Lightning and surge protection for PV installations

Lightning is a low-probability, high-consequence hazard for photovoltaic installations. A direct strike to an array or a nearby structure can destroy inverters, string combiners, and monitoring electronics within microseconds, while indirect effects from nearby strikes routinely degrade equipment over years without a single obvious failure event. This article is for installation designers and O&M engineers who need to size and place surge protective devices (SPDs) correctly rather than bolt on a generic surge arrester and call it compliant.

The core design task is threading two related but distinct disciplines: lightning protection (structural, per IEC 62305) and surge protection of low-voltage electrical systems (per IEC 61643-32 for PV-specific SPD selection). Getting the risk assessment wrong in either direction is expensive - underprotection leads to recurring inverter failures and insurance disputes, overprotection wastes budget on unnecessary Type I arresters and superfluous earthing runs.

We cover risk assessment, strike types, SPD classes and where they belong in a PV circuit, the physics of induced loop voltages, and how earthing systems must be coordinated rather than treated as an afterthought.

Lightning risk assessment

IEC 62305-2 provides a formal risk assessment method (R1-R4) based on structure dimensions, ground flash density (Ng, strikes/km2/year from lightning location network data), and the presence of nearby taller structures or overhead lines. For a rooftop PV array, the relevant question is usually not "will the array itself be struck" but "how much energy from a nearby strike will couple into the DC and AC wiring."

For ground-mounted PV plants without a taller structure nearby, the array is often the tallest object in the field, particularly with tracker rows or elevated fixed-tilt frames on flat land. That materially raises the collection area for direct strikes and justifies a formal external lightning protection system (LPS) analysis rather than reliance on SPDs alone.

Direct vs indirect strikes

A direct strike deposits currents that can exceed 100 kA into the point of attachment, following an idealised 10/350 microsecond waveform used for Type I SPD testing. Direct strikes to the array itself are rare for most rooftop and small ground-mount systems if an air termination system (LPS) intercepts the current and conducts it to earth via down-conductors kept at a safe separation distance from PV conductors.

Indirect effects arise from strikes to nearby structures, overhead lines, or ground within a few hundred metres, inducing transient overvoltages on DC strings and AC cabling through resistive and inductive coupling. These follow the faster 8/20 microsecond waveform used to test Type II SPDs, with peak currents typically in the 5-40 kA range at the point of installation. Indirect events are far more frequent than direct strikes and account for the majority of lightning-related PV equipment damage.

SPD types and placement

SPDs are classified by test waveform and intended duty:

On the DC side, SPDs are wired between each polarity and PE (or a functional earth reference for transformerless inverters, respecting manufacturer isolation requirements) at the string combiner and again at the inverter DC input if the cable run between the two exceeds roughly 10 m. Voltage protection level (Up) must sit comfortably below the inverter's rated impulse withstand voltage, and continuous operating voltage (Uc) of the SPD must exceed the maximum open-circuit string voltage at the coldest expected temperature, with margin - a common error is sizing Uc against STC Voc rather than the low-temperature Voc that can be 15-20% higher.

On the AC side, a Type II SPD is standard at the inverter AC output or at the main distribution board, with a Type I/II combined device at the service entrance when the building has an external LPS with down-conductors within a few metres of the electrical intake. Three-phase installations need SPDs across all live conductors to PE and, where the system is TT, a separate assessment of neutral-earth protection.

Cable loop area

Induced voltage from a nearby strike scales with the rate of change of current (di/dt) and the area enclosed by the conductor loop: V = -L(di/dt), where L depends on loop geometry. A DC string cable routed with positive and negative conductors bundled tightly together (twisted or run in the same conduit) minimises enclosed area and can reduce induced transients by an order of magnitude compared with routing the two polarities on opposite sides of a roof structure. This is one of the cheapest and most frequently ignored mitigations - it costs nothing beyond cable management discipline during installation.

Combiner box and inverter wiring should similarly avoid large loops between the SPD and its earth connection; SPD earth leads should be as short and straight as practically achievable, since a poorly routed 1 m earth lead with sharp bends can add several hundred nanohenries of inductance, materially raising the let-through voltage at the protected terminal during a fast transient.

Earthing coordination

SPDs are only as effective as the earthing system they discharge into. IEC 62305-3 requires down-conductors and earth electrodes to maintain a minimum separation distance (s) from PV metalwork and cabling to avoid dangerous side-flashes; where that separation cannot be achieved, the PV structure must be bonded directly to the LPS, converting an isolated system into an integrated one with equipotential bonding at every crossing point.

