Search

Abstract Avoiding financial losses requires preventing catastrophic oil‐filled power transformer breakdowns. Continuous online transformer monitoring is needed. The authors use paper insulation to evaluate transformer health for continuous online transformer monitoring. The study suggests a new artificial intelligence method for estimating paper insulation residual life in oil‐immersed power transformers. The four artificial intelligence models use backpropagation‐based neural networks to predict paper insulation lifespan. Four primary transformer insulating paper failure indices—degree of polymerisation, 2‐furfuraldehyde, carbon monoxide, and carbon dioxide—form the basis of these models. Each model, including the backpropagation‐based neural networks, estimates paper insulation life using one failure index, along with moisture and temperature data. Optimisation techniques enhance hidden layer neurons and epoch count for improved performance. Results are validated against literature‐based life models, establishing a precise input–output correlation. This method accurately predicts the remaining useable life of power transformer paper insulation, enabling utilities to take proactive measures for safe and efficient transformer operation.

Abstract The lumped element network model has been proven to be an efficient tool for the interpretation of transformer Frequency Response Analysis. However, it is challenging to obtain parameters of the model if transformer design information is unavailable. In such a case, optimisation algorithms can be used for gray‐box model parameter estimation. A methodology is developed by the authors to establish a transformer network model without winding design data, and instead, end‐to‐end open circuit Frequency Response and other terminal test results are utilised; and Genetic Algorithm is applied to approximate the unknown parameters of the model. The modelling approach developed is independent of the unit number of the network model and therefore guarantees the accuracy of optimisation. Furthermore, it can deal with transformers with a complicated winding structure such as interleaved disc type winding. Two single windings, helical and interleaved disc type, and a single‐phase 144/13 kV 60 MVA transformer are used to demonstrate the method. FRA spectra produced by the best estimated gray‐box model and the corresponding white‐box model are compared. The main features in amplitude and phase spectra are well matched, with low values of Relative Standard Deviation, for both single windings and transformer. Estimated electrical parameters show a high consistency with reference values, which are calculated based on winding design data. This validates the methodology and gives confidence to apply grey‐box modelling for FRA interpretation.

Few-shot semantic segmentation (FSS) is a challenging task that aims to segment new classes in query images with a few annotated support samples. One inherent challenge in FSS is the intra-class variation resulting from the limited availability of support samples and the diversity of query data. Current methods frequently employ prototype-support techniques to tackle this issue. However, it is important to note that a single support prototype shares limited commonalities with query features, and increases the difficulty of accurate segmentation. In this paper, we propose a lightweight and effective framework named Siamese Transformer (SiaT) with a mere 0.68M learnable parameters to enhance commonalities between prototypes and query features. The SiaT framework consists of two key modules: the Siamese Transformer Module (STM) and the Query Activation Module (QAM). The STM integrates two shared Transformer decoders for information propagation to generate two enhanced prototypes. One Transformer decoder facilitates the propagation of target-related information from the query to the original support prototype, while the other propagates foreground information from the support to the initial query prototype. The QAM utilizes the two enhanced prototypes from the STM as support information, engaging with query features through channel-wise allocation and concatenation, termed query activation. Moreover, SiaT showcases competitive performance on the widely used benchmarks PASCAL- 5 i and COCO- 20 i , which demonstrates its effectiveness in addressing the intra-class variation challenge within FSS tasks. [ABSTRACT FROM AUTHOR]

Avoiding financial losses requires preventing catastrophic oil‐filled power transformer breakdowns. Continuous online transformer monitoring is needed. The authors use paper insulation to evaluate transformer health for continuous online transformer monitoring. The study suggests a new artificial intelligence method for estimating paper insulation residual life in oil‐immersed power transformers. The four artificial intelligence models use backpropagation‐based neural networks to predict paper insulation lifespan. Four primary transformer insulating paper failure indices—degree of polymerisation, 2‐furfuraldehyde, carbon monoxide, and carbon dioxide—form the basis of these models. Each model, including the backpropagation‐based neural networks, estimates paper insulation life using one failure index, along with moisture and temperature data. Optimisation techniques enhance hidden layer neurons and epoch count for improved performance. Results are validated against literature‐based life models, establishing a precise input–output correlation. This method accurately predicts the remaining useable life of power transformer paper insulation, enabling utilities to take proactive measures for safe and efficient transformer operation. [ABSTRACT FROM AUTHOR]

This paper presents a comprehensive analysis of failures in low-voltage terminal connections within power transformers and proposes a fault classification methodology based on infrared thermography (IRT). Low-voltage terminal connections play a critical role in the reliable operation of power transformers, and their failures can lead to severe operational issues. In this study, we employ IRT as a noninvasive and efficient diagnostic tool to identify and classify various types of failures, including loose connections, overheating, and corrosion. The research involves the collection of infrared thermograms (IRT images) from the low-voltage terminals of power transformers under different operating conditions. The proposed methodology demonstrates its effectiveness in detecting and classifying low-voltage terminal connection failures, thereby enabling timely preventive maintenance and minimizing the risk of transformer malfunctions. This research contributes to enhancing the reliability and longevity of power transformers, reducing downtime, and optimizing maintenance practices in the power industry. [ABSTRACT FROM AUTHOR]

In this work, a mathematical model of a three-phase nonlinear transformer is suggested. The model enables simulating the transformer operation with allowance for its nonlinearity and covers needs of the relay protection. Our model has been developed on the basis of a mathematical model with phase coordinates, where differential equations are composed by the Kirchhoff's phase-voltage law. Based on this model, we first compose a mathematical model for simulating steady-state operation modes of a transformer, taking into account the asymmetry and nonlinearity of its ferromagnetic core. In this model, the initial values of inductances and mutual inductances of loops are determined from the main phase inductance calculated by the experimentally found no-load current, and their current values are determined from the currents in windings and the magnetic fluxes in legs of the transformer core. The magnetic fluxes are calculated by the nodal-pair method. This improved mathematical model is verified through a comparison between the calculated harmonic components of the phase currents and the experimental results. The harmonic components are calculated with the use of Fourier expansion of the calculated phase currents. Their experimental values are determined with a spectrum analyzer. The calculated and experimental harmonic components of the currents of phase A during no-load and rated-load operation of the transformer are tabulated. The comparison of these results shows that the mathematical model of a three-phase transformer we suggest makes it possible to simulate currents in transformer windings under steady-state operation modes with accuracy acceptable for relay protection. [ABSTRACT FROM AUTHOR]

