Solar fire raises questions about panel safety

[W6SJK]

Should rooftop PV disconnects be provided?:

http://www.nctimes.com/business/article_8a32fb03-9e3f-58ca-b860-9c7fe1e28c7e.html

Steve

 

[e57]

I can see the issue - and I can see the contradiction...

Sue Kateley, the executive director of the California Solar Energy Industries Association, a trade group, said DC cutoffs create danger because they mislead firefighters.
The solar panels remain energized as long as the sun is up, and they even retain some electricity when it's dark. Putting a cutoff on the roof only neutralizes the wire running from the panels to the inverter box, but it doesn't eliminate the danger to firefighters.
"It's that perception of safety we want to avoid; they think they're safe, when there's potential from getting a shock, and falling off the roof," she said.

And from the comfort of my chair - well this flash light too... I can see a solution... Although a tree pruner may have been put artfully to effect in the situation. The real tool would be standardization of the solar industry, and yes - some standards for disconnecting means of multiple sets of series panels...

That said - I don't do solar... But I have examined a lot of installations up close, and it's like the wild west in the 1800's. Most are bizzare. Until recently in California at least - contractors installing solar required no licensing, (It's one of the youngest licenses) and for that matter no experience. Some AHJ had about as much experience with it as some of the contractors that may have started doing it the day before. Most of them former roofing and pool contractors.... And some of them (Contractors and Inspectors alike) didn't even see it as "Electrical" until after the inverter.

 

[ty]

I would question the actual installation of this system.
This type of fire is rare, but can happen.

And, if there are more than two source circuits, there must be fuses on the roof for each string.

 

[DownRiverGUy]

Rather than requires a disconnect why not show fire fighters the faster way to cut power to a PV panel.... put a tarp over them.

Cover panels.... disconnect AC main... no power coming in

Just a thought rather than requiring a DC disconnect that would still not truely kill the power coming from the panels.

Thoughts?

 

[SeanD]

It's kind of ironic, I was just talking to Sue about the proposed changes in 2011 for DC arc-fault protection and I pretty much told her the same thing I wrote below.

This is definitely one of the areas of the solar industry where I hope to see changes happen fairly quickly. I can see how a DC disconnect on the roof might seem like a good solution but from where I sit it is far from a good one. Regardless of the how likely or often it happens, Safety is the #1 priority. Sue mentions one reason the DC disconnect isn't a great idea.

I have another. It is practically difficult to do, specifically the way some AHJ are requesting it be done. The few places I have talk to about requiring this want a handled disconnect, which as far as I am aware are required by there listing to be installed vertically. Vertical surfaces on residential roofs tend to be hard to find.

Right now I really only have one attractive solution if a AHJ requires this. Micro-Inverters. But I hoping in the next year or 2 the solar and electrical equipment manufacturers will have a come up with a less expensive option.

Regulation in some way is coming regarding this and it is always better to be a head of the game than playing catch up.

 

[SeanD]

Rather than requires a disconnect why not show fire fighters the faster way to cut power to a PV panel.... put a tarp over them.

Cover panels.... disconnect AC main... no power coming in

Just a thought rather than requiring a DC disconnect that would still not truely kill the power coming from the panels.

Thoughts?

How many tarps do you expect them to carry?

Some roof top commercial systems can be pretty big. And asking the fire department to identify which array the fault is in is probably asking too much.

Residential systems vary. It maybe easy to throw a trap over the whole array or there could be multiply array's on a number of roof surfaces. On large high end houses, you could have an obvious array on the front of the house. Then another array on a section not obviously visible.

In my experience in talking to Fire Chiefs regarding safety and solar. They generally don't want to leave safety up to a firefighters critical thinking skills. In that first adrenalin filled moment when they show up the Chief wants them doing what they are trained to do.

I think the simplest solution would be some kind of disconnect in the module.

 

[SeanD]

I would question the actual installation of this system.
This type of fire is rare, but can happen.

And, if there are more than two source circuits, there must be fuses on the roof for each string.

As would I. My thoughts in looking at the photo: What a mess. Where was the company that installed the system? We have really only had one emergency call but we were there ASAP. (Corroded meter socket, which failed when the utility changed the meter)

What's the code section on the 2 source circuits? I wanted to re-read it again but I don't have my code book handy.

 

[DownRiverGUy]

Valid point on commercial systems... I was thinking of residential.

Granted most large commercial installations have DC combiner boxes with disconnected on the roof with the panels already.

