Statistical analysis of diagnostic indicators during an accelerated ageing experiment for XLPE cable specimens

An accelerated ageing experiment has been performed, in order to age XLPE cable samples in a wet environment to advanced levels of water tree degradation. During this accelerated ageing experiment, semi-periodic measurements were performed using high voltage Frequency Domain Spectroscopy (FDS). This paper presents the rigorous statistical analysis that was performed on these populations of measurements. Various diagnostic indicators will be compared and assessed with respect to their correlation with the duration of the ageing time. In addition to this, the effect of environmental variables, such as humidity, will also be analysed using linear regression techniques. It will be shown that certain diagnostic indicators perform better than others at following the expected positive relationship with the time of accelerated ageing. It will also be shown that the humidity plays an important role in the low frequency dielectric response, even in the presence of a guard terminal, and therefore will show the need of taking such effects into account when analysing data. Comments will also be made about the level of fluctuation of the dielectric response encountered within the sample set.