Statistical analysis of pit formation in an artificial void with Weibull distribution

Abstract

In this study, an author made an artificial void model which was based on the Whitehead abc model and discussed properties of pit formation in the artificial void. In order to investigate influence of residual charges in the artificial void on the pit formation, the distribution was measured with dust figure technique. The results suggest that the negative charges on the void surface play an important role in the pit formation. Voltage dependence of time to the pit formation (V-t characteristics) was obtained. This dependence is realized as linear characteristics which bend at 11 kV and 16 kV. Each linear characteristic is expressed as t/spl prop/ V/sup -n/. An interesting point is that the V-t characteristics become discontinuous at certain voltage between 12 kV and 13 kV. This is because surface discharge patterns in the artificial void change from Polbuschel to Gleitbuschel. Furthermore, the time to the pit formation was analyzed with Weibull probability distribution in order to investigate the pit formation process. In the case of Polbuschel, transition from fatigue-failure-type degradation process (shape parameters m > 1) to random-failure-type degradation process (m = 1) occurs when applied voltage exceeds 11 kV. In the case of Gleitbuschel, the similar transition takes place at 16 kV: the transition from fatigue -failure-type degradation process to early/random-failure-type degradation process (m /spl les/ 1).