A credible ESS payback study starts with the load profile
An energy storage system can charge when the site has room and discharge when demand approaches a costly peak.
That operating idea is simple, but the financial result is not determined by battery nameplate energy alone.
Actual peak reduction is bounded by PCS power in kW, usable battery energy in kWh, the usable state-of-charge window, round-trip efficiency, peak duration, degradation, availability, load-forecast error, and control execution.
A large kWh value cannot overcome an undersized PCS, while a large kW rating cannot sustain a long event without enough usable energy.
Monthly peaks also occur under different production, weather, and occupancy conditions, so one annual maximum is not a complete design basis.
Power in kW
Defines how much site load the system can replace at one moment.
Energy in kWh
Defines how long the selected power can be sustained.
Whole-life cash flow
Includes initial, recurring, inspection, degradation, and replacement effects.
This calculator places two EPC or vendor quotes on the same twelve-month load boundary and the same economic assumptions.
It then separates a requested peak target from the reduction that the entered power and energy can technically support.
Capacity fade, a planned replacement, recurring operations cost, demand-value escalation, and discounting are carried through annual cash flow.
The displayed amounts and performance values are fictional defaults that demonstrate the inputs, not market averages, recommended system sizes, tariffs, or guaranteed returns.
Replace every default with site interval data, bills, scope sheets, warranties, and written cost evidence before using the result in an investment review.
What the peak-shaving model calculates
Step 1 · Requested reduction by month
Requested reduction equals baseline peak minus target peak for each month.
A baseline of 1,000 kW and a target of 700 kW create a 300 kW request.
The tool requires twelve unique months and prevents a target from exceeding its baseline.
This monthly structure preserves seasonal production, cooling, heating, occupancy, and operating differences instead of applying one annual peak to every bill.
Step 2 · Deliverable energy and technical peak limit
Deliverable discharge energy equals rated kWh multiplied by retained capacity, usable SOC, and round-trip efficiency.
The energy-limited power equals deliverable kWh divided by representative peak duration in hours.
The technical peak limit is the lower of PCS rated kW and energy-limited kW.
A power constraint therefore identifies an undersized PCS, while an energy constraint identifies insufficient usable duration.
The model uses one representative duration for an MVP comparison and does not claim to optimize every 15-minute dispatch interval.
Step 3 · Expected reduction and demand value
Technical reduction is the lower of requested reduction and the technical peak limit.
Expected reduction multiplies that technical amount by the operating realization factor.
The realization factor can represent forecast error, control delay, maintenance downtime, conservative operations, and other differences between a modeled dispatch and actual delivery.
Monthly demand-value savings multiply expected kW by the entered monthly marginal value per peak kW.
That value is escalated annually by the common tariff-value change assumption.
Step 4 · Energy arbitrage and recurring cost
Annual discharged energy uses deliverable energy, operating days, equivalent full cycles per day, and realization.
Charging energy is larger than discharged energy when round-trip efficiency is below 100 percent.
Energy arbitrage equals discharged energy times the avoided rate minus charging energy times the charging rate.
The result may be negative when the rate spread does not cover conversion losses, and the model preserves that negative value rather than forcing a benefit.
Annual operations, maintenance, insurance, and inspection costs are escalated separately and deducted from cash flow.
Operating headroom equals the year-one technical peak limit minus the maximum requested reduction.
It is a planning margin for the entered load and technical assumptions.
Operating headroom is not fire-safety headroom and does not establish electrical, structural, interconnection, or emergency-response compliance.
Build an evidence pack before entering numbers
- Collect at least twelve recent electricity bills and the matching 15-minute interval load file.
Mark shutdowns, expansions, temporary construction loads, abnormal weather, or other periods that should not be treated as a normal baseline without adjustment.
Keep the original file and document each normalization decision so a reviewer can reproduce the baseline.
- Identify the interval that drives the relevant billing demand in each month.
Determine how long site demand remains above the intended target around that interval, not merely how long one isolated meter reading is high.
Use a representative duration that is conservative for recurring events, and run separate cases when summer and winter shapes differ materially.
- Establish a monthly marginal value for reducing one billed peak kW from the applicable tariff, contract, and actual bills.
