TEV vs Ultrasonic PD Detection: A Technical Comparison for Ring Main Unit Monitoring

Introduction

Understanding these differences is essential for asset managers and condition monitoring engineers who must specify monitoring systems, interpret measurement data, and make maintenance decisions based on PD survey results. A monitoring strategy that relies exclusively on one method will inevitably miss certain classes of PD activity, while a dual-mode approach that intelligently combines both techniques offers substantially higher detection probability and fewer false positives.

Britop Electric’s SCYC Integrated Online Monitoring Device implements precisely this dual-mode philosophy, incorporating both TEV and ultrasonic sensing channels with edge-computing-based correlation algorithms. This article provides a rigorous technical comparison of the two methods, examines their complementary nature, and discusses practical deployment considerations for RMU monitoring applications.

Physical Principles of Operation

Transient Earth Voltage (TEV) Sensing

When a partial discharge event occurs within the insulation structure of a switchgear panel, the rapid collapse of the electric field across the discharge site generates a current pulse with a rise time typically in the nanosecond range. This high-frequency transient current propagates along the internal conductors and the metallic enclosure of the switchgear.

At any discontinuity in the metallic enclosure — panel joints, gaskets, ventilation grilles, viewing windows — the current path is interrupted, and a transient voltage appears on the external surface of the panel relative to earth potential. This is the Transient Earth Voltage. The physical mechanism is capacitive coupling: the internal discharge current induces a voltage on the external surface through the stray capacitance between the internal live parts and the external enclosure surface.

TEV sensors are capacitive coupling probes that are placed in direct contact with the external surface of the switchgear panel. The sensor detects the time-varying electric field associated with the TEV signal and converts it to a voltage output proportional to the PD magnitude. Typical TEV sensors have a bandwidth extending from approximately 3 MHz to 100 MHz, capturing the high-frequency content that distinguishes PD from lower-frequency phenomena such as switching transients and power system harmonics.

Key characteristics of TEV sensing:

• Non-invasive: sensors are externally mounted with no requirement to open panels or de-energize equipment
• Wide bandwidth: capable of capturing the full spectral content of PD pulses, enabling sophisticated pulse-shape analysis
• Quantitative: TEV amplitude in dBmV or dBμV can be trended over time to track PD severity progression
• Sensitive to internal PD: TEV signals are strongest for PD occurring within solid insulation (voids, delamination) where the discharge site is in close proximity to the grounded enclosure

Ultrasonic Acoustic Emission Detection

The physical process of partial discharge involves the rapid ionization and recombination of gas molecules within the discharge channel. This produces a localized, impulsive pressure wave that propagates through the surrounding medium — air, SF6 gas, oil, or solid insulation — as an acoustic emission. The frequency spectrum of this acoustic emission extends from the audio range into the ultrasonic region, with significant energy content between 20 kHz and 100 kHz.

Ultrasonic PD sensors are typically piezoelectric transducers tuned to a specific frequency band within this range. Contact ultrasonic sensors are mechanically coupled to the switchgear enclosure or directly to insulation surfaces, while airborne ultrasonic sensors detect PD through the air gap within the switchgear compartment.

The propagation characteristics of ultrasonic emissions are fundamentally different from electromagnetic TEV signals:

• Directional propagation: acoustic waves travel at the speed of sound in the medium (approximately 343 m/s in air at 20°C, 1500 m/s in oil, and 3000-5000 m/s in solid insulation), which is many orders of magnitude slower than electromagnetic propagation
• Attenuation: acoustic waves experience geometric spreading loss and absorption, with attenuation increasing with frequency. Higher-frequency components (above 60 kHz) attenuate more rapidly in air, limiting detection range but improving spatial resolution
• Reflection and refraction: acoustic waves reflect from impedance discontinuities — material boundaries, air gaps, structural members — creating complex multipath propagation patterns within switchgear enclosures

Key characteristics of ultrasonic sensing:

• EMI immunity: completely unaffected by electromagnetic interference, making it ideal for electrically noisy substation environments with high levels of corona, switching transients, and radio-frequency interference
• Source localization: the relatively slow propagation speed and directional nature of ultrasonic emissions enable physical localization of PD sources through time-of-flight triangulation or amplitude mapping
• Surface PD sensitivity: ultrasonic methods excel at detecting surface tracking, corona, and floating electrode discharges, which produce strong acoustic signatures
• Non-electrical faults: ultrasonic sensors can also detect mechanical defects such as loose components, bearing wear, and gas leaks, providing additional diagnostic value

Comparative Sensitivity Analysis

Detection Sensitivity by PD Type

The relative sensitivity of TEV and ultrasonic methods varies significantly depending on the physical nature of the partial discharge. This variation is fundamental to understanding why a dual-mode approach is necessary.

