Where to ground the DC

[bwat]

When using inverters that allow a DC pole from the PV array to be grounded, where and how does this grounding (bonding) usually take place? And is it always at the same location?

I'm envisioning almost like an MBJ at each inverter (assuming string), but I suppose it could be somewhere else. What if it's a larger central inverter?

 

[jaggedben]

Only tiny off-grid systems are solidly grounded on the DC side these days. Basically systems that don't have an inverter. I don't think there's any convention or requirement other than being on the line side of disconnecting means for loads.

Everything else is either ungrounded, non-isolated (i.e. effectively grounded on the AC side), or functionally grounded with a device supplied and connected according to the inverter manufacturer. That would include a large central inverter, whatever you mean by that.

 

[pv_n00b]

Assuming you have an isolated inverter that allows one pole of the PV array to be grounded the NEC (690.41) has two choices:

1. If the array voltage is greater than 30V or the Isc is more than 8A then you must have ground fault protection. This is usually provided by the inverter where the array pole is connected to ground internally through a GFDI fuse.
2. If the PV array has not more than two modules in parallel and is not on a building then it can be solidly connected to ground. This connection can be anywhere in the PV array but close to the inverter is better.

Larger central inverters usually still use the GFDI fuse to connect a PV array pole to ground inside the inverter.

 

[bwat]

Thank you both. I hadn't physically seen where this connection occurred with central inverters (we have mostly have been using ungrounded with transformerless string inverters since I started getting involved). It makes sense that I didn't see where this was occurring now that I know it usually happens internally through GFDI fuse.

Insightful as always.

 

[ggunn]

Thank you both. I hadn't physically seen where this connection occurred with central inverters (we have mostly have been using ungrounded with transformerless string inverters since I started getting involved). It makes sense that I didn't see where this was occurring now that I know it usually happens internally through GFDI fuse.

That's the way it worked with transformer coupled inverters but inverters now are mostly (all?) transformerless and don't have GFI fuses.

 

[bwat]

That's the way it worked with transformer coupled inverters but inverters now are mostly (all?) transformerless and don't have GFI fuses.

Are you seeing that with 2-3MW central inverters as well? My perspective is limited, but I believe all of the central inverters that I've seen are transformer coupled where you had the option to ground or not.

 

[ggunn]

Are you seeing that with 2-3MW central inverters as well? My perspective is limited, but I believe all of the central inverters that I've seen are transformer coupled where you had the option to ground or not.

My experience tops out at 1.4MW done with transformerless string inverters.

 

[pv_n00b]

There are a lot of people in the PV industry today who have never worked with an isolated, transformer-based, inverter. Having to talk about GFDI fusing again reminds me of the early days of PV, having to explain it to AHJs. A blast from the past.
The only place we will run into this now is retrofitting old PV systems that have isolated inverters and in central inverters. Central inverters now are around 2MVA and only start to show up in PV systems over 5 to 10MW.

 

[gene6]

If there is no isolation transformer then pardon my ignorance but what happens when an inverter malfunctions ? Say you have a 480 Volt AC side connected to a typical 500kVA pad mount and a i don't know 750 Volt DC side? What prevents one side from leaking to the other? Caps? A micro chip? There could be a ton of available fault current on the 480 side, not much on the DC side. Could one have a situation with over 600 volts DC leaking thru to the AC side?

 

[tallgirl]

If there is no isolation transformer then pardon my ignorance but what happens when an inverter malfunctions ? Say you have a 480 Volt AC side connected to a typical 500kVA pad mount and a i don't know 750 Volt DC side? What prevents one side from leaking to the other? Caps? A micro chip? There could be a ton of available fault current on the 480 side, not much on the DC side. Could one have a situation with over 600 volts DC leaking thru to the AC side?

I assume the larger transformerless inverters have the same power topology as the smaller ones I worked on. There are large capacitors which are charged from the DC input and those are switched, usually via pulse width modulation, to the output past some filter elements.

Which part of that inverter fails is going to depend on which part of the output section failed.

The more dangerous situation would seem to be where a FET failed so that it was conducting ("on") and the array section was continuing to charge the capacitors. In that situation you'd potentially have a lot of very high voltage DC on an output, potentially up to the open circuit voltage from the array.

 

[gene6]

I was hoping for a post that said something re-assuring to make me feel better.
I don't have much experience with caps other than motors and I know they can dead-short, so hopefully this is not the case with inverters.

 

[pv_n00b]

I was hoping for a post that said something re-assuring to make me feel better.
I don't have much experience with caps other than motors and I know they can dead-short, so hopefully this is not the case with inverters.

In theory, it's possible but I am not aware of it ever happening in the field. That I think would be odd enough for us to hear about it. Just like it's possible for a grid-interactive inverter to keep running after the grid disconnects if the conditions are just perfect.

 

[tallgirl]

In theory, it's possible but I am not aware of it ever happening in the field. That I think would be odd enough for us to hear about it. Just like it's possible for a grid-interactive inverter to keep running after the grid disconnects if the conditions are just perfect.

Excellent example. It actually takes a bit of smarts for some very keenly balanced loads to result in a disconnect by the anti-islanding controls.

 

[jaggedben]

Every block diagram of an inverter I've ever seen (that was bigger than a micro-inverter) has relays that physically connect and disconnect the AC output. You also hear them click when an inverter starts operating. So if some fluky catastrophic failure happened in the power electronics, I think the control electronics would still almost immediately open the relays and prevent any kind of raw PV voltage from going to the AC side. They'll use normally open relays so that they most likely fail safe open if the controls stop working. So you'd need something that caused catastrophic failures in both power electronic and control components, but at the same time not enough to destroy the inverter so much that power to the relay coils was interrupted. So like pv_noob said, in theory is possible. But only barely. Put another way, I think they thought of that.

 

[pv_n00b]

Every block diagram of an inverter I've ever seen (that was bigger than a micro-inverter) has relays that physically connect and disconnect the AC output.

I have not seen this. I'm sure there might be some inverters out there with isolation relays but they are not common.