I work at GridSense. Specifically, I'm the quality/brand compliance manager. Technical specifications, data sheets, quote packages—anything that goes out to a customer, I review before it leaves. Roughly 200 items a year. In 2023, I rejected about 18% of first deliveries for spec mismatches or lack of traceability.
In November 2023, a customer came to us while they were looking at power factor correction equipment suppliers. They brought a quote they'd already received: $14,200 for a 480V capacitor bank, rated around 200 kVAR. On the load profile line of the quote, it said: "Based on design peak load: 380 kW."
I did not sign off. Not because of the price. Because that 380 kW number came from a 2022 construction drawing—for a plant expansion that included CNC machining centers, a larger air compressor, and a 500 kW induction furnace.
Those units hadn't been powered in 2022. Design load is not real load. This is the outsider blindspot: most buyers focus on the per-kVAR price of power factor correction equipment and completely miss what actually determines what you need—the 15-minute and 60-minute load shape. Design load isn't measured load.
How to Size a Demand Response Management System (And Why Design Loads Don't Cut It)
This problem surfaced because the same 380 kW design number that was driving the PFC equipment sizing was also being used to size the demand response management system the customer was planning.
That was wrong. Specifically wrong.
The U.S. Department of Energy's Federal Energy Management Program (FEMP) material on power factor correction says it plainly: low power factor gets penalized, and correcting to 0.95 or higher requires measured load data at the point of installation, not a data sheet number. DOE FEMP also warns that capacitor banks interact with harmonics—so specifying a capacitor bank without a harmonic analysis is how you end up solving a problem and creating a new one at the same time.
For sizing a demand response management system, a Lawrence Berkeley National Laboratory report on U.S. demand response potential found roughly 32 GW of dispatchable load already enrolled in U.S. demand response programs, with far more potential. The value isn't in the total; it's in what can be curtailed per 15-minute or shorter interval. That requires a power quality analyzer logging 15-minute intervals—not a capacity table.
(I should mention: this wasn't a new discovery. I'd read the DOE material back in 2019. But in 2023, I started applying that spec to how we review PFC quotes—not just how we explain PFC to customers.)
What the Analyzer Reading Changed
The customer eventually agreed to hold off on the decision and let us do an analyzer recording. We connected a GridSense power quality analyzer to the 480V distribution panel and left it running for a full production month.
The peak demand wasn't 380 kW. It was 412 kW, at 2 p.m. on a Tuesday. The power factor dropped to around 0.86 in the same interval—roughly 245 kVAR of reactive power, on a 412 kW load. Correcting to around 0.97 would require about 145 kVAR of capacitor bank.
But that wasn't the worst part.
The worst part was the harmonics. The analyzer picked up 5th and 7th harmonic distortion levels that would exceed IEEE 519 limits once any capacitor bank was installed (IEEE 519 is the standard for harmonic control in electric power systems). Meaning the "standard" capacitors in the original quote would almost certainly have failed. Not because they were undersized—because they weren't detuned. And the vendor's quote didn't include a harmonic analysis at all.
This is what I call the rookie mistake. In 2022, when I first owned a significant project spec, I made the same one: I treated power factor correction and demand response as a single sizing exercise. They're not. PFC equipment corrects power factor and harmonic distortion. DR equipment manages curtailment and time-of-use value. They share data—usually from the same power quality analyzer—but not the same size calculation.
The customer's existing energy management system had been giving them monthly peaks, but it hadn't captured the 15-minute harmonic spikes. That's the gap between having data and having the right data.
Two Prices—And What Drove the Gap
The original $14,200 quote was equipment-only. No analyzer, no study, no installation. The terms simply said "site survey available on request."
We re-quoted. $19,200, covering: power quality analyzer recording, 15-minute load profiling, harmonic analysis, PFC equipment sized against the measured load curve (with detuned capacitor banks), and installation.
My first reaction: 35% more.
But look at what the $14,200 quote left out:
- Analyzer recording—needed a separate vendor or separate purchase
- Harmonic analysis—not included; "available on request" means not in your budget
- Installation—not included
- Capacitor sizing—"standard," meaning no detuning reactors
- Measured load profile—completely absent
We roughly estimated that the first quote would have landed between $22,000 and $24,000 once the study, the proper installation, and a detuned capacitor bank re-order were added in—before counting the schedule delay from the re-order.
(Should note: that's my estimate based on the customer's situation at the time, not a final invoice.)
GridSense's $19,200 was the final bill. No mid-project re-quote. Delivered inside the committed schedule.
Why Transparent Quotes Cost Less—And What I Ask Now
My position on this is shaped by the role: transparent pricing—even when the number looks higher—usually lands at a lower total cost. Because you can compare it. A "we'll study it later" quote doesn't compare. You're just hiding an unknown behind a lower headline number.
I now ask these questions on any PFC or demand response management quote:
- "Can I see the 15-minute load curve?" If the supplier doesn't have one—or didn't run it before quoting—they're guessing at sizing.
- "What load profile are you assuming?" If it says "based on design load," flag it. Design load isn't measured load.
- "Is the study included, or is it extra?" If it's unclear, assume it's not included.
- "Is harmonic analysis included?" IEEE 519 limits matter, and capacitor banks amplify 5th and 7th harmonics if you don't analyze first.
- "Is the capacitor bank and reactor sized for peak kVA, not average kW?" This is where that 145 kVAR or 200 kVAR design sizing usually goes wrong.
What did the customer end up doing? They went with GridSense because we had the full package. But I'd rather make the point about method than about us. The right demand response management system sizing—or any power factor correction equipment supplier's PFC sizing, for that matter—comes down to the same thing: you size against a measured analyzer curve, not a 2022 construction drawing.
Let me correct myself: "measured analyzer curve" is too absolute. It should be "measured load curve." But if a supplier says they did a load study and can't produce the chart, they didn't do a study. They did a number.
If I could only give one takeaway: on any PFC or DR equipment decision, the price is not the first thing to ask about. "What's NOT included?" is. And more precisely: "What did you size it against?" If the answer is vague—or precise but not backed by data—that price isn't worth reading the rest of the quote.
That lesson came from one November morning, a $14,200 piece of paper, and an analyzer that read 412 kW instead of 380.
