. (Acesso em: 18 mai. 2025).
Konstantinou M, Peratikou S and Charalambide AG (2021) Solar photovoltaic forecasting of power output using LSTM networks’, Atmosphere, vol. 21, http://doi.org/10.3390/atmos12010124.
LABENS. ‘Energia gerada pelas estações EPESOL’, Disponível em: https://lookerstudio.google.com/reporting/d4eeef94-0967-45c488f2087e92bf81e3/page/OnwhB?s=rIzDee2c8X0. (Acesso em: 18 mai. 2025).
Leo DP, Ciocia A, Malgaroli G and Spertino F (2025) ‘Advancements and challenges in photovoltaic power forecasting: a comprehensive review’, Energies. vol. 18. http://doi.org/10.3390/en18082108.
Nakahata BY (2023) Análise de desempenho de quatro tecnologias aplicadas a sistemas fotovoltaicos conectados a rede elétrica instalados em diferentes campus da UTFPR no Paraná. Dissertação de Mestrado, Universidade Tecnológica Federal do Paraná, Curitiba.
Palacios LEO, Guerrero VB and Ordones H (2022) ‘Machine Learning for Solar Resource Assessment Using Satellite Images’, Energies, vol. 15. https://doi.org/10.3390/en15113985.
Polo ANB and Santos JAA (2025) ‘Previsão da potência elétrica de um sistema fotovoltaico: uma análise comparativa entre modelos de aprendizado de máquina’, Revista FSA. vol. 22. http://doi.org/10.12819/2025.22.3.9.
Santos JA and Chaucoski, Y (2020) ‘Previsão do consumo de energia elétrica na região sudeste: um estudo de caso usando SARIMA e LSTM’, Revista CEREUS, vol. 12. https://10.18605/21757275/cereus.v12n4p93-104.
Santos JAA (2021) ‘Aplicação de Redes Neurais na Previsão de Custos de Produção de Frangos de Corte no Estado do Paraná’, Revista FSA, vol. 18. http://doi.org/10.12819/2021.18.6.9.
Shams MY, Elshewey AM, El-Kenawy EM, Ibrahim A, Talaat FM and Tarek, Z (2024) ‘Water quality prediction using machine learning models based on grid search method’, Multimedia Tools and Applications, vol. 83. http://doi.org/10.1007/s11042-023-16737-4.
Srinivas P and Katarya R (2022) ‘hyOPTXg: OPTUNA hyper-parameter optimization framework for predicting cardiovascular disease using XGBoost’, Biomedical Signal Processing and Control, vol. 73, http://doi.org/10.1016/j.bspc.2021.103456.
Suyambu MR, Vishwakarma PK and Gupta S (2025) ‘Predicting Solar Power Based on Machine Learning Approaches and Their Efficacy’, Disponível em: https://ieeexplore.ieee.org/document/9917846. (Acesso em: 20 jun. 2025).
Villar A, Robledo C and Andrade V ‘Supervised machine learning algorithms for predicting student dropout and academic success: a comparative study’, Discover Artificial Intelligence. vol. 4, http://doi.org/10.1007/s44163-023-00079-z.
Villalva MG and Gazoli, JR (2012) ‘Energia solar fotovoltaica - sistemas conectados à rede elétrica’. Revista o Setor Elétrico, Disponível em: https://www.osetoreletrico.com.br/capitulo-x-energia-solar-fotovoltaica-sistemas-conectados-a-rede-eletrica/. (Acesso em: 20 jun. 2025).
Vrablecova P, Ezzeddine AB and Rozinajová V (2018) ‘Smart grid load forecasting using online support vector regression’, Comput. Electr. Eng., vol. 65. https://doi.org/10.1016/j.compeleceng.2017.07.006.
Zhou M, Li X and Zhong Z (2023) ‘Research on Photovoltaic Power Generation Power Prediction Based on RF-XGBoost Model’, Disponível em: https://ieeexplore.ieee.org/document/10393154. (Acesso em: 20 jun. 2025).
Downloads
Published
Issue
Section
License
Copyright (c) 2025 Latin American Journal of Energy Research
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
O autor, no ato da submissão do artigo, transfere o direito autoral ao periódico.
