Transformer No-Load Test Analysis: Principles, Procedures, and Diagnostic Evaluation

The no-load test is an essential routine test for power and distribution transformers, quantifying core (iron) losses and excitation current to verify design quality, manufacturing workmanship, and energy-efficiency compliance. This guide explains the principles, applicable standards, measurement procedures, and result interpretation of transformer no-load testing, including how elevated loss or current readings reveal core defects. Engineers, quality inspectors, and asset managers can use this analysis to support factory acceptance, commissioning, and long-term condition assessment.

Technical Overview

The no-load test (also called the open-circuit or excitation test) is a routine test defined in IEC 60076-1 and its national adoptions such as GB/T 1094.1 and EN 60076-1. Rated voltage at rated frequency is applied to one winding while all other windings are left open-circuited, and the input power (no-load loss, P0) together with the input current (no-load current, I0) are measured. Because no load current flows in the secondary circuit, the measured input power represents almost entirely the losses generated in the magnetic core, with only a negligible resistive component in the energized winding.

No-load loss is composed of hysteresis loss and eddy-current loss in the core laminations, supplemented by stray losses in clamping plates, tie rods, and tank walls. The no-load current consists of a reactive magnetizing component and a small active component that supplies these losses; for modern units it typically ranges from about 0.1% to 3% of rated current depending on rating, flux density, and core material. Both parameters are highly sensitive to core material grade, lamination insulation quality, joint design, and stacking workmanship, which makes the test a powerful screening tool for core condition.

Since no-load loss is present whenever the transformer is energized (8,760 hours per year), it dominates the lifetime energy consumption of lightly loaded units and is regulated by energy-efficiency standards such as IEC 60076-20 and GB 20052. Accurate measurement and analysis of P0 and I0 therefore serve both reliability and economic objectives.

Testing Samples

Oil-Immersed Distribution Transformers

Units in the 10 kV and 35 kV voltage classes with typical ratings from 30 kVA to 2,500 kVA are subjected to the no-load test as a routine factory test and again during site acceptance. Reference no-load losses generally fall between 0.1% and 0.3% of rated power for conventional grain-oriented silicon steel cores.

Large Power Transformers

Generator step-up and interconnecting transformers rated 110 kV, 220 kV, 330 kV, and 500 kV and above are tested in the factory and retested on site after transport and assembly, because the heavy core structure is vulnerable to displacement and clamping relaxation during shipment.

Dry-Type, Cast-Resin, and Amorphous-Core Transformers

Dry-type units are tested with the same principles, with attention to their open magnetic structures. Amorphous-alloy core transformers exhibit substantially lower no-load loss (roughly one quarter to one third of conventional silicon steel designs) but distinct magnetization behavior, so reference expectations and harmonic signatures must be evaluated against material-specific baselines.

Repaired, Aged, or Post-Event Units

After winding replacement, core restacking, through-fault exposure, overexcitation incidents, or transport shock, the no-load test is repeated and trended against original factory values to confirm that the magnetic circuit remains sound.

Testing Items

The following quantities are determined and analyzed during a complete no-load test program:

• No-load loss P0 at rated voltage and frequency, corrected for waveform distortion.
• No-load current I0, expressed in amperes and as a percentage of rated current, recorded per phase and as an average value.
• Excitation (voltage-current) characteristic from approximately 40% up to 110% to 115% of rated voltage, identifying the saturation knee.
• Harmonic spectrum of the excitation current, predominantly the 3rd and 5th harmonics.
• Three-phase unbalance of excitation currents, interpreted against the known asymmetry pattern of three-limb cores.
• Compliance verification against guaranteed values and the tolerance limits of the applicable standard.

Each item carries diagnostic weight. The loss value validates material quality and workmanship, the current value reflects magnetic circuit reluctance and air gaps, the excitation curve exposes saturation margin, and the harmonic and unbalance data help localize anomalies to individual limbs or joints.

Testing Methods

Test Circuit and Instrumentation

The low-voltage winding is normally energized at rated voltage while the high-voltage winding is left open; rated flux is induced regardless of which side is excited, but LV excitation permits safer, lower-voltage connections. A regulated, low-distortion supply such as a dedicated generator or filtered variac is used. The measurement chain comprises voltage and current instrument transformers of accuracy class 0.1 to 0.2, a power analyzer accurate at very low power factor (excitation power factor is typically 0.1 to 0.5), and both an average-responding voltmeter and a true-RMS voltmeter for waveform assessment. Three-phase units are measured with the three-wattmeter method or the two-wattmeter (Aron) method, while single-phase units require only one wattmeter circuit.

Standard Test Procedure

• Disconnect all external attachments such as surge arresters, cables, bus ducts, and voltage transformers so that only transformer losses are measured.
• Energize gradually, raise the voltage to the rated value, and allow several minutes for readings to stabilize as the core reaches steady magnetization.
• Adjust the voltage using the average-responding voltmeter as the reference, and record the simultaneous RMS reading for waveform correction.
• Record total power, per-phase currents, voltages, and frequency, then compute P0 and the I0 percentage.
• Where required, capture additional points at approximately 80%, 90%, 100%, 105%, and 110% of rated voltage to trace the excitation curve.
• Demagnetize the core at the end of the test by smoothly reducing the applied voltage to zero.

