Transformer Fault Types Research Articles

Research on transformer fault diagnosis based on genetic algorithm optimized neural network

Aiming at the problem of transformer fault diagnosis in power system, a fault diagnosis method based on genetic algorithm optimization BP neural network is proposed. The BP neural network structure was established based on the traditional three-ratio fault diagnosis method. The concentration of dissolved gas in transformer oil was taken as the input value of the neural network, and the fault working state of transformer was taken as the output value. The established neural network is used to diagnose transformer faults. In order to avoid BP neural network falling into the problem of local optimal value, the weights and thresholds of the neural network are optimized by genetic algorithm. By using the genetic algorithm toolbox in Matlab to establish the genetic algorithm network structure, the transformer fault data as input for the performance of the neural network test. The results show that the neural network optimized by genetic algorithm has a high classification effect on transformer fault types, and the neural network optimized by genetic algorithm has a higher diagnostic efficiency and accuracy than BP neural network.

Power Transformer Fault Diagnosis Based on DGA Using a Convolutional Neural Network With Noise in Measurements

Fault type diagnosis is a very important tool to maintain the continuity of power transformer operation. Dissolved gas analysis (DGA) is one of the most effective and widely used techniques for predicting the power transformer fault types. In this paper, a convolutional neural network (CNN) model is proposed based on the DGA approach to accurately predict transformer fault types under different noise levels in measurements. The proposed model is applied with three categories of input ratios: conventional ratios (Rogers’4 ratios, IEC 60599 ratios, Duval triangle ratios), new ratios (five gas percentage ratios and new form six ratios), and hybrid ratios (conventional and new ratios together). The proposed model is trained and tested based on 589 dataset samples collected from electrical utilities and literature with varying noise levels up to ±20%. The results indicate that the CNN model with hybrid input ratios has superior prediction accuracy. The high accuracy of the proposed model is validated in comparison with conventional and recently published AI approaches. The proposed model is implemented based on MATLAB/toolbox 2020b.

Enhancing Diagnostic Accuracy of Transformer Faults Using Teaching-Learning-Based Optimization

The early detection of the transformer faults with high accuracy rates guarantees the continuous operation of the power system networks. Dissolved gas analysis (DGA) is a technique that is used to detect or diagnose the transformer faults based on the dissolved gases due to the electrical and thermal stresses influencing the insulating oil. Many attempts are accomplished to discover an appropriate technique to correctly diagnose the transformer fault types, such as the Duval Triangle method, Rogers' ratios method, and IEC standard 60599. In addition, several artificial intelligence, classification, and optimization techniques are merged with the previous methods to enhance their diagnostic accuracy. In this article, a novel approach is proposed to enhance the diagnostic accuracy of the transformer faults based on introducing new gas concentration percentages limits and gases' ratios that help to separate the conflict between the diverse transformer faults. To do so, an optimization model is established which simultaneously optimizes both gas concentration percentages and ratios so as to maximize the agreement of the diagnostic faults with respect to the actual ones achieving the high diagnostic accuracy of the transformer faults. Accordingly, an efficient teaching-learning based optimization (TLBO) is developed to accurately solve the optimization model considering training datasets (Egyptian chemical laboratory and literature). The proposed TLBO algorithm enhances diagnostic accuracy at a significant level, which is higher than some of the other DGA techniques that were presented in the literature. The robustness of the proposed optimization-based approach is confirmed against uncertainty in measurement where its accuracy is not affected by the uncertainty rates. To prove the efficacy of the proposed approach, it is compared with five existing approaches using an out-of-sample dataset where a superior agreement rate is reached for the different fault types.

Transformer fault types and severity class prediction based on neural pattern-recognition techniques

Dissolved gas analysis (DGA) is used to diagnose power transformer fault based on the concentration of dissolved gases and the ratios between them. These gases are generated in oils as a result of electrical and thermal stresses, but these DGA techniques cannot identify the severity of the fault types. In IEEE Standard C57.104, the maintenance action is taken based on the total dissolved combustible gases, which is not sufficient because it ignores the importance of the gas type and its change rate. Thermodynamic theory using different starting decomposing materials, namely, n-octane (C8H18) and eicosane (C20H42), is used to estimate the severity of transformer fault types. Two scenarios are suggested with different data transformation techniques to enhance neural pattern-recognition (NPR) method accuracy for predicting transformer fault types and their severity classes. The proposed scenarios are built based on 446 samples collected from the laboratory and literature. Results refer to the role of the starting decomposing material on the severity of the transformer fault and illustrate that the proposed model has a higher accuracy (92.8%) compared with other DGA methods for diagnosing transformer fault types and superior accuracy (99.1%) to predict fault severity class.

