Rapid shutdown options

[jaggedben]

I was in a session about this (along with other things) in the NABCEP conference a couple of days ago. The presenter talked about all the changes to the RS rules over the past few code cycles, and that he had discussed it with many firefighters. I asked the question of when firefighters approach a burning building that has a rooftop PV system, what will they assume is the situation? What code cycle was the PV array built under, which (if any) RS protection does it have, and if it does, is it working like it is supposed to? Or would they just assume the worst case scenario, that there is no RSD installed and functional. He answered that they would probably assume the worst.

Alternatively, they won't even notice there's a PV system on the building and will saw right through a live DC conduit. I say this only because a fire inspector once told me that he'd watched one of his firefighter colleagues do exactly that. (I was humoring him while trying to convince him our signage for a project was adequate. He started telling me stories.)

I all seriousness, I think your presenter is probably right.

I especially chuckle at the notion that firefighters will comprehend the difference between the red and yellow labels in the 2017 NEC.

 

[pv_n00b]

One of the problems that triggered RSD was firefighting districts in a few cases had taken a stance that fires in buildings with PV would be contained but not fought, and that thinking was expanding. This caused problems with insurance companies who started to question if they would insure buildings with PV and how much it would cost. This was a big wake-up call for the PV industry. It did not get much news exposure, so not many people in the PV industry caught on to what was happening but the people involved in codes and standards knew something had to be done quick. That led to the rather poorly written RSD in the 2014 NEC. The purpose of that was to show the firefighters and insurance companies that the industry acknowledged that something had to be done and they were making progress towards that end.
The 2014 RSD is pretty much useless and was just a placeholder. The 2017 RSD took shape as a more solid solution, but it required a disconnect at each module which few in the PV industry liked. Now with the UL 3741 listing, we can have reduced hazard arrays without module-level disconnection. Hopefully, this is good enough that firefighting districts will not take a no-fight stance and insurance companies will continue to insure buildings with PV without imposing higher rates.

 

[ggunn]

One of the problems that triggered RSD was firefighting districts in a few cases had taken a stance that fires in buildings with PV would be contained but not fought, and that thinking was expanding. This caused problems with insurance companies who started to question if they would insure buildings with PV and how much it would cost. This was a big wake-up call for the PV industry. It did not get much news exposure, so not many people in the PV industry caught on to what was happening but the people involved in codes and standards knew something had to be done quick. That led to the rather poorly written RSD in the 2014 NEC. The purpose of that was to show the firefighters and insurance companies that the industry acknowledged that something had to be done and they were making progress towards that end.
The 2014 RSD is pretty much useless and was just a placeholder. The 2017 RSD took shape as a more solid solution, but it required a disconnect at each module which few in the PV industry liked. Now with the UL 3741 listing, we can have reduced hazard arrays without module-level disconnection. Hopefully, this is good enough that firefighting districts will not take a no-fight stance and insurance companies will continue to insure buildings with PV without imposing higher rates.

I confess that I still do not know a whole lot about the UL3741 listing, but it was talked about in a few of the presentations I saw at the NABCEP CE conference last week. I didn't get a clear impression of exactly what in a PV system has to be listed to qualify. The only data point I have is that it is apparently how the SolarEdge P1101 can have up to 125VDC between conductors for two modules in series in shutdown conditions.

 

[BackCountry]

I confess that I still do not know a whole lot about the UL3741 listing, but it was talked about in a few of the presentations I saw at the NABCEP CE conference last week. I didn't get a clear impression of exactly what in a PV system has to be listed to qualify. The only data point I have is that it is apparently how the SolarEdge P1101 can have up to 125VDC between conductors for two modules in series in shutdown conditions.

The whole thing is very much so in flux, what I’ve taken away from it is this:

From a firefighting perspective, large commercial flat roof buildings are challenging to ventilate to begin with, you add a gigantic solar array on the top and it becomes very challenging because vertical ventilation requires locating the framing structure by feel and visualization using a chainsaw in unfavorable conditions. Add the complexity of a bunch of solar panels, wire, and conduit, and you may not even attempt the ventilation tactic. Too much risk. Further, to assert that the fire department is going to be knowledgeable enough at 2AM to locate a RSD switch or fused disconnect is probably far fetched also, considering the competing priorities of a structure fire.

