Commercial Solar Inverter Sizing for 100kW+ Projects

Set the Commercial Sizing Basis Before Selecting Inverter Capacity

Before you pick an inverter, get the basics right: the point of interconnection, interval load data, your PV production model, and the operating capacity the project actually needs. Don’t size off the PV nameplate alone . For anything above 100 kW, nail down these constraints before you start comparing inverter topologies or sending out procurement packages.

Start With the Load Side

Pull at least 12 months of interval demand data. Look at daytime demand, seasonal peaks, any planned electrification, tenant changes, and new equipment or processes coming online. At the same time, confirm the export cap, any reverse power restrictions, curtailment requirements, and the grid code that applies to your project.

Six Things to Settle Before Comparing Topologies

Your PV production model needs to account for irradiance, orientation, shading, module temperature, losses, degradation, and curtailment. Alongside that, test the DC to AC ratio at a few different levels and see how each one affects clipping, export limits, inverter loading, and installed cost.

The string electrical window matters too. Check that cold-condition open-circuit voltage and hot-condition operating voltage both sit within your inverter’s limits. And when you’re allocating MPPTs, keep roof planes, shading patterns, and modules with different operating conditions grouped with the right inputs.

Grid compliance is its own checklist: voltage, frequency, reactive power, ride-through, remote control, and export limits. Last but not least, model the final planned AC capacity for your growth case, not just what you need for phase one.

Export Limit vs. Inverter Rating: Two Separate Decisions

Here’s something worth pausing on: your export limit and your inverter’s AC nameplate rating are two different decisions, not one. You could pair a higher-rated inverter with active export curtailment, which changes where and how clipping happens. Or you could size the inverter’s AC rating to match the export cap directly, that gives you a different outcome. Both approaches can satisfy the same interconnection limit, but they land on different installed costs, control complexity, and clipping-loss profiles. It’s worth modeling both before you commit.

A quick example helps show why this matters. Say you have a 150 kWp array and a 100 kW export limit. Your production model shows 12,000 kWh a year landing above that 100 kW AC threshold, after accounting for temperature and losses. In that case, a 100 kW AC design will clip that energy, unless you add export control or find a way to absorb the extra load. Whether that loss is worth accepting comes down to three things: how much that clipped energy is worth, what it costs to add more AC capacity, and where your interconnection limit sits.

Whatever you decide, write it down. Every input and decision you record becomes the reference point later — for transformer sizing, redundancy planning, expansion, warranty coverage, and the financial model.

String vs Central Inverter Commercial: A TCO Perspective

String inverters spread the conversion job across multiple units. Central inverters concentrate it in one or a few larger blocks instead. Neither one is automatically cheaper for a 250 kW to 1 MW project. The choice ripples through DC collection, AC feeders, site access, outage risk, and how easily you can expand later.

Building an Equivalent Cost Case

To compare fairly, build cost cases with the same DC array, weather file, export limit, and project term for each option. Include everything: inverter supply, DC and AC BOS, installation labour, access equipment, commissioning, planned replacement, spare units, and outage costs. In short, don’t just compare inverter sticker price, that alone tells you very little.

Here’s why the fault-impact row matters in practice. A 1 MW plant built from ten 100 kW inverter blocks only loses 10% of nominal AC capacity if one block goes down. Build the same plant with a single 1 MW conversion block, and one outage takes out 100% of your capacity. So calculate the actual revenue hit using site-specific production during the repair window and your real energy value, not a rough guess.

Design Inverter Banks for Availability, Fault Isolation, and Critical Loads

An N+1 inverter bank simply means there’s enough installed capacity to still hit your operating target even after the largest block goes down. But that protection is only real if the surviving capacity, thermal conditions, protection settings, and site layout all actually support it.

N, N+1, and N+2 — What the Labels Mean

At N, every installed block is needed just to meet the target, there’s no slack. Add one extra block and you’re at N+1: lose a single designated block, and you can still hit the target. Go to N+2 and you can lose two blocks, as long as that matches a defined operational requirement.

Calculating Surviving Capacity

Work out firm capacity from the blocks that survive a failure, not your total nameplate capacity. Here’s a quick example: a 100 kW load at 0.8 power factor needs 125 kVA, because kVA is just kW divided by power factor — 100 ÷ 0.8 = 125. That 0.8 is just an illustration, though, not a number to assume by default. Motor-heavy or mixed commercial loads often run at a lower power factor, so pull the actual site load list and its real power factor before you size surviving kVA. Once you have that, check the available inverter kVA against ambient-temperature derating, transformer loading, and whatever operating mode you need.

One thing worth flagging: standard grid-tied inverters disconnect automatically during a utility outage. That means N+1 grid-tied capacity doesn’t give you backup power on its own. It only does if you’ve built in approved islanding controls, a grid-forming or backup-capable architecture, protection coordination, and the right authority approvals.

