On-line Partial Discharge Research Articles

Decision tree rules for insulation condition assessment of pre-molded power cable joints with artificial defects

A decision tree method is presented that can judge the initial and final stages of insulation degradation for pre-molded power cable joints with air gap and void defects. According to the partial discharge (PD) of eight experimental subjects, it is validated that the discharge phase changes toward the voltage zero-crossing area and becomes wider. This well-known phenomenon can be applied to assess the transition of insulation status by the cluster theory. Subsequently, the Gini index coefficient and information gain of the decision tree algorithm yield the insulation status diagnostic rules. The result indicates an overall recognition rate of 85% in training data. Furthermore, the proposed diagnostic method has successfully detected the alert of the final stage prior to insulation breakdown for two additional experimental subjects. These decision tree rules are feasible and economical for an online PD monitoring device to achieve condition-based maintenance in distribution power cables.

Online Partial Discharge Condition Monitoring of Complete Networks: Challenges and Solutions Explained through Case Studies for the Pulp and Paper Industry

Unplanned outages within the paper industry lead to significant costs to operators. To reduce unplanned outages and losses, condition monitoring (CM) techniques can be implemented to determine asset health and help target maintenance to where it is required most. The challenges faced by the paper industry to implement condition-based maintenance (CBM) on their medium-voltage (MV) and high-voltage (HV) networks and solutions for online partial discharge (OLPD) monitoring are due to the range and number of assets present and the geographical spread coupled with problems of retrofitting the technology on pre-existing plants. The implementation of OLPD testing or monitoring can vary based on several factors: network topology, connected assets, switchboard type, and the owner's budget and desired local engineer involvement. This article discusses different approaches to implementing an OLPD test program, ranging from simple handheld survey units to complete network monitoring. Its focus is on case studies of the deployment of OLPD testing and monitoring solutions at several U.S. paper sites. The case studies present diagnostic testing of unshielded MV cable at one paper mill up to full permanent OLPD monitoring systems at two others.

Measured and simulated time-evolution PD characteristics of typical installation defects in MV XLPE cable terminations

In condition monitoring of medium voltage (MV) power cable accessories, there are still challenges in the recognition of partial discharge (PD) defects. In the present paper, typical installation defects were deliberately introduced into MV XLPE power cable terminations. Accelerated ageing under PD activity was achieved by applying voltage at elevated frequency for 900 hours. At suitable time intervals, off-line PD measurements were conducted for each test specimen at 50 Hz test voltage. The results show that the initial PD signatures in the form of Partial Discharge Phase-Resolved-Patterns (PDPRP) are unique for each defect type. The signatures in turn evolve uniquely under continuous voltage application. A modified 3-Capacitor PD model based on the physiochemical processes in the discharge area was implemented in MATLAB Simulink™. The measured initial PD characteristics and the subsequent time-dependent changes were interpreted by comparison with the computer numerical simulation results. It is confirmed that the conductivity of the discharging surface area as well as the voltage at which the cavity breaks down, both change with ageing under PD activity, are the dominant parameters that influence the time-evolution of the PD signatures. Furthermore, the geometry and ventilation of the discharge area is a function of the installation defect type and therefore each defect type gives a unique PD signature. The ability to recognise installation defect through PD tests immediately after installation as well as during the course of on-line PD monitoring is desirable in condition monitoring of power cables. The findings strengthen the credibility of PD diagnosis as an after-installation test as well as in-service condition monitoring technique of MV power cable terminations.

A Noise Reduction Based Wavelet Denoising System for Partial Discharge Signal

Partial Discharge (PD) checking is a standout amongst the best technique for insulation condition evaluation of HV power system. PD releases inside a power transformer energize electromagnetic transients that can be identified utilizing sensors working in the ultra-high frequency band. Despite, on-line PD estimations are influenced by elevated amounts of electromagnetic interference that makes delicate PD identification exceptionally troublesome. In our proposed research present an approach which can reduce the noise level. Increase the quality of the PD signal. Here using an approximation logic which can reduce the complexity in terms of time and able to improve the quality. The proposed Double DWT able to get batter result as compares to existing PD techniques.

