Evaluation of three-phase voltage dips classification algorithms based on computational intelligence

In this paper are evaluated three different three-phase voltage dip classification algorithms developed by the authors. All them are based in the information provided by a sequence detector developed in a previuos work. The first algorithm is a logical classifier, the second it is based on fuzzy logic, and the third algorithm is a neural network adequately trained for this purpose. This evaluation is part of a wider work, and it will be used to the potential development of a power quality monitor. The performance of the algorithms under test in the classification of voltage dips, in ideal conditions and with variations in the phase and frequency, and high harmonic content is assessed. The conclusions of this analysis are the basis for the development and implementation of some of the functions of this instrument.