Author 
Message 
Jen

Post subject: Arc Flash Analysis accounting for decline in Fault Current over Arcing time? Posted: Tue Jul 23, 2013 6:01 am 

Joined: Tue Jul 23, 2013 5:41 am Posts: 3

Hi, I'm doing some analysis in which I am modelling the Arc Flash Hazard of 600V Switchgear with a 56kA Fault Current. However I am struggling to understand how modelling at a fixed bolted 3phase fault current using the IEEE equations can give a realistically value. This is becuase the fault current is not fixed in this case, it has an initial peak of 56kA due to the sub transient effect but then has a exponential decline until steady state over the arcing time, until the breaker trips at 0.2 s in this case. What I'm basically asking is if there are a more complex set of equations that can take into a decline in fault current over arcing time? Rather than a linear set which the IEEE equations seem to be as they simple give you a value for worst case over the arcing time?


Top 


Jim Phillips (brainfiller)

Post subject: Posted: Wed Jul 24, 2013 4:54 am 

Plasma Level 

Joined: Mon Sep 17, 2007 5:00 pm Posts: 1501 Location: Scottsdale, Arizona

Jen wrote: Hi, I'm doing some analysis in which I am modelling the Arc Flash Hazard of 600V Switchgear with a 56kA Fault Current. However I am struggling to understand how modelling at a fixed bolted 3phase fault current using the IEEE equations can give a realistically value. This is becuase the fault current is not fixed in this case, it has an initial peak of 56kA due to the sub transient effect but then has a exponential decline until steady state over the arcing time, until the breaker trips at 0.2 s in this case. What I'm basically asking is if there are a more complex set of equations that can take into a decline in fault current over arcing time? Rather than a linear set which the IEEE equations seem to be as they simple give you a value for worst case over the arcing time? At the present time the IEEE 1584 model uses the bolted rms symmetrical short circuit current. It does not account for the decrement which for a strong source, should not be much of an issue. It also is not based on the asymmetrical current which is X/R dependent. However, there can be a significant decay/decrement of short circuit contribution from motors and smaller generators. In this case the only option would be to perform calculations over specific short time intervals i.e. a piecewise solution. (although I know of no one that does this) We have considered language about the piecewise method for the future edition of IEEE 1584 but nothing has been decided yet.
_________________ Jim Phillips, P.E. Brainfiller.com


Top 


Jen

Post subject: Posted: Wed Jul 24, 2013 5:27 am 

Joined: Tue Jul 23, 2013 5:41 am Posts: 3

Jim Phillips (brainfiller) wrote: At the present time the IEEE 1584 model uses the bolted rms symmetrical short circuit current. It does not account for the decrement which for a strong source, should not be much of an issue. It also is not based on the asymmetrical current which is X/R dependent. However, there can be a significant decay/decrement of short circuit contribution from motors and smaller generators. In this case the only option would be to perform calculations over specific short time intervals i.e. a piecewise solution. (although I know of no one that does this) We have considered language about the piecewise method for the future edition of IEEE 1584 but nothing has been decided yet. Hi Jim Thank you for your reply, the system of which I am referring to is a marine vessels power systems model so you can see my dilemma, I shall consider a piece wise solution to try to get a more accurate result for the system in question. Jen


Top 


Beo

Post subject: Posted: Mon Jul 29, 2013 6:52 am 

Joined: Fri Jun 08, 2012 10:53 am Posts: 39

Jen, I have written a masters thesis about arc flash in marine vessels, and found the same problem as you. There are no guidlines for how to model the decrementing currents in the IEEE 1584. I did some stepwise simplified calculations just to show the effects of decrementing currents, and the incident energy with decrementing currents for that particular generator was found to be under 50% compared to using only the 1/2 period subtransient values of the short circuit. It is also a problem if the steady state currents for a generator delivers so little fault currents that the resulting arcing current is too low to release the breaker. I am still investigating if my methods can be used, but I did model the steady state currents and used that as input for the minimum arcing currents, and made sure that all breakers are set below this point. In my case the generator breakers were delayed 0.5s and the steady state arcing currents was below the threshold of the breakers. This could lead to a potential catastrophic incident where the arc can do much damage to the switchboards and subsequently a blackout of the entire ship. It seems that arc flash is a fairly new phenomenon in the marine industry, and I can find no other classification society than Lloyds Register that requires arc flash calculations, even though the consequence of an arc flash in a vessel can be significantly worse than for shorebased installations. I have contacted several classification societies about this, and there seems to be a general unwillingnes to include arc flash calculations as part of the power system documentation.


Top 


Jen

Post subject: Posted: Mon Jul 29, 2013 8:38 am 

Joined: Tue Jul 23, 2013 5:41 am Posts: 3

Hi Beo, Thank you for your reply I totally agree I have also found a significant lack of documentation with respect to how the IEEE 1584 can be applied to marine vessels so I have tried to do some calculations myself, I would love to read your thesis and talk to you more about it if you would permit me? Jen


