LWPET: a novel framework based on Legendre multi-wavelet patch embedding transformer for partial discharge detection

Research on characteristics of acoustic signal of typical partial discharge models

Partial discharge (PD) detection is crucial for ensuring the reliability and longevity of high-voltage insulation systems, as undetected PD can lead to catastrophic failures. This paper presents a novel hybrid detection system integrating Convolutional Neural Networks (CNN) using the ResNet architecture for feature extraction and k-Nearest Neighbors (KNN) for classification. The proposed model achieves an impressive 98.79% accuracy, 99.41% precision, and 98.74% F1-score, significantly outperforming traditional methods such as SVM and CNN-SVM hybrids. The system demonstrates robust noise resistance, maintaining a detection rate above 98%, even in adverse operational conditions. Experimental validation on both synthetic and real-world PD datasets confirms the model’s scalability, computational efficiency, and ability to classify PD signals with minimal false positives and negatives. Additionally, it quantifies critical PD characteristics, achieving a mean true charge of 19.73 pC and a mean apparent charge of 3.15 pC, enabling improved insulation diagnostics. This work establishes a new benchmark in PD detection by addressing key challenges such as noise interference, real-time monitoring, and fault classification accuracy, paving the way for industrial deployment in high-voltage equipment monitoring systems. Future research will be directed toward extending the applicability of the model to other types of electrical insulation faults, enhancing its real-time performance, and integrating it into industrial monitoring systems for comprehensive fault diagnostics. These advancements position the proposed hybrid model as a pivotal tool for improving the safety and reliability of high-voltage insulation systems.

Partial discharge detection on power equipment using a magneto-resistive sensor

This paper presents a novel self-sensing partial discharge (PD) detection method for power transformers that leverages the transformer's iron core as an ultra-high frequency (UHF) sensor. The method aims to address the limitations of traditional PD detection techniques, which either require intrusive internal sensors or are susceptible to external interference. The experiments conducted on a 220kV transformer model with a suspension discharge defect model demonstrate the high detection sensitivity and anti-interference performance of the proposed method. The results show that the iron core can effectively sense PD UHF signals, providing a reliable and practical approach for on-site PD detection in transformers. This PD detection method solves the problem that the internal UHF sensor has security risks while the external sensor is susceptible to interference. It provides a new method for PD detection of transformers in field.

Partial Discharge (PD) detection and its estimation is a vital event for all the power equipment. Acoustic sensor based PD detection is more advantageous than the other available PD measurement technique. Acoustic PD detection technique has noise immunity ability to perform online partial discharge detection. In this work, a real life transformer model (RLTM) has been developed in the laboratory. Experimental procedure using acoustic sensor has been adopted to generate PD within RLTM. Several acoustic PD data have been acquired through acoustic sensors by placing the artificially created PD source at various locations inside the RL TM. Each of the sensor readings can be thought of as a waveform (time-series data) whose structure changes depending on the fault location. The novelty of the work lies in the fact that a number of features have been extracted using the Time Series Feature Extraction on Basis of Scalable Hypothesis tests (tsfresh- A Python Package) from the acoustic partial discharge data followed by classification using Random Forest classifier. Feature Selection is executed using Mutual Information Method. Percentage of identification accuracy for individual and coupled partial discharge sources are 94.4% and 91.5% respectively

Capacitor voltage transformers (CVTs) composed of aluminum foil/oil-impregnated paper insulation system have been operated over 30 years in substations. For the condition monitoring and diagnosis of aged CVTs, partial discharge (PD) detection is expected. However, PD characteristics from PD inception and progress to breakdown (BD) have not yet been clarified. Thus, PD characteristics of a fundamental insulation model composed of aluminum foil/oil-impregnated paper insulation systems are discussed in this paper. Experimental results under a fixed applied voltage revealed that PD characteristics can be classified into 3 periods. In Period 1, PD signals were intermittently generated mainly around the peak phase of applied voltage. In Period 2, PD signals were continuously generated around zero crossing (void discharge) to peak phase. In Period 3, PD signals larger than 10,000 pC were generated. Together with the above electrical PD measurement, the optical measurement of PD light emission images, bubble behavior, and their time transmission were also carried out in order to discuss the mechanism of PD progress.

This paper describes an on-line partial discharge (PD) detection technique in power cables. The technique uses a digital spectrum analyzer with a high-frequency preamplifier and several types of high-frequency inductive sensors. In extruded and impregnated cables, the PD detection is made in the VHF range. The suitability and the sensitivity of the VHF technique were checked against two other conventional PD detecting methods (Pulse Phase Analyzer and Fast Digital Scope) by performing PD measurements in EPR-insulated cable under laboratory-controlled conditions. The suitability and sensitivity of the VHF method were as good as the two conventional PD detection method. The VHF method runs used to conduct on-site PD measurement in several power cables. These measurements were conducted while the cables were in service. Special care was given to distinguish between external electromagnetic noise and electromagnetic signals produced by the PD activities inside the cable insulation.

