Peak Demand Reduction With Solar and Battery

A refrigerated warehouse can add rooftop solar and still face a large demand charge. Its highest grid import may occur as afternoon solar output falls, or during an overnight refrigeration cycle. The relevant question is when the meter records the billing peak, not how much solar the roof generates over a year.

Peak demand reduction with solar and a battery starts with the tariff and measured load. This guide explains the difference between power and energy charges, battery sizing, dispatch controls, seasonal operation, and incremental payback. The warehouse, restaurant, and school scenarios below are illustrative design examples, not documented customer projects.

Last updated: November 2026

SurgePV publishes this post and sells solar design software.

Peak Demand Reduction 2026: What to Check Before Modeling

Use a dated tariff, equipment quote, and tax assessment for a 2026 investment decision. A historical research result can explain how demand reduction works, but it cannot price a specific site or establish eligibility for an incentive.

The practical starting point is the customer’s electricity account. Record the meter, rate schedule, service voltage, billing dates, demand measurement interval, and any contract demand. Ask whether a planned battery or new solar installation changes the applicable tariff. A model built on a superseded rate can produce a convincing cash flow that the next bill never reflects.

For US tax treatment, the IRS Clean Electricity Investment Credit overview describes a 6% base credit and an increased rate of up to 30% where the applicable conditions are met. Additional bonuses require their own eligibility checks; they are not automatic discounts on every battery quote.

The IRS Form 3468 instructions include restrictions involving prohibited foreign entities and material assistance for storage construction beginning after 2025. Have the tax adviser document the project’s construction and service dates, ownership, equipment supply chain, eligible basis, and qualification for any increased rate. Model an unsubsidized case alongside any adviser-approved credit case.

The analysis should also separate who benefits from who pays. A tenant may pay the electricity bill while the landlord owns the roof and electrical infrastructure. A school district may have several meters, each with its own billing peak. Neither a portfolio total nor a tax-credit headline substitutes for a meter-level cash flow and an agreed ownership structure.

Keep a dated assumptions register with the model. Each input should point to a bill, quote, specification, or signed decision. When an input changes, the reviewer should be able to identify which results need recalculation without rebuilding the entire proposal.

Demand Charges vs Energy Charges: What Your Utility Actually Bills

An energy charge prices electricity consumed in kilowatt-hours (kWh). A demand charge prices a measure of power in kilowatts (kW), as defined by the tariff. The billing interval and time window are part of that definition.

Solar can reduce both charges when its output overlaps the relevant load. A battery changes when electricity is imported by charging during one period and discharging during another. The benefit depends on the whole sequence: charging must not create a new expensive demand peak or consume more energy value than discharge saves.

How Each Charge Behaves

Consider an illustrative 30-day month with a constant 200 kW load. Consumption is 200 × 24 × 30 = 144,000 kWh. At an assumed energy rate of $0.10/kWh and demand rate of $15/kW, the energy charge is $14,400 and the demand charge is $3,000, before other charges.

Now consider a 100 kW baseline with one 15-minute interval at 600 kW. Consumption is 72,125 kWh, including the extra 125 kWh during the spike. At the same assumed rates, the energy charge is $7,212.50. If the tariff measures the maximum 15-minute demand, the demand charge is $9,000. The short event materially changes that second line item.

These examples deliberately omit taxes, fixed fees, ratchets, and time-of-use pricing. Their purpose is to distinguish the units. A production estimate in annual kWh cannot be directly subtracted from a billing demand measured in kW.

The Demand Charge Components to Identify

Different tariffs measure different peaks. Read the actual schedule rather than assuming every commercial account has the same structure.

This comparison follows the distinction between tariff design and load profile in Berkeley Lab’s 2017 commercial demand-charge study. It is a checklist, not a statement that every utility uses all five components.

Ratchet Clauses: Why One Missed Peak Can Matter

A ratchet may preserve part of a previous peak as a future billing floor. In an illustrative tariff with an 80% ratchet, a 500 kW reference peak creates a 400 kW floor for the specified lookback period. Reducing measured demand below that floor may produce no immediate demand-charge saving.