Integrated decision‐making method for power transformer fault diagnosis via rough set and DS evidence theories

Precise power transformer fault diagnosis involves incorporating multi-source monitoring information. Uncertain information, missing data, usually occurs in transformer fault cases and diagnosis tasks. To address these challenges, the authors proposed an integrated method of comprehensive transformer fault diagnosis. Diagnostic transformer rules extracted from fault cases form a decision-making table, whereby the main transformer monitoring information and fault types serve as conditional and decision attributes, respectively. Different fault-warning symptoms of the conditional attributes and corresponding decision attributes constitute diagnostic rules. Each obtained symptom in a diagnostic task is evidence supporting different fault types. A modified basic probability assignment (BPA) calculation method is proposed to determine the fault type probability by the obtained symptom. To address contradictory evidence, the symptom significance is introduced to design an improved combination rule incorporating all calculated BPA values to accomplish fault diagnosis. The obtained diagnostic results indicate that more symptoms and a higher symptom significance enable reliable transformer fault diagnosis. The recognition rate of the authors’ method reaches 91.2% with 12–14 symptoms and 94.3% for a 0.9 symptom significance coefficient. It is demonstrated that compared with other combination rules, their method attains a suitable performance (contradiction coefficient K = 0.9 at an 81.3% recognition rate) in realising contradictory information fusion.

Hybrid feature selection–artificial intelligence–gravitational search algorithm technique for automated transformer fault determination based on dissolved gas analysis

Dissolved gas analysis (DGA) is commonly used to identify the fault type in power transformers. However, the available DGA methods have certain limitations because every method depends on the concentration of the dissolved gases. Therefore, in this work, hybrid feature selection–artificial intelligence–gravitational search algorithm (GSA) techniques were proposed to determine the fault type of power transformers based on DGA data. The artificial intelligence (AI) methods applied include support vector machine and artificial neural network. Both AI methods were optimised by GSA to enhance the accuracy of the results. Feature selections using stepwise regression and robust regression were applied to utilise only significant gases. The accuracy of the results was tested with various ratios of testing and training data. Comparison of the results using the proposed method with other optimisation methods and the previous works was performed to validate the performance of the proposed technique. It was observed that the proposed hybrid feature selection–AI–GSA technique yields reasonable accuracy although fewer types of dissolved gases were used. Therefore, the proposed method can be recommended for the application of automated power transformer fault type detection based on DGA data in practice.

Low-cost and Efficient Fault Detection and Protection System for Distribution Transformer

Distribution transformers are a vital component of electrical power transmission and distribution system. Frequent Monitoring transformers faults before it occurs can help prevent transformer faults which are expensive to repair and result in a loss of energy and services. The present method of the routine manual check of transformer parameters by the electricity board has proven to be less effective. This research aims to develop a low-cost protection system for the distribution transformer making it safer with improved reliability of service to the users. Therefore, this research work investigated transformer fault types and developed a microcontroller-based system for transformer fault detection and protection system using GSM (the Global System of Mobile Communication) technology for fault reporting. The developed prototype system was tested using voltage, current and temperature, which gave a threshold voltage higher than 220 volts to be overvoltage, a load higher than 200 watts to be overload and temperature greater than 39 degrees Celsius to be over temperature was measured. From the results, there was timely detection of transformer faults of the system, the transformer protection circuits were fully functional, and fault reporting was achieved using the GSM device. Overall, 99% accuracy was achieved. The system can thus be recommended for use by the Electricity Distribution Companies to protect distribution transformers for optimal performance, as the developed system makes the transformers more robust, and intelligent. Hence, a real-time distribution transformer fault monitoring and prevention system is achieved and the cost of transformer maintenance is reduced to an extent.

