Solar + Storage Design: Battery Sizing & Coupling

Storage has moved from an optional upsell to a core part of solar design. In California, the net billing tariff pays less for exported energy than the retail rate, which makes a battery central to many project economics. Elsewhere, time-of-use rates and export rules vary by utility, so the tariff decides whether a battery earns its cost.

Homeowner economics also lost a federal credit. The IRS says the residential clean energy credit under Section 25D is not available for property placed in service after December 31, 2025, and the credit had covered battery storage of at least 3 kWh. Bill savings and resilience now have to carry the case, which makes correct sizing and honest payback math more important.

Last updated: November 2026

This guide is for installers and EPCs who design battery-backed systems. It covers architecture, sizing, backup panels, code, inverters, modeling and economics. It assumes grid-tied residential and light commercial projects, which is where most storage work sits. Off-grid systems share the math but change the targets, and the text says where they differ.

In this guide:

• Why storage design is now a core installer skill
• AC vs DC coupling tradeoffs, with a decision framework
• Battery sizing math with a worked residential example
• Backup load panel design and transfer equipment
• NEC 706 and 705 compliance essentials for energy storage
• Inverter selection for hybrid new builds and AC-coupled retrofits
• Production and consumption modeling with interval data
• Storage economics, including when a battery does not pay
• 7 common storage design mistakes and how to avoid them

Why Solar + Storage Design Matters in 2026

Three forces push storage into the design brief: export compensation, the loss of the homeowner credit and resilience.

Export compensation. The CPUC’s net energy metering page says that under the Net Billing Tariff, export compensation is based on Avoided Cost Calculator values that are usually lower than the retail rate, and it recommends batteries to store solar for the late-summer evening hours when export credits can exceed retail. A solar-only system exports midday energy at that lower value, while a battery moves the same energy to the evening, when the customer would otherwise buy at the retail rate. Other utilities set their own rules, so read the tariff for each project.

The residential credit. The IRS residential clean energy credit page says the credit is not available for property placed in service after December 31, 2025. With no federal credit softening the price, project value must come from self-consumption, peak shaving and backup capability.

Resilience. Outage frequency varies widely by region, and many homeowners now ask for backup by name. A battery is then a reliability product as well as a bill-savings device, and it is sold that way.

Rate design matters too. Where utilities cut export rates or widen the gap between peak and off-peak prices, value moves from exporting energy to storing it. Check each utility’s current tariff, because those changes happen on the utility’s schedule.

For installers, this changes the job. Storage projects carry larger tickets than solar-only work, and more design risk. A mis-sized battery or a rejected permit plan set can erase the margin on an otherwise clean install. Commercial clients arrive with demand charge printouts and expect a designed answer, not a catalog page.

AC vs DC Coupling: Which Architecture to Choose

Every battery-backed system stores DC energy and serves AC loads. The design question is where the battery connects. AC coupling places a battery with its own inverter on the AC side of the system. DC coupling connects the battery to the DC bus of a hybrid inverter that also serves the PV array.

The difference comes down to conversion count. In a DC-coupled system, PV energy passes through 1 inversion before reaching the home or the battery. In an AC-coupled system, PV energy is inverted to AC, rectified back to DC for storage, then inverted again for use. Each conversion costs efficiency.

That extra conversion shows up in round-trip efficiency (RTE). A DC-coupled system usually shows a higher RTE than an AC-coupled one because fewer conversions occur, but the figures are product-specific. Read each manufacturer’s datasheet for the rated RTE and the conditions behind it. Over a 10-year life, the gap is real but rarely decisive on its own.

Retrofits flip the math. An AC-coupled battery connects without touching the existing PV inverter or array wiring. It works with string inverters and microinverters alike. For the installed base of grid-tied systems, AC coupling is usually the only sensible option.

New builds tilt the other way. DC coupling means 1 inverter instead of 2, which cuts hardware cost, wall space, and conduit runs. It also captures clipped energy on high DC-to-AC ratio designs, which an AC-coupled system would lose. For most ground-up residential projects, DC coupling is the cleaner design.

A simple rule of thumb: for a retrofit, AC couple. For a new build with a single inverter location, DC couple. For a new build with microinverters already specified, AC couple. The architecture should follow the project, not the other way around.