Read whether the ratchet applies to all months, only a season, or a particular demand component. Also check whether the contract permits a reset after a permanent operating change. A battery cannot erase an already-recorded peak simply by discharging in a later month.

This changes commissioning priorities. The control system needs correct meter mapping and the correct billing calendar from the start. A successful demonstration on a mild day says little about whether the controller will preserve energy for the expensive event later in the month.

US Utility Demand Charges by Territory: Build a Site-Specific Table

There is no single demand rate for an entire utility territory. Rate class, service voltage, season, and contract terms can change both the charge and the demand to which it applies. The former territory-level price ranges have been removed because they were not tied to dated schedules.

Build the reference table from the customer’s actual account. The most useful comparison is between their eligible tariffs, with the same load and equipment assumptions held constant. A neighboring business may be on a different schedule even when it has a similar monthly bill.

Reproduce an actual bill before forecasting savings. Match the billed energy, demand, billing dates, and adjustments. If the model and bill differ, resolve the difference rather than absorbing it into an unexplained correction factor. Document whether estimated meter readings or missing intervals affect the comparison.

A simple sensitivity calculation illustrates why the rate matters. A sustained 100 kW reduction in billed demand would save $1,000 per month at an assumed $10/kW-month or $2,500 at an assumed $25/kW-month. Those are arithmetic examples, not published utility rates. They also assume no ratchet floor and no offsetting charging peak.

Demand Charge Share of the Bill by Industry

Industry labels are screening clues, not reliable percentages. Refrigeration, manufacturing, schools, hotels, and offices can each have very different operating schedules within the same category. A cold store with staggered compressor operation may have a smoother profile than a smaller building with synchronized equipment starts.

Calculate the demand-charge share directly from bills and show its seasonal variation. Then distinguish the share that is technically reducible from the share fixed by contract or ratchet. A large demand line item is an invitation to investigate, not proof that a battery will pay for itself.

Ask the facilities team to explain the largest events. A one-off maintenance procedure should be handled differently from a recurring production requirement. Future electric-vehicle charging, heat pumps, or added production equipment may matter more than the building’s historic category.

Why Solar Alone May Leave Demand Charges High

Solar reduces billing demand only when generation overlaps the interval that determines the charge. Annual energy savings can therefore be substantial even when the demand-charge reduction is small.

Compare grid import before and after PV for every billing interval. The original gross-load peak may disappear, leaving an evening or cloudy-day peak as the new maximum. Looking only at solar output on the original peak day misses this migration.

Where Solar Alone Still Helps

Sites with daytime loads that align with solar production can reduce billed demand with PV alone. The extent depends on weather, the shape of the load, the size of the array, and the demand window. Avoid treating schools, warehouses, or offices as uniform profiles.

Check operating calendars as carefully as irradiance. A school may have lower summer occupancy, an office may add evening cleaning, and a warehouse may change delivery hours. These patterns affect whether PV production coincides with demand during the expensive billing windows.

If there is useful overlap, retain a PV-only option in the comparison. A battery should earn its incremental cost relative to that option. Giving the battery credit for savings that PV would already deliver inflates the apparent return.

The Sunset Peak Problem

A load can remain high after solar output begins falling. In an illustrative office example, gross load is 480 kW at 3 p.m. and 510 kW at 5 p.m. If PV supplies 380 kW and 140 kW respectively, net import is 100 kW and 370 kW. The later interval sets the higher net demand of the two.

That calculation is not a full monthly bill simulation. Another day may have a higher import, and the tariff may measure a different window. It shows why an attractive midday chart cannot establish the month’s demand saving.

Include cloudy sequences and operational exceptions in the analysis. A controller that uses most of the battery early in the afternoon may have little energy left when solar fades. The required reserve is a dispatch decision tied to the expected remaining load, not a fixed percentage that works for every building.

How Solar Plus Battery Can Reduce Peak Demand

PV supplies energy while the battery can cover selected periods when net load remains high. Their combined effect must be calculated from coincident load, generation, and dispatch rather than by adding separate annual savings estimates.

