1 or 2 arrays grid-tied to different utility transformers & lightning damage

Rich66

I'm looking at a PV install at a farm which currently has two utility services from two different transformers. I am considering a large PV array and in order to prevent the upgrade of either transformer, half of the array would feed one transformer and half would feed the other transformer. What makes me concerned about this is that the two transformers are connected to the distribution line from different places (300' apart). Thus the array's grounding would create an conductive loop about an acre in area. I'm concerned that this large area would create an 'antenna' that would induce catastrophic voltages in the event of any nearby (or possibly even distant) lightning. Even if the arrays were separated by 10' or so and grounded separately, they are still on the same metal roof, so the metal roof would provide the connecting conductive loop. Does anyone have experience with this sort of problem?

Ethan Brush

Rich66 wrote: »

I'm looking at a PV install at a farm which currently has two utility services from two different transformers. I am considering a large PV array and in order to prevent the upgrade of either transformer, half of the array would feed one transformer and half would feed the other transformer. What makes me concerned about this is that the two transformers are connected to the distribution line from different places (300' apart). Thus the array's grounding would create an conductive loop about an acre in area. I'm concerned that this large area would create an 'antenna' that would induce catastrophic voltages in the event of any nearby (or possibly even distant) lightning. Even if the arrays were separated by 10' or so and grounded separately, they are still on the same metal roof, so the metal roof would provide the connecting conductive loop. Does anyone have experience with this sort of problem?

First I will give the disclaimer that I am not a lightning expert. Second I will say that there is a lot of hogwash and data not backed up by statistics and science by many people/companies selling lightning advice and equipment.

I think problems from lightning are extremely rare and in most cases not worth worrying about. Look at utilities: they have thousands of miles of "antennae" and have very very few problems. One may counter that they are very well grounded, but there are many ungrounded lines and even on an MGN (multi grounded neutral) system there are still the 3 hot lines that can't be grounded and have no "protection" at all.

As to your specific situation, i am not totally clear on your specific concern about " the array's grounding would create an conductive loop about an acre in area." Actually I think the metal roof provides a good equipotential plane that could help in a lightning event. It's probably much more likely that a (non lightning caused) fire will destroy the barn and that nice PV system than a lightning event.

jonr

My guess is that if you can't avoid two ground points, then run a ground wire between them (creating a preferred path) and don't run this wire alongside other wires.

BB.

Given that solar panel frames are ground bonded to metal roofs--There is no real difference for lightning flash over paths (frames of panels are closer to electrical paths than the metal roof itself). Plus metal roofs themselves are not really considered that good of lightning conductors (steel is not a great conductor, metal is thin, poor grounding from roof panel to roof panel).

Regarding capacitive current flow from Solar Panel to Roof... Probably depends to a degree on the type of GT Inverter used. The historical GT Inverter uses a transformer to isolate the DC panel side from the AC side. And usually, the solar panel electrical circuit is "lightly ground bonded" (usually through a 1 amp fuse or so) to earth/green wire ground... So, the direct ground bonding is "more intimate" than the distributed flat plate capacitance of the array). Plus, you will probably find a mix of capacitors from DC Panel leads to earth ground for radio frequency grounding too (to reduce radio frequency emissions that get the FCC upset).

For newer GT Inverters, they do not use an isolation transformer (save money, weight, possibly slightly more efficient). The panels are (more or less) directly connected to the AC mains. I would still expect inductors and various capacitor in the converter circuitry. And remember that the AC mains (transformer center tap or variations of that if three phase power) is usually earth bonded to at least two locations (at the distribution transformer, and at the main panel for the building/site). And the system green wire/safety ground is (usually) tied to the same ground wire, electrical boxes, metal conduit, etc..

So--I would be hard pressed to see the common metal roof between the two distribution transformer/separate utility feeds as causing any "unusual" common mode AC Capacitive Coupling issues regarding lightning.

Of course, if you have two solar power systems right next to each other, a close by electrical strike is going to affect both systems because of physical proximity.

One could calculate the flat plate capacitance of the solar panels... And compare that with the distributed capacitance of the electrical wiring inside of the conduit and see if they are of similar magnitude, or if the solar panels have much higher capacitance than the wiring (would give us some ideas.

If Google is correct, the formula for flat plate capacitance is:

C = e₀ * (A/d)

Where:

http://faculty.wwu.edu/vawter/physic...ParallCap.html

e_o = Permittivity of empty space (e_air ~ e_o) = 8.854x10^-12 C²/N.m²
e = Permittivity of material between plates
A = Surface Area of one of the plates (SI: m²)
d = Separation of the plates (SI: m)

Given that the distance from the panels to the metal roof is fairly large, the actual per panel capacitance is very small.

Would also need to look at the wiring inductance. A series inductor with capacitor(s) to ground is a standard low pass filter (blocks higher frequencies). Again, can look at the numbers and see what the frequency response would be...

Again, I would not worry (lightning does not like to travel more than a few tens of feet down standard electrical wiring--The inductance is high enough that lightning will want to find another path--arc to conduit, electrical connections in switch box, etc.).

But if still a concern, we have some more paths to follow and see what turns up (panel common mode capacitance, inductance of AC wiring (or possibly transmission line impedance would be more appropriate) in conduit, in free space, etc.).

-Bill