Outage behavior deserves a closer look. An AC-coupled battery must island the existing PV inverter and manage its output through frequency shifting. That coordination is common, but it adds a control dependency that a hybrid system handles natively on the DC bus.

DC coupling also gives you more room on array sizing. Because clipped energy can flow straight into the battery, you can push the DC-to-AC ratio harder without wasting production. On tight roofs where every module counts, that flexibility has real value.

Service and monitoring differ as well. A hybrid system reports PV and battery through 1 app and 1 support channel. An AC-coupled system splits that across 2 manufacturers, which matters on year 7 when the homeowner calls you first.

AC coupling also scales cleanly for whole-home retrofits. Because the battery inverter is independent, you can stack units over time as the budget allows. Many homeowners start with 1 battery for critical loads and add a second unit a year later.

One more tradeoff rarely mentioned in spec sheets: fault isolation. With 2 inverters, a battery failure does not take down PV production, and vice versa. With a hybrid unit, 1 failure can idle the whole system until the replacement arrives.

Battery Sizing: Critical Loads, Autonomy, and Daily Usage

Battery sizing starts with 3 numbers: critical load energy per day, desired autonomy, and system losses. Get those right and the capacity picks itself. Get them wrong and no inverter setting will save the design.

The core formula is:

Usable capacity (kWh) = Daily critical loads (kWh) × Autonomy (days) ÷ (Depth of discharge × Round-trip efficiency)

Work through it in 6 steps:

1. Audit the critical loads with the homeowner. List every circuit the battery must serve during an outage.
2. Convert each load to daily energy. Running watts multiplied by realistic hours per day gives Wh per day.
3. Choose an autonomy target. One day is a common starting point for grid-tied backup, and outage-prone regions may justify more.
4. Adjust for depth of discharge (DoD) and round-trip efficiency. Take the usable capacity and DoD limit from the battery datasheet, because chemistry and warranty terms differ.
5. Check power, not just energy. The inverter’s continuous kW rating must exceed the largest coincident load.
6. Check surge. Motors draw several times their running current for the first seconds after start, so use the locked-rotor or surge figure from the motor’s nameplate or datasheet.

Here is a worked audit for a typical US home on partial backup:

Apply the formula. A 1-day autonomy target with a 90% DoD battery at 95% RTE gives 6.0 ÷ (0.90 × 0.95) = 7.0 kWh of nameplate capacity. A single battery of about 10 kWh usable energy or more covers this home with margin for degradation and cloudy stretches.

Now the power check. If the refrigerator, blower, well pump and lighting all run together, the running load is about 2.1 kW (150 + 600 + 1,200 + 100 W). The battery inverter must exceed that continuously. It must also start the well pump, whose start surge is several times its 1,200 W running load.

Battery products come in a few size classes, and usable energy, continuous power and surge differ by product. Pick the class from your calculation, then read the datasheet for the exact figures. A small partial-backup audit like the one above lands near 7 kWh, while the whole-home example below lands near 21 kWh. Off-grid designs are a different problem. They often need several days of autonomy plus generator support, which pushes capacity far higher.

Do not forget the recharge side. The PV array must serve daytime loads and refill the battery on a typical day. A useful rule: daily PV production in the worst month should exceed daily consumption plus planned battery charging, so check the modeled winter output and not the annual average.

A whole-home example shows how fast the numbers scale. Take a home using 30 kWh per day with 18 kWh of overnight and outage-critical consumption. At 1 day of autonomy, 90% DoD, and 95% RTE, the nameplate requirement is 18 ÷ 0.855 = 21 kWh. That points to the larger size classes, with power ratings checked against the coincident load as before.

Stretch the same home to 2 days of autonomy and the requirement doubles to about 42 kWh nameplate. That is several stacked units, or a different product class entirely. Autonomy is the cheapest number to change and the most expensive to underestimate.

Battery chemistry belongs in the sizing conversation, too. LFP packs tolerate deeper daily discharge and higher cycle counts, which supports using the full nameplate figure in the formula. NMC packs often carry lower recommended DoD, so the same nominal size delivers less usable energy per day.

Seasonal drift is the last check. Winter production can fall well below summer output in northern climates. If the design depends on daily recharging in December, verify it against winter-specific yield, not the annual average.

A common undersizing failure is a power failure, not an energy failure. The battery has enough kWh and cannot start the air conditioner. Always run the surge check before you run the payback check.

