How to Size and Model a Commercial Solar System

Why Sizing Starts With Load, Not Roof Area

The residential habit is to look at a roof, count the modules that fit, and call that the system. On a commercial building that logic runs backward. A warehouse with four acres of flat roof might draw almost nothing during the day, while a cold-storage facility on half the footprint runs compressors around the clock. The roof tells you how big the array can be. It says nothing about how big it should be. That answer lives in the meter.

So the first file a designer opens is not a site plan. It is the facility's interval-load data and its rate schedule. Those two together describe what the building actually spends money on and when, and they define the shape the array is trying to match. A commercial bill is rarely a flat price per kilowatt-hour. It stacks an energy charge, a demand charge tied to the single highest peak in the month, and often time-of-use periods that reprice the same kilowatt-hour depending on the hour. The U.S. Energy Information Administration lays out how retail prices reflect fuel costs, grid investment, distribution, and regional supply and demand, and commercial tariffs slice those costs into parts that respond very differently to solar.

Get the order right and everything downstream holds together. Size to the load and the rate, confirm the roof or the ground can hold the result, then model the production and read the money. Size to the roof first and you build a system that fits beautifully and offsets the wrong thing.

Inputs to Gather Before You Size Anything

A commercial sizing exercise is only as good as the data behind it, and most of the delay on a project comes from chasing inputs late. Pull this set up front, before a single module is placed, and the rest of the work moves without backtracking.

• Twelve months of interval-load data, ideally at 15-minute resolution, pulled from the utility or the building's advanced meter.
• The exact rate schedule, including the energy rate, the demand charge, and any time-of-use periods and seasonal changes.
• Twelve months of utility bills to cross-check annual kilowatt-hours and the monthly billed demand peaks.
• Roof plans with age, membrane type, structural capacity, and the locations of every rooftop unit, skylight, and vent.
• For ground mount, a site survey with usable acreage, grading, setbacks, and soil or geotechnical notes.
• Orientation and available tilt for each roof plane or racking row, since these shift the production curve against the load.
• A year-round shade and horizon study, not a single clear-day snapshot.
• The utility's interconnection rules and the service or transformer rating that caps how much the site can export.
• Any planned load changes, such as added refrigeration, EV charging, or a production line, that will move the profile.
• The customer's goal stated plainly: lowest bill, a target offset percentage, a demand-charge ceiling, or a return threshold.

Reading Interval Load Data

Interval data is the record of how much the building drew in each short window across the year, usually every 15 minutes. A monthly bill flattens all of that into one number. The interval file keeps the shape, and the shape is where sizing decisions come from. Two facilities can buy the identical annual kilowatt-hours and need completely different systems because one draws steadily through daylight and the other spikes for three hours each afternoon.

Start by plotting a few representative days across seasons. Look for the daytime base load, the hours when consumption lines up with sunlight, and the timing of the monthly peak that sets the demand charge. A building whose peak lands at two in the afternoon is a strong match for a south-facing array. One that peaks at seven in the evening, after the sun has dropped, will not see its demand charge move much from solar alone, which is the first signal that storage belongs in the conversation.

Watch for data gaps, meter rollovers, and obvious anomalies, because a single bad week can distort an annual model. Reconcile the interval total against the summed bills so you know the file is complete. The EIA publishes average retail electricity prices by sector, which is a useful sanity check on the implied rate, but the real value of interval data is the timing it preserves, not the average it produces.

It also helps to separate weekday profiles from weekends and to flag any single-shift, two-shift, or round-the-clock pattern, since the way a facility runs changes which hours the array needs to cover. A single-shift operation that goes quiet by late afternoon rewards a very different system than a plant that holds steady load overnight. Reading those patterns out of the raw file, rather than assuming a generic commercial curve, is what keeps the rest of the sizing grounded in how the building actually behaves.

How the Rate Structure Steers the Array

The tariff decides what a kilowatt-hour of solar is worth, and that worth is not constant. Three components matter, and each pushes sizing a different way.

The energy charge, billed per kilowatt-hour, is the straightforward one. Every kilowatt-hour the array produces and the building uses avoids one purchased at that rate. If energy charges dominate the bill, sizing to annual consumption gets you most of the value.

The demand charge is billed on the single highest short-interval peak in the month, in dollars per kilowatt, and it rewards a very different array. Demand charges can be a large slice of a commercial bill, and solar only trims them when production is high at the exact moment the peak occurs. A midday peak is reachable. An evening peak usually is not, which is where battery storage earns its place by discharging into the spike.

Time-of-use periods reprice the same kilowatt-hour by hour and season. Under these schedules, production during a peak afternoon window is worth far more than the same energy at midday off-peak. That changes the optimal orientation, sometimes favoring a west-facing tilt that lines production up with the expensive hours rather than with maximum daily output. The EIA notes that retail prices move with regional supply, demand, and the cost of delivery, and time-of-use rates are the utility passing that hourly variation straight through to the customer.

The Inputs That Actually Drive Sizing

Every sizing input earns its place by changing the answer in a specific way. The table below ties each one to what it controls and where it comes from, so nothing gets guessed at the desk.