An On-Load Tap Changer (OLTC) that regulates transformer voltage is one of the most important and strategic components of a transformer. Detecting faults in this component at early stages is, therefore, crucial to prevent transformer outages. In recent years, Hydro Quebec initiated a project to monitor the OLTC's condition in power transformers using vibro-acoustic signals. A data acquisition system has been installed on real OLTCs, which has been continuously measuring their generated vibration signal envelopes over the past few years. In this work, the multivariate deep autoencoder, a reconstruction-based method for unsupervised anomaly detection, is employed to analyze the vibration signal envelopes generated by the OLTC and detect abnormal behaviors. The model is trained using a dataset obtained from the normal operating conditions of the transformer to learn patterns. Subsequently, kernel density estimation (KDE), a nonparametric method, is used to fit the reconstruction errors (regarding normal data) obtained from the trained model and to calculate the anomaly scores, along with the static threshold. Finally, anomalies are detected using a deep autoencoder, KDE, and a dynamic threshold. It should be noted that the input variables responsible for anomalies are also identified based on the value of the reconstruction error and standard deviation. The proposed method is applied to six different real datasets to detect anomalies using two distinct approaches: individually on each dataset and by comparing all six datasets. The results indicate that the proposed method can detect anomalies at an early stage. Also, three alarms, including ignorable anomalies, long-term changes, and significant alterations, were introduced to quantify the OLTC's condition. [ABSTRACT FROM AUTHOR]

Lane detection plays a pivotal role in the successful implementation of Advanced Driver Assistance Systems (ADASs), which are essential for detecting the road's lane markings and determining the vehicle's position, thereby influencing subsequent decision making. However, current deep learning-based lane detection methods encounter challenges. Firstly, the on-board hardware limitations necessitate an exceptionally fast prediction speed for the lane detection method. Secondly, improvements are required for effective lane detection in complex scenarios. This paper addresses these issues by enhancing the row-anchor-based lane detection method. The Transformer encoder–decoder structure is leveraged as the row classification enhances the model's capability to extract global features and detect lane lines in intricate environments. The Feature-aligned Pyramid Network (FaPN) structure serves as an auxiliary branch, complemented by a novel structural loss with expectation loss, further refining the method's accuracy. The experimental results demonstrate our method's commendable accuracy and real-time performance, achieving a rapid prediction speed of 129 FPS (the single prediction time of the model on RTX3080 is 15.72 ms) and a 96.16% accuracy on the Tusimple dataset—a 3.32% improvement compared to the baseline method. [ABSTRACT FROM AUTHOR]

A 500-kV power transformer substation was focused on for seismic performance examination. Finite element models of the power transformer, isolation layers, site soil, bus bars and support structure were established under various analysis conditions. By inputting typical ground motions into the models, seismic time history responses were obtained. First, the seismic performance, critical position and failure modes of a single power transformer without isolation were examined. Then, the effect on the power transformer by arranging isolation bearings was confirmed. Moreover, the influence of site soil dynamic characteristics on the isolating performance of the power transformer was examined. Finally, the influence of the length margin of the bus bars on the seismic response of the isolated power transformer was analyzed. It was revealed that the bottom of the transformer bushings which were made of porcelain was a weak part. Installing isolation bearings effectively reduced the maximum stress response of the bushing for most ground motions. The dynamic characteristics of the site soil significantly affected the isolation efficiency of the power transformer, and the expected isolation effect was achieved when the soil was soft. The displacement response of the transformer was largely increased by isolation, the pulling effect of the bus bars on the porcelain bushing was significant. When the length margin of the bus bars was enlarged to 1.0 m, the influence on the bushing decreased to a low level that could be ignored. [ABSTRACT FROM AUTHOR]

A multi-step approach for the fast calculation of the magnetic field inside transformer tank shields, based on the 2D FEM, is presented in the paper. Due to the limitations of the 2D FEM, the proposed approach utilizes several 2D FEM models and calculates the magnetic field in multiple steps to account for the 3D geometry of the problem. In the first step, a distribution of the magnetic flux density that enters the tank shields is calculated using the quasi-3D model of the transformer. This quasi-3D model is obtained by superimposing the solution of multiple axisymmetric 2D FEM models, and assumes a considerably simplified transformer geometry. To account for the tank shield geometry that is neglected in the quasi-3D FEM model, an additional 2D FEM model with tank shields is introduced. After the distribution of the magnetic flux density that enters the tank shields is calculated, it is imposed in the final 2D FEM model with a non-linear tank shield which is used to calculate the magnetic flux density distribution inside the tank shields. The proposed approach enables a fast calculation of magnetic field distributions, both in the vertical and horizontal directions. The results of the proposed approach are compared against the 3D FEM. The relative error of the maximum magnetic flux density is under 2%, while the NRMSE of the magnetic flux density distribution within the tank shields is under 10%. The key contribution of the proposed approach is a low computation time. In the presented case study, the total computation time of the proposed approach is ~30 s, while the computation time of the 3D FEM is ~1 h. As the computation time is significantly reduced, while the accuracy is acceptable, the proposed approach can be a good alternative to the 3D FEM for design purposes. Therefore, it has industrial value. [ABSTRACT FROM AUTHOR]

