Energy monitoring in a company – what data should you measure and how can you use it to manage energy?

At many sites, energy monitoring stops at the main meter, a few submeters and a dashboard with charts. That provides data, but it is not yet energy management. For monitoring to genuinely help improve energy performance, every measurement should answer a specific question: what do we want to detect, what will we compare the result against, what level will we treat as a deviation, and who is expected to respond.

A well-designed monitoring system shows not only how much energy a company uses, but also when, where and under what operating conditions the consumption occurs. Only by combining energy and power data with process data — production output, operating time, temperature, pressure, flow or equipment status — can normal operation be distinguished from losses, anomalies and gradual deterioration in efficiency.

Energy monitoring at a glance

• Start with business and technical questions, not with selecting a platform.
• Build a metering hierarchy: main meter → areas and systems → lines → significant energy users.
• Measure not only energy, but also the data that explains its use: production, operating time, temperatures, pressures, flows and equipment status.
• For each KPI, define a reference value, response threshold, owner and response procedure.
• Analyse baseload consumption, operation outside production hours, deviations from the expected profile and changes in efficiency under comparable conditions.
• BMS, SCADA and EMS platforms can provide data and perform control functions, but the IT system itself does not replace the energy management process.

Energy monitoring vs energy management – where is the dividing line?

A meter records a value. A dashboard displays it. Management begins only when the data is used in a repeatable process:

1. measurement,
2. assigning operating context,
3. comparison with an expected or baseline value,
4. identifying a deviation,
5. assigning responsibility,
6. corrective action,
7. verification of the result.

This distinction is consistent with the approach used in energy management systems. ISO 50001 provides a framework for systematic management of energy performance, while ISO 50006:2023 develops the topic of energy performance indicators and energy baselines. In this article, ISO 50001 is used only as an organisational context. The rest of the article focuses on technical monitoring of energy and utilities and on the practical use of data in operations.

A similar conclusion can be drawn from U.S. Department of Energy materials on EMIS: an information system can centralise data, analyse profiles, detect faults and support verification of results, but its value depends on regular use by a team that can turn a signal into action.

Auditor’s comment: The most common problem is not a complete lack of data. A site has utility bills, a main meter, machine controllers and sometimes a BMS or SCADA system. The problem starts when the data does not make it clear what exactly should be checked the next day and who should do it.

First determine which decisions require data

Before expanding the metering system, it is worth defining several specific monitoring use cases. For example:

• we want to know how much energy is used outside production hours;
• we want to separate the energy use of the production hall from the compressor room, cooling and HVAC;
• we want to assess the energy intensity of a line per unit of production;
• we want to detect increases in energy demand while the load remains unchanged;
• we want to check whether compressed-air pressure is being maintained above the level required by the process;
• we want to assess refrigeration-system operation against outdoor temperature and load;
• we want to confirm the effect of an upgrade after replacing equipment or changing the control strategy.

For each such use case, four questions need to be answered: what data is required, where will it come from, how often will it be analysed, and what decision should result from it. This helps avoid installing submeters “just in case”.

How should a company’s metering structure be designed?

The most useful structure is hierarchical. Measurements at a lower level should explain what cannot be understood at a higher level.

This does not mean that every motor needs its own meter. A metering point is useful when its result can change a technical or operational decision. If the twentieth submeter does not lead to any action, its value is limited regardless of the quality of the dashboard.

In practice, a good starting point is an energy balance covering the main energy carriers and the largest and most variable loads. Prior assessment carried out as part of a company energy audit or installation measurements can be helpful, because it shows where additional data granularity will actually be needed.

What data should be measured? Energy is only one layer

The choice of data should follow the physics of the system. A kWh reading alone will not explain why a compressor, refrigeration plant or heat substation has started to perform worse.

In compressed-air systems, combining energy data with flow and pressure is particularly important. Compressor power demand alone does not show whether the energy was used to meet actual process demand or, for example, losses in the distribution network. This is why such a system should be assessed as a whole: supply, controls, distribution network and end uses.

Electrical energy monitoring: energy and power answer different questions

Energy tells you how much was used over a given period. Power shows how intensively and when energy was drawn. In practice, both views are needed to manage a site.

The daily energy total may look normal even though an unusual short-term peak occurs. Conversely, the power peak alone does not show whether it was justified by production output. Time profiles should therefore be compared with shift schedules, line status and other operational data.

If the purpose of the analysis is to select distribution parameters, contracted power or a tariff, that is a separate topic related to electricity cost optimisation. Here, the power profile is used primarily for technical and operational diagnostics.

How should the data interval be selected?

There is no single correct interval for an entire site. Main-meter data at a resolution of a dozen or so minutes may be sufficient to analyse the facility profile and operation outside production hours. A dynamic process, a machine cycle or a rapid pressure change may require one-minute data or even shorter intervals.

