Partial discharge is both a symptom and a cause. A void in insulation produces PD; that PD erodes the surrounding insulation further, enlarging the void and increasing PD magnitude. Catching partial discharge early interrupts this self-amplifying cycle before the insulation system fails completely.
What Is Partial Discharge?
Partial discharge is a localized dielectric breakdown within an insulation system that does not completely bridge the gap between conductors. When the electric field strength in a void, delamination, or surface contamination exceeds the breakdown strength of the material in that defect, a small discharge occurs. The discharge is "partial" because it bridges only the local defect, not the full conductor-to-conductor or conductor-to-ground gap.
PD activity is measured in picocoulombs (pC), which quantifies the charge transferred during each discharge event. A single defect can produce thousands of individual discharge pulses per second, each contributing to cumulative insulation erosion. The presence of PD does not indicate immediate failure, but sustained PD activity accelerates deterioration at a rate proportional to discharge magnitude and repetition rate.
High-voltage equipment operates under electrical stress levels where small defects, which would be benign at low voltage, create localized fields sufficient to initiate discharge. This is why PD testing is specific to medium-voltage and high-voltage equipment.
Why Partial Discharge Damages Insulation
Each discharge releases energy that bombards the surrounding insulation material with electrons, ions, and ozone. This erodes and degrades organic insulation materials such as epoxy, polyester, and varnish. Over time, the defect grows, the discharge magnitude increases, and the erosion accelerates.
The chemical attack is significant. Ozone produced by PD reacts with polymer chains in organic insulation, breaking them down and creating surface cracks. Nitric acid can form in the presence of moisture, further attacking organic materials and any metallic components nearby. These byproducts of repeated discharge events accumulate in the defect region, creating conditions that sustain and intensify future discharges.
In high-voltage machines and cables, unchecked PD progresses from the initial defect to tracking paths across the insulation surface or through the bulk material. The endpoint is a disruptive discharge (flashover) or insulation puncture. The timeline between measurable PD activity and failure depends on the voltage level, discharge magnitude, duty cycle, and insulation material. In some cases, deterioration spans years; in others, weeks.
Types of Partial Discharge
Different defect types produce distinct PD signatures. Identifying the discharge type guides the assessment of severity and the likely location of the defect within the equipment.
| PD Type | Location | Cause | Characteristic Pattern |
|---|---|---|---|
| Internal void PD | Within the insulation bulk | Voids from manufacturing, thermal cycling, moisture ingress | Activity symmetric around voltage peaks |
| Slot discharge | Between stator bar and grounded core | Loose coil wedges, degraded slot coating | Activity on positive half-cycles in large machines |
| Surface PD / tracking | Along insulation surface | Contamination, moisture, surface degradation | Irregular, high-magnitude bursts |
| Corona | In air adjacent to sharp conductors | Sharp conductor edges, inadequate clearance | Activity on negative half-cycles for positive conductors |
| Delamination PD | Between insulation layers | Thermal cycling, poor adhesion | Similar to internal void but at a layer interface |
Phase-resolved PD (PRPD) patterns are the primary tool for distinguishing these types. Each defect category produces a recognizable distribution of discharge events relative to the AC voltage waveform. Software analysis tools overlay thousands of discharge pulses onto a phase map, and the resulting cluster shapes indicate the likely defect mechanism.
When PD Testing Applies
PD testing is primarily relevant for medium-voltage and high-voltage equipment where the insulation stress is high enough for discharges to form. The practical threshold is approximately 3 to 4 kV for rotating machines with form-wound stator coils, and 5 kV and above for cable systems and switchgear.
Equipment categories where PD testing is routinely applied:
- Motors and generators above 3-4 kV: form-wound stator coils develop slot discharge and internal void PD as the groundwall insulation ages or sustains mechanical damage
- Medium-voltage switchgear (5 kV and above): bus insulation, vacuum interrupter assemblies, and cable terminations within switchgear are PD sources
- Power cables (medium and high voltage): joints and terminations are the highest-risk locations; cable insulation itself develops PD at defects introduced during installation or manufacturing
- Transformers and potential transformers: oil-paper insulation systems and dry-type transformers both produce measurable PD as insulation degrades
- Gas-insulated switchgear (GIS): PD in SF6 gas-insulated equipment indicates particle contamination, protrusions, or insulator defects
- High-voltage cable terminations and splices: the transition between cable insulation and external hardware is mechanically and electrically stressed and prone to PD at installation defects
Low-voltage equipment (below 1 kV) typically does not develop partial discharge under normal service conditions. The field stress at these voltages is insufficient to initiate localized breakdown in standard insulation systems.
PD Measurement Methods
Offline PD Testing (IEC 60270)
The machine or cable is de-energized, isolated from the power system, and connected to an external test voltage source. PD magnitude is measured using a coupling capacitor and calibrated detector circuit. IEC 60270 defines the standard methodology, including calibration procedures and measurement bandwidth requirements. Results expressed in picocoulombs are directly comparable between laboratories and across test campaigns on the same asset.