To study the influence of the voltage total harmonic distortion (THDV) and its spectrum on the harmonic loss factor (FHL) and the maximum allowable load capacity ratio (Imax (pu)) for the transformer of the LHCD system, we use MATLAB software to model the LHCD system and study the winding loss and maximum allowed load capacity of the LHCD system transformer when the supply voltage source is sinusoidal and non-sinusoidal, respectively. The calculation is carried out by the method of IEEE standard C57.110 and the method of considering the skin effect. The calculation results of both methods clearly show that the THDV value of the supply voltage has a significant effect on the harmonic dissipation factor (FHL) and the maximum allowable load capacity ratio (Imax (pu)) of the transformer. And the calculation method considering the skin effect increases the maximum allowable load capacity ratio of the LHCD system transformer by about 2%. The research results have important reference value for the future retrofit design of LHCD system transformers. [ABSTRACT FROM AUTHOR]

Transformer oil plays a critical role in ensuring the insulation and normal operation of power transformers. However, during operation, the transformer oil would undergo oxidation and degradation, leading to the formation of acids and polar compounds that increase its acidity and degrade its performance. Hence, the recycling and regeneration of aged transformer oil is an essential practice in transformer maintenance, contributing significantly to the extension of a transformer's operational lifespan and the reduction of potential risks. Adsorption separation is a preferred method for regeneration, offering simplicity and high efficiency. This study focuses on impregnated γ‐Al2O3 as an adsorbent for acidity removal in waste transformer oil. We conducted static adsorption experiments to investigate the optimal loading amount of the active component and the best calcination temperature for acid removal. XRD technique is employed to monitor the phase changes of the carrier during the calcination process. BET analysis is used to observe variations in the specific surface area and pore volume of the adsorbent during the modification process. DTG‐TGA is utilized to assess the weight loss during the calcination process. FTIR is employed to observe changes in functional groups during modification and adsorption processes. Along with these investigations, a study is also carried out to assess the dynamic acid removal performance of the acid‐removing adsorbent loaded into a fixed bed. The findings contribute to the development of sustainable methods for regenerating transformer oil, ensuring efficient and environmentally friendly operations. [ABSTRACT FROM AUTHOR]

Dissolved gas analysis (DGA) in transformer oil, which analyzes its gas content, is valuable for promptly detecting potential faults in oil-immersed transformers. Given the limitations of traditional transformer fault diagnostic methods, such as insufficient gas characteristic components and a high misjudgment rate for transformer faults, this study proposes a transformer fault diagnosis model based on multi-scale approximate entropy and optimized convolutional neural networks (CNNs). This study introduces an improved sparrow search algorithm (ISSA) for optimizing CNN parameters, establishing the ISSA-CNN transformer fault diagnosis model. The dissolved gas components in the transformer oil are analyzed, and the multi-scale approximate entropy of the gas content under different fault modes is calculated. The computed entropy values are then used as feature parameters for the ISSA-CNN model to derive diagnostic results. Experimental data analysis demonstrates that multi-scale approximate entropy effectively characterizes the dissolved gas components in the transformer oil, significantly improving the diagnostic efficiency. Comparative analysis with BPNN, ELM, and CNNs validates the effectiveness and superiority of the proposed ISSA-CNN diagnostic model across various evaluation metrics. [ABSTRACT FROM AUTHOR]

Because large oil-immersed transformers are enclosed by a metal shell, the on-site localization means it is difficult to achieve the accurate location of the patrol micro-robot inside a given transformer. To address this issue, a spatial ultrasonic localization method based on wavelet decomposition and PHAT-β-γ generalized cross correlation is proposed in this paper. The method is carried out with a five-element stereo ultrasonic array for the location of a transformer patrol robot. Firstly, the localization signal is decomposed into wavelet coefficients of different scales, which would realize the adaptive decomposition of the frequency of the localization signal from low frequencies to high frequencies. Then, the wavelet coefficients are denoised and reconstructed by using the semi-soft threshold function. Second, a modified phase transform-beta-gamma (PHAT-β-γ) method is used to calculate the exact time delay between different sensors by increasing the weights of the PHAT weighting function and introducing a correlation function. Finally, by using the proposed method, the accurate localization of the transformer patrol micro-robot is achieved with a five-element stereo ultrasonic array. The simulation and test results show that inside a transformer experimental oil tank (120 cm × 100 cm × 100 cm, L × W × H), the relative error of transformer patrol micro-robot spatial localization is within 4.1%, and the maximum localization error is less than 3 cm, which meets the requirement of engineering localization. [ABSTRACT FROM AUTHOR]

Owing to the diverse connection configurations of dual active bridge converters, a multiplicity of low-voltage DC port structures are anticipated to emerge in the independent DC bus structure multiport power electronic transformer (IDBS-MPET). An inadequate low-voltage DC port structure exacerbates the power imbalance in IDBS-MPET, presenting a risk of overmodulation even when transmitting relatively low levels of power. To overcome this limitation, a design scheme of IDBS-MPET topology based on the maximum power transmission capability of the low-voltage DC ports is proposed in this paper. Three topology design rules are derived from the maximum power transmission capability results of more than 80 typical IDBS-MPET topologies. The symmetrical triple cross-phase connection structure, the symmetrical double cross-phase connection structure and the single-phase connection structure are sequentially identified as the three most optimal structures of low-voltage DC ports. By employing the proposed design methodology, each low-voltage DC port achieves its maximum power transfer capability relative to other configurations. The effectiveness of the proposed design scheme is validated by an optimal designed IDBS-MPET topology with six low-voltage DC ports. [ABSTRACT FROM AUTHOR]

