PV grounding configuration

josealjim

Hi and thanks in advance

Id like to give a brief explanation about my situation in order to make u understand my ignorance on these issues. I have 2 years of working experience, 1 year as internship working on solar thermal plants simulations and troubleshooting on site, on my second year (already with an engineer contract) ive been developing the automation of biogas generation and biogas upgrading plants, the executives told me that they wanted to introduce photovoltaic systems in our portfolio, so i started to study theory and simulation making simulations. There is no technical supervisor for pv projects here and we dont have any consultant company for advices, and the electrical engineers here (Thailand) dont care about how installers do. Thanks to this forum and information given by suppliers Im having a better idea about how a good installation should be made, so I really appreciate every suggestion given.

The engineers where I work told me that the lighting ground system and and main service ground system were not but after digging a bit I saw that we have one earth ring system for the building where LPS and MDB busbar are connected. There is a lot of controversy about the best PV grounding system when you dont rely on any code and I would like to have the PV parking we are building grounded with the best configuration possible.

After knowing that we have a common earth ring for LPS and building services and with all the information provided by this forum i made two configurations:
Attachment not found.
On this configuration, all the modules and structure are bonded and connected to a busbar at the array. From the busbar there is an PVC 16sqmm cable going to a 18 m far handhole, from there to the building ring and connected to the SPDs, the same PVC 16sqmm continues to the inverters for ground fault protection.
Although i dont need to rely on NEC code, it says that we can not use EGC as GEC. I dont know if this config I accomplish that

Is this config correct? Would it compromise the groundfault protection of the inverter?

What about this other one?

Attachment not found.
Here there is a GEC to the handhole and from there to the building ring. SPSs are attached to the GEC at the handhole. An EGC goes from the array busbar directly to the inverter for the groundfault protection. Is this config better?
There are no junction box for parallel connection outside the inverters, each string to its inverter dc input but the thai legislation makes you include a circuit breaker right before the inverter. If I size the EGC according to NEC 250.122. it would be an AWG14 wire but the installers already ran a pvc 16sqmm (they havent connected yet, we are waiting the inverters).

On this second config, the connection from the handhole to the earth ring of the building, should it be by bare copper with the same size than the ring?
Is correct to use both GEC and EGC with the same size 16sqmm although with a much smaller one for EGC would be ok?


In brief,
Which of both would you choose?
What would you change to make it better?

THANKS A LOT

BB.

Re: PV grounding configuration

If I understand your questions (and I am certainly not a lightning ground expert)--My choice is #1, and to not make the connection from the "local" PV grounding rod to the GEC on the inverter. The reasons:

1) this brings the potential for lightning energy directly into the building/GT inverter.
2) creates AC power ground loops through your equipment--I would try to avoid that.

The second drawing simply makes it even easier for lighting to be brought into the building/GT inverter/Electrical system (directly attached to PV array safety/lightning ground).

In both drawings, you have a separate green wire from the building single point AC ground rod brought directly into the GT Inverter GEC bar. Plus you are bringing the local Green Wire ground to the inverter chassis (which is probably connected to the GEC ground bar too).

Again--multi-point grounding. While not usually a bad thing for lighting--It does give you more chance of a ground loop, and bringing in energy (somehow) from different points in the building). Not neccesarly bad to multi-point ground, but, from my point of view, voltages are "local" to the GT inverter. You have the local green wire going with the AC leads together--They both will have the same surge voltages present (if any)--And that is "OK". Voltage differences are what cause the problems. With two different ground paths (and potentially two different ground reference voltages/locations), you now are bringing in two "ground" wires that can have large voltage differentials--Because of different ground bond locations (i.e., voltage gradient in ground from lighting or AC voltage injection) or because of the "loop" nature of the two ground paths (wires that follow the same physical path/bundled together do not develop differential voltages. Wires that follow different paths/have different reference points will develop differential voltages).

I have seen 60 VAC differences between grounds in a large salt water tank pumping/filtering system 75' apart (enough to give me a shock). With lightning, the voltage differences can be much more.