Backup Load Panel Design and Transfer Equipment

The critical loads panel is where backup design becomes physical. You move the circuits the battery will serve into a dedicated subpanel. Everything else stays on the main panel and goes dark during an outage.

Partial backup is the cheapest clean design. A dedicated critical loads subpanel feeds refrigeration, lighting, a furnace blower, and a few convenience circuits. One battery handles it in most homes, and the install stays simple.

Whole-home backup is a power problem, not an energy problem. A 200 A service at 240 V can draw up to 48 kW at the bus (200 x 240). A residential battery stack rarely covers that continuously. Realistic whole-home designs pair several batteries with smart load management that sheds the air conditioner, water heater, and EV charger when the battery runs low.

Transfer equipment isolates the home from the grid during an outage. Modern battery systems use an integrated gateway or an automatic transfer switch (ATS) rated for the backed-up load. The NEC has separate rules for optional standby systems. Grid isolation must be positive, and the AHJ decides which listed transfer devices are acceptable, so ask before you specify a manual interlock instead of a listed device.

Meter-main combination services complicate the layout. When the meter and main breaker share 1 enclosure, the transfer device usually goes between that unit and a new backed-up panel. Plan the physical stack-up on the wall before quoting, not during rough-in.

Generator pairing is the other common request. Most modern battery systems accept a generator input or coexist with one on a separate transfer switch. Confirm the exact topology with the battery manufacturer before you quote, because wiring a generator in the wrong place can defeat the transfer scheme.

A smart electrical panel can replace the subpanel entirely. It gives per-circuit control, monitoring, and automated load shedding. The tradeoffs are cost, a longer install day, and a homeowner app that someone has to explain.

Load selection deserves a real conversation, not a checkbox. Ask the homeowner what they actually need during an 8-hour outage. Most pick the refrigerator, some lights, the furnace blower, and Wi-Fi. Almost none pick the second freezer in the garage until the first outage.

Panel placement affects cost more than most quotes admit. Every circuit moved to the critical loads panel needs its conductors rerouted or spliced. A subpanel placed beside the main panel keeps those runs short; a battery location two floors away turns the same job into a day of fishing wire.

Commissioning closes the loop. Before you leave, simulate an outage with the homeowner watching. Confirm the transfer happens in the specified time, the critical circuits stay live, and the large loads shed in the right order.

Keep 1 spare breaker space in the critical loads panel. Homeowners add loads, and a full panel turns a 30-minute circuit addition into a panel replacement. Future-proofing costs almost nothing at install time.

Label every circuit in the critical loads panel with its backed-up status and measured draw. The next service call goes faster. The homeowner also stops plugging a space heater into the backed-up bedroom circuit.

NEC 706 and 705 Compliance for Energy Storage

NEC Article 706 covers energy storage systems. It sits alongside Article 690 for PV and Article 705 for interconnection. Storage plan sets tend to fail review on disconnects, signage and working space, so check those first. Section numbers, limits and exceptions change between code editions, so this guide names the topics and leaves the numbers to the edition your authority having jurisdiction (AHJ) has adopted.

Typical topics the article addresses include a disconnecting means for the storage equipment, overcurrent protection matched to the equipment listing, a directory identifying the power sources on the premises, and working space around the equipment. Read the adopted edition for the exact wording.

Equipment listing is the foundation. UL 9540 is the safety standard for energy storage systems, and UL 9540A is the test method that produces thermal runaway propagation data. AHJs may ask for 9540A reports during plan review, so ask the manufacturer for them before you submit.

NFPA 855 adds installation rules on top of the NEC, including limits on unit size, aggregate capacity, spacing and location. Those limits depend on the edition and on where the system sits, such as a garage, utility closet or outdoors. Look up the current limits in the standard and in your AHJ’s local amendments, and solve placement in design.

NEC Article 705 covers the interconnection side. Busbar limits on load-side connections can cap how much PV and storage a panel accepts, because a battery breaker and a PV breaker share the same busbar allowance. A power control system (PCS) is one route that some code editions offer to monitor currents and limit backfeed, which can avoid a service upgrade on an existing panel. Check whether your adopted edition permits it and whether the equipment listing supports it.