That’s why I think MLPE’s are a waste of time, money, and add a lot of unnecessary complexity. I buy it at the residential level where the PV footprint is relatively small; however, on a flat roof installation in the commercial sector we almost always cover as much as is physically possible to achieve the highest offset.

This is why the UL3741 standard to me, makes sense.

I agree that a listed zip tie is ridiculous compared to a PV wire clip.

For very small commercial (sub 25kw) we stick with micro’s — it’s just easier.

For greater than 25kw, it’s UL3741 all day.

 

[pv_n00b]

I confess that I still do not know a whole lot about the UL3741 listing, but it was talked about in a few of the presentations I saw at the NABCEP CE conference last week. I didn't get a clear impression of exactly what in a PV system has to be listed to qualify. The only data point I have is that it is apparently how the SolarEdge P1101 can have up to 125VDC between conductors for two modules in series in shutdown conditions.

The basic premise of 3741 is to prevent a fully kitted-out firefighter from being exposed to a current through their body more than a maximum value. That's the goal.
Getting to that goal is a complex process. The standard is not the usual list of tests that the array needs to pass. It's a process where the likelihood of various firefighter interactions with the array are evaluated. Then the likelihood that an interaction will expose the firefighter to a voltage that will induce a current through their body that is greater than the maximum allowed is evaluated. These are combined and ranked and scored based on the likelihood of occurrence and danger to the firefighter. An interaction that has a high score has to be mitigated in the design and installation of the array. Careful wire management is high on the list of things that will lower a score.
Because it's impossible to know the path that was taken through the listing to get to the end it's not really possible to know exactly how a particular array is listed. It's very likely that the SE P1101 DC Optimizer was listed because a fully kitted-out firefighter coming into contact with 125V would not result in a dangerous current flow in their body, so the listing was pretty quick for them and array agnostic.

 

[pv_n00b]

The whole thing is very much so in flux, what I’ve taken away from it is this:

From a firefighting perspective, large commercial flat roof buildings are challenging to ventilate to begin with, you add a gigantic solar array on the top and it becomes very challenging because vertical ventilation requires locating the framing structure by feel and visualization using a chainsaw in unfavorable conditions. Add the complexity of a bunch of solar panels, wire, and conduit, and you may not even attempt the ventilation tactic. Too much risk. Further, to assert that the fire department is going to be knowledgeable enough at 2AM to locate a RSD switch or fused disconnect is probably far fetched also, considering the competing priorities of a structure fire.

While not all firefighters will agree, the codes around array layout and RSD have significantly increased firefighter safety and allowed for building ventilation options. Firefighters now have a seat on relevant code-making panels and their input carries a lot of weight. Not many people are going to go up against a firefighter face to face and tell them it's okay if they are less safe if it makes a contractor more money. So now we have roof layouts with built-in areas for ventilation that are AHJ-approved and reduced electrical hazards in and around the array. Firefighters don't have to pull up modules to ventilate a building anymore because a large roof is covered end to end. There is significant training being given to firefighters on how to work with PV systems. Firefighters are smart folks and have to deal with a lot of hazardous sites. So yes, at 2 AM a site commander and their firefighters are going to know where the RSD switch is for a commercial building.
If you talk to the firefighters who sit on the panels you will find that they are fairly happy with the direction things are going since 2014.

 

[BackCountry]

While not all firefighters will agree, the codes around array layout and RSD have significantly increased firefighter safety and allowed for building ventilation options. Firefighters now have a seat on relevant code-making panels and their input carries a lot of weight. Not many people are going to go up against a firefighter face to face and tell them it's okay if they are less safe if it makes a contractor more money. So now we have roof layouts with built-in areas for ventilation that are AHJ-approved and reduced electrical hazards in and around the array. Firefighters don't have to pull up modules to ventilate a building anymore because a large roof is covered end to end. There is significant training being given to firefighters on how to work with PV systems. Firefighters are smart folks and have to deal with a lot of hazardous sites. So yes, at 2 AM a site commander and their firefighters are going to know where the RSD switch is for a commercial building.
If you talk to the firefighters who sit on the panels you will find that they are fairly happy with the direction things are going since 2014.