Besides a micro-inverter system I doubt that we will see PV panels with disconnected built into the panel.

 

[Sierrasparky]

Valid point on commercial systems... I was thinking of residential.

Granted most large commercial installations have DC combiner boxes with disconnected on the roof with the panels already.

Besides a micro-inverter system I doubt that we will see PV panels with disconnected built into the panel.

Personally I would be wondering why these systems are catching fire not so much as to installing some cutoff switch.
I feel the wireing from the panel to the first junction box is not adequate, either by insulation or size. The connectors seem so cheap to me.

In an analogy about 25 years ago I customized a sports car. I added electric fans and removed the belt driven fan. I kept buring out the wiing from the fan relay to the battery terminal block. I used the parts that came with the kit. Several years later that same company came out with some beefy relay and harness. Never had a problem since.

Build it undersized and that what happens. The solar wiring needs to be appropriate for High temps in summer and cold in winter. It needs to withstand the UV. It needs to last probably more than 40 years.
There is a reason that California mandated that the warranty for all PV hardware be for at least 10 years. THe industry standard was only 5. Stuff failed.

 

[DownRiverGUy]

And I 100% agree. I would like to see them require conduit to conduit connections between panels and allow the electrcian to provide the wire connection on a manufactured provided J-box on the panel itself.

The cables/connections on panels can be... touchy.... the new MC connectors are better but still I'd feel more comfortable with screw terminal connections on the panels themselves.

 

[e57]

The connectors seem so cheap to me.

Hook stick operated plug at each set of panels... (That was my first idea - short of a tree pruner... )

Some roof top commercial systems can be pretty big. And asking the fire department to identify which array the fault is in is probably asking too much.

Residential systems vary. It maybe easy to throw a trap over the whole array or there could be multiply array's on a number of roof surfaces. On large high end houses, you could have an obvious array on the front of the house. Then another array on a section not obviously visible.

In my experience in talking to Fire Chiefs regarding safety and solar. They generally don't want to leave safety up to a firefighters critical thinking skills. In that first adrenalin filled moment when they show up the Chief wants them doing what they are trained to do.

I think the simplest solution would be some kind of disconnect in the module.

You're right! The fire Dept should not have to be trained in figuring out the fault - they should also not have to be subjected to a source of energy that lacks a disconnect or means of de-energizing. Nor should the owner of said system. It should be installed with a means if not to completely de-energize - isolate. Think big red button - or handle - 'break glass in case of emergency' etc.

If solar were held to the basic standard for other sources - it would need to be less than six throws of the hand... At least one of those being the AC utility.

If a UPS system is any example - many of those are available with easy to identify - or even automated means to de-energize - not just the inverter - but the batteries feeding it as well. The dreaded EPO switch... :roll:

If worse comes to worse - to offer sequential and safe shut-down - cont actors to drop out isolating the panels. From what little I understand of solar - many of these can be wired in series to get rather high voltages of DC, then paralleled to obtain the necessary amperage capacity. IMO since shutting off the sun is not feasible - there should be a means in which to break these series and paralleled connections to to limit voltage and capacity to a level acceptable by the fire official, to render the system safe - or safe enough - and make it a standard. And either accessible at the roof - or better yet - from the ground right next to the utility service disconnect.

IMO - it could be easily done by simple design of a branch series connected and paralleling switch...

 

[Sierrasparky]

I like the hook style disconnect. Sounds very practical.
Does not need to part of the panel just a part of the hook-up wiring.
I recently saw a install where the Solar installer just laid a nema 3r style junction box on it's back right on the roof. The wires leading through the roof and to the panel below went through the back of the 3R box. The slope was at least 4/12. I don't think that type pf box is rated for that position. The conductors cam right through the roof no conduit at all just romex. I think it was UF type though

 

[e57]

I like the hook style disconnect. Sounds very practical.
Does not need to part of the panel just a part of the hook-up wiring.
I recently saw a install where the Solar installer just laid a nema 3r style junction box on it's back right on the roof. The wires leading through the roof and to the panel below went through the back of the 3R box. The slope was at least 4/12. I don't think that type pf box is rated for that position. The conductors cam right through the roof no conduit at all just romex. I think it was UF type though

A prime example of standards & methods....

I saw one with THHN under CATV clamps down to the invertor - right down the wall on the surface. I couldn't see the rest of it - not sure I wanted to.

