[RE-wrenches] Target Fire

[Kent Osterberg]

William,

Thanks for posting the info on the Target fire.  We all need to learn 
some lessons from it.  I'm sure that it is going to lead to changes in 
our industry.

Some things that the report brings to my mind:

There will be fire inspectors, electrical inspectors, all kinds of other 
experts, and lawyers too.  Be aware that an investigation into a problem 
will reveal any imperfect aspect of the contractor's work.  Any 
imperfection, related to the cause of the problem, or not related, will 
be used by the lawyers try to show incompetence or negligence by the 
contractor.  Examples in this report:
1) 4 AWG wire in lugs rated down to size 2 AWG.  Probably of no 
consequence, but prominently visible in this report.
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.  

Workmanship:
Freedom of the conduit associated with the expansion joint to move over 
its full length and possibly loose or loosened conduit fittings are 
workmanship issues the lawyers will have a heyday with.  Those may well 
be the most direct cause of this fire.

Faults are going to happen:
Even if the lawyers pin the cause of this fire on the contractor, 
electrical faults can and will occur even if there aren't any 
workmanship issues.  The insulation of a windblown, machine vibrated, or 
earthquake shaken wire can abrade and then fault.  Age, temperature, UV 
light, and so forth will degrade insulation over time too.

There has been a lot of discussion about meggars lately.  Meggaring only 
proves the wires aren't faulted when you meggar them. It is slight 
evidence of good workmanship but no protection for when a fault 
eventually does occur.

Listing Issues:
Are PV modules as currently built, adequately fire resistant? 
Are their exposed wires an accident waiting to happen?

Code Issues:
De-rating conductor ampacity for hot roof-top areas is now part of the 
2008 NEC.  Better be on top of that in your (and my) system 
documentation.  Looks like temperature de-rating wasn't required  for 
the version of the code in force in California at the time of 
installation.  It probably will be soon because this fire has brought it 
to everyone's attention; even though it may not have been a direct cause 
of this fire.

Roof top disconnects won't de-energize a PV array but they sure would 
have helped open faulted circuits while the fire was happening.  The 
ability to de-energize the PV combiner output circuit makes sense 
because the PV combiner output circuits are poorly protected with 
existing circuit protection technology.  I predict that this will soon 
be a code requirement.  Now, we don't put put disconnects at the array 
because it doesn't make the array any safer to service.  In the future, 
will be installing them so they can be used during fires or other accidents.

Did you notice that the report says the ground fault fuse in inverter A 
did open? It probably opened after the fault turned into a fireball with 
a current of hundreds amps.  Look at Exhibit N.  Cause related or not 
cause related to this fire, Exhibit N exposes a serious limitation in 
the capability of the currently used ground fault protection scheme.  
And it's a well know limitation - if the PV system has more than one 
ground, the ground fault system sensitivity is seriously reduced or 
totally ineffective.  Many manufacturers warn in their manuals that the 
PV system must only be grounded in one location - the inverter's GFP.  
In this case, the first fault could have been a fault between the 
grounded conductor and ground which essentially deactivated the the GFP. 
The GFP doesn't detect the small current that flows though the ground 
when a grounded conductor is grounded again by way of a fault.  The GFP 
isn't even able to provide a warning that there is a potential hazard.  
The system is just waiting for the second fault to happen.  When the 
ungrounded conductor faulted, the current didn't have to go through the 
GFP. When the GFP did open, it accomplished nearly nothing; current was 
flowing back to the source, back to the PV array, through the first 
fault.  The double fault behaves just like a line to line fault, there 
is no stopping the current except to disconnect at the source.  That GFP 
doesn't really provide very much protection afterall, does it?

"Know your job; then do the best job you know how to do." - Unknown

Kent Osterberg
Blue Mountain Solar, Inc.
70177 Ponderosa Lane
Cove, Oregon  97824