It is currently Mon Mar 08, 2021 5:15 pm



Post new topic Reply to topic Go to page 1, 2  Next
Author Message
ekstra   ara
 Post subject: Arc Flash Barrier
PostPosted: Tue Sep 25, 2012 7:03 am 
Sparks Level
User avatar

Joined: Tue Apr 17, 2012 8:19 am
Posts: 241
Location: Charlotte, NC
I am involved in a project where we are tying an inverter through a isolation transformer to the power grid. What I am trying to find is information for building an arc flash barrier on the AC connection to the 3000A main disconnect for the enclosure. I can not seem to find information on standards for material thickness and incident energy ratings. It is conceivable that the tie to the grid could remain live with the main CB off and the incident energy at the top of the breaker could exceed 75 cal/cm2. I need a way to partition off the top side of the CB to be able to service the enclosure.


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Tue Sep 25, 2012 12:44 pm 
Sparks Level
User avatar

Joined: Tue Apr 17, 2012 8:19 am
Posts: 241
Location: Charlotte, NC
The intent is to build an internal barrier inside the enclosure to accomplish this. I was trying to find the material/thickness requirement to achieve 75 or 100 cal/cm2.


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Tue Sep 25, 2012 1:18 pm 
Arc Level

Joined: Mon Jan 18, 2010 11:35 am
Posts: 525
Location: Wisconsin
It is much more complicated than you think.

Protecting against 75-100cal/cm² that occurs within .050 sec requires you to build something that is literally bomb proof.
However, protecting against 75-100cal/cm² that occurs within 500 sec, may not require much more than a heavy sunscreen.

True arc rated equipment includes channels and ducting as well as blow out vents.


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Tue Sep 25, 2012 1:56 pm 
Sparks Level
User avatar

Joined: Tue Apr 17, 2012 8:19 am
Posts: 241
Location: Charlotte, NC
JBD wrote:
It is much more complicated than you think.

Protecting against 75-100cal/cm² that occurs within .050 sec requires you to build something that is literally bomb proof.
However, protecting against 75-100cal/cm² that occurs within 500 sec, may not require much more than a heavy sunscreen.

True arc rated equipment includes channels and ducting as well as blow out vents.


It is probably about as complicated as I think it is. I had no delusions that it was a simple thing to begin with - NOTHING about Arc Flash that I have encountered has been simple thus far anyway.

I can only guess about tA at the moment - since the end customer has only told me what he calculated the cal/cm2 to be, and has not revealed anything regrding his calculations yet. But I feel pretty certain tA is likely > 0.02 S, but < 0.1 S. So, that narrows the range down a bit.


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Wed Sep 26, 2012 3:53 am 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
Easiest way to do it is to make it remote operated (electronic close and release) and remote racking. "Arc resistant" gear (which is only available for MV) uses proof testing to demonstrate that it works and uses a heavy front to fully contain the blast with chutes/blast doors to release it in another direction and ONLY protects the operator when the equipment is closed and not under service. Various "arc detectors" operate by reducing the time delay to as short as possible to the point where the limit is on the breaker opening time. There are also current limiting breakers/fuses but this only helps downstream. Also there are arc "terminators" (ABB, Square D, others) that trip "close" onto the bus and kill the arc in under 1 cycle while the main breaker trips out in 2-5 cycles later and if there's space can be retrofitted. These work by dead shorting the bus temporarily and killing the arc by providing a lower impedance path. Arc flash blankets WILL do what you suggest (absorb the arc flash and blast) BUT you have to set up and pay for a test to be run at a high voltage lab where you actually test the worst case scenario to perform proof testing for the intended protection scenario. Expensive but the only practical way to do it in manholes.

Recommendation is to go for vacuum breaker and reduce the instantaneous trip below the arcing fault current first and check to see if that gets you anywhere. Use a DC trip unit so you get 3-4 ms trip coil times. If that doesn't work, there are high speed "arc terminator" type devices that shunt close onto the bus and trip in <1 cycle and are fairly cheap to retrofit into an existing system. ABB, Square D, and a couple others are manufacturers. Otherwise, go down the path of remote racking/remote control and deal with the fact that you cannot work with the equipment while energized except remotely.


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Wed Sep 26, 2012 10:41 am 
Sparks Level
User avatar

Joined: Tue Apr 17, 2012 8:19 am
Posts: 241
Location: Charlotte, NC
PaulEngr wrote:
Easiest way to do it is to make it remote operated (electronic close and release) and remote racking. "Arc resistant" gear (which is only available for MV) . . . .