Summary: To accurately and efficiently calculate the frequency response of the lumped parameter equivalent circuit model (LPECM) of a transformer winding is the basis for studying the simulation method of the frequency response analysis (FRA). In this paper, two different cut‐set matrix (CSM) methods, including fundamental cut‐set (FCS) and independent cut‐set (ICS) matrix methods, are proposed to calculate the FRA curves of the LPECM of a three‐winding power transformer. First, based on graph theory and electrical circuit laws, the basic CSM equation is deduced. Second, four different FCS matrix selection methods of the LPECM are introduced and the best FCS matrix is determined by deeply analyzing their characteristics. Then, an ICS matrix method is proposed to obtain the ICS of the LPECM. Furthermore, FCS and ICS are taken into the CSM equation to calculate the FRA results of the LPECM, respectively. Finally, the accuracy of the proposed methods is validated by Multisim software and the performance of these two methods is compared with node voltage matrix (NVM) method by analyzing the calculation time. The results show that FCS and ICS matrix methods both can be used to accurately obtain the FRA results of the LPECM of the three‐winding power transformer, but ICS is more simple than FCS matrix method due to avoiding the procedure for the selection of the tree. In addition, since ICS matrix method has the shortest running time among these methods, it can be used to improve the calculation efficiency of the FRA curves of the LPECM of a three‐winding power transformer. [ABSTRACT FROM AUTHOR]

Summary: Traditional multiple‐output AC‐DC power supply with power factor correction (PFC) function usually employs a two‐stage cascaded topology that is composed of prestage PFC converter and N independent poststage DC‐DC converters to achieve PFC and voltage or current regulation of N outputs. Hence, the conventional two‐stage cascaded multiple‐output AC‐DC power supply needs N + 1 magnetic components as power inductors or transformers and N + 1 controllers, which increases the cost and size of the multioutput converters. The conventional multiple‐string AC‐DC LED driving circuits usually employ a two‐stage topology to realize PFC and independent current regulation of each LED string. To address the issue of complexity, high cost, and large size of the two‐stage multistring LED driving circuit, a single‐stage triple‐output Cuk PFC converter is proposed by cascading boost converter and single‐inductor triple‐output buck converter with sharing main power switch in this paper. The topology operation principle, control strategy, design consideration, and operation characteristics of the proposed circuit are investigated in detail. With the switch sharing and inductor multiplexing technologies at the same time, the proposed converter utilizes only single‐stage converter to achieve PFC function, independent regulation of triple‐channel output current, and flicker‐free with low twice line frequency LED current ripple. The proposed converter is verified by a 50‐W experimental prototype of LED driver for triple string LEDs. The proposed converter only uses two power inductors and a set of controller to achieve greater than 0.945 PF, up to 90.5% efficiency, low LED current ripple, and independent current regulation for triple LED strings with different forward voltage, which may significantly reduce the cost and size of the traditional AC‐DC multistring LED driving circuit. [ABSTRACT FROM AUTHOR]

The lumped element network model has been proven to be an efficient tool for the interpretation of transformer Frequency Response Analysis. However, it is challenging to obtain parameters of the model if transformer design information is unavailable. In such a case, optimisation algorithms can be used for gray‐box model parameter estimation. A methodology is developed by the authors to establish a transformer network model without winding design data, and instead, end‐to‐end open circuit Frequency Response and other terminal test results are utilised; and Genetic Algorithm is applied to approximate the unknown parameters of the model. The modelling approach developed is independent of the unit number of the network model and therefore guarantees the accuracy of optimisation. Furthermore, it can deal with transformers with a complicated winding structure such as interleaved disc type winding. Two single windings, helical and interleaved disc type, and a single‐phase 144/13 kV 60 MVA transformer are used to demonstrate the method. FRA spectra produced by the best estimated gray‐box model and the corresponding white‐box model are compared. The main features in amplitude and phase spectra are well matched, with low values of Relative Standard Deviation, for both single windings and transformer. Estimated electrical parameters show a high consistency with reference values, which are calculated based on winding design data. This validates the methodology and gives confidence to apply grey‐box modelling for FRA interpretation. [ABSTRACT FROM AUTHOR]

Permanent magnet synchronous motors (PMSM) are very popular in the industrial market. Recently, robots are installed to compensate for the shortage of labor. Due to the increase in motors, the number of power lines and position sensor cable wiring are increased; this increase in complicated wiring is a problem. As a method to solve these problems, power superimposition technology in the power line of a PMSM is expected to reduce the distribution cables. In this paper, the power superposition technology method, and circuit equations of mounting transformers in power lines are presented and are then verified by experiment, simulation, and linear analysis. The circuit equation is proposed as a three‐phase circuit equation, which is transformed into a dq‐circuit equation by a transformation matrix. The characteristics of the electric angular frequency versus the dq‐axis voltage demonstrated satisfactory results, with errors within 1% in linear analysis and simulation. The linear analysis and experiment are shown to be similar, although there are some errors. The validity of the circuit equations could be demonstrated by the verification results. [ABSTRACT FROM AUTHOR]

The analysis of gases dissolved in oil is the main technique used to predict possible failures in power transformers. Currently, there are some assessment algorithms for the predictive analysis, but all of them have in common the fact that they are based on the classical gas formation theory. This work aims to analyze the use of decision trees (DTs) for the analysis of gases dissolved in mineral oil, also establishing a relationship associated with decision making during the formation of trees and the formation of gases. In this way, some important aspects involving the insulating mineral oil can be better understood and quantified. For this purpose, one dataset containing 201 samples is used in the study. The results provided by the tree are arranged in a Cartesian plane, and the gases are analyzed based on the ratios used by Doernenburg. DT is discussed in detail to validate the algorithm performance based on the classical gas formation theory. [ABSTRACT FROM AUTHOR]

