Evaluation of Electrical Tree Length Information Based on Partial Discharge Signal

Electrical treeing is a common phenomenon associated with insulation degradation. When these trees extend to the ground side, cable breakdown occurs. To ensure safe transmission in power systems, it is crucial to quantify the growth length of electrical trees using specific indicators. However, existing PRPD spectra exhibit limited capability in characterizing the length of electrical branches, as phase parameters become indistinguishable under varying branch lengths. Currently, the most widely used methods for characterizing partial discharge phenomena are the T-F map and PRPD diagrams. According to experimental results, the <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$T-F$</tex> map did not show significant differences as the electrical tree grew. In the experiment, the PRPD diagrams only exhibited amplitude variations, which were heavily influenced by voltage levels. Therefore, these methods were deemed unsuitable for accurately characterizing electrical tree growth. Given that the growth of electrical trees exhibits chaotic characteristics, this study integrates both nonlinear and linear quantities in partial discharge analysis to identify differences in partial discharge behavior at various growth stages. In this study, after normalizing the original partial discharge signal, amplitude information and sign information were extracted. The amplitude information was treated as a nonlinear quantity, while the sign information was considered a linear quantity. Both were computed using the DFA algorithm and subsequently plotted on a graph, with the amplitude information represented on the <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\boldsymbol{x}$</tex>-axis and the sign information on the <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$y$</tex>-axis. The results indicate that as the electrical tree progresses, the data tend to migrate towards the upper-left quadrant. The experimental results demonstrate that this method can, to some extent, determine the growth of electrical trees by measuring partial discharge signals during their growth process.