Daily PV yield estimate

What a single day of production means when it is an average day

This multiplication is the sizing page’s division run backwards, and the pack’s anchor vectors close the loop exactly: feed in the array that page demanded, at the same sun and the same performance ratio, and the day that returns is the very daily consumption the sizing began from. That round trip is the cluster’s internal consistency check — if the two pages are run with different sun or loss figures, the loop breaks, and the break is a sign the assumptions have quietly diverged.

The average day is a statistical fiction, and at day scale the fiction shows more than it does over a year. Annual figures genuinely smooth out; a day does not — a bright June day can run far above this estimate and a grey December one far below it, while the average day itself may never actually occur. The number is a centre of gravity, not a forecast for tomorrow.

Day scale is the right resolution for a specific set of questions. Battery autonomy is reckoned in days of stored energy against days of production; self-consumption is a matter of whether the household’s daytime habits line up with the hours the energy arrives; and a monitoring app reports its verdict nightly in exactly these units. For bills, offsets and payback — questions that settle over a year — the annual estimate in this cluster is the honest tool instead.

This calculator runs forward only. It is the quickest arithmetic in the cluster — three inputs, one product — and that is its role: the daily sanity check a reader can hold against the monitoring screen, against a neighbour’s system, or against an installer’s conversational claim, before reaching for the heavier pages.

When this calculation is used

• Turning a proposed or installed array size into the average day of energy it should deliver at the site.
• Checking a monitoring app’s daily readings against expectation, once enough days have accumulated to average over.
• Feeding battery sizing: the average production day is one side of the autonomy arithmetic the storage pages carry.
• Sanity-checking conversational claims — an installer’s promised daily output must be consistent with the array, the site’s sun and an honest performance ratio.

Worked example

Close the cluster’s loop: the five kilowatt-peak array the sizing page’s scenario demanded, at the same site with five and a half peak sun hours, keeping the assumed performance ratio of eight tenths.

The day that returns is, to the digit, the daily consumption the sizing page’s household started from — the round trip through the two calculators hands the load back. Read it as the average day the system balances the household, remembering that the actual days scatter widely around it: the array will beat this figure through the bright months and miss it through the dark ones, and both are consistent with the estimate being right.

The figure is produced by the verified engine at load, and this scenario is one of the declared test vectors in the signed pack, alongside runs at other sizes and sites and a declared refusal at a performance ratio of exactly one. It is a planning estimate, not a promise of output — the day a real system delivers depends on weather, equipment and workmanship, and an installer’s site-specific assessment is where this number should lead.

What each input represents

Assumptions and limits

• One annually averaged sun figure produces one average day, so the estimate is silent about the seasonal spread that day scale makes most visible.
• The performance ratio is held constant, although its largest component — heat — varies through the very day being estimated.
• The day is energy only: nothing here says when in the day the energy arrives, which is what self-consumption and export actually turn on.
• The estimate is degradation-free; an ageing array’s average day shrinks slowly, and the pack’s degradation siblings carry that arithmetic.

What the guards protect against

• The performance ratio must lie strictly between zero and one, and the pack declares the refusal at exactly one as a test vector — a lossless system does not exist, and the same ceiling catches a percentage typed where a decimal belongs.
• Peak sun hours are bounded to the band real sites occupy, from a single equivalent hour to nine; a daylight-hours figure or an annual total lands outside it and is refused as a unit mistake rather than inflating the day.
• The array size must be positive and below a utility-scale cap, so an entry in watts rather than kilowatts is refused instead of returning a day wrong by a factor of a thousand.

Provenance

Standard simplified PV yield form at day scale, daily energy = array size × peak sun hours × performance ratio

Product of the array nameplate, the site’s daily peak sun hours and the performance ratio, returning the estimated average daily energy in kilowatt-hours; forward workflow only.

Screening and reference material, to be checked against site-specific resource data and a qualified installer or engineer; not a promise of performance in any form. The signed pack carries its own citation, and the page reports the verification state of the release it mounted.