An industrial plant receives two letters each year: one from the electricity company and one from the gas supplier. With these two figures, and a few more that depend on a conversion factor, 90% of energy managers know how much it costs to run the factory. Many fewer know where and when this energy is consumed within the plant itself. And this second question is what decides whether the energy audit you have on the calendar will be a mere formality or a map of savings.
What the Directive already requires, even though Spain has not transposed it
Article 11 of the Directive (EU) 2023/1791 requires all companies with an average annual consumption exceeding 10 TJ (~2.8 GWh) to conduct an energy audit —or implement a management system— before 11 October 2026. Spain is behind: the new royal decree on audits is in public consultation and RD 56/2016 remains in force as a reference standard. But the deadlines of the Directive do not change, nor does the technical criterion.
The fundamental change is this: the audit ceases to be a document to become a process. The analysis of savings opportunities can no longer come from an estimate of total bills; it must come from a disaggregated reading of consumption. And a disaggregated reading of consumption is exactly what submetering makes possible.
Energy balance: what it must have to hold up
A well-made energy balance is not an Excel sheet with the kWh of the year. It is an engineering exercise with six components:
1. Three years of data, all the vectors. Electricity, natural gas, diesel, biomass, purchased heat, purchased cold, compressed air if it comes from a third party. All expressed in the same unit —tonne of oil equivalent, kilowatt hour, joule— to be able to sum it up. The Directive makes it explicit: it looks at the aggregated consumption of all vectors, not just electricity.
2. Normalised production alongside. Annual kWh do not say anything by themselves. kWh per tonne of product, per unit produced or per hour of operation do. The energy intensity series reveals whether an excellent year was real efficiency or a mix of technical downtime and a mild summer.
3. Segmentation by consumption centres. Not by line, not by building: by homogeneous energy consumption centre. Compressed air is one centre. Process cooling is another. Lighting is another. Each with its own behaviour, its own curve and its own opportunities.
4. Real load curves, quarter-hourly. An annual average is not useful for deciding anything worthwhile. The load curve reveals avoidable peaks, coincidences that are charged to the power term, unnecessary operation on weekends or night hours. The new regulation of the power term has made it critical.
5. Performance of critical equipment. COP of cooling, boiler efficiency, compressor efficiency, power factor. Not nameplate ratings: measured at real load and updated. The difference between a nominal COP of 4 and a real COP of 2.3 completely changes the business case for a replacement.
6. Comparable reference state. If the goal is to measure savings, a baseline built with the same method as the subsequent measurements is required. Without a baseline, what is called "savings" ends up being a collection of climate anecdotes.
Submetering: where the balance stops being arithmetic
Submetering is what turns an aggregated figure into a useful image. Without submetering, an energy balance can only tell you what you already know: how much you consume in total. With submetering, you can answer the questions that determine the investment: how much each process consumes per unit of product, what each cost centre really costs, which consumptions could be shifted to renewable generation hours, what percentage of the power term each line pays.
The implementation method we apply has three levels:
Level 1 — Major cost centres. Those that represent more than 10% of total consumption: process electrical power, thermal per product, auxiliary services (compressed air, cooling, lighting). Few measurements, high impact.
Level 2 — Critical decision centres. Those involved in imminent investment projects: heat pumps to replace gas boilers, LED lighting, high-efficiency motors. Here, the weight in the balance does not matter, what matters is that the measurement decides.
Level 3 — M&V verification points. The meters that will serve to demonstrate the savings after the measurement, following a methodology IPMVP or equivalent. Without this layer, a project does not generate CAE nor is it auditable by a bank.
Why this determines access to aid, not just the audit
The Spanish system of energy efficiency obligations —the Energy Saving Certificates, expanded by the Order TED/635/2026— requires demonstrating savings with certifiable methodology. The funds from the INNOVAE 2026 Programme and the industrial heat auctions of the European Innovation Fund demand the same: not an estimate, but a sustainable measurement over time. The revised Industrial Emissions Directive —which changes the logic of the AAI towards a transformation plan— also.
In all these cases, what decides who has access and who does not is the same: those who have the measurement organised before they are asked, and those who have to build it quickly when the requirement arrives.
How we do it at AUMA
At AUMA, energy balances are data and process engineering, not a document to comply with. The work we do has four ordered steps: prior diagnosis of the existing measurement system (what is there, what is missing, what is poorly calibrated), design of the minimum viable measurement plan to answer the relevant questions, coordinated implementation with maintenance and production so as not to stop the plant, and integration of the results with cost accounting and with the management system if there is one.
In 33 years we have learned that the best investment in energy efficiency is never the one a catalogue promises. It is the one your own plant points out to you when you measure it well. If you have an investment decision on the table, or an audit pending for the 11th of October, start here. Shall we talk?