Transformer Paper Insulation: Kraft vs. Thermally Upgraded Paper (TUP), Aging Byproducts, Moisture Impact & End-of-Life Indicators

Abstract

The solid insulation system of a power transformer — electrical-grade cellulose paper and pressboard — is the component that ultimately determines the transformer's service life. Unlike the copper conductors and silicon steel core, which do not degrade under normal operating conditions, the cellulose polymer chains undergo progressive scission through hydrolysis, oxidation, and pyrolysis. The choice between standard kraft paper (IEC 60554-3-1) and thermally upgraded paper (TUP, IEC 60554-3-5, typically kraft that has been chemically modified with nitrogen-containing compounds such as dicyandiamide, melamine, or polyacrylamide) affects not only the initial cost and thermal class but also the aging trajectory of the transformer over 30-50 years. This article explains the chemistry of cellulose degradation, the quantitative relationship between moisture content and aging rate, the aging byproducts that serve as diagnostic markers (furfural, carbon monoxide, carbon dioxide, and the CO₂/CO ratio), and how to interpret these markers to make informed end-of-life decisions.

1. Kraft Paper — The Baseline Material

1.1 Composition and Manufacturing

Electrical-grade kraft paper is manufactured from high-alpha-cellulose wood pulp (typically softwood — pine, spruce) through the sulfate (kraft) pulping process. The cellulose polymer is a linear chain of β-D-glucose units linked by β-1,4 glycosidic bonds, with a typical degree of polymerization (DP) of 1,000-1,200 for new pressboard and 900-1,100 for new paper (paper experiences more mechanical working during manufacturing).

Key properties of new kraft paper:

Property	Value	Test Method
Degree of polymerization (DP)	900-1,200 (paper), 1,000-1,200 (pressboard)	IEC 60450
Tensile strength (machine direction)	80-120 MPa	ISO 1924-2
Tensile strength (cross direction)	40-60 MPa	ISO 1924-2
Moisture content at manufacture	0.5-1.0%	IEC 60814
Thermal class (IEC 60076-2)	105 °C (Class A)	IEC 60216
Thickness (conductor paper)	50-125 μm	-
Thickness (pressboard)	0.5-8 mm	-

1.2 Aging Mechanism

Kraft paper ages through three simultaneous chemical mechanisms:

Hydrolysis (dominant mechanism): Water molecules attack the glycosidic bonds, cleaving the polymer chain:

Rate driven by: Moisture content in the paper (primary), temperature (secondary), and acid concentration (catalytic)
Water is consumed but also generated — approximately 3 water molecules are released per broken glucose unit (auto-catalytic)
At 2% moisture and 98 °C, the hydrolysis rate is approximately 4× the rate at 0.5% moisture

Oxidation: Molecular oxygen dissolved in the oil attacks the cellulose, generating carbonyl (C=O) and carboxyl (-COOH) groups:

These groups are acidic, accelerating hydrolysis (the acid-catalyzed mechanism)
Oxidation is promoted by: high dissolved oxygen in oil (poor sealing, air-breathing conservator), copper ions catalyzing the oxidation reaction, and high temperature (>100 °C)

Pyrolysis (thermal decomposition): At temperatures >140 °C, cellulose thermally decomposes independent of moisture and oxygen:

Produces CO, CO₂, water, and furan compounds
Activated only under severe overheating (hot-spot >140 °C, blocked cooling duct, core hotspot)
Not significant under normal operating conditions

2. Thermally Upgraded Paper (TUP)

2.1 Chemical Modification

TUP is standard kraft paper that has been impregnated or chemically reacted with nitrogen-containing stabilizing compounds. The three common stabilizers:

Dicyandiamide (DICY, most common): Adds amine groups (-NH₂) that neutralize the organic acids generated by cellulose oxidation, suppressing acid-catalyzed hydrolysis. The amine groups also react with water to form ammonium hydroxide, consuming water and reducing the hydrolysis rate.
Melamine: Similar mechanism — amine groups neutralize acids. Melamine-modified paper has slightly higher thermal stability.
Polyacrylamide: Amide groups provide acid-scavenging capability.

