Projeto ótimo de redes coletoras de média tensão utilizando programação binária

Abstract

In this work, the problem of medium voltage collector networks optimal planning, with focus on wind farms, is solved using an optimization model based on binary programming. The model is taken from a relevant work in the literature and refined in this dissertation with the proposal of some contributions that aim to increase the model’s fidelity to reality and decrease the optimization algorithm execution time. The collector networks of generation plants have a great number of feasible configurations, which increases in combinatorial explosion with the amount of machines existing in the farm and the number of cable types considered, which makes impossible the analysis of all configurations in order to determine the best network topology. The proposed optimization model seeks to minimize the sum of investment cost and grid energy losses cost over a planning horizon, considering radiality, connectivity, voltage limits and cable ampacity constraints, and was tested with farms of 25, 50 and 57 wind turbines, in which competent solutions were obtained. The proposed method was validated based on tests with real plants of 25 and 57 wind turbines, in which competent solutions were obtained. The use of the improved model with contributions given by this work leads to new optimal collector network solutions for the same problem, with lower total costs and shorter optimization processing time.