General Machine Learning-Based Approach to Pulse Classification for Separation of Partial Discharges and Interference

Abstract

This paper describes a complete approach to filtering Partial Discharge (PD) pulses from interference in High Voltage electrical equipment using supervised Machine Learning (ML) techniques. The PD signals are registered in Ultra High Frequency radiation band with a multisensor acquisition system composed of 4 antennae. The proposed methodology focuses on the implementation ML algorithms and proposes a novel in this field approach to the onset detection of incoming signals. The goal was to achieve high accuracy of filtering with reasonably low compilation times of the ML classifier. That would allow to use the model on edge sensor devices. In the paper, different models and training variants of the ML framework are tested. The presented results are based on a robust measurement campaign performed in laboratories of GEIRI Europe. The methodology is validated through tests on 3 separate test scenarios. Each represents a different complexity of the problem with an increasing number of active sources. The results show high potential for utilization of the ANN and other classifiers for PD filtering problems as the accuracy achieves the desired threshold of 80% for most of the tested variants. The methodology is a step forward toward a fully online PD and interference filter.

Spatial Pattern of the Seismicity Induced by Geothermal Operations at the Geysers (California) Inferred by Unsupervised Machine Learning

“…Currently, there are countless implementations of clustering algorithms and the definition of this metric can be approached from numerous different perspectives. In this research activity, we choose to proceed with Hierarchical Agglomerative Clustering (HAC), which can provide information on the similarities between objects in the data space at all levels of aggregation, giving a clear structure of the data space [51] and has been successfully applied in different research areas to identify groups of signals emitted from a common source [52]- [57].…”

Optical signals measurement of discharge plasma in air switch cabinet using a fluorescent optical fiber as the sensor

“…The reason is that fluorescent molecules have a certain lifespan after being excited. At the moment when the excitation source disappears, however, the fluorescence does not immediately disappear completely, and its intensity will decay exponentially over time [32], which can be expressed by equations (8) and (9).…”

“…Researches have shown that the type and discharge quantity can be inferred by extracting the key parameters of the PD signals, thus enabling the diagnosis of the type and severity of internal insulation defects [3][4][5][6]. However, the conventional PD detection methods, such as the ultrasound method and the ultra-high frequency (UHF) method [7][8][9][10], are not applicable to switch cabinets. On the one hand, switch cabinets have metallic walls that can impede the propagation of PD signals; one the other hand, switch cabinets have a complex structure with many components in a small space, which can lead to refraction, reflections, and significant attenuation of the signals' intensity.…”