In reality--Lighting energy wants to only follow wire for relatively short distances 10-20 feet maximum or so, before impedance becomes high enough that it may want to find a different path. Taking the shortest/straightest path to ground with lighting bonds (large diameter, flat/woven cables, large bend radius, etc.), vertical ground rounds at corners/opposite sides of buildings (parallel current flow creates same polaritiy magnetic fields, lightning current flow wants to flow at outside edges of buildings on opposite parallel wires, not parallel wires next to each other)--So you don't want to bring lightning grounding (i.e., solar panel frames) down the middle/inside of the building--As the lighting may jump towards the the outside walls via any "better" conductive path (lightning is something that NEC has not always be good at).

The ground ring around the building is a good idea--With ground rods (as needed) around the perimeter of the building.

6 AWG cable seems to be a good minimum gauge of wire--I read about a church in Germany that had 8 and 6 awg equivalent gauge wiring and the 8 gauge blew after a lightning strike and the 6 awg did not.

Remember that grounding has two functions. The first function (NEC primarily) is to provide a safe return path for short circuit current from a branch circuit. The return wire only has to be heavy enough to "short" the hot wire current until the breaker/fuse blows. A single ground path is all that is needed--rated for the maximum available current needed to be shunted to ground (actually neutral bond at main panel typically). What we generally do is ensure there is only one grounding path. You do not want a 6 AWG cable in parallel with a 14 AWG cable as it is possible for current return through the 6 awg cable to find a 14 awg to be a lower resistance path--and overheat the 14 awg cable (shorter run, bad connection in 6 awg cable path, etc.).

The other reason is for lighting/surge protection. Lighting is uncontrolled current--So the actual cable diameter is a bit more magically based. Experiences from actual strikes provides lots of useful information here. There are many sources that discuss the current capabilities required of the cabling.

Also, remember that lightning is a "high frequency" current--Not DC. So "impedance" (inductive/capacitive/complex impedance) is important. The maximum frequency (again from a little reading) is something like 7kHz or so, meaning that you need to do RF (radio frequency) techniques for proper lightning grounding systems.

Here are some more discussions about lightning (and additional links in the threads):

BB. wrote: »

A couple threads about Lightning:

Off Grid Grounding Technique?
Another Question, this time about Lightning

Note, the above are discussions, not a do A, B, and C--and you will be "safe". There probably is no such thing with lightning. Several different techniques are discussed--and a few of those posters even have experience with lightning. :cool:

And our host's FAQ:

Lightning Protection for PV Systems

From other past posts here, Windsun (admin/owner of NAWS), he said that most of lighting induced failures he saw were in the Inverters' AC output section.

Towards the end of this thread is a very nice discussion of proper generator grounding.

-Bill

-Bill

BB.

Re: PV grounding configuration

Improvements... Hmmm...

If I was aiming for "bullet proofing" against lightning (really no such thing), I would be placing arrestors at the Solar Array Combiner box (if array is >10-20 feet from building surge suppressor).

And I would be installing suppressors at the AC mains panel. Just as good of chance of lightning hitting the AC mains as the solar array.

And, I would look at the Midnite suppressors that use MOVs vs the "standard" Delta type suppressors (arc gap+sand).

http://www.midnitesolar.com/products.php?menuItem=products&productCat_ID=23&productCatName=Surge%20Protection%20Devices

-Bill

josealjim

Re: PV grounding configuration

The connection from inverter to ground will be finally made by a combined AC/DC cable though the main service busbar as NEC 690.47 (3) indicates.

As you said, "Lighting energy wants to only follow wire for relatively short distances 10-20 feet maximum or so, before impedance becomes high enough that it may want to find a different path. Taking the shortest/straightest path to ground with lighting bonds".

So in the first configuration, being the shortest and best path to earth right before entering into the building (shortest connection to the building ring which has a very low impedance,about 1 Ohm ), Why is the lighting energy going to cross the building, pass through the inverter and through the main service busbar that is connected to the ring in order to reach earth?