Fuzzy reinforcement learning based intelligent classifier for power transformer faults

In this work a fuzzy reinforcement learning (RL) based intelligent classifier for power transformer incipient faults is proposed. Fault classifiers proposed till date have low identification accuracy and do not identify all types of transformer faults. Herein, an attempt has been made to design an adaptive, intelligent transformer fault classifier that progressively learns to identify faults on-line with high accuracy for all fault types. In the proposed approach, dissolved gas analysis (DGA) data of oil samples collected from real power transformers (and from credible sources) has been used, which serves as input to a fuzzy RL based classifier. Typically, classification accuracy is heavily dependent on the number of input variables chosen. This has been resolved by using the J48 algorithm to select 8 most appropriate input variables from the 24 variables obtained using DGA. Proposed fuzzy RL approach achieves a fault identification accuracy of 99.7%, which is significantly higher than other contemporary soft computing based identifiers. Experimental results and comparison with other state-of-the-art approaches, highlights superiority and efficacy of the proposed fuzzy RL technique for transformer fault classification.

Oil Chromatographic Analysis and Fault Diagnosis Case Analysis for Transformers

The production principle of characteristic gas in transformer oil and its corresponding transformer fault type are discussed in this paper. Aiming at the fault of multiple characteristic gases exceeding the standard in a 110 kV transformer oil, this paper analyzed the fault reasons according to the characteristic gas method and the three-ratio method, and combined with relevant tests, it was judged that the transformer core multiple grounding fault caused the fault. The on-site inspection verifies the correctness of the failure analysis results and the corresponding measures were adopted to eliminate the transformer faults.

A Power Transformer Fault Diagnosis Method Based on Hierarchical Classification and Ensemble Learning

Since conventional machine learning methods often result in low diagnostic accuracy and non-negligible recognition difference among fault types with imbalanced class distribution among transformer fault types, a multi-level hierarchical power transformer fault diagnostic model is proposed based on hierarchical classification and ensemble learning, where classifiers are hierarchically constructed for level-by-level diagnosis according to the imbalance extent on each level. The Level I neural network classifier extracts 3 generalized feature labels of normal, discharge fault and thermal fault for feature fusion with original data input, to guide classification among 9 detailed operation status under DL/T 722-2014 standard. The Level II classifier adopts EasyEnsemble, generating balanced training subsets by undersampling majority classes and training sub-classifiers in parallel for parameter synthesis in ultimate classifier, to balance information between major fault types and minor ones. Experimental result shows that: compared to traditional methods, our proposed method improves the generalization ability on minority class faults and the overall accuracy by 7%.

A Transformer Fault Diagnosis Method Based on Bayesian Network

This paper summarizes the fault types and fault hazards of the transformer, and combs the corresponding relationship between the transformer fault types and the fault symptoms, which provides reference for the transformer fault diagnosis and state maintenance. Based on the analysis of the existing transformer fault diagnosis, the transformer fault diagnosis based on Bayesian network is mainly studied. Bayesian network has advantages in machine learning, data reasoning and so on. It has been widely applied in the field of fault diagnosis. In this paper, the QMR model of transformer fault diagnosis is established by using the principle of Bayesian network and the relationship between fault type and fault feature, and the model is used to diagnose the specific fault. Then the paper further analyzes the result of fault diagnosis in order to prove the correctness and validity of Bayesian network.

Identification of transformer fault based on dissolved gas analysis using hybrid support vector machine-modified evolutionary particle swarm optimisation.

Early detection of power transformer fault is important because it can reduce the maintenance cost of the transformer and it can ensure continuous electricity supply in power systems. Dissolved Gas Analysis (DGA) technique is commonly used to identify oil-filled power transformer fault type but utilisation of artificial intelligence method with optimisation methods has shown convincing results. In this work, a hybrid support vector machine (SVM) with modified evolutionary particle swarm optimisation (EPSO) algorithm was proposed to determine the transformer fault type. The superiority of the modified PSO technique with SVM was evaluated by comparing the results with the actual fault diagnosis, unoptimised SVM and previous reported works. Data reduction was also applied using stepwise regression prior to the training process of SVM to reduce the training time. It was found that the proposed hybrid SVM-Modified EPSO (MEPSO)-Time Varying Acceleration Coefficient (TVAC) technique results in the highest correct identification percentage of faults in a power transformer compared to other PSO algorithms. Thus, the proposed technique can be one of the potential solutions to identify the transformer fault type based on DGA data on site.

