Water Tree Degradation Impacting on the Insulation Performance of Highly-aged MV/HV-class XLPE Cable

This study applied a pre-breakdown discharge detection test to HV/MV XLPE cables decommissioned after approx. 30 years of service, to identify water trees that govern insulation performance. An AC high voltage was increased stepwise and stopped immediately upon partial discharges (PD) detection to prevent breakdown. The PD source was located, insulation sliced, and defects were observed. The sample set comprised 129 HV XLPE cables rated 66/77 kV (7

- 42 years, dry-curing) and 38 MV XLPE cables rated 22/33 kV (5 - 40 years, 33 dry-curing, 5 wet-curing), both without water-barrier structure. Water trees were identified as the determinants of the insulation performance in 81 HV and 24 MV cables. Some PD-origin segments were excised and re-tested to expand the dataset of performance-relevant water trees.

Moreover, 63 HV XLPE cables rated 66/77 kV were similarly investigated and BTTs were identified as the determinants of insulation performance as well.

Then, insulation performance was defined as the electric field at PD inception, calculated by dividing the applied voltage at first PD by the insulation thickness where the PD occurs. Results show a clear inverse correlation: longer water tree, lower insulation performance and narrower its variability for both BTT and VT. Moreover, the relationship between BTT length and insulation performance was similar for MV cables with 6–8 mm nominal insulation thickness and HV cables with 9–13 mm, indicating nominal voltage has limited impact compared to water tree length. Consequently, water tree length is the key determinant of degradation and should be the primary target in diagnostic assessments of XLPE cable insulation.