Partial Discharge Activity Diagnosis in Electrical Cable Terminations Using Neural Networks

Abstract

The electrical cable is an essential equipment in electricity transportation and distribution which are generally carried out using its types consisting in overhead lines or underground cables. The terminations are among the primordial accessories in electrical cable, they are intended to cover and protect the ends presenting the weak point of failure of an electrical cable because in these zones the electrical stress increases so much. Consequently, partial discharge activity will eventually occur and develop into arcing (Elaggoune et al. in Deuxieme Conference Internationale Sur La Maintenance, La Gestion, La Logistique Et L’electrotechnique, Enset D’oran, Algeria pp 1–5, 2012; Elaggoune in Localisation des sources potentielles des decharges parasites creees, par un dispositif experimental de detection des Decharges Partielles dans un câble de l’energie electrique.memoire de magister, Universite de Laghouat, Algerie, 2012; Goyal in Ind Eng Lett 4:15–20, 2014). Therefore, the terminations must be diagnosed periodically in order to avoid an insulation system breakdown, which causes a power outage and consequently excessive economic losses. The objective of this work is to assess the health of electrical cable terminations usingthedissolved gas analysis which is an effective chemical method toguarantee a detection of partial discharges in cable terminations (Samsudin et al. in Incipient fault detection in 33/11kV power transformers by using combined dissolved gas analysis technique and acoustic partial discharge measurement and validated through untanking, Modern Electric Power Systems (MEPS), Wroclaw, Poland, pp 1–6, 2010; Faisal et al. in Int J Electr Eng Inform 4: 186–201, 2012; Kumpulainen in Aspects and directions of internal arc protection, Doctoral Thesis, University Of Vaasa, Finland 2016; Vahamaki in Arc protection as integrated part of line protection relays, See http://www.cigaproject.ch/wp/docs/Arc%20protection%20integrated%20in%20protection%20relays.pdf , 2020). However, this method failed to identify this phenomenon in some cases. To remedy with this problem, Radial Basic Functions neural network is applied as simulation tool. Therefore, the Radial Basic Functions neural network, based on dissolved gas analysis, can be used successfully in monitoring electrical cable terminations against partial discharges phenomenon.