VA-TEK Engineering A withstand test answers one question: did the cable survive. It cannot tell you whether the insulation is degrading, or how quickly. For that you add a diagnostic, and the most direct one is partial discharge — the small, localised breakdowns that precede most cable failures.
Partial discharge can be measured at VLF, and doing so has real advantages over measuring at power frequency. It also has limits that are easy to overstate. This article covers how VLF PD measurement works, what the recorded numbers mean, and where the method stops being a diagnosis and becomes a comparison.
Key takeaways
- Partial discharge is a precursor, not a failure — detecting it early is what makes it valuable.
- At 0.1 Hz the discharge events spread out over a ten-second cycle, so they separate more cleanly than at 50 Hz.
- PD measured at VLF is not numerically interchangeable with PD at power frequency. Its value is comparative: same cable over time, or between phases.
- Record PDIV, PDEV and PD magnitude, and always against the test voltage as a multiple of U0.
- VLF PD is strongest at locating a defect in a joint or termination; it is weakest at grading distributed ageing such as water trees.
1. What partial discharge is, and why it matters
Partial discharge (PD) is a localised electrical discharge that does not bridge the insulation between the electrodes. In cable it occurs at defects: voids left in the extruded insulation, protrusions or contamination at the semiconductor screens, and damaged or poorly installed joints and terminations. Each discharge erodes a little insulation, and the erosion accelerates, until the remaining wall can no longer hold the voltage and the cable fails.
Because PD is a precursor rather than an end state, it is one of the few signals that can give warning of a cable problem before the lights go out. That is the whole reason to measure it.
2. Why measure PD at 0.1 Hz
Measuring PD at VLF instead of power frequency brings two useful differences.
First, time resolution. One cycle at 0.1 Hz lasts ten seconds. The discharge events spread across that cycle, so a time-resolved detector can separate individual pulses far more easily than at 50 Hz, where a cycle is twenty milliseconds and the pulses pile on top of one another. This is particularly helpful for distinguishing discharge sources.
Second, practicality in the field. The same physics that lets a small set charge a long cable at 0.1 Hz applies to PD: the coupling and detection equipment can be lighter, and the test runs from the same source as the withstand.
The governing document is IEEE 400.3, which covers PD testing of shielded power cable systems; VLF sets with a PD option detect discharge through a coupling capacitor or a high-frequency current transformer (HFCT) clamped around the shield-to-ground connection.
3. The three numbers to record
A PD measurement is only useful if it is recorded consistently. Three values matter:
- PDIV — the partial discharge inception voltage, the voltage at which discharge first appears as the test voltage is raised.
- PDEV — the extinction voltage, the (lower) voltage at which discharge stops as the voltage is reduced again.
- PD magnitude — the discharge level, in picocoulombs (pC) or, for an HFCT, in millivolts, together with the test voltage at which it was measured.
The relationship between PDIV and PDEV, and how magnitude grows with voltage (the “PD slope”), often says more than the absolute number. A defect that lights up at low voltage and grows steeply is a different problem from one that appears only near the top of the test.
4. The comparison trap: VLF PD is not power-frequency PD
Discharge inception depends on the rate at which the electric field changes, so PD measured at 0.1 Hz is not the same number as PD measured in service at 50 Hz. Magnitudes differ, and inception voltages shift. A reading of 200 pC at VLF is not a reading of 200 pC in service.
This is why VLF PD is used comparatively. The defensible readings are the ones taken on the same cable, with the same setup, over time — a rising trend across successive outages — or the relative level between the three phases of the same circuit. An absolute threshold quoted from a power-frequency standard should not be applied directly to a VLF reading.
5. What VLF PD can and cannot do
| It does well | It does less well |
|---|---|
| Locating an active defect in a joint or termination | Grading distributed ageing such as water trees, which are diffuse and may not discharge measurably |
| Showing a rising trend on a cable across outages | Giving a pass/fail verdict on its own — PD is a symptom, not a limit |
| Comparing phases on the same circuit | Producing values transferable to 50 Hz standards without adjustment |
| Confirming that a repair eliminated the original discharge source | Detecting defects that only discharge above the test voltage |
The honest summary: VLF PD is a strong defect locator and a useful trend tool. It is a weak pass/fail gate. Treat a positive reading as a reason to investigate and a clean reading as reassurance only when it is consistent with the trends.
6. Practical setup
Three things make the difference between a usable PD reading and noise:
- Calibrate the detector with a known charge injector before the test, so the pC scale is anchored.
- Know the noise floor. Establish the background with the set energised but at low voltage, and record it. A reading near the floor is not a measurement.
- Ground and shield correctly. The shield-to-ground path is where the HFCT sits; a poor ground adds interference and can hide discharge.
Localising a defect by time-of-flight is limited at VLF, because the propagation and reflection of pulses over kilometres of cable are harder to resolve than at power frequency. PD usually tells you that there is a defect and roughly how serious it is; pinning it to a specific metre often needs a second method.
7. Where this fits
PD is one of two diagnostics commonly added to a VLF test; the other is tan delta. When to add a diagnostic at all, and whether a monitored withstand is the better compromise, is the subject of VLF withstand vs diagnostic testing. For the dielectric-loss view of the same cable, see VLF tan delta testing. And for how PD and tan delta sit within the overall procedure, see the practical VLF guide.
