Transformer Winding Frequency Response Testing: Principles, Methods, and Applications

Transformer winding frequency response testing (SFRA) is a powerful, non-invasive diagnostic technique used to detect mechanical deformation, winding displacement, and internal structural damage in power transformers. By comparing a transformer's frequency response signature against baseline fingerprints or sister-unit records, engineers can assess winding integrity without costly detanking or visual inspection. This article explains the technical principles, testing procedures, applicable standards, and practical applications of frequency response analysis for transformer windings.

Technical Overview

Transformer winding frequency response testing, commonly known as Sweep Frequency Response Analysis (SFRA), is a condition assessment technique that evaluates the mechanical and electrical integrity of transformer windings. A transformer winding behaves as a complex distributed network of resistances, inductances, and capacitances (R-L-C). When a winding experiences mechanical deformation—such as axial displacement, radial buckling, tilting, or hoop distortion—the geometric relationships between conductor turns, insulation, and the core change, which in turn alters the distributed inductance and capacitance parameters of the network.

By injecting a low-voltage sinusoidal signal across a defined frequency range (typically 10 Hz to 2 MHz, and in some instruments up to 10 MHz) and measuring the transfer function between input and output terminals, the test captures a unique frequency response "fingerprint" for each winding. Any deviation between the current response curve and a reference measurement indicates a change in the internal structure of the transformer. Because the method is non-destructive, repeatable, and highly sensitive to subtle mechanical changes, it has become the industry-standard approach for detecting winding deformation after short-circuit events, transportation, or as part of routine condition monitoring programs.

The governing international standard is IEC 60076-18, which defines measurement procedures, connection configurations, and interpretation criteria. In North America, IEEE C57.149 provides equivalent guidance, while the Chinese power industry standard DL/T 911 is widely applied in domestic grid applications. These standards establish consistent measurement conditions so that results remain comparable over the transformer's service life.

Testing Samples

Power Transformers

Large power transformers used in generation plants, transmission substations, and receiving substations are the primary application for SFRA. Both new units (to establish baseline fingerprints) and in-service units (for condition assessment) are tested. Typical voltage classes range from 35 kV up to 1000 kV, with capacities from a few MVA to over 1000 MVA.

Distribution Transformers

Medium-voltage distribution transformers suspected of winding damage after through-faults, lightning strikes, or overloading can be screened using SFRA to confirm internal integrity before re-energization.

Instrument Transformers and Reactors

Shunt reactors, series reactors, and certain voltage transformers with layered or interleaved windings can also be evaluated, provided the test instrument supports the appropriate impedance and frequency ranges.

Transformer Components After Manufacture or Repair

Windings and complete active parts are tested during factory acceptance testing and after on-site repair or relocation to verify that no deformation occurred during handling, transport, or reassembly.

Testing Items

The frequency response test evaluates several characteristic features of the response curve, each corresponding to specific structural elements of the transformer:

• Low-frequency region (10 Hz – 2 kHz): reflects the core laminations, core clamping condition, and bulk inductive behavior; changes may indicate core movement, core saturation differences, or shorted turns.
• Mid-frequency region (2 kHz – 20 kHz): reflects bulk winding behavior including overall winding deformation, axial displacement, and changes in series capacitance between windings.
• High-frequency region (20 kHz – 2 MHz): reflects detailed internal winding structure such as turn-to-turn capacitance, inter-disk capacitance, and lead geometry; deviations suggest localized deformation, tilting of conductors, or clamping structure relaxation.
• Cross-comparison between phases: identifies asymmetries between windings of the same transformer.
• Time-based comparison: compares current measurements with baseline fingerprint records taken at factory acceptance or initial commissioning.
• Construction-based comparison: compares against identical sister transformers of the same design and manufacturer.
• Open-circuit and short-circuit response measurements: distinguishes winding deformation from core-related changes.

Testing Methods

Test Principle

A variable-frequency sinusoidal excitation signal of constant amplitude (typically 1 to 10 V peak) is applied to one end of a winding, and the response signal is measured at the other end. The transfer function H(f) is calculated as the logarithmic amplitude ratio of output to input voltage in decibels: H(f) = 20 × log₁₀(Vout / Vin), plotted against frequency. Both amplitude and phase characteristics may be recorded. Deviations in resonance peaks and anti-resonance troughs indicate changes in the distributed R-L-C parameters, which are directly linked to mechanical geometry.

Standard Connection Configurations

IEC 60076-18 defines several canonical measurement setups:

• End-to-end (open-circuit) measurement: signal applied at one bushing terminal and measured at the neutral or opposite terminal; the primary configuration for winding deformation detection.
• End-to-end short-circuit measurement: the far end of the winding is shorted to neutral; this suppresses the core's influence and highlights inter-winding and winding-to-ground capacitance changes, useful for distinguishing winding deformation from core issues.
• Capacitive inter-winding measurement: signal applied to one winding and response measured on an adjacent winding to assess inter-winding structure.
• Inductive inter-winding measurement: response measured across another winding to evaluate common-mode magnetic coupling paths.

