A monthly statement is a summary of roughly 2,880 fifteen-minute intervals, of which exactly one set the demand charge. The summary tells you the number. It does not tell you which interval, what day, what time, what was running, or whether that interval was an outlier or the third-highest of forty similar ones.
Those questions decide whether peak management is worth attempting at your site and, if it is, which measure to buy. They are answered by interval data, and interval data is usually available at no cost.
What it is
Where advanced metering is installed — which is now most commercial and industrial service in the United States — the meter records consumption over each successive interval and stores it. The utility uses the resulting series to calculate the demand determinant. The same series is what you want.
A typical file has one row per interval, with a timestamp and a consumption figure in kWh for that interval. Multiply a fifteen-minute kWh figure by four to get the average kW during that interval, which is the quantity the demand charge is based on.
Getting hold of it
- Start with the utility's online business account portal. Many publish 15-minute or hourly data for download without any request at all.
- If it is not in the portal, email the account representative. Ask specifically for interval consumption data at the metering interval used for billing, for the maximum retained history, in CSV.
- Give the account number and the service point or meter identifier. A site with several meters needs each one separately, and asking for "the data" produces one meter at random.
- Ask whether the data is billing-quality or estimated, and whether any intervals are flagged.
- If the utility is unresponsive, look up your state commission's rules on customer access to usage data. Citing the rule usually resolves it faster than escalating.
- If a consultant or broker holds it on your behalf, ask them for the raw file, not a report built from it.
Two practical notes. First, ask for data at the billing interval, not whatever is convenient. Hourly data smooths away exactly the spikes a fifteen-minute demand charge prices, and analysis done on it will understate your peak and overstate how easy it is to manage. Second, get at least twelve months. Seasonal shape is the whole story at many sites, and a summer-only extract hides it.
The four things to calculate first
Once the file is in a spreadsheet, four calculations answer most of the strategic questions before any equipment is considered.
1. The monthly maxima, with their timestamps. Sort by consumption and take the top interval for each billing month. Then look at the timestamps. If they cluster — same weekday, same hour, same season — the peak is systematic and controllable. If they scatter, it is incidental, and controlling it is much harder and much less certain.
2. The gap between the highest interval and the tenth-highest. This is the single most useful diagnostic there is.
How much peak sits above the rest of the load
Rank every interval in the month and compare.
- Highest interval average1,412 kW
- Tenth-highest interval average1,268 kW
- (Gap between them)144 kW
- Value of an avoided kW-month, all riders included$19.05
Value of shaving down to the tenth-highest$2,743 / month
A wide gap means very few intervals are setting the charge, so a small, cheap, occasional intervention can capture most of the value. A narrow gap means the peak is the whole load and you are looking at a structural change instead. Figures illustrative.
That distinction decides the entire strategy. A wide gap points at scheduling, sequencing and short curtailments. A narrow gap points at storage, generation or a different tariff.
3. Load factor by month. Energy divided by peak times hours, calculated per month rather than annually. A load factor that collapses in one season tells you where to look.
4. The average day, by day type. Average each interval across all weekdays, then across all weekends, and plot both. The shape that emerges is the thing every control strategy on this site is trying to change, and seeing it is usually the moment the problem becomes obvious.
What the shape tells you
The measures worth considering follow directly from the shape:
- A single sharp morning spike points at simultaneous startup. The fix is sequencing, and it is nearly free: staggered startup.
- A broad afternoon plateau in summer points at cooling. Pre-cooling and setpoint scheduling apply, and thermal storage becomes worth costing if the plateau is large and predictable: HVAC scheduling against the peak.
- Irregular short spikes on a flat base point at batch processes colliding. Demand limiting controls suit this: demand limiting controls.
- A peak that is simply the whole plant running means there is no shape to fix. Look at choosing a different rate schedule or at storage: sizing a battery for peak shaving.
Using it to test a tariff before you switch
The second major use of interval data is modeling. With a year of intervals you can price the same consumption under any published tariff, because a tariff is a set of arithmetic rules and your intervals are the inputs.
This is the only honest way to decide whether to move between schedules. It is covered step by step in how to choose a rate schedule, and it is worth doing before any switch, because switching rules typically impose a minimum stay that makes a mistake expensive to reverse.
Keep it
Interval data is also the baseline for measurement and verification. If you install a measure and want finance to accept the savings, you need the load shape before as well as after, adjusted for anything else that changed. Starting that record now costs nothing and makes the eventual claim defensible: measurement and verification that finance will accept.
