Energy charges and demand charges are different things
An industrial electricity bill generally separates two quantities. The energy charge is applied to how much electricity was consumed over the billing period, measured in kilowatt-hours. The demand charge is applied to how fast electricity was being consumed at the site's highest point, measured in kilowatts or kilovolt-amperes.
The distinction matters because the two respond to different actions. Reducing total consumption reduces the energy charge. It has no necessary effect on the demand charge, because a site can consume less overall and still record the same instantaneous peak.
How billed demand is usually determined
Utilities meter demand as an average over a fixed interval, commonly 15 or 30 minutes, and the billed demand is normally the highest of those intervals within the billing period. A single interval can therefore set the demand charge for the entire month.
Some tariffs go further with a ratchet clause, where the billed demand for subsequent months cannot fall below a defined percentage of the highest demand recorded over the preceding twelve months. Under a ratchet, one abnormal interval can influence charges for a year. Time-of-use structures add another dimension, applying different rates depending on when the peak occurs.
The exact definitions vary by utility and by rate class. Reading the applicable tariff document is not optional detail — it determines what an intervention would actually be worth.
Why industrial sites produce sharp peaks
Heavy industrial loads are rarely steady. Starting large motors, crushers or mills, coincident startup of multiple drives after a shift change or an outage, batch process steps, compressor and pump cycling, and ordinary seasonal or production variation all create short excursions well above the site's average load.
These excursions are usually a small fraction of total energy consumed but can define the demand charge. That asymmetry — brief events with a disproportionate billing consequence — is the reason peak demand management is treated as a distinct engineering topic rather than as part of general energy efficiency.
What peak shaving actually does
Peak shaving means supplying part of the load from a source other than the utility connection during the intervals that would otherwise set the peak. A battery energy storage system charges during periods of lower demand and discharges into the site load when demand approaches a defined threshold, so the metered import stays below that threshold.
The process consists of measuring and forecasting site load, setting a target demand threshold, discharging when load approaches it, and recharging when doing so will not create a new peak. Recharging is not incidental: charging at the wrong moment can create the very peak the system was installed to avoid.
- Interval load monitoring at the point of supply
- A defined demand threshold derived from the tariff and the load profile
- Automated discharge control fast enough to respond within the metering interval
- Controlled recharge scheduled to avoid creating a new peak
Sizing: power and energy are separate requirements
Two independent quantities determine a storage system for demand management. The power rating, in kilowatts, must cover the difference between the actual peak and the target threshold. The energy rating, in kilowatt-hours, must cover that power for the full duration of the peak event, including the longest event that must be managed and any events that occur close together.
A system that has enough power but insufficient energy will hold the threshold and then run out mid-event, and the peak is recorded anyway. A system with ample energy but insufficient power cannot hold the threshold at all. Both ratings follow from the same interval data: how high the peaks are, how long they last, how frequently they recur and how much recovery time exists between them.
The control system decides whether it works
Demand management is a controls problem as much as a hardware problem. The energy management system must see site demand in real time, anticipate the interval average rather than react to the instantaneous reading, dispatch the battery before the interval average is already committed, and manage state of charge across the day so capacity is available when it is needed.
It also has to coexist with the rest of the plant: interlocks, existing SCADA or PLC systems, generator control if present, and any protection scheme at the point of common coupling. Where storage is also used for backup or renewable integration, those objectives compete for the same state of charge and the priority order has to be defined explicitly.
When peak demand management is worth assessing
The characteristics that make a site a plausible candidate are consistent: a demand charge that forms a material share of the bill, a load profile with peaks that are short relative to the billing interval count, peaks that are repeatable and identifiable rather than random, and a connection point where storage can be installed and interfaced without disproportionate electrical works.
Conversely, a site with a flat load profile, a tariff with little or no demand component, or peaks that last many hours will usually see limited benefit from peak shaving, and the assessment should say so.
What is required to evaluate a site
A useful assessment starts from data, not from a proposal. The minimum set is twelve months of interval load data where available, the applicable utility tariff including any ratchet or time-of-use terms, recent bills showing how demand was billed in practice, the single-line arrangement and available connection points, details of any on-site generation, the controls environment, and the physical space and site conditions where equipment could be installed.
From that, the peaks can be characterised, a credible threshold established, the power and energy requirement derived, and the value of the intervention estimated against the tariff as written rather than against a general assumption.
Related applications
Peak demand management rarely stands alone. The same storage asset is often assessed for load shifting under time-of-use pricing, for supporting defined critical loads through interruptions, for absorbing variable renewable output, and — on remote or islanded sites — for reducing generator run hours and fuel consumption. Each objective changes how the system is sized and how the control strategy is prioritised.