Eventos Anais de eventos
ENCIT 2020
18th Brazilian Congress of Thermal Sciences and Engineering
A COMPARATIVE STUDY OF FORECASTING TECHNIQUES FOR WIND ENERGY GENERATION
Submission Author:
Naylene Fraccanabbia , PR , Brazil
Co-Authors:
Naylene Fraccanabbia, VINICIUS MARTINS DUARTE, Sinvaldo Rodrigues Moreno, Gilberto Reynoso Meza, Viviana Mariani
Presenter: VINICIUS MARTINS DUARTE
doi://10.26678/ABCM.ENCIT2020.CIT20-0467
Abstract
Alternative energy sources are becoming more and more frequent, aiming to reduce environmental pollution, besides being ideal to overcome the energy crisis. Wind energy is renewable and clean, besides being a source of energy that is permanently available to human beings, that is, it is inexhaustible and occupies the fourth place in the national electric energy matrix. Due to the high level of uncertainty of the factors that directly interfere in the generation of wind energy, such as wind speed, for example, make wind energy predictions with high precision is a big challenge. Therefore, the objective of this article is to develop a model of forecasting through time series that makes it possible to forecast wind energy production. For the development of this comparison, the Auto Regressive with eXogenous Input (ARX) model combined with forecasting models were compared using the error performance measures absolute mean percentage (MAPE) and determination coefficient (R²). Finally, it is noted that for the training data the Support Vector Machines with Radial Kernel (SVM-RK) and Decision Trees Regression (DTR) models presented the best results, and for the validation data, the SVM-RK model presented the best results.
Keywords
wind energy, Forecasting, ARX topology, machine learning, Time Series
DOWNLOAD PDF VIEW PRESENTATION
