Partial discharge testing can identify insulation activity before it develops into a more serious electrical problem. But detecting a signal is only part of the challenge. In substations, plants, factories, and other electrically active environments, interference can overlap with genuine partial discharge signals and make results harder to interpret.
Accurate partial discharge testing therefore depends on more than the measurement itself. Sensor selection and placement, background-noise assessment, phase synchronization, filtering, multi-channel measurement, and trend analysis all influence how confident PD activity can be distinguished from interference.
So, how do you know whether the signal on the screen comes from the asset or from the electrical environment around it?
How is partial discharge testing done?
The exact test setup depends on the asset and measurement method, but a typical partial discharge measurement follows four basic steps:
- Prepare the asset and test setup: Define the measurement objective, operating or test condition, and suitable measurement points.
- Connect appropriate sensors: Select and position partial discharge sensors based on the asset and measurement method.
- Measure: Acquire partial discharge signals while also observing the background-noise environment and relevant phase reference.
- Analyze: Evaluate signal patterns, magnitude, phase relationship, repetition, trends, and other available information to distinguish meaningful PD activity from interference.
IEC 60270 provides the principal framework for charge-based partial discharge measurements, including measurement circuits, calibration, test procedures, and guidance on distinguishing partial discharge from external interference. Higher-frequency and non-conventional measurements may require other applicable procedures or standards.
Why is noise a problem in partial discharge testing?
Partial discharge signals can be small compared with electrical interference in the surrounding environment. Noise can mask genuine partial discharge activity or resemble discharge signals, making it harder to determine whether the measured activity originates from the asset.
What causes noise in partial discharge testing?
Common sources of interference include:
- External corona or discharge activity: Signals from outside the asset can enter the measurement.
- Drives and power electronics: Switching operations can generate electrical interference.
- Radio-frequency interference: Communication systems and other nearby sources can introduce unwanted signals.
- Electrical switching: Switching activity elsewhere in the system can appear in the measurement.
- Grounding and measurement connections: The test setup itself can couple unwanted interference into the measurement.
The type and level of interference vary by site. Identifying the background-noise environment is therefore an important step before interpreting partial discharge measurements.
How can you reduce noise in partial discharge testing?
Reducing noise starts with the test setup and continues through signal acquisition and analysis. Key practices include:
- Check background noise: Establish the existing interference environment before interpreting measured signals.
- Select the right sensor: Match the sensor and frequency range to the asset, measurement method, and test objective.
- Position and ground correctly: Proper sensor placement, grounding, connections, and cable routing help limit unwanted interference.
- Use phase-synchronized measurement: A phase reference supports PRPD analysis and helps identify recurring signal patterns.
- Apply gating and filtering: Band-pass filters and noise gating can suppress or exclude unwanted signals during analysis.
- Compare synchronized channels: Multi-channel measurements help compare signal behaviour across different sensors or measurement points.
- Review trends: Compare repeated measurements under relevant operating and test conditions rather than relying on one result alone.
These methods work together to improve signal interpretation, but no single technique can confirm by itself that a measured signal is genuine partial discharge.
How do you separate partial discharge from noise?
No single characteristic should be used alone to classify a signal. Instead, combine pattern, phase, repeatability, channel behavior, and other available evidence.
| Indicator | Genuine partial discharge | Noise / interference |
| Pattern characteristics | May form recognizable, recurring discharge patterns | May produce patterns unrelated to an insulation defect |
| Phase relationship | Can show repeatable phase-related behavior | May be asynchronous, although some interference can also be phase-related |
| Repeatability | May recur consistently under comparable electrical conditions | May change with external equipment or interference sources |
| Across synchronized channels | Relative signal behavior may help associate activity with measurement points | Common interference may appear across several channels in a different relationship |
These indicators support interpretation; they should not be treated as standalone proof of partial discharge origin.
Partial discharge testing by asset
The sensor arrangement and measurement method vary with the asset being tested.
Transformers: A partial discharge test of a transformer can use multiple sensors and synchronized channels to compare signal behavior across measurement points and support noise discrimination.
Power cables: Cable partial discharge testing requires sensors and a measurement arrangement suited to the cable system and test objective. See this underground cable PD measurement example for an application in the field.
Generators: Generator partial discharge testing measures discharge activity associated with stator winding insulation using an appropriate sensor and measurement arrangement.
Offline testing vs continuous monitoring
Offline partial discharge testing is performed under a controlled test arrangement when the asset is out of normal service. Continuous monitoring tracks partial discharge behavior while the asset operates over time.
The appropriate approach depends on the asset, diagnostic objective, and how frequently condition information is required. For a deeper comparison, see online vs offline partial discharge testing.
Where Rugged Monitoring fits
Rugged Monitoring supports both portable partial discharge testing and continuous monitoring, with noise-management capabilities suited to different measurement requirements.
For periodic diagnostics and field measurements, the HPM601-P provides four synchronous measurement channels, with an optional eight-channel configuration. It supports Enhanced Noise Management, a sampling rate of 250 MS/s per channel, a 0.01–100 MHz frequency range, and an IP65-rated enclosure.
For applications requiring ongoing visibility, the PD201 supports four or eight synchronous measurement channels. Its noise-management capabilities include noise gating with built-in filters, and the system provides EMI/ESD immunity. It is compatible with HF sensors including HFCT, TEV, capacitive couplers, acoustic, and ultrasonic sensors.
The appropriate setup depends on the electrical asset, measurement environment, and type of visibility required.
Dealing with a difficult partial discharge measurement environment? Contact Rugged Monitoring to discuss the testing or monitoring approach for your application.
Frequently Asked Questions
What is partial discharge testing used for?
PD testing assesses discharge activity associated with insulation defects or deterioration in high-voltage electrical assets and supports condition assessment and diagnostic decisions.
What causes noise in partial discharge testing?
Noise can come from external corona, switching activity, drives and power electronics, radio-frequency sources, grounding, and other electrical activity around the measurement setup.
How do you separate partial discharge from electrical noise?
Signal interpretation can combine background-noise assessment, phase-resolved patterns, filtering and gating, repeatability, synchronized channel comparison, and trends. No single characteristic should be relied on alone.
How is a partial discharge test of a transformer done?
The test uses sensors and a suitable measurement system to acquire PD signals from selected transformer measurement points. The specific sensor arrangement and test condition depend on the transformer and measurement method.
How does multi-channel measurement help distinguish PD from noise?
Synchronized channels allow signal behavior at different measurement points to be compared. These relationships can provide additional evidence for distinguishing localized PD activity from interference common to several channels.
