Engineering Deep Dive: Sizing Solar Arrays for Maximum ROI in Commercial Projects

Reading time: ~12 min read

By Dr. James Okonkwo, PE — Licensed Electrical Engineer, in-house, 12 years commercial solar design. Senior Design Engineer at PES Supply.

Commercial solar is not residential solar scaled up. The engineering decisions are fundamentally different: interconnection costs that scale nonlinearly with system size, demand charge structures that reward peak-shaving rather than maximum energy, utility tariffs with time-of-use components that change the value of every kilowatt-hour depending on when it's generated, and DC/AC ratios that interact with climate, module technology, and inverter topology in ways that can swing project IRR by 3–5 percentage points. This article is my engineering methodology for sizing commercial solar arrays — the calculations I run, the tools I use, and the sensitivities I model — illustrated with a real 500 kW commercial rooftop case study.

For the equipment specifications referenced throughout this analysis, browse our commercial solar panels, commercial inverters, and racking and mounting systems. PES Supply stocks 50,000+ SKUs from 169 authorized brands with delivery in 7–10 business days.

Commercial array sizing is an optimization problem with four interdependent variables:

• DC array capacity (kWp): The total nameplate rating of all modules on the roof or site.
• AC inverter capacity (kW): The total AC output rating of all inverters in the system.
• DC/AC ratio (inverter loading ratio): The ratio of DC array capacity to AC inverter capacity — the single most powerful economic lever in commercial design.
• Interconnection limit (kW): The maximum AC export capacity permitted by the utility's interconnection agreement.

The objective function is not maximum energy production. It is maximum net present value (NPV) — which accounts for energy revenue, demand charge reduction, installed cost, O&M expenses, and financing. A system that produces 5% more energy but costs 12% more to install may have a lower NPV. This distinction is where engineering meets finance, and it's why I always model IRR with sensitivity tables rather than presenting a single-point answer.

The DC/AC ratio — also called the inverter loading ratio — is the ratio of installed DC array capacity to inverter AC capacity. A ratio of 1.30 means the array is rated 30% higher than the inverter. This matters because solar arrays rarely operate at full STC nameplate: the sun is low for much of the day, modules heat up and lose efficiency, and real-world conditions rarely match the 1000 W/m², 25°C cell temperature test standard ([SurgePV](https://www.surgepv.com/blog/central-inverter-design)).

By oversizing the DC array relative to the inverter, you capture more energy during morning and evening shoulder hours — when irradiance is below STC but still meaningful — at the cost of some clipping at solar noon when DC output exceeds the inverter's AC capacity. The economic question is whether the shoulder-hour gains exceed the midday losses.

Source: DC/AC ratio recommendations from commercial inverter design analysis ([SurgePV](https://www.surgepv.com/blog/central-inverter-design)).

The optimal DC/AC ratio is not uniform across geographies. Desert sites (Arizona, Nevada, Southern California) can justify higher ratios than temperate sites (Pacific Northwest, Northeast) for two reasons:

• High irradiance: More annual hours near STC means the DC array spends more time at capacity, but the absolute shoulder-hour gains are larger because the irradiance curve is broader.
• High module temperature: Desert heat raises cell temperatures to 60–70°C, reducing module output by 15–20% from STC (at a temperature coefficient of approximately -0.35%/°C above 25°C). This naturally caps DC output below the inverter limit, reducing clipping even at higher ratios.

In cooler, cloudier climates, the irradiance curve is narrower and module temperatures stay closer to STC, meaning a high DC/AC ratio produces more clipping relative to shoulder-hour gains. For a commercial rooftop in Phoenix, I might model DC/AC ratios of 1.30, 1.35, and 1.40. For the same rooftop in Seattle, I'd model 1.15, 1.20, and 1.25.

Clipping is the energy lost when DC output exceeds the inverter AC limit around solar noon. The common misconception is that clipping is pure loss. It is not: the extra DC capacity produces more energy in shoulder hours than it loses at midday. Annual clipping of 2–3% is economically acceptable on most commercial projects because the capital cost savings of a smaller inverter plus the shoulder-hour gains more than compensate ([SurgePV](https://www.surgepv.com/blog/central-inverter-design)).

The design error to avoid is setting the ratio without modeling the specific site. I run each candidate ratio through PVsyst or SAM with site-specific TMY (Typical Meteorological Year) data, then compare LCOE across ratios of 1.25, 1.35, and 1.45. The selection criterion is the ratio that minimizes LCOE — not the ratio that minimizes clipping.

Accurate shading analysis is non-negotiable for commercial design. The two industry-standard tools take different approaches:

PVsyst is the gold standard for bankable yield estimates, particularly on utility-scale projects above 5 MW where lenders specify PVsyst reports by name. It produces 8,760-hour simulations with detailed soiling, mismatch, and wiring loss models. The output is an independent-engineer-acceptable report format that lenders and tax equity investors trust ([Heaven Designs](https://heavendesigns.in/blog/helioscope-vs-pvsyst/)). PVsyst is desktop-based, steeper to learn, and more granular in its loss parameterization.

