Sizing from an AC capacity

The normal way round is to fill the site: place as many modules as the usable area holds and read the resulting capacity off the summary. 🔆 Simulation with AC Capacity works the other way. You state the AC capacity you are contracted for and the DC to AC ratio you want to build at, and the plant is sized to match it.

That is the shape most utility-scale projects actually arrive in. The connection agreement fixes the megawatts you may export, the tariff is set against them, and the design question is not how much the site could hold but how much of the site you should use to deliver exactly what you have sold.

What it needs

Without both files the simulation has nothing to size a string or an inverter from, so both are requirements rather than refinements here — unlike a plain layout run, which works from typed dimensions alone.

How it sizes

The string geometry comes first, then the inverter count, then the DC target.

Modules in series are worked out by the same string sizing method used elsewhere in the application — the module's voltage behaviour at the site's temperature extremes against the inverter's voltage window. See String sizing.

Parallel strings per inverter come from the inverter's own limits: its input current limit and its DC power limit, whichever binds first.

The inverter count follows from the target:

num_inverters = ceil( AC_capacity / Pac_rated )

installed_ac  = num_inverters × Pac_rated  ≥  AC_capacity

Rounding up is the important behaviour: it never under-sizes. The installed AC capacity is always at or above your target, never below it. Where the inverter rating does not divide the target exactly you get the next whole inverter and a slightly higher installed figure — deliberately, because the alternative is a plant that cannot deliver what it was contracted for.

DC overload and the DC to AC ratio

The DC to AC ratio is installed DC capacity divided by installed AC capacity. A ratio of 1.4 means 1.4 MWp of modules behind every 1 MW of inverter.

Both figures are the design overload limit, and both are adjustable by you.

Why a ratio above 1 is normal

An array almost never produces its nameplate DC output. Nameplate is measured at standard test conditions, STC, and a real array is hotter than the STC cell temperature whenever irradiance is high, and below the STC irradiance level for most of the hours it generates at all. The result is that an inverter sized one to one with the array sits well below its rating for nearly all of the year.

Deliberately fitting more DC than the inverter can pass fills that gap. The array reaches the inverter's limit more often and for longer, the inverter runs nearer its efficient operating region, and on the handful of hours when the array would exceed the limit the inverter clips the surplus — it holds its output at the rating and the extra DC is not converted. The energy given up in clipping is small; the energy gained across every other hour is not. This is why utility-scale plants are routinely DC-overloaded rather than matched.

The two outcomes

Once the target DC capacity is known, it is compared against what the first layout actually placed.

The target is at or below the first run's capacity

The site can deliver it. You are offered a regenerated layout capped to the target: the same site, the same settings, with placement trimmed so the installed DC capacity lands on your target instead of filling the ground.

This is the normal path. The capped layout is the one to review, export and issue, because it is the plant matching the connection you hold.

The target is above the first run's capacity

The site cannot deliver it, and no regeneration can change that — the first run already used all the usable area there is. You are advised to reduce the AC capacity or the DC to AC ratio.

Before doing either, check whether the site is really the binding constraint. The usable area is the boundary shrunk by the perimeter road width, with obstructions, corridors and terrain exclusions subtracted, so a target that misses by a little may be reachable through the layout inputs rather than the commercial ones — a narrower perimeter road, half tables enabled, or Maximize placement on an awkward boundary. See Maximize placement and half tables.

Why a capped plant has empty ground

A capped layout looks wrong at first sight. There is bare ground inside the fence that tables would obviously fit on, and the instinct is that the trim overshot. It did not, and the behaviour is worth understanding before you explain the drawing to anyone else.

The delivered DC must land at or just above the target, never below. If the trim undershot, the installed DC would be less than the ratio you designed to, and the delivered DC to AC ratio would fall short of the figure the whole exercise was built around. Landing slightly high is a rounding artefact; landing low is a failure to deliver the design.

Trimmed tables are held back and returned afterwards. The stages that run after placement clear more ground: control rooms are re-placed against the reduced capacity, lightning arresters take their footprints, inverter pads take theirs. Ground that was occupied when the trim was calculated may be free by the time those stages finish. So the trimmed tables are kept in reserve and put back after them, and any whose ground has since been taken by other equipment is skipped.

Half units are weighted at half throughout. A half table counts as 0.5 wherever tables are counted in this process, so a plant using half tables is trimmed on the same weighted basis as one without them.

The consequence: the empty ground on a capped plant is deliberate. It is the capacity you chose not to build, and it is where a future extension goes if the connection is ever increased.

Where the result appears

Three columns at the right-hand end of the summary table carry the outcome:

Read all three together, and read them against the target you typed. DC/AC is the check that matters: it is the delivered ratio, computed from what was actually placed, and it is the figure that tells you the cap did what you asked. A delivered ratio marginally above the target is the expected result of rounding inverters up and landing DC just above target. See Summary columns for the full column set.

The capped layout carries into everything downstream unchanged — the report, the drawing exports, the bill of materials and the energy calculation all describe the capped plant, so run the energy calculation again after a regeneration.

Where it fits in the design sequence

Where to go next