Transformer Fault Detection Research Articles

GCA-CNN based transformer digital twin model construction and fault diagnosis and condition evaluation analysis

Oil-immersed power transformer is the most important piece of equipment in the transmission system, and the stable operation of this equipment is of great significance to the normal work of the power system. At present, deep learning has been widely used in transformer condition evaluation and fault detection. About the shortage of deep learning algorithm models, this paper proposes a transformer digital twin model construction and fault diagnosis and condition evaluation analysis based on gray clustering algorithm (GCA) and conventional neural network (CNN). The data are first collected and filtered by combining the operation data, condition information amount and faulty features of power transformers, and then the condition features and fault features are combined to evaluate and detect the condition and faults of power transformers using GCA-CNN respectively. The whitening weight function is determined by expert scoring, the state evaluation matrix is established, the evaluation coefficients are obtained to calculate the evaluation weights, and the transformer state is obtained according to the clustering coefficients; 2000 pieces of raw data are input into the model to obtain the output fault types. Finally, the results are derived and compared with the real results. This paper uses real data from a power plant in Yunnan, and according to the results, the model established in this paper has higher accuracy and better evaluation and detection effects.

Online Sensitive Turn-to-Turn Fault Detection in Power Transformers

Power transformers are of the most critical and expensive equipment in the power system industry. Meanwhile, they are exposed to a variety of faulty and abnormal operating conditions. Therefore, to prevent extension of these events they are protected by different protective schemes. Among the various types of the transformer faults, detecting turn-to-turn faults is the most challenging one for the protection systems. Industrial experiences and standard documents reveal that the existing protection systems include severe shortcomings, and still more efforts are required to offer a reliable method for detecting such faults. To this end, this article puts forward the proposal of a novel turn-to-turn fault detection method to satisfy the basic protection requirements of dependability, security, and speed. This method properly detects the low-level turn-to-turn faults while is stable during system transients and external faults. Performance of the proposed method is evaluated through various experimental and simulation studies. The obtained results verify superior operation of the proposed method under different faults and conditions.

Cellulose Degradation and Transformer Fault Detection by the Application of Integrated Analyses of Gases and Low Molecular Weight Alcohols Dissolved in Mineral Oil

This article presents a method for quantification of methanol and ethanol integrated in the same gas-chromatographic run with a quantification of gases dissolved in mineral oil, making it an integrated tool in transformer diagnostics. The results of aging experiments at 120 °C and 60 °C of Kraft paper, copper, barrier, and pressboard immersed in mineral oil, as well as the aging of thermal upgrade paper in mineral and natural ester oil at 140 °C are presented, in order to investigate correlations between different aging markers and to evaluate their partitioning between oil and cellulose at defined conditions. The results of partitioning experiments at 60 °C showed that re-absorption of methanol from oil to the cellulose materials is faster than the re-absorption of furans. This means that methanol is a paper-degradation marker that can be used in diagnostics over shorter equilibrium times and for the detection of developing faults at broader temperature ranges. Furthermore, a statistical overview of methanol concentration from a database and two transformer fault diagnostic cases are presented. Therefore, in addition to an analysis of gases dissolved in oil, the use of methanol and ethanol in transformer fault and failure investigations should be explored and verified through transformer fault investigations and postmortem analyses.

Power Transformer Fault Detection Based on Improved Capsule Coverage Algorithm

In the process of transformer fault detection, oil chromatographic analysis is usually used to observe the change of the dissolved gas content in the transformer oil, warn and judge the internal fault of the transformer. However, in the actual monitoring environment, there inevitably are the problems of missing data, redundancy, and uncertain dimensions. In order to solve the problem of low recognition efficiency or even incorrect recognition caused by the above problems in transformer fault detection, the idea of nuclear is introduced to the improved capsule coverage algorithm. The experimental results obtained by this improved algorithm indicate that the optimized capsule coverage algorithm can realize the accurate identification of transformer faults, it can not only improve the fault maintenance efficiency, but also ensure the transformer has a good power supply level, which is beneficial to improve the construction of national comprehensive power grid.

