Transformer Noise Measurement: Standards, Test Methods, and Procedures

Transformer noise measurement is a standardized acoustic test used to verify that power transformers, distribution transformers, and reactors satisfy contractual sound-level guarantees and environmental noise regulations before commissioning. Conducted by accredited third-party laboratories and manufacturer test bays in accordance with IEC 60076-10 and its national adoptions such as GB/T 1094.10 and EN 60076-10, the test quantifies A-weighted sound pressure and sound power levels under no-load, load-current, and cooling-system operating conditions. This article explains the mechanisms of transformer noise generation, sample and test item requirements, measurement procedures, and practical applications of transformer sound level determination.

Technical Overview

Transformer noise determination (sound level measurement) is the process of quantifying the acoustic output of a transformer and expressing it as an A-weighted sound power level in decibels, dB(A). The test follows IEC 60076-10 and equivalent national standards such as GB/T 1094.10 and EN 60076-10, with measurement techniques rooted in the ISO 3740 and ISO 9614 series for sound power determination.

Transformer noise originates from three principal sources:

• Core noise, caused by magnetostriction of the grain-oriented electrical steel laminations. The core changes dimension slightly with each cycle of magnetization, producing a characteristic low-frequency hum at twice the supply frequency, typically 100 Hz for 50 Hz systems, together with harmonics at 200, 300, and 400 Hz.
• Winding noise, generated by electrodynamic forces between current-carrying conductors. These forces are proportional to the square of the load current and become significant at high loading levels.
• Auxiliary cooling equipment noise, produced by fans and oil pumps in forced-cooled units, which contributes broadband, mid-to-high frequency sound and often dominates when the cooling system is in operation.

The relative contribution of each source determines the acoustic signature of the unit. Modern noise reduction measures, including domain-refined (laser-scribed) core steel, step-lap core joints, amorphous alloy cores, optimized clamping pressure, damped tank panels, and low-noise fan designs, can reduce sound levels by 5 dB to 20 dB compared with conventional constructions. Accurate measurement is therefore essential for verifying design improvements and demonstrating compliance with increasingly strict noise specifications in urban substations and building installations.

Testing Samples

Oil-Immersed Power Transformers

Large generator step-up, transmission, and interconnection transformers, typically rated from tens of MVA to several hundred MVA with cooling classes such as ONAN, ONAF, OFAF, and ODAF. The sample should be a fully assembled unit including bushings, conservator, radiators, and any cooler banks mounted on or beside the tank, mounted on a hard reflecting floor in the test bay.

Dry-Type Transformers

Cast-resin and open-wound units used indoors in commercial buildings, infrastructure projects, and wind or solar plants. Because these transformers operate close to occupied spaces, they are frequently subject to stringent noise guarantees and are tested under the same acoustic procedures as oil-immersed units.

Distribution Transformers

Pole-mounted, pad-mounted, and compact secondary substation transformers. Their sound levels are commonly specified directly as sound pressure levels measured close to the unit, making routine noise verification part of factory acceptance testing.

Shunt Reactors and Specialty Units

Shunt reactors, converter transformers for HVDC links, and traction or rectifier transformers. These units may exhibit elevated magnetostrictive noise or harmonic-rich winding currents that increase electromagnetic force noise, and they therefore require careful acoustic characterization.

In all cases, the sample is accompanied by nameplate data including rated power, voltage ratio, frequency, vector group, and cooling class, and the tap position used during excitation, normally the principal tap, is documented in the test record.

Testing Items

The principal quantities determined during a transformer noise test are:

• A-weighted sound pressure level (LpA) in dB(A), measured at prescribed microphone positions on contours surrounding the transformer.
• A-weighted sound power level (LwA) in dB, derived from the averaged sound pressure level and the area of the measurement surface, and reported as the rated acoustic output of the unit.
• No-load (excitation) sound level, measured with the transformer energized at rated voltage and frequency and the secondary winding open-circuited, characterizing core magnetostriction noise.
• Load-current sound level, measured with one winding short-circuited and rated current flowing, characterizing winding electromagnetic noise; when measured at reduced current, results are converted to rated conditions using the correction formulas of the standard.
• Sound level with cooling equipment operating, together with the incremental contribution of fans and pumps obtained by comparing measurements with auxiliaries on and off.
• Octave-band or narrowband frequency analysis, used to identify discrete tonal components, typically the 100 Hz fundamental and its harmonics.
• Background sound pressure level and environmental correction factors, recorded to validate the measurement accuracy.

Testing Methods

Test Environment and Preparation

Measurements are preferably performed indoors in a large test hall or semi-anechoic chamber with a hard, reflective floor plane and no significant obstacles near the prescribed contours. Background noise is measured immediately before and after the test; it should ideally be at least 6 dB, preferably more than 10 dB, below the combined measured level at every position. Ambient temperature, humidity, and barometric pressure are recorded, and all instrumentation is verified with an acoustic calibrator compliant with IEC 60942 before and after measurement.