The biggest problem is that this is not a switchgear (there is nothing to be racked in/out), the breaker is already defined and installed and the current design (which can be remotely operated) is not easy to retrofit a different type of breaker, the AC connection and busbars are already built and retrofit - again - is not easily accomplished. And the best part is that the proposed arc flash barrier shroud may need to be installed in the field on one of the enclosures.

Which is why the path I was hoping to take involves partitioning off part of the enclosure or the internal AC connection, busbar to the circuit breaker, and top lugs of the breaker itself.

We have several alpha units that need this retrofit - I agree the best solution is to disconnect the grid from the isolation transformer if you want to work on the enclosure, but the customer insists that is not feasible due to interruption of power generation on other inverters connected to the same grid transformer.


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Wed Sep 26, 2012 11:01 am 
Plasma Level

Joined: Mon Jun 02, 2008 11:58 am
Posts: 1103
Location: Charlotte, NC
Here is another thread you may want to look at
http://www.arcflashforum.com/threads/1345/


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Thu Sep 27, 2012 4:00 am 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
Larry Stutts wrote:
Which is why the path I was hoping to take involves partitioning off part of the enclosure or the internal AC connection, busbar to the circuit breaker, and top lugs of the breaker itself.


From a practical point of view at 480 V, I have found that once you get up past around 600-800 A, no matter how fast you can trip, you reach a point of diminishing returns and it becomes almost impossible to reduce incident energy to acceptable levels. It can still sometimes be done but rarely. There is also a "magic" cutoff of around 600-800 A on a single vacuum interrupter. Above those values, you are going to be looking at fuses (such as S&C "smart fuses"), air breakers, or SF6 breakers. Joslyn does build parallel vacuum interruptors to double the range but that is always questionable as an approach.

So...your choices then become:
1. Design your "blast shield" and test it according to IEEE medium voltage rules for arc resistant gear. The standard doesn't support 480 V gear but the test works the same regardless of the voltage. For design purposes, use the existing arc blast pressure calculation. It's grossly conservative but it's all we have without test data. You will spend several tens of thousands at a testing lab.
2. Use portable arc flash blankets. Again, there's no design information so you spend thousands at a testing lab.
3. Use or install additional upstream equipment, whether it's a disconnect or an "arc terminator" type device, where the current levels are reasonable. It avoids the testing lab but the construction costs are not very different.
4. Complete gear replacement with something else which has already had the design changes made to it.

Do you see the theme here? If the goal is to completely avoid having to switch upstream, and you've exhausted all the cheap solutions (reduce trip time or increase distance), the available solutions are about the same price as complete gear replacement. Your "blast shield" idea from a practical design cost point of view is no different than any other because it will require performance testing which has a price tag about the same as 100% gear replacement.

The inherent problem is that somebody put in "cheap" very large and slow 1000+ A breakers years ago instead of spreading the loads out and keeping their overall energy density down to a reasonable value to save money. At that time the only concern was that it wouldn't fly apart magnetically during a bolted fault. There was no arcing fault data available for design purposes. This was acceptable design practice 20 years ago but in today's environment no longer considered good design practice. 20-30 years ago we also didn't hesitate to put in draw out gear and go in and rack it out annually for maintenance. In today's environment that would be unthinkable for a lot of reasons, so now the search begins to find more reliable equipment that needs racking out (if racking is the best way...) only 2 or 3 times in the life of the equipment.


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Thu Sep 27, 2012 5:51 am 
Sparks Level
User avatar

Joined: Tue Apr 17, 2012 8:19 am
Posts: 241
Location: Charlotte, NC
PaulEngr wrote:

. . . . Do you see the theme here? If the goal is to completely avoid having to switch upstream, and you've exhausted all the cheap solutions (reduce trip time or increase distance), the available solutions are about the same price as complete gear replacement. Your "blast shield" idea from a practical design cost point of view is no different than any other because it will require performance testing which has a price tag about the same as 100% gear replacement.

The inherent problem is that somebody put in "cheap" very large and slow 1000+ A breakers years ago instead of spreading the loads out and keeping their overall energy density down to a reasonable value to save money. At that time the only concern was that it wouldn't fly apart magnetically during a bolted fault. There was no arcing fault data available for design purposes. This was acceptable design practice 20 years ago but in today's environment no longer considered good design practice. 20-30 years ago we also didn't hesitate to put in draw out gear and go in and rack it out annually for maintenance. In today's environment that would be unthinkable for a lot of reasons, so now the search begins to find more reliable equipment that needs racking out (if racking is the best way...) only 2 or 3 times in the life of the equipment.


