A transformer dies when its paper dies. The oil can be cleaned, dried, even replaced — the paper can’t. So the real question in condition assessment isn’t “how’s the oil?” It’s “how much life is left in the cellulose?” And that question has a frustrating catch: you can’t measure it directly.
The number that answers it is the degree of polymerization — DP. It’s a direct measure of how intact the cellulose chains still are. But getting it means cutting a physical sample of paper out of the windings, which you can’t do to a sealed, in-service transformer. So the industry does the next best thing: it reads the breakdown products the aging paper sheds into the oil. The main one is a family of compounds called furans, and chief among them, 2-furfural.
This article is about that chain of reasoning — DP, why it matters, how furans stand in for it, and the long list of reasons the stand-in can lie to you.
Table of Contents
What DP actually measures
Cellulose is a polymer: long chains of glucose rings. The degree of polymerization is, literally, the average number of rings in a chain. Fresh Kraft paper comes in high — above roughly 1,000 to 1,200. As the paper ages, the chains break and the average falls.
When DP drops to around 200, the paper is considered finished. Not because it stops insulating — here’s the part people miss — but because it stops holding together mechanically. At low DP the paper has lost most of its tensile strength, and a transformer with brittle paper can pass every electrical test right up until a through-fault or inrush event shakes the windings and the embrittled insulation tears. The failure is mechanical before it’s electrical. That’s why DP, and anything that tracks it, sits at the centre of life assessment. DP itself is measured by a viscometric method under IEC 60450.
Why furans, and why 2-furfural
Three processes age the paper — heat (pyrolysis), water (hydrolysis), and oxygen (oxidation) — and all three break glucose rings. Among the by-products are furanic compounds, which dissolve in the oil where you can sample them.
There are five in the family: 2-furfural, 5-hydroxymethyl-2-furfural, 5-methyl-2-furfural, 2-acetylfuran, and 2-furfuryl alcohol. In practice, only 2-furfural (2-FAL) is used as the aging marker. The other four are less stable under transformer conditions — they decompose or convert, often into 2-FAL itself — so 2-FAL is the one that survives long enough to mean something. It’s measured in the oil by the method in IEC 61198.
So the logic is: paper ages → chains break → 2-furfural appears in the oil → measure it → infer DP. Clean in theory.
The furan-to-DP correlation
The link between 2-FAL concentration and DP is roughly logarithmic — a little furan corresponds to a lot of remaining DP, and the number climbs steeply as the paper degrades. Several researchers have fitted equations to it. The two most cited:
- De Pablo: DP ≈ 7100 / (8.88 + 2FAL)
- Chendong: a log-linear fit, log(2FAL) decreasing as DP rises.
There are others — Pahlavanpour, Scholnick, various lab-specific fits from a CIGRE round-robin. And here’s the first warning sign: they don’t agree. For the same 2-FAL number they hand back DP estimates that can differ by a wide margin. The fits depend on the paper type, on whether the aging was natural or lab-accelerated (lab-aged paper tends to produce more 2-FAL for the same degradation), and on the oil. Most of the equations are only meant for 2-FAL up to about 5 ppm.
The honest position, shared by most labs: a single 2-FAL reading does not give you a trustworthy absolute DP. What it gives you is a point on a trend. Watch the number climb over years and you’re watching the paper age; convert one reading to a hard DP and quote it as fact, and you’re overreaching.
The traps that make furans lie
This is the part worth tattooing on the inside of your eyelids. 2-FAL is a useful marker surrounded by ways to misread it.
Oil regeneration wipes it. This is the big one. Furans live in the oil, so when the oil is reclaimed, filtered, or changed, the 2-FAL goes with it. The number drops to near zero and slowly rebuilds — meaning a transformer with badly aged paper can show a reassuringly low 2-FAL for a year or two after an oil change. Always check the oil history before trusting the number. A low 2-FAL on recently regenerated oil tells you nothing about the paper.
