Shading analysis and mitigation in PV arrays

Shading is disproportionately damaging to PV output relative to the fraction of the module surface it covers, because of how solar cells are wired in series within a module and how modules are wired in series within a string. A small shadow — a single cell partially covered by a chimney's shadow for an hour a day — can cut a string's output by far more than the shaded area's share of total surface would suggest, unless the electrical design accounts for it. Understanding why this happens, and what mitigates it, is central to both site assessment and system architecture decisions.

This article distinguishes near shading (chimneys, vent pipes, adjacent buildings, vegetation close to the array) from far shading (distant terrain, mountains, tree lines), explains bypass diode behaviour and its limits, covers sun-path analysis as the standard assessment method, and reviews module-level electronics — power optimisers and microinverters — as mitigation. It complements the orientation discussion in solar panel orientation and tilt, since shading and orientation should always be assessed together rather than sequentially.

The core message is that shading assessment is not optional due diligence — it is often the single largest source of the gap between a system's theoretical and actual yield, larger in many real installations than orientation or tilt sub-optimality.

Why partial shading costs more than its area suggests

A conventional crystalline silicon module wires 60-72 cells in series. Current flows through the series string at whatever the lowest-current cell allows, because series-connected elements must carry equal current. A single fully shaded cell in an otherwise sunlit string of 60 cells does not reduce output by 1/60; without mitigation it can reduce the substring's contribution to near zero, because a heavily shaded cell also risks operating in reverse bias, dissipating power (hot-spot heating) rather than generating it, while forcing the whole series path's current down to match its severely reduced output.

This series-string vulnerability, not simple area coverage, is why shading assessment focuses heavily on which cells or module sections are shaded and when, rather than treating shading as a simple percentage-of-area irradiance reduction.

Near shading versus far shading

Near shading comes from objects close to the array — chimneys, satellite dishes, vent pipes, adjacent roof sections, overhead wires, or nearby trees — and tends to cast a sharp-edged, fast-moving shadow that sweeps across the array over minutes as the sun moves. Near shading typically affects only part of an array at any given time but can affect different modules at different times of day, making its total energy impact harder to intuit without formal analysis.

Far shading comes from distant obstructions — hills, mountain ridges, tree lines, or tall buildings some distance away — and tends to cast a soft-edged shadow (due to penumbra effects over distance) that, once it begins, often affects the entire array simultaneously since the obstruction subtends a similar angle from every module's position. Far shading is generally easier to model from a single reference point using a horizon profile, while near shading requires point-by-point analysis, sometimes per module, because relative geometry changes meaningfully across even a small roof.

Bypass diodes: what they do and where they fall short

Nearly all modern crystalline modules include bypass diodes wired across substrings, typically dividing a 60-cell module into three substrings of 20 cells each with one diode per substring. When a substring's shaded cells force current below what the rest of the module can supply, the bypass diode conducts and routes current around that substring rather than through it, sacrificing that substring's roughly one-third contribution rather than collapsing the whole module's output to near zero.

Bypass diodes are effective and essential, but they operate at the substring level, not the cell level. A shadow covering even a single cell in a 20-cell substring can trigger the whole substring's bypass, losing a third of that module's output regardless of whether one cell or all twenty in that substring are shaded. This coarse granularity is why the number and arrangement of bypass diodes matters, and why shading a narrow strip that clips one cell in each of a module's three substrings simultaneously (a common effect of a horizontal chimney shadow, for example) can lose most of a module's output even though the shaded area is small.

At the string level, bypass diodes protect individual modules from each other but do not prevent a fully or heavily shaded module from still depressing the whole string's operating point if the inverter's MPPT tracks the string as a single unit, which is the underlying reason module-level electronics exist.

Sun-path analysis

Sun-path analysis maps the sun's position (azimuth and elevation) across the day and across the year for the site's latitude, then overlays it against a horizon profile of surrounding obstructions measured from the array location, typically using a fisheye-lens tool, digital site-survey app, or 3D model derived from site photographs or LIDAR data. The output is an estimate of what fraction of annual, monthly, or hourly irradiance is lost to shading at that specific point, often expressed as a "shading factor" reducing modelled yield.

For an array with multiple planes or a complex near-shading environment, sun-path analysis is usually done per module or per small module cluster rather than for the array as a whole, since near obstructions affect different parts of a roof differently through the day. This granularity is exactly what module-level electronics are designed to exploit: knowing which modules are shaded when allows a designer to decide whether string reconfiguration or module-level power management resolves the loss most cost-effectively.

Worked example: quantifying a shading loss

Consider a 20-module string (72-cell modules, three bypass diode substrings each) where a chimney shadow crosses two modules for approximately 90 minutes around midday during winter months (roughly November through February, about 120 days), fully triggering one substring's bypass diode on each of those two modules during that period. Outside winter, the sun's higher elevation clears the chimney and no shading occurs.

Assume each module is rated 400 W, with the affected substring contributing roughly one-third of that module's output when unshaded (approximately 133 W each), and that unmitigated string-level MPPT tracking during the shaded period causes the whole string of 20 modules to briefly track to a lower operating point set by the two impaired modules, rather than only losing the impaired modules' own substring output — a common real-world effect with a single string inverter and no module-level electronics.

