Transformer Dielectric Dissipation Factor (Tan δ) Testing: Physical Meaning, Temperature Dependence, Pass/Fail Criteria for New and In-Service Oil, Bushing Tan δ & Moisture Correction

Abstract

The dielectric dissipation factor (DDF), universally known in the transformer industry as tan δ (tangent delta), is the most sensitive bulk indicator of insulation degradation. Unlike insulation resistance (DC) or partial discharge (localized), tan δ integrates the condition of the entire insulation volume between the measurement electrodes — the oil, the paper, the pressboard — into a single number. An elevated tan δ indicates moisture, conductive contamination, or chemical degradation (oxidation byproducts) somewhere in the insulation system. This article explains the physics of dielectric loss in transformer insulation, the mandatory temperature correction to a reference temperature (typically 20 °C) without which tan δ comparisons are meaningless, the pass/fail criteria for factory, commissioning, and in-service measurements per IEC 60247 and IEEE C57.12.90, the special case of bushing tan δ (C1 and C2 measurements using the test tap), and how to interpret tan δ trends over the transformer's life.

1. Physical Meaning of Tan δ

1.1 The Loss Angle

A perfect capacitor, when an AC voltage is applied, draws a current that leads the voltage by exactly 90° (purely capacitive — no real power dissipation). A real insulation system — transformer oil + paper + pressboard — is not a perfect capacitor: it has a resistive component (due to ionic conduction in the oil, moisture in the paper, and polarization losses) that causes the total current to lead the voltage by slightly less than 90°.

The phase angle by which the current falls short of the ideal 90° lead is the loss angle δ:

I_total = I_capacitive + I_resistive

tan δ = I_resistive / I_capacitive = (Real Power Dissipated) / (Reactive Power Stored)

For mineral oil-paper insulation, tan δ is typically 0.001-0.010 (0.1-1.0%) for new insulation, increasing to 0.01-0.10 (1-10%) as the insulation ages and accumulates moisture and oxidation byproducts.

1.2 The Equivalent Circuit

At power frequency (50/60 Hz), the transformer winding insulation can be modeled as a capacitor C with a parallel resistance R_p:

tan δ = 1 / (2πf × R_p × C)

Where:

• C = geometric capacitance of the insulation (determined by winding geometry — essentially constant over the transformer's life)
• R_p = parallel resistance representing the dielectric losses (decreases as the insulation degrades — lower R_p → higher tan δ)
• f = test frequency (50/60 Hz, or 0.1 Hz for VLF testing — but the physics is different at very low frequencies)

Practical interpretation: As moisture enters the insulation or acids form from oil oxidation, the ionic conductivity of the insulation increases → R_p decreases → tan δ increases. The capacitance C remains essentially unchanged (the dielectric constant of cellulose paper is 3.5-4.5 and does not change significantly with moisture until extreme wetness).

1.3 Relationship with Insulation Resistance (DC)

Tan δ (AC) and insulation resistance (DC) measure different aspects of the same insulation property:

• DC insulation resistance: Measures the steady-state leakage current after the absorption current has decayed. Dominated by the DC resistivity of the bulk insulation.
• AC tan δ: Measures the total dielectric loss at power frequency, including polarization (dipole orientation) losses that are not present in a DC measurement. Tan δ is more sensitive to polar contaminants (water, acids) than DC IR.

A transformer can have acceptable DC IR but elevated AC tan δ — indicating polar contamination that does not significantly affect the DC leakage path but contributes to AC losses.

2. Temperature Dependence and Correction

2.1 The Exponential Temperature Dependence

Tan δ increases exponentially with temperature because:

• The ionic conductivity of the oil and paper increases (more mobile ions)
• The polarization relaxation time of polar molecules (water, acids) decreases — more molecules can follow the 50/60 Hz field
• Moisture partitioning shifts — at higher temperatures, water moves from the paper to the oil, increasing the oil's conductivity

Temperature correction formula (IEC 60247 / IEEE C57.12.90):

tan δ(20°C) = tan δ(T) × 1.5^((20 - T) / 10)

Where T is the insulation temperature in °C at the time of measurement. The factor 1.5 means tan δ approximately doubles for every 10 °C increase (or halves for every 10 °C decrease).

Alternative correction factors:

Temperature (°C)	Correction Factor K (to 20 °C)	Example: Measured tan δ × K = tan δ at 20°C
0	4.5	0.002 × 4.5 = 0.0090
10	2.0	0.004 × 2.0 = 0.0080
20	1.0	0.010 × 1.0 = 0.0100
30	0.5	0.020 × 0.5 = 0.0100
40	0.25	0.035 × 0.25 = 0.0088
50	0.12	0.060 × 0.12 = 0.0072
60	0.06	0.100 × 0.06 = 0.0060
70	0.03	0.150 × 0.03 = 0.0045

Critical warning: The temperature correction is an approximation. At temperatures above 60 °C, the correction becomes unreliable because moisture partitioning between paper and oil changes rapidly, invalidating the assumed correction factor. Tan δ measurements should be performed when the insulation temperature is between 10 °C and 40 °C for reliable temperature correction.