Top 


Marcelo

Post subject: Posted: Tue Jul 30, 2013 6:50 am 

Joined: Wed Dec 02, 2009 4:46 pm Posts: 29 Location: CT

Beo wrote: Jen, I have written a masters thesis about arc flash in marine vessels, and found the same problem as you. There are no guidlines for how to model the decrementing currents in the IEEE 1584. I did some stepwise simplified calculations just to show the effects of decrementing currents, and the incident energy with decrementing currents for that particular generator was found to be under 50% compared to using only the 1/2 period subtransient values of the short circuit. It is also a problem if the steady state currents for a generator delivers so little fault currents that the resulting arcing current is too low to release the breaker. I am still investigating if my methods can be used, but I did model the steady state currents and used that as input for the minimum arcing currents, and made sure that all breakers are set below this point. In my case the generator breakers were delayed 0.5s and the steady state arcing currents was below the threshold of the breakers. This could lead to a potential catastrophic incident where the arc can do much damage to the switchboards and subsequently a blackout of the entire ship. It seems that arc flash is a fairly new phenomenon in the marine industry, and I can find no other classification society than Lloyds Register that requires arc flash calculations, even though the consequence of an arc flash in a vessel can be significantly worse than for shorebased installations. I have contacted several classification societies about this, and there seems to be a general unwillingnes to include arc flash calculations as part of the power system documentation. I suggest that this identifies the most important issue. The issue is not so much correctly calculating incident energy. Decrementing voltage will lower incident energy available per unit time so incident energy calcculations may be conservative if interrupting time is properly accounted for. However, the lowering of arcing current over time also affects the ability of the protection system to detect the current and react. Failure of protection to react as expected due to a lower arcing current than expected can result in significantly more incident energy, equipment damage, downtime and hazard. I think creating a protection system that reacts quickly, even if arcing currents are lower than theoretically predicted is very important to have a robust system.


Top 


Beo

Post subject: Posted: Wed Jul 31, 2013 10:27 am 

Joined: Fri Jun 08, 2012 10:53 am Posts: 39

Jen, the thesis is not public available. Drop me your contact information in a pm and I will see what can be done. It was written for the company that I work for and there might be some limitations and restrictions. I am always up for discussing arc flash, especially in the maritime industry. The minimum arcing currents certainly is something that needs to be considered. It can be difficult to obtain accurate decrement curves for the generators though. The protection system from JHU/APL that is now commercially available from DRS Technologies is especially developed for marine switchboards and contains a thermal ionisation detector that can detect an arc hours or even days before an arc incident. There are other interesting protection systems out there but the one developed by JHU/APL is the only one with the thermal detection system.


Top 


PaulEngr

Post subject: Posted: Fri Aug 02, 2013 3:58 am 

Plasma Level 

Joined: Tue Oct 26, 2010 9:08 am Posts: 2174 Location: North Carolina

ETAP does a piecewise calculation but applies only global parameters to it. SKM also does a piecewise calculation but you can adjust the settings on an individual basis. With both of them, the contributions are considerably simplified. They are definitely capable of handling motor transients but I can't remember if they also model subtransients or not. IEEE 1584 also does not really specify how to calculated a bolted fault. There are at least 4 or 5 methods out there although only the vast majority are done only using two models. A power system analysis that takes into account more detail would derive an even better model and typically (but not always) produces a lower bolted fault current than predicted such as via the ANSI calculation method. Where things get tricky is that the IEEE 1584 equations are based on measured arcing currents and voltages. A method is used to estimate the arcing current based on the bolted fault current. Where this falls down is that there are assumptions present about how the bolted fault current is calculated. If you deviate from the theoretical bolted fault calculation, then plugging in a different bolted fault current may produce a higher or lower result that may not reflect the actual arcing current (and thus incident energy) experienced in the real world. Would be a fairly easy thesis project for someone since in this case we just have to produce arcing faults and measure the currents/voltages and not get into making calorific measurements.


Top 


Beo

Post subject: Posted: Fri Aug 02, 2013 12:02 pm 

Joined: Fri Jun 08, 2012 10:53 am Posts: 39

Onboard a ship you will almost always find that the arc flash boundary zone for the main switchboard exceeds the available space inside the switchboard room. Another thing to consider besides decrementing currents is multiple branches, all the softwares I have tried fail at calculating arc flash when multiple branches are involved.


Top 


PaulEngr

Post subject: Posted: Sun Aug 18, 2013 10:26 am 

Plasma Level 

Joined: Tue Oct 26, 2010 9:08 am Posts: 2174 Location: North Carolina

Beo wrote: Another thing to consider besides decrementing currents is multiple branches, all the softwares I have tried fail at calculating arc flash when multiple branches are involved. Not true at all. They do just fine. In fact, it is one of the most fundamental features of power system analysis software. In switchgear or MCC's or panelboards, you always have a "bus". Although the bus has some small impedance this is rarely modeled. The "bus" represented on the screen in the modeling software typically has zero impedance and as many branches as you want (even if it's just a passthrough). Just because insulated splices are not typically thought of as "buses" from a power distribution modelling point of view, they are identical. Each "branch" that point connects individual cables or busbars. Where you get into trouble at least with SKM is it's implementation of a "panelboard" or "NCC". The purpose of this function is to model a "load". You can get TCC's, arc flash, etc., from each device attached within the object but this model works best specifically when you have an MCC or panelboard and downstream devices are "simple" and you are not interested in modelling them to any great degree. Otherwise, you need to use a bus and develop all the objects downstream from there. The one complication here at least with SKM is that you are not allowed to have two protective devices backtoback with only a bus in between. This commonly occurs when for instance you connect a bus directly to a main circuit breaker and then connect the downstream circuit breakers to the bus. Inserting a small impedance downstream of the "main" circuit breaker solves the issue.


Top 


Beo

Post subject: Posted: Mon Aug 19, 2013 1:20 am 

Joined: Fri Jun 08, 2012 10:53 am Posts: 39

The scenario I calculated was four generators in parallel with a bus tie between. Two generators on each bus. All generator breakers are delayed 0.5s and the bus tie trips at 40ms. When analysing this in SKM PTW and Paladin DesignBase the fault is left until cleared by the generators after 0.5s. However, the bus tie will clear half the fault in 40ms and thus resulting in a much lower incident energy. I had to calculate those scenarios manually in order to get the correct level of PPE.


Top 



Page 1 of 1

[ 11 posts ] 


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

