An integrated monitoring system and automatic data analysis to correlate lightning activity and faults on distribution networks

Abstract

Overhead power distribution lines are usually subject to voltage transients during thunderstorms. Lightning strokes, whether direct or indirect, are among the main causes of such power quality disturbances. In order to assist the analysis of lightning effects on power quality indices, a monitoring network and an integrated database analysis, including voltage waveform recorders, local electric field measurements and information from lightning location systems (LLS) is presented. Based on the results of the monitoring systems, a data series of lightning-correlated network faults collected over a period of 15 months within the area covered by the monitoring system was obtained. In order to identify whether a lightning occurrence could have caused a network fault, two pattern recognition methods based on novelty detection approaches have been implemented and compared using the collected data: Parzen density estimator (PDE) and support vector data description (SVDD). The SVDD method is able to identify the impacts of a given lightning occurrence with a 85% success rate when examples from non-related-fault strokes are included in the training procedure. With the proposed data-based approach, it is possible to correlate faults and lightning events with reduced computational effort, include positive discharges in the correlation process, and comprise general features (e.g. soil, slopes etc.) of the region under analysis. These can be highlighted as the main advantages of the proposed procedure in comparison with typically used Monte Carlo-based methods.