A Complete Guide to Partial Discharge Acoustic Detection

Introduction

Partial discharge in current transformer insulation systems is the most reliable early warning of impending insulation failure — and acoustic emission detection is the most practically deployable method for identifying active partial discharge in installed power distribution CTs without taking the equipment out of service. A CT that is actively discharging internally is communicating its deteriorating condition through ultrasonic acoustic signals that propagate through its insulation medium and housing — signals that are detectable with the piezoelectric sensor¹ equipment, interpretable with the right methodology, and actionable with the right maintenance response, all without a single minute of planned outage.

The direct answer is this: partial discharge acoustic detection in power distribution CTs works by detecting the ultrasonic pressure waves — typically in the ultrasonic frequency range² — that are generated each time a partial discharge event occurs within the CT insulation system, and the technique is uniquely valuable for installed CT maintenance because it is non-invasive, does not require secondary circuit disconnection, can be performed under energized conditions, and provides location information that electrical partial discharge measurement methods cannot — enabling maintenance teams to distinguish between internal CT insulation defects that require urgent replacement and external corona sources that require no CT intervention.

For power distribution maintenance engineers, insulation condition assessment specialists, and reliability teams responsible for CT fleet management, this guide provides the complete technical framework for acoustic emission partial discharge detection — from the physics of acoustic signal generation through sensor selection, measurement methodology, signal interpretation, and maintenance decision-making.

Table of Contents

• What Is Partial Discharge in CT Insulation Systems and How Does Acoustic Emission Detection Work?
• How to Select and Position Acoustic Emission Sensors for CT Partial Discharge Detection?
• How to Execute a Structured CT Acoustic Partial Discharge Measurement Campaign?
• How to Interpret Acoustic Emission Signals and Make CT Maintenance Decisions?
• FAQs About Partial Discharge Acoustic Detection in Power Distribution CTs

What Is Partial Discharge in CT Insulation Systems and How Does Acoustic Emission Detection Work?

Partial discharge is an electrical discharge that bridges only part of the insulation between conductors — it does not constitute a complete breakdown path between the high-voltage conductor and earth, but it progressively degrades the insulation material surrounding the discharge site until a complete breakdown path eventually forms. In CT insulation systems — whether oil-paper, cast resin epoxy, or SF₆ gas — partial discharge is the primary degradation mechanism that converts an insulation system from serviceable to failed over a timescale that ranges from months to years depending on discharge intensity and insulation type.

The Physics of Partial Discharge in CT Insulation

Partial discharge occurs at sites of insulation weakness — voids in cast resin, gas bubbles in oil-paper insulation, delamination interfaces, metallic inclusions, and regions of locally elevated electric field stress. At these sites, the local electric field exceeds the breakdown strength of the insulation medium within the defect — typically a gas-filled void where the dielectric strength is much lower than the surrounding solid or liquid insulation.

When the local field exceeds the void breakdown strength, a rapid discharge occurs within the void — lasting nanoseconds to microseconds. This discharge:

• Electrically: Produces a current pulse in the primary circuit and a corresponding induced pulse in the secondary circuit — the basis of electrical PD measurement methods
• Thermally: Deposits energy at the discharge site, carbonizing the surrounding insulation material and enlarging the void over successive discharge cycles
• Acoustically: Creates a rapid local pressure change — a mechanical impulse — that propagates outward from the discharge site as an acoustic wave through the surrounding insulation medium and CT housing

The acoustic emission from a partial discharge event is a broadband pressure pulse with significant energy content in the ultrasonic frequency range of 20–500 kHz. The signal propagates through the CT insulation medium — oil, resin, or gas — and through the CT housing walls, attenuating with distance and reflecting at material interfaces, until it reaches the outer surface of the CT where it can be detected by a contact piezoelectric sensor.

Key technical parameters defining CT acoustic partial discharge detection:

• Acoustic emission frequency range: 20–300 kHz for internal CT PD; peak energy typically at 80–150 kHz for oil-paper CT insulation; 100–250 kHz for cast resin CT insulation
• Signal propagation velocity: 1,400–1,500 m/s in transformer oil; 2,500–3,500 m/s in cast resin epoxy; 5,100 m/s in steel housing — velocity differences enable source location by time-of-arrival methods
• Signal attenuation: 6–12 dB per 100 mm in oil; 15–25 dB per 100 mm in cast resin; attenuation increases with frequency — lower frequency components propagate further from the discharge source
• Detection threshold: Minimum detectable PD charge equivalent approximately 100–500 pC for contact piezoelectric sensors on CT housing; electrical PD measurement is more sensitive (5–10 pC) but requires secondary circuit access
• Sensor frequency response: Broadband piezoelectric sensors: 20–300 kHz flat response; resonant piezoelectric sensors: peak sensitivity at 150 kHz ±20%; resonant sensors provide higher sensitivity at design frequency but miss signals outside the resonant band
• Applicable standards: IEC 60270³ (electrical PD measurement — reference method), IEC 62478 (high voltage test techniques — acoustic emission), IEC 60599 (dissolved gas analysis — complementary diagnostic method)