An embedded wide-band sensor for PD on-line monitoring of XLPE cable joint

Monitoring partial discharge (PD) activities in the joint of cross-linked polyethylene (XLPE) cables has attracted much attention due to its extensive application and high fault rate. In this paper, we designed a wide-band sensor with simple configuration, which was embedded in the outer semi-conductive layer of the cable. Based on the analysis of circuit and antenna models, the structural parameters of sensor were determined, and it was found that the effective bandwidth reached up to 500 MHz. In addition, the sensor had a quick response to the pulse with a rising time longer than 2 ns by testing its square response, as well as a linear input-output characteristic and a sensitivity of 3 pC. At last, the sensor was applied to insulation defect type identification in the cable joint. As for three common defects, both the single PD pulse and PD patterns showed distinct. These indicated that the sensor had a potential application prospect in the PD online detection for cable joint.

Partial Discharge Localization Inside Transformer Windings via Fiber-Optic Acoustic Sensor Array

Partial discharge (PD) is the main sign of the insulation deterioration; therefore, the online PD localization of the power transformers can provide the equipment degradation characteristics and has great value for the power grid. Because of some unique merits, acoustic PD detection is one of the best options for such applications. However, when discharges happen inside the windings, precise PD localization by acoustic method becomes very challenging. As the acoustic PD signal can be reverberated and distorted along its propagation, the type and number of sensors, with their installation, should be carefully studied and selected. In this paper, the acoustic wave propagation inside the transformer is explored via numerical simulation. Based on the understanding of wave propagation, a novel-structure fiber-optic acoustic sensor array is properly designed and installed into one phase winding, which is from a real 35 kV transformer. Experimental verification shows that, by using this proposed design, PD localization with less than 5-cm error can be achieved. Therefore, the novel sensor array developed in this paper, together with the installation and localization method, unveils a novel access to accurate online PD localization, especially for those PDs which happen inside transformer windings.

Partial Discharge Classification Using Deep Learning Methods—Survey of Recent Progress

This paper examines the recent advances made in the field of Deep Learning (DL) methods for the automated identification of Partial Discharges (PD). PD activity is an indication of the state and operational conditions of electrical equipment systems. There are several techniques for on-line PD measurements, but the typical classification and recognition method is made off-line and involves an expert manually extracting appropriate features from raw data and then using these to diagnose PD type and severity. Many methods have been developed over the years, so that the appropriate features expertly extracted are used as input for Machine Learning (ML) algorithms. More recently, with the developments in computation and data storage, DL methods have been used for automated features extraction and classification. Several contributions have demonstrated that Deep Neural Networks (DNN) have better accuracy than the typical ML methods providing more efficient automated identification techniques. However, improvements could be made regarding the general applicability of the method, the data acquisition, and the optimal DNN structure.

Noise Rejection and Detection of Partial Discharges Under Repetitive Impulse Supply Voltage

Partial discharge (PD) detection is a fundamental tool for the design and diagnosis of insulation systems, particularly organic materials that must operate throughout their life without PD. While the methods for PD measurements under sinusoidal supply voltage are well-established, the same does not hold under repetitive-pulse voltage waveforms, such as those provided by inverters. Electromagnetic noise generated by electronic switch commutation in inverters can overlap with PD pulses, making it difficult to measure, separate from noise, and identify PD. This is particularly more cumbersome when fast rise-time switches are employed. To address these issues, this paper investigates methods to detect PD pulses on enameled wires under repetitive impulsive voltages with different rise times of 60, 500, and 1000 ns, proposing techniques which allow separation of PD from noise even when they overlap in time and frequency domains. This paper employs a time-domain approach using standard deviation curve acquired from a set of measurements to discriminate PD from repetitive commutation disturbance. These techniques can be implemented to achieve unsupervised noise rejection, which would be an important goal for both offline and online PD testing. The proposed approach is validated successfully through tests carried out on a low-voltage motor.

