Misleading fault indications [RE-wrenches]

Tue Mar 4 20:55:02 PST 2008

William, John, and all,

I've been silent on this issue for years, as have others I'm sure, and I'm
glad to see it coming to light. The method of providing ground fault
protection of PV arrays by ungrounding the grounded conductor is an anomoly
in the NEC, as best I can tell. Discussions I've had with John Wiles suggest
that this approach has been used  because it's the most cost-effective
approach to accomplishing the task at hand, which is to reduce the
possibility of a PV ground fault leading to a fire. But it does beg
questions that William has raised about alternative approaches. Trouble is,
it may be a touchy issue with respect to the rest of Article 250. I'm not
sure how many inspectors really understand and appreciate PV GFP and the
implications for a "grounded" PV array, and I'm not sure that the PV
industry wants to invite further scrutiny on the matter.

Perhaps John Berdner can speak to some other possible GFP scenarios, based
on his depth in this industry. I know that several different approaches were
proposed and abandoned due to the expense of implementing them. I believe
that the "net current difference" approach used in GFDI's was one such.
Placing a fuse or a breaker in place of a main bonding jumper allows the
array to be called "grounded" for sake of argument and code, while providing
a low-cost approach to reducing the possibility of arcing and starting a
fire in the event of a ground fault. The issue is, without ungrounding the
array, how do you deal with the arcing potential of a fault to ground? How
do you "shut off the array" to interrupt this fault to ground? GFP does
nothing about line-to-line faults, and the potential hazards they pose, but
they're less likely that a ground fault. Ungrounding the grounded conductor
seems like the simplest approach. The question is what's the best method to
do this.

One irony I've seen in this is the insistence on having a DC grounding
electrode conductor run to the inverter because the DC side of the system is
a "separately derived system." Given that the grounded conductor on the DC
side is only nominally grounded, I do not understand the pressing need for
the GEC -the grounded DC conductor is allowed to become ungrounded under a
fault condition, and should NEVER be treated as a "grounded" conductor for
safety reasons. Requiring the DC GEC is an onerous burden in some
installations. I'd be interested in John's thoughts on its functional
utility.

William, I've considered the exact problem you raised in your inverter
internal ground fault scenario. I've thought that a GFP device that's able
to sense current flowing in the main bonding jumper (like the Outback GFP)
and open both array positive AND array negative, isolating the array
negative from the grounded negative in the rest of the system, would be
fairly straightforward to produce. It would be similar to the Outback GFP,
except the 1 amp sensing-tripping pole would not open -just sense the
current and trip the other two poles open. This device WOULD still trip in
the inverter fault scenario you've put forward (which would unground the
array admittedly), but since negative would remain bonded to ground, the
inverter breaker would also trip, indicating the cause of the ground
fault -a better result, I think. And with the present GFP devices, the array
becomes ungrounded with any ground fault anyway, so no greater danger is
created by this approach. At least the battery side of the system would be
unambiguously grounded.

My $1.05

Dan Rice
Abundant Sun, LLC.

----- Original Message ----- 
From: "William Miller" <wrmiller at charter.net>
To: <RE-wrenches at topica.com>
Sent: Tuesday, March 04, 2008 7:47 PM
Subject: RE: Misleading fault indications [RE-wrenches]

