Instalação estratégica de medidores de qualidade de energia elétrica para identificação de variações de tensão de curta duração causadas por curtos circuitos.

Abstract

A method for strategic power quality (PQ) meter placement capable of identifying short duration voltage variations (SDVV) caused by short-circuits is proposed in this thesis. The technique is divided into two main parts: offline, with the choice of the number and location of the PQ meters; online, with the estimation of voltage at the non-monitored buses and search space reduction of the short-circuit location. The technique is developed and validated in a database built automatically, using a test system and considering the fault type, fault location, fault resistance, and loading. The remaining voltages at each of the buses are obtained by a detection and classification methodology via redundant discrete wavelet transform (RDWT), which is validated using real oscillographic records from a power quality meter and simulated signals. The strategic meter placement considers, in addition to the installation cost, the presence of symmetric conditions and uniquely identifiable events for the formulation of a new objective function, which presents a weight factor for each of the three parameters that compose it. The restriction of the problem includes the observability matrix, obtained by the database built and the problem is solved via the binary particle swarm optimization (BPSO) method. The method validation is performed from stochastic simulations, using the Monte Carlo method in a set of short-circuit scenarios that represent one year of simulations. The method presented good results, showing fast detection responses for the most diverse types of SDVV. For estimation, the method was able to estimate voltages at the non-monitored buses with mean errors of 0.71%, mean reduction of the search space to 16% of the feeder, and a success rate of 97.55% for a solution with 5 meters for the IEEE 34-bus test system.