The rule is simple: the interval should be short enough to reveal the phenomenon you want to respond to, but not shorter merely because the technology allows more data points to be recorded.

How can baseload consumption and operation outside production hours be identified?

Baseload consumption is recurring energy demand that occurs when the main process is not operating or is running only at a minimum level. Not all baseload is a loss — some equipment has to operate continuously. The aim is to distinguish justified demand from demand that has no technological reason.

The simplest analysis combines:

• the power profile from the main meter,
• the production and shift calendar,
• status of key systems,
• information on planned shutdowns.

Hypothetical example: after production ends, the site continues for 6 hours to draw 60 kW more than would be expected from the list of loads required for night-time operation. That is 360 kWh per night. Over 220 such nights, this amounts to 79,2 MWh of energy per year that needs to be explained.

This does not yet mean 79,2 MWh of savings. First, it is necessary to determine which equipment creates the difference and whether switching it off or reducing its operation is technically permissible. This is precisely why area submeters are usually more useful than trying to meter every individual load.

How can anomalies and deterioration in efficiency be detected?

An anomaly does not have to be a dramatic spike on a chart. Small but recurring deviations are often more important.

1. Deviation from the time profile

Equipment operates at times when it should be off according to the schedule, or starts earlier than the process requires.

2. Deviation from operating conditions

Two days have similar production output and external conditions, but energy use differs significantly. This is a signal to check changes in setpoints, operating mode, faults or operator overrides.

3. Increase in energy intensity

Total energy use may not increase if production has fallen at the same time. A kWh-per-unit indicator can therefore reveal deterioration that is not visible on the energy bill.

4. Change in the relationship between energy and a process parameter

For example, a compressor room may require more energy to deliver a comparable flow at a similar pressure, or a refrigeration system may draw more power at a similar load and outdoor temperature. Such a trend is often more important than a single alarm.

Automatic deviation detection can be useful, but the threshold should reflect the nature of the process. A threshold set too tightly generates hundreds of alerts that people stop reading after a few weeks.

Which KPIs actually help manage energy?

A KPI should answer a technical question. It does not have to be complicated.

Example indicators:

• kWh per tonne, piece or batch of product;
• kWh per hour of actual line operation;
• site baseload power in kW outside production hours;
• share of energy used outside planned operating hours;
• compressor energy use relative to the measured quantity of compressed air delivered;
• cooling or thermal energy relative to the scale of the process, where such a denominator is meaningful;
• maximum power under defined operating states;
• number and duration of deviations from the agreed profile.

For a simple and stable process, a single specific indicator may sometimes be sufficient. If outdoor temperature, product mix, line speed or the number of shifts has a significant impact on consumption, the comparison needs to be normalised for these conditions. Otherwise, the system may report “deterioration” that is actually caused by different production conditions or weather.

ISO 50006 specifically highlights the role of EnPIs and energy baselines in evaluating energy performance. In practical monitoring, this means that a chart without a reference point describes history; it is not an evaluation tool.

The most common mistake: vast amounts of data without a defined response process

Modern BMS and SCADA systems, PLCs and IoT platforms can provide thousands of data points. This is technically impressive, but from an energy management perspective it can lead to information overload.

For each key KPI or alarm, the following should be clear:

Data quality also matters: consistent units, time synchronisation, clear descriptions of metering points, handling of missing data, and periodic checks of sensor reliability. Without this, the system starts comparing values that only appear to describe the same phenomenon.

From alarm to action – a simple workflow for an industrial site

A good energy monitoring and management system does not end with an email notification. In practice, it should create a closed loop:

1. deviation detection — the KPI moves outside the agreed range;
2. data validation — check whether the problem is caused by a sensor or communication error;
3. context check — production, maintenance, start-up, recipe change, weather;
4. technical diagnosis — identify the probable source of the deviation;
5. action — adjust a setpoint, repair equipment, change a schedule, switch off unnecessary load;
6. verification — compare data after the change with the expected value;
7. record the conclusion — update the threshold, procedure or alarm logic if necessary.

In its materials on EMIS operations, the DOE describes a similar model: issues identified by the system need to be verified and prioritised, corrective actions then implemented, and the results subsequently checked. Simply generating more alarms does not create savings.

How can monitoring be used to verify the results of an upgrade?

Monitoring is particularly valuable after an upgrade because it makes it possible to verify not only the initial result after commissioning, but also whether the result is sustained.

Before making the change, define:

• the baseline period,
• the scope of data,
• conditions that affect energy use,
• the normalisation method,
• the date the upgrade is introduced,
• the assessment period after the change.

