Isolated and bonded earth rods in PV systems

[josealjim]

There are many opinions regarding grounding and earth equipotential rods in PV systems installations. I hope you could help me with this topic guys.
The installer and inverter company gave me some ideas about earthing pv systems and I have set SPD protections right before entering into the building. I though that 3 independent rods in triangle right near the solar array to connect PV and SPDs would be a better idea that running any EGC throughout the building or bonding GEC to the building rods. Im a bit worried about voltage induction between isolated earth rods and also about closed loops if we equipotential them.


WHAT DO YOU THINK ABOUT THE CONFIGURATION? I NEED YOUR HELP.





The SPDs will be installed outside with IP65 box, they will be connected by EGC to the PV electrodes bar (as you guys can see on the pic). To ground that box, can I use the same EGC or should I run another wire for it?


Looking forward to hearing from you,


thank in advance

 

[ggunn]

There are many opinions regarding grounding and earth equipotential rods in PV systems installations. I hope you could help me with this topic guys.
The installer and inverter company gave me some ideas about earthing pv systems and I have set SPD protections right before entering into the building. I though that 3 independent rods in triangle right near the solar array to connect PV and SPDs would be a better idea that running any EGC throughout the building or bonding GEC to the building rods. Im a bit worried about voltage induction between isolated earth rods and also about closed loops if we equipotential them.


WHAT DO YOU THINK ABOUT THE CONFIGURATION? I NEED YOUR HELP.



The SPDs will be installed outside with IP65 box, they will be connected by EGC to the PV electrodes bar (as you guys can see on the pic). To ground that box, can I use the same EGC or should I run another wire for it?


Looking forward to hearing from you,


thank in advance

One thing about your drawing, and I don't know if it is germane, is that in the inverter all the grounds are (usually, anyway) tied together. Your drawing shows isolation between the EGC and the GEC in the inverter.

 

[jaggedben]

First of all, your drawing doesn't show an EGC connection between the array and the inverter. This is a serious problem for the functioning of the ground-fault protection device in the inverter and the safety of the array regarding both fires and shock hazard. The idea that you can isolate the ground at the array from the ground in the building is probably a non-starter, although possibly you can put some surge protection in between.

Secondly, the code requires a bonding jumper connection between the DC and AC grounding electrodes, which could be read to include the one at your array. Depending on which code cycle you are on, the electrode at the array may or may not be required. If it is required to be connected to the AC electrode, it requires a GEC size conductor, not EGC. Something to ask your AHJ about.

Thirdly... what are you trying to protect against? Lightning surge entering the building?

As far as grounding the array, depending on how it is installed (carport?) the structural steel supports going into the earth may provide a much better electrode than any amount of ground-rods, arranged in a triangle or not.

 

[josealjim]

First of all, your drawing doesn't show an EGC connection between the array and the inverter. This is a serious problem for the functioning of the ground-fault protection device in the inverter and the safety of the array regarding both fires and shock hazard. The idea that you can isolate the ground at the array from the ground in the building is probably a non-starter, although possibly you can put some surge protection in between.

Secondly, the code requires a bonding jumper connection between the DC and AC grounding electrodes, which could be read to include the one at your array. Depending on which code cycle you are on, the electrode at the array may or may not be required. If it is required to be connected to the AC electrode, it requires a GEC size conductor, not EGC. Something to ask your AHJ about.

Thirdly... what are you trying to protect against? Lightning surge entering the building?

As far as grounding the array, depending on how it is installed (carport?) the structural steel supports going into the earth may provide a much better electrode than any amount of ground-rods, arranged in a triangle or not.

Following suggestions I have made another config.
- SPDs right near the PV earth rods
- Metallic conduit grounded at both ends
- Bare cooper grounding electrode bonding jumper





- What do you all guys think about this one?
- The distance between array and building is about 20m. I worried about lighting strikes induced voltages into the dc wires along the metallic conduit, I dont want any high voltage entering into the building. By grounding both ends of the conduit will they be avoid?
- What kind of EGC and GEC wire should I use (section,etc)





We dont want to break the asphalt to much in order to set the wires. Could i place the bare cooper bonding jumper as indicated in the picture?


Thanks

 

[ggunn]

First of all, your drawing doesn't show an EGC connection between the array and the inverter. This is a serious problem for the functioning of the ground-fault protection device in the inverter and the safety of the array regarding both fires and shock hazard. The idea that you can isolate the ground at the array from the ground in the building is probably a non-starter, although possibly you can put some surge protection in between.

Secondly, the code requires a bonding jumper connection between the DC and AC grounding electrodes, which could be read to include the one at your array. Depending on which code cycle you are on, the electrode at the array may or may not be required. If it is required to be connected to the AC electrode, it requires a GEC size conductor, not EGC. Something to ask your AHJ about.