<div class="section abstract"><div class="htmlview paragraph">Partial discharge (PD) detection has been always a fundamental tool, potentially, for the design, quality control, commissioning, and reliability monitoring for the of insulation systems. The word “potentially” stems from the objective consideration of the intrinsic limitations suffered by the existing partial discharge, PD, measurement technologies, especially the need of experts to interpret results and the lack of clear correlation between PD-related quantities, and the condition-based maintenance approach. On the whole, a thorough revision of insulation systems design procedures and of the tools to evaluate aging and failure risk is needed, especially in components of electrical assets which are critical in terms of reliability, resilience, and safety. This paper focuses on critical asset components, such as ships, aircrafts, aerospace, and any type of vehicles, where the coming electrification is significantly increasing nominal voltage, power density and efficiency, and where power electronics is the operative core. A new design approach, PD-free and reliability redundant, is briefly discussed, highlighting the importance to develop an innovative PD detection and analytics approach where measurements and interpretation are fully automatic and unsupervised, and experts are replaced by artificial intelligence. The aim is to remove the adverb “potentially” applied to the usefulness of PD detection and monitoring. The application of this innovative PD technology to measurements on dedicated test objects, able to incept either internal or surface discharges, or both, show how the capability to identify automatically the source typology of PD can help both in insulation system design and in devising health-condition monitoring systems for the asset components.</div></div>

Partial discharge (PD) measurements are universally accepted as an effective and reliable tool for on-line condition monitoring of high voltage (HV) equipment. A relatively new application is the monitoring of falling/leaning trees on the covered-conductor (CC) overhead distribution lines. Rogowski coil is used as a sensor for the detection of PDs produced due to falling trees on the CC overhead lines. This sensor is non-intrusive and superior to the conventional PD detector methods. In the next stage, this wired sensor will be converted into a wireless one. This paper gives the brief introduction to the concept and methodology of PD detection and sensor characteristics for localization of falling trees on the CC lines.

Partial discharge (PD) pattern identification is an important tool in high‐voltage (HV) insulation diagnosis of power systems. Based on an extension method, a PD identification system for HV power transformers is proposed in this paper. A PD detector is used to measure the raw three‐dimensional (3D) PD patterns of epoxy resin power transformers using an L sensor, according to which two fractal features (the fractal dimension and the lacunarity) and the mean discharge are extracted as critical PD features that form the cluster domains of defect types. The matter–element models of the PD defect types are then built according to the PD features derived from practical experimental results. The PD defect type can be directly identified by the correlation degrees between a tested pattern and the matter–element models. To demonstrate the effectiveness of the PD features extraction and the extension method, the identification ability is investigated on 144 sets of field‐test PD patterns of epoxy resin power transformers. Compared with a multilayer neural network and K‐means methods, the results show that a high accuracy together with a high tolerance in the presence of noise interference is reached by use of the extension method.

Six models of typical insulation defects are designed and lab-simulated system used for partial discharge (PD) detection of cables and accessories has been constructed to study PD characteristics of internal insulation defects in 110 kV XLPE cable system. By employing CPDM, a kind of on-line PD detector, test data of various models including related PD diagrams are obtained for our research. Based on these characteristic parameters, discharge characteristic analyses of a single PD pulse are presented and statistical operators can be obtained by numerical calculation. The experimental results show that CPDM on-line detecting method is effective for detecting PD signals from internal defects of cable accessories. In addition, there are significant differences in the some statistical parameters (e.g. discharge phase, discharge repetition rate, etc) and characteristic parameters of time-domain PD waveform which can be used as reference points for PD patterns identification in future research.

Wavelet-based de-noising of on-line PD signals captured by Pearson coil in covered-conductor overhead distribution networks

UHF is one of the effective methods which could detect partial discharge in GIS. However, more and more GIS have insulator of metal enclosed structure which leads UHF electromagnetic wave caused by partial discharge could not pass through the insulator. Fortunately, there are little slits between the metal band and the GIS shell which maybe lets higher frequency electromagnetic wave pass. It is possible to detect the partial discharge for GIS with metal enclosed insulator structure by higher frequency. In this paper, we measured the GIS equipment in the laboratory, set up the 1:1 model of GIS equipment based on the measured data. By the simulation software, we simulated the GIS partial discharge by giving a Gaussian pulse signal which is similar to the partial discharge, got the electromagnetic wave spectrum which can spread through the slits between the metal band and the GIS shell. On the basis of the data of the electromagnetic wave spectrum which can be detected, we designed an antenna which could be used for the detection of GIS with metal enclosed insulator structure. With usage of HFSS electromagnetic simulation software, we built up models of two kinds of sensors and repeatedly changed the size of the antenna to the optimal structure according to the simulation optimization data. At last, we got the standing wave ratio direction, radiation pattern, and 3D polar plot of the two sensors. Considering factors such as structural features and portable use, we compared the two kinds of antenna performance difference. The results show that the electromagnetic wave spectrum which can be detected of GIS partial discharge is 0.5-4 GHz, and in this band, the micro-strip antenna is more suitable than equiangular spiral antenna in the detection of partial discharge in GIS.

This paper proposes two alternative methods for online partial discharge (PD) detection and localization of a 6.2/7 kV XLPE insulated medium voltage (MV) power cable. The artificial defect as surface discharge was made in the middle of the cable at the joint connecting between two of 10 m length cables with 50 sq.mm diameter. Therefore, the total length of the cable is 20 m. In addition, the corona discharge was also made at the end of cable termination. The PD measurement was performed under uncontrollable noises with limitations of cable characteristics. UHF and HFCT sensors together with their supplemented techniques as Trending method for UHF sensor as well as Time Domain Reflectometry (TDR) technique for HFCT sensor are applied in the experiment in order to identify and locate the PD types. The results for PD identification and localization of both UHF and HFCT sensors are satisfied.

Acoustic partial discharge (PD) detection method has been developed as an important non-destructive testing technique adopted for monitoring the material processes and for testing, detection, and characterization of any defect in an electrical apparatus like power transformers. This method has advantages like the placement of several sensors on the outside of the tank allows a localization of PD source and measurements can be performed on-line and on-site. The paper describes a combination of UHF/electrical and acoustic method because combination of methods more increases plausibility of PD activity than pure acoustic method. This method can find PD sources with higher sensitivity especially for hidden PD defects. An experimental validation has been done on the transformer insulation model using an acoustic sensor at first in this paper. The study was complemented with measurements on transformers in test field.