Voltage sag state estimation based on l_1-norm minimization methods in radial electric power distribution system

Abstract

Voltage sags have a high impact on the proper equipment operation and the electric power end-user processes continuity. Economic losses are a growing problem for the electric utilities, regulators and electric energy final customers and therefore, the formulation of new mathematical methods for voltage sags diagnosis are needed. In this sense, the state estimation methods seek the determination of the frequency or the number of voltage sags that an end-user would experience. In this research area, optimization problems based on techniques such as singular value decomposition, voltage profile curve fitting and voltage sag source location have been formulated. The results of these approaches may be inaccurate when the pre-fault currents, non-zero fault impedances and unbalanced conditions are considered. We will evidence that the results from singular value decomposition method are inaccurate considering these real fault conditions. Also, a new mathematical formulation of the voltage sag state estimation problem based on ℓ1-norm minimization is proposed in this work. The proposed method is applied and validated to the IEEE 33-node test distribution network. Voltage sags caused for network faults are only considered. The results validate a remarkable improvement in comparison with the singular value decomposition method and show an innovative tool for voltage sags state estimation in radial electric power distribution systems