Partial Discharge Testing: UHF, HFCT, and Acoustic Methods for Substation Equipment

Partial discharge (PD) is a localized electrical discharge that occurs within a void, crack, or contamination site in insulation without completely bridging the gap between conductors. The discharge ionizes the gas in the void, releasing energy that slowly erodes the surrounding insulation material through chemical attack, thermal damage, and physical bombardment by ions. A single PD event is measured in picocoulombs (pC) of charge displacement, a very small quantity. But at power frequency, PD events may occur thousands of times per second at the same site, and the cumulative erosion over months and years produces carbonized tracks, voids that grow larger, and ultimately a failure path through the insulation.

The diagnostic value of PD testing is that it detects this mechanism while it is still in its early stages, well before power factor or insulation resistance measurements show any change. Bulk dielectric tests measure the average condition of a large volume of insulation; PD testing detects the localized defect within that volume. A transformer bushing that tests clean on power factor may already have a PD source developing at a void in the condenser section that will produce a power factor increase only after substantial erosion has occurred. PD testing finds that void first.

What a partial discharge measurement captures

A PD event produces three detectable signatures: an electrical current pulse in the circuit connected to the insulation system, an electromagnetic wave radiated from the discharge site, and an acoustic pressure wave (sound) emitted by the rapid gas expansion in the void. Each of these forms the basis for a different PD measurement method, and each has different sensitivity, frequency range, and suitability for different equipment types.

The conventional PD measurement method defined by IEC 60270 detects the current pulse by connecting a coupling capacitor and a measuring impedance in series with the test object, amplifying the resulting voltage pulse, and recording its magnitude in picocoulombs. IEC 60270 measurements are calibrated in absolute pC and provide repeatable, comparable results between test facilities and test instruments. This method is most commonly used in factory acceptance testing of new equipment, where test leads can be connected directly to the equipment terminals in a controlled environment. Field application of IEC 60270 is possible but requires careful management of background electrical noise, since any PD activity in the surrounding environment couples into the measuring circuit alongside the signal of interest.

UHF and TEV detection for GIS and metal-clad switchgear

For gas-insulated switchgear (GIS) and metal-clad medium-voltage switchgear, non-conventional PD detection methods are the standard field approach. Ultra-high frequency (UHF) sensors detect the electromagnetic pulse radiated by a PD event, which travels through the SF6 gas in a GIS bus or through the air inside metal-clad switchgear at frequencies between 300 MHz and 3 GHz. The UHF signal attenuates rapidly with distance, which is both a limitation (sensitivity drops over long bus sections) and an advantage (it helps localize the PD source to a specific bay or compartment by comparing signal levels at sensors in adjacent locations).

UHF sensors are mounted on inspection windows, drain plugs, or purpose-built sensor ports in the GIS enclosure. For retrofitting UHF monitoring to existing GIS without factory-installed sensor ports, external capacitive sensors can sometimes be clamped to dielectric inspection windows. The measurement result is reported as a signal level in dBm or millivolts rather than in calibrated picocoulombs, because the coupling between the UHF sensor and the PD source is not precisely known. Trending over time and comparison to a baseline from commissioning is the primary interpretation method.

Transient earth voltage (TEV) detection is the dominant non-invasive method for metal-clad medium-voltage switchgear. A PD event inside a metal-clad switchgear panel causes a transient current to flow on the outer surface of the enclosure, measurable with a TEV probe (a capacitive sensor pressed against the panel exterior). TEV detection requires no access to the interior of the switchgear and can be performed while the equipment is fully energized and in service, making it the preferred method for in-service condition monitoring surveys. TEV readings are compared between phases, between bays, and against baseline readings to identify switchgear that warrants further investigation or de-energized inspection. A switchgear bay with TEV levels significantly higher than its neighbors or significantly higher than its own prior baseline is a candidate for a more detailed examination.

