Analysis of discharge spectra for cavities in impregnated paper insulation under AC stresses

Establishes whether measured discharge impulses in paper-insulated cables can correctly be assumed to come from cavities where breakdown is most likely to start, and whether the discharge magnitude can be used as a valid measure of failure probability. In such cables the suspect cavities are often those formed by an unimpregnated butt-gap between the turns of a wound paper layer next to a conductor or under a screen. In the present experiments discharges were generated in model voids like these, about 2 mm wide and 0.15 mm deep and bounded on one side by a metal. The aim was to determine whether the discharge magnitudes corresponded to values that are measured in service, and to seek a relation between magnitude and life. The latter requires life-tests at high voltages. This paper discusses the voltage dependence of discharge frequency, charge transfer, and the statistical distribution of discharge magnitudes