Partial Discharge Testing for Motor and Cable Insulation is a diagnostic technique that detects localized electrical breakdown inside voids, cracks, or interfaces of an insulation system, without bridging the full distance between conductor and ground. On medium voltage (MV) assets, typically motors, cables, and switchgear rated above about 3.3 kV, partial discharge (PD) is one of the earliest and most reliable indicators of insulation degradation, often visible months or years before an insulation failure would otherwise occur.
Insulation is never perfectly homogeneous. Voids from manufacturing, delamination between winding layers, contamination on cable terminations, or thermal cycling stress create weak-strength regions inside otherwise sound insulation. When the electric field across one of these regions exceeds the local breakdown strength, a tiny discharge occurs, a partial breakdown that does not span the full insulation thickness. Each event releases energy as a fast current pulse, plus heat, light, and byproducts such as ozone.
A single PD event is harmless, but repetition is not. At operating voltage a weak spot can discharge thousands of times per second, and each pulse erodes the surrounding material a little further. Over time this treeing process carves conductive channels through the insulation until a full breakdown occurs. PD testing catches this while it is still developing, not after a trip or a burnout.
PD activity is a MV and HV phenomenon. Low voltage insulation, below roughly 1 kV, rarely generates measurable discharge because field stress across typical voids is too low. PD testing is standard practice on:
For motors, PD is closely tied to the winding's insulation class, since class defines the thermal and dielectric margin before degradation accelerates.
Offline testing takes the asset out of service and applies a controlled test voltage from a separate supply while PD sensors capture pulse activity across a voltage ramp. This gives a clean measurement free from plant noise, and finds PD inception and extinction voltages directly, but the machine must be shut down and isolated.
Online testing measures PD while the asset runs under normal load, using capacitive couplers, high-frequency current transformers, or antenna-based sensors for switchgear. It captures real operating stress offline testing cannot replicate, but needs careful noise separation since plant interference sits near the same frequency range as genuine PD.
Most reliability programmes use both: periodic offline tests as a baseline, and online monitoring for trending in between.
PD magnitude is measured in picocoulombs (pC), the apparent charge of each pulse as seen at the measurement terminals, not the actual charge at the defect site. A single reading in isolation says little; what matters is pulse magnitude, repetition rate, pattern, and the trend over successive tests. Typical guidance bands, though thresholds vary by manufacturer and test method, are summarized below.
| PD Level (apparent charge) | General Interpretation | Typical Action |
|---|---|---|
| Below 100 pC | Low, consistent with sound insulation | Continue normal monitoring interval |
| 100 to 1,000 pC | Moderate activity, may indicate minor voids or contamination | Increase monitoring frequency, trend closely |
| 1,000 to 10,000 pC | Significant activity, degradation likely underway | Schedule inspection, plan corrective action |
| Above 10,000 pC | Severe activity, elevated failure risk | Prioritize outage, consider derating or replacement |
A rising trend across tests is a far stronger indicator than any single value, since some assets run with naturally higher baseline PD without elevated risk. Dated test records logged through a condition monitoring programme, alongside vibration and temperature data, give a clearer failure picture than PD data alone.
IEC 60270 is the foundational standard for PD measurement technique, defining calibration, apparent charge measurement, and test circuit requirements for high voltage equipment generally.
IEC 60034-27 addresses PD testing for rotating electrical machines specifically: offline testing under Part 27-1 and online monitoring under Part 27-2, covering motors and generators. It accounts for winding geometry, multiple discharge sources within a single stator, and the noise environment of running machines.
Recurring causes include voids from manufacturing defects or thermal cycling, surface contamination and tracking on terminations, damaged stress control layers on cable joints, slot discharge from loose coils vibrating against the stator core, and insulation aging from sustained overheating.
Elevated PD alongside abnormal vibration often points to loose windings rather than dielectric aging, while PD rising with load current suggests thermal involvement.
PD testing delivers the most value as a recurring, trended measurement tied into the maintenance workflow, not a one-off snapshot. Recording PD results and follow-up actions in a CMMS such as Fabrico alongside vibration and temperature data lets reliability teams see the full picture for each critical asset and schedule interventions before an unplanned trip.
For teams evaluating how condition monitoring data can drive maintenance planning directly, book a Fabrico demo to see how PD trends and other diagnostic inputs fit into a single asset record.
Rarely with meaningful results. PD activity requires enough field stress to ionize the local gas, and low voltage insulation, below roughly 1 kV, typically does not reach that stress level in normal operation. PD testing is standard practice on MV assets from about 2.3 kV upward.
Not directly. PD testing shows degradation is occurring and its severity relative to typical bands, but remaining life depends on load history, thermal exposure, and defect type. Trending PD levels over time gives a far better indication of rate of change than any single test.
Not necessarily. Sensor placement, cable routing, and noise floor all affect detection sensitivity, and some defect types are harder to detect. A zero reading should be corroborated with other diagnostics such as polarization index testing.
Zaloguj
Blog