What Published Research Actually Shows

The 2017 Berkeley Lab and NREL study, Solar + Storage Synergies for Managing Commercial-Customer Demand Charges, reported these median reductions under a basic non-coincident demand charge. Its reference comparison used a one-hour demand averaging interval, simulated buildings and weather, three-hour batteries, and perfect foresight.

These are historical simulation results across a range of configurations. They are not a current commercial savings guarantee or measured results from SurgePV installations. Actual forecasts need the customer’s tariff, equipment limitations, and realistic control behavior.

What Drives the Difference Between Sites

Load shape determines how much power and energy the battery must supply. A narrow spike and a broad plateau can share the same peak kW while requiring very different usable capacity. PV can change both the height and duration of the remaining net-load peak.

The frequency of events matters because the battery must recharge between them. The model should track state of charge continuously, including weekends and holidays. Resetting to a full battery at the start of every simulated day can hide a genuine shortage of charging opportunities.

Control behavior matters as well. A forecast can reserve energy for a later event, but forecasts can be wrong. Compare a simple threshold policy with a forecast-based policy using the same data and constraints. Report the economic difference without claiming a universal performance premium.

Finally, the tariff determines which successful discharges are valuable. A reduction outside the billed time window may save energy without changing demand charges. A reduction below an existing ratchet may have little immediate value. Keep those effects visible rather than reporting a single blended percentage.

Sizing the Battery: kW First, Then kWh

The most common sizing error in commercial peak shaving is starting from kWh. Vendors quote cabinets in kWh. EPC firms model in kWh. Customers ask about kWh first. But the demand charge prices kW, not kWh, so the design must start with the kW shape and back into the kWh.

Step 1: Pull 12 Months of Interval Data

Request 15-minute interval data from the utility through the customer’s account portal. Use the utility’s approved data-export or customer-access process. The dataset should cover one full year to capture seasonal variation.

A complete 365-day dataset at 15-minute resolution contains 35,040 intervals; check leap years, time-zone changes, and missing records. The analysis identifies the highest 15-minute average power draw in each calendar month , the 12 monthly peak events that priced the prior year’s demand charges.

Step 2: Set the Target Demand Ceiling

The target ceiling is the kW level you want to hold the meter under. It is not zero. It is a chosen tradeoff between battery size and savings. Holding the meter under 200 kW on a facility that historically peaks at 500 kW requires roughly 300 kW of battery power for the worst event.

The optimal ceiling balances three things: incremental capital cost per kW of battery, demand charge savings per kW of reduction, and the cost of missing the most demanding billing event.

Step 3: Compute Required Battery Power (kW)

Subtract the target ceiling from each interval’s net load after solar. The largest positive difference is the required discharge power for full coverage of the modeled events. A percentile may be useful for screening, but the excluded peak can still determine the bill.

An illustrative net demand of 480 kW and a target ceiling of 280 kW gives 200 kW of required battery power. Compare the result with available equipment and its temperature-dependent power limits.

Step 4: Measure Sustained Event Duration

Pull the longest sustained event from the interval data , the event where demand stayed above the target ceiling for the longest continuous window. Separate short spikes from sustained plateaus. Also inspect repeated events that leave too little time to recharge.

The duration determines battery energy capacity.

Step 5: Compute Required Battery Energy (kWh)

Integrate the power deficit over the event to calculate usable AC energy. For an illustrative constant 200 kW deficit lasting 3 hours, the requirement is 600 kWh delivered to the load. If the assumed usable state-of-charge fraction is 90% and discharge-path efficiency is 95%, nameplate capacity would be 600 ÷ (0.90 × 0.95), or about 702 kWh before aging and reserve allowances.

These assumed fractions are calculation inputs, not generic battery specifications. Use the selected vendor’s definition of usable capacity, reserve requirements, conversion losses, and end-of-life guarantee. Round-trip efficiency includes charging losses and should not be used interchangeably with discharge efficiency.

