Modelo fuzzy evolutivo para detecção e classificação em tempo real de distúrbios na qualidade de energia elétrica

Abstract

Power quality disturbances lead to several drawbacks such as an increase in line and equipment voltageand consequent ohmic losses; limitation of the production capacity; higher operating temperatures, premature fails, and reduction of life expectancy of machines; malfunction of equipment; and unplanned outages. Real-time detection and classification of disturbances areof great importance for power systems. This paper proposes an evolving Takagi-Sugeno fuzzy model (eTS) framework for disturbance detection combined with a hybrid Hodrick-Prescott and Fast Fourier Transform feature selection method applied over a sliding window of voltage signals. The spike, notch, inter-harmonic, short interruption and harmonicdisturbanceswere considered. Classificationperformance in terms of the root mean squared error (RMSE) and non-dimensional error index (NDEI) have shown encouraging results. Moreover, the proposed data stream-based eTS disturbance monitoring system has shown to be able to learn new disturbance patterns automatically by online adapting the parameters and structure of fuzzy rules.