Dissolved gas analysis of insulating oil for power transformer fault diagnosis with deep belief network

Dissolved gas analysis (DGA) of insulating oil can provide an important basis for transformer fault diagnosis. To improve diagnosis accuracy, this paper presents a new transformer fault diagnosis method based on deep belief networks (DBN). By analyzing the relationship between the gases dissolved in transformer oil and fault types, the Non­code ratios of the gases are determined as the characterizing parameter of the DBN model. DBN adopts multi-layer and multi-dimension mapping to extract more detailed differences of fault types. In this process, the diagnosis parameters are pre-trained. A back-propagation algorithm adjusts them with the labels of the samples and optimizes the parameters. To verify the effect of the proposed method, the diagnostic DBN model is constructed and tested using various oil chromatographic datasets collected from the State Grid Corporation of China and previous publications. The performances of the DBN diagnosis model are analyzed by different characterizing parameters, different training datasets and sample datasets. In addition, the influence of discharge and overheating multiple faults on the diagnosis model is studied. The performance of the proposed approach is compared with that derived from support vector machine (SVM), back-propagation neural network (BPNN) and ratio methods respectively. The results show that the proposed method significantly improves the accuracy of power transformer fault diagnosis.

A new approach of DGA interpretation technique for transformer fault diagnosis

Dissolved Gas Analysis (DGA) is one of the most common techniques to detect the incipient faults in the oil-filled power transformers. In this paper, a new approach of DGA technique is proposed to overcome the conflict that takes place in the traditional interpretation techniques for transformer fault diagnosis. The new approach is based on the analysis of 386 dissolved gas samples data set that collected from the Egyptian electric utility chemical laboratory as well as from credited literatures. These data sets are used to build the technique model and also as a tested data set to get the technique’s accuracy. The new approach DGA diagnoses the transformer fault types based on the gas concentration percentage limit of the sum of main five gases (Hydrogen (H2), Methane (CH4), Ethan (C2H6), Ethylene (C2H4), and Acetylene (C2H2)) and some suggested gases ratios depending on the sample data set analysis. The validation of the proposed approach of DGA technique is satisfied by comparing its results with the results of the IEC Standard Code, Duval triangle and Rogers methods for the collected data set. The results refer to the ability and reliability of the new approach in transformer faults diagnostic.

Transformer Incipient Fault Prediction Using Combined Artificial Neural Network and Various Particle Swarm Optimisation Techniques

It is important to predict the incipient fault in transformer oil accurately so that the maintenance of transformer oil can be performed correctly, reducing the cost of maintenance and minimise the error. Dissolved gas analysis (DGA) has been widely used to predict the incipient fault in power transformers. However, sometimes the existing DGA methods yield inaccurate prediction of the incipient fault in transformer oil because each method is only suitable for certain conditions. Many previous works have reported on the use of intelligence methods to predict the transformer faults. However, it is believed that the accuracy of the previously proposed methods can still be improved. Since artificial neural network (ANN) and particle swarm optimisation (PSO) techniques have never been used in the previously reported work, this work proposes a combination of ANN and various PSO techniques to predict the transformer incipient fault. The advantages of PSO are simplicity and easy implementation. The effectiveness of various PSO techniques in combination with ANN is validated by comparison with the results from the actual fault diagnosis, an existing diagnosis method and ANN alone. Comparison of the results from the proposed methods with the previously reported work was also performed to show the improvement of the proposed methods. It was found that the proposed ANN-Evolutionary PSO method yields the highest percentage of correct identification for transformer fault type than the existing diagnosis method and previously reported works.

Internal fault fast identification criterion based on superimposed component comparison for power transformer

Abstract Due to the interference of false differential current caused by the saturation of the iron core, the differential protection for the power transformer may suffer mal-operation and long response time. According to the investigation for the characteristic of transformer core, there should be a time interval between the sudden change of phase voltage and the emergence of differential current. And it can be utilized to distinguish between the external and the internal faults. A new time-interval based criterion for transformer differential protections is proposed, aiming at identifying transformer fault type by virtue of highly accurate time interval. Furthermore, a novel filter algorithm is proposed to determine the time interval for its excellent singularity detecting and effective noise-repressing ability as well as high-speed response. The EMTDC-based simulation, which covers various fault situations, is analyzed, of which the results confirm that the criterion can effectively prevent the differential protection from mal-operations and meanwhile guarantee the high-speed response to the internal fault.

