Recurrence Quantification Analysis as a Novel LC Feature Extraction Technique for the Classification of Pollution Severity on HV Insulator Model

Recently, Recurrent Plot (RP) was introduced to study Leakage Current (LC) for polluted insulator performance monitoring. Based on complex graphical representations, RP only provides a qualitative overview of the insulator state. To overcome this issue, we present in this paper a novel technique, named Recurrence Quantification Analysis (RQA) able not only to indicate RP structures, but also to quantify LC dynamics during the contamination process. RQA is introduced to investigate RP structures, quantify LC dynamics and extract features from LC waveforms for polluted insulator monitoring and performance diagnostic. For this purpose, LC acquisition is firstly carried out on a plan insulator model uniformly polluted with saline solution. Eight RQA indicators are presented to investigate LC waveforms under various pollution conductivities. Finally, mean values of RQA indicators are proposed as input for three well-known classification methods (K-Nearest Neighbors, Naive Bayes and Support Vector Machines) in order to classify the contamination severity into five classes. Results show excellent correlation between RQA indicators and the pollution severity level.