Transformer fault diagnosis is very important for the stable operation of transformers, but it is still difficult to recognize transformer faults. The main obstacle is attributed to the lack of labeled data for transformer faults. Due to the limited data of transformer faults, a great deal of unlabeled data in real life is not difficult to obtain and contains a lot of effective information. Semi‐supervised self‐training can significantly enhance the classification accuracy and performance of the base classifier by applying unlabeled data. However, one of the reasons for the degradation of semi‐supervised self‐training is the lack of initial labeled data to train the initial classifier. In this paper, we propose an improved semi‐supervised SMOTE (SS‐SMOTE), which uses the potential information contained in unlabeled data to generate synthetic samples, and then reasonably increases the labeled data set. Meanwhile, we combine SS‐SMOTE with self‐training into an algorithm framework. Finally, a series of experiments are carried out to evaluate the performance of the proposed algorithm, and compared with other existing state‐of‐the‐art semi‐supervised algorithms. The results show that the performance improvement of the classifier using the semi‐supervised improved SMOTE algorithm framework is effective, and has a good effect on solving the problem of small sample classification. © 2023 Institute of Electrical Engineer of Japan and Wiley Periodicals LLC. [ABSTRACT FROM AUTHOR]

Abstract Since compact high temperature superconducting (HTS) transformer has a small size and high efficiency, it can meet the requirements of high‐voltage transformers for vehicle‐mounted mobile substation. However, due to the small internal insulation size of the compact HTS transformers, it is necessary to improve the insulation strength to ensure the insulation safety of the transformer. Reasonable main insulation structure can greatly improve the insulation level of transformer. Based on the design theory of main insulation structure, this paper designs a main insulation structure model “small liquid nitrogen gap—barrier” suitable for compact HTS transformer. The insulation withstand voltage experiment of the main insulation structure shows that when the liquid nitrogen gap distance is less than 8 mm, the insulation strength of liquid nitrogen in the structure is improved. When the liquid nitrogen gap distance is too small, the size and thickness of the barrier will affect the insulation strength of the structure. According to the insulation strength of main insulation structures with different liquid nitrogen gaps, the feasibility of “small liquid nitrogen gap—barrier” main insulation structure is verified. The small liquid nitrogen gap structure can be adopted for the main insulation of compact HTS transformer.

Abstract This paper addressed the challenges associated with the complexity, numerous parameters, computational resource demands, and slow processing speed of transformer fault identification models based on deep learning technologies. Sparse convolutional neural network (CNN) approach is proposed for identifying faults related to dissolved gases in oil. Leveraging an improved Gramian angular field, one‐dimensional fault samples are converted into two‐dimensional feature images and data augmentation is implemented to meet the input requirements of deep learning models. Building upon visual geometry group (VGG)19 and residual networks (ResNet)50 networks for fault diagnosis, sparsity techniques are introduced through pruning, the fusion of convolution layers and batch normalization layers, and parameter quantization. Numerical experiments and performance evaluations on dissolved gas in transformer oil fault data demonstrate that the proposed method effectively reduced model complexity, minimized parameter count, conserved computational resources, and improved processing speed while maintaining a considerable level of fault identification accuracy. This made it applicable to edge computing platforms characterized by small form factors and low power consumption in the power industry.

Reliable operation of oil-immersed power transformers is crucial for electrical transmission and distribution networks. However, the aging of high voltage assets including power transformers along with the increasing of load demand have heightened the importance of adopting cost-effective asset management strategies. Dissolved gas analysis (DGA) has been recognized as a valuable diagnostic tool for detecting potential faults and monitoring the condition of oil-immersed power transformers. Traditional offline DGA method involves periodic sampling and laboratory analysis, which often results in delayed detection and response to emerging faults. To address these limitations, online DGA approach has been emerged to provide real-time monitoring and continuous data acquisition. This paper presents a new asset management approach for mineral oil-immersed power transformers by analysing the online DGA data using convolutional neural networks. The proposed approach provides real time solutions to classify emerging fault type and predict transformer health deterioration level with high accuracy. Results show that the accuracy of fault diagnostics of the proposed approach is approximately 87%.

Abstract Cascaded H‐bridge power electronic transformers (CHB‐PET) can efficiently connect AC/DC distribution networks, realizing power transmission and electric energy conversion. Once a pole‐to‐pole DC fault occurs on the low‐voltage direct current (LVDC) side of CHB‐PET, the fault current has a fast rise and high amplitude, which easily leads to the rapid blocking of power electronic devices and makes it difficult for DC line protection to obtain fault information. A stepwise blocking scheme of PET is proposed for LVDC network protection. When the fault on the LVDC side of PET is detected, cascaded H‐bridge and low‐voltage H‐bridge of dual‐active‐bridge are blocked first, and then high‐voltage H‐bridge of dual‐active‐bridge is blocked after a 5 ms delay. The stepwise blocking can provide fault information for DC line protection. On this basis, wavelet packet transform is employed to extract different frequency components of fault current. The energy of the specific frequency band corresponding to the switching frequency of DAB and the total energy of all frequency bands contained in the fault current is calculated; then their ratio can be utilized to construct the protection scheme. Finally, the PSCAD/EMTDC simulation results verify the feasibility and effectiveness of proposed protection scheme based on stepwise blocking of CHB‐PET.

Abstract Transformer winding turns often consist of multiple parallel strands. The spatial position variation of each strand affects the leakage inductance of each branch, resulting in an uneven distribution of short‐circuit currents within the winding turns. And this unevenness persists even when transposition structures are implemented. Traditional methods in transformer analysis frequently overlooked the distribution characteristics of short‐circuit currents when calculating electromagnetic forces. A frequency‐domain calculation method for analysing the current distribution in winding turns was proposed, with a deviation of less than 3% compared to existing analysis methods. Two typical 110 kV transformer models were utilised to investigate the influence of uneven current distribution on the spatial distribution of electromagnetic forces. The spatial distribution of short‐circuit electromagnetic forces in low‐voltage (LV) windings exhibited significant changes, with maximum change rates of 10% and 61.2% for axial and radial electromagnetic force, respectively, in a LV winding with 4 parallel strands. The research also analysed how strand radial width and axial height affect current distribution unevenness and proposed specific design principles to mitigate these disparities in winding design. The findings offer valuable insights for selecting structural parameters and assessing short‐circuit stability during transformer design.

