Takeshi Endoh

Pattern recognition of partial discharge in XLPE cables using a neural network

Describes an experimental study on pattern recognition of a partial discharge (PD) in a cross-linked polyethylene (XLPE) cable by using a neural network (NN) system. The NN adopted was a three-layer artificial neural system with feedforward connections, and its learning method was a back-propagation algorithm incorporating an external teacher signal. Input information of the NN was a combination of the discharge magnitude, the number of pulse counts, and the phase angle of the PD. The PD measurement was carried out using a 66 kV XLPE cable with an artificial defect under an AC applied voltage of 38 kV. After learning 30 typical input patterns, the NN discriminated unknown patterns, with and without PD, with an accuracy of 90%. In a long-term performance test of a 66 kV XLPE cable with a defect, the NN alarm processor was able to recognize the presence of a PD about one hour before breakdown of the cable, and successfully alerted the operator.<<ETX>>

Experimental study on the causes of electrical breakdown of EHV XLPE insulation using 'prebreakdown partial discharge detection method'

The causes of electrical breakdown in EHV XLPE (cross-linked polyethylene) cables were clarified by carrying out the prebreakdown partial discharge (PBPD) detection test under electric stresses from 40 to 70 kV/mm. Ten causes of electrical breakdown were observed in the PBPD test. As expected, it was found that they were associated with very small contaminants and protrusions. The electrical tree inception stress was estimated to be 270 kV/mm on the average. The prebreakdown time was estimated to be on the order of 1 hour under an applied electric field of 5 kV/mm.<<ETX>>