Evaluation of spaces for the separation of signals due to multiple PD sources

Contin, A.; Pastore, Stefano; Paganin, R.

doi:10.1109/icacact.2014.7223600

Cited by 3 publications

(8 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The DWT has shown potential in discriminating between PD signals within the noisy environment of on-line applications [9] as well as accurately maintaining distinctions between PD sources from a variety of equipment [5]. It is generally followed by a dimensional reduction technique, typically principal component analysis (PCA), to reduce the dimension of the feature vector to three [1,4,5].…”

Section: Feature Extractionmentioning

confidence: 99%

“…More recently, OPTICS has been used to evaluate different spaces for PD signals [5], and showed some promising results.…”

Section: Feature Discriminationmentioning

confidence: 99%

“…OPTICS [19] is broadly similar to DBSCAN but has seen use in CM applications only very recently [5]. It has been chosen as it produces results that are visually interpretable and thus it can provide user insight into a data set that would not be possible with other clustering algorithms.…”

Section: Opticsmentioning

confidence: 99%

“…The current method of selecting the most appropriate classification (or clustering) algorithm for a specific data set remain a largely experimental-based approach [2]. Of the plethora of available techniques, PD source classification has been achieved using k-means [1], fuzzy-classifiers [3], Density Based Spatial Clustering of Applications with Noise (DBSCAN) [4], Ordering Points to Identify Clustering Structure (OPTICS) [5], support vector machines [6], and probabilistic neural networks [7]. By contrast, the feature extraction stage has seen a more restricted selection, with the main methods using either a time-frequency (TF) approach [8], or the use of a discrete wavelet transform (DWT) [9].…”

Section: Introductionmentioning

confidence: 99%

“…By contrast, the feature extraction stage has seen a more restricted selection, with the main methods using either a time-frequency (TF) approach [8], or the use of a discrete wavelet transform (DWT) [9]. More recently, the current methods of feature extraction have been classified into so-called T1, T2, and T3 spaces (using global, mean, and single valued parameters respectively), and an evaluation of their performance carried out [5]. The results of this analysis were that the T2 (such as the DWT) and T3 spaces showed similar performance, although T3 was marginally better.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Methods for wavelet-based autonomous discrimination of multiple partial discharge sources

Nimmo

George

Lewin

2017

IEEE Trans. Dielect. Electr. Insul.

View full text Add to dashboard Cite

Recent years have seen increased interest in the application of on-line condition monitoring of medium voltage networks as the need to maintain and operate ageing cable networks increases. Detection and analysis of partial discharge (PD) activity is generally used as an indicator of pre-breakdown processes that may be indicative of insulation degradation over time. A significant challenge for on-line monitoring is discrimination between multiple partial discharge sources that will often naturally exist in the data. To discriminate between PD sources each PD signal is represented as a feature vector and a clustering algorithm is used to identify clusters in the resulting feature vector space, often after dimensional reduction. Each cluster identified in the data corresponds to a distinct PD source. In this work a comparison of clustering algorithms and dimensional reduction techniques is performed to identify clusters for a variety of PD data sets, in all cases the feature vector is created using wavelet decomposition energies. The three clustering algorithms used were Density Based Spatial Clustering of Applications with Noise (DBSCAN), Ordering Points to Identify Clustering Structure (OPTICS) and Simple Statistics-based Near Neighbour clustering technique (SSNN). The two dimensional reduction techniques considered were Principal Component Analysis (PCA) and t-Distributed Stochastic Neighbour Embedding (t-SNE). At present the most commonly used combination of dimensional reduction technique and clustering algorithm for PD data are PCA and DBSCAN respectively. From the comparison performed in this work it was found that t-SNE combined with OPTICS or SSNN were the most successful at clustering PD data. For use in practical situations SSNN is preferred over OPTICS as it requires only a single input parameter, which generally be hardcoded, leading to a completely autonomous technique. It is therefore suggested that a combination of t-SNE and SSNN is particularly appropriate for discriminating PD sources.

show abstract

Section: Feature Extractionmentioning

confidence: 99%

“…More recently, OPTICS has been used to evaluate different spaces for PD signals [5], and showed some promising results.…”

Section: Feature Discriminationmentioning

confidence: 99%

Section: Opticsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Methods for wavelet-based autonomous discrimination of multiple partial discharge sources

Nimmo

George

Lewin

2017

IEEE Trans. Dielect. Electr. Insul.

View full text Add to dashboard Cite

show abstract

Automatic Evaluation of Partial Discharges Presence in Surge Arresters and Current Transformers Installed in Substations Without Equipment Shutdown

Rodrigues

Amorim

Tanscheit

et al. 2020

2020 IEEE 3rd International Conference on Dielectrics (ICD)

View full text Add to dashboard Cite

A New Feature Space for Partial Discharge Signal Separation Based on DWT Coefficient Variance

Uriel¹,

Martín²

2018

JEE

View full text Add to dashboard Cite

Abstract:In this paper, a new feature space for PD (partial discharge) signal separation is presented. Three typical PD defects were experimentally reproduced in a laboratory for obtaining independent PD sources. Signals were acquired with a digital storage oscilloscope and then post-processed with DWT (discrete Wavelet transform) for de-noising. The new feature space for PD source separation was constructed with the variance of each Wavelet coefficient vector and was compared with an established feature space for PD source separation; based on the energy of DWT coefficient vectors. After a space reduction by mean of PCA (principal components analysis), the separation capability among them was measured by comparing the final classification error after training a neural network. Results showed that with this new feature space it is possible to separate different sources of PD signals. Later, the feature space proposed was used to separate two PD sources from a real equipment tested. Further analysis on the reduced feature space has shown the band location of PD signals information for separating purpose.

show abstract

Evaluation of spaces for the separation of signals due to multiple PD sources

Cited by 3 publications

References 14 publications

Methods for wavelet-based autonomous discrimination of multiple partial discharge sources

Methods for wavelet-based autonomous discrimination of multiple partial discharge sources

Automatic Evaluation of Partial Discharges Presence in Surge Arresters and Current Transformers Installed in Substations Without Equipment Shutdown

A New Feature Space for Partial Discharge Signal Separation Based on DWT Coefficient Variance

Contact Info

Product

Resources

About