A peak-shaving calculation needs two inputs: a time-series load profile and the electricity payment terms. The first determines the power and energy a BESS must provide. The second determines whether doing so reduces the bill. The worked example below uses hypothetical values that can be changed in the interactive calculation.
Step 1. Determine support power
Assume site demand rises to 180 kW and the desired grid-import limit is 140 kW. The BESS supplies the difference: 180 − 140 = 40 kW. This is the steady peak-support requirement. Short motor-starting events need a separate assessment using measurements and the PCS overload characteristic. Billing-interval averages may hide these events.
Step 2. Count energy between charging windows
The peak lasts 30 minutes, or 0.5 hours. Required AC output energy is 40 × 0.5 = 20 kWh. If two such peaks occur before recharging, the requirement becomes 40 kWh. Nominal DC capacity must account for the usable SOC window, discharge efficiency, ageing and emergency reserve. Do not substitute round-trip efficiency for the efficiency of the discharge path alone.
Step 3. Check reserve recovery
At an assumed round-trip efficiency of 90%, restoring 20 kWh of delivered energy requires about 22.22 kWh of charging energy. At 10 kW charging power, the idealised duration is about 2.22 hours; actual constraints can extend it. Check site demand plus charging power throughout that period. Charging must not create a replacement peak above the agreed limit.
Step 4. Translate the physical peak into money
With illustrative prices of UAH 4/kWh for charging and UAH 8/kWh for displaced purchases, the energy-cost difference per event is 20 × 8 − 22.22 × 4 ≈ UAH 71.11. This is before maintenance, degradation and capital costs. At the same UAH 4 price for both periods, the result is negative: approximately −UAH 8.89 because of losses. Add demand-charge savings only if the actual contract contains such a charge and its billed maximum is reduced.
Step 5. Replay the full period
Do not multiply one convenient day's result by a year. Replay shifts, weekends and seasons with SOC limits and available charging windows. Compare storage with rescheduling controllable loads. An enquiry to USK Energy should include the profile, tariff, import limit and required reserve. Compare PCS, batteries, EMS, metering and other supplied components separately in the offer. Use the same assumptions for each supplier so that the comparison remains reproducible.
Decide which maximum is being limited
A momentary starting event, average power over a measurement interval and the contractual billing maximum are different quantities. If the task concerns a fifteen-minute average, the energy model needs matching intervals. That does not establish whether the PCS can support a brief current surge; starting behaviour requires separate measurements and equipment evidence. One hourly file should not be used to claim that both tasks have been verified.
Define where the limit is measured. A controller monitoring only part of the facility may regulate that part correctly while failing to limit the main supply. The diagram should identify included loads, solar generation and auxiliaries. Agree power signs and timestamps before calculating a battery size.
Also distinguish the target from the existing connection capacity. The selected operating target may be lower, but it must be justified by the actual task. Changing it by a few kilowatts can substantially change required discharge energy when many intervals sit just above the target. Keep it as an explicit model input.
Validate the file before selecting equipment
Each row needs a timestamp, interval duration and value with a known unit. If the file contains kWh accumulated over fifteen minutes, divide energy by 0.25 hours to obtain average kW. For example, 35 kWh over that interval means 140 kW, not 35 kW. If the supplied value is already in kW, dividing again creates an error.
Check missing rows, duplicates, time zones and calendar changes. Do not replace a missing observation with zero: that creates a false charging opportunity. Keep a data-quality flag and explain how doubtful sections are treated. Include normal shifts and known demanding days without assuming the demanding day repeats throughout the year.
As a basic check, total the file's energy and reconcile it with available metering over the same period and boundary. Explain discrepancies before using the data for quotations. A profile may already contain the action of an existing battery or generator; in that case it is not the underlying load alone. Record that distinction rather than sizing a new battery against a partially controlled signal.
A six-interval example
Consider an illustrative ninety-minute sequence with fifteen-minute steps and a 140 kW import target. For clarity there is no solar generation, no additional auxiliary load and no charging-converter restriction below available connection headroom. Required battery support is demand above the target. Charging headroom is the positive difference between the target and demand.
| Interval | Load, kW | BESS support, kW | AC output, kWh | Possible AC charging input, kWh |
|---|---|---|---|---|
| 1 | 100 | 0 | 0 | 10 |
| 2 | 160 | 20 | 5 | 0 |
| 3 | 180 | 40 | 10 | 0 |
| 4 | 120 | 0 | 0 | 5 |
| 5 | 200 | 60 | 15 | 0 |
| 6 | 100 | 0 | 0 | 10 |
Maximum support is 60 kW and total AC discharge is 30 kWh. Available AC charging input is only 25 kWh. Repeating this sequence without another charging window reduces stored energy. A battery cannot repeatedly deliver more energy than it receives while also covering conversion losses.
The table is not a complete equipment specification. It isolates the arithmetic so each requirement can be checked. A real dataset must add the site's sources, auxiliaries, battery limits and control constraints using consistent boundaries.
Carry battery state through every interval
For illustration, assume separate charging and discharging efficiencies of 95%. These are hypothetical values, not a product specification. Charging adds 25 × 0.95 = 23.75 kWh on the DC side. Delivering 30 kWh removes 30 / 0.95, approximately 31.58 kWh. The net change is about minus 7.83 kWh. Under these assumptions, round-trip efficiency is 0.95 × 0.95 = 90.25%, not 95%.
If the initial available DC energy above the agreed lower boundary is 50 kWh, and the upper boundary does not restrict charging, about 42.17 kWh remains at the end. But the endpoint alone is insufficient. The minimum within the sequence occurs after interval five, at approximately 32.67 kWh. Every interval must respect the limits even when subsequent charging restores energy.