The kW-to-kWh ratio at this point in these illustrative options falls between 1:2 and 1:4. A 1:2 ratio (200 kW / 400 kWh) suits short, sharp peaks. A 1:4 ratio (200 kW / 800 kWh) suits broad plateau loads.

This sizing methodology aligns with the framework laid out in our commercial battery storage sizing guide, which walks through the same five steps with a 1.2 MW industrial worked example. For the underlying tariff concepts, see the demand charge and peak shaving glossary entries, both of which feed the dispatch logic in cloud solar design software.

Automation: Predictive Control vs Threshold Dispatch

Controls determine whether stored energy is available during the billed peak. Compare control policies against the same load, PV, battery, and tariff assumptions before paying for a more complex controller.

Threshold Control: A Clear Starting Point

Threshold control requests discharge when net import exceeds a selected limit. If the illustrative limit is 300 kW, a 340 kW net load calls for 40 kW of discharge, subject to equipment power and available energy. A control model must also define charging behavior and a minimum state of charge.

A threshold is easy to explain, but it does not by itself reserve energy for a later event. A long morning plateau can deplete the battery before a larger afternoon peak. Conversely, a conservative threshold may leave capacity unused. The best setting should be evaluated over the whole billing period.

The measurement system matters. Check sensor placement, sign convention, meter latency, and how the controller handles lost communications. A controller responding to the wrong meter can make its own chart look successful while the utility records a different peak.

Predictive Control: Reserving Energy for What Comes Next

Forecast-based control uses expected load and generation to schedule charging and discharge. Its potential advantage is the ability to preserve energy for a later event or exploit an available charging window. Its limitation is forecast error.

Evaluate ordinary days and exceptions separately. A holiday schedule, equipment fault, weather change, or unexpected charging session may create the event that determines the bill. Ask how the controller falls back when forecasts or communications are unavailable, and who can change its operating limits.

A perfect-foresight simulation is useful as a theoretical benchmark, but it gives the controller information that an operating system does not possess. Label that benchmark clearly. Compare it with a deployable policy before converting the result into a customer savings commitment.

What Predictive Control Costs

Obtain a written quote covering licensing, communications, monitoring, integration, updates, and support. Some costs recur even if the battery is unavailable. Others depend on dispatch, enrollment, or a share of savings. The contract should explain how savings are measured and who can audit the calculation.

Payback for the controller should use its incremental benefit relative to the simpler alternative, not all battery savings. Include integration and staff effort. If the vendor will not provide an exportable operating history or explain the baseline, the owner may struggle to verify performance later.

Acceptance should include response behavior, alarms, reserve protection, and recovery after an outage. Agree on the commissioning evidence before equipment arrives. A screenshot of one successful discharge is insufficient to assess a year-round commercial operating policy.

Three Illustrative Design Examples: Warehouse, Restaurant, School

These examples describe design decisions, not completed projects. No customer records support the earlier project claims, so their installation dates, measured savings, tax awards, and internal rates of return have been removed. Use the examples to structure a site investigation.

Illustrative Design Example 1: Refrigerated Warehouse

A refrigerated warehouse may have a steady base load with sharper compressor and defrost events. Start with the actual operating sequence. Ask which loads are essential, which can be staggered, and whether product-temperature requirements limit any proposed load shift.

The first comparison is between gross building load and net load after PV. If the residual peak is late in the day, additional modules may improve annual energy savings without materially reducing that peak. The battery study then needs the duration of the residual deficit, not just the original building maximum.

For an arithmetic screening example, assume a 400 kW power conversion system paired with 1,000 kWh of usable AC energy. At full power, that energy supports 2.5 hours of discharge before reserve and operating restrictions. It does not support a 400 kW deficit for an entire afternoon. A lower power deficit can last longer, but only within the actual energy budget.

A warehouse review should inspect the operating day before the peak too. A battery discharged for arbitrage overnight may enter the next day with insufficient reserve. Charging it during compressor startup could create a new maximum. Model the sequence rather than treating the peak day as an isolated event.

Procurement should include environmental conditions and service access. The owner needs a clear maintenance route, alarm responsibility, emergency procedures, and a plan for periods when a cabinet is unavailable. These costs and constraints belong in the decision alongside electrical savings.

