Question 1

What does this tool actually show?

Accepted Answer

It applies per-gas Ostwald partition to a five-gas DGA fingerprint and visualises how the resulting point moves on Duval Triangle 1 and the Unified Pentagon (CIGRE 2023) as headspace volume V_G changes. The shift demonstrates how a real fault can be misclassified by sampling error alone.

Question 2

Why does the drift differ per fault type?

Accepted Answer

Different faults emit different gas mixes. A T2 thermal fault is dominated by C2H4 (Ostwald k_O ≈ 1.7); a partial discharge is dominated by H2 (k_O ≈ 0.05). The same V_G/V_L ratio removes a small fraction of C2H4 but a large fraction of H2, so the drift direction depends on which gases the fault chemistry produces.

Question 3

Which sources back the calculation?

Accepted Answer

Ostwald coefficients from IEC 60567:2023 Annex A Table A.1; Duval Triangle 1 zone boundaries from IEC 60599:2022 Table 1 and Fig B.3; Pentagon Unified zone set from Paulhiac & Duval (2023), "Unified pentagon for DGA in mineral insulating oil", CIGRE / Centre of Excellence for Transformers paper, with background in CIGRE TB 771 (2019).

Question 4

Are the example fingerprints real measurements?

Accepted Answer

No — they are illustrative. Only the T2/T3 boundary case used in the companion blog is a verified fingerprint. The other selector entries are typical placeholder values to demonstrate drift mechanics; they are not specific case data.

Gas	True (dissolved in oil)	Measured (after headspace loss)	$Δ (%)$
H2	30.0	15.4	-48.6%
CH4	200.0	178.1	-10.9%
C2H2	1.0	1.0	-4.1%
C2H4	180.0	175.0	-2.8%
C2H6	60.0	58.9	-1.8%

Gas	True (dissolved in oil)	Measured (after headspace loss)	$Δ (%)$
H2	30.0	15.4	-48.6%
CH4	200.0	178.1	-10.9%
C2H2	1.0	1.0	-4.1%
C2H4	180.0	175.0	-2.8%
C2H6	60.0	58.9	-1.8%

Sampling Drift on Duval Triangle + Pentagon

Triangle 1 — drift between true and measured

Sources

Related

Sampling Drift on Duval Triangle + Pentagon

Triangle 1 — drift between true and measured

Sources

Related