Generation · reviewed August 2026
1 MW Solar Power Plant Generation
Last reviewed August 2026 · annual, monthly and daily units, CUF, losses, verification
A 1 MW solar power plant produces about 16.5 to 17.2 lakh units a year in Tamil Nadu, roughly 4,500 to 4,700 units on an average day, at a capacity utilisation factor of 18.9 to 19.6 per cent. Across India the band runs 16.5 to 18.2 lakh units, at 18.8 to 20.8 per cent. In Gujarat or Rajasthan the same plant produces 17.9 to 18.2 lakh units. Output falls about 0.5 per cent a year, so year twenty-five delivers around 14.4 lakh units.
The annual number is easy. The monthly shape is what breaks models. March produces 10 per cent of the year and November 6.7 per cent. A flat monthly assumption is wrong in both directions by a third.
Baseline on this page
1,000 kWp of modules on fixed tilt in Tamil Nadu, no shading, monthly cleaning in the dry season, inverter availability above 99 per cent, 19.0 per cent capacity utilisation in year one.
Read alongside
Get a free feasibility read for your 1 MW site
Send the generation figure you have been given and the district, or twelve months of meter readings, and we will tell you whether it holds up.
The arithmetic
One formula, and the only variable that matters
Generation is nameplate capacity multiplied by the hours in a year multiplied by capacity utilisation. Everything interesting is in that last term, which is a property of the site and the plant together.
Worked example, 1 MW in Tamil Nadu
That is year one. Divide by 365 for the average day, which is 4,560 units. Multiply by the tariff for revenue, and remember that the tariff applies only to units somebody actually takes.
Year one
16.6 lakh units
At 19.0% capacity utilisation on fixed tilt.
Average day
4,560 units
Annual divided by 365. No single day is average.
Best day, March
5,600–6,000 units
Clear sky, cool morning, clean modules.
Worst day, October
1,300–2,100 units
Heavy overcast during the northeast monsoon.
Generation estimator
Your site, your year, your cleaning schedule
Four inputs change the answer materially: where the plant is, what it is mounted on, how often it is cleaned, and how old it is. The monthly shape below is the part most models get wrong.
Monthly profile
Ten per cent in March, six point seven in November
This is the Tamil Nadu shape for a fixed-tilt plant. February to May carry a third of the year; October and November together carry a seventh. Cash flow, cleaning schedules and maintenance windows should all follow this curve.
| Month | Share | Units | What is happening |
|---|---|---|---|
| January | 8.8% | 1,46,432 | Clear skies, cool mornings, good conversion efficiency. |
| February | 9.2% | 1,53,088 | The best conversion conditions of the year: high irradiation, moderate temperature. |
| March | 10.0% | 1,66,400 | Peak month. Long days, clear skies, and modules still cool enough to perform. |
| April | 9.8% | 1,63,072 | High irradiation but rising cell temperature starts to bite into efficiency. |
| May | 9.4% | 1,56,416 | Very high irradiation offset by the hottest cell temperatures of the year. |
| June | 7.9% | 1,31,456 | Southwest monsoon cloud cover begins to reduce output. |
| July | 7.6% | 1,26,464 | Lowest irradiation of the southwest monsoon period. |
| August | 7.7% | 1,28,128 | Cloud cover persists, though rainfall gives some free module cleaning. |
| September | 8.1% | 1,34,784 | Recovery as monsoon cloud clears and temperatures moderate. |
| October | 7.3% | 1,21,472 | Northeast monsoon arrives. Heavily overcast days and the weakest fortnight of the year. |
| November | 6.7% | 1,11,488 | The lowest month. Short days and persistent cloud cover in coastal districts. |
| December | 7.5% | 1,24,800 | Improving through the month as the northeast monsoon withdraws. |
What to do with this
Schedule preventive maintenance and inverter servicing for October and November, when a day of downtime costs the least. Schedule cleaning for the end of the dry season, when soiling has accumulated and the units are worth most. Both decisions are worth more than they sound and neither costs anything to plan.
