Among the various types of transformer failures, partial discharge (PD) stands out as the most insidious and destructive culprit. It does not trigger an immediate trip like a short-circuit fault, but quietly erodes the insulation system over months or even years, eventually leading to a sudden catastrophic breakdown.
For operation and maintenance personnel, “invisible discharges” are far more dangerous than “visible faults.” This article delves into the physical mechanisms of partial discharge, its typical types, damage mechanisms, and provides a systematic overview of current online monitoring technologies along with key engineering considerations.
Partial discharge is a localized electrical discharge that occurs in a defect within an insulation system when the electric field strength exceeds the local breakdown strength, but without causing an immediate complete breakdown of the entire insulation.
Inside a transformer, the insulation system consists of oil-impregnated paper (oil-paper composite). Due to manufacturing imperfections, long-term aging, or external stresses, defects such as air gaps, bubbles, metallic particles, or moisture may exist. The permittivity of these defects differs from that of the main insulation, causing local electric field distortion. When the local field strength exceeds the breakdown strength of the medium at the defect, partial discharge occurs.
The danger of partial discharge lies not in the energy of a single event (ranging from microjoules to millijoules), but in its cumulative effects:
Thermal effect: The temperature at the discharge point can reach hundreds or even thousands of degrees Celsius, causing thermal decomposition and carbonization of the insulating material.
Chemical corrosion: Discharge decomposes the insulating oil, generating characteristic gases such as hydrogen, acetylene, and carbon monoxide, as well as conductive carbon particles and acidic byproducts, further degrading the insulation properties.
Physical erosion: Charged particles generated by the discharge continuously bombard the insulation surface, causing material ablation and progressively enlarging the defect, creating a vicious cycle.
This process follows a typical evolution chain: “defect → partial discharge → insulation deterioration → defect growth → intensified discharge → breakdown.” The time from initial discharge to final failure can span years, but once it enters the acceleration phase, failure often occurs within hours to days.
Based on the location and mechanism, partial discharges in transformers can be classified into three main categories. They differ significantly in waveform characteristics, root causes, and severity.
Location: Air voids inside solid insulation, bubbles between oil-paper layers, or delaminated insulation.
Mechanism: The permittivity of a gas-filled void (≈1) is much lower than that of oil-paper insulation (≈2.2–4). Under an AC field, the electric field stress across the void is much higher than across the main insulation. When this stress exceeds the breakdown strength of the gas, discharge occurs.
Identification features:
Discharge pulses typically appear near the peaks of the power frequency waveform (symmetric or asymmetric between positive and negative half-cycles).
Discharge magnitude is relatively stable and increases stepwise with rising voltage.
Dissolved gas analysis (DGA) shows a marked increase in acetylene (C₂H₂) and hydrogen (H₂) , and gas ratios point to “discharge-type faults.”
Severity: ★★★★★ (most destructive, directly threatening the main insulation)
Location: Along the surface of insulation, commonly found at winding ends, lead insulation, the root of bushings, etc.
Mechanism: Contamination, moisture, or burrs on the insulation surface cause a high tangential electric field component, leading to flashover-like discharges along the surface.
Identification features:
Discharge pulses have relatively large amplitudes and a wide phase distribution.
Often accompanied by visible corona glow or faint audible noise.
DGA shows higher proportions of ethylene (C₂H₄) and methane (CH₄) relative to other discharge types.
Severity: ★★★★☆ (prone to cause phase-to-phase or phase-to-ground flashover)
Location: Metal tips, burrs, or floating potential components (e.g., loose bolts, shielding covers).
Mechanism: Under a strong electric field, the field stress at a metal tip is extremely concentrated, ionizing the surrounding gas and causing discharge.
Identification features:
Discharge pulses are concentrated near the peak of the negative half-cycle of the power frequency waveform (asymmetry characteristic of a needle–plane gap).
Waveform shows typical “pulse burst” characteristics.
Acoustic emission (AE) signals appear as continuous high-frequency noise, distinct from the discrete signals of internal discharges.
Severity: ★★★☆☆ (less harmful initially, but prolonged existence can contaminate the oil and trigger more severe faults)
Traditional preventive tests (such as off-line induced voltage tests and PD tests) cannot reflect the actual insulation condition under operating conditions. In recent years, the maturity of online monitoring technologies has enabled real-time assessment of transformers operating in service.
