Peak shaving reduces the highest level of electricity a site draws from the grid during relevant measurement periods. A business can do this by limiting certain loads, changing operations or supplying part of the load from an on-site resource such as a battery. The objective is to lower a demand peak, which is different from reducing total electricity consumption.
That distinction matters because electricity bills can charge for both energy used and demand. Lowering one does not automatically lower the other.
Kilowatts describe the peak; kilowatt-hours describe energy
Demand is a rate of electricity use, expressed in kilowatts. Energy is the amount used over time, expressed in kilowatt-hours. A short period when several large loads run together can produce a high demand reading even if the facility’s monthly energy use is modest.
DOE’s battery implementation overview identifies peak shaving as a storage application aimed at reducing peak demand. Whether that creates a bill saving depends on the site’s actual tariff: measurement intervals, billing periods, seasonal charges and any demand ratchets or minimums must be examined.
Do not assume a business’s highest instantaneous reading is the number used on its bill. The utility may calculate demand over a defined interval and apply additional billing rules.
How peak shaving works
A controller can monitor the site’s grid import and act when demand approaches a chosen target. It might reduce EV charging, defer a flexible process or discharge a battery to supply the difference between building load and the target grid draw.
An illustrative example: a facility draws 300 kW for 30 minutes, and a battery supplies 50 kW throughout that period. Grid import would be 250 kW during the event, assuming no other changes. Supplying that contribution requires 25 kWh delivered on the AC side:
50 kW × 0.5 hours = 25 kWh.
The installed battery needs sufficient usable energy after the relevant losses and reserves, plus adequate output power. This example does not establish savings. If another unaddressed event sets the billing peak, the expected demand-charge reduction may not occur.
What makes a site a useful candidate?
Start with interval data and the tariff, not a battery size. The shape and repeatability of the peak determine whether a strategy can work.
| Pattern | Question to investigate |
|---|---|
| Brief, predictable peaks | Can a controlled resource cover the full billing interval? |
| Long daily plateaus | How much stored energy would sustained reduction require? |
| Irregular process starts | Can controls respond without disrupting production? |
| New charging demand | Can vehicle charging be managed before adding storage? |
| Several peaks close together | Is there time and capacity to recharge between events? |
The DOE storage evaluation method provides a framework for evaluating system performance. An economic assessment must add the site’s bills, dispatch assumptions and costs rather than treating technical performance as proof of financial benefit.
Peak shaving versus load shifting
Peak shaving targets maximum demand. Load shifting changes when electricity is used or imported. The two can overlap: moving charging away from a busy afternoon may lower a peak as well as change timing. But shifting energy into a new crowded period can create another peak.
Charging a battery also adds electrical demand. The dispatch plan must account for charging, not just discharge. A strategy that trims the afternoon peak but creates a larger overnight billing peak can defeat its purpose.
Storage introduces operating trade-offs
A battery reserved for outage backup may have less energy available for demand management. Cycling also affects operation and warranty considerations. The assessment should explain how backup reserve, demand reduction and other uses are prioritized.
Ask for a comparison with operational measures. Staggering equipment starts or managing flexible loads may address part of the issue without a battery. The appropriate solution depends on whether those loads can change without harming comfort, production or service commitments.
Soleil’s commercial battery storage guide provides a next step for Maryland businesses evaluating storage in a broader project context.
Start by identifying the billed demand quantity
A demand-management project needs to target the quantity that actually affects the bill. The site’s highest observed power, the highest interval average and the demand used for billing may differ. The applicable tariff defines the commercial treatment, so the analysis should begin with that document and the customer’s account information.
An owner can ask the analyst to point to the bill line being addressed and explain how it is calculated. If the model reduces a plotted peak but does not reproduce the relevant billing treatment, its financial conclusion is incomplete. A clear explanation should connect the interval data, billing rule and modeled result.
Check whether the analysis reflects the site’s actual customer class and current arrangement. A tariff found for a nearby business may not apply to the subject account. Likewise, a historic bill may not describe a future arrangement after a significant service or operating change.
This step prevents an attractive control concept from being evaluated against the wrong commercial problem. The technical ability to reduce demand matters, but the bill must recognize that reduction for the proposed saving to occur.
Interval data reveals the duration of the problem
A monthly bill can show that demand is costly without showing how long each peak lasts. Interval data adds the time dimension needed to evaluate a control strategy. The analysis should examine the relevant measurement period and enough operating history to capture meaningful variation.
A short, repeatable event may require a different response from a long plateau. If a battery must supply a reduction for several hours rather than a few minutes, the delivered energy requirement changes substantially even when the target power reduction is the same. This is why sizing from one maximum kW value is inadequate.
Check data quality before relying on the pattern. Missing intervals, changed meter identifiers, clock inconsistencies or an unusual closure can distort the picture. The analyst should explain which records were used and how uncertain periods were handled.
Also speak with the people who run the facility. A spike may correspond to an identifiable process, while an apparently regular pattern may disappear after a planned equipment change. Operating context gives the data a useful meaning.
Distinguish flexible loads from essential loads
Some electricity use can move in time without affecting the facility’s purpose. Other loads have firm operational constraints. A peak-shaving strategy should identify that difference before proposing curtailment or rescheduling.
For example, an illustrative site might have EV charging sessions with several hours of parking time, alongside a production process that cannot be interrupted safely or economically. Both contribute to demand, but they do not offer the same flexibility. The control plan should respect the actual service requirement rather than treating every kilowatt as interchangeable.
