Estimador de estados trifásico híbrido para sistemas de distribuição de energia elétrica: uma abordagem considerando a viabilidade técnico-econômica

Abstract

This work presents a new technically and economically viable methodology for the state estimation in electric energy distribution systems using synchrophasor measurements and RMS (root mean square) measurements together. For the proposed methodology, the currents passing through the lines of the system are used as state variables in rectangular coordinates and the measurements are obtained by sensors located along the branches of distribution feeders. RMS values of voltage, active and reactive power supplied through smart meters are considered, in addition to synchrophasor measurements at the main substation (busbar located at the beginning of the feeder) for the definition of the angular reference. The approach is three-phase and represents the intrinsic characteristics of a distribution network such as load unbalance, radial topology, existence of mutual impedances and lack of sensors dedicated to their continuous monitoring. The methodology is based on the solution of a non-linear optimization problem which aims to minimize an objective function associated with the sum of squared differences between measured values and their corresponding estimated values. Virtual measurements are considered as zero current injections for the null power injection buses, treated directly in the objective function. The formulated problem is solved at each instant of time ‘t’ in order to estimate operating states along the time, enabling the tracking of daily load profiles. Additionally, load constraints for buses that are not monitored in real time are incorporated into the problem, being determined based on load estimates obtained in a previous instant of time ‘t-1’, correlating consecutive estimation time intervals. The optimization problem is solved by the modified Safety Barrier Interior Point Method so that the solution is always determined within a feasible search space which contains the most probable operating state of the monitored system. For computational simulations, the measurements used by the proposed algorithm are emulated from a Three-Phase Power Flow via Current Injection (TPFCI). Several simulations were carried out varying the percentage of loads for analysis considering daily loads profiles and the results obtained were compared with another existing methodology in the literature and with the true values that were obtained through the TPFCI. The difference between the estimated and the true values is very small, attesting to the efficiency of the proposed method. It also presents a concept based on data centralization through an asynchronous message broker for the integration of all modules necessary for the complete operation of the state estimator in practice, making the proposal applicable and viable for monitoring unbalanced three-phase networks.