Towards the Text Compression Based Feature Extraction in High Impedance Fault Detection

High impedance faults of medium voltage overhead lines with covered conductors can be identified by the presence of partial discharges. Despite it is a subject of research for more than 60 years, online partial discharges detection is always a challenge, especially in environment with heavy background noise. In this paper, a new approach for partial discharge pattern recognition is presented. All results were obtained on data, acquired from real 22 kV medium voltage overhead power line with covered conductors. The proposed method is based on a text compression algorithm and it serves as a signal similarity estimation, applied for the first time on partial discharge pattern. Its relevancy is examined by three different variations of classification model. The improvement gained on an already deployed model proves its quality.

On-line partial discharge monitoring system for underground MV cables – Part II: Detection and location

Condition based maintenance constitutes a useful tool towards upgrading aging infrastructure in the context of asset management in smart grids. Towards that aim, on-line partial discharge (PD-OL) monitoring systems can provide useful information regarding insulation degradation, that can be used to prevent destructive faults. In this study, a set of novel algorithms is proposed for utilization in such a system. More specifically, an enhanced approach is proposed to denoise partial discharge (PD) associated measurements, by the utilization of the discrete wavelet transform (DWT). Moreover, an overall process is presented for the efficient detection of peaks in denoised signals, and the respective calculation of the PD apparent charge. Furthermore, algorithms are proposed for the detection and location of PDs in practical applications, taking also into account possible pulse reflections, and disregarding the corresponding false positive detections. All proposed methods are rigorously tested, and the respective results are presented.

Research and Design of Cable Partial Discharge Online Monitoring Technology Based on GPS Clock Wireless Synchronization

A prototype of cable partial discharge online monitoring system was presented in this paper, including partial discharge sensor, partial discharge collector, GPS wireless synchronization system, centralized monitoring host and operation software. Real-time cable partial discharge monitoring data transmitted wirelessly, analyzed by platform software, and output information of time domain waveform, frequency waveform, discharge quantity, PRPD spectra of discharge number-discharge quantity-power frequency phase, locating information of partial discharge, history record and query of partial record and alarm severity of partial discharge.

Online Partial Discharge Insulation Condition Monitoring of Complete High-Voltage Networks

Renforth et al. in 2017 present a novel online partial discharge (OLPD) high-voltage (HV) insulation condition monitoring (CM) solution to monitor complete HV networks in petrochemical and other critical industry facilities. Continuous, 24/7 OLPD insulation monitoring of in-service cables, switchgear, transformers, and rotating machines operating at 3000 V and above provides operators with an early warning against incipient HV insulation faults that manifest themselves through high PD activity. Knowledge of the type, severity, and source of any site(s) of PD across the HV network (including remotely connected plant such as Ex/ATEX HV motors located in hazardous gas zones) enables pinpointed preventative maintenance interventions to be made to the HV insulation system during planned maintenance turnarounds to avoid unplanned outages. To monitor a complete HV network, OLPD sensors and monitoring nodes are distributed at strategic locations across the network. The CM data from the distributed nodes are passed via Ethernet/LAN to a partial discharge monitoring server (PDMS) located at the control center of the facility with a secure internet connection. The PDMS operates an advanced OLPD monitoring database for the logging, display, benchmarking, trending, and visualization of the PD monitoring data on a graphical user interface. The insulation condition criticality (from 0% to 100%) of each phase of all of the individual HV plant items monitored is presented on a series of mimic screens of the HV network single-line diagram. This paper concludes with a case study of a complete HV network OLPD monitoring design solution implemented at a large, crude-oil processing facility in Eurasia.