Large-Power Transformers: Time Now for Addressing Their Monitoring and Failure Investigation Techniques

Several review studies exist in the literature about monitoring, fault detection, and diagnosis of power transformers. However, they are general approaches in terms of power transformers. Some only focus on applying a specific class of techniques, but again, for general power transformers. Other reviews focus on applying different technologies such as fiber optics, thermal cameras, and vibration sensors, but all within the perspective of general power transformers. A significant question remains: among all types of power transformers, which specific techniques should be used, and why are they more adequate? What are the uncertainties that can decrease their precision? What about the balance, in terms of costs, associated with applying a certain technique and the return needed for a particular type of transformer? In this context, this paper is not only a literature review of well-known problems related to power transformers. Here, we do not just center on large power transformers (100 MVA or higher). Still, we describe a case study of a phase-shifting 1400 MVA-400 kV three-phase transformer that currently connects two European countries that began to show signs of abnormal operating conditions in 2012. In this way, the need to detect and identify anomalies in their initial stage of development for a possible preventive maintenance action is more than justified, which is essentially achieved with continuous monitoring models of the transformer, as concluded in this paper.

Transformer vibration fault detection based on spectrum analysis

Measuring the vibration signal of transformer is one of the ways to verify whether the transformer is faulty. In this paper, taking the vibration detection of transformer as the technical starting point, the signals before and after the vibration fault of transformer are compared by frequency spectrum analysis, and the vibration fault detection of transformer is realized. The vibration test system is constructed and tested in the field. The results show that the non-contact vibration sensor can be used to detect the uninterruptible power failure of the transformer through spectrum analysis and processing, which is simple, safe and reliable, and has good application value.

Detection of ultra-low concentration acetylene gas dissolved in oil based on fiber-optic photoacoustic sensing

Dissolved acetylene gas in oil is the characteristic gas to distinguish overheating and discharge faults of transformer. To realize the detection of ultra-low concentration acetylene in transformer oil, a photoacoustic (PA) gas detection system based on ultra-high sensitive fiber-optic acoustic sensing technique was designed. In order to amplify the power of the laser, an Erbium-doped fiber amplifier (EDFA) was utilized, thereby increasing the intensity of the PA excitation signal. A cantilever based fiber-optic acoustic sensor was utilized to detect the generated PA signal. the interference of carbon dioxide and water vapor on the detection of ultra-low concentration acetylene was studied through analyzing the absorption spectrum. By optimizing the laser modulation parameters, the interference of the interfering gases on the acetylene absorption signal is reduced. Experimental results show that the response time of the system is 9 min with the headspace degassing based oil–gas separation method. The ultra-low concentration of dissolved acetylene gas in oil was tested, and the detection limit reached 0.05 μL/L. The technical solution provides a new solution for transformer fault detection.

An innovative method of fault detection in power transformers

The Bayesian classifier is a priori the optimal solution for minimizing the total error in problems of statistical pattern recognition. The article suggests using the classifier as a regular tool to increase the reliability of defect recognition in power oil-filled transformers based on the results of the analysis of gases dissolved in oil. The wide application of the Bayesian method for solving tasks of technical diagnostics of electrical equipment is limited by the problem of the multidimensional distribution of random parameters (features) and the nonlinearity of classification. The application of a generalized feature of a defect in the form of a nonlinear function of the transformer state parameters is proposed. This simultaneously reduces the dimension of the initial space of the controlled parameters and significantly improves the stochastic properties of the random distribution of the generalized feature. A special algorithm has been developed to perform statistical calculations and the procedure for recognizing the current technical condition of the transformer using the generated decision rule. The presented research results illustrate the possibility of the practical application of the developed method in the conditions of real operation of power transformers.

Acoustic Emission Based Fault Detection of Substation Power Transformer

Fault detection of Substation Power Transformer by Non-contact measurement is important for the safety of machines, instruments, and human beings. To make non-contact measurement as convenient as possible, it is desirable that efficient algorithms based on AE (acoustic emission) discrimination are developed. This paper presents a system for quick and effective fault detection of substation power transformer, based on AE signals collected by non-contact single microphones. In the experiment, collected data were preprocessed in multiple ways and three machine learning algorithms were designed based on classifiers (Convolutional Neural Network (CNN), support vector machine (SVM), and k-nearest neighbors (KNN) algorithm) trained and tested by a tenfold cross-validation technique. After comparison among the designed classifiers, the results show the two-dimensional principal component analysis (2DPCA) preprocess combined with SVM achieved the best comprehensive effectiveness.