The two ratios
Capacity utilisation is about the site. Performance ratio is about the plant
These get used interchangeably and they measure different things. Confusing them is how a badly built plant in a sunny district passes for a good one.
Capacity utilisation factor, CUF
Actual generation divided by generation at nameplate capacity for every hour of the year. It bundles weather and plant quality together, so it cannot be compared across locations. Tamil Nadu delivers 18.9 to 19.6 per cent on fixed tilt; Rajasthan 20.3 to 20.8. A CUF quoted without a location is meaningless.
Performance ratio, PR
Actual generation divided by what the irradiation measured at the array should have produced. It isolates the plant from the weather, which is what you want when judging a contractor. Expect 78 to 82 per cent in year one on a well-built fixed-tilt plant, and require it in the EPC contract with liquidated damages.
The clause worth insisting on
A performance ratio guarantee of 78 to 80 per cent measured over the first twelve months, with liquidated damages, and a pyranometer on site to measure the irradiation the guarantee is calculated against. A peak-watt guarantee protects the contractor; a performance ratio guarantee protects you.
The loss stack
From 20 lakh units on paper to 16.6 in the meter
The irradiation falling on a 1 MW array in Tamil Nadu is worth about 20 lakh units a year in theory. Seventeen per cent of it is lost in ways that are mostly ordinary and partly controllable.
| Loss | Typical | Controllable? |
|---|---|---|
| Cell temperature above 25°C | 7–9% | Partly. Row height, ventilation and module temperature coefficient all help. |
| Soiling | 3–5% | Yes, entirely. This is a cleaning frequency decision, not a fact of nature. |
| Inverter conversion | 1.8–2.5% | Slightly. Better inverters and correct DC to AC sizing. |
| DC cabling and mismatch | 2–3% | Yes, at design stage. Cable sizing and string matching cost money once and pay every year. |
| AC cabling and transformer | 1.2–1.8% | At design stage only. |
| Downtime and availability | 0.8–1.5% | Yes. Spares holding and a contracted response time. |
| Shading and horizon | 0–2% | Yes, at layout stage. After construction it is permanent. |
| Total losses | 16–20% | Performance ratio of 80 to 84 per cent in a good year one |
Where the recoverable units are
Soiling and downtime together are five to seven per cent, and both are operating decisions rather than design ones. On a 1 MW plant that is 80,000 to 110,000 units a year, worth ₹2.5 to ₹8.5 lakh depending on the tariff. Recovering them costs a cleaning schedule and a spares shelf.
Verification
How to tell whether your plant is doing what it should
Most underperformance is invisible for months because nobody compares generation against irradiation. Four checks, done monthly, catch almost everything.
Read the meter, not the app
The bidirectional meter is the commercial truth. Reconcile it against the inverter portal every month, and investigate any gap above one per cent rather than assuming a software issue.
Calculate performance ratio monthly
Generation divided by irradiation from your own pyranometer, normalised. A month-on-month decline of two per cent with steady irradiation means soiling or a fault, and it is worth finding.
Check string currents quarterly
A dead string in a 1 MW plant is about a per cent of output and can sit unnoticed for a year. String level monitoring makes this a report; without it, it is a physical inspection.
Thermography once a year
Hotspots, bypass diode failures and connector degradation appear in an infrared scan long before they show in generation. A drone survey of a 1 MW plant is a morning of work.
The number to hold your contractor to
If year one performance ratio is below 75 per cent on a fixed-tilt plant with no shading, the plant has a defect or the commissioning was incomplete. That is a contractual conversation, and it is much easier to have in month three than in year three. Insist on the data being collected from day one.
Where we fit, and where we do not
What we will check, for free
Blues Renewables works at MW scale as a subcontract EPC and balance-of-system contractor inside other developer projects, and as a turnkey contractor at the rooftop and ground-mount sizes we have built. We are not a yield consultant and we do not issue P50 or P90 reports. What we can do is tell you whether a generation figure you have been given is plausible.
What a generation read covers
- Whether the units per MW you have been quoted imply a capacity utilisation that state actually delivers.
- Whether the layout has a shading problem that will cost you generation permanently.
- Whether your existing plant is underperforming, from twelve months of meter readings.