Setting and Testing the Redundancy Level

Base your redundancy level on a realistic repair-time assumption, how easy spares are to get locally, access constraints, and what lost output would actually cost you. A spare that’s sitting on a shelf but can’t be isolated, configured, tested, and put into service within your recovery window isn’t really usable redundancy, it just looks like it on paper. So commission the failed-block scenario for real. Isolate the designated inverter, confirm the rest of the bank stays within voltage and thermal limits, check that protection selectivity still works, and write the restoration steps into the handover package.

How Much Redundancy Is Justified for a Hospital Solar Installation?

For hospital projects, build redundancy around the solar system’s actual job and the facility’s emergency-power setup. Don’t just build around a general sense that “more redundancy is safer.” An N+1 PV inverter design can keep a specified generation level running after one block fails, but it won’t keep life-safety loads powered during a grid outage on its own. That’s the emergency power system’s job, not the PV array’s.

Start by defining what needs to happen after a failure. That might be a minimum grid-connected PV export level, a cap on acceptable revenue loss, or a target contribution to an engineered microgrid. From there, calculate available kVA after removing the largest block, and check it against five things:

• critical-load kW and its governing power factor, taken from the actual facility load list rather than an assumed value
• inverter thermal derating at expected equipment temperature
• transformer and switchgear capacity
• PV input limits under credible irradiance conditions
• planned maintenance operating requirements

Then weigh three procurement options: a cold spare, an installed spare inverter block, or a contracted service response. The financial comparison should cover spare cost, storage or installation cost, repair duration, commissioning labour, and the value of the energy you’d lose. Pick whichever option actually meets the recovery-time requirement the facility and its electrical design authority have set.

Build for Load Growth With Modular Capacity Expansion

A scalable 1 MW design reserves the electrical and control interfaces you’ll need for the final build-out, and it does this before phase one is even energized. Empty switchgear space alone won’t cut it, though, if the upstream bus, transformer, protection, communications network, or utility agreement can’t carry the future load.

Reserve five things before you energize phase one:

• Switchgear and feeders: cubicle positions, busbar rating, breaker capacity, feeder routes, and fault-duty capability
• Transformer: final-capacity thermal loading, tap range, impedance effects, and MV protection capacity, checked against the ultimate build-out rather than just phase one
• DC infrastructure: conduit, trench, combiner locations, and MPPT allocation for the later array blocks
• Controls and protection: network ports, addresses, relay inputs, CT capacity, and protection settings for each phase
• Interconnection: model the ultimate export case early, since later capacity changes can trigger new utility studies or metering requirements

If your release schedule is still uncertain, standardize the inverter blocks now. Consistent feeder arrangements, protection settings, communications templates, and commissioning records mean fewer engineering changes down the line. A centralized architecture tends to suit a single land parcel where you’re installing collection infrastructure once. Distributed blocks, on the other hand, work better for staged rooftops or expansion driven by individual tenants.

At every phase gate, test against the final design case: breaker interrupting duty, voltage rise, transformer loading, export controls, network capacity, and protection selectivity. That way, a finished PV expansion never runs into a wall built by an AC system that was only ever designed for phase one.

Specify Fleet Features That Lower Portfolio O&M Cost

Fleet features only lower O&M cost when they actually cut down a measurable task: diagnosis time, truck rolls, configuration rework, replacement time, that sort of thing. So your procurement spec should say exactly how each capability gets tested before it’s accepted, not just describe it in marketing terms.

A good alarm system tells grid events, communications loss, insulation faults, thermal derating, and internal faults apart from each other. A generic “offline” notification, by contrast, tells you nothing about what tools, parts, or skills the fix actually needs.

For remote fleet visibility, weigh EPEVER Products for Remote Monitoring & Control against your project’s real communications setup and supported equipment — not a generic feature list. And wherever remote access is allowed, insist on named accounts, role-based access, credential management, and logs of every configuration change.

Before handover, validate the telemetry properly. Check timestamp accuracy, whether communications recover after an interruption, alarm routing, exported event records, and the documented procedure for restoring a replacement inverter’s configuration.

What Inverter Features Actually Reduce O&M Costs Across a Portfolio?

Portfolio-scale O&M costs come down when the operator can do three things quickly: spot the affected device, work out a likely cause, and send someone out with the right information. So the spec should focus on what evidence is available before a technician even leaves the office.

Actionable alarms need a code, timestamp, severity level, and the affected inverter or input, that’s enough to tell a passing grid event apart from something that actually needs a truck roll. Device-level operating data should also capture DC, AC, temperature, insulation, and communications readings per inverter, so service teams can compare the problem unit against the healthy ones. Configuration control needs named access, approval workflows, audit records, and configuration backups, so nobody makes an undocumented parameter change. Data access should be interoperable too. Documented communications mapping and exportable event history for whatever monitoring or SCADA system the project uses. And finally, the replacement process needs serial-number traceability, a clear RMA route, approved configuration records, and an agreed spare strategy.

During commissioning, walk through the whole workflow. From the initial fault, through alarm receipt, diagnosis, isolation, replacement configuration, and back to service. Time each step. That baseline is what your operations contract should be measured against.

