Localização de falta de alta impedância em sistemas de distribuição: uma abordagem no domínio da frequência e estimação de parâmetros através do MMQP

Abstract

High impedance fault location still represents a challenge for protection engineers of electric power utilities. Such a challenge remains a very critical point and of great concern to utilities, as this type of fault can cause serious damage. This work presents an analytical formulation for the location of high impedance faults in power distribution systems, based on the estimation of the system parameters using the voltage and current signals recorded in the substation bus by the digital protection relay. The proposed method is based on models in the frequency domain, considering the representation of the behavior of the high impedance fault through DC sources, antiparallel diodes, and an electric arc resistance. The proposed model is an overdetermined algebraic system formed by nonlinear equations, in which the solution is obtained through a weighted least squares estimator. A control variable for residual analysis is proposed and applied for the selection of the best estimates. The proposed method considers in the mathematical formulation the capacitive effect of the aerial distribution networks. The proposed technique for identifying the faulted section is based on the behavior of the distance estimates of the fault, where through the calculated coefficient of determination (R2), the correct section is identified based on the low adherence of the estimates to the coefficient. The validation is performed through numerical simulations using the 13-bus and 34-bus IEEE test systems, a real distribution system of the RGE Sul Power Utility and a real high impedance fault recorded from a digital relay. Comparative test results with state-of-the-art models highlight the accuracy of the proposed technique, indicating potential for application in real systems. Mean errors below 0:500% were obtained in the estimation of the fault distance considering the presence of phase-to-ground linear faults, while for high impedance faults the mean error was below 1:678%. The variability in high impedance faults and the effects of line capacitances in the system modeling and applied mathematical equations were considered.