One number is not enough
A remaining-life estimate for a power transformer is a forecast, and every forecast is uncertain. Reporting it as a single number, such as 14 years, invites two mistakes: trusting it more than the data allows, or dismissing it because it once turned out wrong.
GENESYS AMS & LES reports remaining life as a median with an 80% band. For the sample transformer in bay three that is 14.2 years, with an 80% band from 11.8 to 18.0 years. The band is part of the answer, not a footnote.
The inputs are named
Each estimate names the model that produced it and the data it used. For a transformer that typically means a thermal ageing model driven by load and ambient temperature, the trend of dissolved gases from oil tests, and inspection findings. The sample estimate for that transformer uses 12 inputs, and the screen lists them.
When a new oil test arrives, the estimate moves and the record shows which input moved it. Planners can see whether a change came from a measurement or from a change in the model.
Reading the band
A wide band says the record is thin: few tests, a short history or conflicting readings. The useful response is usually more data, such as an extra oil test or an online monitor, rather than an early replacement. A narrow band that crosses the replacement threshold within the planning horizon is a different signal: the asset belongs in the capital plan.
- Median and 80% band, always together
- Model and inputs listed beside the estimate
- Every change linked to the input that caused it
- Wide band: collect data. Narrow band near the threshold: plan the replacement
Limits
Life estimation does not replace engineering judgement. It summarises the record in a form that can be compared across a fleet. The models are only as good as the tests and inspections behind them, which is why GENESYS keeps those records next to the estimate.
More on this: GENESYS AMS & LES.