Detection and Classification of Lamination Faults in a 15 kVA Three-Phase Transformer Core Using SVM, KNN and DT Algorithms

This paper deals with the detection and classification of two types of lamination faults (i.e., edge burr and lamination insulation faults) in a three-phase transformer core. Previous experimental results are exploited, which are obtained by employing a 15 kVA transformer under healthy and faulty conditions. Different test conditions are considered such as the flux density, number of the affected laminations and fault location. Indeed, the current signals are used where four features (Average, Fundamental, THD and STD) are extracted. Elaborating A total of 328 samples, these features are utilized as input vectors to train and test classification models based on SVM, KNN and DT algorithms. Based on the selected features, the results confirm that the transformer current can be used for the detection of lamination faults. An accuracy rate of more than 84% is obtained using three different classifiers. Such findings provide a promising step toward fault detection and classification in electrical transformers, helping to prevent the system and avoid other related issues such as the increase of power loss and temperature.

Prognostication of Failures Using Signal-to-Noise Ratio to Determine Partial Discharges Activities in Power Transformers

The increase in number of power transformer necessitates increased awareness among power system engineers to monitor transformer health in order to avoid complete failures that may results in prolonged outages. The utilities, nowadays, are exploring different ways of detecting these faults in an early stage. This paper aims to present a real-time monitoring technique for a three-phase transformer using data analytics. Accordingly, the use of signal to noise ratio (SNR) for fault detection in power transformers has been described. This technique determines the partial discharge with the help of sensors needed for unique identification of failed transformers in different test systems. The SNR for different transformer coils installed in field conditions has been compared and discussed in detail. The technique is expected to enable power transformer operators for detection of incipient faults and thus can avoid catastrophic failures and the resultant losses.

Review on Transformer Fault Detection and Diagnosis Algorithms

The method of fault diagnosis based on dissolved gas analysis (DGA) is of great importance to detect possible failures in the transformer and to improve the safety of the electrical system. The DGA data of the transformer in the smart grid has the characteristics of a large amount, different types and a low density of values. Since the power transformer is an important type of power supply in the electrical network, this document provides a complete overview of the power transformer and describes how to diagnose faults. Furthermore, on-line monitoring, the method of fault diagnosis and condition-based maintenance strategy decision-making method as also have been described. The paper presents detailed literature on the recent advancements and methods being adopted by various authors on fault detection.

Research on vibration characteristics and Multi-parameter state recognition of ± 500 kV converter transformer under fluctuating conditions

The vibration of the winding and the iron core is the main source of the vibration of the converter transformer tank, and the vibration measurement and characteristic analysis are the judgment basis for its condition monitoring and identification. Compared with AC transformers, the vibration mechanism and vibration characteristics of converter transformers under fluctuating operating conditions are imperfect. Aiming at the special tank structure and considering the fluctuation of operating conditions, this paper establishes a multi-channel vibration measurement system for online identification of the operating status of the converter transformer, and extracts the characteristics parameters of the vibration signal in the time domain, frequency domain, and energy, and analyzes the differential distribution characteristics of winding vibration under different load currents, iron core vibration under different no-load voltages, and combined vibration during online rated operation. Finally, a method for identifying the operating status of converter transformers based on multi-characteristic parameters is proposed. The research results show that the characteristic parameters such as tank vibration intensity, maximum vibration frequency, and frequency band energy fluctuate with the operating status of the converter transformer, and there are obvious changes. Vibration measurement and characteristics of multi-characteristic parameters can be used to provide a theoretical basis and solution ideas for the operation status identification and early fault detection of converter transformers.

A moving window average method for internal fault detection of power transformers

Power transformers are among the mos]t critical components of power systems and should receive reliable protection against faults. Differential relays are widely used in the protection of power transformers. However, these relays are susceptible to mistaking the inrush currents, which are generated during the turning on of transformers. This paper presents a new method for detecting inrush currents based on the moving window averaging of current. The study found that using the Blackman Harris window rather than other windows in this method significantly reduces fault detection time. The proposed method is robust to strong noises such as white Gaussian noise. Finally, the proposed method was compared with the methods commonly used in industrial protection relays. In this comparison, it was demonstrated that the method offers several advantages over conventional techniques, including the ability to detect faults with high second harmonic and under current transformer saturation and to function properly in modern transformers.

