Can we accurately model transformer oil degradation with mathematical equations?

[vsjas]

As the major contribution to degradation is due to the oxidation of oil which adds more polar constituents to the oil and thus increase the effective dielectric constant of the mixture.Can we assess the oil degradation just by looking at the effective dielectric constant changes over time?
TIA

 

[jim hardy]

Shortcuts seldom work out well.

Before recommending that to my company i'd want to become very skilled at testing transformer oil.

https://constellation.uqac.ca/3036/1/FarzanehDehkordi_uqac_0862N_10090.pdfhttps://www.sdmyers.com/transformer...tical-testing/test-package-library/#test-1527

Liquid power factor is a particularly useful in-service tool for testing and monitoring oil because the test is sensitive to moisture, oxidation of the oil, and contamination from outside sources.

old jim

 

[Windadct]

"oxidation" per se was not a problem in my experience, but in the USA, most transformers are sealed, elsewhere they may have a conservator and breather via desiccant.

In the field we did dielectric and PF (or Tan-Delta - which is a measurement of the Dielectic Constant) using a Doble tests set(on higher Voltage units). But also would sample and return to the lab to have the dielectric, PF as well as Dissolved Gas Analysis - which we did with Gas Chromatography.

So moisture was the first issue, then the various issues that cause the oil to breakdown ( Overloading, Hotspots, tracking, corona) -- etc... it is pretty interesting how each cause https://www.netaworld.org/sites/default/files/public/neta-journals/NWwtr09_Hamrick.pdf.

Of these the DGA was always the most telling.

 

[vsjas]

I am trying to mathematically model a liquid dielectric to study the effect of oxidation on oil.(In order to verify my model) Can anyone suggest me an easy technique to measure Total Polar Compounds content of an oil(Other than chromatography) and also is there a way to find out what are the oxidation products of mineral oil and their concentrations?
TIA

 

[Baluncore]

Transformer oil is usually pure alpha-methyl naphthalene. The oil may be partially burnt by arcs at poor connections, so the oil becomes cloudy with carbon particles. If water is present it forms a hydroxide, a yellow soapy scum that will block filters.

Water has a high dielectric constant only when it is a free liquid, there may be no change in dielectric constant between the good oil and hydrated oil, probably hydro-naphthalenes ? The hydrated oil usually floats to the surface rather than forming an emulsion.

You will need to identify all the contaminant species that might be present. Transformer oil must be clean to function, so the impurities that must be detected will be in very small concentrations, say below 1 part in 1000. The change in dielectric constant will be very very small. You might do better estimating contamination by measuring the scattering of light with some form of turbidity sensor.

There was an old test that plunged a red-hot iron rod into the oil. If it made a cracking noise it was wet and needed replacement. Be aware that some transformer oil is contaminated with PCB carcinogens, those remain from older oil previously used in some equipment.

 

[Windadct]

If I recall - the less known issue is that the insulation is really paper, and if there is moisture, the paper has a much higher affinity for it than the oil. Water, in bulk, or condensed out through temp swings may precipitate out to the bottom, and then hold the oil in saturation, rarely ( again in the US) are there pumps to cause mixing and generate a suspension, still only the largest- high duty service units would have pumps (like Generator step up or Arc Furnace trans), and you would not generally have access to the top surface to inspect for this.

PCBs in transformer oil in the USA has been addressed for many years, and it would be extremely unlikely to find one here ( perhaps a 30+ year old pole mount / pop-top), even retro-filled transformers were typically still replaced. Additionally the PCBs themselves are not particularly hazardous in casual exposure. However their byproducts, particularly from combustion ( dioxin) is the broader concern. Also has a particular smell. So a hazard to be aware of but considering all of the hazards associates with electrical work, this is well down the list IMO.

Still to be able to mathematically model the breakdown, it seems to be so many variables, that this would be difficult to have a practical application. So I am sure you could come up with a model of pure oil breakdown, but in reality, how does moisture affect that analysis, how does carbon or cellulose - or high concentrations of the various dissolved gasses... as a simulation of this case in isolation, I am sure best answered by a petroleum chemist.