Water treeing in XLPE medium-voltage cables — the creeping ageing mechanism behind cable failures after years of service
Water treeing in XLPE medium-voltage cables — the creeping ageing mechanism behind cable failures after years of service
The guide on partial discharge and tan-delta measurement and the guide on VLF cable testing cover two test methods used to assess aged medium-voltage cables. This article covers the underlying degradation mechanism these test methods are precisely trying to detect: water treeing in the XLPE (cross-linked polyethylene) insulation of medium-voltage cables, standardised within the scope of IEC 60502-2.
What water treeing is
Water treeing is a slow, time-progressive degradation process that occurs when moisture and an electric field act on the XLPE insulation simultaneously. Under that combination, microscopically small, branched structures form within the insulation that resemble the branches of a tree in shape — hence the name. These "trees" are usually not a direct breakdown path in themselves, but they gradually weaken the local dielectric strength of the insulation over a period that can extend from years to decades, until the remaining breakdown strength drops below a critical threshold.
Two types: bow-tie trees and vented trees
The literature distinguishes two main types of water trees, which form at different locations within the insulation and carry a different risk profile:
- Bow-tie trees: form within the insulation itself, typically around a microscopic contaminant or void where moisture has accumulated. They grow outward from that central point in two directions at once (hence the bow-tie-like shape) and generally remain limited in extent.
- Vented trees: form at the interface between the insulation and the conductor- or insulation-screen layer, often at microcracks that fill with moisture. They grow inward from that interface into the insulation and are generally considered more risky than bow-tie trees in the literature, since they can penetrate deeper into the insulation from a point of higher local field strength.
Note: the precise relationship between water tree type, remaining breakdown strength and practical failure risk is a subject of ongoing materials research and depends on the specific insulation compound, cable construction and operating conditions. This guide article describes the general mechanism, not a quantitative lifetime prediction for a specific cable.
Why this degradation mechanism stays invisible for a long time
Water treeing affects the insulation without any externally visible damage: the cable sheath and the outer surface of the insulation can look entirely normal while an extensive network of water trees has already formed internally. This makes the mechanism particularly difficult to detect with a purely visual inspection alone, and is precisely why diagnostic techniques that assess the electrical properties of the insulation itself — such as partial discharge measurement, tan-delta measurement and VLF testing — carry practical value when estimating the remaining service life of older medium-voltage cables.
Mitigation: water-blocking construction and TR-XLPE
Two practical measures are used to limit water treeing:
- Water-blocking construction (water-blocking tape or longitudinal water blocking within the cable construction) limits the ingress and spread of moisture along the cable, reducing the availability of the moisture needed for water tree formation.
- Tree-retardant XLPE (TR-XLPE), an insulation compound specifically formulated to slow the formation and growth rate of water trees compared to standard XLPE, is applied in newer cable constructions as an additional measure — particularly in "wet" underground applications where prolonged moisture exposure cannot be fully excluded.
Neither measure eliminates the mechanism entirely; they slow it down and thereby extend the practical service life of the cable under moist operating conditions.
Practical relevance
When assessing the failure risk of an older medium-voltage cable (see also the guide on PD and tan-delta measurement and the guide on VLF cable testing), it is important to recognise that water treeing is an internal, slowly progressing mechanism that is not detected by a purely visual inspection of the sheath and terminations — periodic electrical diagnostics of the insulation itself is precisely what is needed to gain an indication of remaining condition, particularly for cables that have already sat in wet ground for a long time (years to decades).
Common mistakes
- Assuming a cable without visible external damage is also in good internal condition — water treeing affects the insulation precisely without any externally visible sign.
- Treating bow-tie trees and vented trees as equally risky, while vented trees are generally considered the riskier type in the literature due to their growth direction from an interface with higher local field strength.
- Viewing water-blocking construction or TR-XLPE as a complete guarantee against water treeing, while both measures slow the process but do not eliminate it entirely.
- Skipping periodic electrical diagnostics on older cables with an uneventful operating history, while precisely the gradual, creeping nature of water treeing means a cable can function for years without failure before remaining breakdown strength reaches a critical point.
Related
Further reading
- IEC 60502-4 (kabeleindsluitingen MS)Stress cone at a medium-voltage cable termination — why the cut-back screen itself creates an electric field problem
- IEC 60840 / Praktijk (kabeltrajecten)Cable sheath cross-bonding — why long single-core high-voltage cables do not simply earth the sheath at both ends
- §522.8Underground cables — burial depth and mechanical protection (§522.8)
- IEEE 400.2 (VLF-beproeving)VLF cable testing — why a medium-voltage cable is tested at 0.1 Hz instead of power frequency after installation or repair
- IEC 60364-5-52Aluminium cables — cross-section equivalence relative to copper and the 1.6× rule of thumb
- NEN-EN 1366-3Fire-resistant cable penetrations — NEN-EN 1366-3