The 3000A breaker here is for a single 1.25MW inverter. The breaker is on the 350V side of an Isolation transformer. The problem is more related to the customer using a single upstream breaker on the high-voltage side of 5-8 Isolation Transformers that tie to the grid.

But yes, I do see your point. There is no practical way to get the Line-Side connection of the 3000A breaker below 40 cal/cm2. As far as I'm concerned the best option is no live work on the enclosure.

Perhaps the best way to achieve that would be to house the breaker in a separate enclosure altogether and just bring the Load-Side connections into the Inverter Enclosure. That is what I was trying to achieve with the blast shield - effectively an enclosure within an enclosure to isolate anything energized from the 1.25 MW inverter enclosure.


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Thu Sep 27, 2012 7:18 am 
Arc Level

Joined: Mon Jan 18, 2010 11:35 am
Posts: 525
Location: Wisconsin
Have you considered using a "virtual main" scheme, sensing a fault on the secondary and then tripping the primary breaker
(e.g. CTs mounted as close to the transformer secondary coils as possible or optical sensing devices in each drive compartment in conjunction with CTs)?


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Fri Sep 28, 2012 6:10 am 
Sparks Level
User avatar

Joined: Tue Apr 17, 2012 8:19 am
Posts: 241
Location: Charlotte, NC
JBD wrote:
Have you considered using a "virtual main" scheme, sensing a fault on the secondary and then tripping the primary breaker
(e.g. CTs mounted as close to the transformer secondary coils as possible or optical sensing devices in each drive compartment in conjunction with CTs)?


The breaker on the primary side of the transformer actually controls 5-8 transformers, and the customer does not want to trip that for a problem in a single inverter enclosure. The 3000A main breaker in the inverter enclosure trips for that.

The main problem is that it is not possible to service anything in the inverter enclosure because the Line-Side of the circuit breaker is still live and has an incident energy of 75 - 100 cal/cm2 which is why I am trying to incorporate an arc flash barrier for the live components - Essentially an enclosure within an enclosure to allow technicians access to the inverter enclosure interior.

The customer does not want the loss of revenue from shutting down 8 inverters when there may only be a problerm with one of them. In this case the only action possible may be to leave the one inverter that has a problem offline until the others can be shut down to allow access for maintenance.


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Mon Oct 01, 2012 7:07 am 

Joined: Sat May 30, 2009 6:21 pm
Posts: 19
i have to say that sometime what customer want and what he need are two thing and your the one that is suppose to make him realise that...
that's i nice phrase...but if you push it to much the customer will go somewhere else

i like the vitual main option...could be implemented only when work are being done on equipement

for the materiel...unless you want to do some intensive testing and design...i would say use the same thing as the switchgear is built of

if it need's to be isolated 1/4 inch GPO3 well bolted is not that bad

oki this is not scientificaly approve but it's a start, and if the primary side of the breaker is now inaccesible the chance of creating an arc are very low


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Mon Oct 01, 2012 7:15 am 
Arc Level

Joined: Wed Jun 04, 2008 9:17 am
Posts: 428
Location: Spartanburg, South Carolina
You could have the customer install a maintenance switch on the upstream breaker so that it will only trip for a fault that occurs on the 350V side while live maintenance is being performed. If this is unacceptable, then live maintenance is not practical.


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Mon Oct 01, 2012 9:30 am 
Sparks Level

Joined: Mon Jun 07, 2010 9:40 am
Posts: 119
WHen I first read the original post, I assumed he was talking about "what was needed to create a barrier" in the sense of a location that was originally exposed to the line side terminals of the protective device, and hence at the IE rating of the line side terminals.
Is that NOT what he was asking?

[And on the other hand, it is always interesting how the "it is not possible to do "X", changes when actual $$ (or whatever currency) numbers start to ramp up for the work around you have to do.]


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Mon Oct 01, 2012 9:57 am 
Sparks Level
User avatar

Joined: Tue Apr 17, 2012 8:19 am
Posts: 241
Location: Charlotte, NC
The upstream breaker is on the grid-side of an isolation transformer - actually 5 to 8 transformers ganged together. The low-voltage connection is connected to a 1.5 MW step-up (or step-down depending on how you look at it). The low-voltage connection there goes to the 3000A Main Circuit breaker in the inverter enclosure.

I am interested in a physical barrier that encloses from the 350 Volt AC connection to the enclosure to the top of the 3000A breaker because the incident energy at that point is 75-100 cal/cm2.

