Fleet DGA Triage: rank your transformers by DGA urgency, with the reason on every row
Fleet DGA Triage is a free Seetalabs® tool for transformer asset managers. Load a CSV of laboratory DGA results and it ranks up to five transformers per session from most to least urgent, with the IEEE C57.104-2019 status, the IEC 60599:2022 assessment, the probable fault type and the reason for each position. Your data stays in your browser.
Below the tool: "Try it first with the built-in demo fleet of 11 transformers, which is not subject to the five-unit limit."
How it works
- Export the DGA results from your laboratory or spreadsheet as CSV, one row per oil sample, or start from the template the tool provides.
- Load the file. The tool reads comma, semicolon or tab separators, decimal points or commas, and dates as YYYY-MM-DD or DD/MM/YYYY.
- Read the fleet table, ordered from most to least urgent. Each row shows the IEEE status, the IEC category, the Duval zone, the recommended action and any warning about the data, such as a single sample, a latest sample older than 18 months, a missing O2/N2 ratio or C2H2 with a communicating tap changer.
- Open a transformer to see Duval Triangles 1, 4 and 5 and Pentagons 1 and 2 with its sample plotted. To look at one transformer in more detail, including whether its Duval zone holds within laboratory uncertainty, use the DGA Calculator.
- Download the history file before you close the page. Next time, load it together with the new laboratory results: new samples are added to each transformer, and a sample with the same date is updated.
How the ranking is built
Priority follows what the standards ask you to do next, not a score the tool invents.
- High: IEEE C57.104-2019 Status 3.
- Medium: IEEE Status 2, a confirmation sample requested by IEEE, or gas levels above the IEC 60599:2022 typical values, where the action the standard gives is to sample more often.
- Low: none of the above.
- Not assessed: a fluid the tool does not cover, or a fault gas missing in the latest sample.
Natural and synthetic esters reach medium priority at most, because IEEE C57.155-2014 §6.2 asks for closer monitoring with a confirmation sample and no ester reference defines a higher level.
Within the same priority, transformers are ordered by IEEE status, then IEC category, then whether C2H2 is present, then how far the worst gas is above its limit.
Method and standards
Mineral oil. IEEE C57.104-2019 §6.1.1 Steps 1 to 6, with Tables 1 and 2 (gas levels by age and O2/N2 ratio), Table 3 (increments) and Table 4 (rates over 3 to 6 samples within 24 months). When the O2/N2 ratio is missing, the tool uses the more cautious "> 0.2" column and says so. IEC 60599:2022: typical values of Annex A (upper bound by default, lower bound selectable), rates, the gas ratios of Table 1 and the CO2/CO ratio of §5.5 as indicative. The tables were checked cell by cell against the published standards on 28 September 2026.
Fault type. Computed when the gases are significant: IEEE Status 2 or above, a confirmation requested, or IEC values above typical. Duval Triangle 1 with the 2022 arcing sub-zones in oil (-H) and paper (-P), Duval Triangles 4 and 5 as activated in CIGRE TB 771 (2019) Appendix H.1, and Duval Pentagons 1 and 2. The oil-and-paper sub-zones are indicative, and the page says so next to the result.
Esters. IEEE C57.155-2014 Table 1 and IEC 63585:2026 Annex B typical values; Duval Triangle 3 or 3b, Triangles 6 and 7 for soybean ester, Pentagons 3, 4 and 4b. A gas counts as rising only when the change exceeds the analytical uncertainty of the two samples; below 10 ppm the tool makes no judgement.
Worked example
Three transformers, one CSV
Three mineral-oil transformers of unknown age, each with one sample and a sealed conservator (O2/N2 at or below 0.2). All share the same background gases (H2 10, CH4 5, C2H6 5, CO 200, CO2 2000, O2 1000 and N2 50000 ppm) and differ in C2H4 and C2H2:
| Transformer | C2H4 (ppm) | C2H2 (ppm) | IEEE status | Priority |
|---|---|---|---|---|
| A | 300 | 5 | 3 | High |
| B | 70 | 0 | 2 | Medium |
| C | 5 | 0 | 1 | Low |
The tool ranks them A, B, C. For this age and O2/N2 column, IEEE C57.104-2019 places C2H4 between its Table 1 and Table 2 levels at 70 ppm, which gives Status 2, and above Table 2 at 150 ppm, which gives Status 3.
When the gases are significant, the tool also names the fault. A sample with H2 60, CH4 90, C2H6 30, C2H4 300 and C2H2 2 ppm falls in zone T3 of Duval Triangle 1 and of Pentagon 1, and in T3-H of Pentagon 2: a thermal fault above 700 °C in oil only. The 2 ppm of C2H2 next to gases in the hundreds does not make the point unreliable.
What the CSV needs
Required columns: transformer, date, H2, CH4, C2H6, C2H4 and C2H2, in ppm. Optional: CO, CO2, O2, N2, age or year of manufacture, whether a tap changer communicates with the main tank, the fluid type and the solid insulation. English and Italian column names are recognised, and units in the header, such as "H2 (ppm)", are ignored. A file without the required columns is rejected with a message.
An empty cell, "-" or "NA" means "not measured". A value such as "<1" keeps its limit: the tool runs each method at 0 and at the limit, and when the two give different results that method is not assigned and a warning names the gas. "ND" without a number stays unknown. The limit is kept in the history file and written back as "<1" in the CSV export.
Limits
- Five transformers per session in the free version. If a file brings more, the tool lists which units it will analyse and which it will leave out, and you choose to load the first five or cancel.
- CSV and history files only: no Excel import. No trend chart and no generated PDF report (use your browser's print function).
- Main tank, laboratory samples. Silicone and unknown fluids are read but marked "not assessed". IEC typical values apply to power transformers only.
- With a communicating tap changer, the IEEE status remains the one the standard gives, even though part of the C2H2 may come from the tap changer; the tool flags it.
- No health index and no remaining life. The tool ranks by DGA urgency only.
- The result is an indication for planning, not a diagnosis.
Your data stays in your browser
Everything is calculated in your browser. The page's Content Security Policy sets connect-src 'none' and form-action 'none', so the values cannot be sent to any server. The tool uses no cookies, no analytics and no browser storage. Your history exists only in the file you download; if you try to close the page with unsaved data, the browser warns you.
The tool runs in your browser. Seetalabs® does not store or see the values you enter.
Frequently asked questions
Why is a transformer with high gas levels ranked "medium"?
Because its gases are above the IEC 60599:2022 typical values or at IEEE Status 2, and for both the action the standards give is closer monitoring. High priority is reserved for IEEE C57.104-2019 Status 3.
My transformers are filled with natural ester. Does the ranking still work?
Yes. Esters are compared with IEEE C57.155-2014 and IEC 63585:2026 typical values and use the ester Duval methods. They reach medium priority at most, as IEEE C57.155-2014 §6.2 gives no higher level.
Does the tool calculate a health index?
No. It ranks transformers by DGA urgency. A health index combines DGA with oil quality and other data, and is part of Ronin AI.
Can I keep the history between sessions?
Yes, in a history file you download. Load it next time together with the new laboratory CSV, and the new samples are added to each transformer.
When the fleet is larger than five
The file you use here opens in Ronin AI as it is: its columns are recognised and every row is shown for review before any calculation. From there each transformer gets the route for its fluid, its history and, where the data support it, a health index that states which method produced it. Tell us how many transformers your file covers and which laboratories produce it.
Tool version 1.5.3. Seetalabs®.
