A Study of Partial Discharge Localization in Transformer

Radio frequency (RF) detection of partial discharges (PD) is a long-established principle, being described in the international standard for partial discharge measurement. The growing market for mobile communication devices in recent years has led to big reductions in the cost of electronic hardware operating in the VHF and UHF bands, making advanced measurements of PD in these bands increasingly affordable. Some 20 years ago, ultra-high frequency methods for PD detection were pioneered in gas-insulated substations (GIS) and have since become widely accepted. Locating PD in GIS is essentially a one-dimensional problem. More recent research into PD location in power transformers, also using the UHF method, has had to address the more challenging problem of PD location in three-dimensions. Both the GIS and transformer applications monitor enclosed volumes, where the PD is within a metalclad system that traps the signal and provides good screening of external RF interference. At the same time, other researchers have developed RF detection and location systems for open environments, such as complete air-insulated substations. When deployed in a fixed installation or a surveying vehicle equipped with GPS, this system is capable of monitoring insulation defects as they develop, and provides a location capability using sophisticated signal processing. This paper begins with an overview of some aspects of radiometric monitoring techniques and goes on to outline some research challenges which it is hoped will attract the attention of the many researchers who are continuing to advance the areas of radiometric PD detection, measurement and location technology.

Detection of winding deformation and partial discharge (PD) is essential for monitoring and failure risk assessment of power transformers. Recently, an online method based on electromagnetic waves has been proposed for detecting radial deformation (RD) in high-voltage (HV) transformer winding using ultra wide band pulses, called the hyperboloid method. In this method, a short pulse is transmitted toward HV winding and its reflection captured using a set of antennas. By comparing and analyzing signals in sound and deformed conditions, the RD can be detected. On the other hand, signals radiated from PD defect in transformers are generally appeared in the range of UHF band (0.3–3-GHz frequencies). In this paper, the hyperboloid method is proposed for implementation in UHF band, and the PD effect on the result of the hyperboloid method is investigated. Thus, a set of UHF sensors can be used for the detection of PD and detection and localization of RD defect. Results show that PD signals can affect the signal of the hyperboloid method and lead to wrong localization of RD. In order to solve this problem, the UHF stepped-frequency method is proposed, utilized, and tested on the HV winding of a 1.6-mega volt ampere (MVA) distribution transformer. Experimental results show the capability of the proposed method for detecting both defects using only one set of UHF antennas.

Rapidly increasing growth and automation of power are given the opportunity to power engineers to incorporate new ideas in the field of online monitoring of power apparatus for reliable operation of electrical system. As a part of new innovation in power system, photonics play an important role for driving innovation toward the sensor development, optical data communication to imaging, lighting and displays. In any power apparatus, degradation of insulation occurs due to the accumulation of mechanical, thermal and electrical effects for a long period. As the insulation of high voltage (HV) power apparatus are deteriorated by partial discharges (PDs), detection and measurement of PDs are of utmost important for the quality assessment of HV power equipment. In this chapter, an attempt has been made to detect the PDs with the help of photonic technology which eliminates the electrical interference; a challenge is being faced by power engineers in recent decades. It is understood from the several studies, PDs of HV power apparatus can be detected by electrical, chemical and acoustic methods. Electrical detection focuses on capturing the electrical impulses created by the current streamer in the void and impurities. Electrical detection method has some limitations. The primary limitation is its susceptibility to noise. The high voltage transformer (HVT) environment contains high levels of electrical noise, both narrowband and broadband. Chemical detection has also some drawbacks. Firstly, this detection method does not give any information about the location of the PD and the insulation condition. Secondly, the main problem is that chemical tests cannot be performed continuously when the HV equipment are in working operation. To explore the detection of PDs through acoustic method, acoustic phenomenon of PDs has been studied, and it is found that it leads to the propagation of acoustic waves having frequencies typically between 30 and 1000 kHz inside the transformer. The propagation of acoustic wave modulates the refractive index of the medium, which modulates the intensity of the laser beam in sympathy with the acoustic waves. Acoustic methods have many advantages over other methods in that they are immune to electromagnetic interference (EMI), which can greatly reduce the sensitivity of electrical methods and they can be applied to find failure spot and location of discharges. Optical technique is also capable to couple PDs with the optical signal. To capture the acoustic waves, an optical method is demonstrated in this chapter for detection and localization of PD sources by using photo detector inside the HV power apparatus. In this work, a new concept such as direct optical partial discharge detection technique has been introduced, where the energy exchange mechanism of electrical to acoustic is used for online monitoring of PDs. The acoustic signals are collected directly through the photonic technology, which is free from electrical interference, and finally they are converted to electrical signals for detection of amplitude as well as location of PDs occurring inside the HV power apparatus. This is a cost-effective tool for online monitoring of PDs and can be used as a predictive maintenance tool for utilities.

