[RE-wrenches] Target Fire

[Kent Osterberg]

Bob-O,

Opening the ungrounded circuits at (near) the source is the only thing 
that is going to provide real protection.  Moving the GFP to the 
combiner would help in that sense.  But the current sensing scheme would 
still be dependent on there being a single ground connection.  An 
undetected ground fault in a grounded wire would still de-sensitize the 
GFP.  There could be hundreds of amps flowing into a fault before you 
got; you pick the number 1, 2, 5 amps; flowing through the GFP.  There 
could be a lot of fire happening before the GFP would trip.

This incident is making me think about ungrounded arrays; at least there 
could be an alarm (for someone to ignore) when any ground fault occurs.  
Or maybe I'll be a micro-inverter convert.

Do you have any 600-volt dc-rated arc-fault sensing breakers?  We need 
some, 15-amp ones should do for most strings.

Kent Osterberg
Blue Mountain Solar



Bob-O Schultze wrote:
> Kent,
> Sort of makes the case for the GFPs to be in the combiner boxes on 
> large systems, doesn't it? Obviously nearly worthless at the inverter 
> in this case.
> Bob-O
>
>
> On May 6, 2009, at 3:52 PM, Kent Osterberg wrote:
>
> Richard,
>
> I 100% agree with you.
>
> The thing that concerns me most is that the last exhibit in the report 
> shows that the GFP scheme that our industry has relied on to provide 
> fire protection doesn't work very well.  As PV system size gets larger 
> there are more and more opportunities for an undetected fault on a 
> grounded conductor that disables the GFP.  Then a second fault, on an 
> ungrounded conductor, isn't cleared.  There will be more fires.  And 
> they can't lay all the blame on the contractor. 
>
> Kent Osterberg
> Blue Mountain Solar
>
>
> Richard L Ratico wrote:
>> Kent, 
>>
>> I'm trying not to lose sight of the forest for the trees here. I agree that the
>> expansion joint is probably completely functional. It also appears to be a well
>> intentioned effort by the installer to anticipate and compensate for the
>> inevitable movement in the conduit run. The installer took the initiative, after
>> all, since the design did not call for it.
>>
>> My point is, even had the installer used the correct EMT type fitting,
>> functional or not, the fault would still have happened. With the EMT restrained
>> by overly tight straps, it would still have pulled apart at the loose coupling,
>> the nearest, weakest link.
>>
>> But, had the installer followed through, using the appropriate RMC to match the
>> choice of expansion coupling, the fault may well have been prevented. Even when
>> very loose, threaded fittings are unlikely to pull apart completely. Of course,
>> using 423' of 3" RMC instead of EMT, would have resulted in a more expensive
>> job, though as a percentage of the overall cost, maybe it wouldn't have amounted
>> to much.
>>
>> IMHO, Monday morning quarterbacking, EMT was a poor choice for this system,
>> subject to nearly 4" of movement. On a job of this scale, there are simply too
>> many opportunities for a loose coupling to occur, whether initially on
>> installation, or over time after repeated thermal cycling. S... happens. Kind of
>> analogous to why we pull an equipment grounding conductor in metal conduit
>> systems, even though the conduit itself provides an equipment ground. RMC would
>> provide for much better sleeping at night. PVC, nasty as it is in some respects,
>> in this case, would have been significantly less likely to start a fire. It's
>> harder to forget to glue a joint than to overlook wrenching an EMT coupling. If
>> one does pull apart, since PVC is nonconductive, it's less likely to cause a
>> fault. Better yet, as the report recommends, is to design a means to clear the
>> fault, should one occur.
>>
>> In a perfect world, EMT might be fine.
>>
>> Dick
>>
>>
>> Richard,
>>
>> I think you are right, the weakest link in the chain was a compression fitting. 
>> Maybe one that was not properly tightened.   An EMT connector will thread right
>> into a RMC coupling and it doesn't look like it would interfere with the
>> expansion joint motion, so I still think the expansion joint may be functional. 
>> It should slip with modest hand pressure.  Wouldn't the installer notice if an
>> expansion joint wouldn't work?  And then not use it.  it just wouldn't make any
>> sense to install it in if it didn't work.
>>
>> Kent
>>
>>
>> Richard L Ratico wrote: 
>>
>>
>> Kent,
>>
>> I looked at the photo of that expansion coupling and initially thought, "That'll
>> work". But....  if the conduit was RMC OR IMC all the joints would have been
>> THREADED into couplings, and even if by mistake they were not wrench tight, they
>> would not have pulled apart. Even if strapped too tightly, the roof blocks would
>> probably pull over first. With EMT, contraction pulled apart the weakest link in
>> the chain, a loose slip fitting.
>>
>> Thanks William.
>>
>> Dick
>> Solarwind Electric
>>
>> --- You wrote:
>> 2) The expansion joint fabricated from adapters for different conduit 
>> systems probably works fine and may not have been a contributing factor 
>> to this fire, but it is prominently visible in the report.  
>> --- end of quote --- 
>>