HelioScope, built by Folsom Labs, is the cloud-native tool of choice for commercial and industrial projects between 100 kW and 5 MW, where speed and collaborative review matter more than the .PRJ file format. Simulations complete in 30–60 seconds, and the platform has been DNV-validated to within 1% of PVsyst for typical C&I projects ([Solar Apps Directory](https://apps.list.solar/tools/helioscope/)). It also earned bankable recognition from Sunstone Credit in October 2025 — the first web-based solar platform to do so.

Source: Tool comparison from solar design software analysis ([Heaven Designs](https://heavendesigns.in/blog/helioscope-vs-pvsyst/), [Solar Apps Directory](https://apps.list.solar/tools/helioscope/)).

For commercial rooftop design, I use HelioScope for initial layout and rapid scenario iteration, then validate the final design in PVsyst when the project requires a bankable yield report for financing. For projects under 500 kW where financing doesn't require a PVsyst report, HelioScope alone is sufficient.

Before investing time in PVsyst or HelioScope modeling, I use NREL's PVWatts calculator for initial feasibility assessment. PVWatts provides a quick estimate of annual energy production based on location, system size, module type, array tilt and azimuth, and system losses. While it doesn't model site-specific shading, it's the right tool for the first 10 minutes of a project — confirming that the solar resource and system size produce enough energy to justify deeper analysis.

PVWatts uses a default system loss of 14% (soiling, mismatch, wiring, availability), which I typically increase to 18–20% for commercial rooftops with partial shading and inverter loading ratios above 1.25. The tool draws on NREL's National Solar Radiation Database (NSRDB), which provides TMY3 data for most U.S. locations.

The interconnection process is where many commercial solar projects stall. Before purchasing equipment, the installer or developer must apply for and receive an approved interconnection agreement from the local serving utility. The application package includes a completed interconnection form, site plan, electrical single-line diagram, and equipment specification sheets showing the exact inverter model and its UL 1741 certification listing ([Solar Permit Solutions](https://www.solarpermitsolutions.com/blog/utility-interconnection-guide-for-solar-installers)).

Utilities use this submission to run a hosting capacity screen — checking whether the local distribution transformer and feeder can absorb the proposed generation without upgrades. For systems under the utility's simplified review threshold (typically 10–50 kW, varying by utility), approval is often administrative. For larger commercial systems, a supplemental or full interconnection study may be required, which can take 4–12 months and cost $5,000–$50,000.

Every inverter interfacing with the grid must carry UL 1741 listing, which verifies the inverter meets the interconnection requirements defined in IEEE 1547 — the industry standard for interconnecting distributed energy resources with electric power systems. The 2018 revision of IEEE 1547 fundamentally changed the requirements:

• Anti-islanding: Inverters must disconnect from the grid within a fraction of a second during a power outage, protecting utility line workers.
• Voltage and frequency ride-through: The 2003 edition directed DERs to disconnect on any abnormal voltage or frequency. The 2018 revision replaced this with three ride-through categories, requiring Category III inverters to stay online at 0 per-unit voltage for up to one full second — shifting the burden from passive disconnection to active grid support ([REIG Solar](https://www.reig-us.com/ieee-1547-solar-interconnection-voltage-ride-through-guide/)).
• Reactive power support: The 2018 standard requires inverters to provide reactive power support within specified ranges, participating in local voltage regulation.
• Size cap removed: The 2003 edition applied only to resources rated 10 MVA or less; the 2018 revision removed that fixed cap ([Entogo](https://entogo.ca/insights/ieee-1547-der-grid-interconnection/)).

Your interconnection agreement specifies the required ride-through category. Verify the inverter is configured correctly during commissioning — incorrect grid profile settings are a common cause of PTO delays.

For commercial customers, demand charges often represent 30–50% of the total electric bill. These charges are based on the highest 15- or 30-minute power draw during the billing period — not total energy consumed. A solar array that reduces peak demand can deliver value far beyond its energy production.

My demand charge analysis methodology:

1. Obtain 12 months of interval meter data from the customer's utility. This is non-negotiable — you cannot model demand charge reduction with monthly billing data alone.
2. Identify the demand charge structure: Is it a single peak demand charge, or time-of-use demand (e.g., on-peak vs. mid-peak vs. off-peak)? Are there ratchet clauses that lock in a minimum demand based on a historical peak?
3. Overlay the solar production profile (from PVsyst or HelioScope) against the load profile for each interval. Calculate the net demand (load minus solar) for each 15-minute interval.
4. Determine peak demand reduction: Compare the maximum net demand with solar to the maximum gross demand without solar for each billing period.
5. Value the reduction: Multiply the kW reduction by the demand charge rate ($/kW) for each billing period and sum across the year.

