On this page13 sections
- 01What a production forecast is made of, and which term carries the honesty
- 02Concepts to hold first
- 03From nameplate to the meter
- 04Why yield, not size, decides the economics
- 05How the method works
- 06Try it, verified
- 07What each input represents
- 08Worked example
- 09Reading the result
- 10Common mistakes
- 11Questions readers arrive with
- 12When this calculation is used
- 13Assumptions and guards
What a production forecast is made of, and which term carries the honesty
This is the sizing calculation turned around. Sizing divides a load by sun and losses to demand an array; the estimate multiplies an array by the same sun and the same losses to promise an energy. The two share their vocabulary deliberately — same peak sun hours, same performance ratio — so an array sized on this cluster’s first page and estimated on this one tells a consistent story, and any change of assumption has to be made in both places or the inconsistency shows.
The product of nameplate and peak sun hours alone would be the IDEAL year: the energy the array would deliver if the field behaved like the laboratory. The performance ratio is the term that carries the honesty. Multiplying by it concedes cell temperature, inverter conversion, wiring, soiling and mismatch in one factor — so the estimate is only as credible as that single number, and quoting a yield without stating the ratio behind it is how solar forecasts mislead politely.
An annual figure smooths what a year actually looks like. Production concentrates in the bright months and thins in the dark ones, and the estimate deliberately averages that swing away: it is the right quantity for bills, offsets and payback arithmetic, which settle over a year, and the wrong quantity for asking whether the lights stay on in December. The pack carries monthly and seasonal siblings for that question.
The relation is linear in every input, and linearity cuts both ways. It makes the estimate transparent — double the array, double the year — and it means every input error passes straight through to the answer undamped. An optimistic sun figure or a flattering performance ratio does not average out; it propagates, and then it propagates again through the payback calculation this estimate feeds.
Concepts to hold first
The energy an installation delivers across a whole year, at the meter. It is the figure a bill responds to, a payback calculation consumes, and a warranty dispute is argued over — the output side of the whole solar case.
Yield per unit of installed capacity — the great equaliser between systems of different sizes. Two installations can only be compared fairly through their specific yield, because raw yield mostly measures who bought more panels.
Sunshine is a human observation; irradiance is an energy measurement. Solar arithmetic runs on irradiance, and every hour contributes exactly the energy it carried — a bright haze counts for something, a clear winter afternoon for less than its brilliance suggests.
The deliberate discounting of a rated figure to account for real conditions. In a yield estimate it is carried by the performance ratio; the discipline is stating the discount once, explicitly, instead of sprinkling optimism through several numbers.
From nameplate to the meter
A panel’s rating describes a laboratory: fixed light, fixed temperature, a moment of perfect alignment. A roof offers none of these for long. The journey from nameplate to meter passes through the site’s actual solar income, then through every loss a real installation imposes — cell temperature above the laboratory’s, dust and pollen, wiring resistance, inverter conversion, module mismatch, the occasional shadow.
The estimate compresses that journey into three numbers: capacity, peak sun hours, and the performance ratio that carries the whole loss budget. The compression is the method’s strength and its limit at once — it makes the calculation checkable in your head, and it means seasonal shape, degradation over years and unusual loss mechanisms are deliberately out of scope.
When a quoted yield disagrees with this estimate, the disagreement is diagnostic. If the quote is higher, one of its three numbers is more optimistic than yours — and it is nearly always the performance ratio, because sun data is public and capacity is on the invoice. Asking which number moved turns a sales conversation into an engineering one.
Energy narrows from laboratory rating to delivered yield as real losses apply
The loss staircase between a rating and a meter. The performance ratio states the whole staircase as one honest number.
Why yield, not size, decides the economics
Everything financial about a solar installation — payback, savings, export income — is priced in delivered kilowatt-hours, not installed kilowatts. Two households buying the same array can experience payback periods years apart purely because one roof faces the sun and the other tolerates it. Yield is where geography, orientation and installation quality all cash out.
This is also why the yield estimate deserves conservative inputs. An optimistic array size costs roof space; an optimistic yield estimate costs money, because every financial decision downstream — including the payback calculation at the end of this journey — inherits it silently.
How the method works
The installed capacity is multiplied by the site’s peak sun hours, giving the energy the array would deliver each day if it converted perfectly.
That figure is multiplied by the performance ratio, applying the whole loss budget in one declared step.
The daily result is extended across the year. The estimate is annual by construction — seasonal shape is deliberately not modelled, and a site with strong seasons will overshoot this figure in one half of the year and undershoot it in the other.
The certified engine performs this calculation. This page explains what it does; it does not reproduce it — a second implementation is a second answer waiting to disagree.
Try the worked scenario
The engine below arrives pre-filled with the same reference system the array-sizing lesson used — the pack’s own worked example. Try your own capacity first. Then hold everything else still and move only the performance ratio a few points in each direction: the yearly swing you see is the money that quiet number controls.
