Shielded and Covered Cable Testing

Covered fails at the surface; shielded fails in the volume.

Few terminology confusions cost distribution engineering as much as the one between covered cable and insulated cable. Both are medium voltage cables with extruded polymer over the conductor, both can use XLPE, and both appear in catalogues with 15 kV, 25 kV and 35 kV classes. The similarity ends there. A covered cable for compact overhead construction carries a covering intended to reduce the consequences of accidental contact with vegetation and structures. It has no insulation shield, so the electric field is not confined: the cable surface stays electrically active with respect to its environment, and withstand capability belongs to the system of cable plus spacer plus tie plus ambient conditions. A shielded cable is a complete coaxial system — conductor shield, insulation sized for U0, insulation shield, and grounded metallic shield. The field is radial, calculable and entirely contained.

That structural difference dictates different failure modes and therefore different test philosophies. On covered cable the dominant degradation is at the surface: leakage current, dry-band arcing, erosion and tracking. The decisive test measures the material's resistance to that mechanism on an inclined plane. On shielded cable the dominant degradation is internal: voids, contaminants and protrusions at the semicon-insulation interface start partial discharge and electrical trees. The decisive test is partial discharge measurement on 100% of manufactured lengths. The teaching shortcut — covered fails by tracking, shielded fails by PD — is powerful, but it needs one caveat: water trees in shielded cable can grow for years with practically no detectable PD, and IEEE 400.2-2024 states explicitly that PD does not directly detect water trees, with dissipation factor the quantity most sensitive to diffuse water ageing.

What you will be able to do

• Distinguish a compact-network covered cable from an extruded shielded cable layer by layer, and explain why a covering is not full insulation.

• Build the complete routine, sample and type test matrix for a medium voltage shielded cable per IEC 60502-2, with correct levels, times and numerical criteria: 1.73·U0 with no discharge above 10 pC, and 3.5·U0 for 5 minutes applied voltage.

• Recognise that numerical regimes change between IEC 60502-2, IEC 60840 and IEC 62067, and avoid transplanting criteria across voltage classes.

• Describe the tests specific to covered cable — inclined plane tracking per IEC 60587, UV weathering, cable-to-spacer compatibility — and their acceptance criteria.

• Justify physically why DC testing condemns extruded insulation and why VLF at 0.1 Hz exists, from P ≈ ω·C·U².

• Interpret field diagnostics — VLF withstand, tan delta at 0.1 Hz, damped AC partial discharge — against IEEE 400.2-2024.

• Explain the fault location chain: time domain reflectometry, surge generator, and the arc reflection method.

The radial field and why interfaces decide cable quality

In a shielded cable the field in the insulation is that of a coaxial capacitor: E(x) = U0/[x·ln(R/r)], where r is the radius over the conductor shield, R the radius under the insulation shield, and x the radial position. Maximum stress occurs at the conductor shield: E_max = U0/[r·ln(R/r)]. For a 12/20 kV cable with r = 4.5 mm and 5.5 mm of insulation, E_max is about 3.34 kV/mm and E_min about 1.50 kV/mm. For a 76/132 kV design with r = 20 mm and R = 40 mm, E_max reaches about 5.48 kV/mm.

Two degradation mechanisms compete inside the insulation. A vented water tree starts at the semicon interface, where protrusions and inorganic contaminants act as documented initiators, and grows inward; a bow-tie starts at a contaminant or void inside the insulation and grows both ways. Both are diffuse regions of aqueous microchannels, they typically produce no measurable PD, and they take years. The critical engineering point is the transition: once a water tree converts into an active electrical tree with PD, the process accelerates sharply toward breakdown. In accessories the dominant mechanism is different again — workmanship. An irregular shield termination, insufficient contact pressure or interface contamination creates localised PD.

