Métodos heurísticos aplicados no ajuste de curvas de Weibull em energia eólica

• Ano de defesa: 2015
• Autor(a) principal: Azevedo, Danilo César Rodrigues
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Engenharia mecânica; Energia eólica; Métodos heurísticos; Métodos de otimização
• Link de acesso: http://www.repositorio.ufc.br/handle/riufc/13024

The application of heuristics optimization has proven quite effective when compared to traditional optimization methods by differentiation. It is known that in some cases, directly minimizing the functions involved in the process can be complicated or even impossible. This work aims to develop a Weibull curve fitting methodology, using Ant Colony Optimization method and the Particle Swarm Optimization method as far as the hybridization of these. The tipical stochastic characteristic should provide good results for any wind distribution, concentrated or dispersed, which would make it valid to use in coastal conditions, flat or complex terrain or even urban. The result obtained by the heuristic approach of two SONDA wind samples, referring to Petrolina, Pernambuco, Brazil and Sao Martinho da Serra, Rio Grande do Sul, Brazil was compared with eight other known methods and commercially applied: the least squares method, the moment method, empirical method, the maximum likelihood method, the modified maximum likelihood method, energy pattern method, equivalent energy method and the chi-squared method and the goodness of fit will be evaluated by RMSE tests, MAPE, R2, and the deviation in the forecast power density. Heuristic methods have proven competitive, with power forecast error values around 10−14%