Do not assume that a posted demand-charge number automatically equals the savings from one measured kW reduction.
Billing-demand ratchets, contract demand, power factor, seasonal windows, and other terms may change the realized value.
- Transfer rated kW, rated kWh, usable SOC, efficiency, annual fade, warranted cycles, and test conditions from each quote into one comparison sheet.
Ask vendors to restate values on equivalent AC or DC boundaries, temperatures, discharge rates, and end-of-warranty conditions when the proposals use different definitions.
A higher headline number is not better evidence when its measurement boundary differs.
- Reconcile initial scope beyond the battery and PCS.
Include engineering, civil and electrical work, transformers or switchgear, interconnection, communications, EMS integration, fire-safety systems, commissioning, training, permitting support, and internal labor where applicable.
Record every exclusion that one bidder expects the owner or another contractor to supply.
- Reconcile annual and periodic cost.
Include preventive and corrective maintenance, software, communications, monitoring, insurance, inspection, testing, augmentation, disposal, and planned replacement where supported by the project boundary.
Enter only confirmed written support or discounts attributable to the quote, not a possible future award.
Core formulas and financial interpretation
ESS peak-shaving formulas and decision meaning| Measure | Model structure | Decision use |
|---|
| Retained capacity | Rated kWh × (1 − annual fade) ^ battery age | Shows how the usable energy boundary changes before and after a planned replacement.
|
| Deliverable energy | Retained kWh × usable SOC × round-trip efficiency | Provides the conservative energy available to support the representative event.
|
| Technical peak limit | min(PCS kW, deliverable kWh ÷ event hours) | Identifies whether power or sustained energy limits the dispatch.
|
| Expected peak reduction | min(requested kW, technical kW) × realization | Converts a nameplate proposal into an economic delivery assumption.
|
| Energy arbitrage | Discharged kWh × avoided rate − charging kWh × charge rate | Exposes conversion-loss cost and preserves a negative spread.
|
| Annual net cash flow | Demand savings + energy arbitrage − recurring cost − replacement cost | Shows the replacement-year dip and annual economic contribution.
|
| NPV | Net initial outflow plus discounted annual cash flows | Compares two quotes over one equal study boundary.
|
Simple, first, and sustained payback
Simple payback divides net initial investment by year-one net cash flow and ignores later variation.
First payback marks the first cumulative crossing above zero, but a later replacement can make cumulative cash flow negative again.
Sustained payback is the crossing after which cumulative cash flow does not fall below zero again within the study period.
Discounted payback and NPV
Discounted payback applies the selected annual discount rate before testing recovery.
NPV includes net initial investment at time zero and every discounted annual cash flow over the selected horizon.
Apply the same discount rate, study period, inflation boundary, and replacement convention to both quotes.
The planned replacement cost occurs once in the entered year and is escalated by the recurring-cost change assumption.
The replacement-year cash outflow is included in that year, while battery age resets in the following year.
If replacement occurs in the final study year, the model includes the cost but cannot display later post-replacement benefits.
That result is a boundary warning, not a reason to omit the replacement.
Read several outputs together
Technical limit and operating headroom
A quote may meet the target in year one but miss it later as capacity fades.
Review the annual technical limit, the degradation warning, and the capacity-down sensitivity rather than relying on the first-year label alone.
A small positive margin also deserves a dispatch study using the actual load curve.
Break-even monthly value per peak kW
The break-even value is the monthly marginal demand value that makes study-period NPV equal zero under all other inputs.
Compare it with the value supported by the applicable bills and contract.
If the evidence-based value is lower, renegotiate scope, system price, target, operating strategy, or risk allocation.
Cycle utilization and replacement reserve reference
Equivalent full cycles equal operating days times cycles per operating day.
The maximum utilization measure compares each pre-replacement or post-replacement interval with warranted lifetime cycles.
The annual replacement-reserve reference merely divides planned replacement cost by replacement year and does not determine accounting treatment or an actual reserve contribution.
Capacity and demand-value sensitivity
Capacity-down isolates performance and availability exposure, while demand-value-down isolates tariff and billing exposure.
They are changed separately so one favorable assumption cannot hide the other risk.
A positive base NPV with a negative downside result should shift attention toward performance guarantees, measurement rules, and price protection.