Waveform, Frequency, and Voltage Corrections

Because the core is a nonlinear element, supply voltage distortion changes the measured loss. The IEC 60076-1 and IEC 60076-8 methodology bases the voltage setting on the average-responding (mean-value) voltmeter and applies a correction factor derived from the deviation between the mean-scaled and RMS indications, accounting for the voltage exponent of the core loss. If the test frequency differs from rated, the voltage must be adjusted to keep the volts-per-hertz ratio constant so that rated flux density is maintained, with appropriate frequency corrections then applied. Unlike the load loss test, no temperature correction is required, since magnetic losses are essentially temperature independent; only the tiny winding I2R component of the excitation current depends on temperature and may be subtracted in high-precision work.

Supplementary Single-Phase Test for Defect Localization

When three-phase results are suspicious, a single-phase diagnostic test is performed by applying voltage across winding pairs (U-V, V-W, and U-W), typically on a delta-connected winding. In a healthy three-limb core, the two readings involving symmetric magnetic paths should agree closely while the third follows its characteristic relationship. An individual reading that deviates markedly from this pattern localizes the defect to a specific limb or joint region.

Result Interpretation and Fault Diagnosis

• Both P0 and I0 significantly high: interlaminar insulation failure, shorted core laminations, burrs at sheared edges, failed insulation of core tie bolts, or accidental two-point grounding of the core.
• P0 moderately elevated with nearly normal I0: circulating currents in clamping structures or magnetic shields, localized coating degradation, or non-uniform joint clamping pressure.
• I0 sharply elevated with only slight loss increase: enlarged air gaps at core joints, missing or misaligned laminations, loose core clamping, or lower-grade magnetic material.
• Phase-current unbalance exceeding the normal pattern for a three-limb core: limb-specific joint defects, foreign magnetic debris, or displaced laminations.
• Growth of current harmonics or a shift of the saturation knee: altered core material condition, previous overexcitation, or residual magnetization requiring demagnetization before retest.

Results are compared with the factory test certificate and the guaranteed values. Under IEC 60076-1, the tolerance for no-load loss is +15%, for no-load current +30%, and for total losses +10%; readings beyond these limits trigger engineering investigation before acceptance.

Safety Precautions

Although the test power is small, the open-circuited high-voltage terminals develop full rated voltage by induction and must be treated as live, with clear barricading. The supply must be de-energized and grounded before any connection changes, and the integrity of core and tank grounding must be verified prior to energization.

Applications

• Routine and type testing of new transformers in manufacturing plants in accordance with IEC 60076-1 and GB/T 1094.1.
• Site commissioning and handover acceptance, comparing measured values with factory certificates.
• Energy-efficiency grading and regulatory compliance verification under standards such as IEC 60076-20 and GB 20052.
• Loss capitalization and total-cost-of-ownership evaluation during procurement of distribution and power transformers.
• Condition assessment after transport, seismic events, through-faults, or overexcitation incidents.
• Verification of core repairs, restacking, or rewinding work in repair workshops.
• Research and development benchmarking of core materials, including hi-B and domain-refined silicon steel, amorphous alloys, and step-lap joint technologies.

Witnessed testing by accredited third-party laboratories is frequently specified for high-value units and energy-efficiency certification, providing independent confirmation of no-load performance.

FAQ

Why is the low-voltage winding normally energized during the no-load test?

Rated flux density is achieved regardless of which winding is excited, but energizing the LV side allows lower test voltages and simpler, safer connections. The open HV terminals nevertheless develop rated voltage and must be treated accordingly.

What tolerances apply to no-load results?

IEC 60076-1 permits no-load loss to exceed the guaranteed value by no more than +15%, no-load current by +30%, and total losses by +10%. Corresponding national standards specify comparable limits, and contracts may impose tighter requirements.

Does ambient or winding temperature affect the no-load test?

No-load loss is dominated by magnetic hysteresis and eddy-current effects, which are essentially temperature independent, so no correction is applied. This contrasts with the load loss test, where temperature correction is mandatory.

Why can the no-load current rise after a DC winding resistance measurement?

Direct current leaves residual magnetization in the core, which biases the magnetic operating point and increases apparent excitation current asymmetry. The core should be demagnetized by slowly reducing an AC voltage to zero before the formal no-load measurement.

Can the no-load test detect winding inter-turn faults?

A complete inter-turn short causes a dramatic rise in excitation current and abnormal noise and will be caught immediately. However, incipient inter-turn insulation deterioration is better revealed by frequency response analysis, impedance comparison, or partial discharge testing, with the no-load test serving as an effective screening step.

What supply capacity is needed for the test?

Only a small capacity is required, since the supply delivers merely the excitation current, typically below 3% of rated current, at rated voltage. The practical constraint is voltage waveform quality rather than power, so a clean, low-distortion source is essential for accurate loss measurement.