Modeling of stress-enhancement at defects inside cable insulation

AC breakdowns are commonly used as a performance indicator for power cables, and yet the data generated can be misinterpreted if the test section of cable is found to have a manufacturing defect. As an example, in a recent publication, dielectric a breakdown value on a 17-year field aged cable was treated as a "suspension" by the authors after the discovery of a conductor shield skip at the failure location; and yet, others have taken the same data set and have analyzed the uncensored data as indicative of the material performance. In an effort to resolve the differences, the same data set is considered here, with an assumption that degree of aging is not significantly impacted by any enhancement of field-aging stresses. Statistical analysis is performed to determine if the questionable breakdown value can be considered an outlier. A 2-dimensional finite element analysis based upon the shape of the defect enables an estimate of the local stress enhancement factor, and a "corrected" breakdown value is calculated. The original authors conservative treatment of the questionable breakdown value as a "suspension" is supported by analysis with inclusion of a stress-corrected breakdown value, and the two approaches yield similar failure distributions. Use of the uncorrected value in discussions related to failure probabilities on the low-stress side of the distribution is shown to substantially underestimate failure stresses