This paper discusses the application of the first peak and cumulative energy methods to calculate the time difference of arrival (TDOA) of UHF signals emitted by a partial discharge source in order to locate it. We used a 200 × 200 × 97 cm3 dimensional arrangement composed of four UHF printed monopole antennas, which was bio-inspired by the Jatropha mollissima (Pohl) Baill leaf. In addition, an oil cell composed of a pair of immersed flat-tip electrodes was used as a source of partial discharge (PD). The oil cell was positioned at four different coordinates in order to determine the efficiency of source localization based on the signals detected by the antennas and the applied methods. Finally, we evaluated the effects of the addition of noise filtering techniques and the sampling error of the measuring instrument on the localization results. Both TDOA extraction methods converged to satisfactory location results considering the dimensional arrangement evaluated, resulting in absolute errors of 7.20 cm up to 9.00 cm and 5.40 cm up to 10.96 cm, regarding the real PD position, for the cumulative energy and first peak methods, respectively.

This paper addresses the critical challenge of detecting, separating, and classifying partial discharges in substations. It proposes two solutions: the first involves developing a signal conditioning system to reduce the sampling requirements for PD detection and increase the signal-to-noise ratio. The second approach uses machine learning techniques to separate and classify PD based on features extracted from the conditioned signal. Three clustering algorithms (K-means, Gaussian Mixture Model (GMM), and Mean-shift) and the Support Vector Machine (SVM) method were used for signal separation and classification. The proposed system effectively reduced high-frequency components up to 50 MHz, improved the signal-to-noise ratio, and effectively separated different sources of partial discharges without losing relevant information. An accuracy of up to 93% was achieved in classifying the partial discharge sources. The successful implementation of the signal conditioning system and the machine learning-based signal separation approach opens avenues for more economical, scalable, and reliable PD monitoring systems.

This paper presents an evaluation of the effectiveness of envelope detection in detecting partial discharge signals in ultra high frequency range using a commercial circuit. Experimental tests were conducted on a defective potential transformer with a voltage class of 36 kV. The objective is to demonstrate that the approach is viable and simplifies the hardware required for signal sampling. The results show that the circuit was able to detect the signal envelope, preserving essential informations with a reduced sampling rate. This indicates that the approach is promising and can help minimize costs associated with instrumentation systems used for radiometric monitoring of partial discharges in high-voltage equipment.

Partial discharge (PD) detection in the UHF range is valued because of its high sensitivity, noninvasive nature, and strong immunity to substation noise. However, the requirement for high sampling rates substantially increases application costs, especially for sensor networks and real-time monitoring. This paper presents the development of a compact and low-cost electronic device for monitoring PD sources in high-voltage substations. The system integrates analog stages for amplification, envelope detection, filtering, and digitization, combined with an FPGA-based data acquisition interface. Unlike conventional methods, it operates independently of oscilloscopes, thus enabling long-term deployment in substations. The experimental results under controlled conditions demonstrated that the acquired signals could be distinguished based on energy, kurtosis, and crest factor, indicating the potential to classify PD sources with an average accuracy of 96% using a Random Forest model. Furthermore, the heat maps generated from multiple measurement points enabled the identification of regions with higher irradiation. These results suggest the viability of the proposed device as a practical, low-cost, and scalable solution for continuous substation monitoring and future sensor network applications.