Voltage Event Classification Method Based on Symmetrical Components Models for Extended ABC Classification Criterion

Abstract

The identification, analysis and mitigation of voltage events are relevant research areas in the context of smart grids. The analysis can be conducted by using predefined models that are representative of certain grid faults. One of the criteria that represents the largest number of fault types is the extended ABC, which is used as the basis of this paper. Some of the algorithms proposed in the specific bibliography that use this classification criterion have an acceptable performance under ideal conditions, but are affected by different disturbances related to the same fault or the power grid itself. Present work proposes a new algorithm that improves the classification performance evaluating the mean squared error of each event type using two different symmetrical components estimations. The algorithm is compared with different proposals reported on the bibliography and evaluated with several types of common electrical network disturbances by simulation. Results show a significant reduction of classification errors.