Principle: Partial discharge generates steep current pulses that excite electromagnetic waves up to several gigahertz. UHF antennas are installed at insulation flanges, valve openings, or dedicated sensor ports on the transformer tank to capture these electromagnetic signals.
Advantages:
Strong anti-interference capability: Most on-site interference (e.g., mobile signals, power line carrier) lies below 300 MHz, so the UHF band has very low background noise.
Locatable: Multi‑sensor time‑difference of arrival (TDOA) techniques enable spatial localization of the discharge source.
Limitations:
Signal attenuation in oil‑paper insulation can be significant, limiting sensitivity for deep internal discharges.
Requires pre‑installed sensor ports on the transformer, which can be costly for retrofit.
Typical applications: Long‑term online monitoring of large power transformers and GIS.
Principle: The energy released by a partial discharge generates mechanical stress waves (ultrasonic waves) in the insulating medium. Piezoelectric sensors attached to the outer tank wall detect these ultrasonic signals.
Advantages:
Non‑intrusive: No internal modifications to the transformer are required.
Precise localization: By measuring the time differences of arrival of the acoustic wave at multiple sensors, the discharge source can be located in 3D space, typically with centimeter‑level accuracy.
Limitations:
Ultrasonic waves attenuate rapidly in transformer oil, limiting the detection range (typically 3–5 meters from the sensor).
Susceptible to mechanical vibrations and electromagnetic noise; requires filtering and advanced algorithms.
Typical applications: Precise fault location, investigation of sudden discharge events.
Principle: The transient current pulses generated by partial discharge propagate along ground leads, the neutral point, or bushing tap circuits. A clamp‑on HFCT sensor couples these high‑frequency signals.
Advantages:
Easy installation: Can be installed live without de‑energizing the transformer.
High sensitivity: Can detect apparent charges down to a few picocoulombs (pC).
Limitations:
Susceptible to external electromagnetic interference; requires signal discrimination algorithms.
Cannot provide spatial location; indicates only the presence and intensity trend of PD activity.
Typical applications: Routine inspections, portable surveys, trend‑based early warning.
| Technology | Sensitivity | Anti‑Interference | Localization Capability | Ease of Installation | Typical Application |
|---|---|---|---|---|---|
| UHF | High | Very strong | Possible (multi‑sensor) | Requires pre‑installed port | Long‑term online monitoring |
| AE | Medium | Medium | Precise 3D localization | Simple (external) | Fault pinpointing |
| HFCT | Very high | Medium | Not possible | Very simple (clamp‑on) | Routine inspection, early warning |
In engineering practice, a single technology rarely covers all fault types. A multi‑sensor fusion approach combining UHF, HFCT, and AE is recommended to improve diagnostic accuracy through data integration.
PD detection and assessment are primarily guided by the following standards:
IEC 60270: The fundamental standard for PD measurement, defining the calibration method for apparent charge (pC).
GB/T 7354: The Chinese national standard equivalent to IEC 60270.
IEEE C57.113 (or regional equivalents) – Guide for Partial Discharge Measurement in Power Transformers.
In practice, a comprehensive diagnosis should consider:
Trend of discharge magnitude: Sustained growth (e.g., >20% per month) warrants attention.
Phase‑resolved patterns: The distribution of discharges across the power frequency cycle helps identify the discharge type.
Multi‑source information fusion: Simultaneous triggering of UHF and AE signals helps exclude external interference.
Partial discharge is a “precursor signal” of insulation degradation, not a fault in itself. Through scientific online monitoring and diagnosis, it is entirely possible to detect defects at an early stage, schedule planned maintenance, and avoid the enormous losses caused by sudden failures.
For operation and maintenance personnel, shifting from periodic off‑line testing to continuous condition‑based monitoring, and from reactive repair to proactive early warning, is essential for improving equipment reliability. With the rapid advancement of the Internet of Things (IoT), edge computing, and artificial intelligence–based diagnostic algorithms, partial discharge monitoring is evolving toward holistic sensing, intelligent diagnosis, and predictive trending, providing solid data support for the entire lifecycle management of transformers.
