Peak shaving limits the power a site draws from the grid during demand peaks. The BESS supplies the difference between demand and the chosen import limit. Whether this reduces the bill depends on the contract; sometimes the main benefit is staying within available connection capacity.
A peak has height and duration
Illustrative example: demand is 180 kW, the desired import limit is 140 kW and the peak lasts half an hour. Support must provide 40 kW and 20 kWh at the output. This is not yet a battery nameplate rating: allow for losses, reserve and discharge limits. Motor starting requires a separate check of PCS capability.
Find the charging window
If demand stays close to the limit all day, the reserve may recover slowly. Check consecutive peaks rather than only the highest instant. A monthly bill cannot replace a time-series profile with suitable resolution. The charging schedule must also respect the same connection constraints.

Identify the financial effect
Do not assume a maximum-demand charge exists if it is absent from your contract. Assess avoided connection expansion, tariff shifting and other evidenced benefits separately without counting the same energy twice. An equipment enquiry needs the load profile, import limit, peak durations and permitted charging windows.
Define the peak before selecting storage
A facility can use the word peak for very different events: a motor starting, compressors running together for ten minutes, or several hours of maximum production. These events require different power-to-energy ratios. Hourly data may hide a short starting event, while the highest instantaneous reading does not describe the energy needed through a long production peak. Start by stating the limit that must be maintained and the period over which it applies.
The reason for limiting power matters as well. A business may want to remain within an agreed connection capacity, avoid overloading a particular network component or reduce a specific charge on its bill. These are different economic cases. Do not include savings from a maximum-demand charge unless that mechanism exists in the actual supply contract. Conversely, enabling several production lines to run together may have value even without a separate peak-demand tariff.
Calculate the excess-load profile
For each time interval, subtract the chosen limit from facility demand. Replace negative differences with zero: spare capacity is not automatically energy that can be used for charging. The largest positive difference indicates the required power support. Summing each positive difference multiplied by its duration gives the energy delivered to the load. Check separately that the measurement resolution is adequate for the task you are trying to solve.
Consider a hypothetical 250 kW limit. Demand is 310 kW for twenty minutes and then 280 kW for forty minutes. Storage support is 60 kW followed by 30 kW. Required delivered energy is 60 × 20/60 + 30 × 40/60 = 40 kWh. The system needs enough power for the first interval and enough energy for both. Selecting equipment from the 40 kW average excess would overlook the higher power requirement at the start.
The resulting 40 kWh is not the battery's nameplate capacity. The next step accounts for the usable SOC window, losses at the defined boundary, any retained emergency reserve and available capacity at the agreed end-of-life condition. Avoid adding unexplained contingency several times in different rows. A comparable proposal should show the energy required by the task and the assumptions used to convert it into a physical battery configuration.
Repeated peaks can be harder than one large event
Look at the time available before the next peak. If the factory operates close to its import limit after the first event, there may be little capacity available for recharging. Permitted charging power is constrained by the gap between the import limit and current demand, the charging equipment and battery limits. Overnight charging helps only if the stored energy can cover all the daytime events that occur before another useful charging opportunity.
In the example, energy must be restored after delivering 40 kWh. With an illustrative effective round-trip efficiency of 90%, the grid input needed to repeat the same delivery is about 44.4 kWh. If only 15 kW of spare capacity is available for two hours, the input is limited to 30 kWh. That schedule does not restore the starting reserve. Buying a more powerful PCS does not resolve a shortage of available charging energy.
Evaluate at least two consecutive difficult days. A model beginning every morning with a full battery can hide a persistent energy deficit. It is reasonable to define a planned recovery period, but that period must appear in the operating schedule and in the calculation. Where production does not leave such a window, discuss a different reserve target, additional generation or changes to the duty before increasing battery capacity automatically.
Control settings need explicit acceptance criteria
| Operating condition | What to establish before purchase |
|---|---|
| Rapid load increase | Response time of the complete measurement and control chain |
| Low SOC | Priority of retained backup energy relative to peak reduction |
| Missing meter data | Defined fallback, alarm and recovery behaviour |
| Solar generation changes | Charge and discharge response to a sudden generation reduction |
| Grid supply returns | Whether simultaneous recharging creates a new maximum |
An EMS import limit is not equivalent to a protective device. The controller manages power within its measurement and response capabilities; electrical protection serves a different function. Agree acceptable excursions, their duration and failure behaviour with those responsible for the electrical system. The supplier should substantiate the performance of the proposed package rather than promise an unspecified absence of every possible exceedance.
Check the input file before modelling
Review timestamps, missing records, units and power direction. kW and kWh per interval are different quantities. If a file contains energy for fifteen-minute periods, divide by 0.25 hours to obtain average power. Peaks within those intervals remain unknown; converting units does not recover the missing time resolution.
Mark outage days and non-working days separately. They need not automatically be removed, but their operating context should be understood. A large accumulated reading after a data gap may not represent a genuine load step. Ask the person responsible for metering to confirm the format before selecting equipment power.
Keep the original file, cleaned dataset and adopted scenario separately. Another supplier can then repeat the calculation, and the business can understand differences between offers. Request the most demanding intervals as well as final kW and kWh: BESS support, SOC before and after each event, and charging available between events. This reveals an implicit full-battery assumption more quickly than comparing price summaries.
Agree how missing data is treated and label estimated values as estimates. If the largest apparent peak falls in an incomplete period, obtain better information before ordering. That provides a concrete reason for additional measurement instead of increasing the entire package by an unexplained percentage.
Compare storage with changes to the operating sequence
Some peaks can be reduced by starting equipment sequentially, moving vehicle charging or using an existing thermal buffer. That does not necessarily remove the need for storage. It may allow a smaller battery to operate under a clearer duty. Compare the cost and impact on production output. A measure that reduces throughput or disrupts a process is not a free alternative simply because it avoids purchasing a battery.
For an enquiry to USK Energy, prepare profiles from several representative working days, the desired limit, recurring peak events and available charging windows. Include a day with the most demanding combination of loads, not just an average day. Request comparison by power support, usable energy, EMS behaviour and ability to repeat the cycle. The outcome should be equipment suited to your operating schedule, with assumptions that remain visible when production changes.
Keep the original profile with the accepted specification. It becomes a reference for later performance checks and expansion decisions. If the plant adds another shift or changes a compressor sequence, the original calculation can then be rerun on the new data. This is more useful than retaining only a quotation with battery kWh and a broad statement that the system will reduce peaks.
Check a single demand peak
Change the inputs and save the result to discuss the equipment package.
Pᴮᴱˢˢ = max(P − limit, 0); E = Pᴮᴱˢˢ × t. A rectangular peak without losses. Total energy across several peaks is daily delivery, not automatically battery capacity: recharge may occur between peaks. Procurement requires a time series and an SOC reserve.
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.