Sizing the Array to Energy and Demand

With the load and the tariff in hand, sizing becomes a matching problem. You are fitting a production curve against a consumption curve, hour by hour, and scoring the fit in dollars rather than in kilowatt-hours offset. Three target modes cover most commercial jobs.

Sizing to annual energy aims the array at total consumption, typically landing somewhere short of full offset so that midday overproduction does not spill onto the grid at a poor export rate. This mode suits buildings where energy charges carry the bill and daytime load tracks the sun. Sizing to the daytime base load is the more conservative version, where the array is held to what the building reliably consumes during production hours, so nearly every kilowatt-hour is self-consumed at the full retail energy rate.

Sizing against the demand charge is a separate exercise. Here the question is how much of the monthly peak the array can reliably shave at the hour the peak occurs, and the honest answer is often less than the nameplate suggests, because peaks and clear skies do not always coincide. When the peak sits outside production hours, solar alone will not touch it, and the model should show that plainly instead of crediting a reduction that will not appear on the bill. That gap is the engineering case for pairing the array with storage, which is where Enact's commercial tools model the array and the battery against the same load profile rather than in isolation.

Roof and Site Limits as a Ceiling

Only after the load and tariff define a target does the physical site come in, and its job is to confirm or constrain, not to originate. On a rooftop, the usable area is what remains after fire setbacks, access pathways, and the footprints of every rooftop unit, skylight, and duct are subtracted. A roof that looks enormous on an aerial image often loses a third of its area to mechanical equipment and code clearances. Membrane age matters too, because mounting a 25-year array on a roof with eight years of life left forces an expensive removal and reinstall partway through.

Structural capacity is the harder ceiling. The roof has to carry the dead load of the racking and modules plus wind and snow, and many older commercial buildings need a structural review before anything is added. Ground mount trades those limits for acreage, grading, setbacks, and soil conditions that affect foundation type. In both cases the constraint is the same: the site tells you the largest system that is physically and legally buildable, and if that ceiling sits below the load-driven target, the array is capped there and the model is rerun against the real number.

The reverse case is just as common and easier to get wrong. When the buildable area far exceeds the load-driven target, the temptation is to fill it anyway. Resist that unless the economics of the extra export actually hold up under the tariff, because capacity the building cannot use is capital spent to sell cheap power back to the grid. The site ceiling is a boundary to respect, not a quota to hit.

Modeling Production Honestly

A number of panels is not a production estimate. Real output is what the array delivers after the site's irradiance, the module orientation and tilt, shading across the full year, temperature, soiling, and inverter and wiring losses all have their say. Multiplying nameplate watts by a generic sun-hours figure skips every one of those and produces a number the meter will not back up.

The U.S. Department of Energy explains that a photovoltaic system's real-world performance is shaped by factors including the available solar resource, orientation, temperature, soiling, and shading, rather than by the nameplate rating on its own.

U.S. Department of Energy, Solar Performance and Efficiency

Build the estimate on validated weather data for the specific location, apply a temperature coefficient, include soiling appropriate to the local climate, and account for inverter clipping when the array is deliberately oversized to its inverters. Run the shade analysis across the whole solar year, since a low winter sun and seasonal canopy strip output that a summer check never sees. The Department of Energy's Solar Energy Technologies Office documents how these performance factors and improving hardware shape project outcomes. Present the result as a modeled estimate with its assumptions stated, because on a commercial deal the output feeds a financial model that a finance team will stress-test line by line.

Offset, Self-Consumption, and Export

Once production is modeled, the meaningful step is laying that hourly production curve directly over the interval-load curve. Offset percentage alone hides the story. A system can offset 90 percent of annual kilowatt-hours on paper while dumping a large share onto the grid at midday and leaving the evening load untouched. What matters commercially is how much of the production the building consumes on site, because self-consumed energy is worth the full retail rate while exported energy is worth whatever the tariff pays for it, which is frequently much less.

Overlaying the two curves shows the self-consumption rate hour by hour and exposes the overproduction shoulders where the array is making more than the building can use. On many commercial loads the daytime base is high enough that self-consumption stays strong, which is one reason commercial solar economics often read well. Where production and load diverge, the overlay points to the fix, whether that is trimming array size, shifting orientation toward the priced hours, or adding storage to carry midday generation into an evening peak. National cost benchmarks such as EnergySage's solar cost data give a directional sense of installed pricing before a site-specific quote, but the sizing decision is driven by the overlay, not by a cost per watt.

Reading the Result as a Financial Picture

A sized system and a production model are inputs, not conclusions. The commercial decision is financial, and the sizing work only pays off when it is read in dollars. Translate the modeled production into avoided energy cost at the actual energy rate, avoided demand charges at the kilowatts the array reliably shaves during peaks, and any export credit the tariff allows. That yearly bill reduction is the top line everything else sits on.

