Agrupamento incremental de fluxo de dados para análise e monitoramento da qualidade de energia

Abstract

The concept of Power Quality is related to a set of changes that can occur in the electrical system. Power quality problems can be defined as problems that manifest in voltage and current signals or as variations in frequency. These result in flaws or bad consumer equipment operation. Such changes (disturbances) can occur in many parts of the power system – be it in the consumer electrical wiring or in the supply system, causing financial losses to both. Thus, real -time automatic detection and classification of disturbances, based on a large volume of data generated by monitoring equipment, is of fundamental importance. In this study, evolving intelligent models, that is, models equipped with incremental online learning algorithms capable of changing their parameters and structure according to new information that emerge from a data stream, are considered for pattern recognition and classification. In particular, an evolving Takagi-Sugeno (eTS) fuzzy model, and an evolving fuzzy set -based evolving model (FBeM) are taken into consideration. A Hodrick-Prescott filter combined with a Fast Fourier Transform technique and mean voltages are considered for pre -processing measured data and extracting variables that indicate the presence of disturbances. The models developed in this study have reached classification performance comparable to that of stateof-the-art models in the field of power quality. Detection and classification of disturbances such as voltage sag and swell, inter-harmonics, sub-harmonics, harmonics, short-term interruption, oscillatory transient, spikes and notching, possibly occurring simultaneously, were reached with an accuracy of about 85-95%. In addition, the evolving models adopted, combined with the above-mentioned pre-processing techniques, have shown to be superior in terms of computational memory and time.