The solid insulation in a transformer ages, and the windings lose mechanical strength long before the fluid around them gives any obvious sign. CIGRE Technical Brochure 738 is the working group's synthesis of how that ageing happens in liquid-impregnated cellulose, and how to read it from the samples we can actually draw. It is a guidance document — the considered findings of a study group, not a normative standard — and it carries weight precisely because it gathers decades of laboratory and field evidence into one place. It is a revision of an earlier brochure, broadened to cover ester liquids alongside the long-established mineral-oil database.
What it covers
The brochure treats the degree of polymerisation as the master indicator of cellulose condition: new paper sits at a high chain length and falls through chain scission until it reaches a levelling-off value where only crystalline regions remain. Around that core it develops the Arrhenius ageing model, stresses that its activation energy and pre-factor are material- and process-specific rather than universal, and quantifies the accelerating roles of water, low-molecular-weight acids and oxygen. It compares ester and mineral-oil systems, correcting earlier over-optimistic life claims drawn from unrealistically dry test conditions. Its most operationally useful chapter catalogues the chemical markers read from oil — carbon oxides, furanic compounds and methanol — with their partition behaviour, temperature-correction approach, and the protocols for direct paper sampling and DP measurement.
Why it matters in practice
The brochure's central caution is that markers do not translate cleanly into a single life number. A general dependence between furan content and average DP is unlikely to hold across paper types, transformer designs and operating histories, and laboratory thresholds do not transfer reliably to real units. Its water findings are equally consequential: esters dissolve far more water than mineral oil, so the partition curves and moisture limits built for mineral-oil systems cannot be lifted across to ester-filled units. The working group's view that esters reduce cellulose ageing rate meaningfully, but not by the extravagant factors once claimed, is the realistic baseline we carry into life assessments.
How we use it
This brochure is our foundation for cellulose-ageing interpretation. When a client asks how much life remains in a unit, we use its kinetic framework and marker partitioning rather than a single furan reading taken at face value, and we apply its temperature corrections before comparing samples drawn at different loads. For ester-filled units it is the reference that keeps us from importing mineral-oil moisture expectations, and its post-mortem sampling guidance shapes how we advise on investigating a decommissioned transformer — where field judgment about hot-spot location and sampling region matters as much as the laboratory result.