A novel adaptive wavelet threshold estimation based on hybrid particle swarm optimization for partial discharge signal denoising

Partial discharge (PD) of high voltage equipment detected by online monitoring is regarded as a key indicator of insulation status. Threshold estimation exerts an important influence on the denoising effects of PD signals after wavelet-based denoising. In order to improve the adaptive performance of wavelet-based denoising and reduce the distortion of denoised signal, an approach of hybrid particle swarm optimization (HPSO) adaptive wavelet threshold estimation (HPSOTE) is presented in this paper. To obtain the global optimum thresholds, a novel hybrid particle swarm optimization algorithm merging crossover mutation and chaos is proposed in HPSOTE which could solve the premature convergence problem of common threshold selection methods. Genetic algorithm (GA) and particle swarm optimization (PSO) algorithm are adopted to compare the effects with HPSOTE. The denoising results of simulated and actual PD signals show that HPSOTE has a faster convergence rate and more effective global optimization ability than the others and significantly improves the credibility and calculation speed. It gives better mean square error (MSE) and amplitude error performance. This HPSOTE can remove the noise effectively and has comparatively high application value in PD online monitoring for smart grid based on big data.

Research on Measurement and Evaluation of Ultra High Frequency Partial Discharge Monitoring Device

In order to evaluate the partial discharge Ultra High Frequency (UHF) detection device, the author conducted a laboratory simulation experiment on UHF Partial Discharge (PD) online monitoring system. The UHF signal is simulated by the pulse signal generated by the ns steep pulse generator. By comparing the characteristics of the generated UHF signal with the spectral features detected by the UHF partial discharge on-line monitoring system, the detection capability, pattern recognition ability, and anti-jamming capability of the UHF partial discharge on-line monitoring system are measured and evaluated. Through the research on the testing and evaluation of power equipment, it can ensure the normal operation of the partial discharge online monitoring system.

A Capacitive Strip Sensor for Detecting Partial Discharge in 110-kV XLPE Cable Joints

The joints of cross-linked poly ethylene (XLPE) power cables often suffer the frequent faults and latent defects, which can be detected effectively by partial discharge (PD). In this paper, a novel capacitive strip sensor for the PD measurement is proposed and testified under the context of prefabricated 110-kV XLPE cable joints. First, a simulation model based on finite integration technique is established. Then, the transmission mode and the frequency spectrum of the electromagnetic wave related to the PD signal are investigated. Furthermore, the energy distribution of the PD signal among the cable joint is analyzed to identify the optimal position of installation. Finally, a real capacitive strip PD detection sensor is deployed on a 110-kV XLPE cable system, and comparative tests with a standard sensor are conducted. Consistent experimental results have been observed, and a sensitivity better than 5 pC has been achieved on the proposed capacitive strip sensor, which clearly shows its potential on online PD monitoring for high voltage power cable joints.

Field Applications of Ultra High Frequency Techniques for Defect Detection in GIS

The reliable and stable operation of power apparatus is important for the development of GIS. It is important to utilize condition monitoring technologies and anticipate possible failures in advance. Many papers have been published about the partial discharge detection with UHF or X-ray in laboratory, but seldom in field application. Thus, many engineers at project sites are not familiar with the current professional diagnosis techniques. Recently, during the GIS routine data analysis obtained by partial discharge online monitoring system, it was found that the UHF monitoring signals’ developing trend of the 220 kV GIS No. 2 high-voltage side of transformer in phase C at an actual station was abnormal and needed further detection. In order to precisely investigate the problem and then guide the operation and maintenance activities, a series of professional diagnoses were conducted. Three new types of partial discharge detection and positioning methods were applied for accuracy, including UHF partial discharge detection based on multi-stage amplified signal demodulation and multiple weighted averages processing; the partial discharge detection based on the signal radiation hole of insulation disk at the ground connection; and the positioning method based on UHF-SHF. After a series of troubleshooting works, the partial discharge defects have been diagnosed, and the case can be referred in the field monitoring of GIS.