|
| John:
|
| Thanks for the detailed background on the fault scenario and the
| code-making process.
|
| I try to differentiate between a fault in which the negative current
| returns on the equipment grounding conductor, the positive current returns
| on the grounding conductor or a plain old pos. to neg. short circuit.  I
| don't know if it is important to differentiate these faults here but I
| suspect it is.
|
| I have some questions:
|
| 1. John, do you have any concerns about the scenario I presented?  It just
| does not seem right that a fault in an inverter indicates as a fault in
the
| PV and that the inverter continues to have power supplied to it.
|
| 2. Also, do you share my gut reaction that un-bonding a neutral just to
| accomplish ground fault detection is somehow not the straightforwards
| approach?  The drawing that indicates this scenario is located at:
| http://mpandc.com/case_studies/ground_fauilt/ground_fault.html  It is the
| second study from the top.
|
| 3. Has anyone attempted to detect ground faults by monitoring the
| difference in current on the positive lead versus the negative?
|
| BTW, the top study on that page indicates a scenario as you mentioned in
| your reply and illustrates the importance of ground fault detection.
|
| Sincerely,
|
| William Miller
|
|
|
| At 05:19 PM 3/4/2008, you wrote:
|
| >Mark/Mark/Wrenches:
| >
| >I don't really have anything to add to the original question but
| >perhaps some background would help frame the issue - or at least muddy
| >the water further ;^)
| >
| >The issue of a low impedance fault in the grounded PV conductor is an
| >interesting one.
| >It is hard to detect this type of fault since a very low impedance
| >fault looks remarkably like an intact fuse or closed breaker between the
| >grounded conductor and ground.
| >During the development of UL 1741 we had some spirited discussions on
| >the topic and it is clear that the intent of the GFDI requirement (per
| >NEC and UL1741) was to cover all types of ground faults including those
| >in the grounded conductor.  Not all implementations of GFDI circuits I
| >have seen will do this and, because the NEC or 1741 are not explicit on
| >this point these devices continue to be Listed.  I am not saying anyone
| >is wrong or right here only that the Standards do not explicitly address
| >the issue.
| >
| >If we look for guidance it is sometime useful to go back to the
| >original purpose of the GFDI.   The GFDI is there to eliminate the
| >possibility of ground fault currents flowing in "unintended conductors".
| >   The idea is that if you have a ground fault you will see currents
| >flowing in support structures, conduits, hangers, and all kinds of metal
| >bits that might be connected between the fault and ground.  Conduits for
| >example are supposed to carry the fault current long enough to open the
| >over current device feeding the circuit.  This is part of the reason we
| >use conduit, it is normal, and the fault current is usually only very
| >short term.
| >
| >The big BUT here we size our normal PV over current devices at 1.56 Isc
| >so it is possible for these ground faults currents to flow continuously
| >for indefinite periods of time without tripping the normal PV over
| >current devices.  None of this "metal stuff" is intended or evaluated to
| >carry current for long periods of time and so it can possibly over heat
| >and thereby create a fire hazard.  Originally the NEC only required GFDI
| >for rooftop mounted arrays on dwellings (which includes more than just
| >homes by the way).  A couple of years ago there was a ground fault on a
| >ground mounted medium sized commercial system and the fault melted
| >through the side of the metal conduit. This was the seminal event which
| >prompted the recent NEC requirement for GFDI in all grounded PV systems
| >(which I support and believe is a resonable requirement).
| >
| >If you have a low impedance fault in the grounded conductor you can
| >clearly get fault currents to flow in the "unintended conductors" and
| >therefore the GFDI should detect it and interrupt the fault current.
| >Nicked cables in the grounded conductor are certainly not unheard of and
| >this would normally blow the GFDI fuse or trip the GFDI breaker.  This
| >type of fault normally has a high enough impedance to be detected and
| >trip the GFDI circuit as well.  It is possible, however, (although some
| >would argue only theoretically) to have a low impedance fault that will
| >trip the GFDI over current device but remain undetected.  In that case
| >there is a possibility, albeit remote, for currents to be flowing in the
| >unintended conductors.  Will it be detected ?  This all depends on the
| >nature of the ground fault and how the GFDI detection circuitry was
| >implemented.
| >
| >Best Regards,
| >
| >John Berdner
|
|
| - - - -
| Hosted by Home Power magazine
|
| To send a message: RE-wrenches at topica.com
|
| Archive of previous messages:
http://lists.topica.com/lists/RE-wrenches/read
|
| List rules & how to change your email address:
www.mrsharkey.com/wrenches/etiquette.php
|
| Check out participant bios: www.mrsharkey.com/wrenches/
|
| Moderator: michael.welch at homepower.com
|
|
|

- - - -
Hosted by Home Power magazine

To send a message: RE-wrenches at topica.com

Archive of previous messages: http://lists.topica.com/lists/RE-wrenches/read

List rules & how to change your email address: www.mrsharkey.com/wrenches/etiquette.php

Check out participant bios: www.mrsharkey.com/wrenches/

Moderator: michael.welch at homepower.com
--^----------------------------------------------------------------
This email was sent to: michael.welch at re-wrenches.org

EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Qcs.bz9JC9.bWljaGFl
Or send an email to: RE-wrenches-unsubscribe at topica.com

For Topica's complete suite of email marketing solutions visit:
http://www.topica.com/?p=TEXFOOTER
--^----------------------------------------------------------------