It would have been prudent to house that particular circuit breaker in a separate enclosure so that when it was open, the entire inverter enclosure could be safe to work on. If the system HAD been connected to a single transformer and did not need the breaker from the transformer to the grid to be able to handle the inrush from 8 transformers, the incident energy would have been less than 40 cal/cm2, but given the nature of alpha builds - things change.

It is likely that all future inverter enclosures will have a separate totally enclosed section with a blowout panel for the main circuit breaker.


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Mon Oct 01, 2012 6:02 pm 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
Larry Stutts wrote:
It is likely that all future inverter enclosures will have a separate totally enclosed section with a blowout panel for the main circuit breaker.


Funny you should say that. A Canadian underground substation manufacturer does just that. They use dry transformers (very common in underground gear because they are vertically more compact) which necessitates a set of ventilated doors. The main circuit breaker has a very heavy front door but a "wireway" which is wide open leading out into the transformer compartment. The second purpose for this opening is really to ensure that should the breaker fail, it will vent out into and through the transformer compartment. This kind of design won't work for unitary substation and padmount style transformers which use either shielded cable or duct banks but could be adapted for instance to isolation transformers connected to drives.

In another application I have, the secondary side of a transformer (1 MVA) is well above 40 cal/cm^2. There are molded case circuit breakers mounted directly on the enclosure walls of the transformer (again, dry type). So they really can't be worked on at all for a variety of reasons but outside of the things inside this enclosure, the incident energy runs <1.2 cal/cm^2. So putting in main CB's on the MCC's provides a place to do LOTO and work downstream. For the transformer itself only the dsconnect upstream of it is ever used. The circuit breakers are molded case so they are not serviceable except to inspect annually and test every 3 years.


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Tue Oct 09, 2012 9:16 am 
Sparks Level
User avatar

Joined: Sun Jan 04, 2009 5:00 pm
Posts: 286
Location: Louisville, KY
There is a standard for such a barrier in blanket form. It has different fault currents and the highest rated to date are 40,000A.

ASTM F2676 - 09 Standard Test Method for Determining the Protective Performance of an Arc Protective Blanket for Electric Arc Hazards

These have to be connected to a structure which can take the blast. They are tested with such a structure and with cords (normally Kevlar) which can hold the pressure wave for the clearing time required.

The total calories are NOT the issue. Fault current for a clearing time. The max fault current and clearing time is the most critical information. The blankets are tested for the max and for other combinations which equal the same energy to assure they work from the maximum fault current with that clearing time but also with lower fault currents and longer clearing times. Most must pass 5kA for 120 cycles (2s). They are commonly available on the market. Most are 25kA but a few are 40kA.


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Tue Oct 09, 2012 12:58 pm 
Sparks Level
User avatar

Joined: Tue Apr 17, 2012 8:19 am
Posts: 241
Location: Charlotte, NC
elihuiv wrote:
These have to be connected to a structure which can take the blast.


But how do you know for certain a structure can take the blast (i.e. the gist of my original question)?


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Tue Oct 09, 2012 4:23 pm 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
Larry Stutts wrote:
But how do you know for certain a structure can take the blast (i.e. the gist of my original question)?


You build a simulation and test them. The other problem with the blankets is that again, no design data. They have to be tested for the application. It's not like you can buy a "25 kA" blanket, hang it, and you're done.


Top
 Profile Send private message  
Reply with quote  
 Post subject:
PostPosted: Wed Oct 10, 2012 5:43 am 
Sparks Level
User avatar

Joined: Tue Apr 17, 2012 8:19 am
Posts: 241
Location: Charlotte, NC
PaulEngr wrote:
You build a simulation and test them. The other problem with the blankets is that again, no design data. They have to be tested for the application. It's not like you can buy a "25 kA" blanket, hang it, and you're done.


Yes, I realize you test it. But there has to be some starting point. I would think that going for way-obvious-overkill would be just as inappropriate as ridiculously-undersized. I am sure it takes several evolutions of design once you are actually in the ballpark.

At some point you have to actually test it real-world, right? If you have done a few previous tests you might have an idea of what it will take.

I understand that it just takes what it takes. It just seems it would be a lot more efficient and less time consuming if you had a general idea of what size materials you were going to use when you pick the cal/cm2


Top
 Profile Send private message  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 32 posts ]  Go to page 1, 2  Next

All times are UTC - 7 hours


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to:  
cron
© 2019 Arcflash Forum / Brainfiller, Inc. | P.O. Box 12024 | Scottsdale, AZ 85267 USA | 800-874-8883