Illustrative Design Example 2: Restaurant Portfolio

A restaurant portfolio should be analyzed meter by meter. Cooking, ventilation, refrigeration, and customer arrival patterns can create different peaks at otherwise similar stores. A portfolio average can hide a site where the battery has little economic value.

Assume a candidate store has a 60 kW battery power limit and 150 kWh of usable AC energy. Its ideal full-power duration is 2.5 hours before other operating restrictions. If the evening peak lasts longer, the controller must either discharge at a lower level or reserve power for the most valuable intervals. The correct choice depends on the tariff.

Lease rights may determine feasibility before economics. Establish who controls the electrical room, who approves equipment placement, and what happens when the lease ends. Check whether the electricity account stays with the operating company and whether equipment can be transferred or removed.

For a portfolio tender, standardize the evidence package while allowing different site sizes. Request a load-data quality check, demand-component breakdown, battery assumptions, and monthly savings schedule for every store. Compare performance using the same conventions for charging losses and service costs.

Demand-response participation is a separate decision. Do not assume a program payment is available simply because another store participates. Obtain the applicable enrollment terms, measurement method, dispatch obligation, and any conflict with onsite demand management. Revenue belongs in the committed case only when the contractual basis is clear.

Illustrative Design Example 3: School District

A school district needs separate occupancy and billing calendars. Teaching days, holidays, sports facilities, kitchens, and summer maintenance may create different load patterns. A sunny summer holiday does not necessarily produce the highest financial value for either PV or storage.

Build the model for each meter, then aggregate financial results. A district-wide battery total cannot automatically reduce demand at an electrically separate campus. Similarly, a portfolio control system cannot move physical battery energy between sites unless the relevant electrical and market arrangements allow it.

An initial design comparison might hold PV constant and evaluate a smaller battery against a larger one. For each option, show which events are fully covered, which remain above the target, and which months are constrained by a billing floor. Explain whether the extra capacity improves ordinary operation or mainly protects against a rare event.

Schools may also want emergency power. Identify essential loads and their required outage duration separately from the demand-charge case. The controller must preserve the agreed reserve, and the financial model must use only the energy available above it. The same stored energy cannot be fully committed to both purposes at the same time.

A district procurement should define ownership, service response, remote access, data retention, and responsible staff. Include the cost of keeping the system useful after personnel changes. An operator who cannot understand the dispatch schedule or retrieve the meter data will find it difficult to challenge a missed savings target.

How to Compare the Examples Without Inventing Results

These are practical investigation questions derived from the distinction between power, energy, and billing demand. They are not performance rankings. A site with a less attractive label can still have better economics if its tariff and operating profile are more favorable.

Keep a source register for any eventual customer case study. It should include permission to publish, measured bills, installation scope, baseline methodology, and the period covered. Until those records exist, present the outcome as a design estimate with its assumptions.

Summer-Only vs Year-Round Peak: Why It Matters for Sizing

Seasonality changes how often a battery can earn savings, but it does not by itself determine the required kWh. Power follows the size of the deficit; usable energy follows its duration and the opportunities to recharge.

Summer-Only Peak Profile

A cooling-dominated building may have its largest peaks in summer. Inspect the rest of the year rather than assuming the battery is idle. The tariff may retain a ratchet, offer different energy spreads, or change the valuable dispatch window by season.

Size against the relevant summer events and then calculate the annual economics. A battery can be technically well sized for the hottest day yet expensive relative to the limited number of months in which it earns substantial savings. A smaller system or operational load change may be preferable.

Year-Round Peak Profile

A continuously operated building may have valuable demand reduction opportunities throughout the year. It may also have broad plateaus that require substantial energy to reduce. More operating days do not automatically mean better payback.

Track charging opportunities, temperature effects, and cumulative throughput. A model should carry the battery state from one event into the next. Repeated daily discharge can affect warranty coverage and future usable capacity, so evaluate the selected equipment’s published terms.