The acoustic emission detection advantage over electrical PD measurement in field maintenance applications:

Electrical PD measurement per IEC 60270 is the reference method for PD quantification — it provides calibrated charge measurements in picocoulombs and is the method used for factory acceptance testing. However, electrical PD measurement in the field requires access to the CT secondary circuit, a calibrated coupling capacitor, and a noise-free measurement environment — conditions that are rarely achievable in an energized power distribution substation. Acoustic emission detection requires only physical access to the CT housing surface — it can be performed with the CT fully energized, under load, without any secondary circuit modification, and in the presence of the electromagnetic noise environment that makes electrical PD measurement impractical in the field.

How to Select and Position Acoustic Emission Sensors for CT Partial Discharge Detection?

Sensor selection and positioning are the two most influential variables in acoustic PD detection quality — a correctly selected sensor at the wrong position will miss internal PD signals, and a correctly positioned sensor with the wrong frequency response will detect external interference rather than internal discharge.

Sensor Selection for CT Acoustic PD Detection

Piezoelectric Contact Sensors (Primary Method):
Contact piezoelectric sensors are pressed against the CT housing surface and detect acoustic waves transmitted through the housing wall. They provide the highest sensitivity for internal PD detection and are the standard method for CT acoustic PD surveys.

Selection criteria:

• Frequency range: 50–200 kHz for oil-immersed CTs; 80–300 kHz for cast resin CTs — the higher attenuation of resin requires higher frequency sensitivity to detect signals from the discharge source before they attenuate to noise floor
• Sensitivity: Minimum -65 dB ref 1 V/μbar for reliable detection of PD sources at distances up to 300 mm through oil; minimum -55 dB for cast resin applications
• Housing compatibility: Magnetic mounting base for ferromagnetic CT housings — provides consistent coupling force and repeatable sensor positioning for trend monitoring; adhesive coupling for non-ferromagnetic housings

Airborne Ultrasonic Sensors (Supplementary Method):
Non-contact ultrasonic sensors detect airborne acoustic emission from surface corona and external PD sources. They are used to distinguish external corona — which produces strong airborne signals but weak contact signals — from internal PD, which produces strong contact signals but weak airborne signals.

Sensor Positioning for Different CT Types

Oil-Immersed CT (Porcelain or Composite Bushing):

• Primary sensor position: Lower tank wall, 50–100 mm above the tank base — oil-borne acoustic signals from internal PD sources propagate downward and concentrate at the tank base; this position maximizes signal-to-noise ratio for internal PD detection
• Secondary sensor position: Mid-tank wall at 90° to primary sensor — enables two-dimensional source location by time-of-arrival comparison
• Avoid: Bushing surface — external corona on the bushing surface produces strong acoustic signals that will mask internal PD signals if the sensor is positioned on the bushing

Cast Resin CT (Epoxy Encapsulated):

• Primary sensor position: Base of the CT body, directly on the epoxy surface — cast resin has higher acoustic attenuation than oil, requiring sensor placement as close as possible to the expected PD source location (typically the high-voltage conductor interface or the core-resin interface)
• Secondary sensor positions: At 120° intervals around the CT body circumference — enables three-point source location for resin-encapsulated CTs
• Coupling medium: Acoustic coupling gel mandatory for cast resin — the surface roughness of epoxy creates air gaps that severely attenuate high-frequency signals without coupling gel

Coupling Quality Verification

Before recording PD measurements, verify acoustic coupling quality:

$SNR_{coupling} = 20 \times \log_{10}\left(\frac{V_{signal}}{V_{noise}}\right) \geq 6 \text{ dB}$

Apply a pencil lead break (Hsu-Nielsen source) on the CT housing surface 100–200 mm from the sensor — this produces a broadband acoustic impulse that verifies the sensor is correctly coupled and the signal path is intact. A correctly coupled sensor will show a clean impulse response with SNR ≥ 6 dB above the background noise floor.

How to Execute a Structured CT Acoustic Partial Discharge Measurement Campaign?

A structured acoustic PD measurement campaign for a power distribution CT fleet requires a defined measurement protocol that enables comparison between CTs, between measurement periods, and between the CT under test and a known-healthy reference — because absolute acoustic signal levels are meaningless without context; it is relative levels and trends that identify deteriorating insulation.