PD Type	TEV Sensitivity	Ultrasonic Sensitivity	Notes
———	—————–	————————	——-
Internal void discharge in epoxy	High	Low-Medium	Solid insulation attenuates acoustic emissions significantly; TEV couples efficiently
Surface tracking on insulation	Medium	High	Tracking produces strong acoustic emissions at the surface; TEV signal depends on proximity to enclosure
Corona from sharp conductor	Low-Medium	High	Corona in air produces characteristic “frying” ultrasonic noise; TEV coupling is weak unless discharge is near panel surface
Floating electrode discharge	High	High	Large discharge magnitude produces strong signals in both domains
Particle discharge in GIS	Medium	High	Moving particles produce distinctive acoustic signatures; TEV signals are attenuated by gas insulation
Delamination in paper-oil insulation	High	Low	Internal discharge in oil-impregnated paper couples well electromagnetically but is acoustically shielded

Signal-to-Noise Ratio Considerations

In real substation environments, the limiting factor for PD detection is rarely the absolute sensitivity of the sensor but rather the signal-to-noise ratio (SNR). The noise environment for TEV and ultrasonic sensing is fundamentally different:

TEV noise sources:

• Corona from adjacent equipment producing broadband electromagnetic interference from 1-300 MHz
• Radio-frequency interference from communication equipment, particularly in the 30-300 MHz VHF band
• Switching transients from circuit breaker and contactor operations
• Partial discharge in adjacent switchgear panels (cross-coupling)
• Mobile phone and wireless data signals in the 700 MHz to 2.6 GHz bands (typically out of band but can cause intermodulation in poorly designed front-end electronics)

Ultrasonic noise sources:

• Mechanical vibration from transformers (magnetostriction at 100/120 Hz and harmonics)
• Cooling fan noise (typically below 20 kHz but with harmonics extending into the ultrasonic range)
• Wind-induced vibration of external structures
• Rain and hail impact on outdoor equipment enclosures
• Rodent and insect activity within switchgear enclosures
• Compressed air leaks in pneumatically operated switchgear

The SCYC device addresses these noise challenges through several signal processing strategies:

• Frequency-domain filtering: the TEV channel employs a bandpass filter that rejects both the low-frequency switching transient energy (below 1 MHz) and the high-frequency communication bands (above 100 MHz). The ultrasonic channel uses a narrow bandpass filter centered at 40 kHz, which is above most mechanical noise sources yet within the primary PD acoustic emission band.
• Pulse-shape discrimination: genuine PD pulses have characteristic rise times and pulse widths that differ from noise sources. The on-board DSP implements pulse-shape analysis algorithms that reject pulses whose temporal characteristics fall outside the expected PD envelope.
• Coincidence detection: a PD event that produces simultaneous TEV and ultrasonic signatures is almost certainly genuine, as it is statistically improbable for independent noise sources in both domains to produce coincident triggers. The SCYC device implements a configurable coincidence window (typically 100 μs to 1 ms) within which TEV and ultrasonic triggers are correlated.

Spatial Resolution and Source Localization

One of the most significant practical differences between TEV and ultrasonic methods is their capability for PD source localization.

TEV-Based Localization

TEV signals propagate electromagnetically throughout the interconnected metallic structure of a switchgear assembly. Because the propagation velocity is near the speed of light, time-of-arrival differences between sensors on different panels are extremely small (on the order of nanoseconds) and require specialized high-speed sampling hardware to resolve.

In practice, TEV-based localization relies on amplitude comparison: the TEV amplitude is measured at multiple positions on the switchgear enclosure, and the PD source is assumed to be within the panel showing the highest amplitude. This method provides panel-level localization but cannot pinpoint the exact location within a panel (e.g., which bushing or cable termination).

Ultrasonic-Based Localization

Ultrasonic localization is substantially more precise due to the much slower propagation speed of acoustic waves. With a propagation velocity of approximately 343 m/s in air, a 100 μs time-of-flight difference between two sensors corresponds to a path length difference of only 34 mm. This enables centimeter-level localization accuracy within a switchgear panel.

Practical localization methods include:

• Time-of-flight triangulation: using three or more sensors with known positions, the PD source location is calculated from the relative arrival times of the ultrasonic pulse at each sensor
• Amplitude mapping: systematically moving a single sensor and recording the ultrasonic amplitude at each position, with the highest amplitude position indicating the closest proximity to the source
• Phase-array techniques: using an array of closely spaced ultrasonic sensors and applying beamforming algorithms to electronically steer the array’s sensitivity pattern toward the PD source

For ring main unit monitoring, the SCYC device typically deploys one ultrasonic sensor per cable compartment, providing compartment-level localization that is sufficient to guide maintenance personnel to the correct section of the RMU for detailed inspection.

Practical Deployment for Ring Main Units

Sensor Installation Considerations

Ring main units present specific challenges for PD sensor installation due to their compact construction and the need to maintain the equipment’s original insulation coordination and ingress protection ratings.