Sizing Implications

Do not select the warranty solely by calendar years. Compare energy throughput, operating temperatures, state-of-charge restrictions, and retained-capacity definitions. Ask how a change in dispatch strategy affects coverage. The owner should know whether adding a second revenue service changes the maintenance or warranty obligations.

Payback Math: Incremental Battery Cost Dedicated to Demand Reduction

Evaluate the battery against a defined baseline, such as PV without storage. Include only the extra costs and savings caused by adding the battery. This avoids giving both assets credit for the same avoided purchase.

Incremental Cost Components

Request a complete installed quote covering cabinets, power conversion, controls, electrical work, design, permits, commissioning, and any network upgrades. Ask which costs are provisional. Space preparation, fire-safety measures, communications, and ongoing support can affect the decision even when they are absent from an initial equipment price.

Keep taxes and incentives as separate lines. An incentive may arrive after the owner has paid the installation invoices, so cash timing and financing costs matter. A tax-credit percentage should apply to the adviser-approved eligible basis, not automatically to every line on the contractor’s invoice.

Incremental Annual Savings

Calculate demand savings as the difference between baseline and proposed bills for each billing period. Then add energy-charge changes, including charging losses. Count program revenue only where the operating schedule and contract support it.

For a transparent screening calculation, assume a 100 kW reduction in billed demand in every month at an assumed $15/kW-month. Annual demand savings are 100 × 15 × 12 = $18,000. If installed cost is assumed to be $120,000 and incremental annual service cost is $3,000, simple payback is $120,000 ÷ $15,000 = 8 years. No subsidy, arbitrage revenue, tax effect, or replacement is included.

This is arithmetic, not a market-price estimate. Change the inputs to the actual quote and simulated monthly bills. A financing decision also needs discounting, degradation, replacement timing, financing costs, and residual value. Simple payback leaves those effects out.

When the Math Does Not Work

A battery may fail the economic screen when the reducible demand charge is small, the event lasts too long, or a contract floor prevents billed demand from falling. There is no universal rate threshold that separates good projects from bad ones.

Stress the result against missed peaks and lower available capacity. Show the customer which assumption most changes the outcome. If the project only works with every incentive, full availability, and ideal control, that dependency should be visible before procurement.

Compare operational alternatives too. Staggered equipment starts or managed charging may reduce the required battery size, but assess their cost and operational consequences. A production or comfort constraint can make an apparently inexpensive load shift unacceptable.

Common Implementation Mistakes

The most useful design review asks where the proposed system can fail to reduce the actual bill. Check capacity, metering, control priorities, and the tariff together.

1. Undersizing on kWh by Using a Vendor Default

Calculate usable energy from the load deficit. An illustrative 300 kW deficit lasting 3.5 hours requires 1,050 kWh delivered to the load. A cabinet offering only 900 kWh of usable AC energy would cover three hours at that power, leaving half an hour uncovered. Nameplate capacity alone cannot resolve the shortfall.

Integrate a variable deficit rather than multiplying the maximum power by every event’s full duration. The rectangular approximation can be useful for screening, but it may overstate or understate the design requirement when reserve, efficiency, and repeated events are ignored.

2. Sizing the Power Conversion System Too Small

A battery with 1,000 kWh of energy and a 250 kW power conversion system cannot deliver a 350 kW discharge merely because enough energy remains. Power and energy are separate limits. Check output at the site’s temperature and electrical conditions, including any shared inverter constraint.

3. Ignoring Capacity or Generation Demand Charges

A controller focused on the building’s own maximum may not address a charge based on a wider system event. Identify every demand-related component and its measurement rule. Where an aggregator offers another service, reconcile its dispatch instructions with the site’s own bill-management objective.

4. Not Verifying Ratchet Clauses Before Sizing

An illustrative 600 kW reference peak under an 80% ratchet creates a 480 kW floor for the applicable period. The model should carry that floor forward according to the tariff. Do not restart the ratchet at zero each month or assume a new battery removes an earlier obligation.

5. Combining Backup Power and Peak Shaving Without a Reserve

Backup requirements need their own essential-load schedule and duration. Agree on the reserve that must remain available. Then run the peak-shaving case with that energy withheld. If the owner chooses to release reserve during expensive events, document who authorizes the tradeoff and how reserve is restored.

