Sizing & Yield

Sizing a system means choosing how much generating capacity to install, and it is best understood as reconciling three limits that pull in different directions.

The first is the load - the amount of electricity the building actually uses over a year, and, just as important, *when* it uses it. In principle you might size the array to cover all or a chosen fraction of that annual demand. But annual totals hide the timing that 2.1 showed matters: a system sized to match annual energy may still export heavily at midday and import heavily in the evening. So load-based sizing starts from a real picture of consumption, not a guess.

The second is the area - how much suitable surface exists to carry modules. 'Suitable' is doing a lot of work: it means unshaded for the useful hours, decently oriented and tilted, structurally able to carry the modules, and actually available (not cluttered by plant, not needed for other uses). A large roof of poor orientation may offer less usable area than a small well-oriented one. For BIPV the area question expands to the whole envelope - facades, glazing, canopies - which can unlock generating surface a roof-only view would miss, especially on tall buildings where facade area dwarfs the roof, but at lower per-surface yield.

The third is the budget - what the client will actually pay. Since BIPV is costlier per watt than BAPV, budget bites harder, and the honest designer sizes to what pays for itself or meets the client's real goals, not to the maximum the surface could hold.

These three rarely agree. A home may have the load and the budget but too little good roof; a tower may have facade area and load but a budget that covers only part of it; a factory may have vast roof and budget but a daytime load that rewards a big array. The art of sizing is finding the size that best reconciles the three for *this* building - and it is genuinely a design act, because it decides how much of the envelope generates and how the system reads. The binding sizing calculation, matched to metered loads and verified yield modelling, is the engineer's; the designer's job is to understand the negotiation and shape the brief and the envelope so a good answer is possible.

Here is the distinction every solar-literate designer must hold, because confusing the two is the single most common error in the field. Rated power (kWp) and delivered energy (kWh) are different kinds of number, and one does not equal the other.

Rated power, in kilowatts-peak (kWp), is the output a module or array produces under Standard Test Conditions (STC): a defined, generous laboratory setup - full bright irradiance, a cool cell temperature, a clean spectrum, perfect facing. It is a label, and a useful one: it lets you compare modules fairly and describe a system's size. But STC almost never occurs on a real building - the cells are hotter, the sun weaker or at an angle, the glass dusty, the sky partly cloudy - so the array almost never actually produces its rated power. kWp is a nameplate, not a forecast.

Delivered energy, in kilowatt-hours (kWh), is what the system actually produces over time - a day, a year - and it is what genuinely matters, because energy (power multiplied by time) is what runs the building and what the meter counts. A 5 kWp system does not deliver '5 units'; it delivers however many kWh accumulate as the real sun rises and sets, weakened and interrupted by every real-world factor, across the whole year.

A helpful bridge between the two is specific yield - the kWh delivered per year for each kWp installed (kWh per kWp), which bakes the site's climate and losses into a single figure that varies enormously by location and design. Sunny, well-designed installations achieve high specific yield; shaded, hot or poorly oriented ones far less. It is a far more honest way to compare real prospects than nameplate alone - but it, too, is site-specific and belongs to proper modelling.

The practical discipline is simple and non-negotiable: whenever you hear a system quoted in kW, ask what energy (kWh) it is expected to deliver on this site, and treat any figure as a modelled estimate, not a guarantee. A designer who conflates kWp with kWh will oversize expectations, mis-sell the building, and be caught out when the meter tells the truth. The rated number is where you start; the delivered number is what you owe the client honestly - and only an engineer's site modelling can give it credibly.

If nameplate is never delivered, what governs how much you actually get? The losses - and they are captured in one honest figure, the performance ratio (PR). Roughly, PR is the energy the system actually delivers divided by the energy it would deliver if every module ran at its rated efficiency in the site's sunlight with no losses at all. It is always less than one, and it is the single most useful sanity-check on a yield claim: a plausible, well-designed system lands in a sensible range below one; a claim implying a PR near one, or above, is a red flag.

What eats the gap? Walk down the losses. Orientation and tilt: a surface not squarely facing the useful sun collects less - a real issue for facades and any non-ideal roof. Shading: even partial shade, as 2.2 showed, can cost far more than its area suggests, especially on a shared string. Temperature: PV cells lose efficiency as they heat, and hot climates (much of India) and poorly ventilated integrations (a sealed BIPV facade) push cell temperatures up and yield down - a genuine, sometimes large, loss that this course flags repeatedly. Soiling: dust, pollen, bird droppings and (in dusty Indian conditions) grime cut output until cleaned. System losses: wiring resistance, inverter conversion, mismatch between modules, and the DC-to-AC clipping from 2.2. Downtime: any hours the system is off for faults or maintenance. Degradation: modules slowly lose output over decades, so year-twenty yield is below year-one.

Each loss is modest alone; stacked, they turn nameplate into a considerably smaller delivered figure. That is not a defect to hide but the normal, honest physics of real installations - and it is exactly why a credible yield number cannot be guessed from the kWp and a sunshine figure. It requires proper simulation that models this specific site's irradiance, the array's real geometry and shading, the local temperature, expected soiling, and the chosen equipment - the work of an engineer with the right tools and data. For BIPV the stakes are higher, because integration deliberately accepts some of these losses (non-ideal orientation for architecture, higher temperature from tight integration) in exchange for its benefits, so an honest, modelled yield is essential to judge whether the trade is worth it. Any figure in this lesson is illustrative, to teach the shape of the losses - never a specification.

Put sizing and yield together and a clear discipline emerges for the designer - one that earns trust and avoids the classic embarrassments.

First, size to a real picture, not a fantasy. Start from the building's actual load and timing, the genuinely suitable area, and the honest budget, and find the reconciliation - not the biggest array the surface could theoretically hold. An oversized system that exports cheaply, or a system sized to nameplate expectations that the site cannot deliver, both disappoint. For BIPV, this means being clear-eyed that generating surface bought for architecture (a shaded or non-ideal facade) may add far less energy per rupee than a smaller, well-placed array - and sometimes the honest answer is a modest BIPV feature plus efficient BAPV elsewhere, or simply less PV.

Second, speak in delivered energy, and always as an estimate. Never let a client believe the kWp is the energy. Frame expectations in modelled annual kWh (or specific yield), state the assumptions, and label every figure as a site-specific estimate pending engineering modelling - because it is. This honesty is not weakness; it is the difference between a designer clients trust and one who is caught out by the meter.

Third, respect the losses in the design. Many of them are partly within the designer's influence: orientation and tilt (shape the envelope to catch useful sun where you can), shading (place and detail to avoid self-shading and plan for what neighbours may build), temperature (ventilate integrations - a recurring BIPV theme, Module 6.4), and soiling access (make surfaces cleanable). You cannot eliminate the gap, but thoughtful design keeps the performance ratio healthy rather than needlessly poor.

Fourth, and firmly, know where your role ends. The binding sizing - matched to metered loads, structural capacity and budget - and the binding yield - from proper simulation of this site and system - are the engineer's, with the manufacturer's verified data and the governing standards (module ratings to IEC, and the codes). Your job is to understand the negotiation and the nameplate-versus-delivered distinction well enough to brief it, shape the envelope for it, and hold the client to honest expectations. Nameplate is never delivered energy; the honest designer says so, designs to narrow the gap, and defers the binding numbers to those equipped to give them.