The amine-modified paper retains the stabilizer within the cellulose fiber structure — it is not leached into the oil over time, providing a durable "acid buffer" for the life of the insulation.

2.2 Performance Comparison: Kraft vs. TUP

Parameter	Standard Kraft	TUP (DICY-modified)	Improvement Factor
Thermal class (IEC 60216)	105 °C (Class A)	120 °C (Class E)	+15 °C
Aging rate at 110 °C (relative)	4.0×	2.0×	2× slower
Tensile strength retention after 500 h at 130 °C (sealed tube, 2% moisture)	30-40%	60-70%	~2× better
Furfural generation rate (relative)	1.0	0.4-0.6	Lower diagnostic value
Cost premium over kraft	-	10-20% on paper cost (2-4% on total transformer cost)	-

The practical consequence: A transformer built with TUP instead of standard kraft at the same hot-spot temperature will have approximately double the expected insulation life. Alternatively, for the same design life, a TUP transformer can be operated at approximately 8-10 °C higher hot-spot temperature — enabling a smaller, lighter design for the same rating or a higher overload capability.

2.3 Important Diagnostic Caveat

TUP generates 40-60% less furfural per unit of DP loss compared to standard kraft — because the amine stabilizer suppresses the acid-catalyzed pathways that produce furans. This means:

Furfural-based DP estimation formulas are NOT valid for TUP unless specifically calibrated for the TUP type
A TUP transformer with 2-FAL = 0.5 mg/kg may have significantly lower actual DP (closer to end-of-life) than the same furfural concentration in a kraft transformer
For TUP transformers, rely more heavily on CO/CO₂ trends, moisture in paper, and mechanical strength indicators rather than furfural alone

3. Aging Byproducts as Diagnostic Markers

3.1 Furfural (2-FAL) — The Leading Chemical Marker

Generation mechanism: During the hydrolysis of cellulose, the glucose rings are partially dehydrated to form furan compounds. 2-furaldehyde (2-FAL) is the most abundant furan, accounting for 80-90% of the total.

Diagnostic value per IEC 61198 and IEC 60422:

2-FAL (mg/kg oil)	Estimated DP (Kraft)	Condition	Action
<0.1	>600	Minor aging	Routine DGA + furfural every 2-3 years
0.1-0.5	400-600	Moderate aging	Increased monitoring (annual)
0.5-2.0	250-400	Significant aging	Intensive monitoring (6-monthly); plan for end-of-life
2.0-5.0	150-250	Advanced aging	End-of-life within 3-5 years; do not subject to through-faults
>5.0	<150	Critical	Replace or re-insulate; risk of in-service failure

3.2 Carbon Monoxide (CO) and Carbon Dioxide (CO₂)

These gases are monitored through DGA. Their generation rates indicate total paper decomposition:

CO generation pathways:

Thermal decomposition of cellulose at >150 °C (pyrolysis)
Partial oxidation of cellulose
Breakdown of oil under thermal stress (oil can also generate CO, particularly with oxidation)

CO₂ generation pathways:

Low-temperature oxidation of cellulose (<150 °C) — the dominant CO₂ source in normally aging transformers
Complete oxidation of cellulose

CO₂/CO Ratio Interpretation:

CO₂/CO Ratio	Interpretation
>10	Normal aging at moderate temperature (<120 °C). CO₂ dominates.
7-10	Mildly elevated aging temperature (120-140 °C)
3-7	Elevated hotspot (>140 °C) — significant CO production from pyrolysis. Investigate.
<3	Severe localized hotspot (>200 °C). Paper carbonization is occurring. Immediate investigation required.

