Submetering for Factories: Design, Data and Allocation
Factory energy submetering is the practice of installing measurement devices at points inside the plant electrical distribution system, downstream of the utility incomer, so that consumption can be attributed to lines, processes, buildings or tenants. The utility meter at the incomer gives one number for the whole site; submetering breaks that number down into quantities that can be acted on. This page sets out practical guidance for electrical engineers and plant buyers who are planning a submetering system or reviewing an existing one: where to place meters, how to collect data, how to allocate costs, and which errors are most common in the field.
Why the Utility Incomer Total Is Not Enough
A single incomer meter answers one question: how much energy did the site import in the billing period. It cannot answer the questions that drive operating decisions.
- Which line or process consumes the most energy? Without sub-metered data, a high-consumption line can only be identified by estimate or by shutting equipment down and observing the incomer, which interrupts production.
- Where did consumption change? An increase in the monthly total may come from added shifts, a failing motor, a new chiller, or a tenant expanding their operation. The incomer total cannot separate these causes.
- How should internal costs be allocated? Departments, product lines and tenants usually pay for energy on some internal basis. Allocation requires measured quantities at the boundary of each cost centre.
- Is a specific load behaving abnormally? Efficiency problems, unauthorised loads and equipment left running outside shift hours show up as deviations at the submeter, not in a site-wide total.
The incomer total remains necessary — it is the reference for billing reconciliation and for overall benchmarking. Submetering does not replace it; it makes the total explainable.
Designing the Metering Hierarchy
A factory energy submetering system is normally organised as a hierarchy rather than a flat list of meters. Each level answers a different class of question, and the levels should sum, within measurement tolerance, to the level above.
Level 1: Incomer and main distribution
Meters at the utility incomer and at the outgoing feeders of the main LV or MV switchboard. These establish the site total and the split between major distribution boards, transformers or buildings. In many plants the protection relays or the utility meter already provide some of this data; confirm accuracy class and register set before relying on it.
Level 2: Distribution boards and area panels
Meters at sub-distribution boards feeding a workshop, a production hall, a utility area (compressed air, chilled water, water treatment) or a tenant space. This level is where most allocation decisions are made, because it usually matches the cost-centre structure of the plant.
Level 3: Individual large loads
Dedicated meters on significant individual loads — large motors, compressors, chillers, furnaces, injection moulding machines, test benches. The threshold for "significant" is a design decision; a common approach is to meter any load above a chosen rating, or any load whose consumption is expected to be a material share of its parent board.
Design rules that prevent rework
- Define the reporting structure (site → building → cost centre → line → load) before specifying hardware, so that meter addresses, tagging and naming follow the same tree.
- Leave spare metering positions and spare communication capacity at each level; adding a meter to a live board later is more expensive than reserving space and a Modbus address now.
- Record the CT ratio, meter serial number, firmware version and installation date for every point at commissioning. This record is the reference for every later accuracy investigation.
Choosing Metering Points
Not every cable needs a meter. The choice of points should follow the decisions the data will support. The main metering strategies are:
By line or production process
Meters on the distribution board or boards that feed a defined production line or process step. This supports energy-per-unit-of-output reporting and comparison between shifts or between similar lines. It requires the electrical distribution to follow the process boundaries reasonably closely; where it does not, partial metering plus allocation is the practical compromise.
By tenant or cost centre
In multi-tenant industrial buildings, meters at each tenant's supply point are the basis for recharging. The meter should be at the boundary the lease defines, and the lease should reference the meter, its accuracy class and the reading method.
By utility system
Compressed air, chilled water, cooling towers, dust extraction and wastewater treatment are often metered as systems, because their consumption is driven by plant-wide demand rather than by a single line. System-level metering supports leak detection and efficiency work on the utility itself.
By large individual load
For loads that dominate a board, a dedicated meter gives direct data for maintenance and replacement decisions. For smaller loads, group metering at the panel is sufficient and cheaper per point.
A practical checklist for selecting points: does this point correspond to a cost centre or a management decision; can the meter be installed and maintained safely; is the load current within a sensible CT range; and will the data be read and used, or merely recorded?
Multi-Tariff Registers and Time-of-Use Data
Many industrial tariffs price energy differently by time of day and season, and some include demand charges based on the maximum import in a window. Meters that support multi-tariff registers record consumption separately in each tariff period, which allows the plant to compare its load profile against the tariff structure and to verify the utility bill line by line.
- Confirm that the meter's real-time clock is synchronised and that daylight-saving or time-zone handling matches the utility's billing convention; a one-hour offset misassigns consumption between tariff periods.
- Demand registers (for example, maximum demand over a fixed window) are useful for checking whether operational changes — starting large motors sequentially rather than together — affect the billed demand.
- Where the utility bill is the cost basis, the submeter tariff configuration should mirror the utility's periods, not a simplified internal version, if the submetered data is to reconcile with the bill.
Data Collection: RS485 Modbus Networks and Gateways
Most industrial submeters expose data over an RS485 serial bus using the Modbus RTU protocol. A typical factory energy submetering installation connects meters in daisy-chain or bus topology to one or more gateways, which poll the meters and forward data to a local server or an energy management platform.
Network planning points
- Bus length and device count. RS485 segments have practical limits on cable length and number of devices; long plants are usually split into several segments, each with its own gateway or serial port.
- Addressing. Every Modbus slave needs a unique address on its segment. Agree an addressing plan that follows the metering hierarchy, and keep it in the commissioning record.
- Termination and wiring. Use twisted-pair cable, correct termination at the segment ends and consistent A/B polarity. Most intermittent "meter offline" faults are wiring or termination faults, not meter faults.