Ive been posting on this forum (at the beginning without bringing the EGC) and on Mike holt's forum and they agreed that i should bring the EGC into the house:

- Cariboocoot moderator from this forum http://forum.solar-electric.com/showthread.php?23934-SPD-protections-in-37KWp-installation-out-of-LPS

- http://forums.mikeholt.com/showthread.php?t=164312&p=1594557#post1594557

- http://forums.mikeholt.com/showthread.php?t=163167

- http://forums.mikeholt.com/showthread.php?t=156506&page=2


And if I dont bring an EGC to the inverter, wouldnt I compromise the GFDI of the transformerless inverter? (well explained by Solarpro and jaggedben on the previous links )


Thanks

josealjim

Re: PV grounding configuration

BB. wrote: »

Improvements... Hmmm...

If I was aiming for "bullet proofing" against lightning (really no such thing), I would be placing arrestors at the Solar Array Combiner box (if array is >10-20 feet from building surge suppressor).

And I would be installing suppressors at the AC mains panel. Just as good of chance of lightning hitting the AC mains as the solar array.

And, I would look at the Midnite suppressors that use MOVs vs the "standard" Delta type suppressors (arc gap+sand).

http://www.midnitesolar.com/products.php?menuItem=products&productCat_ID=23&productCatName=Surge%20Protection%20Devices

-Bill

Sorry, i SPDs at the dc side of the inverter are included but i forgot to include them into the drawing

BR, and thank you for your support

BB.

Re: PV grounding configuration

Jose,

My question/concern is that you are "double connecting" your AC system grounds. You have the green wire ground from the Main AC Panel going to the inverter, and you have a second ground from a local ground rod/building ground ring going to the inverter. You show the grounds as (sort of) being separate inside the inverter--But in real life (generally) these grounds are all one in the same (i.e., all of the grounds eventually connect to the sheet metal chassis of the GT inverter).

For a lightning situation--This is like RF (radio frequency) grounding--Having lots of short ground connections that go directly to the building ring ground/local ground rods is ideal.

However, when you follow standard NEC (I know that does not apply in Thailand) practices--You have your green wire ground (and AC hot wires) coming from your main AC panel and "snaking" through the building to the point of use (GT Inverter in this case). When you tie the GT inverter ground directly to the ground ring/rod outside the building (and I would suggest placing a GT inverter on an exterior wall--Inside or Outside of building to avoid bringing possible lighting energy from the array into the middle of an occupied building--But you have SPD as building wall on DC wiring--So that is a good thing--Lots of caveats here) you have created a "ground loop".

Ground loops have several different issues.

One is that if you have a typical AC power short--The current no longer has a single path back to the AC main panel to trip a breaker/fuse. There are now several paths--The original green wire path, and a second (or multiple) path through the lighting ground system. You need to ensure that both paths are capable of managing the grounding current in a fault situation (typically 6 AWG wire is "good enough" for a standard home sized electrical panel--But need to check NEC or your local code for your application).

Another issue is that buildings have "physical size". When you have multiple ground rods (and grounding sources) a lightning strike will create a voltage gradient (high voltage near the strike dropping towards zero the farther away). So--now (for example) you have two ground rods 20 meters apart and one rod is connected to your GT inverter "ground point" and the other rod is connected to your AC main panel. And a lighting strike can now "energize" one rod relatively to the other with thousands of volts--Which will have to flow through the building green wire ground system and through the building ring ground. The impedance will be different between the two ground systems--So a different voltage will have to be generated.

And, there is the Loop Antenna effect. You have a loop antenna that is dozens(?) of meters across and it can "receive" lightning energy through the "loop" (the area of the loop is the "receiving area"--that is why we run wires in "pairs" and even twist them--To minimize loop area).

And that is what Cariboocoot and others have warned you about:

Cariboocoot wrote: »

BTW you may have an inadvertent ground loop formed by the inverter if the ground bus bar is in fact connected to the (case) ground terminal (as it probably is).

