Engineering question
How should an industrial site turn meter data into reliable energy visibility without claiming guaranteed savings?
Energy monitoring is an information system, not a saving by itself. The architecture must preserve measurement context from the electrical single-line and meter through communication, timestamps, storage, calculations and dashboards. Without meter hierarchy and production context, a polished dashboard can still produce misleading conclusions.
Begin with decisions: which tariff, demand, process or equipment questions should operators and managers answer? Then select measurement points and data resolution. A main incomer may support billing reconciliation; a feeder meter may isolate a production area; a machine-level meter may support operating-state or specific-energy analysis.
Calculation basis
Formulas and units
Energy from interval power
E(kWh) = Σ Pavg(kW) × interval hours
Use synchronized intervals and preserve gaps; do not silently treat missing data as zero.
Specific energy
SEC = energy / accepted production quantity
Define product, boundary, scrap/rework and time alignment before comparing shifts.
Demand
Demand = energy accumulated over the tariff interval / interval hours
The utility's interval and method must be represented accurately.
Worked example
Apply the formula
A line averages 320 kW for 45 minutes and produces 18 accepted tonnes during the aligned interval.
- 1Energy = 320 × 0.75 = 240 kWh.
- 2Specific energy = 240/18 = 13.33 kWh/t.
- 3If 2 tonnes are later rejected, the KPI basis must state whether accepted tonnes are revised.
Result: The interval SEC is 13.33 kWh per accepted tonne on the stated boundary; it is not directly comparable until quantity and time conventions match.
Reference architecture
- Revenue, incomer and feeder meters located from the approved single-line diagram
- RS-485/Modbus or Ethernet network with documented addresses and polling budget
- Edge gateway or PLC for protocol handling, buffering and quality state
- Time-series or relational storage with UTC timestamps and site-time display
- Calculation service with versioned KPI definitions
- Role-based dashboards, alarms, reports and export
- Backup, retention, cybersecurity and maintenance ownership
Data-quality controls
Store quality alongside value. Detect counter reset, rollover, flatline, impossible jumps, duplicate timestamps, communication loss and time drift. Reconciliation from submeter totals to the parent meter can reveal gaps, but expected unmetered loads and transformer losses must be documented.
- Do not interpolate billing or compliance data without an approved rule.
- Version meter multipliers, CT/PT ratios and replacements.
- Preserve raw cumulative energy as well as derived intervals.
- Align production events and meter intervals using a governed time source.
From visibility to action
Use the system to identify questions for engineering investigation: abnormal idle load, demand peaks, compressor or pump scheduling, power-factor behaviour, production-specific energy or maintenance drift. Savings require a baseline, an implemented measure and measurement/verification appropriate to the project. The dashboard alone does not guarantee ROI.
Common mistakes
- Starting with dashboards before defining decisions
- Ignoring CT/PT ratios and meter replacement history
- Treating communication loss as zero consumption
- Comparing SEC values with different production boundaries
Troubleshooting checks
- Negative or implausible power: check CT polarity, phase association and meter wiring.
- Missing intervals: inspect poll budget, gateway buffer and time sync.
- Submeters exceed parent: verify hierarchy, interval alignment, multipliers and duplicated meters.
- Demand differs from utility: reproduce the exact tariff interval and aggregation method.
Assumptions
- Meters and CT/PT circuits are suitably specified and installed
- Time and production context can be aligned
- KPI boundaries are approved
Limitations
- Not a utility revenue-metering design
- Does not guarantee energy savings or payback
- Power quality and harmonic studies may require dedicated instruments and methods
Topic cluster