Thirdly... what are you trying to protect against? Lightning surge entering the building?

As far as grounding the array, depending on how it is installed (carport?) the structural steel supports going into the earth may provide a much better electrode than any amount of ground-rods, arranged in a triangle or not.

Is the DC side grounded or ungrounded? It makes a difference; inverters for ungrounded systems do not (cannot) use reference to ground to detect ground faults.

 

[josealjim]

What do you think about this new config? Im working in thailand and i dont trust installers here so Id like arrange everything following NEC

I do not want to bond both end of the metallic conduit to different earths by running cable throughout the building




The two conduits config was given to me by the installer, so lets make it one:



Should the connection between conduit and array earth be made by bonding the array end of the conduit by EGC to the array bus bar and the house end of the conduit by bonding to the electrodes bonding jumper as attached?
The house end of the conduit should be bonded by GEC or EGC?

THANKS A LOT

 

[jaggedben]

Is the DC side grounded or ungrounded? It makes a difference; inverters for ungrounded systems do not (cannot) use reference to ground to detect ground faults.

An EGC to the inverter is still required to make sure that the GFDI detects ground faults on the array. There has to be an adequate path for fault current in order for a fault to be detected.

To me this is a paramount safety issue and the safety reputation of the entire industry is at stake if not paid attention to, regardless of what the code says.

What do you think about this new config? Im working in thailand and i dont trust installers here so Id like arrange everything following NEC

I still don't see an equipment ground running directly to the inverter. In your most recent config, the bonding jumper would probably provide an adequate fault current path, but under the NEC it not allowable to use this as a substitute for an EGC.

I don't see any reason you couldn't just put surge protection on the EGC itself, especially if the NEC is not at stake. Why not just go back to your original config and do that?

If you're not being held to the NEC, you don't need a bonding jumper between the electrodes at the array and at the building. You can dispense with that.

 

[ggunn]

An EGC to the inverter is still required to make sure that the GFDI detects ground faults on the array. There has to be an adequate path for fault current in order for a fault to be detected.

FWIW, many inverters are now transformerless and do not (cannot) use the EGC to detect ground faults.

 

[josealjim]

An EGC to the inverter is still required to make sure that the GFDI detects ground faults on the array. There has to be an adequate path for fault current in order for a fault to be detected.

To me this is a paramount safety issue and the safety reputation of the entire industry is at stake if not paid attention to, regardless of what the code says.



I still don't see an equipment ground running directly to the inverter. In your most recent config, the bonding jumper would probably provide an adequate fault current path, but under the NEC it not allowable to use this as a substitute for an EGC.

I don't see any reason you couldn't just put surge protection on the EGC itself, especially if the NEC is not at stake. Why not just go back to your original config and do that?

If you're not being held to the NEC, you don't need a bonding jumper between the electrodes at the array and at the building. You can dispense with that.

Thanks for your reply.

- So in order to allow the inverter to detect ground fault i should install and EGC from the array earth bus bar throughout the metallic conduit to the DC bus of the inverter (the inverters are transformerless).
- But we also have to install the grounding electrode bonding jumper between rods as NEC 2011 says.
- How do I bold the metallic conduit?
* I do use the metallic conduit as electrode bonding jumper, connecting the array end to the earth array bus bar and the other end (the one near the house) to the house earth rods by GEC?
* I install the electrode bonding jumper between array and house rods and i connect the array end of the metallic conduit to bus bar of the array and the house end to the bonding jumper? (In this case the metallic conduit wont be a substitute of the bonding jumper)
* I connect the array end of the conduit to the array bus bar and the house end of the conduit to the EGC mentioned before that goes to the inverter DC bus bar? (The bonding jumper between both earths would be installed as well)

Thanks again, u guys are helping me a lot

 

[jaggedben]

FWIW, many inverters are now transformerless and do not (cannot) use the EGC to detect ground faults.

FWIW, you are completely wrong.

First of all, the pre-operation test (Riso) performed by TL inverters measures the resistance between each conductor and ground, as explained at 2:30 in this SMA video. There is only one set of grounding terminals in any inverter, and that's what he means. So anything connected to the grounding terminals in the inverter, including an EGC, can and does get 'used' to detect a ground fault.

Secondly, the residual current device used during operation also has a (high impedance) connection to the grounding terminals (i.e. EGC if properly installed), as shown in this image from this article. Notice how the EGC is used to detect the ground fault in the example.

If the trip limit is 300mA, as stated in the figure, then that's about 1500 ohms on an array operating at 450V. That might not be too much for two electrodes at the array and inverter on a ground mounted array. But for safety, one should be measuring the resistance of the insulation between the PV conductors and the racking, not the resistance of the earth. Furthermore, for a rooftop mounted array on even a small wood framed residence, the resistance of the structure probably will be more than 1500 ohms. Which means that the GFDI will not and cannot detect a fault between PV conductors and racking without an EGC.

