Real-time insulation status assessment based on standard deviation for underground cable joints

The purpose of this paper is to analyze the relationship between the partial discharge and insulation status of underground cable joint. The insulation diagnosis mechanism is established by the characteristics of the probability distribution of discharge parameters. The data process of partial discharge consists of filtering and reduction, parameter extraction and parameter estimation of probability model. Four probability models were utilized, such as, Normal distribution, Log-Normal distribution, Weibull distribution and Gamma distribution. The failure of the goodness-of-fit test accompanied with huge variation of discharge parameters and striking distribution parameters. These outliers might imply that a structural variation might happen and could be used to analyze the insulation mechanism. Finally, the Normal distribution parameter for the weight of discharge region is chosen to be a feature. The threshold of insulation status is the upper limit of confidence level calculated by three standard deviations. The result shows that the outliers appeared at two sites which divide deterioration of the insulation into tree partitions, such as initial term, midterm, and final term.