Two federal benefits then reshape the picture for a tax-paying business. The commercial Investment Tax Credit covers a share of eligible project cost, and accelerated depreciation lets the owner recover most of the asset's basis on a short schedule. The Solar Energy Industries Association describes the ITC as one of the central federal mechanisms supporting solar deployment and treats accelerated depreciation of solar property as a separate benefit that lowers the effective cost of the asset. Keep these general and attributed, and label every return figure as a modeled output tied to stated assumptions. The sizing choice feeds straight into these numbers: a system sized to self-consumption at a high retail rate reads very differently in the model than one sized past the load that exports cheap kilowatt-hours. Right-sizing is the lever, and the financial read is where you confirm you pulled it correctly.

C&I Sizing Mistakes to Avoid

The same errors sink commercial sizing across markets and crews. Each one traces back to skipping the load or the rate and starting somewhere easier.

Sizing to the roof instead of the load. Filling the available area produces a system that fits and offsets the wrong thing, usually overproducing at midday and exporting at a poor rate.

Working from monthly bills alone. A monthly total hides the shape. Without interval data you cannot see the coincident peak, which means you cannot honestly estimate demand-charge savings.

Ignoring the demand charge. Treating a commercial bill as pure energy overstates the value solar delivers, because demand charges only fall when production is high at the peak hour.

Assuming a single orientation is best. Under time-of-use rates, maximum daily output is not the goal. Lining production up with the priced afternoon window can be worth more than a larger total at the wrong hours.

Estimating production as watts times sun hours. Skipping temperature, soiling, losses, and year-round shade inflates the kilowatt-hours and sets up a dispute when the meter reports the real figure.

Confusing offset with self-consumption. A high annual offset that leans on heavy midday export is worth far less than a tighter system that consumes nearly everything on site.

Checking interconnection last. Designing past the service rating or the utility's export cap forces a redesign after the customer has already seen the bigger number.

From Load File to Finished Model in One Place

These steps break down when they live in separate tools. Interval data in a spreadsheet, the tariff in a billing calculator, the array in a layout app, and production in a fourth program means every handoff is a chance for a number to drift, and the version the customer signs rarely matches the one the model produced. On a commercial deal, where a finance team will test each assumption, that drift is expensive.

Enact keeps the chain connected. You bring in the load profile and the rate structure, size the array against both, model production on validated site data, and read the energy, demand, and financial result in a single workflow built for commercial solar, with the same design and proposal platform carrying the model through to the document the customer sees. Because the load, the production estimate, and the savings share one model, the system you size is the system you propose, and the proposal rests on engineering a reviewer can follow back to the meter.

Frequently asked questions

How do you size a commercial solar system?

You size it from the facility's interval-load data and rate structure first, then test the result against roof or ground limits. Read the consumption shape and the mix of energy, demand, and time-of-use charges, size the array to match production against that load, model the output on validated site data, and confirm the site and interconnection can hold it. The roof caps the system, but the load and tariff set the target. The EIA explains the components behind commercial electricity prices that the sizing has to respond to.

How is commercial solar different from residential?

Commercial sizing works from interval data and a multi-part tariff, while residential often works from a roof and a flat or simpler rate. A commercial bill stacks energy charges, demand charges on the monthly peak, and frequently time-of-use periods, so the array has to match production to the load in time, not just in annual total. Self-consumption and demand-charge timing drive the design. See Enact's commercial solar tools for how load, production, and finance are modeled together.

What is a demand charge and how does it affect solar sizing?

A demand charge bills the single highest short-interval peak in the month, in dollars per kilowatt, and solar only reduces it when production is high at the exact hour of that peak. A midday peak is reachable by a well-oriented array, while an evening peak usually is not, which points toward battery storage. This is why interval data matters: without it you cannot see when the peak lands. The EIA's sector price data helps sanity-check the rate behind the bill.

Do you need interval data to size a commercial array?

For any building with meaningful demand charges or time-of-use rates, yes. Monthly bills give you an annual total but hide the consumption shape and the timing of the peak, and both decide how much value solar actually delivers. Interval data at 15-minute resolution lets you overlay production against load hour by hour and estimate demand savings honestly. Accurate production modeling then depends on the site factors the Department of Energy lists, applied to that load.

How much does commercial solar production depend on orientation and shading?

Enough that two arrays of identical size can produce very different output and very different value. The Department of Energy notes that performance is shaped by the solar resource, orientation, temperature, soiling, and shading rather than the nameplate rating (DOE). Under time-of-use rates, orientation also shifts production into or out of the priced afternoon window, so the best tilt for the bill is not always the one that maximizes daily kilowatt-hours.

Should a commercial array cover 100% of the load?

Not usually. Full annual offset often means overproducing at midday and exporting those kilowatt-hours at a low rate, which is worth far less than energy the building consumes on site at the full retail rate. Sizing to the daytime base load keeps self-consumption high, and the right target depends on the tariff and the customer's goal. The modeled production and the resulting economics, including the federal ITC, confirm where to land.

Sources

1. EIA - Electricity prices and factors affecting prices
2. EIA - Average price of electricity (FAQ)
3. DOE - Solar Performance and Efficiency
4. DOE - Solar Energy Technologies Office
5. SEIA - Solar Investment Tax Credit (ITC)
6. SEIA - Depreciation of Solar Energy Property
7. EnergySage - Solar Panel Cost
8. Enact - Commercial solar solutions
9. Enact - Design and proposal platform