Algoritmo para um estimador de estado trifásico linear baseado no método de soma de correntes aplicado às redes de distribuição de energia elétrica

Abstract

The way in which the energy companies have been operating their substations and supervising their feeders in the distribution networks has undergone a major modernization process in the past two decades. As the number of consumers and the physical extension of the infrastructure increased, there was a greater demand for equipment and monitoring techniques that would guarantee the reliability of these systems. This work presents the development of a new algorithm based on currents summation method, for implementation of a linear formulation of the state estimation problem in three-phase distribution networks, applied in the normal operation regime and under fault condition. In the normal operating regime, the modules of branch and node currents are estimated based on the weighted least squares method and a load adjustment algorithm, which makes it possible to obtain an estimative of technical losses, active and reactive power flows and the system voltage profile. For validation, data from two 13.8 kV real distribution systems were used to test the method. The meters located in the analyzed feeders provide the values of the measured quantities and an approximation of the average power factor of the loads located downstream from these measurement points, while pseudomeasures are used to make the system observable. In the proposed method, an iterative reestimation algorithm based on system scanning was implemented, where the parameters estimated in the first estimate are used to update the loads, branch currents and new variances. In the second stage of this thesis, an algorithm for fault location in radial DS was developed using a system sweeping method associated with the proposed state estimator. In this case, in addition to considering the short-circuit current measured at the substation, the currents of each load were estimated during the fault to verify their influence on the fault location process. In this condition, a real distribution system was modeled in the ATP software in order to emulate measurements of voltages and currents at the substation, and voltage magnitudes registered by other meters during the fault. Factors such as the influence of the fault resistance, the type of fault, the system loading and the fault location were tested. The results obtained show that the developed method is able to provide the system status with adequate precision for distribution systems with a reduced number of measurement points along the feeder, besides presenting a low computational processing time and simplified modeling. The obtained results show that the developed method can provide the state of the system with adequate precision for distribution systems with a reduced amount of measurement points along the feeder, besides presenting a low computational processing time and simplified modelling. The results obtained by the proposed estimator and the tested reestimation method showed similar errors in all the simulations performed, proving that, in the context of the analyzed networks, the load adjustment process already provides pseudo-measurement values close to the estimated values. Under a fault condition, the algorithm showed promise in its function, locating faults with a small error margin in most simulations.