A demand charge is not an energy charge
A demand charge applies a tariff price to a measured or billing demand quantity. The measurement window, kW or kVA unit, ratchet, season, and billing determinant can differ by utility and class.
Toronto Hydro's current page is one named 2026 example. Its General Service 50-999 kW class lists distribution demand in CAD per kVA per 30 days and transmission demand elements tied to kW. It is not an Ontario-wide tariff.
- OfficialToronto Hydro's 2026 General Service 50-999 kW page distinguishes a distribution charge stated per kVA per 30 days from transmission demand elements tied to kW.Verified 2026-07-17
Power sets the height and energy sets the duration
Battery power limits how far the load can be reduced at an instant. Usable energy and round-trip efficiency limit how long the reduction can continue. A dispatch target that exceeds either bound must be reduced or suppressed.
The signature diagram shows a planned cap crossing a single load peak. Real intervals can contain several peaks, rebound charging, and tariff-specific billing windows.
Evidence needed before annualizing
Use interval demand for the billing period, the exact tariff sheet, battery power and usable energy, efficiency, state-of-charge limits, operating reserve, cycling limits, and the dispatch rule.
Confirm whether billing demand is kW, kVA, coincident demand, non-coincident demand, or a ratcheted value. Preserve those units through the calculation.
Measure the billed result, not the control signal
A controller may hit a target while the billed determinant remains unchanged. Measurement and verification should compare the utility interval and billing calculation with the counterfactual load, then account for recharge and losses.
Enerwav never annualizes a storage saving from annual load alone.
Identify the billed demand determinant
A demand charge is based on a measured demand quantity defined by the controlling tariff. The unit may be kW or kVA. The interval, billing period, peak definition, minimum demand, ratchet, seasonal rule, power-factor treatment, and applicable service class can change the amount. The customer's bill and tariff sheet are therefore primary inputs.
Toronto Hydro's business-rate page is a named example because it distinguishes demand-related charge units for a stated service class. It is not an Ontario-wide tariff and it is not loaded as an Explorer default. A site served by another distributor or another class needs its own charge labels, units, effective date, and billing determinants.
Rebuild the billed peak from interval data
The analyst should obtain interval demand covering enough billing periods to show recurring peaks, seasonal operations, shutdowns, production changes, and data gaps. The interval length must match or be reconciled to the tariff measurement. The bill should be used to check whether the reconstructed maximum agrees with the billed determinant.
A single annual maximum is not enough for dispatch design. Storage may need to respond to several peaks in one day, a long process plateau, or a peak that repeats before recharge. The interval record should identify event start, event duration, ramp, frequency, surrounding load, operating cause, and whether the event can be changed operationally without storage.
Power clips height; energy sustains duration
Battery power limits how much demand can be reduced in an interval. Usable energy limits how long the reduction can continue. Controls, response delay, state of charge, reserve, efficiency, temperature, auxiliary load, warranty, and inverter limits can reduce the simple nameplate capability. A target below the power-limited residual peak is not feasible.
The illustrative load profile makes this boundary visible. The declared battery can hold the planned cap during several intervals but reaches its power limit at the highest interval, leaving load above the cap. The example computes only interval dispatch and energy. It does not annualize savings or imply that the equipment size is suitable for a real site.
Test tariff ratchets and recharge consequences
A tariff can preserve part of an earlier peak through a ratchet or minimum billing demand. In that case, clipping one current peak may not reduce the billed determinant by the same amount. The model should calculate the charge rule before assigning savings. It should also check whether another meter, season, or transmission determinant sets a separate peak.
Recharge creates load and may create a new peak if controls respond only to the original event. The dispatch study needs a recharge window, maximum charge power, energy price, round-trip efficiency, reserve requirement, and forecast of later events. Solar charging can help only when production, site load, and available storage capacity align.
Define measurement and verification before procurement
The owner should state how baseline demand, target demand, battery dispatch, state of charge, meter data, tariff changes, outages, and operational overrides will be recorded. The method should distinguish a lower peak caused by storage from one caused by weather, reduced production, maintenance, or a process change.
A useful monthly review compares the modeled event set with actual dispatch and the final billed determinant. Missed events, false triggers, unavailable capacity, power limits, recharge peaks, communication failures, and tariff changes should be recorded. This evidence supports controller tuning and prevents a forecast value from being reported as achieved savings.
Know when the screen should stop
A project should remain at review-required status when interval demand is unavailable, the tariff class is uncertain, the demand unit is unclear, the peak is too long for the declared energy, the peak repeats before recharge, or the charge is controlled by a ratchet the model has not represented.
Operational changes may be cheaper or more reliable than storage for a controllable process peak. Solar may reduce a daytime peak but cannot be credited without coincident production evidence. Storage may serve resilience or energy-shifting objectives while providing little demand value. The decision should name the primary objective and keep secondary value conditional.
Use the named tariff sample only to explain units
The Toronto Hydro example demonstrates why a model must preserve kVA and kW rather than converting every demand-related line to one unit. The analyst should copy the exact current charge label, unit, service class, and effective date only when that tariff controls the account. The example does not establish the customer's billed peak or avoided amount.
If apparent power controls a charge, a real-power reduction may not produce the same billed reduction because power factor also matters. If a transmission charge uses a different peak definition from distribution, each determinant needs a separate baseline and dispatch test. The storage screen should suppress any line whose measurement rule has not been represented.
Taxes, riders, fixed amounts, and energy charges should remain separate from demand savings unless the tariff explicitly connects them to the same determinant.
The calculation record should cite the tariff page and the customer's bill period.
Retain both records.