ABSTRACTThis article deals with the methods of finding partial discharge (PD) location in power transformers using ultra high frequency (UHF) measurements. The UHF technique utilises two methods to find the PD location, that is, the shortest path method and hyperbolic method. The shortest path method works based on the comparison of the measured data and the ones in the database. In the hyperbolic method, a hyperbolic equation is obtained between each two element subset of sensors. The coordinate that best fits all equations is known as the PD location, and can be obtained in three different ways, that is, iterative algorithms, the Fang method and Chan method. The convergence of iterative algorithms is limited by poor initial estimate, overshoot, mitigation of non‐convergence etc. The Fang and Chan methods are two closed‐form solutions that are used in the communication system to find the radiation source location. This article explains how to use these two methods to obtain the PD coordinate inside the power transformer. These two methods can find exactly the coordinate that best fits all hyperbolic equations. At the end of this article, several tests are carried out through CST software and the PD locations is estimated by all presented methods. The simulation results show how the Fang and Chan methods can overcome the limitations of the iterative method.

Location of a partial discharge (PD) fault inside a transformer can be determined using the Ultrahigh Frequency (UHF) PD measurement technique. Hereby, the electromagnetic (EM) waves that propagate due to the PD are captured using specially designed probes that are inserted into the transformer tank through its oil drain valves. The difference in the arrival time (AT) of these signals is the base of an equation system, where the PD location and the PD inception time are unknown. In this contribution, the influence of erroneous determined ATs on the localization error is investigated using a two-dimensional analogy model. Furthermore, a novel method based on detection of abrupt changes in a certain statistical property of the signal for the AT determination is proposed. A test setup is utilized to evaluate the localization algorithm. Finally, the results are presented and the precision of the proposed method for PD localization using the UHF measurement technique is discussed.

To locate partial discharge (PD) in power transformer, a combined array sensor is manufactured which consists of a cross-shaped ultrasonic array with 25 elements and an ultra-high frequency (UHF) array with 2*2 elements. Algorithms for PD location are studied based on the sensor arrangement, and a total of 193 ultrasonic and 13 UHF virtual elements are achieved via virtual expansion. At first, two-sided correlation transformation based on the UHF virtual expanded array is employed to pre-estimate the PD’s direction, and then rotated signal sub-space based on the ultrasonic array is used to direct accurately. To enhance the accuracy and success rate of an array, ray tracing technique which consists of geometrical optics and uniform geometric theory of diffraction is studied to backtrace the wave emitted by PD, considering the UHF signal as an example. The PD location experiments in a transformer model indicate that the location algorithms based on the proposed combined array sensor adapt the PD’s broadband signal well, and ray tracing of the electromagnetic wave emitted by PD makes the location much accurate and successful than that of a single array. The PD location based the proposed combined array takes the advantages of the high sensitivity of UHF array and the high accuracy of the ultrasonic array.

We constructed an on-line partial discharge (PD) monitoring system (PDOM) with a CT sensor, an analogue-digital converter, and a tablet PC. A sampling frequency for data acquisition is 100 MHz, and we exchange on-site PD data through a cloud system. We successfully measured PDs in the early stage of insulation deterioration of a cross-linked polyethylene (XLPE) cable in a field site. Additionally, following PD detection by using a PDOM, we can locate a PD position by using a PD locator (PDL) system. By locating a PD, we can estimate the actual PD magnitude considering attenuation factor of a high frequency pulse. We show how the PDL system works on a XLPE cable network with time synchronization between two positions by using an injection current pulse. The locating accuracy of the PDL system is evaluated.

Power transformers are important nodes in the electrical power grid. Common reasons for breakdown are problems in the insulation system. Detecting partial discharge in the insulation system of a power transformer at an early stage reduces the risk of total breakdown of power apparatus. One method to detect partial discharge is the acoustic measurement. With this technique a detection and localization of partial discharge is possible by placing acoustic sensors on the surface of the transformer tank. The low level of electrical interferences from outside the measurement setup constitutes one of the strengths of this method. A further advantage is the ability of identifying the position of the partial discharge source, which is needed to estimate the risk and to enable a fast and effective repair.

One of the main methods for partial discharge (PD) source localisation in power transformers is acoustic emission measurements. This study describes a new method for detection and location of two simultaneous partial discharge sources in three‐phase power transformer. In this method, acoustic signals are detected by sensors first and are then denoised using a wavelet transform. Finally, the two PD sources are localised using the modified binary partial swarm optimisation (MBPSO) method. To prove the efficiency of the two simultaneous PD localisations, the proposed algorithm is used to localise PD sources of the arc furnace transformer at Isfahan's Mobarakeh steel company. For this purpose, the PD localisation problem converts to an optimisation problem. To prove the efficiency of the MBPSO algorithm, its performance is compared with a genetic algorithm. The PD localisation results confirm the efficiency of the proposed method for the detection and location of PD sources.