If the string's normal midday output is 20 × 400 W = 8000 W, and the shading event depresses the whole string to roughly 70% of that value for the 90-minute window (a plausible order of magnitude for string-level MPPT mismatch loss, since the healthy 18 modules are forced away from their own optimum), the loss during the event is:

Loss power = 8000 W × 0.30 = 2400 W, for 1.5 hours = 3.6 kWh per shading event

Over 120 winter days: 3.6 kWh × 120 = 432 kWh/year lost to this single shading source, against an assumed annual string yield of roughly 10,000 kWh — a loss of about 4.3% from what looks, on paper, like a small chimney shadow on two of twenty modules. This is the kind of result sun-path and mismatch analysis is meant to surface before installation, when it can still be designed around cheaply, such as by adjusting string boundaries so the two affected modules share a string with a power optimiser or their own micro-inverter rather than a plain series string inverter.

Module-level electronics as mitigation

Power optimisers are DC-DC converters fitted to each module (or module pair), each performing its own maximum power point tracking before feeding a shared DC bus to a central string inverter. They limit shading impact to the affected module(s) rather than letting mismatch depress the whole string's operating point, addressing exactly the mechanism in the worked example above, though the string's total current is still ultimately determined by the inverter's overall DC input handling.

Microinverters go further, converting each module's DC output to AC individually, fully decoupling each module's performance from its neighbours; there is no shared string operating point at all. This gives the strongest per-module shading immunity but at higher per-watt cost and more devices to install and, over the system's life, potentially service.

When mitigation is not worth it

Module-level electronics add cost, complexity, and additional points of potential failure over a 25-year system life. Far-shading losses that affect the whole array simultaneously gain little from module-level electronics, since there is no mismatch between modules to resolve — every module is shaded together, so the loss is simply lower irradiance across the board, not solvable by better MPPT granularity. In these cases, accepting the modelled loss (or, where the far shading is severe, reconsidering the site or that portion of the array) is more appropriate than paying for module-level electronics that address a different problem than the one present.

The decision typically comes down to whether shading is symmetric (whole array together) or asymmetric (some modules shaded while others are not) at a given time — asymmetric shading is where module-level electronics earn their cost, while symmetric shading is a straightforward yield reduction no amount of clever electronics can recover.

Standards and references

• IEC 61215 — Module design qualification, including bypass diode functional requirements under partial shading test conditions.
• IEC 61730 — PV module safety qualification, covering hot-spot and reverse-bias endurance relevant to shaded-cell failure modes.
• IEC 62109 — Safety of power converters, applicable to power optimisers and microinverters as module-level power electronics.
• IEC 62548 — Array design requirements, including guidance on string configuration decisions that interact with shading mitigation choices.

Frequently asked questions

Can trimming a few tree branches meaningfully change a shading loss estimate?

Yes, near-shading sources such as vegetation are often the most cost-effective to mitigate directly, since removing or trimming a shading branch can eliminate a specific, quantifiable loss identified in sun-path analysis at far lower cost than compensating with module-level electronics. This is usually the first option to evaluate, particularly where the vegetation is on the system owner's own property and shading is concentrated at a specific time of day traceable to specific branches.

Do bypass diodes wear out or fail?

Bypass diodes can fail, most commonly from repeated thermal cycling caused by frequent shading events forcing them into conduction, which generates heat. A failed-short diode permanently bypasses its substring (a clear underperformance signal detectable via IV curve tracing), while a failed-open diode removes shading protection entirely, risking hot-spot damage during future shading events. Modules in chronically shaded locations are worth including in periodic thermal or IV-curve inspection routines specifically because of this failure mode.

Is a small amount of soiling the same problem as shading?

Soiling (dust, pollen, bird droppings) behaves similarly to shading at the cell level when concentrated, such as a bird dropping covering part of a single cell, and can trigger the same substring bypass and mismatch effects as physical shading despite covering a tiny area. Uniform, light soiling across a whole module is a more benign, roughly proportional irradiance reduction. This is one reason concentrated soiling is addressed with occasional cleaning rather than treated as a design-stage shading problem.

How accurate are smartphone-based shading assessment tools?

Modern site-survey apps using a phone's camera and horizon-detection algorithms provide reasonably accurate far-shading horizon profiles and useful approximate near-shading assessments for typical residential surveys, generally adequate for informing design decisions. For complex commercial arrays or borderline economic cases, professional 3D modelling using site-specific drone or LIDAR survey data gives materially better precision and is usually justified by the scale of investment involved.

Does shading loss scale linearly with the number of affected modules?

Not necessarily. Whether a shading event on two modules in a twenty-module string costs roughly the two modules' worth of output or considerably more, as in the worked example above, depends heavily on inverter architecture, string length, and whether module-level electronics are present. This non-linearity is exactly why per-site sun-path and mismatch analysis, rather than a simple area-proportional estimate, is the standard practice for any array with a non-trivial shading environment.

Sources

• PVEducation — Shading and bypass diodes
• NREL — PV system performance and losses
• Fraunhofer ISE — Photovoltaics report
• IEC — Standards for photovoltaic energy systems