A Probabilistic Calculation Method of Transformer Temperature Fault Based on Photoelectric Infrared Sensor

Aiming at the defect of the lack of quantitative detection probability evaluation method, we design two models which affect the detection probability of image fusion system. They are spectral contrast modeling of target and background and probability model of transformer fault detection and recognition under the influence of multiple factors. Based on the infrared optical imaging method, we established the fault identification model of multi-mode transformer or its connected parts; Studied several infrared optical imaging fusion test systems, and deduced the calculation method of transformer fault detection based on false alarm probability; Analyzed the correlation between the detection ability of the detection system and the characteristics of the target itself, environmental illumination, detection distance and so on. According to the relationship among SNR, detection range and detection probability, we derived the calculation model of target detection probability of optical imaging system under complex background. Through calculation and experimental analysis, the results show that the proposed method can effectively evaluate the detection probability of photoelectric detection imaging system under different background in practical application.

A Novel Condition Monitoring Procedure for Early Detection of Copper Corrosion Problems in Oil-Filled Electrical Transformers

The negative impacts of catastrophic fire and explosion accidents due to copper corrosion problems of oil-filled electrical transformers are still in the spotlight due to a lack of effective methods for early fault detection. To address this gap, a condition monitoring (CM) procedure that can detect such problems in the initial stage is proposed in this paper. The suggested CM procedure is based on identified measurable variables, which are the relevant by-products of the corrosion reaction, and utilizes an Early Fault Diagnosis (EFD) model to detect and solve the copper corrosion problems. The EFD model includes a fault trend chart that can track a fault progression during the useful life of transformers. The purpose of this paper is to verify and validate the effectiveness of the suggested CM procedure by an empirical study in a power plant. The result of applying this procedure was early detection of copper corrosion problems in two transformers with suspected copper corrosion propagation from a total of 84. The corrective action was adding an optimized amount of a passivator, an anticorrosion additive, to suppress the corrosion reaction at the correct time. The main conclusion of this study is the importance of early detection of transformer faults to avoid the negative impacts on societal, company, and individual levels.

Design of sound monitoring and fault diagnosis system for dry-type transformers

This paper focuses on several common fault phenomena and possible causes of dry-type transformer. MATLAB programming realizes the processing of acoustic signal. Through the encapsulation of GUI interface program and the processing and analysis of a large number of fault signals, fault sets are formed. Through the actual transformer fault acoustic signal processing, the main fault frequency of the acoustic signal is extracted and used as an important criterion for transformer fault detection.

A Multinomial DGA Classifier for Incipient Fault Detection in Oil-Impregnated Power Transformers

This study investigates the use of machine-learning approaches to interpret Dissolved Gas Analysis (DGA) data to find incipient faults early in oil-impregnated transformers. Transformers are critical pieces of equipment in transmitting and distributing electrical energy. The failure of a single unit disturbs a huge number of consumers and suppresses economic activities in the vicinity. Because of this, it is important that power utility companies accord high priority to condition monitoring of critical assets. The analysis of dissolved gases is a technique popularly used for monitoring the condition of transformers dipped in oil. The interpretation of DGA data is however inconclusive as far as the determination of incipient faults is concerned and depends largely on the expertise of technical personnel. To have a coherent, accurate, and clear interpretation of DGA, this study proposes a novel multinomial classification model christened KosaNet that is based on decision trees. Actual DGA data with 2912 entries was used to compute the performance of KosaNet against other algorithms with multiclass classification ability namely the decision tree, k-NN, Random Forest, Naïve Bayes, and Gradient Boost. Investigative results show that KosaNet demonstrated an improved DGA classification ability particularly when classifying multinomial data.

Interpretation of Frequency Response Analysis for Fault Detection in Power Transformers

Frequency response analysis (FRA) is a method of monitoring a power transformer’s mechanical integrity. However, identifying the type of fault and its severity by comparing measured responses is still challenging and mostly relies on personnel expertise. This paper is taking one step forward to standardize the FRA interpretation process by proposing guidelines based on various international standards and FRA case studies. In this study, the FRA signature is divided into three regions: low-, mid- and high-frequency regions. The deviation from the fingerprint signature for various faults is classified into small, large, and no variations, based on the calculation of the correlation coefficient. The proposed guidelines are developed based on the frequency regions, and the level of variation is represented using a simple arrow method to simplify the interpretation process. A case study is conducted on a three-phase 11/0.433 kV, 500 kVA distribution transformer with a short circuit winding fault to validate the proposed guidelines.

Application of Infrared Thermography in Fault Detection and Preventive Maintenance in Three-Phase Distribution Transformers

Application of Infrared Thermography in Fault Detection and Preventive Maintenance in Three-Phase Distribution Transformers