Predição da Estabilidade Transitória em Sistemas Elétricos de Potência: Uma Avaliação do Conjunto Mínimo de Ciclos de Medição

Abstract

In an electric power system, the assessment and prediction of dynamic security are essential to avoid interruptions in the supply of electricity to consumers, in addition to ensuring that the system operates reliably. Regarding the transient safety assessment, predicting the behavior of the electrical power system in the post-fault period within a short time interval is essential so that preventive and, if necessary, corrective actions are taken. In view of this, the present work proposes to develop an approach based on artificial neural networks, multilayer perceptron, for the prediction of transient stability in an electrical power system through measurements that can be easily acquired by phasor measurement units, such as magnitude and voltage angle of the buses. In the proposed structure, the smallest number of measurement cycles necessary to accurately assess the state of the grid was investigated. The results obtained in the IEEE 68 bus system show the efficiency of the predictor that obtains an accuracy of 97.5% in the classification from 6 consecutive measurement cycles of the system response. When only 1 measurement cycle was provided to the predictor, a high accuracy was also achieved (96.1%) by the proposed methodology.