Algoritmo evolutivo para reconfiguração de redes e posicionamento eficiente de geradores distribuídos em sistemas de distribuição de energia elétrica

Abstract

Much of the electricity is produced far from consumer centers, requiring complex transmission and distribution systems containing long feeders to supply the demand. Therefore, losses inherent to the electrical components are associated with this transportation process. Thus, modifying the network configuration and installing distributed generators near the load centers are alternatives to reduce losses in these distribution systems. In this work, an algorithm based on the Differential Evolution optimization method is proposed to reduce the active power losses in radial power system feeders by network reconfiguration combined to the placement of Distributed Generators (DG). Both power systems reconfiguration and the positioning of DG are mixed integer nonlinear problems with the size of their sets solutions susceptible to a combinatorial explosion depending on the dimensions of the problem variables. For this reason, the metaheuristic method was adopted. In addition, an efficient data structure called the Node-Depth Encoding (NDE) is used to represent the electrical network. The function whose objective is the minimization represent the total losses in the lines, value calculated by means of the power flow by backward/forward sweep method by sum of power. The proposed algorithm was constructed in a modular way, in other words, one module is responsible for the generating units installation, and another is responsible for the generating units installation. Thus, it is possible to carry out the study and analysis of the loss reduction percentage by performing the operations separately or together. Under these circumstances, five algorithms were produced, two with the execution of the isolated operations, two with sequential operations in different orders, and the proposed iterative algorithm. All of them implemented in Matlab. Simulations were performed with two different networks widely used in the literature, containing 33 and 69 bus respectively, and the results obtained by the evolutionary method for the final state of the network were verified with the results obtained by the exhaustive search. The obtained values of the reduction of the losses and buses voltage were compared, and it was observed that the order of execution of the operations influences the performance of the reduction of the losses. Finally, the proposed iterative evolution algorithm obtained the greatest reduction of the active power losses