Define the conditions under which a load can change: the permitted duration, recovery time, user commitment and relevant operating authority. A nominally flexible load may be unavailable during a particular event. Model that limitation instead of assuming perfect participation.
The resulting strategy may combine measures. Managing a flexible load can reduce part of the peak, while a battery covers another portion. The combined design should be evaluated as one operating plan so that the resources do not work against one another.
Battery power sets the reduction; energy supports its duration
The battery’s output capability determines how much demand it can offset at a given moment within its operating limits. Available energy determines how long that contribution can continue. A system needs both capabilities for the intended event.
Consider an illustrative facility with a target reduction of 40 kW. Maintaining that reduction for 15 minutes requires 10 kWh delivered to the relevant AC boundary; maintaining it for two hours requires 80 kWh. These figures are simple energy arithmetic, before determining the installed capacity needed for losses, reserves and permitted operation.
The comparison shows why two sites with the same desired kW reduction can need very different storage arrangements. It also shows why a proposed battery cannot be evaluated from its energy capacity alone. The output and duration must match the actual demand event.
A useful model records events the system cannot fully cover. Those unmet events may determine the remaining billing demand. Hiding them behind an average annual performance figure can overstate the practical value of the strategy.
Recharging can create a second demand problem
Discharged energy must be replenished if the battery is to respond again. Recharging adds load at its connection point, so it belongs in the same demand analysis as discharge. An otherwise effective strategy can create a new peak if charging is scheduled without regard to the facility’s other demand.
The controller needs an appropriate charging plan and operating limits. The proposal should explain how it restores available energy, what happens when another event occurs before recharge is complete, and how a backup reserve affects the usable dispatch range.
An illustrative warehouse might experience a morning peak, a later charging opportunity and an afternoon peak. The system’s ability to address both events depends on the energy available after the first event and the permitted recharge between them. Looking at each peak independently can make the combined requirement appear easier than it is.
Ask the analyst to show the complete charge and discharge schedule alongside the resulting grid demand. That view makes trade-offs visible and helps identify whether the control merely moves the problem.
Forecasting and operating limits belong in the assessment
Some demand events are predictable; others are not. A strategy based on an expected schedule can differ from one that responds to measured demand. The chosen approach must fit the site, available data and control capability, with clear treatment of uncertainty.
Owners should ask what information the controller uses and what happens when that information is missing. A design should explain the relevant fallback behavior rather than assuming that every measurement and communication path remains available. The appropriate behavior depends on the equipment and operational requirements.
Operational limits should also be visible in the financial model. If the battery preserves a minimum reserve, cannot provide the requested output under a particular condition or is unavailable during maintenance, those constraints affect demand reduction. The model should not assume more flexibility than the installed system provides.
DOE’s BESS evaluation method supports evaluating measured storage performance. For an owner, that evidence becomes most useful when it is connected to the dispatch objective and the actual billing calculation.
Evaluate the result against a defensible baseline
After installation, a lower bill does not automatically prove the control caused the difference. Weather, production, occupancy, operating hours and tariff changes can all affect demand and cost. A performance review needs an appropriate comparison that identifies the factors being held constant or adjusted.
A useful report can show when the resource operated, its contribution during relevant intervals and the resulting grid import. It should distinguish measured quantities from modeled estimates of what would have happened without control. Those categories should not be blended into an unexplained savings total.
For an illustrative factory that reduces production during the same month a battery enters service, comparing only the previous bill with the new bill could credit the battery for a change caused partly by production. The evaluation needs operating context to avoid that mistake.
Agree on the performance evidence before purchase. If the business expects a particular analysis, specify the data access and reporting necessary to support it. A dashboard showing battery activity is not automatically a verified savings report.
Test the business case against changing assumptions
A peak-shaving investment can depend on demand charges, operating patterns, equipment costs and achievable control. Sensitivity analysis shows how the result changes when those assumptions differ from the base case. It is more informative than presenting one precise payback period as certain.
Ask which assumptions matter most. A planned expansion could make the historical profile unrepresentative. A change in charging demand could alter the value of load management. A different electricity arrangement could change the bill component the project was designed to reduce.
Include ongoing responsibilities as well as equipment cost. Controls, monitoring, maintenance and the effort needed to keep the operating strategy aligned with the facility have practical implications. The assessment should identify who owns those tasks rather than assuming the system will manage itself indefinitely.
The strongest proposal explains both why the strategy is promising and the conditions under which its benefit weakens. That gives the owner a basis for deciding whether to proceed, revise the scope or first address the peak through operational changes.
Questions before approving a project
Will peak shaving reduce my energy use?
Not necessarily. It can reduce demand while leaving energy use similar. A battery introduces losses, so total purchased energy can increase even if a demand-charge component decreases.
Can solar alone shave peaks?
It can reduce grid demand when production coincides with the relevant peak. Cloud cover, timing and the billing method determine how dependable that contribution is.
How much history is useful?
Gather available interval data and bills covering seasonal and operational variation. Identify upcoming equipment or schedule changes; historical peaks may not represent the future facility.
What should a proposal show?
Ask for the tariff assumptions, modeled demand before and after control, charging behavior, unmet events and sensitivity to changing operations. Require savings and backup benefits to be explained separately.
Start with the demand profile
For a commercial storage discussion, bring interval data, recent bills and planned load additions to Soleil Energy’s battery storage team. Those records help establish whether peak shaving is an appropriate project objective.
Project requirements depend on the site, equipment and applicable rules. Confirm design and installation decisions with the appropriate qualified professional.
