// Energy efficiency

Monitor energy consumption — and actually cut it.

Measuring energy is not adding one more meter: it is getting to answer «how much did this action save?» with a number that holds up in front of a client or an auditor. Here is what to measure, how to collect it without rebuilding the plant, and where people usually stop.

Why measure before optimising

Without a starting measurement, every saving is an opinion.

You cannot cut what you do not measure: the bill tells you how much you spent, not where or when.
The energy audit (mandatory for large firms under EU rules) asks for measured data, not estimates.
Whoever sells efficiency — ESCOs, energy performance contracts — must prove the saving, not declare it.
Consumption drifting from the norm is often the first sign of a failure on its way.

What to actually measure

Electrical, downstream: Voltage and current per line or machine: active power and power factor tell you not just how much you use but how well.
Every carrier: Not only electricity: gas, water, district heating, diesel, LPG. Real efficiency shows when you sum the carriers, not one at a time.
The context: Output, degree-days, running hours: a kWh means nothing without the driver that explains it.
The asset: Which machine or plant the consumption belongs to. Without an asset register, the figure is a number with no address.

How to collect the data without rebuilding the plant

The data is almost always already there: PLCs, meters, inverters. The work is collecting it, not generating it.

From existing PLCs and meters over Modbus TCP or OPC-UA: read what is there.
From an MQTT broker already in place, if the plant is connected.
Where measurement is missing, downstream nodes (e.g. Raspberry Pi with an ADC) reading voltage and current directly.
Outbound-only, per-device authenticated traffic: no inbound ports opened on the customer firewall.

From measurement to provable savings

This is what separates a dashboard of charts from a tool that closes a contract.

Set a baseline: expected consumption as a function of the drivers (output, weather), not a flat average.
Compare actual consumption with the baseline: the gap is the saving, measured per the M&V protocol (IPMVP).
Track EnPIs — energy performance indicators — over time, by department and by carrier.
Report: in an energy performance contract, the proven saving is exactly what you invoice.

The costly mistakes

«The bill is enough»

The bill is downstream of everything and monthly: it does not say which machine wastes or when. It is for paying, not for deciding.

«First install every sensor»

Most of the value comes from data you already have (PLCs, meters). New sensors go where measurement is missing, not everywhere.

«We will estimate the saving»

An estimate does not hold in a guaranteed-results contract or an ISO 50001 audit: you need baseline plus measurement.

«Energy and maintenance are two worlds»

A drifting electrical signature is both abnormal consumption and a failure on its way: splitting them loses both signals.

Have plants or buildings to monitor and want to know where to start? Let us look together at what is already measurable with what you have in the field, before buying a single sensor.

Let’s talk

Page updated 23 August 2026.