Translate Inverter Design Choices Into 20-Year Project IRR

If IRR (internal rate of return) is new to you, here’s the short version: it’s the annualized return that a project’s cash flows imply, once you account for the upfront investment and every year of income after that. It’s the single number most investors and lenders use to judge whether a solar project clears their bar. So an inverter that looks cheaper on paper still needs to earn that reputation in the IRR model. Cost alone doesn’t settle the question.

Inverter selection feeds into a 20-year IRR through installed cost, energy yield, availability, replacement timing, and service cost. Whatever financial model you build needs to use the same technical assumptions as the EPC’s single-line diagram and operating plan, otherwise the numbers won’t mean much.

Five Inputs the Model Needs to Handle Consistently

Treat five categories of input the same way every time. Year-zero CAPEX covers inverter supply, BOS, transformers, switchgear interfaces, installation, commissioning, and spares. Production draws on site irradiance, module temperature, orientation, shading, losses, degradation, clipping, and curtailment. Availability turns each fault scenario into affected capacity and outage hours. O&M and replacement covers monitoring fees, the planned replacement year, warranty scope, labour, access equipment, and spare logistics. And revenue applies the contracted export value, tariff schedule, incentives, and demand-charge treatment, wherever interval demand data supports it.

A Worked Fault-Cost Example

Here’s how to calculate lost generation from a fault: multiply the modeled production during the affected block’s downtime by how long the outage lasts. The block’s nameplate capacity only sets the ceiling on that loss. What actually matters is modeled output, not nameplate capacity, since the block would rarely be running at full rated output for the whole outage anyway.

Say a 50 kW block is down for a cumulative 20 daylight hours across the repair window, spread over two partial days. If your production model forecasts an average of 30 kW during those hours, lost generation comes out to 30 kW × 20 h = 600 kWh. That 50 kW nameplate figure only tells you the theoretical ceiling. It doesn’t enter the actual math. From there, multiply 600 kWh by your applicable energy value and add the restoration cost.

When N+1 Actually Pays for Itself

Only include an N+1 design once you’ve weighed its extra CAPEX and carrying cost against the outage cost it would actually avoid. Run sensitivity cases across installed CAPEX, availability, replacement year, tariff value, clipping, and curtailment. A topology decision holds up when the model clearly shows which assumption is driving the result — not when that driver is buried inside a blended contingency line.

Before financial close, reconcile everything against warranty terms, service-level commitments, approved interconnection limits, and the final protection and monitoring design.

Choose a Scalable Commercial Inverter Architecture With a Defensible Investment Case

The way to choose a commercial inverter architecture is to prove it meets four things at once: the interconnection limit, the operating objective, the expansion plan, and the lifecycle financial case. For projects at 100 kW and up, that means working through four steps in order. Start with the sizing basis. Then compare topology-level installed and service costs, define what degraded operation looks like, and finally test the full build-out in the financial model.

Use interval load data and a site-specific production model to set your AC capacity and DC to AC ratio. Compare distributed and centralized architectures under identical PV, export, and weather assumptions. Wherever availability really matters, calculate surviving kVA after pulling out the designated inverter block, and be clear about the difference between grid-connected generation redundancy and actual outage backup.

For phased projects, reserve the final AC, protection, transformer, communications, and interconnection capacity before phase one goes live. At handover, verify export controls, protection selectivity, telemetry, alarm routing, and the documented replacement workflow.

The EPEVER SPT Series On-Grid PV Inverter is worth evaluating as a modular block option for the right commercial projects. As with any candidate, though, run it through the same checks: approved electrical ratings, grid-code compatibility, monitoring integration, access conditions, service coverage, and expansion compatibility.

FAQ

How much energy will a 100 kW solar system produce per month?

You can’t get monthly production from the 100 kW rating alone, it just doesn’t tell you enough. Instead, run a site-specific simulation that accounts for irradiance, orientation, shading, module temperature, DC to AC ratio, soiling, electrical losses, degradation, curtailment, and availability. And use modeled monthly numbers, not an annual average, when you’re weighing tariff value and export limits.

Which technical specifications matter most for a commercial on-grid inverter?

Start with AC voltage, kVA rating, maximum DC voltage and current, MPPT operating range, and input allocation. Then check grid-code functions, thermal derating, protection functions, communications protocols, monitoring compatibility, warranty scope, service process, and replacement lead time. Every one of these needs to line up with your approved single-line diagram, string design, interconnection agreement, and actual operating environment.

When is N+1 inverter redundancy justified on a commercial PV project?

N+1 earns its keep when the modeled cost of lost generation, contractual availability exposure, or a defined operating requirement outweighs the extra cost of another inverter block and its supporting infrastructure. Start by calculating surviving kVA after one block fails, using your actual site load list and power factor rather than an assumed number. Then test repair time, spare availability, thermal derating, transformer limits, and how ready the site is to commission a replacement.