6. Failing to Consider Building Controls

Managed vehicle charging, equipment sequencing, and suitable thermal load shifts can change the residual peak. Investigate them with the facilities team before selecting battery capacity. Include communications, comfort, production, and recovery-load effects rather than describing curtailment as free.

How SurgePV Fits the Project Workflow

SurgePV supports solar array layout, shading analysis, energy-yield modeling, financial modeling, and proposal generation. Those documented capabilities help organize the PV side of a commercial project and explain reviewed assumptions to the customer.

A tariff-specific battery dispatch study needs its own verified tool and engineering review. Do not assume that interval-data import, forecast-control simulation, ratchet modeling, or capacity-market optimization is included simply because a platform offers financial modeling.

Use the financial modeling workflow to discuss the supported scope in a demo. Bring the actual tariff, load-data format, PV assumptions, and required outputs. Ask the team to demonstrate the intended workflow and identify which analyses require a separate specialist.

Keep the handoff explicit: the PV model provides generation assumptions; the battery study evaluates constrained dispatch; the finance model compares costs and benefits; the proposal identifies who approved each input. This separation makes revisions easier to audit and prevents a sales estimate from becoming an unsupported engineering guarantee.

What Commercial Solar Buyers Often Miss

The cheapest or largest system is not automatically the best investment. Compare options against the same baseline and the same operating constraints.

Bigger PV Does Not Guarantee Bigger Demand Savings

More modules may offset additional energy while the residual peak remains outside the solar window. Evaluate PV-only and PV-plus-storage alternatives separately. The next module and the next battery cabinet should each earn their incremental cost.

A Cheaper Battery Can Have a More Expensive Shortfall

Compare usable energy, deliverable power, warranty, service access, and controls alongside purchase price. A lower-cost option may still be right if its missed events have little financial consequence. The issue is the value of the coverage gained by spending more, not size alone.

Demand Response Is a Separate Service

Program availability payments and event performance payments have different rules from avoided retail charges. A dispatch instruction can also conflict with preserving energy for a later onsite peak. Model both services together before adding their independently calculated revenues.

When Peak Shaving Is Not the Right Strategy

Peak shaving deserves less emphasis when the reducible demand charge is small relative to the required installed cost. It may also be unsuitable where long load plateaus exhaust affordable storage or contractual minimums prevent savings.

Major planned load changes require a forward-looking scenario. Added production equipment, electric heating, or vehicle charging can change both peak power and duration. Keep the measured historical case, but do not present it as a forecast of a substantially different operation.

An unresolved tariff transition is another reason to compare alternatives. Ask what happens to the investment if the rate schedule changes. Energy arbitrage, self-consumption, resilience, and operational load management may have value, but each needs its own evidence and cannot simply replace a missing demand-charge benefit.

A useful screening output is a reasoned go/no-go decision with the missing information listed. It should identify whether the obstacle is technical, contractual, economic, or simply poor data. That allows the owner to address the actual constraint before buying more modeling work.

Stacking Revenue: Demand Response and Capacity Markets

A battery may support several services, but the same power and energy cannot be committed twice at the same moment. Revenue stacking requires a feasible combined dispatch schedule and compatible contracts.

How Stacking Works

Start with the primary service and its reserve requirement. Then add each proposed service with its event window, availability obligation, measurement baseline, payment formula, and penalty terms. Ask whether enrollment changes export permissions or equipment-control rights.

Realistic Revenue Stacking

Use a separate line for each income source and identify its status: contracted, eligible but uncontracted, or exploratory. Keep exploratory revenue outside the committed investment case. A published program headline is not evidence that a particular site can enroll or receive the maximum payment.

Review the net benefit after aggregator fees, added cycling, communications, and service costs. Test a coincident event where the program requests discharge before the building’s own peak. Show how the controller chooses and which savings are forgone.

What Stacking Requires from the Battery

Read the actual warranty and operating agreement. Throughput, depth of discharge, temperature, and operating mode may affect coverage. Preserve enough operating data to demonstrate compliance and explain disputes over performance.

