Application of the Weibull Distribution for the Recognition of Transformer PD Phenomena

The problem of identification of partial discharge phenomena occurring in a transformer insulating system is addressed. Under the assumption that different PD sources can produce pulse of different PD distributions, five typical artificial defects have been designed to compare the PD distributions coming from different sources, such as internal voids, surface and corona discharges. In this aim, Weibull distribution has been introduced to estimate the parameters and the least square methods (LSM) have been used to estimate the parameters. The goodness of fit to a Weibull distribution is checked by the Cramer-von-Mises (CVM) test. It is shown that the probability function fits well the partial discharge height distribution. By this way,different discharge sources can be distinguished. by Weibull parameter. In addition, the recognition of two or more PD phenomena simultaneously active is approached by means of an additive Weibull distribution. It allows quantities related to each of the two or more superimposed PD phenomena to be estimated, according to the Weibull parameters and the standard mean quantities.

Application of the Weibull Distribution for the Recognition of Transformer PD Phenomena

The problem of identification of partial discharge phenomena occurring in a transformer insulating system is addressed. Under the assumption that different PD sources can produce pulse of different PD distributions, five typical artificial defects have been designed to compare the PD distributions coming from different sources, such as internal voids, surface and corona discharges. In this aim, Weibull distribution has been introduced to estimate the parameters and the least square methods (LSM) have been used to estimate the parameters. The goodness of fit to a Weibull distribution is checked by the Cramer-von-Mises (CVM) test. It is shown that the probability function fits well the partial discharge height distribution. By this way,different discharge sources can be distinguished. by Weibull parameter. In addition, the recognition of two or more PD phenomena simultaneously active is approached by means of an additive Weibull distribution. It allows quantities related to each of the two or more superimposed PD phenomena to be estimated, according to the Weibull parameters and the standard mean quantities.