English

Power transformers has an important role in electrical power transmission and its interruption has financial losses, thus its condition monitoring is essential and performance of this equipment is effective for power system reliability. In this paper, proposed method has advantages of both probabilistic neural network (PNN) and Interval Type-2 Fuzzy Support Vector Machine (IT2FSVM). Firstly, main feature is extracted from primary and secondary three phase currents and search coils differential voltage by wavelet transform and this information is used as probabilistic neural network inputs. AI techniques are applied to establish classification features for faults in the transformers based on the collected gas data. The features are applied as input data to PNN and IT2FSVM combination of classifiers for faults classification. The experimental data from NTPC Korba-India is used to evaluate the performance of proposed method. The results of the various DGA methods are classified using AI techniques. In comparison to the results obtained from the AI techniques, the PNN plus IT2SVM has been shown to possess the most excellent performance in identifying the transformer fault type. The test results indicate that the PNN plus IT2SVM approach can significantly improve the diagnosis accuracies for power transformer fault classification. In addition, the study aims to study the joint effect of PNN and IT2SVM on the classification performance when used together.

Simulation and Analysis on Winding Deformation of a Power Transformer in Current Transformer Connecting Manner

Transformer winding deformation is one of the main types of transformer faults. To check if a power transformer is being under winding deformation, the transformer can be connected in a current transformer during its testing. A transformer winding simulating model is set up under this connecting manner. Then the simulation has been performed with current source which is the frequency sweep power. The simulation results show that the winding deformation can be reflected effectively with current source method. This method lays the foundation for the realization of on-line monitoring and diagnosis of the transformer neutral directly grounded side winding.

Research on Transformer Fault Diagnosis Method Based on Artificial Immune Network and Fuzzy C-Means Clustering Algorithm

The transformer is one of the indispensable equipment in transformer substation, it is of great significance for fault diagnosis. In order to accurately judge the transformer fault types, an algorithm is proposed based on artificial immune network combined with fuzzy c-means clustering to study on transformer fault samples. Focus on the introduction of data processing of transformer faults based on artificial immune network, the identification of transformer faults based on fuzzy c-means clustering, and the simulation process. The experimental results show that the proposed algorithm can classify power transformer fault types effectively, and the algorithm has a good application prospect in the transformer fault diagnosis.

Dissolved gas analysis (DGA) of natural ester insulating fluids with different chemical compositions

This study presents experimental results of the application of the dissolved gas-inoil analysis (DGA) technique to natural ester insulating fluids. The investigations covered two common types of transformer faults, discharge and overheating, and four different commercial insulating oils, one mineral and three natural esters: Biovolt A (from corn oil), Biovolt HW (from sunflower oil with high oleic acid) and Envirotemp FR3 (from soya oil). The natural esters insulating fluids (NEIFs) analyzed had different chemical compositions. Experiments were performed in the laboratory with the test oils at three different moisture levels (low, medium and high). The purpose was to see the effect of moisture content and oil composition on the quantities of fault gases dissolved in the different oil types. The overheating tests were performed by heating the oils in an oven at 150 °C. Discharges were generated using a flat plate-plate electrode arrangement immersed in oil. Five key dissolved hydrocarbon gases - hydrogen (H2), methane (CH4), ethane (C2H6), ethylene (C2H4) and acetylene (C2H2) - generated by the two fault types in NEIFs were analyzed and compared with mineral oil (MIO) fault gases. The classic Duval triangle 1, modified Duval triangle 3, Roger's ratio, Dornenburg and IEC 60599 methods were applied to predict the fault type. The findings confirmed that the direct application of classic DGA fault interpretation methods, developed for mineral oils, to non-mineral oils needs some modifications. IR and RMN spectroscopy analyses were performed to observe differences in triacylglycerol group compositions before and after heating tests.