Why a flat last-year baseline fails ISO 50001

A plant that beat last year’s kWh total looks efficient until someone asks what the weather and the throughput did. ISO 50001 energy performance is not a year-on-year bar chart. It is expected consumption given the relevant variables, then a fair comparison with the meters.

A flat baseline treats every extra tonne of product as waste and every mild winter as a triumph. The energy team cannot defend that story in an energy review, and the board should not sign it.

AEYS trains EnRC models from the same mappings operations already use: timestamp, consumption, and relevant variables per unit and energy carrier. Actual-versus-predicted tables, WMAPE bands, CUSUM, and intensity indices sit next to the model so operators and reviewers share one picture.

Start with the SEUs that already have a year of tagged data. File uploads are enough for the first models. Connector Hub takes over when the same columns have to arrive every month.

The company energy report then quotes those runs instead of a second spreadsheet. That is how energy efficiency work survives surveillance year, not only certification week.