The king of indicators: specific energy consumption

The starting point for almost all industrial energy management is specific energy consumption: energy per unit produced — kWh per tonne, per part, per batch, per square metre treated. It is the sector’s EnPI par excellence because it speaks the factory’s language: it relates energy to what the plant exists to do.

Below it sit the process-specific indicators that submetering makes possible: the oven’s consumption per tonne baked, the compressor’s per cubic metre of air delivered, refrigeration’s per hour of cold store held at set point. And, with yesterday’s instalment, its version in euros: energy cost per unit produced — the figure a management committee understands without translation.

The four mistakes that invalidate an indicator

Most EnPIs that fail don’t fail for lack of data: they fail by design. Four patterns keep recurring.

1 · Comparing without normalising. January’s kWh/tonne is not comparable with August’s if the plant is heated in winter, nor is a month at full load comparable with one at half load. A serious indicator is normalised for its explanatory variables —production, outdoor temperature, product mix, operating hours— just as required by the baseline of a CAE (Spain’s Energy Savings Certificates) application (in Spanish). Without normalisation, the indicator measures the calendar, not efficiency.

2 · Ignoring partial load. Almost no process consumes proportionally: an oven at half load uses far more than half of what it uses at full load. Specific consumption gets worse when production falls, even though nobody is operating any worse — and improves when production rises, even though nobody is operating any better. Anyone who doesn’t know this ends up praising and scolding the production team for commercial decisions.

3 · Changing the formula without leaving a trail. An EnPI that is recalculated differently every quarter —a different boundary, a different unit, a different allocation criterion— cannot demonstrate improvement to anyone: not to management, not to the ISO 50001 auditor, not to a verifier. Methodological consistency is worth more than sophistication.

4 · Publishing it late. A monthly indicator that arrives on the 20th of the following month describes history. The same indicator, daily or weekly, on the production whiteboard, drives behaviour — which is its only real purpose.

The minimum scorecard that works

For a medium-sized industrial plant, experience keeps showing the same pattern: four or five well-chosen indicators beat forty.

1. Plant-wide specific consumption (kWh/unit), normalised, weekly — the big picture.
2. Specific consumption of the dominant process, daily — where the game is won or lost.
3. Non-production consumption (nights, weekends, shutdowns) — the phantom-consumption detector, and the cheapest saving there is.
4. Energy cost per unit (€/unit, from yesterday’s instalment) — the one you take to the committee.
5. If ISO 50001 applies or there is a CAE application on the horizon: the EnPIs for the significant energy uses required by the standard, with their documented baseline — as we saw when explaining what the auditor asks for (in Spanish).

Each one with an owner, an alarm threshold and a fixed review time. An indicator without those three things is just a chart.

The PowerCloud view

EnPIs almost always die of maintenance: someone has to gather consumption, production and temperature data every week, recalculate, normalise and publish — and by the third month that someone has another urgent priority. PowerCloud keeps them alive: it calculates the indicators from the utility meter and submeter readings, normalises them by the variables you define, alerts you when they go outside their thresholds and keeps the historical series with the methodological consistency that an auditor —or your own committee— will want to go through. The production whiteboard is still the whiteboard; what changes is that nobody has to put together the number written on it by hand any more.

That brings the series to a close: measure by process, cost in real euros, manage by indicators. Three steps, each useless without the one before — and the third is the only one visible from the boardroom.

How many energy indicators does your plant have today — and how many are actually reviewed every week? We help you set up the minimum scorecard: which four or five EnPIs fit your process, which variables to normalise them by and how to keep them alive without anyone recalculating them by hand. We start from your real load profile — we obtain it from Datadis (the official platform of Spain’s electricity distribution companies) with your authorisation. Diagnostic session: 30 minutes, no obligation.

Sources: UNE-EN ISO 50001:2018 (EnPIs and significant energy uses); QUAMA and Gradhoc (industrial energy performance indicators); Cuerva and Apolo Energies (specific consumption and industrial practice); previous posts in this series at powercloud.es/noticias/.

La entrada Energy efficiency indicators in industry: the ones a committee understands and the mistakes that invalidate them se publicó primero en Powercloud | Energy monitoring and management.