[amluto]

There are many different ways to connect solar panels to the inverter. Usually the exposed metallic parts will be “grounded” (bonded to the ground / protective earth / equipment grounding conductor / whatever your jurisdiction calls it at the inverter).

But this is an array that looks a bit remote from any structure, and even if the electronics themselves are utterly, perfectly safe, the bonding wire will cause the panel frames to be at the potential of the inverter’s ground, and the soil will be at whatever its local potential is, and this may be different, resulting in a “touch potential”. You can mitigate this with a supplemental ground (or whatever your jurisdiction calls it) closer to the panels. There are fancy ways to do this particularly well to improve safety in the event of a fault.

If you are contacting the actual power wiring of the solar array, something is wrong. If you are feeling capacitive coupled current from the power wiring of the array, then I imagine this means your inverter has crappy filtering, and again the only way you would feel it is if the panel frame isn’t at the same potential as the local soil.

Oh, and if you live somewhere prone to lightning, you probably want to give a possible lightning strike to the panel a better route to a high quality ground than through the bonding wire to your house.

(In fact, in many parts of the world, including IIRC 48 US states, the power utilities may operate their distribution networks with a current carrying wire (combined neutral and ground) that is grounded at multiple points along its route, which will generate substantial leakage current through the ground. The exceptions are IIRC California and Wisconsin, and IIRC those states have their rules to help protect cattle from becoming part of an undesirable circuit while being milked. In CA at least, all the distribution transformers on the poles have at least two bushings on the primary side, and this is why. I believe this is California General Order 95, rule 33 [0].)

[0] https://ia.cpuc.ca.gov/gos/go95/go_95_rule_33.html

[quickthrowman]

> Usually the exposed metallic parts will be “grounded” (bonded to the ground / protective earth / equipment grounding conductor / whatever your jurisdiction calls it at the inverter).

It’s not usually, all of the panels and also the metal racking must be bonded and grounded.

Supplementary grounds are allowed by the NEC, article 250.54

There are zero requirements for supplemental grounds, you could use a single 24AWG conductor bonded to a brad nail or a #6 bare copper conductor bonded to (3) 5/8” dia’ 10’ ground rods. The fact that there are no requirements should tell you something about the necessity of supplemental grounds.

> 250.54 Auxiliary Grounding Electrodes. One or more grounding electrodes shall be permitted to be connected to the equipment grounding conductors specified in 250.118 and shall not be required to comply with the electrode bonding requirements of 250.50 or 250.53(C) or the resistance requirements of 250.53(A)(2) Exception, but the earth shall not be used as an effective ground-fault current path as specified in 250.4(A)(5) and 250.4(B)(4).

[amluto]

The OP is pretty clearly not in the US and is not subject to the NEC.

And just because the NEC doesn’t say you need to mitigate touch potentials for a specific installation doesn’t mean that you shouldn’t. The NEC is just a bunch of requirements, not a complete guide to building safe, high quality electrical installations. And there appears to be plenty of guidance from IEEE about this if you feel like paying for a copy of the standard:

https://store.accuristech.com/ecia/products/preview/2085911

The NEC does have plenty to say about grounding for purposes of reducing potential differences in other (non-solar) installations, though. Swimming pools are a notable example (with no specific requirement IIRC to connect the fancy bonding to the Earth at the pool, but a typical concrete pool with NEC-mandated bonding is likely an excellent connection to the local soil).

Amusingly, an inspector could plausibly require an “equipotential plane” per NEC 547.44 if the solar array is accessible to livestock.

Alternatively, from first principles: suppose you are standing with bare feet on wet soil 100 meters from a building. That building has a very nice ground rod array, giving an impedance of 50 ohms from the building’s ground to your feet. Now take a long wire that is insulated except at the ends, connect one end to the building ground, and firmly hold the other end in your hand. How do you feel about this situation? You have a nice wire loop, potentially with area on the order of 100m^2, containing your heart. It can inductively couple to time-varying magnetic fields (e.g. the fields around nearby transmission lines) and it can capacitively couple to other conductors. It can resistively couple to power lines with multiply grounded neutral or to electrified railways or to nearby buildings with old-style (earlier NEC revisions) shared neutral/ground feeds And lightning can strike the array. I think it would be wise to ground the array to earth to reduce the potential difference between the frame and earth.