This paper reports on construction and verification of a high space resolution partial discharge (PD) location system for the next generation power module. A novel PD location method (polarity reversion method) is proposed by calculating electromagnetic wave (EMW) propagation characteristics inside the power module using finite difference time domain (FDTD) analysis. In addition, EMW signal characteristics under AC voltage and impulse voltage application are compared. Based on these results, an automatic PD location system has been successfully constructed under the impulse voltage application.

The objective of this work is to assess the efficacy and reliability of a certain technique of acoustic emission measurement for detection and location of partial discharges in power transformers. The possibility of locating a simulated partial discharge was studied in the laboratory. The experimental setup involved an arrangement of increasing complexity, starting with a simple tank filled with insulating oil, up to simulation of an unenergized transformer. It was concluded that, under careful control of experimental parameters, such location is effective. The study was complemented with observations on transformers in the field. In this case, results were less reliable. While the acoustic emission test was not sensitive to intense discharges within the internal winding of a transformer, problems in two tap-switches and one mechanical problem were detected. Considering that the literature reports doubts on the effectiveness of the technique, it was found that measurement of acoustic emission is useful for defects external to the transformer winding. Taking into account its success in locating some kinds of defects, partial discharge measurements by both acoustic and electromagnetic means supplement other usual tests to increase the reliability of the final diagnosis.

Partial discharge (PD) location in power transformers using electrical methods require transformer parameters to estimate the PD location. Previous research using a lumped parameter model of a transformer consisting of inductance (L), series capacitance (K) and shunt capacitance (C) has shown an algorithm for PD location. This algorithm does not require L, K and C values for the transformer in their explicit form. Rather, the products LC and LK are required. This paper presents three methods of estimating LC and LK values for a power transformer, which could then be used for PD location. The paper shows that all three methods give identical results confirming that either of these methods could be used for estimating LC and LK values. Results based on impedance measurements from two transformer windings are also presented.

Medium voltage underground cables are widely used in urban area as distribution power lines due to their advantages over overhead cables. Underground cables may suffer from partial discharge because of insulation degradation after certain period of time. In this paper, an improved multi-end partial discharge location algorithm called segmented correlation trimmed mean algorithm is proposed. The algorithm uses segmented correlation technique and trimmed mean data filtering technique to enhance the accuracy of partial discharge location algorithm. The algorithm had been tested in MATLAB environment which consists white Gaussian noise. First, discrete wavelet transform is carried out to suppress noisy signals. Next, segmented correlation technique is applied to the de-noised segmented partial discharge signals. Segmented correlation technique has manipulates partial discharge signals by performing cross correlation process on segmented partial discharge signals in order to reduce memory usage and algorithm execution time. Lastly, trimmed mean data filtering technique is applied to the estimated partial discharge location values to estimate new partial discharge location value. Trimmed mean data filtering technique does statistical process control on estimated partial discharge location values in order to minimize the error of estimated partial discharge location. The segmented correlation trimmed mean algorithm had been compared with the existing multi-end correlation-based algorithm. The result shown that the segmented correlation trimmed mean algorithm has better accuracy than multi-end correlation-based algorithm.

Two major tasks of partial discharges (PD) measurements may be distinguished, (i) providing general evidence and the type of PD (detection) and (ii) the location of the PD. Dependent on the type of device under test the two issues have changing priority. For the on-line/on-site PD location in power transformers unconventional PD measuring methods like acoustic ultra-sonic measurements or electromagnetic measurements up to the UHF (ultra high frequencies) range are performed, while the Time Domain Reflectometry (TDR) of electric PD signals is a standard technique for the location of PD in power cables.

This paper proposes a new method for improving the partial discharge (PD) localization accuracy by detecting the best places of ultra‐high frequency (UHF) sensors on the power transformer tank. In this way, firstly, the critical sites inside the power transformer that can produce PD are determined. Then the sensors’ positioning is done in such a way that if the PD happens at each point of these critical sites, the distance between the estimated PD location obtained by solving the time difference of arrival (TDOA) equations and the actual PD location is laid in an acceptable range. For this purpose, the TDOA equations are linearized around the critical PD coordinates as a matrix form A.X = B. The configuration of the sensor positions that minimizes the condition number of the matrix A is the best configuration. To evaluate the effectiveness of the proposed method, several simulations and evaluations are performed on a power transformer in the CST software. The obtained results and comparing them with the previous method and also conventional arbitrary sensors positioning confirm the effectiveness of this method.