What a Real Solar Generation Estimate Actually Requires — And Why "kWh per kWp" Alone Isn't It

Almost every solar quotation includes an expected annual generation figure — it is the number that lets a customer estimate their payback period and compare one installer's offer against another's. It is also, very often, computed as a single flat multiplier — "1,400 kWh per kWp per year," applied identically whether the roof is in Jaisalmer or Kochi. That shortcut isn't just imprecise; it throws away most of the information that actually determines how much a specific rooftop will generate.

Irradiance varies by location — and it isn't a small variation

The amount of solar energy a given rooftop receives over a year depends on its latitude, local climate, and typical cloud cover — the same nominal system size can generate meaningfully more in one part of the country than another. A serious estimate starts from measured irradiance data for the site's actual coordinates rather than a single number meant to represent the whole country. PVGIS — the European Commission's solar resource service — is one of the standard sources for this kind of site-specific irradiance data, and it is what a bankable generation estimate is typically built on rather than a rule-of-thumb average.

Orientation and tilt change the answer, not just the aesthetics

A flush array on a sloped roof takes its tilt and its compass direction (azimuth) directly from the roof itself — the panels lie in whatever plane the roof already occupies. A tilted frame on the same roof can face a different direction and sit at a different angle entirely, and the two will not generate the same amount of energy even with an identical panel count. A generation estimate that ignores which of these two situations actually applies is computing the yield for a plane the panels aren't even in.

Shading is not optional to model

A water tank, a parapet wall, a neighbouring building, or even the array's own rows shading each other at a low sun angle all reduce real-world output below what an unobstructed calculation would suggest. Row spacing on a tilted frame exists specifically to keep one row from shading the row behind it — and on a flush array lying coplanar with the roof, that same shading risk simply does not exist, so identical spacing would only waste usable roof area. Treating every mounting configuration the same way, shading-wise, either overstates generation on an obstructed roof or wastes panel capacity on an unobstructed one.

System losses are a real, quantifiable deduction — not a footnote

Even a perfectly sited, perfectly oriented array does not convert 100% of incident irradiance into delivered electricity. Inverter efficiency, wiring losses, module soiling, and temperature-driven derating all reduce the theoretical maximum to a real-world figure, and a credible estimate accounts for these rather than presenting the theoretical ceiling as the expected outcome.

Why this matters for the customer conversation, not just the engineering

A generation number is a promise, even when nobody says the word "guarantee." A customer sizing their payback expectations against an inflated flat-rate estimate is heading toward a disappointing first year of actual bills — and a disappointed customer is the single most expensive outcome in a referral-driven business. An estimate computed from real site coordinates, actual orientation, and modeled shading is not just more technically defensible; it is the estimate that still looks right eighteen months after the system is commissioned.

Solset AI's Design Studio computes generation from PVGIS irradiance data for the site's exact coordinates, taking the array's real orientation and tilt from the roof or the mounting frame as appropriate, rather than a single flat multiplier applied to every design regardless of where it sits.