GFDI PV ground faults

[Jpflex]

what method of testing or troubleshooting would yall recommend for finding a possible ground fault within a field solar array?

On our pv site we have solar arrays feeding inverters to produce to the grid.

Often times a field ground fault trips the inverter fuse wich is in series with the pv negative side. At this point the inverter chassis is bonded to the negative pv arrays similar to service neutral bonding point at one location

Currently my company has technicians test voltage to ground at this GFDI negative point with GFDI fuse open. The negative wire to ground is neg 600 vdc and supposedly will drop 2 volts per second if a ground fault is present in the field.

I would rather wrap an ammeter around the inverter chassis ground and check for current draw appraching 5 ampere limit instead of this voltage drop test.

The voltage test has not always been accurate as faults have been in the inverter on one ocation and not tge field while using this test.

I also dont understand the voltage drop test?

 

[tortuga]

Can you post some more info on the tester? Perhaps a photo? Sounds like a megger

 

[Jpflex]

Can you post some more info on the tester? Perhaps a photo? Sounds like a megger

The company does not want to use a megger. Currently only using a DVOM on 20k and 200k ohm scale.

Checked between ac phase busses and ground. Dc buss and ground

 

[jaggedben]

There are many PV strings? There is a combiner box for each inverter? More details would help.

A basic insight is that in order to find a ground fault on a particular string you need to isolate all strings from each other, both positive and negative conductors. It's useful if you have fuse holders or MC connectors placed at the combiner or inverter to isolate strings for this kind if testing. In both positive and negative.

Once the string with the fault is isolated, it's easy to find it; it will be the one with positive and negative voltage to ground. You can even do the math to find which panel it's at.

 

[ramsy]

The company does not want to use a megger.

Good policy to prevent knuckleheads from destroying the solar array.

 

[Jpflex]

There are many PV strings? There is a combiner box for each inverter? More details would help.

A basic insight is that in order to find a ground fault on a particular string you need to isolate all strings from each other, both positive and negative conductors. It's useful if you have fuse holders or MC connectors placed at the combiner or inverter to isolate strings for this kind if testing. In both positive and negative.

Once the string with the fault is isolated, it's easy to find it; it will be the one with positive and negative voltage to ground. You can even do the math to find which panel it's at.

I contract to repair the inverters which our company provids for the solar array field. We do not ssrvice the owner field but we do assist the owner customer in ground faults at times.


Typically we only varify if ground fault is in the field or in our inverter. There are individual string boxes with disconnects per solar array. The negative feed from pv field is bonded to the inverter at the GFDI.

The company has us open the gfdi at inverter dc pv field feeder input, if not already blown and test negative pole to ground. If voltage drops from negative 600 vdc at 2 volts per second it is assumed fault in field?

 

[jaggedben]

..... If voltage drops from negative 600 vdc at 2 volts per second it is assumed fault in field?

That does not sound like a reliable rule of thumb to me. My experience in these types of systems is with smaller systems and a while back now, but my experience contradicts that such a reading necessarily indicates a ground fault. Also in a more severe fault on a combined output conductor, the reading from the faulted conductor to ground would simply be zero.

Also that type of reading where you see the voltage dropping is in part I believe an artifact of using a high impedance meter so you need to understand the meter characteristics too.

 

[Jpflex]

That does not sound like a reliable rule of thumb to me. My experience in these types of systems is with smaller systems and a while back now, but my experience contradicts that such a reading necessarily indicates a ground fault. Also in a more severe fault on a combined output conductor, the reading from the faulted conductor to ground would simply be zero.

Also that type of reading where you see the voltage dropping is in part I believe an artifact of using a high impedance meter so you need to understand the meter characteristics too.

Well some of those things i brought up to the company. I believe our DVOM is 10 mega ohms of impedance.

The voltage drop test has not proved reliable since this test has been used to "indicate ground fault in the field," but ended up being in the inverter

The reason inmade this post because the concept or theory of how this works to find field ground faults made no sense

 

[tortuga]

Its easier for me to translate electrical systems into IEC terminology even DC systems,
For the IEC terminology;
The first letter indicates the relationship between the DC source (PV panel) and a Grounding Electrode System (GES) or Earth;
"T" = Direct electrical connection of a point on the DC system to a GES or Earth (Latin: terra)
"I" = All live parts isolated from Earth (Latin: īnsulātum), except a high impedance.

The second letter indicates the relationship between the exposed-conductive-parts of the installation, and a GES (Earth):
"T" = Direct connection to a GES independent of a any DC connection or lack there of to a GES.
"N" = Direct connection of the GES to a point on the DC system either a neutral, positive grounded or negative grounded. On DC telcom system's that I am very familiar with its 'positive ground' to make -48V.
-- hyphen --
Then after the hyphen describe the protective conductor typically a separate Equipment Grounding Conductor (EGC)
"S" = If protective functions provided by the EGC alone (or frame).
"C" = If one of the DC current carrying conductors is combined with the EGC protective function' like the negative battery ground on a vehicle frame.

So for types of PV in the NEC 690.42 you'd get either a
Solidly grounded system = TN-S
A functionally grounded system or a ungrounded system = IT-S

I am not sure if there are other types but thats how I would go about understanding the system.
Then depending on the system type you know what tests to do....

 

[jaggedben]

...
So for types of PV in the NEC 690.42 you'd get either a
Solidly grounded system = TN-S
A functionally grounded system or a ungrounded system = IT-S

I am not sure if there are other types but thats how I would go about understanding the system.
Then depending on the system type you know what tests to do....

Honestly I dunno how much this helps. A system with a ground-fault fuse looks like TN-S when the fuse is good and IT-S when the fuse is open, except that if it's open because of ground fault then it may look like TN-S.

 

[Jpflex]

That does not sound like a reliable rule of thumb to me. My experience in these types of systems is with smaller systems and a while back now, but my experience contradicts that such a reading necessarily indicates a ground fault. Also in a more severe fault on a combined output conductor, the reading from the faulted conductor to ground would simply be zero.

Also that type of reading where you see the voltage dropping is in part I believe an artifact of using a high impedance meter so you need to understand the meter characteristics too.