All earthing systems on a site - LPS earth, electrical installation earth, and any dedicated array frame earth - should be bonded into a single earthing system wherever permitted by the applicable wiring regulations, to avoid dangerous potential differences during a strike. A common design fault is installing a separate, unbonded "clean" earth for SPDs in the belief that it improves protection; in practice this creates exactly the potential difference the coordination requirement is meant to eliminate.

Worked example

Consider a 500 kWp ground-mounted plant with 20 string inverters, each fed by DC combiner boxes 40 m from the inverter. Ground flash density Ng = 2.5 strikes/km2/year (typical for temperate Europe), and the plant occupies a footprint of 300 m x 150 m without an external LPS air termination system.

Collection area for direct strikes, per the simplified IEC 62305-2 method for an isolated structure, approximates equivalent area Ae = L x W + 6H(L+W) + 9(pi)H^2, where H is structure height. Taking H = 3 m for a tracker row, L = 300 m, W = 150 m:

Ae = (300 x 150) + 6(3)(300+150) + 9(3.14159)(3^2) Ae = 45,000 + 8,100 + 254.5 Ae = 53,354.5 m2 = 0.0534 km2

Expected direct strikes per year Nd = Ng x Ae x 10^-6 correction is already embedded when Ae is in km2, so Nd = 2.5 x 0.0534 = 0.133 strikes/year, or roughly one direct strike every 7.5 years. That frequency alone justifies Type I SPDs at the main AC intake and Type II devices at every combiner box and inverter, even without a full external LPS, because the plant will statistically experience multiple direct or very-near strikes over a 25-year design life.

Common mistakes

• Sizing DC SPD continuous operating voltage (Uc) against STC Voc instead of the higher low-temperature Voc, leaving insufficient margin.
• Running positive and negative DC conductors on opposite sides of a structure, maximising induced loop voltage instead of minimising it.
• Installing a separate "dedicated" SPD earth electrode instead of bonding into the site's single earthing system, creating dangerous potential differences.
• Treating SPDs as fit-and-forget: thermal disconnectors and status indicators need periodic inspection, and Type I/II devices typically need replacement after known major strike events regardless of visual status.
• Omitting AC-side protection because "the inverter has internal protection" - internal MOVs are rarely sized for repeated Type I/II duty and degrade after a handful of significant transients.
• Ignoring the separation distance requirement between LPS down-conductors and PV array metalwork, risking side-flash into DC cabling.

Frequently asked questions

Do all PV systems need Type I SPDs?

No. Type I devices are justified where the structure has an external LPS with down-conductors near the electrical intake, or where risk assessment under IEC 62305-2 shows a meaningful direct-strike frequency, as is common for exposed ground-mounted plants.

Can one SPD protect both DC combiner box and inverter?

Only if the cable run between them is short (typically under 10 m); beyond that, induced transients between the two points can exceed the protection level at the inverter, so a second SPD stage is needed at the inverter DC terminals.

How often should SPDs be inspected?

At minimum during annual O&M visits, checking the mechanical status indicator and confirming there has been no visible discolouration or thermal disconnect tripping; sites in high-Ng regions benefit from post-storm inspection after any recorded strike within a few kilometres.

Does a metal roof provide any lightning protection on its own?

A continuous, adequately bonded metal roof can act as a natural air termination component, but this must be verified against IEC 62305-3 thickness and bonding requirements - a thin, unbonded profiled sheet can be punctured or can conduct dangerous currents into the building structure.

Is surge protection relevant for battery-coupled systems?

Yes, and arguably more so - battery inverters and BMS electronics are similarly vulnerable to induced transients, and DC coupling topologies can introduce additional loop area between PV, battery, and inverter that needs the same loop-minimisation and SPD placement discipline described above.

Standards and references

• IEC 62305-1 to -4: Protection against lightning (risk management, structural protection, LPS design).
• IEC 61643-32: Low-voltage surge protective devices - SPDs connected to the DC side of PV installations.
• IEC 62109-1: Safety of power converters for use in photovoltaic power systems.
• Fraunhofer ISE: field studies on PV inverter failure modes including surge-related degradation.
• IEA PVPS Task 13: reports on PV system reliability and failure statistics.
• NREL: technical reports on balance-of-system reliability for utility-scale PV.

For related design considerations, see cable sizing and voltage drop, MPPT explained, the system sizing calculator, the inverter sizing calculator, and the glossary for term definitions.