Step 1: Establish Baseline Measurements

Before acoustic PD detection can identify deteriorating CTs, baseline measurements must be established for each CT in the fleet under known-healthy conditions:

• Record baseline at commissioning or last known healthy condition: Measure and document acoustic signal level, frequency spectrum, and phase-resolved pattern for each CT at the time of commissioning or immediately after a confirmed healthy insulation test
• Document measurement conditions: Record primary voltage, primary current, ambient temperature, and weather conditions — acoustic PD signal levels vary with voltage (PD inception voltage) and temperature (insulation viscosity affects signal propagation in oil)
• Establish fleet reference: Identify the statistical distribution of acoustic signal levels across the CT fleet — CTs with signal levels more than 6 dB above the fleet median require investigation regardless of absolute level

Step 2: Define Measurement Sequence and Frequency

• Annual survey for CTs above 15 years service age: Insulation degradation accelerates in the second half of CT service life; annual acoustic PD surveys provide sufficient temporal resolution to detect deterioration before it reaches critical levels
• 6-monthly survey for CTs with known insulation concerns: CTs that showed elevated acoustic levels in the previous survey, CTs with abnormal dissolved gas analysis⁴ results, and CTs that have experienced thermal overload events
• Immediate survey after fault events: Any CT that has been subjected to a through-fault current exceeding 50% of rated short-time current requires acoustic PD assessment within 30 days — fault current thermal stress can initiate insulation degradation that manifests as PD within weeks of the fault event

Step 3: Execute Measurement Protocol

1. Prepare the measurement environment: Record ambient noise level with sensor coupled to CT housing but signal source disconnected — this establishes the noise floor for SNR calculation; if ambient noise exceeds -40 dBV at the measurement frequency band, identify and eliminate noise sources before proceeding
2. Apply sensor at defined positions: Use the CT-type-specific positioning defined in Step 1 of the sensor selection section; apply coupling gel for cast resin CTs; verify coupling quality with Hsu-Nielsen source test
3. Record time-domain waveform: Capture minimum 10 seconds of continuous acoustic signal at each sensor position — sufficient to observe multiple power frequency cycles and identify phase-correlated PD activity
4. Record frequency spectrum: FFT analysis of the captured waveform; identify peak frequency components; compare against baseline spectrum — new frequency components above baseline indicate new PD activity
5. Record phase-resolved pd pattern⁵: Synchronize acoustic measurement with power frequency voltage phase using a reference voltage signal; plot acoustic event amplitude versus phase angle — the PRPD pattern shape identifies the PD source type
6. Apply multi-sensor time-of-arrival analysis: If two or more sensors are deployed simultaneously, record the time difference of arrival (TDOA) of acoustic signals between sensor positions — enables source location calculation

Step 4: Source Location Calculation

For two sensors at known positions on the CT housing:

$\Delta d = v_{oil} \times \Delta t$

Where $\Delta t$ is the measured time difference of arrival and $v_{oil}$ is the acoustic propagation velocity in oil (1,450 m/s). The source lies on a hyperbola defined by the constant path length difference $\Delta d$ — with three or more sensors, the intersection of multiple hyperbolas provides a point source location.

For a CT with known internal geometry, source location accuracy of ±20–50 mm is achievable with three sensors and careful TDOA measurement — sufficient to distinguish between a PD source at the high-voltage conductor interface (most critical), the core-insulation interface (moderate severity), and the tank wall (lowest severity).

Application Scenarios

• Power Distribution Substation Annual CT Fleet Survey: Contact piezoelectric sensors at lower tank wall; single-sensor amplitude and spectrum survey; comparison against fleet baseline; flag CTs with >6 dB increase from baseline for follow-up multi-sensor survey
• Aged CT Insulation Condition Assessment (>20 years service): Multi-sensor deployment with PRPD analysis; TDOA source location; correlated with dissolved gas analysis results; maintenance decision based on combined acoustic and chemical evidence
• Post-Fault CT Insulation Assessment: Immediate single-sensor survey within 30 days of fault event; comparison against pre-fault baseline; elevated signal level triggers accelerated monitoring program
• New CT Commissioning Baseline: Full multi-sensor survey at commissioning; PRPD pattern recorded as reference; frequency spectrum documented; results stored in CT asset management record as lifetime baseline

How to Interpret Acoustic Emission Signals and Make CT Maintenance Decisions?

Signal Interpretation Framework

Acoustic PD signal interpretation requires distinguishing between four signal categories that produce overlapping amplitude ranges but have distinctly different frequency spectra, phase-resolved patterns, and maintenance implications:

Category 1: Internal Void Discharge (Most Critical)

• Acoustic characteristics: Repetitive impulses at 2× power frequency repetition rate (two discharge events per voltage cycle — one on positive half-cycle, one on negative); peak frequency 80–150 kHz; signal stronger on contact sensor than airborne sensor
• PRPD pattern: Symmetric clusters at 45° and 225° phase positions (positive and negative voltage peaks); amplitude distribution follows Gaussian distribution within each cluster
• Maintenance implication: Active internal insulation degradation — schedule replacement within next planned outage; increase monitoring frequency to monthly until replacement