Routine testing: partial discharge as process control

Partial discharge is a routine test on every manufactured length — operationally, 100% of the reels shipped. The circuit is a low-interference AC source, a blocking filter, a coupling capacitor, a measuring impedance and an analyser per IEC 60270 and IEC 60885-3, in an area with electromagnetic noise control. Before measuring, a known charge is injected by calibrator to prove the sensitivity of the whole circuit. The IEC 60502-2 criterion: at 1.73·U0, with proven sensitivity of 10 pC or better, no repetitive discharge above 10 pC. Note the wording — no detectable PD means below the proven sensitivity of the arrangement, not physical absence of discharge.

Routine applied voltage under IEC 60502-2 is 3.5·U0 for 5 minutes in AC — never DC on extruded insulation — with no breakdown as the criterion. To test the reel ends without external surface discharge polluting the measurement, the water termination is used: the cable end is immersed in a column of deionised water that acts as a distributed electrode and grades the axial field along the termination, with conductivity held by a closed-loop deioniser. Without it, the end length that would have to be sacrificed to external flashover would make the test impractical at higher voltages.

The type test sequence: the order is the argument

The electrical type test of IEC 60502-2 is a chained sequence on one sample, and the sequence itself carries the reasoning. First the cable is mechanically abused by bending; then the insulation is interrogated by PD; a dielectric baseline is established with tan delta; the sample is thermally aged under voltage; PD is interrogated again; a lightning impulse is applied hot; and the whole thing closes with a final applied voltage. Only survival of the complete set, in that order, demonstrates the design.

Dissipation factor is measured at U0 with the conductor at the specified elevated temperature, 95 to 100 °C for 90 °C XLPE. Indicative IEC 60502-2 limits are 80 × 10⁻⁴ for XLPE and 400 × 10⁻⁴ for EPR and HEPR. Watch the contrast: the high voltage programmes of IEC 60840 and IEC 62067 use 10 × 10⁻⁴, so writing tan δ < 10⁻³ for any medium voltage XLPE cable makes the requirement unduly severe. Then come 20 thermal cycles of 8 h each under voltage, with PD repeated after the last cycle, the hot lightning impulse, and the long-duration test at 4·U0 for 4 h without breakdown, which is the verdict on the whole assembly.

Why DC condemns extruded cable, and what replaced it

Under DC, field distribution in extruded polymers is governed by conductivity — a strong function of temperature and field — rather than by permittivity. Space charge is injected and trapped in the insulation and at interfaces. When the voltage is removed, or when the cable is re-energised on AC, the abrupt redistribution adds the residual field of the trapped charge to the service field and can precipitate failure of a cable that would have survived in operation. IEC 62067 records that DC testing of extruded main insulation is ineffective and potentially harmful.

IEEE 400.2-2024 replaced the 2013 edition and brought substantive changes: application extended to 138 kV class, preferential use of RMS values for sinusoidal VLF, removal of leakage current and harmonics as primary monitored-withstand parameters, retention of tan delta and PD as diagnostic quantities with PD referred to IEEE 400.3, explicit recognition that PD does not detect water trees, and withdrawal of old universal tan delta limits in favour of population and trend criteria. Recommended durations at 0.1 Hz: 60 min for installation and acceptance of new cable, a minimum of 30 min for a simple withstand on aged cable (60 min on a critical feeder), and 15 to 30 min acceptable in a monitored withstand when the parameter is stable.

Dissipation factor at 0.1 Hz is the screening test for diffuse water ageing. The protocol measures at 0.5·U0, U0 and 1.5·U0 and extracts three figures of merit: mean tan delta, the differential tip-up between steps, and temporal stability. Interpretation is complementary rather than substitutive: high and uniform tan delta suggests diffuse water ageing, while low tan delta does not exclude a severe localised defect, because a joint with localised PD barely moves the global figure. Tan delta and PD are partners, never substitutes.

From the kilometre to the metre: locating the fault

Fault location works as a funnel: electrical pre-location to within metres, then acoustic or electromagnetic pinpointing to within centimetres. The base technique is time domain reflectometry. A pulse injected into the cable reflects at every impedance change, and distance follows from d = v·Δt/2. In XLPE the velocity factor is about 0.52 to 0.57, giving a one-way velocity of 156 to 173 m/µs and a distance factor v/2 of roughly 78 to 83 m/µs. The familiar reference of 80 to 90 m/µs is the factor already divided by two — forgetting that doubles the answer. Signatures are readable: an open circuit reflects positive, a short reflects negative, joints give small changes, and taps give multiple reflections.

