A New Statistical Theory for the Breakdown of Liquid Hydrocarbons

Abstract

Hitherto it has been accepted that breakdown under pulse conditions occurs in liquid hydrocarbons with an insignificant statistical time lag. Theories have been developed in which the whole of any time lag measured has been assigned to a formative time, but these theories are not at all satisfactory. The paper shows that (a) that a significant statistical time lag does in fact exist provided the experimental procedure is correctly interpreted and (b) that a proper statistical analysis of the previous pulse measurements together with a statistical theory of breakdown shows that these measurements provide clear evidence for rather than against, a statistical time lag. Experimental results are given to show how the statistical time lag depends on electric stress and on cathode conditions.