Importance of trending: The CO₂/CO ratio is most diagnostic when it changes significantly from a stable baseline. A transformer that has operated for 10 years with CO₂/CO = 8 ± 1 that suddenly drops to 4-5 has developed a new hotspot. The absolute ratio varies between transformers depending on the oil type, the presence of oxidation inhibitor, and the oxygen content.

4. Moisture: The Accelerator of Cellulose Aging

4.1 Moisture Equilibrium Between Paper and Oil

Moisture partitions between the cellulose insulation (paper, pressboard) and the oil according to temperature-dependent equilibrium curves. At operating temperature (60-80 °C), the paper holds 10-50× more water than the oil on a ppm (mg/kg) basis.

Typical equilibrium: At 60 °C, paper at 2% moisture is in equilibrium with oil at approximately 15-20 ppm water. At 20 °C (cold transformer), the same paper releases water to the oil, raising the oil moisture to 30-40 ppm.

Measuring moisture in paper: Oil moisture is measured by Karl Fischer titration (IEC 60814). The paper moisture is then calculated from the equilibrium curve (Oommen, Fabre-Pichon, or MIT models) using the oil moisture and the oil temperature at the time of sampling. This calculation is accurate only when the transformer has been at stable temperature for several days — do not sample during or immediately after a large load change.

4.2 Impact of Moisture on Aging Rate

Paper Moisture (% wt)	Relative Aging Rate (at 98 °C)
0.5%	0.2× (below baseline)
1.0%	0.5×
1.5%	1.0× (IEC 60076-7 baseline)
2.0%	2.0×
3.0%	5.0×
4.0%	10.0×

Practical consequence: A transformer operating with 3% moisture in the paper instead of 1.5% ages at 5× the normal rate — its design life of 30 years is consumed in approximately 6 years. Moisture control (functional breather, intact conservator bladder, timely oil drying) is the single most important life-extension measure for older transformers.

FAQ

Q: Is TUP always better than kraft paper? Should I always specify TUP?

For new power transformers ≥10 MVA: Yes, specify TUP. The cost premium (typically 2-4% on the total transformer price) is recovered through: (1) longer life or higher overload capability, (2) reduced risk of premature end-of-life — the TUP "buffer" against occasional high-temperature excursions can prevent a forced outage, and (3) most utilities now specify TUP as standard — requesting kraft to save cost is a false economy. For distribution transformers (<2,500 kVA) where the economic model is run-to-failure with a 25-30 year expected life, kraft is acceptable and standard. The decision should be based on the expected life and criticality, not on the material cost difference.

Q: Can furfural be removed by oil treatment?

Yes. Vacuum degasification strips a portion of the furfural from the oil (more volatile than heavy hydrocarbons, less than dissolved gases). Clay treatment (Fuller's earth, activated alumina) during oil regeneration removes nearly all furfural from the oil. This creates a dangerous situation: the oil now shows low furfural, suggesting healthy paper, but the paper DP is unchanged (still degraded). This is why furfural trending is essential — if the furfural drops from 2.0 to 0.1 mg/kg after oil treatment, it signals that treatment was performed, not that the paper recovered. Always document oil treatment events in the transformer's history record and correlate furfural values with the treatment timeline.

Q: How is the moisture content of paper determined without taking a physical sample?

Draw an oil sample for Karl Fischer titration at a time when the transformer has been at stable temperature for 24+ hours. Record the oil temperature at the time of sampling. Use the equilibrium curves (Oommen model, or the modified curves from IEC 60422 Annex B) to calculate the paper moisture from the oil moisture. For example, at 60 °C oil temperature: if the oil moisture is 20 ppm, the Oommen curve gives approximately 2.2% moisture in paper. The accuracy of this method is ±0.5-1.0% absolute — sufficient for aging assessment but not for precision diagnostics. For critical transformers where paper moisture is a concern, online moisture sensors (capacitive thin-film sensors in the oil or at the tank wall) provide continuous trending.