The customer should receive a priority schedule in plain language. It should say which objective wins during conflicts, who can override controls, how reserve is restored, and which party bears the financial consequence of an unavailable system.

Year-Round Operational Considerations

Demand management needs ongoing review because load, tariffs, and available capacity can change. Assign responsibility for that review before commissioning.

Tariff review. Compare new rate notices with the model’s effective schedule. Check demand windows and ratchets as well as prices. Preserve the earlier version so a change in billed savings can be explained.

Load-profile review. Look for changed schedules, new equipment, and unusual peaks. Discuss them with the facilities team before retuning controls. A short-term workaround should not quietly become the permanent operating policy.

Capacity and availability review. Follow the manufacturer’s approved service procedures. Track usable capacity, outages, alarms, and thermal restrictions against the original design assumptions. Update the savings forecast when the system’s available capability changes.

Event review. Compare predicted and actual dispatch around the billed peak. Determine whether a miss came from insufficient power, insufficient energy, a forecast error, an unavailable asset, or an incorrect tariff setting. Those causes need different remedies.

Data retention. Keep meter imports, PV generation, battery charging and discharge, state of charge, and control changes on consistent timestamps. The owner needs access to these records even if the software supplier or maintenance contractor changes.

Financial reconciliation. Compare modeled bills with issued bills and explain differences. Weather, occupancy, and tariff changes complicate before-and-after comparisons. Separate those effects from equipment performance before reporting a savings percentage to management.

The Outlook for Peak Demand Reduction Through 2028

Future battery prices, tariffs, and control performance should be treated as scenarios. A project should not rely on an unsupported forecast that equipment will become cheaper or demand charges will rise by a particular percentage.

Maintain a base case using current documented inputs and alternative cases for the variables that matter most. Show the owner how a different tariff, replacement cost, service fee, or usable-capacity trajectory changes the decision. Avoid combining every favorable assumption into a single headline payback.

The useful long-term question is whether the system can adapt. Review expandability, data access, control settings, service arrangements, and ownership of operating records. A project that can be re-evaluated as conditions change is easier to manage than one whose economics depend on a hidden calculation.

Next Steps

Collect a complete year of bills and interval data, identify every demand-charge component, and reconcile a baseline model to the actual account. Record missing data and upcoming operating changes before sizing equipment.

Compare PV alone, operational load management, and PV with storage using consistent assumptions. Ask the storage engineer to show the hardest events, charging constraints, and reserve policy. Ask the finance reviewer to separate eligible incentives from unverified revenue.

For the solar design and proposal portion, book a project walkthrough. Bring the load and tariff questions with you so the supported software scope and specialist handoffs are clear before procurement.

Frequently Asked Questions

How much can solar plus battery reduce commercial demand charges?

Savings depend on the tariff, load profile, and constrained battery dispatch. The 2017 Berkeley Lab and NREL simulations reported a 42% median reduction for their basic non-coincident demand-charge comparison, but that result is not a forecast for an individual property.

What is the difference between a kW demand charge and a kWh energy charge?

A kW demand charge prices a defined measure of power, while a kWh energy charge prices electricity consumed. Read the tariff to identify the demand averaging interval, time window, and any billing floor.

Why can solar leave commercial demand charges high?

The billed peak can occur when solar output is low or unavailable. Compare every billing interval after PV, because reducing the original peak may expose a different interval as the new maximum.

How do you size a battery for peak demand reduction?

Calculate required discharge power from the net-load deficit and usable energy from its duration. Then account for efficiency, reserve, aging, recharge opportunities, and the selected equipment limits.

What is the payback for a commercial peak shaving battery?

There is no universal payback period. Divide the documented net investment by the modeled incremental annual benefit for a simple-payback screen, then assess financing, replacement, degradation, and tax timing separately.

Are predictive controls better than threshold controls?

Their value depends on the site and forecast quality. Compare both policies under the same constraints, including missed forecasts and unavailable equipment, before assigning a financial premium.