CLASSIFICATION OF POWER QUALITY CONSIDERING VOLTAGE SAGS IN DISTRIBUTION SYSTEMS USING KDD PROCESS

In this paper, we propose a methodology to classify Power Quality (PQ) in distribution systems based on voltage sags. The methodology uses the KDD process (Knowledge Discovery in Databases) in order to establish a quality level to be printed in labels. The methodology was applied to feeders on a substation located in Curitiba, Paraná, Brazil, considering attributes such as sag length (remnant voltage), duration and frequency (number of occurrences on a given period of time). On the Data Mining Stage (the main stage on KDD Process), three different techniques were used, in a comparative way, for pattern recognition, in order to achieve the quality classification for the feeders: Artificial Neural Networks (ANN); Support Vector Machines (SVM) and Genetic Algorithms (GA). By printing a label with quality level information, utilities companies (power concessionaires) can get better organized for mitigation procedures by establishing clear targets. Moreover, the same way costumers already receive information regarding PQ based on interruptions, they will also be able to receive information based on voltage sags.

PQ classification; KDD Process; Artificial Neural Networks; Support Vector Machines; Genetic Algorithms; Quality Label