Q: What's the relationship between the CO₂/CO ratio and the paper insulation condition?

The CO₂/CO ratio is an indicator of the temperature of paper decomposition, not the total quantity of decomposition. A high ratio (>10) means low-temperature oxidation is dominant — normal for a transformer operating at design temperatures. A low ratio (<3) means CO production is disproportionately high, indicating pyrolysis at >300 °C — either a localized hot-spot (partial discharge at a winding conductor, core bolt heating, circulating current in the tank) or carbon tracking. The absolute quantity of CO + CO₂ increases with total paper decomposition but is less reliable as a DP indicator than furfural. The CO₂/CO ratio is most useful for detecting thermal faults in their early stages — before furfural accumulates to alarming levels.

Q: Does the transformer oil type affect paper aging?

Yes, significantly. Mineral oil has a moderate oxygen solubility (approximately 3,500 ppm at 25 °C and atmospheric pressure), which feeds oxidative paper degradation. Natural ester fluids (vegetable oil-based, e.g., FR3) and synthetic esters have significantly higher water saturation limits (1,000-2,000 ppm vs. 50-60 ppm for mineral oil at 20 °C). This means ester fluids can hold more water in the liquid phase, keeping the paper drier — paper moisture in an ester-filled transformer is typically 0.5-1.0% lower than in an equivalent mineral oil transformer. This reduced paper moisture translates to 2-4× slower aging. However, esters also have higher viscosity, affecting heat transfer, and cost 3-5× more than mineral oil — the life-extension benefit must be weighed against the cost.

Q: At what DP value is the transformer paper insulation at end-of-life?

The consensus end-of-life threshold is DP = 200 (CIGRE TB 323, IEC 60422). At this point, the paper retains only 15-25% of its original tensile strength and cannot reliably withstand the mechanical forces of a through-fault (short-circuit). However, DP = 200 is not a cliff — paper does not fail structurally at DP = 200 and remain functional at DP = 210. The degradation is gradual, and the actual risk of failure depends on: (1) the maximum through-fault current the transformer may experience (a transformer connected to a stiff grid with high fault level has higher risk at DP = 250 than a transformer on a weak rural system at DP = 150), (2) the paper's mechanical condition — oxidative embrittlement (cross-linking, reduced flexibility) makes paper more fragile at a given DP than pure hydrolytic degradation, and (3) the transformer's short-circuit history — a transformer that has already experienced a near-withstand fault has less residual strength than one that has never been fault-stressed. The end-of-life decision is probabilistic, not deterministic — use DP as one input among many.

References / Standards

Reference	Title
IEC 60554-1:1977	Specification for cellulosic papers for electrical purposes — Part 1: Definitions and general requirements
IEC 60554-3-1:1979	Specification for cellulosic papers for electrical purposes — Part 3: Specifications for individual materials — Sheet 1: General purpose electrical paper
IEC 60554-3-5:1984	Specification for cellulosic papers for electrical purposes — Part 3: Specifications for individual materials — Sheet 5: Special papers
IEC 60450:2004	Measurement of the average viscometric degree of polymerization of new and aged cellulosic electrically insulating materials
IEC 60422:2013	Mineral insulating oils in electrical equipment — Supervision and maintenance guidance
IEC 61198:1993	Methods for the determination of 2-furfural and related compounds
IEC 60076-7:2018	Power transformers — Part 7: Loading guide
CIGRE TB 323	Ageing of Cellulose in Mineral-Oil Insulated Transformers
IEEE C57.91-2011	IEEE Guide for Loading Mineral-Oil-Immersed Transformers

*Authored by Du Fu, Production Engineer at ZY POWER. The paper insulation is the transformer's Achilles' heel — every transformer maintenance and life-extension program must include paper condition assessment as a primary pillar. Furfural + DGA CO/CO₂ + moisture in paper = the minimum viable diagnostic dataset.*