A busbar check belongs in every storage design. Add the PV breaker, the battery breaker and the main breaker, compare the sum with the panel’s busbar rating and the rule in your adopted edition, and decide early whether a PCS, a supply-side tap or a service upgrade is needed.

PV rapid shutdown still applies with storage attached. The battery must not keep array conductors energized beyond the shutdown boundary, so coordinate the storage shutdown scheme with the PV rapid shutdown initiator during design, not on the roof.

Code editions vary by jurisdiction. One AHJ works from one edition, the next county from another, and the utility adds its own interconnection handbook on top. Check which edition applies before you draw the single-line diagram.

The utility side adds a second layer. Many utilities require an updated interconnection agreement when storage joins an existing PV system, even if nothing about the array changes. Some also require export control settings or a non-export relay.

Commercial storage raises the stakes. Larger systems can trigger provisions for fire department access, signage and dedicated rooms, and the permitting timeline may be longer than for residential work. Ask the AHJ about review time at the start.

Signage is cheap and frequently wrong. The directory must identify every power source, its location and its disconnect. Print it, laminate it, and mount it where the inspector stands, not where it fits aesthetically.

Call the AHJ before design freeze on any storage project. A short call about placement and clearances can save a redesign. Both the NFPA 70 (NEC) and NFPA 855 pages let you read the standards online with a free account, which is the place to confirm any number in this section.

Inverter Selection: Hybrid vs AC-Coupled Retrofits

The inverter decision follows the architecture decision. A hybrid inverter serves the PV array and the battery from 1 chassis in a DC-coupled design. An AC-coupled battery brings its own inverter and leaves the existing PV inverter alone.

Read 3 ratings on every datasheet, not 1. The grid-tied rating sets normal operation. The backup port rating sets what the unit can serve while islanded, which can be well below the grid-tied number. The surge rating sets what it can start. The rule-of-thumb column below gives margins to check against the datasheet, not manufacturer requirements.

Oversizing the battery and undersizing the inverter is the classic rookie error. A large battery behind a small inverter serves lights and a refrigerator. It will not serve a cooktop and a well pump at dinner time.

For retrofits, confirm compatibility with the existing PV inverter before quoting. Most AC-coupled batteries island the PV inverter and manage its output through frequency shifting. That coordination works with most string inverters and microinverters, but verify the pairing on the battery manufacturer’s compatibility list.

On new builds, compare the hybrid unit’s backup port rating against the critical loads sum before price enters the conversation. The cheapest hybrid inverter often has the weakest islanded output. That gap only shows up during the first outage, on your warranty ticket.

Frequency-shift control deserves a compatibility check of its own. When the battery is full during an outage, the AC-coupled battery raises grid frequency to curtail the PV inverter. Most modern inverters ride through this gracefully, but older units can trip offline and stay there.

Three-phase service changes the selection entirely, mainly on C&I projects. Many residential-class batteries are single-phase only, and mixing them onto a 3-phase service creates imbalance problems. Confirm phase configuration before shortlisting equipment.

Firmware matters more than brochures admit. Backup behavior, surge response, and generator coordination all live in software. Check the manufacturer’s release history. A unit with active firmware support ages far better than one abandoned after launch.

Warranty length tracks architecture, too. Compare the inverter warranty and the battery warranty term by term, and match them across the stack so 1 component does not become the weak link in year 11.

One spec worth asking about directly: pass-through current. Some gateways limit the backed-up panel to a lower current than the inverter can supply. Match the transfer equipment rating to the design, or the bottleneck moves to a box you already sold.

Production and Consumption Modeling for Storage

Monthly kWh averages hide the information that sizes batteries. A home using 30 kWh per day with a flat profile needs a different battery than the same home with a sharp evening peak. Use interval data, either 15-minute or hourly, whenever you can get it.

A useful storage model outputs 4 curves: PV production, household consumption, battery state of charge, and grid import/export. From those curves you get self-consumption rate, cycles per year, and export volume. Those 3 numbers drive the economics.

SurgePV’s shadow analysis and energy-yield modeling cover the production side for the actual roof, and its generation and financial tool covers financial modeling. Check during a demo which load-profile and tariff inputs the tool accepts for a storage study, because the published capability list does not name battery dispatch.

