Planning Your Transformer Asset Management Programme for 2025

Year-end is a natural point to step back from individual transformer cases and assess the DGA programme as a whole: what it is covering, how reliably it is being done, whether the analytical methods applied are extracting the available information, and where the most valuable improvements lie.

This is a practical checklist of questions that most commonly reveal high-value opportunities in transformer DGA programmes, not a comprehensive audit framework, but a targeted review of the areas where gaps most commonly exist.

Sampling Coverage: Who Is Not in the Programme?

Are all critical transformers being sampled? Transmission-class equipment with no DGA history is a risk exposure that should not exist. IEEE C57.104-2019 [1] applies to oil-filled transformers in service; there is no category of transmission transformer for which no sampling is appropriate. The starting point for any programme review is a list of all in-service units and a check that each has a current DGA record.

Are transformers filled with alternative fluids in the programme on appropriate terms? Natural ester and synthetic ester transformers cannot be evaluated against the mineral oil threshold values in C57.104 [1] or IEC 60599. CIGRE TB 771 and TB 443 provide the appropriate interpretation frameworks for these fluids. If your fleet includes ester-filled units and your analytical approach applies mineral oil thresholds to them, you have a systematic analytical gap.

Are substation transformers with no backup supply identified for prioritised monitoring? CIGRE TB 812 [2] establishes that consequence of failure is a primary factor in maintenance strategy design alongside condition. A transformer whose failure would create a sustained outage with no available backup justifies a higher sampling frequency and consideration of online monitoring regardless of its current condition rating.

Sampling Frequency: Is It Calibrated to Condition?

IEEE C57.104 [1] recommends sampling frequency by condition level: annual for Condition 1, quarterly to monthly for Condition 3 and 4. The principle is sound, but the implementation requires active management.

Are elevated-condition transformers actually being sampled at their recommended frequencies? A transformer that moved to Condition 2 or 3 in a previous year but is still on an annual schedule is receiving inadequate attention for its current condition.

Is rate-of-change driving frequency increases, or only absolute concentration? A transformer with concentrations below condition level thresholds but a consistently rising trend over six samples is more concerning than one that has been stable at the same elevated level for three years. Rate-of-change should inform sampling frequency as much as absolute level, which requires an analytical method that tracks trajectory, not just current concentration.

CIGRE TB 445 [3] provides the framework for condition-based maintenance strategy, including how monitoring intensity should scale with assessed risk.

Analytical Approach: What Does Your Programme Actually Produce?

This is where the most significant improvement opportunities are typically found.

Does your DGA review produce a fleet-level risk ranking, or a series of individual transformer reports? If your end-of-review deliverable is a list of transformers that exceeded thresholds, you do not have a fleet-level view. Two transformers both below threshold could have entirely different risk profiles depending on their history, their trajectory, and where they sit in the population distribution of transformer gas profiles. Dukarm et al. [4] demonstrated systematically that threshold-only methods both miss genuine risk and generate false alarms.

Transformer Oil Analyst™ (TOA) generates CSEV (Cumulative Severity) and HF (Hazard Factor) metrics for every transformer in the database, population-normalised and history-integrating metrics that produce a ranked fleet view updated each time new results are entered [4]. If your analytical method does not produce this, it is worth evaluating whether it should.

Are rate-of-change signals being reviewed, or only absolute concentrations? If your review focuses on whether a number crossed a threshold, you are seeing one dimension of the data. A transformer generating gas at 20 ppm/month from a previous zero baseline is more concerning than one that has been at 150 ppm for five years. Both the current level and the rate of change matter.

Are carbon gases receiving appropriate attention? CO and CO₂ indicate cellulose insulation degradation, the component that most directly determines remaining transformer life and cannot be replaced without rewinding. A DGA review that focuses on hydrocarbons and treats carbon gases as secondary is systematically underweighting the most life-limiting information in the data.

Data Infrastructure: Is Your History Available?

Is historical DGA data consolidated in a single system? Long-established utility DGA programmes commonly have historical data scattered across laboratory reports, spreadsheets, legacy software, and paper files. That history is analytically valuable, and CSEV improves in reliability with longer records, but only if it is in the system where analysis occurs. A consolidation project for historical data is one of the highest-value data management investments a utility can make.

Is your laboratory relationship providing consistent, reliable results? Interlaboratory variability in DGA is documented and substantial, sufficient to shift results across condition boundaries without any change in actual transformer condition [4]. If your programme has used multiple laboratories over time, or if there are concerns about result variability or turnaround time, beginning of year is the right time to standardise laboratory relationships and document the analytical method used for each transformer's record.

Are sampling notes and provenance recorded alongside gas concentrations? An anomalous result is best investigated with knowledge of who collected the sample, when, under what load conditions, and from which sampling point. Programmes that record only the numbers and not the context lose information needed for informed interpretation.

High-Consequence Assets: Is the Monitoring Level Appropriate?

Do your highest-consequence transformers have adequate monitoring? CIGRE TB 630 [5] provides the selection framework for online monitoring deployment: units combining high operational consequence with elevated condition risk indicators justify continuous monitoring. For transformers where quarterly sampling leaves a 90-day blind window that represents unacceptable risk exposure, Monitor Watch provides continuous DGA visibility with the same R-DGA analytical framework applied to sensor data.

Do the highest-HF transformers in your fleet have current DGA data? If a transformer identified as fleet-highest-risk in the previous year's review has not been sampled since that identification, correcting that is the most immediate priority before any other programme improvement.

Planning for 2025

The improvements with the highest expected value in most transformer DGA programmes, based on where gaps are most commonly found:

1. Move from threshold-flagging to population-normalised fleet ranking (CSEV/HF)
2. Consolidate historical data into the analytical platform to improve CSEV reliability
3. Calibrate sampling frequency to condition trajectory, not just current level
4. Develop fluid-appropriate interpretation for ester-filled transformers in the fleet
5. Deploy continuous monitoring on units combining high consequence and elevated risk rank

For a structured review of your programme with Delta-X Research's support, contact us. For product details, visit the TOA page and Monitor Watch page.