Fractal‐based autonomous partial discharge pattern recognition method for MV motors

On-line partial discharge (PD) monitoring is being increasingly adopted to improve the asset management and maintenance of medium-voltage (MV) motors. This study presents a novel method for autonomous analysis and classification of motor PD patterns in situations where a phase-reference voltage waveform is not available. The main contributions include a polar PD (PPD) pattern and a fractal theory-based autonomous PD recognition method. PPD pattern that is applied to convert the traditional phase-resolved PD pattern into a circular form addresses the lack of phase information in on-line PD monitoring system. The fractal theory is then presented in detail to address the task of discrimination of 6 kinds of single source and 15 kinds of multi-source PD patterns related to motors, as outlined in IEC 60034. The classification of known and unknown defects is calculated by a method known as centre score. Validation of the proposed method is demonstrated using data from laboratory experiments on three typical PD geometries. This study also discusses the application of the proposed techniques with 24 sets of on-site PD measurement data from 4 motors in 2 nuclear power stations. The results show that the proposed method performs effectively in recognising not only the single-source PD but also multi-source PDs.

Noise Discrimination Method for Partial Discharge Current Focused on Damped Oscillation Waveform

SUMMARYOnline partial discharge (PD) measurement is very important for insulation monitoring of high‐voltage equipment. One of the biggest challenges performing online monitoring is discrimination between PD detect signals and external noises. Experimental results using high‐frequency current transformer (CT) sensor show that the current waveforms associated with various PD signals are all damped cosine waveforms, and have unique frequency and decay times of damped oscillation. Therefore, we focused attention on the damped oscillation waveform of the PD current, especially damped cosine waveform having a different number of oscillations. In addition to this, the noise discrimination method using wavelet transform or short‐time Fourier transform is effective. Those noise discriminating processes are applicable to online PD current measurement system.

Monitoreo de Descargas Parciales En-línea en Generadores Eléctricos de Alta Tensión

Rotating electric machines with operating voltage greater than 12 [kV] are an integral part of the installed base of electric generators in different countries. These voltages incorporate new challenges to be solved in maintenance practices. One of the most important challenges is the identification of the condition of the stator winding. For many years, the most commonly used monitoring techniques in electrical generators were off-line insulation measurements, especially in direct current. More recently, the use of on-line insulation monitoring techniques has enabled problems to be detected in advance, such as partial discharge measurement. This article presents the measurement of on-line partial discharges, as a condition monitoring technique for high voltage electric generators. It is also used for risk assessment and decision support for corrective actions. In the case study, it was evidenced that periodic monitoring is key during the operating time of the evaluated generator unit, which can be extended to other generators. The monitoring of on-line partial discharges allowed to know its condition and evolution through the trend in time.Keywords: Partial discharge, electric generator, failure mode, predictive maintenance, void.

Experience With Online Partial-Discharge Measurement in High-Voltage Inverter-Fed Motors

Partial-discharge (PD) testing has long been an important tool for assessing the condition of the high-voltage insulation in motor and generator stator windings. In the past several years, many motors have been powered from inverters, which facilitate variable-speed motor operation. The most common drive used today is the voltage-source pulse-width modulation (VS-PWM) type. VS-PWM drives rated up to 13.8 kV are becoming more common in natural gas processing plants, as well as in other petrochemical facilities. Such drives generate high-voltage impulses in the kilovolt range with rise times in the submicrosecond range. These impulses are a form of severe electrical interference that can make difficult the online detection of PD (with magnitudes 1000 times smaller) due to the overlapping frequency content in PD and in the impulses. Thus, PD detection on medium-voltage VS-PWM systems has been a challenge in spite of the serious stator winding insulation aging that such drives may cause to these motors. This paper discusses the stator winding failure mechanisms that produce PD, including the insulation problems that VS-PWM drives can accelerate. A research project that lasted several years is reviewed. This project culminated in a prototype online PD monitoring system suitable for motors fed by VS-PWM drives.