- What the cleaning frequency should be for your site, and what the current one is costing you.
What we will tell you to get elsewhere
- A P50 and P90 energy yield assessment for a lender. That must be independent.
- Ground-measured resource assessment campaigns.
- Warranty claims against a module manufacturer on your behalf.
- Certification of generation data for a regulatory filing.
Send twelve months of meter readings
If you already have a plant, that is all we need to tell you whether it is performing. If you are still evaluating a project, send the generation figure you have been given and the district, and we will tell you whether it holds up.
The most common overstatement we see is 22 per cent capacity utilisation quoted for a Tamil Nadu site that will deliver 19. On a 1 MW plant over 25 years that is about 65 lakh units of generation that does not exist.
Generation, answered
How many units does a 1 MW solar power plant produce?
About 16.5 to 17.2 lakh units a year in Tamil Nadu, which is roughly 4,500 to 4,700 units on an average day. The figure comes from capacity utilisation: 1,000 kWp multiplied by 8,760 hours in a year multiplied by a capacity utilisation factor of 18.9 to 19.6 per cent. Across India the band is 16.5 to 18.2 lakh units, at 18.8 to 20.8 per cent. In Gujarat or Rajasthan the same plant produces 17.9 to 18.2 lakh units because the irradiation is higher.
How many units does a 1 MW plant generate per day?
Between 4,500 and 5,000 units on an average day in Tamil Nadu, but the average is misleading. A clear March day produces 5,600 to 6,000 units; a heavily overcast monsoon day in October produces 1,300 to 2,100. Any cash flow model built on a flat daily figure will show a deficit in October that is not real and understate March by a fifth.
What is CUF and how is it different from performance ratio?
Capacity utilisation factor is actual generation divided by the generation the plant would produce running at nameplate capacity every hour of the year. It bundles the weather and the plant together, so it varies by location. Performance ratio is actual generation divided by what the irradiation that actually fell on the array should have produced, which measures the plant alone. CUF tells you what the site yields; performance ratio tells you whether the plant is working. Judge a contractor on performance ratio.
Why does generation fall over the years?
Module degradation, typically 2 to 3 per cent in the first year and 0.4 to 0.55 per cent a year after that, as specified in the module warranty. Over 25 years that is a loss of about 12 to 15 per cent, so year twenty-five produces around 14.4 lakh units against 16.6 in year one. A 25-year revenue model using the first-year figure throughout overstates lifetime generation by roughly six per cent.
How much does dirt on the modules cost?
Three to eight per cent of annual generation depending on the site and the cleaning frequency. A dusty site near a road or a cement plant cleaned quarterly loses far more than a rural site cleaned monthly. On a 1 MW plant, five per cent is 83,000 units, worth ₹2.66 lakh at a feed-in tariff and ₹6.5 lakh against an industrial tariff. Cleaning costs a fraction of that, which is why cleaning frequency belongs in the O&M contract as a number.
What generation figure should I put in a project report?
A site-specific P50 figure from an independent yield assessment, and a P90 figure for the lender to stress against. Do not use a round number from a vendor presentation. For an early estimate in Tamil Nadu, 16.6 lakh units per MW in year one on fixed tilt is defensible; anything above 17.5 lakh per MW in Tamil Nadu needs to be explained by tracking or by exceptional irradiation.
Does a higher-wattage module produce more units?
Not per kilowatt-peak, no. A 620 Wp module and a 550 Wp module of the same technology produce the same units per kWp installed. The higher-wattage module needs fewer mounting points, less structure and less labour per kWp, which lowers cost slightly. What changes generation is the technology and the temperature coefficient, not the number printed on the label.
How do I check whether my plant is underperforming?
Compare measured generation against measured irradiation to get performance ratio, month by month. A well-built fixed-tilt plant in Tamil Nadu should hold a performance ratio of 78 to 82 per cent in the first year. Below 75 per cent something is wrong, and it is usually soiling, a string that has been offline for weeks unnoticed, or an inverter derating in the afternoon heat. Without a pyranometer on site you are guessing, which is why the weather station is not an optional line item.