Read the result as the year’s delivered energy under average conditions. It is computed live by the verified engine — this page stores nothing and would rather refuse than estimate for itself.
What each input represents
The nameplate of the array in kilowatts-peak — its rated output under Standard Test Conditions. Use the AS-INSTALLED figure: module count times module rating, after the rounding to whole panels that installation forces, which is generally a little more than the size the load calculation asked for.
The site’s daily solar resource as equivalent hours of full STC-strength sun, annually averaged. It must come from location-specific resource data, and it must be the SAME figure the sizing used — an estimate run on sunnier assumptions than the sizing is not a forecast, it is a wish.
The fraction of the ideal yield the whole installation retains, as a decimal strictly between zero and one. It bundles thermal, inverter, wiring, soiling and mismatch losses into one factor. The default is an illustrative grid-tie value; the measured performance ratio page in this cluster exists to replace assumption with evidence after the first metered year.
Worked example
The scenario
Run the cluster’s reference system forwards: the 5 kWp array the sizing page’s scenario called for, at the same site with 5.5 peak sun hours, keeping the assumed performance ratio of 0.8.
What returns is a year of production in kilowatt-hours. Read it as the central estimate for an average year — the number to set against annual consumption to see how much of the bill the array displaces, and the number the payback page will convert into money. It is not a floor, not a promise for any particular month, and not resilient to a wrong performance ratio.
Nudge only the performance ratio downwards, as a dusty site or a hot summer would. The year shrinks in exact proportion — no other input softens it — which is why the difference between an assumed ratio and a measured one is the first place to look when production disappoints.
Reading the result
The annual figure is the one your electricity bills will feel, spread unevenly across the seasons. Comparing it to your yearly consumption tells you the fraction of your usage the array covers on paper — before storage, export rules and timing decide how much of that paper coverage you actually keep.
Divided by the installed capacity, the result becomes specific yield — the number to compare against neighbours, quotes and published benchmarks for your region. A specific yield far from the local norm means a site condition or an input error, and both are worth finding.
This estimate is the denominator of the next lesson: measure your real yield for a year, divide by what this page predicts, and you have measured your installation’s true performance ratio.
Common mistakes
Quoting yield for the array you wish you had — using nameplate capacity but forgetting the panels that were dropped from the final layout.
Letting optimism in twice: once in generous peak sun hours and again in a generous performance ratio. The loss budget exists to be stated once.
Treating the annual figure as a monthly promise divided by twelve. Winter will not honour it.
Comparing raw yields between differently sized systems. Only specific yield compares fairly.
Questions readers arrive with
Why does my quote show a higher yield than this estimate?
Because one of the three inputs differs — and sun data is public while capacity is on the invoice, so it is nearly always the performance ratio. Ask which value the quote assumed and what it counts. The question is fair, and the answer is informative in either direction.
Does the estimate account for panel degradation?
No. It describes a year, not a lifetime. Panels lose a little capacity each year; a lifetime energy model applies that decline year by year, and this estimate is the honest first rung of such a model rather than a substitute for it.
Can I use it for a shaded roof?
Cautiously. Light, brief shading is one of the losses a realistic performance ratio absorbs. Structural shading — a chimney across the array every afternoon — breaks the assumption that losses are roughly proportional, and deserves site-specific modelling.
Is the calculator’s answer a guarantee?
It is an estimate under declared assumptions, computed by a certified engine from your inputs. The engine guarantees the arithmetic and refuses implausible inputs; the assumptions remain yours, which is exactly why this lesson exists.
When this calculation is used
Forecasting the production of a proposed array before purchase, as the number a quotation’s promises are tested against.
Feeding the economics: the annual yield is the quantity the payback and savings calculations in this pack consume.
Solving backwards for the array a production target demands — the declared reverse workflow, for targets set in energy rather than load.
Comparing candidate sites at a fixed array size, where only the peak-sun-hours figure differs between runs.
Assumptions and guards
One annual-average peak-sun-hours figure stands in for the whole year, so the estimate says nothing about seasonal distribution.
The performance ratio is constant across the year, although the temperature losses inside it are largest exactly when the sun is strongest.
The figure is a first-year, degradation-free estimate. Module output declines slowly with age; the pack’s degradation calculators handle later years.
Losses beyond what the performance ratio bundles — extended outages, inverter clipping under a deliberately undersized inverter — are not modelled here and have their own siblings in the pack.
The performance ratio must lie strictly between zero and one — the pack declares the refusal at zero as a test vector. A zero ratio describes a system that delivers nothing, and a ratio of one or more describes one with no losses; neither is an estimate worth returning.
Peak sun hours are bounded to the band real sites occupy, which refuses annual totals, irradiance in other units and daylight-hours figures instead of folding them silently into a wildly wrong year.
The array size must be positive and below a utility-scale cap, so an entry in watts rather than kilowatts — a thousandfold slip — lands outside the band and is refused.