What Comes After Metering: Moving from Electricity Accounting to Energy Management

Five things metering alone cannot handle

• No visibility into where energy goes: the main meter gets the bill exactly right but cannot say what share belongs to a given workshop, line or machine
• No baseline, so deviation cannot be judged: without knowing what normal consumption looks like at a given output level, all you can do is compare absolute figures month over month
• Peak-period load shifting depends on people: when demand approaches the contracted limit, a duty operator watches the meter and phones each area, so response speed depends on who is on shift
• Water, air and heat sit outside the same report: compressed air, steam and chilled water live in separate systems from electricity, so consumption per unit of product across all carriers cannot be calculated
• Carbon accounting lacks a data source: reporting has to be broken down by energy carrier and accounting boundary, while the granularity and retention period of metering data do not support it

All five share one root cause. The issue is not meter accuracy — it is that the data stops at settlement and never enters the production management loop.

Six layers from metering to management

Moving from "calculate the bill" to "control consumption" is not a system replacement. It is a matter of stacking layers on top of the metering base you already have. We recommend working through the six layers below in order, because the output of each becomes the input of the next.

• Submetering: place measurement points by workshop, process stage and energy-intensive equipment, and bring electricity, water, compressed air, steam and chilled water into one collection layer
• Baseline: establish reference values for consumption per unit of product and for idle-state equipment load, adjusted for output volume, shift pattern and ambient temperature
• Benchmarking: compare shifts on one line, lines within one plant and sites across a group, then narrow any gap down to a specific operation
• Alarms: deviation from baseline, prolonged idling, demand approaching its limit, suspected distribution leaks — rules trigger and route notifications by level of responsibility
• Load scheduling: how load is distributed across tariff periods, start-stop logic for chiller plants and compressed-air stations, and charge-discharge strategy for generation and storage assets
• Accounting: consumption is converted into comparable units by energy carrier, forming the ledger behind carbon accounting and energy-saving calculations — the same data set an ISO 50001 rollout requires

How the metering stage differs from the energy management stage

The difference between the two stages is not how many meters are installed. It is who uses the data, how often, and what decisions rest on it. The table below sets out typical differences across several dimensions.

One caveat: the right-hand column is not a target you reach in a single step. Most plants get there over two or three phases — submetering and baselines first, then alarms and load scheduling.

Data requirements and connecting to an existing metering system

Whether these layers can be built depends on three things: sampling frequency, measurement point coverage, and how the work connects to existing systems. A metering system already in service (known in Russian-speaking markets as АСКУЭ) usually does not need to be torn out — it is brought in as one source of electricity data.

The connection method follows site conditions. If the existing system exposes an OPC UA or Modbus interface, data is read directly; if only a database is available, a scheduled sync is built against the table structure; if neither exists, a parallel collection device is added on the meter side. Utility-side water, air, heat, HVAC and pure-water/wastewater have to enter the same collection layer as well, or a combined consumption definition never really holds.

• Sampling frequency: half-hourly or hourly values are enough for settlement data, while load scheduling and demand control call for minute-level sampling or finer
• Measurement point coverage: start with the equipment and process stages that account for the bulk of consumption; in general manufacturing that usually means compressed-air stations, chiller plants, drying ovens and heat treatment
• Data retention: baselines and benchmarking assume year-over-year comparability, so retention period and archiving method should be fixed during the technical design stage
• Architecture reliability: continuous-production sites can adopt a redundant OPC UA architecture, and data covering a communication outage is backfilled automatically once the link recovers

Load forecasting and large-language-model scheduling suggestions belong to our R&D and joint-validation track and are not part of standard delivery; any such suggestion is executed only after human confirmation. Cross-platform EMS and web management platform formats are assessed against the scope of each project.