As someone who may be climbing on said rooftops to work on systems, I will trouble you to understand the purpose of the EGC and specify it in your designs. :rant:

 

[jaggedben]

- So in order to allow the inverter to detect ground fault i should install and EGC from the array earth bus bar throughout the metallic conduit to the DC bus of the inverter (the inverters are transformerless).

Yes. If you rely just on electrodes you could compromise the effectiveness of the GFDI. The bonding jumper between electrodes would probably also provide effective ground-fault detection, but in the NEC you have to run EGCs with circuit conductors.

- But we also have to install the grounding electrode bonding jumper between rods as NEC 2011 says.

Yes, so says the NEC, but I have read a lot of criticism of this when it comes to lightning protection, including (implicitly, I think) from the host of this forum. If you're not bound to follow the NEC, then maybe you should choose not to follow this part of it. I'm not enough of an expert to tell you which is better, just that there seems to be quite a bit of debate. I'm really not an expert on lightning protection.

- How do I bold the metallic conduit?

With a direct-burial rated bonding clamp, maybe.

* I do use the metallic conduit as electrode bonding jumper, connecting the array end to the earth array bus bar and the other end (the one near the house) to the house earth rods by GEC?

I wouldn't do that; you're inserting more points of failure. I would either run a wire EGC through the conduit or run the separate bonding jumper. Or both.

I also don't see why you need metallic conduit; you could use non-metallic conduit, or even direct-burial rated cable.

 

[ggunn]

FWIW, you are completely wrong.

First of all, the pre-operation test (Riso) performed by TL inverters measures the resistance between each conductor and ground, as explained at 2:30 in this SMA video. There is only one set of grounding terminals in any inverter, and that's what he means. So anything connected to the grounding terminals in the inverter, including an EGC, can and does get 'used' to detect a ground fault.

Secondly, the residual current device used during operation also has a (high impedance) connection to the grounding terminals (i.e. EGC if properly installed), as shown in this image from this article. Notice how the EGC is used to detect the ground fault in the example.

If the trip limit is 300mA, as stated in the figure, then that's about 1500 ohms on an array operating at 450V. That might not be too much for two electrodes at the array and inverter on a ground mounted array. But for safety, one should be measuring the resistance of the insulation between the PV conductors and the racking, not the resistance of the earth. Furthermore, for a rooftop mounted array on even a small wood framed residence, the resistance of the structure probably will be more than 1500 ohms. Which means that the GFDI will not and cannot detect a fault between PV conductors and racking without an EGC.

As someone who may be climbing on said rooftops to work on systems, I will trouble you to understand the purpose of the EGC and specify it in your designs. :rant:

Relax. Of course my designs are all well equipped with ECG conductors. I am not an expert on internal inverter design, but the way ground fault detection has been explained to me by the manufacturers of transformerless inverters is that they use embedded current measuring devices on the positive and negative DC inputs and flag a ground fault if there is a difference in the readings between them. Since the arrays are floating with respect to ground they cannot measure a potential to ground from either the positive or negative conductor as traditional DC grounded inverters can.

 

[jaggedben]

Relax. Of course my designs are all well equipped with ECG conductors. I am not an expert on internal inverter design, but the way ground fault detection has been explained to me by the manufacturers of transformerless inverters is that they use embedded current measuring devices on the positive and negative DC inputs and flag a ground fault if there is a difference in the readings between them.

That's true, but there's still a high impedance connection to ground in the inverter.

Since the arrays are floating with respect to ground they cannot measure a potential to ground from either the positive or negative conductor as traditional DC grounded inverters can.

They can and do, in the pre-operation Riso test.

 

[josealjim]

Yes. If you rely just on electrodes you could compromise the effectiveness of the GFDI. The bonding jumper between electrodes would probably also provide effective ground-fault detection, but in the NEC you have to run EGCs with circuit conductors.



Yes, so says the NEC, but I have read a lot of criticism of this when it comes to lightning protection, including (implicitly, I think) from the host of this forum. If you're not bound to follow the NEC, then maybe you should choose not to follow this part of it. I'm not enough of an expert to tell you which is better, just that there seems to be quite a bit of debate. I'm really not an expert on lightning protection.



With a direct-burial rated bonding clamp, maybe.



I wouldn't do that; you're inserting more points of failure. I would either run a wire EGC through the conduit or run the separate bonding jumper. Or both.

I also don't see why you need metallic conduit; you could use non-metallic conduit, or even direct-burial rated cable.

Thanks jaggedben for your help.

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:



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?



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?

BR and THANKS