As Mike's video on the subject points out, not bonding the auxiliary electrode to your other electrodes to make it a single GES invites lightning induced earth gradients to send current through the EGCs and the equipment.
But not bonding the panel and racking metal to an EGC is a code violation (!!), so the proper solution is to bond the PV ground electrode to the rest of the GES through wire which is outside the building if possible.
Oh, and actual lightning current will go wherever it darned well pleases, unless it is completely diverted from the house by a properly engineered and installed lightning protection system, which has little to do with NEC grounding systems.

So, I am back to #1 with SPD on the DC lines (outside wall of building, would be nice to have SPD on array Combiner Box/Junction Box to protect panels), and I would suggest on your AC main panel bus. And not connect the ground from the GT Inverter Ground to the Local ground rod/ringing directly--But just use the standard building green wire ground system.

In the end, the whole idea for lighting control (as I understand--am not a lighting engineer) is to intercept lightning energy on the outside of the building and divert to your ground rod/ring before it ever gets into the building (Faraday cage concept). The real life problems are that wires and cables have impedance (complex resistance that changes depending on the distances and frequencies of the energies involved).

What I "want" is for the equipment/wiring inside to the house to be all at the same potential (ideally). And when you are dealing with lightning or improperly designed/installed/maintained grounding systems--Is you can literally have hundreds to thousands of volts difference between a metal sink/plumbing, an AC power outlet, a corded telephone, cable TV, outside yard lighting switch, etc. inside the home.

More or less, 120/240 VAC equipment is designed to work with as much as 600 VAC potential between the wiring and the "local ground". And the wiring/equipment (if NRTL/UL/etc. approved) is highpot tested to 1,800 VAC minimum between AC wiring and equipment chassis (or copper wiring insulation to conductive solution( during manufacturing.

The SPD is designed to short any voltages that exceed operational voltages and shunt them to some sort of local safety ground and keep those voltages from entering the home/building.

In the US/North America, we also ground bond our AC wiring (think of a center tapped 240 VAC transformer. The center tap reads 120 VAC to L1 or L2, and the center tap is grounded to the local earth rod/cold water pipe/etc.). The bonding helps prevent the AC wiring from ever being energized over 120 VAC with respect to earth ground (for example, a 12,000 Volt pole top distribution wire falls on the 120/240 VAC transformer output wiring to our homes).

In Europe (and much of the world?), I don't believe it is common to ground bond one of the 230 VAC lines--So the use other methods (I hope) to ensures safety.

You mentioned that you have a transfomerless GT inverter... That means it does not have an isolation transformer between the DC of the solar panels and the AC mains of the building. Is that safer, not safer, or "the same" as an isolated GT inverter? Per code, they are both the same (safety wise) as long as each passes the general UL/IEC/etc. electrical requirements.

One of which is to detect if there is a ground fault in the DC array wiring. The only "electrical" requirement for this to work reliably is that the GT inverter is grounded to a local ground rod (typically the main AC panel ground bus/rod connection) and the solar array conduit (if metal) and racking is also grounded too. From an NEC point of view, running a 6 AWG wire from array frames and metal racking back to the AC panel ground is "good enough". It just needs to manage enough current to detect a DC (or 50/60Hz) electrical fault.

From a Lightning point of view grounding--The array should have its own ground rod to direct energy from lightning to earth ground. The 6 AWG wire from the Array Ground to the AC Panel panel ground is to handle short circuit current so the circuit breaker will trip (rather than energize a piece of metal/equipment). It only needs to manage 50/60Hz or DC current--So complex impedance is not usually an issue.

-Bill

josealjim

Re: PV grounding configuration

Thanks Bill again,

I dont want to trouble u anymore so lets see if we can conclude the thread with this post.

In order to avoid the ground loop and diverge any High induced voltage by lightning i came up with the followings solutions:

First (This config does not diverge lighting but avoid the ground loop)
Attachment not found.

On this first one we dont connect the PVC 16sq ground wire to the closest point of the ground ring, bringing possibles Lightning inside the building (I would not like to do that).

