Central inverter vs string inverter: telling them apart in overhead imagery

A VHR image of a site mid-construction shows boxy shapes and cable runs. Nothing in the frame carries a nameplate or label, so matching an inverter count against the EPC's progress report means reading the layout around each structure.

The two topologies get counted differently, and that difference changes what shows up on a draw schedule. A central inverter block is usually one line item tied to a pad and an MV transformer. String inverters show up in multiples, mounted on or near the racking itself, and the "count" a lender cares about is really a count of inverter locations across dozens or hundreds of rows.

What a central inverter looks like from above

Central inverters read as a cluster, not a scatter. Look for:

• A concrete or gravel pad set apart from the rows of tables, usually near the center or edge of a block
• A rectangular skid or containerized unit, often 20-40 feet long, sometimes paired with a medium-voltage transformer right next to it
• Cable trenches or conduit runs fanning out from the pad into the array, visible as faint straight lines in high-resolution frames
• A single unit (or two or three, for a large block) per several megawatts, not one per row

If the imagery shows one or two boxy structures per array section with trenching radiating outward, you're looking at a central-inverter design. This is the layout you'll see most often on older utility-scale sites and some larger IPP portfolios, where the EPC consolidated power conversion into fewer, bigger units to cut wiring runs.

What a string inverter layout looks like instead

String inverters are smaller and numerous, and they sit close to the modules they serve. From overhead:

• Small rectangular or box-shaped units mounted on the racking posts, roughly one every one to three tables, depending on design
• No central pad or MV transformer cluster, just distributed boxes repeating down each row
• Shorter, less visible cable runs, since each inverter is near its own string of modules
• A pattern that looks almost decorative at full-site zoom. Tiny regular dots along the rows, easy to miss unless you're counting module inventory at the same pass

This is the layout you'll see on most sites built in the last several years. It's also the harder one to verify by eye, because a string inverter is a fraction of the size of a table and gets lost in a wide shot. At typical VHR ground sample distance, you're working with pixel clusters a few pixels across, not a clean outline.

Why the count, not the type, is what the drawdown cares about

Knowing which topology a site uses tells you how to read the image. A milestone tied to "80% of inverters installed" or "DC collection complete" needs a number, pad count for central, unit count for string, measured against what the interconnection or financing schedule expects at that draw. An EPC's self-reported percentage doesn't tell a lender whether that number came from a walkdown, a guess, or last quarter's paperwork rolled forward.

That's the gap a counted inventory from VHR imagery closes: a module, table, and inverter count that's independent of what the EPC submitted, run against the same imagery each quarter so the trend line is comparable draw to draw. Signing off on a draw request or reconciling an as-built against the interconnection schedule is a different job than eyeballing a static aerial frame and hoping the inverter cluster matches the spec sheet.

If your current process is a lender's engineer squinting at a site photo trying to count inverter pads by hand, an independently counted inventory each quarter is built for exactly that handoff. Reach out if you'd like to see how a quarterly count would read against your next draw schedule.