Deep learning–based surface contamination severity prediction of metal oxide surge arrester in power system

Abstract

This paper presents an advanced technique based on cross-Stockwell transform (XST) and sparse autoencoder to predict the surface contamination severity of metal oxide surge arrester (MOSA) employing leakage current signal. Generally, MOSAs in power system network are exposed to different environmental conditions where its condition may degrade due to accumulation of pollutants, which may cause premature failure of it. Hence, system reliability can get affected. Therefore, monitoring the surface condition of MOSA is very important. In this proposed technique, MOSA leakage current signals of different surface contamination severity have been cross-correlated with leakage current recorded at the clean surface in joint time-frequency plane through XST, which is an extended version of ST. Thereafter, sparse autoencoder, a deep learning framework, has been applied to extract potential deep feature from leakage current-converted XST matrices. The extracted deep features have been classified through different classifiers. It has been observed that the proposed technique yields satisfactory accuracy regarding the estimation of surface contamination severity of MOSA. Therefore, the proposed method can be implemented to monitor the surface condition of MOSA, and it may be applied for topologically similar problems.

Leakage Current Measurements of Surge Arresters

“…The work [22] lists innovative extensions created after establishing the content of Appendix D of the standard. These include works [24][25][26][27][28], which analyze various techniques for analyzing the current signal-both with only the total leakage current and additionally with a separate resistive component, allowing the amplitude and shift angles in the current signal for the components to be obtained-total and/or resistive [24][25][26], as well as leading to the determination of the shape and similarity of current signals recorded in a synchronous manner [27,28]. On the other hand, in the work [23], a number of improved classical methods have been compiled that allow its resistive component to be obtained from the current signal, specifying the limitations of their use, of which the presence of higher harmonics in the supply voltage was indicated in the first place.…”

Application of Deep Learning and Symmetrized Dot Pattern to Detect Surge Arrester Status

“…Amorim et al [10] proposed using a high frequency current transformer (HFCT) to sense the grounding cable conductor current of a surge arrester to obtain its PD signal, and used the phase resolved partial discharge (PRPD) spectrum to evaluate its state. Das et al [11] proposed utilizing the leakage current of surge arresters to identify the severity of surface contamination based on cross-stockwell transform and a sparse autoencoder. Metwally et al [12] proposed an online monitoring method using prony analysis-Hilbert transform based on feature extraction of the frequencies, phase angles, and magnitudes of all frequency components for the surge arrester considering pollution and dry conditions when measuring the total leakage current.…”