Methods of Mathematical Statistics in Analyzing Probabilistic Characteristics of External Insulation

The article presents the results of studying the methods of mathematical statistics in analyzing two probabilistic characteristics of external insulation that differ fundamentally from each other in physical nature, which are aimed to solve essentially different and non-overlapping practical problems. It is shown that the positive results of standardized tests of external insulation by the 15-shock method of acting impulse voltage are not able to guarantee reliable operation of the tested samples in operation. It is noted that the objective of ensuring the required level of insulation reliability is solved during its design by increasing the number of acting impulse voltages at each level of the experimental effect curves. The minimum necessary and sufficient number of experiments that have to be carried out at each level of the effect curve has been established, at which the necessary operational insulation reliability is ensured while avoiding an excessive electrical strength margin with a guaranteed positive result of standardized tests. It is concluded that it is unacceptable to identify the effect curve with the probabilistic characteristic of insulation that deals with the discharge voltage as a random variable.