Generation & sizing · India
About four units a day on annual average for a well-sited, unshaded plant — roughly 120 units per kW per month. The number you are quoted matters less than whether it is an annual average or a good-day figure, because only one of those pays your bill.
Indicative, not a promise
| Basis | Typical per kW | What moves it |
|---|---|---|
| Per day, annual average | ~3 to 5 units | Region, shade, orientation, cleaning |
| Per month, annual average | ~110 to 130 units | Same, plus seasonal spread |
| Best months | Above average | Clear pre-monsoon skies |
| Monsoon weeks | Well below average | Extended cloud cover |
| North Indian winter | Below average | Fog and haze, not cloud |
Last reviewed
The capacity printed on a module is measured under standard test conditions — a defined light intensity at a defined cell temperature. Your roof in May is not that laboratory. Four things pull real output below the label, and all of them are normal rather than faults.
Temperature. Cells lose efficiency as they get hotter, and a panel in direct Indian summer sun runs far above the rating temperature. This is why peak output on a blazing day can be lower than on a bright, cooler day.
Soiling. Dust, pollen and bird droppings on the glass block light continuously. Unlike the other losses this one is fully recoverable, which is precisely why cleaning discipline earns its place in maintenance.
Conversion and wiring. Turning DC into usable AC costs a few percent, as does the cable run itself.
Sun angle. Generation follows a curve, not a switch. Early morning and late afternoon contribute far less than the middle of the day, which is also why an appliance running at night is not being run by your panels — the point laboured, for good reason, in solar panels for an AC.
Ask of any generation number: is this an annual average or a good day? The gap between the two is wide enough to change a payback calculation entirely.
A 3 kW plant might produce well over fifteen units on a clear March day. The same plant will produce a fraction of that during a monsoon week. If a savings projection is built from the March figure and presented as typical, the payback period on paper will be considerably shorter than the one you actually experience. A projection built on annual units is the one to trust, and it is reasonable to ask which was used.
Regional differences are real. The dry north-west and the interior Deccan generally do best, combining strong sunshine with comparatively few overcast days; Jodhpur and Gujarat sit at the upper end. Much of Telangana, Karnataka and Maharashtra sits solidly in the middle. The wettest regions sit lower: Kerala contends with two monsoon periods, and West Bengal and the north-east with long humid cloud cover.
Here is the part worth internalising, though. The spread between a good Indian location and a mediocre one is usually narrower than the difference shade makes on one particular terrace. A well-placed array in a humid state can comfortably out-generate a partly shaded array in a sun-rich one. So a regional average is a starting point for a conversation, never a substitute for looking at your own roof — which is also why how much shadow-free area you have is the more useful question.
This is the arithmetic worth doing before anyone visits.
It is tempting to size for the May peak so the bill goes to zero in the hardest month. That normally buys capacity which sits underused for the rest of the year, and every kilowatt beyond the ceiling carries its full cost with no extra assistance.
In most states the mechanism that solves seasonal mismatch already exists: surplus units exported in strong months are credited against consumption in weak ones, which is the whole point of net metering. Where that credit genuinely does not carry forward — and the arrangement varies, as Indore illustrates — sizing changes, which is exactly why the settlement rules deserve a question before you sign.
An installer who answers these precisely is quoting from a model. One who cannot is quoting from a brochure.
Run the free home check with your average monthly units and we will size it honestly, including what your roof can actually hold. If you charge an electric car at home, add that load before sizing — solar for EV charging. Related: cost & payback · roof area needed · 3 kW price · how rooftop solar works.
As a planning figure, around four units a day averaged across a full year for a well-sited, unshaded plant. The honest range across India is roughly three to five units per kW per day depending on where you are, how the array is oriented and how clean it is kept. Treat four as an annual average rather than a daily promise: a clear March day may exceed it comfortably while a monsoon week may fall well below it. Any savings estimate should be built on the annual figure, because that is what actually settles against your bill.
Roughly 110 to 130 units per kW per month as an annual average, which is where the common rule of thumb of about 120 units per kW per month comes from. Individual months vary far more than that average suggests. The pre-monsoon months are usually the strongest, the monsoon weeks are the weakest in most of the country, and in the northern plains late December and January can be poor for a different reason, which is fog and haze rather than cloud.
Because the kilowatt on the label is measured under standard test conditions in a laboratory, not on your roof in May. Real output is reduced by cell temperature above the rating point, by dust on the glass, by wiring and inverter conversion losses, and by the sun being low in the sky outside the middle of the day. A well-built plant in India typically delivers somewhere in the region of three quarters of nameplate at the best moment of a good day, and the daily total is what matters rather than that instantaneous figure.
Broadly, the dry north-west and the interior Deccan do best, because they combine strong sunshine with relatively few overcast days. Rajasthan and Gujarat sit at the higher end. Much of Andhra Pradesh, Telangana, Karnataka, Maharashtra and Madhya Pradesh sit comfortably in the middle. The wettest and most humid regions, including Kerala, coastal Karnataka, West Bengal and the north-east, sit lower because of extended cloud cover during two monsoon periods in some cases. The differences are real but usually smaller than the difference shade makes on an individual roof.
Take twelve months of billed units from your electricity bills and find your average monthly consumption. Divide that by about 120 to get an indicative capacity in kilowatts for a plant intended to offset most of it. A home using 360 units a month therefore points at roughly 3 kW. Then check two constraints before treating that as the answer: whether you have the shadow-free roof area for it, and how the central subsidy ceiling affects the economics of going beyond 3 kW.
Usually not. Sizing to a summer peak buys capacity that sits underused for much of the year, and the extra kilowatts beyond the subsidy ceiling are paid for in full. In most states surplus units are credited against later consumption under net metering, which already smooths seasonal mismatch to a degree. The better approach is to size against annual consumption and let the credit mechanism handle the monthly variation, unless your billing arrangement genuinely does not carry credits forward.
More than most buyers expect, and more than the regional differences in sunshine. Soiling on the glass reduces output continuously and is entirely recoverable with routine cleaning, which is why a cleaning habit matters as much as equipment choice in dusty regions. Shade is worse because it is not recoverable: a single obstruction crossing part of the array during productive hours can drag down output disproportionately depending on how the modules are strung. This is why a shade assessment on your specific roof beats any regional average.