It’s unstable. 2-FAL degrades at high temperature and in the presence of oxygen — a meaningful fraction can disappear over weeks in an oxidising, hot unit. So the oil’s oxygen level and the transformer’s thermal history both bend the reading.
Moisture and acid drive it. The hydrolysis that produces furans is accelerated by water and acids in the insulation — the same water you’d track with Karl Fischer moisture testing and the equilibrium moisture curves. Wet paper ages faster and sheds furans faster, which is real information, but it means 2-FAL reflects the conditions as much as the cumulative age.
There’s a manufacturing baseline. Some 2-furfural can be present from oil processing, not paper aging, so a baseline matters.
Thermally upgraded paper plays by different rules. Modern thermally upgraded (nitrogen-stabilised) cellulose ages more slowly and produces less 2-FAL per unit of degradation, so a furan-to-DP equation built on standard Kraft will read it wrong.
None of this makes furan analysis useless. It makes it a trend tool with an asterisk, not a one-shot DP meter.
The companions: CO₂/CO and methanol
Because 2-FAL has these weaknesses, it’s rarely read alone.
The CO₂/CO ratio from dissolved gas analysis is a useful cross-check. Both gases come partly from paper breakdown, and the ratio is more stable than furans and less disturbed by an oil change. Under IEC 60599 a ratio roughly between 3 and 10 is treated as normal paper aging; a ratio below about 3 points to paper involvement in a fault. It won’t give you a DP, but it corroborates — or contradicts — the furan story.
The newer marker is methanol. Its appeal is exactly where furans fail: methanol is produced only by cellulose chain scission, with no comparable instability problem, so in principle it tracks the very early breaking of chains more faithfully. It’s still maturing as a field tool and the interpretation isn’t as settled as furans, but it’s the direction the diagnostics are heading.
For the authoritative treatment of furanic diagnosis, the reference is CIGRE Technical Brochure 494, “Furanic compounds for diagnosis.”
Why it matters
Put the pieces together and the value is clear. The single most important thing to know about an old transformer is how close its paper is to mechanical collapse — and that’s invisible to every electrical test until it’s too late. DP answers it, you can’t measure DP directly, and furans (backed by CO₂/CO and increasingly methanol) are how you read it through the oil.
Treat 2-FAL as a trend, anchor it to the oil’s regeneration history, cross-check it with the gas ratios, and it’s one of the most valuable numbers in the whole oil analysis. Treat it as a direct DP readout and it will eventually embarrass you.
FAQ
What is the degree of polymerization (DP) in a transformer?
The average number of glucose rings in the cellulose chains of the paper insulation. New Kraft paper is above roughly 1,000–1,200; around 200 the paper has lost most of its mechanical strength and is considered end-of-life. DP is measured viscometrically under IEC 60450.
Why can’t DP be measured directly in service?
Because it needs a physical sample of paper from the windings, which can’t be taken from a sealed, energised transformer. Furan analysis estimates it indirectly from compounds the aging paper sheds into the oil.
What is 2-furfural and why is it the marker?
2-furfural (2-FAL) is one of five furanic compounds produced as cellulose breaks down. It’s used as the marker because it’s the most stable of the five under transformer conditions; the others decompose or convert, often into 2-FAL. It’s measured per IEC 61198.
How reliable is a furan-to-DP estimate?
As a single absolute number, not very — the published equations disagree, and the result depends on paper type and conditions. As a trend tracked over years, it’s valuable. The reading is only meaningful alongside the oil’s regeneration history.
Why did my furan number drop after an oil change?
Because furans dissolve in the oil and leave with it. Reclamation or replacement resets 2-FAL toward zero, then it slowly rebuilds. A low value on recently regenerated oil says nothing about the paper’s actual condition.
What’s the CO₂/CO ratio used for?
It’s a cross-check on paper aging from dissolved gas analysis. A ratio roughly between 3 and 10 (IEC 60599) reflects normal paper aging; below about 3 suggests paper is involved in a fault. It’s more stable than furans and less affected by an oil change.