APY PREMIUM GROUPMost plants know exactly what they paid for electricity last month. Very few can say which department spent it. This article describes a real project in which our team installed an energy monitoring system at a plastic packaging plant, breaking consumption down by department using meters at the main distribution boards. We describe the engineering approach and the transferable lessons only — no client name, no actual consumption figures, no equipment brands, and no commercial details, in line with our confidentiality obligations.
The brief: one bill, dozens of machines behind it
The plant runs several production groups — blow moulding, injection moulding, printing and decoration — plus utilities (chillers, air compressors, cooling) that run continuously. Electricity is one of the largest cost lines in a plant like this, yet the only data management had was a single monthly utility bill.
That left three basic questions unanswerable:
- If the bill rose 8% this month, which department caused it — or was it simply higher output?
- What is the energy cost per unit produced on each line?
- If we invest in efficiency improvements, where should we start to get the best return?
These three questions are the real starting point of almost every energy monitoring project we see. The driver is not a desire for attractive dashboards; it is that management is making spending decisions without data.
Why a single bill is not enough to cut cost
A utility bill is backward-looking and fully aggregated. It tells you that costs went up; it never tells you why. In practice the difference between the two kinds of data is stark.
| Aspect | Utility bill only | Department-level metering |
|---|---|---|
| Data frequency | Monthly, known only after the month has closed | Daily/hourly — issues can be corrected within the same shift |
| Granularity | Whole plant, aggregated | Per department and per distribution board |
| Root cause analysis | Guesswork and supervisor intuition | Compare the abnormal time window directly against the production schedule |
| Target setting | Plant-level only — nobody feels ownership | Per-department targets; each manager sees their own number |
| Proving ROI | Hard — other variables sit in the same bill | Measure before/after on the specific point that was improved |
The last row matters more than most people expect: an energy project that cannot demonstrate its result rarely receives a second round of funding, whether or not the first round actually worked.
What we actually installed: a 4-layer architecture
Energy monitoring systems that stay useful for years almost always share the same structure. Only the scale and the brand of each layer changes.
- Measurement layer (meters at the MDB panels) — energy meters installed at the main distribution board of each zone, with CTs matched to the rating of each panel. Measurement points are placed to match departmental responsibility boundaries, which are not always the same as the existing electrical single-line diagram. This is the single most consequential decision in the whole project.
- Communication layer — meter values are collected over an industrial fieldbus into a gateway, kept separate from the office network so that monitoring traffic never disturbs process control, and so that plant IT never becomes a single point of failure for energy data.
- Historian layer — every point is logged on a fixed cycle using one consistent naming convention across the plant (department / panel / signal type). This layer is what allows later phases to be added without rework.
- Presentation layer (dashboard) — KPI summary, daily trends, energy share by department, and file export so engineering and finance can work with the data in their own tools.
In this project the system covered the plant's main production and utility departments, viewable both at department level and at individual distribution board level.
Five lessons that transfer to other plants
- Place meters along organisational lines, not electrical ones — data only gets used when someone owns the number. A measurement point that straddles two departments ends up owned by neither.
- Agree definitions before pulling cable — does "the department's electricity" mean machines only, or does it include area lighting and HVAC? Settle this at design stage, not during the first monthly review.
- Adopt a naming convention from the very first point — plants that start with ad-hoc tag names hit a wall the moment they want to add meters or produce cross-department reports.
- Data alone does not reduce cost — what reduces cost is a weekly review of the numbers and the actions that follow. A system without an owner becomes an attractive screen nobody opens, usually within three months.
- Always leave spare capacity — nearly every plant that installs this wants more measurement points within the first year. Providing for it up front is cheaper than re-entering the panels later.
A 10-point checklist before installing department-level energy monitoring
- Which business question are you trying to answer (peak reduction / cost per unit / ISO 50001 preparation / internal cost allocation)?
- What is the electrical boundary of each "department", and does it match the current distribution layout?
- Is there physical space in each MDB panel for meters and CTs?
- Does CT installation require a panel shutdown, and what outage window is available?
- Are the electrical drawings current? (Older installations are usually modified in the field without updating the drawings.)
- How frequently will data be logged, and how many years of history must be retained?
- Will data live on-premise or in the cloud, and who is responsible for backups?
- Who owns these numbers, and on what review cycle?
- What comes in the next phase (time-of-use costing, solar integration, automated alerting)?
- After handover, can the plant team add meters and edit screens themselves, or must the contractor be called every time?
Items 1 and 8 determine whether the project delivers value at all. Items 3 and 4 are the ones that most often push the schedule.
What plants like this typically add next
Once the data layer is stable, extensions become far cheaper because no further electrical work is needed. The common next steps are:
- Time-of-use (TOU) costing instead of a flat average tariff, so the value of shifting load out of peak periods becomes visible.
- Peak demand monitoring with advance warning before the configured threshold is exceeded.
- Rooftop solar integration on the same screen, showing self-consumption versus grid import.
- Automated alerts and scheduled reports pushed to the channels the team already uses, so nobody has to watch a screen.
For budgeting purposes, an energy monitoring project in the range of ten to twenty measurement points, including the data layer and dashboard, typically falls somewhere between several hundred thousand and roughly one million baht, depending on point count, cable runs, the condition of existing panels, and reporting complexity. Treat this as an order-of-magnitude reference for early budgeting only — a firm price requires a site survey.
For the underlying principles and payback calculation, see our article on whether power monitoring really reduces factory electricity cost, or visit our power monitoring service page.
Conclusion
A successful energy monitoring project is not measured by how good the dashboard looks. It is measured by whether, six months after handover, people still open it and still make decisions from it. Getting there depends on aligning measurement points with real human responsibility, verifying data accuracy rigorously before handover, and having an owner who uses the data routinely. Of those three, the technology is by far the easiest part.