Feature Selection and Classification of Field Leakage Current Waveforms using Genetic Algorithms

摘要

Leakage current monitoring is a commonly employed tool for the investigation of HVrninsulators’ performance. Several techniques have been applied on in order to extract activityrnindicating information from leakage current waveforms. However, a fully representative valuernis yet to be defined. A recent approach to cope with this problem is to classify waveformsrnusing a feature set comprising from commonly used values from time and frequency domain,rnand different feature selection and pattern recognition algorithms for the classification. In thisrnpaper, Genetic Algorithms are employed to perform both feature selection and classification.rnEleven features were selected (out of the original twenty) and the accuracy achieved rangedrnfrom 82.2% to 88.48%, which is a considerable increase compared to earlier GA classificationrnwith no feature selection. Results are also compared to other feature selection andrnclassification techniques that were previously applied to the same data set. Results emphasizernthe importance of feature selection and showed that Genetic Algorithms may give comparablernresults to nonlinear classifiers, provided that they are also employed for feature selection.