The key insight: a south-facing array peaks at solar noon, but many commercial loads peak between 3:00 PM and 6:00 PM. An array oriented slightly west of south (240° azimuth) can shift peak production to better align with late-afternoon demand, increasing demand charge savings even if total energy production drops slightly. This is a design optimization that pure energy-maximization models miss.

Internal Rate of Return (IRR) is the financial metric that determines whether a commercial solar project gets built. But a single-point IRR is misleading — it assumes every input is known with certainty. Real projects face uncertainty in energy prices, incentive programs, equipment degradation, and financing costs. I model IRR as a sensitivity table across these variables.

For each project, I build a two-way sensitivity table varying two key parameters and calculating IRR for each combination. The most common pairings:

• DC/AC ratio vs. installed cost ($/W): Shows the IRR tradeoff between inverter loading and capital cost.
• Energy price escalation vs. system size: Shows how future energy price uncertainty interacts with the decision to oversize or undersize the array.
• Module degradation rate vs. discount rate: Shows the sensitivity to long-term performance assumptions.

The table that follows illustrates the DC/AC ratio vs. installed cost sensitivity for our 500 kW case study, calculated with a 25-year analysis period, 30% federal ITC, 5-year MACRS depreciation, and 3% annual energy price escalation.

Note: These figures are illustrative for a 500 kW commercial rooftop in a mid-Atlantic climate zone with a demand charge of $18/kW and energy rate of $0.14/kWh. Actual results vary by location, tariff structure, and project-specific factors.

The table reveals the optimization sweet spot: at this site's climate and tariff structure, a DC/AC ratio of 1.30 maximizes IRR across most installed cost scenarios. Beyond 1.30, the incremental shoulder-hour gains are offset by increased clipping and the cost of additional modules. Below 1.20, the inverter is oversized relative to the array, leaving production capacity on the table.

• Location: Richmond, Virginia (Climate Zone 4A)
• Roof area: 35,000 sq ft of usable space
• Module: 550W bifacial TOPCon, 72-cell format
• Array configuration: 909 modules in 27 strings of 33 modules (south-facing, 10° tilt)
• DC capacity: 500.0 kWp
• Inverter: 2 × 200 kW string inverters (400 kW AC total)
• DC/AC ratio: 1.25
• Annual production: 685,000 kWh (modeled in HelioScope, validated in PVsyst)
• Capacity factor: 15.6%

The DC/AC ratio of 1.25 was selected over alternatives (1.20 and 1.30) based on the sensitivity analysis. At 1.20, the IRR was 12.9%; at 1.30, the marginal gain was 0.2 percentage points but required 36 additional modules, increasing roof loading and structural reinforcement costs that outweighed the energy gains. The 1.25 ratio represented the inflection point of the LCOE curve for this specific site.

The array was oriented at 195° azimuth (slightly west of south) rather than due south to better align peak production with the facility's 3:00–6:00 PM demand peak. This reduced annual energy production by approximately 1.2% but increased demand charge savings by $4,800 per year — a net positive for NPV.

• Maximizing energy instead of NPV: The system that produces the most kWh is not always the most profitable. Demand charge reduction and time-of-use arbitrage can favor different orientations and sizes.
• Ignoring interconnection costs: A 500 kW system that requires a transformer upgrade or feeder reinforcement can see installed costs jump by $0.30–0.50/W. Model interconnection early.
• Using default loss parameters: PVWatts' 14% default loss doesn't account for commercial-specific losses like HVAC interference shading, roof access pathways, and fire setback requirements. Use 18–20% for realistic commercial modeling.
• Oversizing the DC/AC ratio without climate-specific modeling: A ratio that works in Arizona can be wrong in Ohio. Always model with site-specific TMY data.
• Neglecting module degradation in Year 1 vs. Year 25: A system sized for Year 1 peak demand may not deliver the same demand charge reduction in Year 15. Model degradation explicitly.

Commercial solar sizing is a multi-variable optimization that requires climate-specific modeling, tariff-aware economic analysis, and sensitivity testing across the parameters that matter most to your project's financial outcome. The tools exist — PVsyst, HelioScope, NREL PVWatts, SAM — but the engineering judgment to interpret their outputs and make the right tradeoffs is what separates adequate design from optimal design.

When you're ready to specify equipment for your commercial project, PES Supply has the inventory to support it: 50,000+ SKUs from 169 authorized brands, including commercial-grade modules, string and central inverters, and complete balance-of-system components. Delivery is 7–10 business days. Browse our solar panels, commercial inverters, racking systems, and electrical disconnects to build your specification package.

For engineering support on commercial project sizing, contact PES Supply's technical team. I review project specifications and provide equipment recommendations based on the methodology outlined in this article.