HFCT clamps for cable PD detection

High-frequency current transformer (HFCT) clamps measure PD activity in medium-voltage cables by clamping around the cable shield earth bond at a termination or cable joint. When a PD event occurs within the cable insulation, the resulting current pulse travels along the cable and returns via the shield, passing through the HFCT clamp. The clamp converts this current pulse to a voltage signal that is recorded and analyzed. Because the HFCT measurement is made at the cable termination rather than requiring any connection to the cable conductor, it can be performed on energized cables at any point where the shield earth bond is accessible, typically at switchgear cable terminations or at above-grade cable joints.

HFCT testing is performed alongside VLF tan delta testing as a complementary diagnostic. VLF identifies bulk insulation degradation across the full cable length; HFCT identifies specific PD sources within the cable that may or may not be caught by the bulk measurement. A cable with a localized void in the insulation at a joint may show a nearly normal tan delta result while the HFCT clamp detects clear PD activity at that joint. Together, the two methods give a more complete picture than either alone.

Acoustic PD detection on transformers

For power transformers, acoustic PD detection uses piezoelectric sensors mounted on the outer surface of the transformer tank to detect the pressure wave generated by PD events inside the tank. The acoustic signal travels through the transformer oil from the discharge source to the tank wall, where the sensor converts it to an electrical signal. Multiple sensors mounted at different positions on the tank can triangulate the approximate location of the PD source, the sensor closest to the source receives the signal first, and the differences in arrival time between sensors allow a three-dimensional position estimate for the discharge site.

Acoustic PD detection on transformers is challenging because the tank environment is acoustically complex, oil flow from the cooling system, fan noise, external vibration from the substation structure, and electromagnetic interference all produce signals that can obscure the PD signature. Experienced practitioners distinguish PD acoustic signatures from interference by their phase relationship to the power frequency voltage, genuine PD activity has a predictable occurrence within the AC cycle (phase-resolved PD) while most mechanical noise is not synchronized to the power frequency. Measurement during a low-load, low-noise period improves signal-to-noise ratio.

Phase-resolved PD patterns (PRPD)

The most powerful PD interpretation tool is the phase-resolved partial discharge (PRPD) pattern, which plots each PD pulse as a point on a chart with phase angle on the horizontal axis (0° to 360° of the AC cycle) and pulse magnitude on the vertical axis. Different PD source types produce characteristic PRPD patterns that allow experienced analysts to identify the nature and location of the discharge activity.

Internal voids in solid insulation produce PD pulses clustered symmetrically around the positive and negative voltage peaks (near 90° and 270° in the cycle), because the electric field stress in the void is maximum at voltage peak. Corona discharge from a sharp metallic point in gas produces a highly asymmetric pattern, heavy activity on one half of the cycle only, because the geometry of the discharge changes with field polarity. Surface discharge along a contaminated insulator surface produces a pattern spread across a wide phase range with irregular magnitude variation. A floating metallic particle in SF6 gas produces a characteristic pattern associated with particle bounce, the particle charges and discharges as it contacts the enclosure, producing random-magnitude pulses across the cycle.

Recognizing these patterns requires instrument software that can display the PRPD plot in real time and accumulated over many cycles, and interpretation experience built from comparing patterns to known defect types. Field PD surveys typically capture the PRPD pattern alongside the raw signal level, providing both a quantitative result and a qualitative signature for the engineering review.

PD testing versus conventional insulation tests

PD testing and conventional insulation tests, power factor, insulation resistance and PI, and dielectric withstand, are complementary, not competing, assessments. A complete acceptance test program for high-voltage equipment runs both. Power factor detects distributed insulation degradation that has affected a significant volume of the insulation system. PD testing detects localized defects that a bulk measurement may not yet register. The equipment that most benefits from PD testing at acceptance is equipment where a localized defect at a specific high-stress point, a bushing condenser void, a cable joint inclusion, a GIS particle, is the most likely failure mode. PD testing has become standard on new GIS equipment per IEC 62271-203 and is increasingly required for high-voltage cable acceptance per IEC 60840 and IEC 62067.