Predicting Solar Yield

A yield estimate is not one calculation but a chain of multipliers, and understanding the chain is what separates a defensible estimate from a guess. It starts with the solar resource: how much solar energy actually lands on the surface in question over a year, usually expressed in kilowatt-hours per square metre per year (kWh/m2/yr). This is not a single global number - it depends on where the building is (a site in Rajasthan receives far more than one in the cloudier north-east), and, crucially, on the *plane* of the surface. The resource on a horizontal roof, an optimally tilted array and a vertical facade are all different, and databases and tools exist precisely to give the resource *on a given plane at a given place*.

From the resource you multiply by the collecting area - the square metres of active PV - and by the efficiency of the modules, which is the fraction of the light energy they convert to electricity (Module 1). Standard crystalline silicon sits around a fifth or so; many BIPV products, trading efficiency for transparency, colour or integration, convert less. Area times efficiency times resource gives a first, optimistic figure: roughly the energy the cells would make if everything were ideal.

Nothing is ideal, so the chain continues. You apply an orientation and tilt factor - how far the surface deviates from the sun-optimal angle for that latitude, which for a vertical BIPV facade can be a substantial penalty. You apply a shading factor - the fraction of the resource lost to trees, neighbouring buildings, parapets or the building's own geometry, which for facades in dense cities can be severe and time-varying. And finally you apply everything else that erodes real output - heat, dirt, wiring resistance, inverter conversion, module mismatch - bundled into the performance ratio. The output of that whole chain is an annual yield in kWh: a genuine estimate, provided every link was set honestly. Set any link with wishful thinking and the error compounds all the way down.

Two numbers let professionals compare systems fairly, and a literate designer should know both. The first is the performance ratio (PR) - the ratio of the energy a system actually delivers to the energy it would deliver if it ran at its nameplate rating under the resource it received, with no losses at all. PR bundles every real-world loss into one figure between 0 and 1: temperature (hot cells make less power, the subject of Lesson 6.4), soiling (dust and dirt on the glass, a serious issue in much of India), wiring and DC losses, inverter conversion losses, module mismatch, downtime and more. A well-designed, well-ventilated, clean rooftop system might reach a PR around 0.8; a hot, dusty, or poorly ventilated facade-integrated array will sit lower. PR is powerful precisely because it is honest - it names, in one number, the gap between the brochure and the meter.

The second is specific yield - the annual energy produced per unit of installed capacity, in kWh per kWp (kilowatt-hours per kilowatt-peak) per year. Because it normalises out system size, it lets you compare a small home system with a large commercial one, or a good site with a poor one, on equal terms. A sun-rich, well-oriented Indian site might deliver a high specific yield; a shaded north facade, dramatically less. Related is the capacity factor, the average output as a fraction of the peak rating across the whole year - low for solar (the sun is not always up) but a useful sanity check.

The reason these metrics matter to a *designer*, not just an engineer, is that your architectural choices move them directly. Choosing a vertical facade over a tilted roof lowers the resource on the plane. Choosing a coloured or semi-transparent BIPV glass lowers the efficiency. Placing modules where the building shades itself lowers the shading factor. Detailing a facade with no air gap raises the operating temperature and lowers the PR. Every one of those is a design decision that shows up in the yield estimate - which is exactly why the designer must understand the metrics, even while the binding numbers come from the model.

In real practice, nobody produces a serious yield figure by hand. Engineers use PV simulation software and solar-resource databases that model the chain in far more detail than any manual estimate. Conceptually, these tools do three things. First, they pull location-specific resource data - typically a 'typical meteorological year' built from long-run satellite and ground measurements - for the exact site, giving hour-by-hour irradiance and temperature rather than a single annual average. Second, they model the geometry: the plane of each array, its orientation and tilt, and often a full shading analysis using a 3D model of the building and its surroundings, so the software can compute how a parapet or a neighbouring tower shades the array at 9am in December versus 3pm in June. Third, they apply a detailed loss model - temperature behaviour from the module datasheet, soiling assumptions, wiring and inverter efficiencies - to turn incident sunlight into an hourly, then annual, energy figure.

The output is far more trustworthy than a hand estimate, and it is what a bankable project relies on. But - and this is the honest heart of the lesson - a simulation is still a model, and a model is only as good as its inputs and assumptions. The resource data is a statistical typical year, not next year's actual weather. The soiling and degradation assumptions are estimates. The shading model is only as accurate as the 3D context fed into it. Real yield in any given year will vary with the weather, with how clean the array is kept, with faults and downtime, and with the slow degradation of the modules over decades.

So the professional stance is twofold. A designer should absolutely *understand* the chain and the metrics well enough to reason about how a design choice will move the yield, and to sanity-check a number for plausibility. But the binding yield figure - the one a contract, a loan or a payback promise rests on - must come from a proper simulation run by a qualified engineer against verified, site-specific data and the manufacturers' datasheets, and even then it is presented as a well-founded estimate with uncertainty, never a guarantee. Any figure in this lesson is illustrative of the principle, not a specification.

Everything above applies to any PV system, but BIPV makes yield estimation genuinely harder and usually lower than the equivalent bolted-on (BAPV) array, and an honest designer builds that into expectations from the start. A standard rooftop BAPV array is close to the textbook ideal: mounted at a chosen tilt and orientation, on open racking that lets air flow behind the modules to keep them cool, and generally sited to minimise shading. Its yield is the easiest case to estimate and often the highest per watt.

BIPV departs from that ideal in several compounding ways, each of which pushes the estimate down. Orientation and tilt are constrained by the building, not the sun: a BIPV facade is vertical and faces whatever direction the wall faces, which for many walls is far from optimal, cutting the resource on the plane substantially. Shading is often worse and more complex: facades in dense urban settings are shaded by neighbouring buildings, balconies, fins and the building's own massing, and because PV modules are sensitive to partial shading, a small shadow can cost disproportionately. Heat is often worse: BIPV integrated tight against an insulated wall with no ventilation gap runs hotter than a well-aired rooftop panel, lowering the performance ratio (Lesson 6.4). And the modules themselves may be less efficient by design - coloured, patterned or semi-transparent BIPV glass converts less light than a plain module, a deliberate trade of yield for appearance and function.

None of this is an argument against BIPV; it is an argument for *honest estimation*. A facade that generates less per square metre can still be worth it - it may use surfaces a bolt-on array never could, replace expensive cladding, and turn a huge vertical area on a tall building into a generator where the small roof cannot meet the load. But the designer who promises rooftop-BAPV yields from a north-ish, self-shaded, unventilated coloured facade is setting up disappointment. Estimate BIPV on its own honest terms, expect lower specific yield than the ideal roof, quantify it with a proper model, and let the architecture justify itself on the fuller picture rather than on an inflated generation number.