TEV sensor installation:

• Externally mounted on the front panel surface using magnetic or adhesive attachment
• Optimal position: center of each functional unit panel, away from edges and joints where TEV signal attenuation occurs
• Must maintain clearance from panel edges to avoid partial discharge from the sensor itself
• One sensor per functional unit (typically 3-4 per RMU) provides adequate coverage

Ultrasonic sensor installation:

• Internal mounting within cable compartments provides optimal acoustic coupling
• When internal mounting is not feasible (equipment must not be opened), contact sensors can be externally mounted on panel surfaces, though sensitivity is reduced by 10-20 dB due to acoustic impedance mismatch at the air-metal interface
• Sensor must be firmly coupled to the mounting surface using acoustic coupling gel or a spring-loaded mounting fixture to eliminate air gaps
• One sensor per cable compartment is typical; additional sensors may be deployed in busbar compartments if accessible

Environmental Factors

RMU deployment environments range from clean indoor substations to pole-mounted outdoor installations exposed to extreme temperatures, humidity, and pollution. These environmental factors affect both detection methods:

• Temperature: TEV sensor sensitivity is relatively stable across the -40°C to +80°C operating range of the SCYC device. Ultrasonic sensor sensitivity may vary by ±3 dB over this range due to changes in piezoelectric element properties and acoustic coupling gel viscosity. The SCYC device compensates for this through factory-characterized temperature compensation curves.
• Humidity and condensation: high humidity can increase surface conductivity on insulation, potentially reducing PD inception voltage but also attenuating TEV signals due to increased surface leakage. Ultrasonic propagation in air is affected by humidity-dependent absorption, with attenuation increasing at higher humidity levels for frequencies above 50 kHz.
• Pollution and contamination: conductive pollution layers on insulation surfaces promote surface tracking and increase PD activity. Contamination on sensor surfaces can affect both TEV (changes in capacitive coupling) and ultrasonic (changes in acoustic impedance matching), necessitating periodic sensor cleaning in heavily polluted environments.

Integration with Asset Management Systems

The value of PD monitoring data is realized when it is integrated with the utility’s wider asset management and maintenance planning processes. The SCYC device supports this integration through standard communication protocols and data formats.

Data Trending and Analysis

Continuous monitoring generates large volumes of time-series data. Effective analysis requires:

• Baseline establishment: a commissioning survey establishes the normal PD activity levels for each monitored asset, providing a reference against which future measurements are compared
• Trend analysis: PD activity that shows a consistent upward trend over weeks or months is more significant than a single elevated reading, which may be caused by transient environmental conditions
• Correlation with operational parameters: PD activity often correlates with load current, voltage, temperature, and humidity. Correlating PD data with these parameters helps distinguish load-dependent PD (indicative of conductor or connection problems) from voltage-dependent PD (indicative of insulation defects)

Alarm Management

Effective alarm management prevents “alarm fatigue” while ensuring that genuinely critical conditions receive prompt attention:

• Multi-level alarming: the SCYC device supports configurable “alert” (advisory) and “alarm” (urgent) thresholds for each monitored parameter
• Persistence requirements: alarms should require the threshold to be exceeded for a configurable minimum duration or number of consecutive measurement cycles before triggering, preventing nuisance alarms from transient events
• Rate-of-change alarming: in addition to absolute thresholds, the rate of change of PD activity can trigger alarms, providing early warning of rapidly developing faults

Conclusion: The Case for Dual-Mode PD Detection

The technical comparison presented in this article demonstrates that TEV and ultrasonic PD detection methods are fundamentally complementary rather than competitive. Each method has distinct strengths and limitations:

• TEV provides broad-area, non-invasive screening with high sensitivity to internal insulation defects but is susceptible to electromagnetic interference and cannot precisely localize PD sources
• Ultrasonic provides precise source localization with immunity to electromagnetic interference and high sensitivity to surface and corona discharges but has limited sensitivity to deeply buried internal defects

A monitoring strategy that combines both methods, as implemented in Britop Electric’s SCYC Integrated Online Monitoring Device, achieves comprehensive PD detection coverage with high confidence and low false-positive rates. The device’s edge-computing capabilities enable real-time correlation of TEV and ultrasonic signals, automatically classifying PD types and providing actionable diagnostic information to maintenance personnel.

For utilities and industrial operators responsible for 10kV and above distribution switchgear, the investment in dual-mode online PD monitoring delivers a compelling return through reduced unplanned outages, extended equipment life, and optimized maintenance expenditure — transitioning asset management from a reactive, time-based model to a predictive, condition-based approach.

---

Related Products

• Eagle-beak Shear Excavator Attachment – QINGDAO BRITOP QUALITY PRODUCT
• Auger Excavator Attachment – QINGDAO BRITOP QUALITY PRODUCT
• Steel Structure Shear Excavator Attachment – QINGDAO BRITOP QUALITY PRODUCT
• Jaw Crushing Pliers Excavator Attachment – QINGDAO BRITOP QUALITY PRODUCT
• Jaw Crushing Bucket Excavator Attachment – QINGDAO BRITOP QUALITY PRODUCT