Accessories deserve the last word, because a joint or termination locally removes everything the cable had going for it. It interrupts the insulation shield, creating the highest tangential stress point in the system, controlled by a stress cone or by geometric and refractive field control; it interrupts the metallic shield; and it creates new interfaces whose quality depends on workmanship. Type tests for accessories under IEC 60502-4 and the associated test methods are therefore run with the accessory installed on representative cable — real diameter over insulation, real cross-section, real preparation tool, real connector and torque — in a sequence analogous to the cable's: PD, AC voltage, impulse, thermal cycles, PD after cycling, short circuit and examination. There is no universal statistic saying a fixed percentage of failures occur in accessories; the defensible statement is that accessories concentrate field stress and human execution, and any percentage must be quoted for a specific population.

Syllabus

• Anatomy compared: nine layers of a shielded cable against the covered-cable stack

• Radial field physics: E(x) = U0/[x·ln(R/r)] and why the conductor shield interface is the birthplace of trees

• Standards map: IEC 60502-2, IEC 60840, IEC 62067, IEEE 400, 400.2, 400.3, IEC 60228, 60229, 62230, 60587

• Water trees versus electrical trees: agents, structure, growth rate, and which diagnostic sees which

• Routine tests on shielded cable: PD on 100% of lengths, applied voltage, the water termination, conductor resistance, sheath spark test

• High voltage routine regimes: 10 pC at 1.5·U0 and 2.5·U0 for 30 min under IEC 60840

• Routine and sampling tests on covered cable, and what the tracking requirement actually contracts

• The type test sequence in one sample: bending, PD, tan delta, thermal cycles, PD again, hot impulse, final applied voltage

• Prequalification: 1.7·U0 for 8 760 h with at least 180 thermal cycles

• Inclined plane testing per IEC 60587, tracking wheel, UV and weathering, cable-spacer-tie system qualification

• Field diagnostics per IEEE 400.2-2024: VLF durations, monitored withstand, tan delta MTD/DTD/TDTS

• Damped AC partial discharge and PD mapping by reflectometry

• Why DC is banned on extruded insulation, and the one place DC survives

• Fault location: TDR velocity factors, thumper energy discipline, arc reflection, acoustic pinpointing

• Accessories: where the system is really tested

Laboratory work

Attend a complete routine test on a shielded cable with a water termination: mounting the reel, connecting the termination, verifying deioniser conductivity, calibrating the PD circuit with a 5 pC injection, ramping to 1.73·U0 while reading the phase pattern of the discharges, and applying 3.5·U0 for 5 minutes. Mandatory observation points: the settling time of the water column, the laboratory background in pC, and the difference between the PD pattern of an internal void (first and third quadrant) and a surface discharge at the termination. Directed study: build the complete routine, sampling and type matrix for a 12/20 kV, 3 × 185 mm² aluminium XLPE cable per IEC 60502-2, stating level, duration, numerical criterion, sample size and clause for every test. Then compare it against the equivalent matrix for a 25 kV covered cable, identifying which tests exist in both families with different criteria, which exist in only one and why, and what the practical error would be of applying one matrix to the other.

Key takeaway

Read the standard for the voltage class you are actually testing. The three numbers that matter most — 10 pC at 1.73·U0 with 3.5·U0 for 5 min at medium voltage, 10 pC at 1.5·U0 with 2.5·U0 for 30 min at high voltage, and 4 kV/mm limited to 10 kV on the sheath only — are not interchangeable, and DC belongs to the jacket, never to the main insulation.

Module 14 of the Atlas Energy Academy · Equipment Families · about 6 hours · Video series (10 × 10–15 min) plus a cable laboratory visit · Prerequisites: Module 11 — Test Taxonomy