Without saying a lot more detail: I’m one of those firefighters, with 17 years on the job.

Not a single member in my battalion, beyond my efforts to educate, knows a thing about rapid shutdown. Clearly I’m not speaking for everyone in every jurisdiction — I’m just giving my perspective having worked on a truck company for 5 years. I’m a licensed electrical contractor in my off time, which has significantly improved my mental health.

 

[ggunn]

Because it's impossible to know the path that was taken through the listing to get to the end it's not really possible to know exactly how a particular array is listed. It's very likely that the SE P1101 DC Optimizer was listed because a fully kitted-out firefighter coming into contact with 125V would not result in a dangerous current flow in their body, so the listing was pretty quick for them and array agnostic.

If that is true, why isn't the threshold voltage 125V instead of 80V?

 

[jaggedben]

If that is true, why isn't the threshold voltage 125V instead of 80V?

Because, as I believe he already said, 80V was thought at the time to be the max Voc for any single PV module, and therefore a practical way to set a limit. Not because it had any objective saftey meaning.

 

[ggunn]

Because, as I believe he already said, 80V was thought at the time to be the max Voc for any single PV module, and therefore a practical way to set a limit. Not because it had any objective saftey meaning.

Well, OK, but doesn't that call into question whether that 80V number ever had any significance whatsoever? If the objective is safety, then the maximum voltage between conductors should be set by that objective safety meaning. If 125V is safe, why should anyone be held to an 80V maximum?

 

[jaggedben]

The significance was that it was (they thought) the lowest they could go without a draconian effect on how modules are manufactured.

I think they would have put it: why should we allow 125V without further study when you can install systems with exisiting equipment at 80V or less?

 

[pv_n00b]

Well, OK, but doesn't that call into question whether that 80V number ever had any significance whatsoever? If the objective is safety, then the maximum voltage between conductors should be set by that objective safety meaning. If 125V is safe, why should anyone be held to an 80V maximum?

You are not alone in questioning the 80V number. There had to be a number chosen, as I said before doing nothing was not an option. RSD was going to happen, and that's just the way it is. So a number was chosen that was as low as reasonably possible without shutting down the whole PV industry by requiring strings of cells internal to the module to be opened to reduce the Voc.
Even choosing 125V would not materially change the design, IMO. MLPE would still be required, maybe someone would come out with a product where one MLPE could service two modules if each module Voc was equal to or less than 62.5V.

 

[pv_n00b]

Without saying a lot more detail: I’m one of those firefighters, with 17 years on the job.

Not a single member in my battalion, beyond my efforts to educate, knows a thing about rapid shutdown. Clearly I’m not speaking for everyone in every jurisdiction — I’m just giving my perspective having worked on a truck company for 5 years. I’m a licensed electrical contractor in my off time, which has significantly improved my mental health.

All I can say is the training is out there and available to all firefighters. Choosing not to take it is an option I guess but it's an active choice not to learn how RSD works. Is it really a good idea to not have increased safety because some firefighters will choose not to learn about it? Probably not.

 

[ggunn]

Even choosing 125V would not materially change the design, IMO. MLPE would still be required, maybe someone would come out with a product where one MLPE could service two modules if each module Voc was equal to or less than 62.5V.

Someone did; it's the SolarEdge P1101.

 

[pv_n00b]

Someone did; it's the SolarEdge P1101.

Yup, and they did it before 3741 because it halved the number of MLPEs needed and the higher voltage had other advantages. There are other MLPE that put two modules in parallel which also cuts the number of MLPE in half, doesn't require listing to 3741, and provides RSD. But MLPE are still needed. Once you put in any MLPE you are just on that path.

 

[BackCountry]

All I can say is the training is out there and available to all firefighters. Choosing not to take it is an option I guess but it's an active choice not to learn how RSD works. Is it really a good idea to not have increased safety because some firefighters will choose not to learn about it? Probably not.

I think this is a classic case of trying to create a solution for a problem that doesn’t necessarily exist. The reason most firefighters don’t have a whole lot of training or desire to look at a placard riveted to a main service panel is because they simply don’t have time. When you secure utilities, you shut off the gas and main breaker. We’re actually trained to leave all breakers in a panel on, aside from the main breaker, to investigate if anything was tripped later once the fire is in the investigation stage.

