How to Size an Energy Management System: Three Scenarios, Three Different Answers

There's no universal EMS size—stop pretending there is

I'm a quality and brand compliance manager at GridSense, an electrical equipment manufacturer. I review every EMS specification package and engineering submittal before it reaches our customers—roughly 200 packages a year. In 2024 alone, I rejected about 15% of first submissions. Not because the equipment was wrong, but because the specification didn't match the site.

The single most common mistake? Someone downloaded a template from a conference, or copied a competitor's spec sheet, without checking whether it fit the plant. An EMS sized for a multi-tenant industrial park rarely works for a single-stamp metal shop—and the reverse is just as true.

There's no “correct” size. There's only the size that matches your metering points, your reporting requirements, and your power quality sensitivity. Below are three scenarios I've seen play out repeatedly, how I size around each one, and how to figure out which bucket you're in.

Scenario A: Predictable single-building operations

This is the most common scenario I encounter: one building, steady production hours, mostly standard loads—HVAC, lighting, motors, a couple of large machine tools. Management wants monthly energy trends and maybe some cost allocation by department.

The mistake I see here is over-instrumentation. A vendor walks in and quotes 30 submetering points because a “smart submetering solution for factories” sounds impressive on a proposal. But if those 30 meters all sit behind one feeder serving one production line, you're not getting granular data. You're getting redundant data—and paying $8,000 more than you need to.

What I'd specify for this scenario:

• A main meter with power quality analyzer capability at the service entrance. Something that can capture sags, swells, and harmonics at 1-second resolution or better
• Four to eight sub-metering points on major distribution feeders (air handling, lighting, production line A, production line B)
• 15-minute interval data recording for energy accounting. Nobody's making billing decisions on 1-second data for this kind of site

That's a modest package. Size it in a day, install it over a weekend, and validate the CT wiring against a portable power quality analyzer within the first week.

In my first year, I made the classic specification error: I assumed “standard” metering accuracy meant the same thing to every vendor. It didn't. One quoted 2% accuracy on current transformers; another quoted 0.5%. We only caught the gap during a post-install audit. The redo cost $600 and a two-week delay.

Let me be clear about context: this worked for us because our site has clean single-tenant distribution and predictable load patterns. If you're dealing with seasonal spikes or shared facilities with adjacent tenants, the calculus changes. You'll need more meters and higher accuracy—don't let anyone tell you otherwise.

Scenario B: Multi-tenant or chargeback environments

Here's where sizing gets serious. If you're allocating costs to tenants, departments, or business units, you need billing-grade accuracy—not trend-level accuracy. The difference is roughly 0.5% vs. 2% on current transformers, and it matters when someone disputes a $9,000 monthly charge.

For this scenario, I size around three constraints:

1. Point count: One meter per billing unit. No exceptions. I've seen people try to divide a shared feeder proportionally—that always ends in a dispute.
2. Accuracy class: Revenue-grade CTs (Class 0.5 or better) and meters rated for billing. Don't let a vendor quote “typical” accuracy and expect you won't notice.
3. Data rate: 15-minute intervals are standard for billing. Some utilities push for 5-minute. Anything faster is overkill for chargeback purposes—it just adds storage costs and maintenance overhead.

What most people don't see is the wiring audit. I've yet to review a multi-tenant EMS package where the CT polarity and phase rotation were 100% correct on first install. Something always gets swapped. Budget for a third-party commissioning audit before you go live—it's cheaper than a tenant dispute.

We tried to keep three existing CTs from an older meter bank instead of replacing them. The accuracy class was unknown and the burden didn't match the new meters. The data drifted by 4% for two months before we caught it. Replacing all fourteen meters was cheaper than recalibrating the original three—a lesson in false economy.

Scenario C: Power-quality-sensitive processes

If you're running welding equipment, large VFDs, CNC machining centers, or precision assembly, your EMS sizing problem looks different. You don't need more meters—you need faster meters.

A 1-second interval meter will tell you how much energy the welding bay consumed. It won't tell you why your CNC controller tripped at 2:47 PM. That's a voltage sag of 22% lasting 180 milliseconds. A 1-second meter averages right over it. For reference, IEEE 519-2022 sets harmonic limits that these operations routinely exceed without visible symptoms until the right meter catches it.

For this scenario, the specs I check are:

• Meter sampling rate: at least 256 samples per cycle (roughly 12,800 samples/sec on 50 Hz, 15,360 on 60 Hz) for basic waveform capture
• Power quality analyzer capability: harmonic analysis to the 50th order, flicker measurement, transient detection
• Compliance: IEC 61000-4-30 Class A if you need regulatory-grade data, Class S if you're doing internal diagnostics

This is where the price jumps. You're looking at node-level monitoring rather than feeder-level. But for plants in this category, losing a batch of precision parts to a 180-millisecond sag costs more than the analyzer does.

The “just add more meters” advice ignores a basic physical fact: most factory circuits aren't wired for individual machine monitoring. Adding a meter between a breaker and a distribution block tells you the block's total draw, not each machine's. If you need per-machine data in this scenario, you're looking at current transformers on each motor starter—a very different project scope.

How to figure out which scenario you're in

I ask three questions before I even open a spec:

1. Is anyone going to dispute the bill? If yes, you're in Scenario B. Revenue-grade accuracy and one meter per billing unit. No shortcuts, no proportional splits.
2. Do you know why your equipment trips? If you've had unexplained faults, product damage, or downtime you can't trace to a mechanical cause, you're in Scenario C. A high-speed power quality analyzer isn't optional.
3. Does anyone act on the data daily? If the answer is “monthly reports and annual budget planning,” you're in Scenario A. Don't pay for performance nobody will use.

When our GridSense EMS energy meter gets specified on a project, these are the questions I ask the integrator before I sign off. The meter itself is neutral—it's the sizing that makes or breaks the deployment.

This sizing guidance reflects our Q1 2025 audits and specifications. Standards and equipment change. Verify current requirements against IEC 61000-4-30 or your local grid code before you finalize a package. And if you're adding a power quality analyzer to the spec, confirm that the CT burden and accuracy match the meter's input requirements. That mismatch causes more install delays than any other single issue I've seen.

The bottom line: an EMS isn't a one-size product. The correct size is the one that survives your first audit. If you can't explain why each meter exists and who will use its data, cut the point count. If you can explain it, don't let a vendor talk you down to a cheaper tier that can't deliver what you need.

Marta Kowalska

Marta Kowalska is an electrical equipment procurement and quality analyst covering distribution gear, automation, generators, pumps, test instruments, and engineered assemblies. She reviews IEC 61439 routine-verification records for assembly current and temperature rise, IEC 60034-1 rating and test data for rotating machines, and ISO/IEC 17025 laboratory scopes for reported calibration uncertainty. She helps buyers compare inspection status, document traceability, total installed cost, lead time, service access, spare-parts support, warranty boundaries, and failure consequences across equipment categories.