Worked interpretation of two different quote shapes
Consider a site requesting a 300 kW reduction for a representative two-hour event.
One quote may offer 500 kW and 1,200 kWh with a higher initial price, while another offers 350 kW and 800 kWh at a lower price.
After usable SOC and efficiency, the first quote may have enough energy to remain power-capable, while the second may approach both its power and energy boundary after degradation.
The lower-price quote can still have the higher NPV if the requested reduction is modest, but it can lose that advantage when actual events last longer or the target rises.
The point of the comparison is not to select the largest unit automatically.
It is to expose which assumptions cause each quote to succeed or fail.
Illustrative quote interpretation checklist| Observation | Possible meaning | Evidence to request |
|---|
| Power constraint appears | PCS output is reached before usable energy is exhausted.
| PCS overload behavior, continuous rating, temperature derating, and dispatch trace.
|
| Energy constraint appears | The entered event duration consumes the usable energy boundary first.
| Usable SOC, AC deliverable energy, duration test, and end-of-warranty condition.
|
| Positive year-one margin disappears later | Capacity fade erodes the target margin before replacement.
| Guaranteed capacity curve, augmentation plan, availability, and remedy.
|
| Base NPV is positive but downside is negative | The decision depends heavily on capacity or demand-value assumptions.
| Performance guarantee, bill baseline, measurement protocol, and risk sharing.
|
| Simple payback is early but sustained payback is late | A planned replacement reverses the first cumulative recovery.
| Replacement scope, timing trigger, price basis, warranty reset, and disposal.
|
Recalculate the case with the actual interval load before final design.
A monthly peak and one duration cannot show recharge conflicts, consecutive events, operational reserve, holidays, outages, export restrictions, or detailed state-of-charge trajectories.
Those items belong in a dispatch simulation and engineering study, with this calculator serving as an auditable screening and quote-normalization layer.
Common ways an ESS business case becomes overstated
- Treating rated kWh as deliverable AC energy omits the usable SOC boundary, efficiency, and degradation.
- Applying the highest annual peak reduction to every month ignores seasonal load and target differences.
- Treating a posted demand-charge rate as the exact marginal value of one measured kW can ignore the billing-demand method and contract terms.
- Assuming energy arbitrage is always positive ignores charging losses and a narrow energy-rate spread.
- Comparing only battery and PCS prices omits electrical work, controls, communications, fire-safety systems, commissioning, insurance, inspection, and owner effort.
- Moving replacement beyond the study horizon or adding unsupported residual value changes the quote boundary and can make NPV appear artificially favorable.
- Counting the same battery dispatch as peak shaving, energy arbitrage, renewable capture, resilience, and grid-service revenue without resolving operational conflicts double-counts benefits.
- Using one vendor forecast without a measurement and verification rule leaves baseline adjustment, availability, and remedy disputes unresolved.
Korean tariff, electrical, fire-safety, and public-sector boundaries
Korea Energy Agency technical information describes an ESS as an integrated system involving a battery, PCS, BMS, and a PMS or EMS, and identifies peak shaving as one use case.
A separate Korea Energy Agency energy-reporting guide distinguishes storage capacity in kWh from PCS capacity in kW and states the basic duration boundary created by dividing stored energy by output.
This calculator uses that dimensional relationship but does not replace vendor evidence for usable energy, temperature behavior, degradation, availability, or controls.
KEPCO basic supply terms define maximum demand around a 15-minute measurement concept and distinguish measured maximum demand from demand used for billing.
The calculator therefore embeds no Korean tariff table and asks the user to enter a bill-verified monthly marginal value per peak kW.
A result is not a utility determination of billing demand or a guarantee that a measured reduction will create the modeled invoice saving.
Financial feasibility is separate from legal and engineering compliance
Article 9 of the Electric Safety Management Act, effective February 1, 2026 in the reviewed version, preserves a separate pre-use inspection boundary for applicable private electrical installations.
Article 6 of the related Enforcement Rule, reviewed in the version effective July 1, 2026, addresses inspection targets, standards, and procedure.
NFPC 607, effective May 4, 2026 in the reviewed current performance standard, and NFTC 607, effective July 1, 2024 in the reviewed technical standard, create separate professional fire-safety design and review boundaries.
