ANÁLISE DE TÉCNICAS DE INTELIGÊNCIA COMPUTACIONAL PARA ALOCAÇÃO OTIMIZADA DE CONCENTRADORES INTELIGENTES SEM FIO EM REDES ELÉTRICAS DE BAIXA TENSÃO

Abstract

As part of the modernization of distribution systems, distribution companies install wireless smart concentrators in low-voltage electrical networks to communicate with smart meters installed in final consumers. Such concentrators record information that the distribution companies can use in their planning and operation studies. However, in extensive electrical networks, the installation of these concentrators requires a high investment on the part of distribution companies. To assist in the study of optimized allocation of wireless smart concentrators, the computational intelligence techniques genetic algorithm, particle swarm and variable neighborhood search are analyzed in this work, as they present a rapid convergence within the search space. The objective function of the problem considers the minimization of the number of installed concentrators to maintain the coverage of all the nodes of the electrical network. Additionally, as the extensive low voltage feeders cover regions with different radio propagation characteristics and each wireless communication technology has a different range, a mathematical model was used that considers that each concentrator may have a different coverage radius, depending on the type of communication device used and the urban obstacles that are in its surroundings. Thus, in the optimization process, the availability of two different communication systems based on ZigBee technology was considered, according to the signal range, propagation characteristics in the study area and operation and maintenance costs.