 

[Bigrig]

Not having installed any of these systems myself, how would you provide disconnects for the fancy shingle-style solar panels? Even covering them may not be possible as it may not be immediately obvious which section of roof contains them.

 

[e57]

Not having installed any of these systems myself, how would you provide disconnects for the fancy shingle-style solar panels? Even covering them may not be possible as it may not be immediately obvious which section of roof contains them.

Me neither - but one would assume a multi-pole disco at what the solar industry likes to call a "combiner box". Many of which look DYI - or made in someones back yard manufacturing and distrubution center...

(Ref: http://www.solarquest.com/squire/installsolar.htm)

Most are better looking than this - but you get the idea... Looks like an accident waiting to happen to me. Better put on your PPE or welding gear if you plan to touch the upper bus with a screw driver in that box.

In the process of finding some of the finer examples of one - I found an interesting article on the topic...

 

[big john]

What about having a standardized load-break elbow on each conductor located at the solar panels? I've never seen a solar panel install, but I imagine we're only talking two conductors unless they're really massive commercial systems.

Firemen carry hotsticks. This isn't medium-voltage. They can yank each elbow from a safe distance, totally isolating the solar panels from the rest of the electrical system.

-John

 

[ty]

As would I. My thoughts in looking at the photo: What a mess. Where was the company that installed the system? We have really only had one emergency call but we were there ASAP. (Corroded meter socket, which failed when the utility changed the meter)

What's the code section on the 2 source circuits? I wanted to re-read it again but I don't have my code book handy.

nicely worded, but I don't think it's as simple as "article xxx.xxx says two or more source circuits"

It would start with 690.9 and go to 240.15, i believe, going off of memory.
There is an exception in 690.9 about the short circuit currents. I believe the exception applies with 2 or less source circuits, basically by process of elimination. After that, I think the exception wouldn't apply.

xxx


As far as covering the array that was mentioned, 690.18 gives 3 possible solutions.

 

[ty]

What about having a standardized load-break elbow on each conductor located at the solar panels? I've never seen a solar panel install, but I imagine we're only talking two conductors unless they're really massive commercial systems.

Firemen carry hotsticks. This isn't medium-voltage. They can yank each elbow from a safe distance, totally isolating the solar panels from the rest of the electrical system.

-John

I don't see that as very plausable in the installation.
And these voltages can typically be up to 600v DC.
Overcurrent Devices Must be rated for DC.

 

[SeanD]

Looking over some of the posts in this thread, I'd like to point out something. Just like there are hack electricians, there are certainly hack solar installers. (They seem to be coming out of the wood work these days).

The solar industry in the last 5 years in CA has changed pretty radically. Many of the items people mentioned have been addressed by code sections, knowledgeable inspectors and manufacturers updating there products . (I.e. Modules use USE-2 wire which is Sunlight resistant, Conduit is required where exposed to physical damage).

Additionally legit solar installers just like electricians understand that a product must not only be listed but must be used as listed. Homemade combiners are hack or homeowner installs.

I was talking with my boss about the particular installation mentioned in the article. It seems like a listed flush roof mounted fused DC combiner would be the simplest solution. It would protect against DC faults inside a structure.

While this would not provide a Firefighter the ability to "Disconnect" the system, an array could "disconnect" itself if compromised. As far as I am aware this is not currently required by any NEC code section.

 

[SeanD]

nicely worded, but I don't think it's as simple as "article xxx.xxx says two or more source circuits"

It would start with 690.9 and go to 240.15, i believe, going off of memory.
There is an exception in 690.9 about the short circuit currents. I believe the exception applies with 2 or less source circuits, basically by process of elimination. After that, I think the exception wouldn't apply.

xxx

As far as covering the array that was mentioned, 690.18 gives 3 possible solutions.

690.9 covers PV source circuits, while I just scanned through it. To paraphrase, Protect the wire, modules and inverter terminals from DC short circuit currents (It can be done by design as well as overcurrent protection) I am pretty sure it doesn't mention requiring fuses (or overcurrent protection) at roof level.

240.15
Is for 2008. It clearly states fuses or over current protection is required for ungrounded conductors. We are currently using the 2005 NEC in CA. But I am certainly making a list of changes for the adoption of 2008 so this is good to know. It does not however mention it being required at roof level.

690.18
Just says cover modules when installing or servicing. I am assuming you mean the Staff notes when you say "3 possible solutions". There are probably more, staff notes are not code requirements.