Second

Attachment not found.

On this second the PV ground wire doesnt come to the inverters so, could I have problems with the isolation monitor of the inverter?

Third
Attachment not found.
On this third one the EGC goes to the inverter but DC busbar and AC ground inside the inverter are not connected by removing the bonding wire (the inverters warranty would be lost if we open the case).
This is my favorite config because the GFCI will work perfectly, its isolation monitoring device between EGC and DC wires + the residual current detector.

Im between (B) and (C). Which one would you choose? I hope i came up with a good understanding

Will the isolation monitor of the transformerless inverter work correctly if the ground to measure the impedance is the one from the MDB busbar? The MDB busbar will be connected to the array by the building ring, so I suppose the the impedance between Array earth connection point and DC cables, and MDB busbar and DC cables would be the same right? If this is correct I could use (B) so Ill not have problems with the warranty.

THANKS a lot,
Jose

BB.

Re: PV grounding configuration

On Number 1, if you wanted to do this way--I would add the bond between the building ground ring and the ground wire from the PV Array frame+mount ground wire (want the lightning energy to go to ground first before it has a chance to enter the building (even if there is a ground loop inside the building--ground loop being a "secondary problem").

Also, if the solar array is any distance from the building (more than 3-6 meters away), I would add a ground rod at the base of the array mounting.

Number 2, what I would suggest is the "standard" ground connection method.

Note, on all drawings, the AC SPD's should be mounted on/near the AC main panel (to protect against energy entering the main AC panel.

Number 3, my second choice--But much better than #1. For lightning control--I really don't know if #2 or #3 is "best". #2 no ground loop vs #3 which may be better at diverting lightning energy to earth ground. I would be able to accept either--I guess (sorry--I am trying to reason out with you at the same time I type this).

In the US--we have the choice between plastic or metal conduit. If you use metal conduit going into the building and plastic conduit (or j-box) to "open" or break the ground loop.

Metal conduit does have the advantage of acting like a capacitor to pickup lightning energy and divert to ground (we use shielded cable/cable in conduit to reduce radio frequency emissions and prevent outside emissions such as electrostatic discharge events from getting into computers and causing them to crash).

If you run metal conduit/pipe from the array to the GT inverter, then only Number 3 is possible (i.e., the conduit shorts all of the grounding together anyway). And metal conduit is very nice for helping with lightning control.

In all cases, adding a ground rod at the base of the array to ground the green wire will be recommended. If the array were on the roof of the house/building, run a green wire (lightning ground) straight down the the outside wall (shortest run) to a ground rod next to the foundation. Run a 6 awg or larger wire from the array ground rod to the building/local ground rod/ground ring to tie all together for AC ground safety. It could be the same green ground wire you already have running from the array to the ground ring.

The GT inverter's ground detection/insulation test system, etc. should not be affected by any of the grounding systems we have discussed here. As long as the +/- lines from the DC Solar Array are not connected to any grounding system.

You should not have any DC GFI/Resistance problems--But if you do, it could be from the SPD's -- Especially if some time in the future they are activated by lightning--You may have to replace them at that point to get the GT Inverter to work again (carbon/damage from lightning energy dissipation).

There is one other issue with the "new style" GT inverters--They have installed DC Arc Fault detection per the new safety requirements. There have been some reports of false tripping by the DC Arc Fault detection circuits and some field retrofits. You might research your GT inverter choice(s) and make sure that they are not affected.

Let us know what you decide. And--I hope--We don't hear anything about it again (all works well, no lighting hits your place, etc.).

-Bill

josealjim

Re: PV grounding configuration

Ill definitely go then for the previous one, ignoring the ground loop as many people have suggested.

Attachment not found.

We wanted at the beginning to install a ground rod right near the array but the impedance to earth (three rods in triangle config) was > than 25ohm. Thai standards dont allow that, it has to be =< 5 ohms so the best and nearest connection to earth was the building ring.

Thanks a lot to everyone for his patience and support. I hope this thread helps other people.