A Setting-Free Differential Protection for Power Transformers Based on Second Central Moment

Esponda, Héctor; Vázquez, E.; Andrade, Manuel A.; Johnson, Brian K.

doi:10.1109/tpwrd.2018.2889471

Cited by 29 publications

(7 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Differential protection has been the main protection scheme in power transformers because of its simple principle and efficiency . When a power transformer is energized initially or experiences a sudden change in its terminals, a typically non-sinusoidal transient current flows in transformer windings which abruptly reaches a large value in the first half cycle [2].Although differential protection has been widely M.Afrasiabi, S.Afrasiabi, B.Parang, M.Mohammadi and H.Samet are with the School of Electrical and Computer Engineering, Shiraz University, Shiraz, Iran (e-mail: shpower77@yahoo, musa.afra@shirazu.ac.ir, benyamin.parang@shirazu.ac.ir, m.mohammadi@shirazu.ac.ir, samet@shirazu.ac.ir).…”

Section: A Motivationmentioning

confidence: 99%

Fast GRNN-Based Method for Distinguishing Inrush Currents in Power Transformers

Afrasiabi

Parang

et al. 2022

IEEE Trans. Ind. Electron.

View full text Add to dashboard Cite

Differential protection, as the key protection element in the power transformers, has always been threatened with sending false trips subjected to external transient disturbances. As a result, differential protection needs an additional block to distinguish between internal faults and external transient disturbances. The protection system should i) be able to perform based on raw data, ii) be able to learn fully temporal features and sudden changes in the transient signals, and iii) impose no assumption on noise. To address these challenges, a fast recurrent neural network, namely fast gated recurrent neural network (FGRNN). By removing reset gate in the gated recurrent unit (GRU), the proposed network is capable of learning abrupt changes in addition to significantly reducing the computational time. Furthermore, a loss function based on an information theory concept is formulated in this paper to enhance the learning ability as well as robustness against non-Gaussian/Gaussian noises. A generalized form of mutual information is also adopted to form a noise model-free loss function, then incorporated with the designed deep network. Simulated and experimental examinations engaging various external factors, in addition to comparison between the proposed fast GRNN, GRU and seven firmly-established methods indicates the faster and more reliable performance of the proposed algorithm.

show abstract

Section: A Motivationmentioning

confidence: 99%

Fast GRNN-Based Method for Distinguishing Inrush Currents in Power Transformers

Afrasiabi

Parang

et al. 2022

IEEE Trans. Ind. Electron.

View full text Add to dashboard Cite

show abstract

“…The new industrial park is an industrial estate with a certain social culture, and its essence is flexible specialization. It is inevitable for companies to choose integration to reduce transaction costs [4][5].…”

Section: Definition Of Industrial Spatial Distributionmentioning

confidence: 99%

Gravity Index Algorithm of Industrial Spatial Agglomeration Center Based on Distributed Collaboration

2021

DPS

View full text Add to dashboard Cite

In order to measure the degree of industrial agglomeration as well as the spatial location of industrial agglomeration, this paper proposes the central gravity index method by integrating two kinds of industrial spatial agglomeration measurement methods based on output value and distance. The purpose of this research is to design a central gravity index algorithm for distributed collaborative industrial space integration. In the present study, the data of the industrial data of North -West, Tianjin, Hebei, Yangtze River Delta and Zhu Jiang Delta were surveyed, and the data were verified and verified using the industrial accumulation measurement and the central gravity index algorithm. The results show that SVM has the strongest learning ability, and has good performance in precision (84%), recall (84%) and f1measure (84%); the learning ability of KNN is second only to SVM, and precision (82%), recall (82%) and f1measure (82%) have good performance. It is concluded that the central gravity index algorithm of spatial agglomeration in this study improves the existing industrial spatial agglomeration measurement methods, and makes up for the defects that the output value measurement method ignores the spatial distribution, while the distance based method does not use enough output weight and cannot determine the geographical location; through the geographical location of virtual center enterprises, it provides a new method for tracking the geographical path of industrial transfer, and contribute to the research of regional integration.

show abstract

“…Several techniques based on artificial intelligence, digital signal processing, and probability theory and statistics have been developed in order to improve the efficiency of transformer differential protection schemes [8]- [11]. For instance, in [8] a method based on the Clarke transform and fuzzy logic to single out internal faults on power transformers is presented.…”

Section: Introductionmentioning

confidence: 99%

“…Research reported in [9] and [11] used high-order statistics and the second central moment techniques to distinguish between internal faults and energizations, respectively. Both methods presented good success rates in the event discrimination, but only [11] reported adjusts with no dependence of the transformer and system parameters. However, these methods did not consider transformer overexcitation, which is an important situation for validating new power transformer protections.…”

Section: Introductionmentioning

confidence: 99%

A Clarke-Wavelet-Based Time-Domain Power Transformer Differential Protection

Medeiros

Silva

Popov

et al. 2022

IEEE Trans. Power Delivery

View full text Add to dashboard Cite

Phasor-based differential protection is widely used as the main protection function of the power transformer due to its reliability and ability to discriminate internal from external faults and inrush currents. However, these methods present some delays due to phasor convergence during the fault occurrence and can fail during challenging situations, such as transformer energizations with low second-harmonic content and turn-to-turn and turn-to-ground internal faults. This paper proposes a novel time-domain power transformer differential protection based on Clarke and wavelet transforms with only one differential unit and with automatic setting to be used in any power transformer. Considering both actual and simulation data, the performance validation reveals that the proposed method is efficient, ultrafast, simple, and independent of the fundamental and harmonic components of the differential current. The method was also implemented in a real-time digital simulator to demonstrate its practical feasibility.

show abstract

A Setting-Free Differential Protection for Power Transformers Based on Second Central Moment

Cited by 29 publications

References 36 publications

Fast GRNN-Based Method for Distinguishing Inrush Currents in Power Transformers

Fast GRNN-Based Method for Distinguishing Inrush Currents in Power Transformers

Gravity Index Algorithm of Industrial Spatial Agglomeration Center Based on Distributed Collaboration

A Clarke-Wavelet-Based Time-Domain Power Transformer Differential Protection

Contact Info

Product

Resources

About