Degradation of insulation paper is a key contributor to the failure of power transformers. Insulation degradation accelerates at elevated temperatures, which highlights the potential for better thermal management to prolong life. While several studies have analyzed the benefits of high thermal conductivity oil for reducing temperatures inside a transformer, this study is an initial assessment of the benefits of high thermal conductivity paper on transformer life. Blending particulates with cellulosic fibers offers a pathway for high thermal conductivity paper (with good dielectric properties), which can reduce internal temperatures. Presently, life extensions that can be achieved by the use of such thermally conducting papers were estimated, with the thermal conductivity of the paper being the key parameter under study. The analytical-numerical thermal model used in this study was validated against experimental measurements in a distribution transformer, adding confidence to the utility of the model. This model was then used to provide estimates of hot-spot temperature reduction resulting from the use of papers with higher thermal conductivity than baseline. Transformer life was predicted conventionally by tracking the degree of polymerization of paper over time, based on an Arrhenius model. Results indicate that increasing the thermal conductivity of paper from 0.2 W/mK (baseline) to 1 W/mK reduces the hot spot temperature by 10 °C. While degradation significantly depends on the moisture and oxygen content, the model shows that such a temperature reduction can increase life for all conditions, by as much as a factor of three.

Tasmanian devil optimization (TDO) algorithm represents one of the most recent optimization algorithms that were introduced based on the nature behavior of Tasmanian devil behavior. However, as a recent optimizer, its performance may provide inadequate balance among the exploitation and exploration abilities, especially when dealing with the multimodal and high-dimensional natures of optimization tasks. To overcome this shortage, a novel variant of the TDO, called improved Tasmanian devil optimization (ITDO), is introduced in this paper. In ITDO, two competitive strategies are embedded into TDO to enrich the scope of the searching capability with the aim of improving the diversification and identification of the algorithm. The effectiveness of the ITDO algorithm is examined by validating its performance on CEC 2020 benchmark functions with different landscape natures. The recorded results proved that the ITDO is very competitive with other counterparts. After ITDO exhibited a sufficient performance, then, it was applied to estimate the parameters of the 1 kVA, 230/230 V, single-phase transformer. Some assessment metrics along with convergence analysis are conducted to affirm the performance of the proposed algorithm. The recorded results confirm the competitive performance of the proposed method in comparison with the other optimization methods for the benchmark functions and can identify the accurate parameters for the single-phase transformer as the estimated parameters by ITDO are highly coincident with the experimental parameters. [ABSTRACT FROM AUTHOR]

Diffusion tensor imaging (DTI) is a non-invasive technique for analyzing the movement of water in the brain. However, the precision of measurements required for tracking white matter pathways can lead to long scan times, which can be challenging for some patient populations such as pediatric patients. To address this issue, researchers have been experimenting with deep learning techniques for faster estimation of DTI parameters, which are helpful in neurological diagnosis, of diffusion-weighted images. Our proposed solution is a transformer neural network-based approach for fast estimation of diffusion tensor parameters using sparse measurements. While there have been attempts to address this problem, our proposed model handles both scalable and generalized estimation of DTI parameters using multiple sparse measurements. Through experimentation on the Human Connectome Project (HCP) Young Adult benchmark dataset, our proposed model demonstrated state-of-the-art results in terms of fractional anisotropy (FA), axial diffusivity (AD), and mean diffusivity (MD) when compared to traditional linear least square (LLS) fitting and 3D U-Net model with 16 × 16 × 16 input size (3D U-Net16). [ABSTRACT FROM AUTHOR]

Partial discharge evaluation is a principal method for assessing insulation conditions in power transformers. Traditional singular value decomposition (SVD) approaches, however, face issues like high residual noise and loss of signal details in white noise suppression. This article introduces an advanced denoising algorithm integrating SVD, variational mode decomposition (VMD), and wavelet thresholding to effectively address mixed noise in on-site power transformer assessments. The algorithm initially employs SVD to suppress mixed noise, specifically targeting narrowband interference by decomposing the noisy signal and nullifying the corresponding singular values. Post-SVD, the signal is further processed through VMD, with its modal components refined via wavelet thresholding. The final reconstruction of these denoised components effectively eliminates white noise. Applied to an input signal with a signal-to-noise ratio of -27.593 dB, the proposed method achieves a postdenoising ratio of 13.654 dB. Comparative analysis indicates its superiority over existing algorithms in mitigating white noise and narrowband interference and more accurately restoring the partial discharge signal. [ABSTRACT FROM AUTHOR]

The reliability and stability of differential protection in power transformers could be threatened by several types of inferences, including magnetizing inrush currents, current transformer saturation, and overexcitation from external faults. The robustness of deep learning applications employed for power system protection in recent years has offered solutions to deal with several disturbances. This paper presents a method for detecting internal faults in power transformers occurring simultaneously with inrush currents. It involves utilizing a data window (DW) and stacked denoising autoencoders. Unlike the conventional method, the proposed scheme requires no thresholds to discriminate internal faults and inrush currents. The performance of the algorithm was verified using fault data from a typical Korean 154 kV distribution substation. Inrush current variation and internal faults were simulated and generated in PSCAD/EMTDC, considering various parameters that affect an inrush current. The results indicate that the proposed scheme can detect the appearance of internal faults occurring simultaneously with an inrush current. Moreover, it shows promising results compared to the prevailing methods, ensuring the superiority of the proposed method. From sample N–3, the proposed DNN demonstrates accurate discrimination between internal faults and inrush currents, achieving accuracy, sensitivity, and precision values of 100%. [ABSTRACT FROM AUTHOR]