Tariff structure drives the cycling strategy. On time-of-use rates, the battery discharges through the utility’s peak window, often in the evening, and recharges from midday solar. Read the tariff for the exact hours. On flat rates with low export compensation, it simply absorbs surplus production for evening use.

Under California’s net billing tariff, export timing matters as much as export volume. Avoided-cost rates change by hour and by month. A model that assumes flat export prices will misstate storage savings by a wide margin in either direction.

Rate escalation assumptions matter just as much. Choosing an annual utility rate escalation instead of a flat rate can change the 20-year savings figure a lot. Document your escalation assumption in every proposal, because it is easy to inflate.

Include degradation in the model. A battery cycling daily loses usable capacity over time, within its warranty terms. Take the warranted end-of-term capacity from the manufacturer and model year 1 and year 10 side by side. If the design only works at year-1 capacity, it does not work.

Getting interval data is easier than it used to be. Many US utilities offer Green Button downloads from the customer portal, and many smart meters expose 15-minute data. Ask for 12 months, because a single summer bill can mislead the whole design.

When real data does not exist, synthesize the profile from known anchors. Start with the monthly kWh from bills, then apply a load shape matched to the home type: EV charging overnight, cooking peaks in the evening, cooling peaks in the afternoon. Flag the model as estimated in the proposal so everyone knows its limits.

Cycle-throughput warranties also belong in the model. Many batteries warrant energy throughput alongside years and cycles, so read the warranty. A design that cycles aggressively twice a day can hit the throughput cap before the calendar limit, which changes the real warranty end date.

Economic Analysis: When Storage Pays and When It Doesn’t

Here is the part most storage marketing skips. A battery is not automatically a good investment. Whether it pays depends on the tariff, the installed cost and how the customer values backup, so model the project with and without storage and let the customer see both.

Storage pays fastest where exports are worth much less than retail energy, where time-of-use spreads are wide, where commercial demand charges are significant and where outages carry real cost. It pays slowest where the grid already acts like a free battery. Under full retail net metering, exported energy comes back at full value, so a battery saves little on the bill. See the glossary entry on net metering for the mechanics.

The expired residential credit changes the math. With Section 25D no longer available for property placed in service after December 31, 2025, homeowners cannot count a 30 percent credit, and every residential case should be modeled without any federal credit. Commercial storage follows a different schedule. Eide Bailly’s summary of the 2025 law says the accelerated phase-out applies to solar and wind, and that credits for energy storage remain on their earlier schedule. Ask a tax professional which rules apply to a given commercial project.

On cost, use current local quotes, because installed storage prices vary by product, region and whether the battery is added to a new solar install or retrofitted. Adding a battery to a new solar install usually costs less than a standalone job because labor, permitting and truck rolls are shared.

Levelized cost of storage is a cleaner metric for skeptical customers. Divide total installed cost by lifetime discharged energy. As a simple illustration with assumed numbers, a battery installed at $1,200 per kWh and cycled 4,000 full cycles delivers energy at about $0.30 per kWh before losses and degradation (1,200 / 4,000). Compare that with the customer’s peak rate. It beats peak prices in some markets and loses badly in others.

Commercial projects stack value differently. A commercial battery can shave demand charges, arbitrage energy prices and provide backup in the same day, so the payback can look very different from a residential one. Use interval data for each case.

Our solar proposal software can show the with-and-without-storage comparison in the customer-facing proposal, which builds more trust than a single optimistic payback line. For cost context, see our solar installation cost breakdown and the battery payback guide.

Financing changes the customer’s math more than most tariffs. A battery rolled into a long solar loan costs less per month and more in total interest. Show both the cash price and the financed payment. Ask for any dealer fee on financed storage in writing, and include it in the cost comparison, not in the fine print.

One honest note on resilience: its value is real, and it is not a spreadsheet number. Some customers pay for backup the way they pay for insurance. Others will not. Offer the option, price it correctly, and never pad the bill-savings case to sell it.

Common Storage Design Mistakes

The same errors keep appearing in storage designs. None of them are exotic, and all of them are avoidable at the design stage.

1. Sizing on energy and ignoring power. A 13.5 kWh battery with a 5 kW inverter cannot start a 4-ton air conditioner. Check continuous and surge ratings against coincident loads before anything else.

2. Promising whole-home backup with 1 battery. One unit backs up a critical loads panel well. Whole-home coverage usually needs several units plus load management, and the proposal should say so.