3. Pass/Fail Criteria

3.1 New Oil (IEC 60296 / IEEE C57.106)

Parameter	New Oil Limit at 90 °C	Test Method
Tan δ for new uninhibited mineral oil	≤0.005 (0.5%)	IEC 60247
Tan δ for new inhibited mineral oil	≤0.010 (1.0%)	IEC 60247

3.2 In-Service Oil (IEC 60422, Measured at 90 °C)

Condition	Tan δ at 90 °C	Action
Good	≤0.10 (10%)	Routine monitoring
Fair	0.10-0.20 (10-20%)	Increased monitoring frequency; consider reclamation
Poor	>0.20 (20%)	Oil reclamation or replacement; investigate

3.3 Transformer Winding Insulation (Factory and Field, Corrected to 20°C)

Condition	Tan δ	Interpretation
New, dry	≤0.005 (0.5%)	Factory acceptance — normal for modern manufacturing
New, borderline	0.005-0.010 (0.5-1.0%)	Acceptable but indicates slightly elevated moisture or minor contamination
In-service, good	≤0.010 (1.0%)	Service-aged insulation within normal range
In-service, investigate	0.010-0.020 (1.0-2.0%)	Increased moisture or polar contamination — correlate with oil moisture, acidity, IFT
In-service, action	0.020-0.050 (2.0-5.0%)	Significant moisture/contamination — perform detailed diagnosis, consider drying/reclamation
In-service, critical	>0.050 (5.0%)	Severe insulation degradation — risk of thermal runaway if tan δ continues to increase (positive feedback: higher tan δ → more dielectric heating → higher temperature → higher tan δ → thermal runaway)

3.4 Tan δ "Tip-Up" — Voltage Dependence

Tan δ is measured at a specific test voltage (typically 10 kV for distribution class, up to the line-to-ground voltage for power transformers). An increase in tan δ as the test voltage increases — "tip-up" — indicates:

• Partial discharge within the insulation: The PD activity adds a resistive current component at higher voltages
• Ionization in voids: Air or gas-filled cavities within the paper or at the paper-oil interface ionize at higher voltage gradients

Acceptance criterion: The increase in tan δ from 0.5 U₀ to 1.0 U₀ should be ≤0.001 (0.1%). A larger increase indicates incipient PD or void ionization — investigate with a dedicated PD measurement.

4. Bushing Tan δ — C1 and C2 Measurements

4.1 Capacitance-Graded Bushing Construction

Capacitance-graded bushings (OIP, RIP, RIS) are constructed with concentric conducting foils embedded in the insulation body, arranged to grade the electric field radially. The test tap (also called the capacitance tap or voltage tap) is connected to the outermost foil.

C1 (Main Insulation): Capacitance between the HV conductor and the test tap. Represents the main dielectric body of the bushing. C1 is measured with the test tap connected to the measuring instrument and the bushing flange grounded.

C2 (Tap Insulation): Capacitance between the test tap and the grounded flange. Represents the outer insulation layer. C2 is measured with the test tap connected to the measuring instrument and the HV conductor grounded.

4.2 Bushing Tan δ Acceptance Criteria

Bushing Type	New C1 Tan δ	In-Service C1 Tan δ Limit	Action if Exceeded
OIP (Oil-Impregnated Paper)	≤0.004 (0.4%)	≤0.007 (0.7%)	Investigate — may indicate moisture ingress through degraded gaskets
RIP (Resin-Impregnated Paper)	≤0.004 (0.4%)	≤0.007 (0.7%)	Investigate; RIP bushings are sealed — tan δ increase indicates manufacturing defect
RIS (Resin-Impregnated Synthetic)	≤0.005 (0.5%)	≤0.008 (0.8%)	Investigate

C2 tan δ is less critical but should be ≤0.02 (2%) for in-service bushings. A high C2 tan δ indicates contamination or moisture ingress at the test tap insulation — often due to a loose or contaminated test tap cover, not an internal bushing problem. Clean the test tap assembly and re-test before concluding a bushing defect.

4.3 Trending — The Most Important Diagnostic

A single tan δ value is less informative than the trend. A bushing with stable C1 tan δ = 0.005 (0.5%) for 10 years is a low-risk asset. The same bushing with C1 tan δ increasing from 0.003 to 0.006 over 2 years is a developing problem, even though the absolute value is still below the 0.007 limit. Tan δ trending — not the absolute value — drives the maintenance decision.

FAQ

Q: What is the relationship between tan δ and power factor (cos φ)?