I’m not saying that we shouldn’t be educated on how to throw a disconnect switch, or a rapid shutdown switch. What I’m saying is: code professionals sometimes implement what they see as a solution and yet it doesn’t actually work out in the field.

How many times have you had an inspector just glance at all of your safety labeling and not care at all? Happens to me almost every time.

I think my overall point here is that most of the time, a main disconnect is going to get thrown. If a truck company sees a solar array, I can almost guarantee you they won’t get near it. They’ll go where there isn’t any PV, which in residential is fairly easy. In the flat roof commercial space, a heat hole has to be over the fire… and large array blocks will prevent that, thus you switch tactics.

I am in agreement that safety is the top priority, where I disagree is that rapid shutdown at the module level is a priority.

 

[jaggedben]

... How many times have you had an inspector just glance at all of your safety labeling and not care at all? Happens to me almost every time.

I would say the order of prevalent inspector behavior at my inspections, from most prevalent, is:

- glance at the labeling and approve like you say
- check the plans and ask me for any missing labels, nitpick similar but not indentical language, insist labels match plans even when plans clearly contain an error or typo
- ask for additional labels that are blatantly incorrect and misleading
- carefully observe the applied labels in context and give appropriate approvals or corrections

I will just add that this applies mainly to building inspectors on residential projects. Fire inspectors on both residential and commercial, in my much more limited experience, tend to be less concerned with the code and more with applying common sense to the labeling. However most inspectors of both types have no clue about interactive inverters and are overly concerned about load side PV AC disconnecting means when there's no backup source.

All that is to say, I can understand why firefighters may not have the most trust in labeling for rapid shutdown.

 

[BackCountry]

I would say the order of prevalent inspector behavior at my inspections, from most prevalent, is:

- glance at the labeling and approve like you say
- check the plans and ask me for any missing labels, nitpick similar but not indentical language, insist labels match plans even when plans clearly contain an error or typo
- ask for additional labels that are blatantly incorrect and misleading
- carefully observe the applied labels in context and give appropriate approvals or corrections

I will just add that this applies mainly to building inspectors on residential projects. Fire inspectors on both residential and commercial, in my much more limited experience, tend to be less concerned with the code and more with applying common sense to the labeling. However most inspectors of both types have no clue about interactive inverters and are overly concerned about load side PV AC disconnecting means when there's no backup source.

All that is to say, I can understand why firefighters may not have the most trust in labeling for rapid shutdown.

I’ve had a similar experience. On commercial jobs we usually get a separate fire inspection for labeling and clearance. They, like you say, just want to see a lot of labels — not even necessarily the correct ones, just a lot of red. And now yellow. The labeling is there for a point: to direct you to what you need to throw to mitigate the hazard, ie. a safety switch or fused disconnect.

Going back to the OP’s question — I personally don’t believe MLPE level shutdown makes sense in the commercial space, and that’s why I am utilizing UL3741 in every large project we have going. In the residential space, I think that’s fine and very simple to implement.

I do not think 500+ MLPE’s on a 500kw project are going to yield anything tangibly different than UL3741 compliant installs where the inverter is 1ft from the array block. I do think they’ll create a lot more failure points, cost, labor, and so on.

 

[ggunn]

Going back to the OP’s question — I personally don’t believe MLPE level shutdown makes sense in the commercial space, and that’s why I am utilizing UL3741 in every large project we have going. In the residential space, I think that’s fine and very simple to implement.

How does one do that?

 

[BackCountry]

How does one do that?

We’re designing all of our large (large to use is 50kw+, we’re a small company) using UL3741 listed assemblies (ie. PanelClaw or Unirac) without MLPE’s using either SMA or Chint inverters. We haven’t had any kickback yet from an AHJ using that method and making sure to highlight 690.12(B)(2)(1) because most plan checkers haven’t ran across it yet.

For real cost savings numbers, a 600kw rooftop project was going to cost $39,000 for MLPE’s. Add on top of that labor and logistics, which would be hard to quantify, I’d guess at least a $45,000 cost in the end.