3. Ignoring PV recharge capacity. A battery paired with an undersized array sits empty through winter. Verify that daily production exceeds daily consumption plus planned charging.

4. Defaulting to 1 architecture. AC coupling on every new build wastes the customer’s money. DC coupling on every retrofit wastes more. Match the architecture to the project in front of you.

5. Treating code clearances as an install-day problem. Fire-code spacing and location rules decide where batteries can physically go. Solve placement in design, not on the wall with the inspector watching.

6. Modeling on monthly averages. Averages hide peaks, and peaks size the equipment. Interval data costs almost nothing to obtain and prevents the most expensive sizing errors.

7. Forgetting degradation. Design for year-10 capacity, not day-1 capacity. The warranty already expects fade, and so should your autonomy calculation.

The pattern across all 7: storage punishes assumptions. Every mistake on this list is a number someone did not measure or a rating someone did not read. The fix is process, not talent.

A short checklist catches these before the quote goes out. Build your own version covering power, surge, recharge, architecture, code and degradation, and require it on every storage design, with no exceptions for experienced designers.

Conclusion

Storage design is now a core installer skill, and the market rewards the teams that treat it as a process. Match the architecture to the project. Size energy and power separately. Design the backup panel before you price the job.

Rate structures keep getting more complex, and customers are asking for backup by name. Installers who master storage design can win that work on quality and not on price.

None of this requires new hardware or exotic tools. It requires discipline: measure the loads, respect the code, model the real tariff. Teams that do those 3 things win against teams still quoting batteries from a brochure.

Code compliance belongs in the design phase, too. NEC 706, NFPA 855, and Article 705 decide placement, disconnects, and interconnection limits. A 10-minute AHJ call before design freeze is the cheapest insurance on the project.

Model the economics honestly. Interval data, real tariffs, and year-10 degradation produce payback numbers that survive the first utility bill. That honesty is what turns a storage customer into a referral source.

Storage done well is also the best marketing you can buy. A battery that performs exactly as promised generates referrals for years, and a battery that underperforms generates the opposite. Design is where that outcome gets decided.

Your action items from this guide:

• Audit critical loads with interval data before quoting any battery.
• Match the coupling architecture to the project: AC for retrofits, DC for most new builds.
• Design to NEC 706, NFPA 855, and Article 705 from day 1, and call the AHJ early.

SurgePV covers solar design, shadow analysis, financial modeling and proposals in one platform, and you can book a demo to see how that fits a storage workflow. If storage is becoming your core offering, the design process above is where the margin lives.

Frequently Asked Questions

What is the difference between AC and DC coupled solar storage?

AC-coupled systems use a separate battery inverter and are easier to retrofit. DC-coupled systems connect the battery to the solar inverter’s DC bus and usually have fewer conversion steps. AC coupling is more flexible for existing systems, while DC coupling is often the cleaner design for new installations.

How do you size a solar battery?

Divide daily critical-load energy times autonomy days by depth of discharge times round-trip efficiency. For a 6 kWh a day critical load, one day of autonomy, 90 percent depth of discharge and 95 percent efficiency, that is about 7 kWh. Then check that the inverter’s power and surge ratings cover the largest simultaneous loads.

What is NEC 706 in solar storage?

NEC Article 706 covers energy storage systems, including disconnects, overcurrent protection and related installation requirements. It works alongside Article 690 for PV and Article 705 for interconnection. Requirements change between code editions, so check the edition your authority having jurisdiction has adopted.

Can any solar system be retrofitted with a battery?

Most grid-tied systems can take an AC-coupled battery without changing the PV inverter. DC coupling usually means replacing the PV inverter with a hybrid unit. Check the battery manufacturer’s compatibility list for the existing inverter before you quote.

How long does a solar battery last?

Lifespan depends on the product, depth of discharge, temperature and usage, so read the manufacturer’s warranty for years, cycles and energy throughput. Warranties often cap whichever limit is reached first. Model year-10 capacity, not day-one capacity, when you size the system.

Is battery storage worth it with net metering?

With full retail net metering, a battery saves little on the bill because exports already earn the retail price, so the case rests on backup value. Where export rates are lower or time-of-use prices are high, a battery can raise self-consumption and improve payback. Model the actual tariff with and without storage.