A DC Series Arc Fault Detection Method Using Line Current and Supply Voltage

Lu, Qijing; Ye, Zeyu; Su, Mengmeng; Li, Yasong; Sun, Yuce; Huang, Hanqing

doi:10.1109/access.2019.2963500

Cited by 46 publications

(17 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To train the RNN, it is necessary to measure the difference between the true value and the estimated value of the RNN output by using the loss function. The cross-entropy function was selected as the loss function of the proposed RNN model, see (1).…”

Section: Rnn Training and Testingmentioning

confidence: 99%

“…In (1), Y is the true value,Ŷ is the estimated value. N is the number of samples, M is the number of labels, y ij is the indicator variable (0 or 1),ŷ ij is the probability that sample i belongs to category j.…”

Section: Rnn Training and Testingmentioning

confidence: 99%

“…Qiwei Lu proposed a DC series arc fault detection method by using the information on line current and supply voltage [1]. Na Qu uses four features of current in time domain and ten features of current in frequency domain as fault feature [2].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Series Arc Fault Diagnosis and Line Selection Method Based on Recurrent Neural Network

Liu

et al. 2020

IEEE Access

View full text Add to dashboard Cite

Series arc fault is a common phenomenon in the power system, it will directly affect the working reliability, but there is no mature method to detect it due to its concealment and chaos. Common detection methods that build on the arc fault eigenvectors obtained by manual analysis are subjective and incomprehensive. A series arc fault diagnosis and line selection method based on recurrent neural network (RNN) for a multi-load system was proposed in this paper. Firstly, a series arc fault experiment under a multi-load system was carried out, the training set and test set were built by using the data obtained from the experiment. Then, the RNN model was built, trained, and tested through the training set and test set. Finally, the fast-continuous detection method and the probability-based classification result correction method were proposed, and the detection speed and accuracy were improved much further. The results show that the proposed method is effective for diagnosing series arc fault and line selection under a multi-load system, without analysis of arc fault characteristics. INDEX TERMS Series arc fault, deep learning, recurrent neural network, RNN, fault diagnosis, fault line selection.

show abstract

Section: Rnn Training and Testingmentioning

confidence: 99%

Section: Rnn Training and Testingmentioning

confidence: 99%

See 1 more Smart Citation

Series Arc Fault Diagnosis and Line Selection Method Based on Recurrent Neural Network

Liu

et al. 2020

IEEE Access

View full text Add to dashboard Cite

show abstract

mentioning

confidence: 99%

“…Series arcs detection methods [4,52]. The method in [4] is based on the comparison of current drop estimates in successive short time intervals (50 µs) and with running average values on longer intervals (50 ms), showing a factor of 2 of difference in the indicators with and without arc; the required sampling is 200 kSa/s, thus potentially exposed to high-frequency pollution from static converters, then reduced by successive averaging. Similarly, in [53], min and max current values are run over optimized window lengths (5 to 25 ms), whose difference is an indicator of intermittence; the influence of system noise was not investigated, although the number of consecutive windows used in estimates can be used to filter out grid transients such as load steps, avoiding them being detected as arcs.…”

mentioning

confidence: 99%

Power Quality Phenomena, Standards, and Proposed Metrics for DC Grids

Mariscotti

2021

Energies

View full text Add to dashboard Cite

This work addresses the problem of power quality (PQ) metrics (or indexes) suitable for DC grids, encompassing low and medium voltage applications, including electric transports, all-electric ships and aircrafts, electric vehicles, distributed generation and microgrids, modern data centers, etc. The two main pillars on which such PQ indexes are discussed and built are: (i) the physical justification, so the electric phenomena affecting DC grids and components (PV panels, fuel cells, capacitors, batteries, etc.), causing, e.g., stress of materials, aging, distortion, grid instability; and (ii) the existing standardization framework, pointing out desirable coverage and extension, similarity with AC grids standards, but also inconsistencies. For the first point, each phenomenon is discussed with quantitative conclusions on relevant thresholds: in many cases some percentage of distortion (ripple) is acceptable (stress on capacitors and storage, impact on fuel cells, and PV panels), whereas in other cases, much higher levels may be tolerated (interference to protection and monitoring devices). Standards are reviewed for indications not only of low-order harmonics and voltage fluctuations typical of old DC grid schemes, but also for high-frequency noise, including thus supraharmonics and common-mode disturbance, and filling the gap with the electromagnetic compatibility domain. However, phenomena typical of EMC and electrical safety (such as various types of overvoltages and fast transients) are excluded. Suitable PQ indexes are then reviewed, suggesting integrations and modifications, to cover the relevant phenomena and technological progress, and to better follow the normative exigencies: ripple is considered in time and frequency domain, in particular with a band limited implementation; for transients and pulsed loads, more traditional indexes based on area, energy, and half duration are confronted with indexes evaluating the power trajectory and its derivative.

show abstract