Natural and synthetic esters are increasingly being chosen instead of mineral oil for environmental and fire safety reasons. However, their use in power transformers is limited due to insufficiently well-understood ageing processes affecting their properties and the properties of cellulosic materials impregnated with them. The research results presented in many scientific papers prove that the use of esters slows the ageing process of cellulosic materials. This article presents the results of research aimed at answering the question of whether the effect of slowing the ageing process will also occur in the case of insulation with very low initial moisture. The answer to this question will allow us to better understand the role of water in the ageing process of the transformer's insulation system. The thermal ageing process was carried out at a temperature of 150 °C in closed systems. The degree of cellulose polymerisation was taken as a measure of the degree of paper ageing. Great attention was paid to measuring the water content in both paper and electro-insulating liquids at various stages of their ageing. Furthermore, measurements of the properties of electro-insulating liquids were made, which are considered markers of ageing. The test results obtained indicate that in the case of a dry insulation system, corresponding to the initial moisture of the new units, the type of liquid used to impregnate the cellulosic material does not significantly affect its depolymerisation process. However, in the case of paper impregnated with natural esters, a lower dynamic of moisture increase in insulation was observed. [ABSTRACT FROM AUTHOR]

Power transformers play a critical role in the performance of power systems. This equipment is costly due to significant copper and iron prices and manufacturing costs. Therefore, maintenance and protection of such equipment is essential. Despite its robust performance, maloperation of differential protection (DP) in transformers may cause operational challenges to power system operators. The differential relay may operate incorrectly after the transformer energization leads to an inrush current (IC) and the relay identifies the event as an internal fault, and consequently issues the trip command. The other case of maloperation includes, but not limited to, a moment when the current transformer saturates due to an external fault. In this paper, a novel approach for DP is proposed, that is based on signal processing methods. In this paper, variational mode decomposition (VMD) and the deep neural network are implemented by using the convolutional neural network (CNN) and bi‐directional long short‐term memory (BiLSTM) models. The VMD decomposes differential current signal (DCS) to intrinsic mode functions with corresponding narrow‐band property frequency spectrums, which provides more detailed information about signal characteristics in different frequency bands. At the next stage, an effective feature for the BiLSTM is extracted by the CNN with the convolutional layers to classify events and proper discrimination. Extensive simulations on a 500 MVA transformer in MATLAB demonstrate the effectiveness of the proposed protection approach to differentiate ICs from internal and external faults with 99.8% accuracy in less than 1/8th of a power cycle. [ABSTRACT FROM AUTHOR]

Since compact high temperature superconducting (HTS) transformer has a small size and high efficiency, it can meet the requirements of high‐voltage transformers for vehicle‐mounted mobile substation. However, due to the small internal insulation size of the compact HTS transformers, it is necessary to improve the insulation strength to ensure the insulation safety of the transformer. Reasonable main insulation structure can greatly improve the insulation level of transformer. Based on the design theory of main insulation structure, this paper designs a main insulation structure model "small liquid nitrogen gap—barrier" suitable for compact HTS transformer. The insulation withstand voltage experiment of the main insulation structure shows that when the liquid nitrogen gap distance is less than 8 mm, the insulation strength of liquid nitrogen in the structure is improved. When the liquid nitrogen gap distance is too small, the size and thickness of the barrier will affect the insulation strength of the structure. According to the insulation strength of main insulation structures with different liquid nitrogen gaps, the feasibility of "small liquid nitrogen gap—barrier" main insulation structure is verified. The small liquid nitrogen gap structure can be adopted for the main insulation of compact HTS transformer. [ABSTRACT FROM AUTHOR]

The inability of power transformers to resist a short-circuit (SC) fault significantly impacts the safety of power systems. During SC faults, the leakage field undergoes significant distortion, resulting in large SC electromagnetic forces (EFs), which results in critical mechanical stresses. The finite element method (FEM) might be one of the best non-destructive numerical methods that enable designers to assess the transformer SC performance. Using the FEM, most SC performance analysis studies assume either the three-phase (LLL) or the single-phase-to-ground (LG) SC fault. Moreover, the existing studies use an excitation method that provides a high error rate due to the sensitivity of transient solvers. Furthermore, no considerable attention has been given to analyzing the influence of grounding the wye-connected windings on the SC performance. In this work, a non-linear-transient field-circuit coupled 3-D finite element model is created to analyze the SC performance of a delta-wye connected distribution transformer under various SC fault conditions. Results show that grounding the secondary windings significantly impacts the performance when a LG fault is applied. Except for the LG fault with wye configuration of the secondary windings, the windings SC current and EFs decay/rise over time with an exponential trend till reaching steady state. [ABSTRACT FROM AUTHOR]

Power transmission systems are made up of transmission lines and substations, and the main equipment in a substation are power transformers. Power transformers correspond to about 60% of the total cost of substations and are the most complex, being composed of several thermal, mechanical, electrical, and chemical elements and processes. Due to this complexity, unscheduled shutdowns occur in these transformers, which leads to the need for periodic maintenance and replacement. Within this context, this article proposes a methodology to analyze the costs generated by additional aging in transformers when subjected to an additional load caused by the shutdown of another transformer in the system. The methodology was validated in a case study carried out in a company in southern Brazil. The results show that the cost of unavailability is considerably higher when scheduled outages occur. It should also be noted that the 5 transformers studied are responsible for 93% of the unavailability cost, and the common fact among these is that the average normal load is above 50%. In this sense, it can be concluded that the methodology can assist in decision-making related to the management and maintenance of this critical equipment causing significant economic consequences in the long run. [ABSTRACT FROM AUTHOR]