- Polling load. Reading every register from every meter at short intervals can saturate a slow segment. Read the registers actually needed (energy registers, demand, instantaneous values) at an interval matched to the reporting requirement.
- Time synchronisation. Gateways should keep meter clocks consistent so that interval data from different meters can be compared and summed.
Where a plant already has a SCADA or BMS network, the submetering gateway can often feed it directly, avoiding a second data infrastructure. Decide early whether the metering system will be standalone or integrated, because it affects gateway selection and network segregation.
Allocation Methods and Common Mistakes
Measured data still has to be turned into cost allocation. The method should be written down before the first invoice or internal report is produced.
Allocation approaches
- Direct allocation. Where a meter covers exactly the cost centre, allocate the measured consumption directly. This is the preferred method and the reason metering points should follow cost-centre boundaries.
- Sub-metered plus shared services. Utility systems (compressed air, chilled water) are metered at the system level and allocated to users by a documented key — for example, by the users' share of direct metered consumption, or by measured flow where air or water meters are installed.
- Residual allocation. The difference between the incomer total and the sum of submeters (distribution losses, unmetered loads) is allocated by a stated rule. Publishing this residual openly avoids disputes and highlights unmetered areas worth adding to the system.
Common mistakes
- CT ratio errors. A meter configured with a CT ratio different from the installed CTs will report scaled consumption — often by a round factor such as 2, 5 or 10. The error is easy to miss because the data looks plausible. Cross-check each submeter against a portable reference or against the parent meter's balance at commissioning.
- Unbalanced three-phase loads. Single-phase loads distributed unevenly across phases cause phase currents and phase voltages to differ. Meters that assume balanced conditions can produce inaccurate power and energy readings on such circuits. When specifying meters for panels with mixed single-phase loads, confirm that the model handles unbalanced loads; the documented DNS 9000 series is noted as handling unbalanced loads.
- Metering the wrong side of a boundary. A meter installed upstream of a tenant's isolation point, or downstream of a shared busbar, will not match the commercial boundary it is meant to support.
- Missing or inconsistent naming. Meters that are physically correct but poorly labelled in the data system produce reports nobody trusts. Enforce a naming convention at commissioning.
- Reading data but not acting on it. A submetering system without a review routine — comparing periods, checking residuals, investigating deviations — generates cost without generating savings.
Payback Logic and a Practical Checklist
Payback for a submetering project is assessed from the costs it makes visible or avoidable, not from a generic saving figure. The logic is straightforward: the system cost (meters, CTs, wiring, gateways, installation labour, software) is compared with the value of the decisions the data enables. Those decisions typically fall into four groups:
- Allocation accuracy. Correct recharging to tenants or departments, and recovery of energy costs that were previously absorbed as overhead.
- Waste identification. Loads running outside production hours, compressed-air leaks, oversized or idling equipment — identified from interval data and corrected.
- Demand management. Verifying the effect of load scheduling on billed demand.
- Verification of efficiency work. Before-and-after measurement of equipment replacement or process changes, using the same meters and the same boundaries.
Because these values are site-specific, the payback calculation should be built from the plant's own tariff, load structure and identified opportunities. A useful discipline is to size the first phase of metering around the two or three decisions with the clearest value, then extend the hierarchy once the data routine is established.
Pre-installation checklist
- Metering hierarchy and naming convention agreed and documented.
- Metering points matched to cost centres and to the lease or reporting boundaries.
- CT ratios, accuracy class and burden checked against the load current range.
- Meter specification confirmed for unbalanced-load capability where single-phase loads are mixed.
- RS485 segments planned: cable route, termination, addressing, gateway capacity.
- Register list and polling intervals defined; tariff periods and clock handling confirmed.
- Commissioning record template ready: serial numbers, CT ratios, firmware, installation dates.
- Allocation rules and residual rule written down before the first report.
Building the System: Meters and Management Platform
The metering hardware and the data platform are two parts of one system, and the choice of each constrains the other. LOGGON TECH supplies industrial metering and monitoring equipment, including meters for panel and feeder installation; a gateway or data collector for Modbus data collection is specified per project. The product range for this application is described on the energy submetering page, which covers meter types and their typical installation points in a factory distribution system. The software side — data collection, tariff handling, allocation reporting and deviation review — is described on the energy management system page. Engineers specifying a new system should read both pages together with the checklist above, since the meter register set and communication options determine what the platform can report, and the reporting requirements determine which meters and gateways are needed.
FAQ
How many submeters does a factory need?
There is no fixed number. The count follows the metering hierarchy: at minimum, meters at the main outgoing feeders, then at distribution boards matching cost centres, then on significant individual loads. A first phase often covers the main feeders and the largest loads; further points are added as the reporting routine develops. The design rules in this page — spare capacity, consistent naming, documented commissioning records — matter more than the initial count.
Can submeters share the existing RS485 network or BMS?
Often yes, provided the segment has capacity for additional devices, the addressing plan allows unique addresses, and the polling load is acceptable. Where the existing network is heavily loaded or serves critical control functions, a separate metering segment with its own gateway, forwarding data to the management platform, is the more robust arrangement.
Why do submeter readings not add up to the utility bill?
Small differences are normal and come from meter accuracy class, CT accuracy, timing differences between meter clocks and the utility's billing periods, and distribution losses. Large differences usually indicate a configuration error — most often a wrong CT ratio — or unmetered loads. The residual should be calculated and tracked routinely; a sudden change in the residual is itself a useful fault indicator.