Check the upper boundary too. A nearly full battery at the beginning cannot accept all the proposed 10 kWh of AC charging input. The final balance would then be lower than calculated above. The model must restrict charging to actual free capacity instead of allowing energy to accumulate beyond the agreed operating window.
Initial charge should also have a reason. It may come from a previous day or an explicitly scheduled preparation period. Beginning each simulation with a full battery without accounting for that charge gives the controller an advantage that real equipment does not receive.
PCS power and battery energy are linked but distinct
The example requires at least a documented 60 kW of AC support under the relevant conditions. That is not yet a final product nameplate rating. Check temperature, SOC, reactive-power requirements and other agreed limits. A PCS with sufficient AC capability cannot deliver the full result if the battery assembly does not permit the required DC current.
Avoid substituting an arbitrary percentage margin for a proper check. State what each allowance covers: confirmed load uncertainty, a future group or documented derating. Energy capacity comes from chronological demand, protected reserve and the accepted battery condition at the end of the operating period. Higher converter power does not automatically supply enough stored kilowatt-hours.
Where reactive power is required simultaneously, ask for the applicable active-power capability rather than treating the converter's apparent-power rating as an unrestricted kW value. The supplier should confirm the particular operating point. This prevents a nominally suitable unit from being selected on a different rating basis.
Charging must not create another peak
At each step, limit charging to the smallest permitted value: PCS capability, battery capability, free capacity and connection headroom. With a 120 kW load and a 140 kW target, only 20 kW remains. A 40 kW charging command would create 160 kW of import and defeat the original objective.
For changing loads, metering, communications delays and the control algorithm matter. An ideal interval model does not establish that no short overshoot will occur in real time. Describe the required regulation and verification criterion in the EMS enquiry. Suppliers and responsible contractors need the same understanding of permissible deviation and measurement boundary.
Define behaviour when measurements or communication are unavailable. A schedule based on stale demand should not be assumed to maintain the target. The required fallback is a site-specific functional requirement to be agreed and demonstrated, rather than an undocumented expectation added after purchase.
Move from the energy balance to money
Calculate actual electricity purchases before and after BESS for each tariff interval. The difference includes extra charging and reduced purchases during discharge. With the same electricity price throughout, losses may increase the energy bill even while the technical maximum falls. That is not necessarily a calculation error: limiting power and reducing energy expenditure are different results.
Add reductions in other payments only where the contract contains the relevant component and the billing condition actually changes. If the chargeable maximum remains in another interval, successfully shaving one peak may not reduce that payment. Include the complete equipment package, auxiliaries, support and agreed replacements in the investment assessment.
Do not turn one favourable day's difference directly into a payback period. Seasonal production, holidays, maintenance and changes in the tariff schedule can alter annual benefit. Retain the chronological annual model or a justified representation of operating periods and explain its limits.
Test the inconvenient days
Check consecutive peaks, an extended shift, a partly charged starting battery and restricted recharge opportunities occurring together. Another scenario should preserve the agreed outage reserve. When the target is missed, report the intervals and cause: insufficient power, energy, charging headroom or equipment availability. This identifies a useful specification change more clearly than one aggregate success percentage.
Compare storage with rescheduling controllable operations, then test the combined option. Moving part of a process can reduce the battery package required, provided the production consequences are acceptable. Keep the original profile, formulae and requirements version together.
For USK Energy, a useful calculation output is not just desired kW and kWh. It also identifies the intervals that determined those values, the energy boundary, the recharge assumptions and the duties the system must support. Those details allow suppliers to propose comparable equipment and expose a configuration that meets a headline rating but fails the actual sequence.
As a sensitivity check, the same profile with a permitted 160 kW target needs support only at 180 and 200 kW. Maximum support becomes 40 kW and total AC discharge becomes 15 kWh. This does not authorise a higher import limit; it shows how strongly equipment requirements depend on the chosen feasible target. Recalculate recharge and economics rather than carrying the previous result across unchanged.
Follow the battery through six intervals
Change the inputs and save the result to discuss the equipment package.
BESS settings and time step
| Step | Load, kW | AC charging, kW | AC delivery, kW | DC remaining, kWh | Above limit, kW |
|---|---|---|---|---|---|
| 1 | — | — | — | — | — |
| 2 | — | — | — | — | — |
| 3 | — | — | — | — | — |
| 4 | — | — | — | — | — |
| 5 | — | — | — | — | — |
| 6 | — | — | — | — | — |
Next-step energy = previous energy + AC charging × t × charging efficiency − AC delivery × t / discharge efficiency. Charging and discharge respect power limits and the usable DC window above protected reserve. An illustrative profile without PV or auxiliary demand. Exceeding the target means additional grid import, not automatic load disconnection.
Before requesting a quotation
Mark questions for which you already have a confirmed answer. Save the list and send unresolved points to the supplier.
Why the tariff spread is not the whole calculation
Compare costs for the same delivered service: energy reaching the load and the share of demand actually supported. Charging prices apply to input energy; savings apply to output energy. Conversion losses, auxiliary consumption and reserve constraints sit between the two. Demand-charge savings are a separate benefit only when your actual supply contract includes that charge.
Change charging cost, displaced energy cost and cycle efficiency in the example below. A negative result is useful: the tariff scenario does not even cover energy losses. A positive result is not a payback forecast. Maintenance, degradation, financing and other costs still need to be deducted before comparing the remaining benefit with the investment.
Further verification
From calculation to equipment
Preparing an equipment purchase? Send the load, backup duration or specification. USK Energy will clarify the configuration, compatibility and delivery terms.




