How many solar panels do I actually need?

Enter your electricity use and see the array size, panel count and the arithmetic behind both — plus how the answer moves when the uncertain inputs move.

Home › Solar › System Sizing

How much solar do I need?

Turn your electricity use into an array size and a panel count — with every step shown.

1 · How do you want to enter your electricity use?
Look for the yearly total on a bill, or add up twelve months of kWh.

Enter kWh for each month. The annual total is their sum.

Add what you actually run. Hours is hours on a day it is used; Days is days per week. Duty is the share of that time the appliance is really drawing power — a fridge is plugged in all day but its compressor runs about a third of it.

A tally almost always comes out below the real bill — commonly by 10–30%. Standby draw gets missed, hours get underestimated, and there is always something nobody thought of. Use this to get a working figure and to see what dominates, then check it against an actual bill. If the bill is higher, the bill is right.
Required — a bill cannot be converted to kWh without it.
This is the least reliable path. Your bill includes fixed charges, taxes and sometimes demand charges that are not energy, so working backwards from the total tends to overstate how much electricity you actually use. The result is marked low confidence.

2 · Anything you plan to add?

Optional. An EV or a heat pump can change the answer more than anything else on this page.

Check your car's figure — real-world efficiency varies by more than double.

3 · Site and equipment

A regional yearly average. Most populated places fall between 3 and 6 — or look it up for your site above.
100% aims to match your yearly use.
Typical real systems land near 80–86%.

If the two least certain inputs are wrong

Continue in Production Estimator → Will it fit my roof? →
Your data stays in this browser. Everything on this page is calculated on your own device. Your electricity use, roof dimensions, prices and equipment details are never uploaded, stored on a server, or included in analytics.
Planning estimate — not an engineering or installation plan. Results depend on the information and assumptions you enter. Output, savings and equipment behaviour vary with weather, shading, tariffs, temperature, ageing and installation. Verify equipment limits against current manufacturer documentation. Permits, structural capacity, electrical protection, conductor sizing, grounding, rapid shutdown, fire access, utility interconnection and code compliance must be reviewed by qualified professionals and the applicable authorities.
How this is calculated

Four steps, in order:

targetAnnualEnergy = annualUse × targetOffset
specificYield = peakSunHours × 365 × performanceRatio
requiredArray_kW = targetAnnualEnergy ÷ specificYield
panelCount = ceil(requiredArray_kW × 1000 ÷ panelWatts)

Generation is then recalculated from the rounded array, not the theoretical one — which is why the achieved offset usually lands slightly above your target. You cannot buy 19.1 panels.

Losses are applied exactly once. If you give an all-in performance ratio, no itemised losses are added on top; if you itemise them, they combine multiplicatively — two 10% losses retain 81%, not 80%. Peak sun hours here means plane-of-array equivalent sun, so if your figure already reflects your roof's orientation and shading, the performance ratio must not include them again.

A worked example

A household uses 12,000 kWh a year and wants to cover all of it. The site averages 5.0 peak sun hours a day, the system is expected to retain 86% after losses, and the panels are 400 W.

Specific yield is 5.0 × 365 × 0.86 = 1,569.5 kWh per kW a year. Dividing 12,000 by that gives 7.65 kW of array. At 400 W a panel that is 19.1 panels, which rounds up to 20 panels — 8.0 kW. That rounded array generates about 12,556 kWh, or 105% of the household's use. The extra 5% is the rounding, not a modelling error.

If you do not know your consumption

Use the From my appliances tab. List what you run, how many hours on a day you use it, and how many days a week — the tool does the arithmetic and shows which items dominate. That last part is usually the more valuable output: in most households two or three things account for well over half the bill, and until you know which, any effort to reduce it is guesswork.

Expect the total to land below your actual bill. A tally misses standby draw and underestimates hours, and typically comes out 10–30% low. Nothing here inflates it to compensate — if you have a bill to check against, the bill wins.

Where the sunlight figure comes from

Peak sun hours is the number this whole calculation turns on, and until you set it, it is a guess. Give the tool a location and it stops being one: it looks up a twenty-year monthly average of the sunlight actually measured over your part of the world, then works out how much of that lands on a panel at your roof's angle rather than flat on the ground.

That second step matters more than people expect. Published sunlight figures are almost always for a horizontal surface, and panels are rarely horizontal. Near the equator the difference is small — Manila gains about 1% from a sensible tilt. In London it is around 15%, and steeper still further north. Feeding a horizontal figure into a tilted array quietly undersizes the answer everywhere outside the tropics.

The tool will also tell you the tilt that collects the most over a year. It is usually well below your latitude, which surprises people who have heard the "set the tilt to your latitude" rule — that rule optimises for the equinox, not for the year, and over-tilts almost everywhere.

What the location lookup does not know

It is a climate figure for a grid cell roughly 110 km across, not a survey of your roof. Specifically, it has no idea about:

Your skyline. A tree, a ridge, a chimney or the building opposite can take a large bite out of the day, and none of it appears here. On a constrained site this is often the single biggest factor and the only way to settle it is to look — from where the panels will actually go, at the times of year that matter.

This year. Individual years commonly run 10% either side of a twenty-year average, in both directions. A system that looks marginal against the average will look worse in a dull year and better in a bright one.

Your microclimate. One degree is a coarse grid. A coastal fog belt, a valley that holds cloud, or the lee of a mountain range can all differ sharply from the cell average — and two adjacent cells can genuinely differ by 10%, as they do across Manila between the bay and the mountains.

None of that makes the figure useless; it makes it a well-founded starting point rather than a measurement. If you have a real quote, an installer's shading study, or a neighbour's actual production, those beat this and you should use them instead.

Your location stays on your device

The place list and the sunlight data are ordinary files downloaded with the page, and the lookup is arithmetic your browser does locally. No address, coordinate or result is sent to us or to anyone else — there is no geocoding request and no weather API call, which is the usual way this feature leaks where you live. The privacy policy spells out the one exception, which belongs to your browser rather than to this site.

What moves this answer most

In order: your electricity use, then peak sun hours, then losses, then panel wattage. Panel wattage matters least — it changes how many panels you need, not how much array. Two of the four are genuinely uncertain, which is why the sensitivity table above shows the answer across a range rather than pretending to one figure.

Limitations

This sizes an array against a yearly energy total. It does not know whether your roof has space (try the panel fit calculator), whether your use lines up with when the sun shines, or what your utility pays for exported energy. Where net metering has ended, a system sized to 100% of annual use may offset considerably less than 100% of your bill — the payback calculator handles that distinction.