In transformer insulation testing, tan δ and power factor are related but measured differently: Power Factor = cos φ = P / (VI), where P is the real power dissipated. Tan δ = W / VAR, where W is the real power and VAR is the reactive power. For small δ (which is the case for transformer insulation — δ < 5°), tan δ ≈ sin δ ≈ δ (in radians), and power factor ≈ tan δ. Numerically, for tan δ = 0.01 (1%), power factor ≈ 1.0% — essentially the same. The distinction matters only for very lossy insulation (tan δ > 0.10) where the approximation diverges. In the transformer industry, "tan δ" and "power factor" are used interchangeably, but instruments calibrated for power factor measurement (e.g., Doble M4000) technically measure power factor and report tan δ as the calculated equivalent.

Q: At what temperature should I measure tan δ?

The measurement should be performed with the transformer insulation as close to 20 °C as possible. This is the reference temperature for factory data comparison and industry criteria. If the transformer cannot be cooled to 20 °C (e.g., testing a transformer just taken out of service at 60 °C), measure at the actual temperature and apply the temperature correction. However, the correction becomes unreliable above 40 °C — if the transformer is hot, allow it to cool for 4-8 hours (with fans off, conservator isolated if possible) before testing. The time investment in temperature stabilization is repaid in measurement accuracy and comparability.

Q: Why does new oil have a tan δ specification measured at 90 °C?

The oil tan δ is measured at 90 °C (per IEC 60247) rather than 20 °C because at room temperature, even severely degraded oil may show acceptably low tan δ — the polar contaminants (water, acids, oxidation byproducts) are relatively immobile at low temperatures and do not contribute to the AC losses. At 90 °C, the mobility of ions and polar molecules increases, and the degradation is revealed. This is analogous to the reason oil acidity is measured at elevated temperature — the test exposes degradation that is masked at ambient temperature. The 90 °C measurement also correlates with the oil's performance at the transformer's operating temperature (60-90 °C top oil).

Q: What does it mean if the tan δ varies with test voltage (tip-up)?

Tip-up — an increase in tan δ as the test voltage increases — is the classic signature of partial discharge activity in the insulation. At low voltage (e.g., 1-2 kV), the insulation behaves as a linear dielectric. As voltage increases, gas-filled voids within the paper, at the paper-oil interface, or in the oil itself begin to ionize. Each ionization event dissipates a small amount of energy — this energy dissipation appears as an increase in the resistive component of the current, i.e., an increase in tan δ. The voltage at which the tip-up begins indicates the PD inception voltage (PDIV). A tan δ tip-up of >0.001 (0.1%) between 0.5 U₀ and U₀ is a criterion for further investigation by dedicated PD measurement (acoustic or UHF). Tip-up is particularly important for new transformers during FAT — it can detect manufacturing defects (incomplete impregnation, voids in the solid insulation) that would cause premature failure in service.

Q: Can I measure tan δ of the entire transformer at once (all windings together)?

The transformer insulation system is tested as discrete sections to localize any defect: (1) HV winding to (LV + Ground) — measures the HV-to-ground and HV-to-LV insulation, (2) LV winding to (HV + Ground) — measures the LV-to-ground and LV-to-HV insulation, and (3) HV to LV (both floating from ground) — measures the inter-winding insulation. Testing "all together" would combine the three insulation paths in parallel and produce a single tan δ value that is the weighted average of all paths — a defect in any one path would be masked by the parallel good paths. Each insulation section has a different geometry, volume, and oil-to-paper ratio, so they can degrade at different rates. The sectional testing is essential for diagnostic localization.

Q: How often should tan δ testing be performed on in-service transformers?

There is no absolute standard interval. Recommended practice: (1) At commissioning — baseline measurement for all windings and bushings, (2) Every 3-5 years for critical power transformers (GSU, transmission substation), annually after 25 years of service, (3) Whenever a significant change is detected in oil DGA, moisture, acidity, or BDV — these changes may indicate developing insulation degradation that tan δ will confirm, and (4) After any major maintenance event (oil reclamation, bushing replacement, internal inspection). The interval should be risk-based — a transformer with stable tan δ over 10+ years requires less frequent testing; a transformer with an increasing trend requires more frequent monitoring.

References / Standards

Reference	Title
IEC 60247:2004	Insulating liquids — Measurement of relative permittivity, dielectric dissipation factor (tan δ) and d.c. resistivity
IEC 60076-1:2011	Power transformers — Part 1: General
IEC 60137:2017	Insulated bushings for alternating voltages above 1000 V
IEEE C57.12.90-2021	IEEE Standard Test Code for Liquid-Immersed Distribution, Power, and Regulating Transformers
IEEE C57.19.01-2017	IEEE Standard for Performance Characteristics and Dimensions for Outdoor Apparatus Bushings
CIGRE TB 445	Guide for Transformer Maintenance

*Authored by Du Fu, Production Engineer at ZY POWER. Tan δ is a bulk insulation quality metric — it reports the average condition of the entire insulation volume and cannot localize a defect. Elevated tan δ must be correlated with other diagnostics (DGA, oil quality, bushing C1/C2, PD) before concluding the cause and making a maintenance decision.*