Algoritmo heurístico construtivo aplicado ao planejamento de redes aéreas de média tensão com a alocação de geração distribuída

Abstract

The solution to distribution networks expansion planning problem seeks to establish updates in the system so that it is able to supply the future demand obeying important criteria that represent the quality in the supply. Considering that in recent years the number of distributed generation connected to the system is increasing, contributing to the solution of some problems in the operation such as the high losses, the poor quality in the energy supplied, the low reliability that can be a reality, among others, this article presents a new algorithm to be applied to expansion planning of medium voltage overhead lines and which also has the ability to establish a plan for the connection of distributed generation in the network. Thus, the algorithm operates in two steps. In the first step of operation, a new topology is established for the network, which meets the future demand and respects the technical criteria that are necessary for electricity to be delivered to consumers with quality. In this process, the problem is represented by a nonlinear mathematical model whose objective function seeks to minimize the cost of network expansion and the constraints represent the physical laws that govern the power flow and ensure that future demand will be met with quality. In this operation step, the solution to the problem is constructed in an iterative way, where in each iteration a specialized sensitivity indicator uses the information obtained through the solution of the mathematical model to aid in decision making. This step of the algorithm ends when a radial topology for the system is determined. In the second step, the algorithm performs an evaluation in the established topology to indicate the capacity and the most interesting buses for connection of the Distributed Generation, seeking the best benefit for the operation of the network. In this process, the algorithm also takes advantage of the information obtained through the nonlinear mathematical model for the evaluation. Computacional tests with the new algorithm were performed considering data from systems available in the specialized literature to evaluate their performance. The results obtained through the simulations showed that the algorithm finds excellent solutions and a good convergence time.