Sound Pressure Method

The conventional method of IEC 60076-10, based on ISO 3744 and ISO 3746, uses a prescribed measurement contour located 0.3 m from the principal radiating surface of the transformer, with the contour geometry adapted when separate cooling banks are present. Microphone positions are distributed along the contour at intervals not exceeding approximately 1 m, at a height corresponding to about half the height of the radiating surface. The averaged sound pressure level is corrected for background noise (K1) and for the acoustic environment (K2), and the sound power level is calculated as LwA = LpA − K1 − K2 + 10 log10(S/S0), where S is the measurement surface area and S0 equals 1 m².

Sound Intensity Method

As an alternative permitted by IEC 60076-10, the sound intensity method of ISO 9614-1 (discrete points) and ISO 9614-2 (scanning) employs a two-microphone intensity probe using instruments conforming to IEC 61043. Because sound intensity measurement suppresses steady background noise and is less sensitive to reverberation, it is particularly suitable for factory environments where background noise cannot be fully controlled. Field indicator checks are performed at each position to verify measurement validity.

Operating Conditions During Measurement

No-load measurements are performed with rated voltage applied at rated frequency, the secondary winding open, and the tap changer at the principal tap, capturing core-related noise. Load-current measurements are made with one winding short-circuited and a reduced voltage applied to circulate the agreed current, capturing winding noise; measurements at partial current are extrapolated to rated current using the conversion rules of the standard. Measurements are additionally repeated with fans and oil pumps running, and, where required, the no-load and load-current sound power levels are combined by energetic summation to obtain the total sound level.

Instrumentation and Accuracy

Sound level meters of accuracy class 1 conforming to IEC 61672, calibrated with an IEC 60942 calibrator, are used with windscreens where necessary. When performed under controlled conditions, the sound pressure and sound intensity methods both achieve engineering-grade (grade 2) accuracy, with typical expanded uncertainties in the range of 1 dB to 3 dB, and the reported result is compared against the guaranteed sound level, with consideration of measurement uncertainty.

Applications

Transformer noise measurement supports a wide range of engineering and regulatory purposes:

• Factory acceptance and type testing to demonstrate compliance with contractual sound-level guarantees before shipment.
• Environmental impact assessment and permitting for urban, residential, and hospital-adjacent substations subject to daytime and night-time boundary noise limits.
• Research and development of low-noise designs, including core material selection, step-lap construction, clamping optimization, tank stiffening, damping treatments, and low-noise cooling fans.
• Selection and qualification of dry-type transformers for installation inside office towers, hotels, data centers, and other noise-sensitive buildings.
• Post-installation verification in substations and troubleshooting of abnormal hum caused by loose core clamps, DC bias effects, or harmonic-rich converter loads.
• Export certification, since IEC 60076-10, EN 60076-10, and GB/T 1094.10 requirements are widely harmonized across international markets.

FAQ

Why do transformers hum at twice the line frequency?

Magnetostriction causes the core laminations to expand and contract once during each half-cycle of the magnetizing current, so a 50 Hz system produces a fundamental tone at 100 Hz, accompanied by harmonics. Electromagnetic forces in the windings and mechanical resonances of tank panels add further components to the acoustic spectrum.

What background noise level is acceptable during testing?

The background level at each microphone position should be at least 6 dB below the level measured with the transformer operating, and preferably more than 10 dB lower. When the difference is between 3 dB and 10 dB, the standard background correction is applied; if the difference falls below approximately 3 dB, the results become unreliable and measures such as night-time testing or shutting down adjacent equipment are required.

Which method is preferable: sound pressure or sound intensity?

The sound pressure method is the classical, straightforward approach and gives excellent results in a quiet, controlled test hall. The sound intensity method is more robust against background noise and reverberation and is better suited to busy factory environments. Both methods deliver comparable engineering-grade accuracy when their respective applicability conditions are satisfied.

What noise levels are typical for transformers?

Small distribution transformers commonly measure between approximately 45 dB(A) and 55 dB(A) as sound pressure close to the unit, while large forced-cooled power transformers may reach 70 dB(A) to 85 dB(A) as sound power or sound pressure depending on rating and cooling mode. Guaranteed values are agreed individually in the purchase specification.

Can noise be measured after the transformer is installed?

Yes. Field measurements are carried out following the guidance of IEC 60076-10-1, although reflections from buildings, other equipment, and uncontrolled background noise increase the measurement uncertainty, so on-site results are generally treated as indicative rather than conformance-grade values.

How can transformer noise be reduced?

Effective measures include using high-permeability or domain-refined core steel, amorphous cores, step-lap core joints, optimized clamping and cushioning, vibration-isolating mounting pads, stiffened or damped tank walls, and the selection of lower-speed fans or naturally cooled operating modes for light-load periods.