Test Procedure

• De-energize and isolate the transformer; apply working-ground procedures and confirm adequate safety clearance.
• Disconnect surge arresters, bushing potential devices, and other auxiliary components connected to the bushings to prevent parallel impedance paths from distorting the measurement.
• Connect test leads with consistent, documented routing; ground the instrument properly and maintain short, direct cable paths to minimize electromagnetic interference.
• Select the measurement configuration and frequency sweep range according to the applicable standard and the purpose of the test.
• Perform the sweep for each phase and each winding group; record amplitude and phase responses.
• Repeat measurements to verify reproducibility; correlation coefficients between repeated traces should meet the criteria in IEC 60076-18.
• Overlay current curves with reference fingerprints and analyze deviations by frequency band.
• Classify results according to interpretation guidelines: good condition, satisfactory condition with minor changes, or suspicious condition requiring further investigation.

Interpretation Criteria

IEC 60076-18 Annex B provides numerical tolerance bands between reference and current measurements, defined separately for low, mid, and high frequency ranges. Statistical correlation metrics are used to quantify similarity between curves. Interpretation should always consider the frequency band in which deviation occurs, the magnitude of deviation, and the transformer's service history (fault events, through-faults, transportation).

Applications

• Post-fault assessment after exposure to short-circuit currents, through-faults, or lightning impulses to determine whether winding deformation has occurred before re-energization.
• Pre-commissioning baseline fingerprinting during factory acceptance testing and site commissioning of new transformers.
• Transport impact verification: comparing measurements before and after shipment detects deformation caused by shock, vibration, or improper handling during road, rail, or sea transport.
• Condition-based maintenance programs: periodic SFRA measurement as part of transformer life management and asset health indexing in transmission and distribution networks.
• Verification after on-site repair, detanking inspection, or active part replacement.
• Detection of specific failure modes including axial winding displacement, radial (hoop) buckling, conductor tilting and spiraling, loss of clamping pressure, partial winding collapse, short-circuited turns, broken winding strands, and core clamping degradation.
• Forensic analysis after transformer failures to establish the root cause and distinguish electrical failure from antecedent mechanical damage.
• Acceptance testing to verify manufacturing quality and design consistency across production batches of identical transformer designs.

FAQ

What is the difference between SFRA and a routine electrical test?

Routine tests such as winding resistance, insulation resistance, and turns-ratio measurement detect gross electrical faults, whereas SFRA is uniquely sensitive to mechanical geometry changes inside the winding. A transformer can pass all routine electrical tests yet show significant winding deformation detectable only through frequency response analysis.

Can SFRA be performed on an energized transformer?

No. SFRA requires the transformer to be completely de-energized, isolated, and grounded, with auxiliary components disconnected from the bushings. Online measurement is not feasible because the injected signal would be overwhelmed by power-frequency voltage and the connection geometry would be unsafe.

How often should frequency response testing be performed?

Best practice is to record a baseline fingerprint at commissioning and repeat measurements after any known through-fault event, after transport or relocation, following repair, and at periodic intervals (typically every 3 to 6 years) as part of condition-based maintenance, or whenever dissolved gas analysis indicates abnormal internal activity.

What standards govern SFRA testing?

The primary standards are IEC 60076-18 (Measurement of frequency response), IEEE C57.149 (IEEE Guide for the Application and Interpretation of Frequency Response Analysis for Oil-Immersed Transformers), and DL/T 911 for Chinese power industry applications. These documents define connection configurations, frequency ranges, repeatability requirements, and interpretation tolerance bands.

What does a deviation in the low-frequency region indicate?

Deviations below approximately 2 kHz are usually associated with the magnetic core—such as core clamping relaxation, residual magnetization differences, or in severe cases shorted core laminations or short-circuited turns. Residual magnetism from a prior DC resistance test is a common benign cause of low-frequency differences, which is why demagnetization before testing is recommended.

How are SFRA results classified?

Following IEC 60076-18, results are typically classified into three categories: good condition (curves match within tolerance), satisfactory condition (minor deviations warranting continued monitoring), and suspicious condition (significant deviations requiring internal inspection or complementary diagnostics such as DGA, winding capacitance measurement, or low-voltage impulse testing by accredited third-party laboratories).

Can SFRA quantify the severity of deformation?

SFRA is a comparative technique that identifies the presence, location, and general nature of structural change, but it does not provide a direct geometric measurement of deformation. Severity is inferred from the magnitude and frequency band of deviation, combined with fault-current records and complementary diagnostic tests.