Banca de QUALIFICAÇÃO: DIEGO JOSE DA SILVA

OPTIMAL ALLOCATION OF CAPACITOR BANKS IN DISTRIBUTION SYSTEMS USING A HIBRID ALGORITHM

The electric power system are composing basically into three areas, generation, transmission and distribution energy. A large part of the energy is transported by the transmission lines and deliv-ered by distributors that transport energy to consumers. It is a stage of distribution, usually occur the great amount of power losses, which is divided into technical end commercial losses. The technical losses in distribution networks consume the big part of produced energy, which can be minimized by means of the optimal allocation and size of capacitor banks, which also contribute to the control of the voltage profile. The allocation of capacitor banks in the electric power system consists of defining a number of banks, the optimal size and local to be installed that will minimize technical losses. The common problem is formulated as considered a mixed integer nonlinear programming (MINPL) prob-lem and is subject to the following constraints: load flow constraints, bus voltage constraints, current limit constraints. The problem has been solved by various techniques of optimization, which is exact and approximate methods. The current work presents an application of bio-inspired technique and an exhaustive search algorithm to solve the problem of optimal allocation and size of capacitor banks in distribution networks. A hybrid version has led to a good convergence improvement process, which coordinated a dedicated power flow, provides as allocation of capacitor banks with fast computational time even for large distribution networks. The results obtained are a financial feasibility analysis, with the purpose of relating or investing as benefits achieved. A discussion was also in consideration the technical feasibility in relation to the local to be installed each capacitor bank. The method was imple-mented in MATLAB®. As the main result, were obtained the best locations, considering the objectives and constraints. The studies performed in distribution networks available in the literature and the re-sults have been proved with other techniques found in the specialized literature.