The primary limitation is that offline testing requires a planned outage and applies a test voltage from an external source. The test voltage may differ from normal operating stress in magnitude and waveform, which can underrepresent or misrepresent the PD activity the insulation experiences in service. Access to the equipment for connection of the test source is also a practical constraint, particularly for large motors or cables with complex termination arrangements.
Online PD Monitoring
Capacitive or inductive sensors are permanently or temporarily coupled to the equipment while it operates at normal service voltage. This approach measures PD continuously without requiring an outage, and captures PD activity under actual operating stress, load, and temperature conditions.
Signal interpretation is more complex than offline testing because the plant electrical environment introduces noise at frequencies overlapping with PD signals. High-frequency current transformers (HFCT), capacitive couplers, or directional couplers installed at cable terminations or motor terminals capture PD pulses during operation. Noise rejection algorithms and directional techniques separate genuine PD from background interference. Continuous online monitoring enables trending, which is more diagnostically useful than a single-point measurement.
Acoustic PD Detection
PD events generate ultrasound in the 20 to 150 kHz range, propagating through solid structures and through gas. Contact acoustic sensors applied to the external surface of transformer tanks, cable termination housings, or switchgear panels can locate the acoustic source and confirm PD activity. Airborne ultrasound detectors work for open-air installations where corona or surface discharge is audible at ultrasonic frequencies.
Acoustic detection is particularly useful for locating the source of PD within a switchgear enclosure when an electrical measurement has already confirmed PD is present. It does not provide the calibrated pC magnitude of electrical methods but adds spatial information that electrical measurements alone cannot provide.
UHF Monitoring
PD in gas-insulated switchgear and oil-immersed transformers emits electromagnetic radiation in the 300 MHz to 3 GHz range. UHF sensors installed in sensor ports on GIS enclosures or transformer tanks capture these signals. UHF methods provide discrimination between PD types based on pulse shape and frequency content, and they reject low-frequency industrial noise that affects conventional electrical measurements. UHF is the preferred method for GIS because it can detect very low-level discharges before they reach magnitudes detectable by conventional IEC 60270 methods.
Interpreting PD Results
PD results are interpreted through several complementary metrics: PD magnitude in picocoulombs, PD repetition rate (pulses per second), the phase position of discharges relative to the AC voltage cycle expressed as a PRPD pattern, and the trend of these quantities over successive measurements.
A single PD reading taken in isolation is difficult to assess without reference data. The same pC magnitude may be acceptable for one machine design and indicate severe degradation in another, depending on voltage rating, insulation class, and the manufacturer's baseline data for that equipment type. Trending over multiple measurements from the same machine at comparable load and temperature conditions provides far more diagnostic value than any single test.
For stator windings, the normalized quantity NQN (or NQ in some standards) standardizes PD results across machines of different ratings and test conditions, enabling comparison across a fleet. IEEE 1434 and IEC 60034-27 provide guidance on measurement procedures and result interpretation for rotating machines. Expert interpretation is required for definitive assessments; published thresholds are guidelines rather than absolute pass-fail criteria.
Results that warrant action typically include: a step increase in PD magnitude between successive tests on the same machine, PRPD patterns consistent with slot discharge in an aging motor, or PD levels exceeding the 90th percentile of a population of similar machines at comparable service age.
PD vs Insulation Resistance Testing
PD testing and insulation resistance (IR) testing serve different diagnostic purposes and detect different insulation conditions. Understanding the distinction prevents over-reliance on either method alone.
IR testing applies a DC voltage (typically 500 V to 5 kV depending on equipment rating) and measures the bulk leakage current through the insulation, expressed as resistance in megaohms. The polarization index (PI) is the ratio of the 10-minute to 1-minute resistance reading, indicating whether the insulation contains mobile charge carriers from moisture or contamination. IR testing is sensitive to surface contamination and bulk moisture content but does not detect internal voids or delamination that have not yet created a conductive path between conductors.
PD testing detects defects in the insulation bulk and at interfaces before they produce a measurable change in bulk resistance. An insulation system can have excellent IR and PI values while containing significant void-based PD activity, because the voids are electrically isolated from each other and from the conductor surfaces. Conversely, surface contamination that suppresses IR to low values may not produce significant PD if the contamination is uniformly distributed rather than concentrated at high-field locations.
The two tests are complementary. A complete insulation health assessment for high-voltage equipment uses both: IR and PI for bulk condition and contamination screening, PD testing for internal defect detection. Neither test replaces the other.
Continuous Electrical Monitoring for Rotating Assets
Tractian's electrical monitoring continuously tracks motor current signatures that indicate developing insulation degradation, complementing periodic offline PD testing programs.
See Electrical Monitoring