The public-institution energy-use rule reviewed in the version effective April 30, 2026 defines ESS and contains a public-sector installation provision with coverage, exclusions, and sizing conditions.
None of those determinations can be made from a kW, kWh, or payback result alone.
Official ESS sources and calculator boundaries| Primary source | Reviewed basis | Use and boundary |
|---|
| Korea Energy Agency EG-TIPS ESS information | Reviewed August 17, 2026 | Supports the system-component and peak-shaving context, not universal performance values.
|
| Korea Energy Agency ESS reporting guide | Reviewed August 17, 2026 | Supports the kWh, kW, and duration boundary, while user evidence supplies SOC, efficiency, fade, and realization.
|
| KEPCO Basic Supply Terms | Reviewed August 17, 2026 | Supports the 15-minute maximum-demand context, while billing savings remain a site-specific user input.
|
| Electric Safety Management Act, Article 9 | Effective February 1, 2026 | Identifies the separate pre-use inspection boundary without automating coverage or approval.
|
| NFPC 607 | Effective May 4, 2026 | Identifies a separate fire-safety performance boundary and supplies no project payback or universal sizing rule.
|
| NIST Handbook 135e2025 | Published August 2025 | Supports equal-boundary present-value comparison of initial, recurring, replacement, and energy cash flows without importing a federal discount rate.
|
The reviewed rule version must control when an agency summary page and the current legal text differ.
In particular, the calculator does not automate public-institution coverage because building status, contract power, exclusions, project timing, and evidence require a separate fact-specific review.
It also does not calculate grants, taxes, accounting classification, depreciation, financing, residual value, or insurance eligibility.
Frequently asked questions
What should I enter for peak duration?
Use the recurring time that interval demand remains above the target around the billing-driving event. Review the 15-minute curve rather than one isolated point, choose a conservative representative duration, and calculate separate seasonal cases when the shape changes materially.
Why does the model apply efficiency to deliverable energy?
The model uses one conservative charge-to-discharge boundary so sustainable output and charging cost remain connected. If a vendor states separate discharge efficiency, auxiliary load, or DC and AC guarantees, convert them to one consistent boundary before entry.
Can the marginal demand value equal the published demand-charge rate?
Only when the tariff, billing-demand method, contract, and actual bill support that marginal relationship. Verify the value against real invoices because measured maximum demand and billed demand can differ.
Which payback measure should control the decision?
Simple payback is useful for a quick screen but ignores later replacement and discounting. A long-term review should consider sustained payback, discounted payback, NPV, downside sensitivity, and the physical target margin together.
What happens when replacement occurs in the final study year?
The model includes the replacement cash outflow in that year and resets battery age in the following year. A final-year replacement therefore adds cost without showing later benefit, making the chosen study boundary visible.
Can I set energy arbitrage to zero?
A peak-only screening can neutralize the separate spread by using equal energy rates or zero operating cycles. Do not silently omit real charging energy needed for peak shaving if that omission would overstate project benefit.
Does positive operating headroom mean the installation is safe?
No. It is only the arithmetic gap between entered technical capability and requested reduction. It says nothing about fire, electrical, structural, interconnection, cybersecurity, emergency response, or legal compliance.
Does the higher NPV quote automatically win?
No. Confirm that both quotes include the same scope, test boundary, availability, warranty, replacement, and owner cost. A financially higher result based on weaker evidence or an excluded safety scope is not an equivalent offer.
Turn the result into an EPC evidence request
Compare more than kW, kWh, and total price.
Review monthly target delivery, the degradation path, operating headroom, first and sustained payback, discounted payback, NPV, downside sensitivity, cycle utilization, and the replacement-year cash-flow dip together.
Ask the preferred bidder for a dispatch simulation using the site interval file, a measurement and verification protocol, capacity and availability guarantees, remedy language, an inclusion and exclusion schedule, and documented safety and replacement scope.
A calculator result becomes decision-grade only when its input boundary matches the contract boundary.
- Save the baseline file and the reason for every target adjustment.
- Normalize both quotes to the same performance and cost checklist.
- Verify tariff, interconnection, electrical, fire-safety, and inspection questions with the responsible parties.