If energy use falls from 12 to 9 MWh per day after an upgrade, but production falls by 30% at the same time, the entire difference cannot automatically be attributed to the investment. Conversely, unchanged total energy use does not mean the upgrade had no effect if the site is producing more.

For this reason, upgrade performance should be assessed under comparable conditions: production level, operating time, outdoor temperature, equipment load or other relevant variables. For more complex processes, a baseline model may be needed instead of a simple “before and after” comparison.

Monitoring can also reveal so-called operational drift: after several months, setpoints are changed, schedules are extended, overrides appear and equipment gradually returns to more energy-intensive operation. The benefit of an upgrade can then be lost without any failure that is visible to production.

BMS, EMS and SCADA – how can existing automation be used?

In many companies, some of the required data already exists.

• BMS/BAS typically collects data from building systems and can control HVAC, heat substations, lighting and other technical systems.
• SCADA provides information on the technological process, machine status and operating parameters.
• EMS/EMIS can act as an integration layer for energy and process data, calculate KPIs, generate alarms and support analysis.
• equipment controllers, variable-speed drives and PLCs often already provide parameters that can be used without adding a separate sensor.

System terminology differs between manufacturers, so the acronym matters less than the question of what function the system performs and what data is available.

A BMS is therefore a valuable part of monitoring, but it is not synonymous with energy management. It can be a data source and an execution layer for automatic control of energy use. Only when it is combined with KPIs, response logic and verification of results does it become part of a system that supports real decisions.

There is also no basis for assuming a universal percentage saving from the implementation of a BMS alone. The effect depends on the starting point, the quality of control, the operating profile and the technical capabilities of the system.

Monitoring and management of thermal energy – a heat substation as an example

Heat-substation automation is a good example of a system where measuring energy alone is not enough. It is worth looking at it more broadly than only from the control perspective.

A heat management system can use, among other things:

• thermal energy,
• flow,
• supply and return temperature,
• outdoor temperature,
• zone temperatures,
• control-valve position,
• pump status,
• building occupancy schedule,
• information on demand from the process or production.

This makes it possible to check whether the system is heating outside occupied hours, whether network parameters are appropriate for the weather, whether the controller is operating unstably, and whether the effect of a setpoint change is sustained over the following days.

The same logic applies to thermal and cooling energy management. In a chilled-water system, electricity consumption needs to be supplemented with temperatures, flow, load, chiller and pump status, and external conditions. Only then can an increase in energy use caused by higher demand be distinguished from deterioration in system performance.

How should energy monitoring be implemented in stages?

There is no need to start with a project covering the entire site. In many cases, a controlled pilot is better.

Step 1. Define 3–5 problems that monitoring should solve

For example: night-time consumption, compressed air, cooling, line energy intensity, verification of an upgrade.

Step 2. Map the data that already exists

Check the main meter, submeters, BMS, SCADA, PLCs, variable-speed drives, heat and gas meters, and flow meters. Often, a significant part of the infrastructure is already in place.

Step 3. Build an energy balance and metering hierarchy

Define the hierarchy at site, area, line and key-load level. Only then identify the metering gaps.

Step 4. Define the process context

Link energy data with production, schedules, operating status, temperature, flow and other variables that genuinely affect the result.

Step 5. Define KPIs and baseline values

Each indicator should have a unit, data sources, a reference point and an owner.

Step 6. Define alarms and the response procedure

Do not start with dozens of notifications. First implement a few alarms that lead to a specific action.

Step 7. Verify one complete cycle: signal → diagnosis → action → result

If the organisation does not close this loop, adding more meters will only increase the amount of unused data.

Step 8. Expand the system where the pilot has demonstrated value

Only after the process has been proven should additional systems be included and analysis automated.

Where should an existing industrial site start?

If a company already has meters, BMS, SCADA or equipment data, the first step should not automatically be to buy another platform. First determine which data is reliable, what is missing and which decisions are supposed to be made on the basis of it.

A practical starting point is to analyse the available profiles, map the main energy users and define 1–3 use cases. As part of energy audits and measurements, the energy balance can be combined with supplementary measurements and the data structure can be planned so that monitoring later supports operation, diagnostics and assessment of upgrades.

If the company has a lot of data but it is difficult to identify which signals require action, the starting point should be an audit of the metering structure and how the data is used, not another dashboard. To discuss the scope of such an analysis, contact Energy Trend.

Sources

• ISO 50001:2018 (with Amd 1:2024) – Energy management systems — Requirements with guidance for use
• ISO 50006:2023 – Evaluating energy performance using energy performance indicators and energy baselines
• U.S. Department of Energy – Energy Management Information System Capabilities
• U.S. Department of Energy – Operations That Support Energy Management Information Systems
• U.S. Department of Energy – Compressed Air Systems