Uma abordagem de mineração de dados para estimativa da velocidade do vento

• Ano de defesa: 2018
• Autor(a) principal: Freitas, Nicksson Ckayo Arrais de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal Rural do Semi-Árido; Brasil; Centro de Ciências Exatas e Naturais - CCEN; UFERSA; Programa de Pós-Graduação em Ciência da Computação
• Palavras-chave em Português: Recursos renováveis; Energia eólica; Velocidade do vento; Mineração de dados; Bancos de dados; Inteligência Artificial; Renewable sources; Wind energy; Wind speed; Data mining; Databases; Artificial intelligence; CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://repositorio.ufersa.edu.br/handle/prefix/918

Renewable sources are the most promising alternatives for power generation, whereas the use of fossil fuels has caused strong impacts on terrestrial ecosystems and the climate. Wind industries, as a power source, have advantages over other sources, as a consequence, wind energy generation capacity had a tremendous growth worldwide. However, energy forecasts are crucial elements for electrical system operators, because they can make better decisions on the electrical market and support operational activities. It is worth emphasizing that the output of energy from wind farms depends on the stochastic nature of the wind, which is a natural, intermittent, uncertain and difficult-to-control resource. In fact, wind speed prediction may avoid economic losses, ensure the safe and sustainable supply of electricity, facilitate regulation of wind systems, and increase the operational efficiency of industries through a more reliable decision making. Wind speed prediction is a complex and challenging problem due to the lack of appropriate tools and the events that influence wind conditions like earth moving, physical effects, and climatic factors. For proposing solutions in this context, we must consider that weather data have accumulated huge volumes of information in spatial databases, demanding the investigation of relevant means for knowledge extraction. Data mining arises as a solution to extract relevant knowledge intelligently and semi-automatically from huge datasets. This paper presents a new and low-cost data mining approach for wind speed forecasting, which incorporates relevant artificial intelligence algorithms and provides effective treatment of datasets. The approach has proven to be flexible, promising, and well-founded in two case studies carried out in Brazil. Neural networks, support vector machines, decision trees, and k-nearest neighbors are methods involved in building the diverse models for wind speed estimation