Category 2: Surface Tracking Discharge (High Severity)

• Acoustic characteristics: Irregular impulse pattern; power frequency correlation present but asymmetric; peak frequency 50–100 kHz; signal detectable on both contact and airborne sensors
• PRPD pattern: Asymmetric clusters — stronger on one half-cycle than the other; irregular amplitude distribution indicating erratic discharge behavior
• Maintenance implication: Surface insulation degradation — typically at bushing-flange interface or core-resin interface; replacement required; do not defer beyond next scheduled outage

Category 3: External Corona (Low CT Severity)

• Acoustic characteristics: Continuous hiss rather than discrete impulses; strong airborne signal; weak or absent contact signal; peak frequency 20–50 kHz
• PRPD pattern: Concentrated at voltage zero-crossing points (90° and 270°); very consistent amplitude distribution
• Maintenance implication: External corona from adjacent conductors, insulators, or hardware — no CT insulation degradation; investigate and correct external corona source; no CT replacement required

Category 4: Mechanical Vibration and Interference (No PD)

• Acoustic characteristics: Continuous signal at power frequency and harmonics (50 Hz, 100 Hz, 150 Hz); no correlation with voltage phase; signal present on contact sensor but not phase-correlated
• PRPD pattern: Uniform distribution across all phase angles — no phase correlation
• Maintenance implication: Mechanical vibration from magnetostriction, loose components, or external mechanical sources — not a PD signal; no insulation concern; investigate mechanical source if vibration level is elevated

Maintenance Decision Flowchart

Correlation With Complementary Diagnostic Methods

Acoustic PD detection provides the most actionable field diagnostic — but its conclusions are strengthened by correlation with complementary methods:

• Dissolved Gas Analysis (DGA): Hydrogen (H₂) and methane (CH₄) generation in oil-immersed CTs confirms active PD; acetylene (C₂H₂) indicates high-energy arcing discharge; correlation between acoustic signal level increase and DGA gas generation rate confirms internal discharge source
• Thermal imaging (infrared): Hot spots on CT housing surface indicate resistive heating from tracking discharge paths; correlation with acoustic signals at the same location confirms surface discharge activity
• Electrical PD measurement (IEC 60270): Provides calibrated charge measurement in pC — required for definitive severity assessment; performed during planned outage with CT de-energized and secondary circuit accessible

Common Interpretation Mistakes

• Attributing all elevated acoustic signals to internal PD: External corona from adjacent hardware is the most common source of false-positive acoustic PD indications in power distribution substations; always compare contact and airborne sensor signals before concluding internal PD is present
• Making replacement decisions based on single-measurement amplitude alone: A single elevated amplitude reading without PRPD pattern analysis, frequency spectrum comparison, and baseline correlation provides insufficient evidence for a replacement decision; acoustic PD assessment requires the complete signal characterization package
• Ignoring acoustic signals below the “alarm threshold”: Progressive insulation degradation produces gradually increasing acoustic signal levels over months to years; a signal that is 3 dB above baseline today and 4 dB above baseline at the next survey is more concerning than a signal that is 6 dB above baseline but stable — trend is more informative than absolute level
• Performing acoustic PD survey immediately after a voltage transient or switching event: Switching operations produce acoustic signals that can persist for minutes in oil-immersed CTs; allow minimum 30 minutes after any switching operation before beginning acoustic PD measurements

Conclusion

Acoustic emission partial discharge detection is the most practically deployable condition monitoring technique available for installed power distribution CTs — it requires no outage, no secondary circuit access, no specialized substation infrastructure, and no modification to the CT or its connected circuits. The technique’s value is not in detecting PD at a single moment in time — it is in establishing a baseline for each CT in the fleet, trending the acoustic signal level over successive measurement campaigns, and using the phase-resolved pattern and frequency spectrum to distinguish the internal void discharge that requires urgent replacement from the external corona that requires no CT intervention. In power distribution CT fleet management, acoustic emission partial discharge detection is the maintenance investment that converts reactive CT failure response — emergency replacement after an unexpected insulation breakdown — into planned asset management, where deteriorating CTs are identified months before failure and replaced during scheduled outages without the safety risk, protection outage, and emergency procurement cost of an unplanned CT failure.

FAQs About Partial Discharge Acoustic Detection in Power Distribution CTs

1. Understand the underlying technology of piezoelectric sensors used in high-frequency acoustic monitoring. ↩

2. Explore the specific ultrasonic frequency characteristics produced by electrical discharge events. ↩

3. Access the official IEC 60270 standard for conventional electrical partial discharge measurement. ↩

4. Learn how dissolved gas analysis identifies insulation degradation through chemical indicators in oil. ↩

5. Detailed guide on how to interpret phase-resolved partial discharge patterns for diagnostic purposes. ↩