Part of power transformers failures are result of great electromagnetic forces acting on its windings due the circulation of high currents, as in case of short circuits. To contribute for a more reliable mechanical design, this paper presents a finite element analysis (FEA) for a complete calculation and evaluation of electromagnetic forces on transformers windings. Three arrangements of concentric windings are analyzed considering three-phase and three-dimensional models, dealing with issues as the influence of magnetic core window and lateral phases presence, ensuring more accurate results. The adopted modeling allows to determinate the total axial and radial forces on all the phases and its distribution over the height of windings and around the core leg. It was possible to identify points of greater and lesser forces and the behavior of forces on different constructive arrangements, allowing its choice improvement during the design stage. The results present good agreement to the analytical method, allowing the application of the developed FEA for transformer design optimization, contributing for a time and costs reduction and avoiding future mistakes in its construction and failures during its operation. [ABSTRACT FROM AUTHOR]

Reliability and long lifetime of on-load tap changers are very important issues for the power transformers. In this work, a novel model switch is developed to simulate the switching process of diverter switches and flexibly adjust the individual contact gap. Then, the contact voltage, contact current and mechanical vibration acceleration are measured simultaneously to reflect the switching characteristics. Further, the complete contact tap time sequences during switching process are obtained based on the joint electrical and mechanical vibration acceleration responses. In addition, the effects of arcing contact gap and transition contact gap on the contact tap time, main vibration frequency and power spectral density are investigated experimentally. The results could be used to discriminate the possible contact wear defects and associated electrical contact fault. [ABSTRACT FROM AUTHOR]

The key to intelligent fault diagnosis for transformer is to find relevant dissolved gas characteristics with the capability of representing different types of faults. However, the existing problem is that using a single traditional ratio method as the inputs of classification module cannot obtain high classification performance. Therefore, a fault diagnosis method for transformer based on neighborhood component analysis (NCA) and capuchin algorithm (CapSA)-regularized extreme learning machine (RELM) is proposed. Firstly, 21-dimensional features of the oil-immersed transformer are extracted based on the traditional ratio methods such as IEC three-ratio and Roger's ratio. Secondly, The NCA is used to select the feature to obtain an effective 6-dimensional feature. Then, these features are utilized as the input of the RELM model for training, and the parameters of model are optimized by CapSA. Finally, the feasibility and validity of the proposed method are verified by actual data sets collected from previous publications. The experimental results show that the proposed method has higher diagnostic precision and stability. [ABSTRACT FROM AUTHOR]

The transformer is a fundamental piece of equipment for power grids. The analysis and optimization of their life-cycle costs are of great importance to reinforce the economic efficiency of electrical networks. This paper constructs a comprehensive transformer life-cycle cost (LCC) model by fusing life-cycle cost theory with relevant transformer expenditure. It proceeds to examine the life-cycle cost aspects of the transformer, delving into its cost dynamics under various influencing factors, establishing interconnections between these factors and analyzing the cost relationship. By employing MATLAB software (Matlab 2021a) along with the whale optimization algorithm (WOA) this paper optimizes the objective function. Through this, it establishes the LCC model for 20 power transformers, obtaining the optimal objective function curve and the maximum value for LCC optimization of the transformer. Unlike previous research, this study adds a detailed analysis of several factors that influence LCC. At the same time, it develops a more complete, scientific and rational LCC optimization model. An illustrative example validates the model and the superiority of the whale optimization algorithm. The algorithm not only shows its scientific basis and superiority, but also serves as a guiding mechanism for LCC management in transformer engineering practices. Ultimately, it emerges as a fundamental tool to improve the efficiency of power grid asset management. [ABSTRACT FROM AUTHOR]

The objective of the paper is to analyse the effects of various geometrical and operating parameters on the liquid flow distribution in OF-cooled power transformers. Our investigation includes two cases: one with a simplified winding geometry and another that closely resembles the actual winding geometry. The analyses were carried out using computational fluid dynamics (CFD) and custom, internally developed thermo-hydraulic models. Our findings confirm that buoyancy forces rather than the pump drive the liquid flow within the windings of an OF-cooled power transformer. The results also show that the liquid flow distribution, which is influenced by the winding geometrical properties and liquid properties, has a significant impact on the hot-spot temperatures of the windings. The comparison between the results of the CFD simulations and the results of the simple model demonstrates a high level of agreement in calculating both the mass flow rates and temperatures. [ABSTRACT FROM AUTHOR]

The insulated core transformer (ICT) power supply is widely employed in electron beam accelerators (EBAs) due to its high power, heightened efficiency, and stable operation. However, the segmented-core structure of the ICT power supply increases magnetic leakage, which leads to it adversely affecting the consistency of the output voltages in the rectifier stages. Currently, numerous studies focus on stage voltage compensation, including turns compensation, capacitor compensation, dummy primary winding compensation, and full-parameter compensation. This paper presents a unified simulation model and an improved gradient-based genetic algorithm, which can also optimize the parameters of the four compensation methods. Based on this, the performance of the power supply using the four compensation methods under different ICT energy levels and power supply requirements is studied, and the selection suggestions are given. This work fills the gap in the performance comparison and application research of various compensation methods. [ABSTRACT FROM AUTHOR]

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. [ABSTRACT FROM AUTHOR]

This paper addressed the challenges associated with the complexity, numerous parameters, computational resource demands, and slow processing speed of transformer fault identification models based on deep learning technologies. Sparse convolutional neural network (CNN) approach is proposed for identifying faults related to dissolved gases in oil. Leveraging an improved Gramian angular field, one‐dimensional fault samples are converted into two‐dimensional feature images and data augmentation is implemented to meet the input requirements of deep learning models. Building upon visual geometry group (VGG)19 and residual networks (ResNet)50 networks for fault diagnosis, sparsity techniques are introduced through pruning, the fusion of convolution layers and batch normalization layers, and parameter quantization. Numerical experiments and performance evaluations on dissolved gas in transformer oil fault data demonstrate that the proposed method effectively reduced model complexity, minimized parameter count, conserved computational resources, and improved processing speed while maintaining a considerable level of fault identification accuracy. This made it applicable to edge computing platforms characterized by small form factors and low power consumption in the power industry. [ABSTRACT FROM AUTHOR]