Performance of Regression Techniques in Predicting Alternating Current Power (ACP) in Solar Plants: Case Study at IFPE Pesqueira

Pedrosa Filho, Manoel Henrique de Oliveira, Lima, Ricardo Ataíde

ISES Solar World Congress 2025 · Fortaleza, Brazil · 2025-11-03
Published by International Solar Energy Society (ISES)
DOI: 10.18086/swc.2025.03.54

Abstract

This study evaluates machine learning techniques for predicting Alternating Current Power (ACP) output in a photovoltaic plant at IFPE Pesqueira campus, Brazil. Models including Multiple Regression, Support Vector Regression (SVR), Decision Tree, Random Forest, and XGBoost were compared using solar irradiance, ambient temperature, and Sandia cell module temperature as inputs. All models achieved high accuracy (R^2 > 0. 8 , RM SE < 0. 17 ), with SVR demonstrating the best performance (R^2 = 0. 8517 ), albeit with a higher computational cost. Faster methods such as Decision Tree (0.0022 s forecast time) proved viable for real-time applications. This work offers practical insights for model selection based on the trade-off between accuracy and efficiency in operational scenarios.

Keywords

Photovoltaic power forecasting, Machine Learning applications, Regression models comparison

← All proceedings papers