It is currently Sat Aug 08, 2020 4:15 pm



Post new topic Reply to topic
Author Message
ekstra   ara
 Post subject: Multiple branch study
PostPosted: Wed Jun 13, 2018 11:03 pm 
User avatar

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

I have been doing arc flash calculations for some time now, and often I get into an issue when I have a system with multiple branches. The softwares I have used don't take this into account either, and we have to calculate manually if we want a more accurate result.

The system is as follows: there are four generators, connected to two separate bus bars that are connected with a bus-tie breaker. The worst case arc energy is when all four generators are running in parallel and the bus tie is closed.

The problem is that the bus-tie have different time settings than the generators. Bus tie often trips at 100ms or less. And generators at 0.5s. If we for simplicity say that all generators are equal (not always the case, but very often), then we are left with only half the fault current for 0.4s and the full fault current for 0.1s.

And how do we now calculate the protection boundary? Is it the full fault current for 100ms, half the fault current for 0.4s (or perhaps half the fault current for 0.5s since that current is actually there the entire time until cleared by the generators), or is it someplace in between?

I guess this would also be the case if considering decrementing currents. Then you would have different fault currents for different amount of times. How would one calculate the protection boundary in such cases?


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Multiple branch study
PostPosted: Mon Jun 18, 2018 8:00 am 

Joined: Tue Apr 17, 2012 9:17 am
Posts: 10
Which software program are you using? ETAP has a method called "subtraction of Incident Energy for Multiple Source Systems". It's oddly named, but it basically does what you're looking for. I don't want to copy-paste the entire help file here (don't want to run into any copyright issues), but here's an excerpt.

"The second method was added to ETAP 7.0.0 to handle those situations for which the first method is not acceptable. Mainly the cases there multiple sources have very different trip times. The new method is called “Subtraction of Incident Energy for Multiple Source Systems”. There is an ETAP Options (Preferences) setting which needs to be set to “True” in order to activate this method. The following image shows the entry and its location in the preferences editor:

~ image left out ~

For example, we can analyze a multiple source system and place an arc fault at Bus “5BM”. Bus 5Bm is fed from two different utility connections with source protective devices (HVCBs) “22” and “12”. Each breaker operates at different fault clearing times.

The difference in the operating times is caused by the different relay time dial settings as shown in the following TCC:

~ image left out ~

The time difference averages about 2 to 3 seconds between the relay operations. In this case RelayA1 operates first (at approximately 1 second) and its arcing current contribution is removed at this time. This constitutes almost 33% of the energy contribution to the fault location. The total incident energy calculated with the removal of CB 12’s contribution is nearly 80 cal/cm2 which is well above the maximum value described for PPE in NFPA 70E 2009. By comparison if we were to run the same calculation without using the incident energy subtraction method, the incident energy results would be as shown in the images below:

~ image left out ~

The incident energy removed between 1 and 3 seconds (33% of the total fault current) amount to about 20 cal/cm2. The new result is nearly104 cal/cm2."


I've used Easypower in the past and they did not have this option in version 9. Not sure about SKM's options.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Multiple branch study
PostPosted: Mon Jun 18, 2018 1:53 pm 
Plasma Level
User avatar

Joined: Wed May 07, 2008 5:00 pm
Posts: 822
Location: Rutland, VT
JeffBlichmann wrote:
I've used Easypower in the past and they did not have this option in version 9. Not sure about SKM's options.


EasyPower has had this function for awhile and it was in Ver 9. It is called Integrated method.

_________________
Barry Donovan, P.E.
www.workplacesafetysolutions.com


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Multiple branch study
PostPosted: Tue Jun 19, 2018 2:07 am 
User avatar

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

I am using Paladin DesignBase 6.1 and can't find a similar function.

Just curious: how do the programs that allows multiple branches with different time settings deal with the arc protection boundary? Is it the initial arc current for the initial time period, the subsequent arc current for the entire time period, worst case of the two or something else entirely?

A few examples to illustrate why I am asking:

Both examples are ungrounded switchgear, 690V, 22mm gap, 610mm working distance.

1. A 50kA system supplied from two equal sources. 0.5s trip of first breaker and 1s trip of second breaker.
50kA for 0.5s = 9331 mm boundary zone
25kA for 1s = 9070 mm boundary zone

2. A 50kA system system supplied from two equal sources. 0.3s trip of first breaker and 0.7s trip of second breaker.
50kA for 0.3s = 6596 mm boundary zone
25kA for 0.7s = 7120 mm boundary zone

So in the first example the initial current for the initial period is worst case. In second example the remaining fault current for the entire period until cleared by last breaker is worst case. Is this way of considering arc boundary correct? What really happens in an arcing event when fault currents are being tripped at various times? Do the arc energy add up for the diffenrent periods? Same for boundary?


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Multiple branch study
PostPosted: Tue Jun 19, 2018 3:45 am 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
Let's simplify this a bit. Arc power mostly depends on current and the relationship is nonlinear. So you cannot use superposition to add source A + source B, etc., nor to determine it using some sort of maximal relationship, and for the same reason trying to take max(A, B...) isn't going to work either. This forces you to look at each combination of conditions and different time points.

Fortunately when we integrate over time to get arc energy (and by extension, incident energy), since time is linear, we can calculate piecewise linear results. Thus if you have scenario A for 0.1 seconds and scenario B for 0.4 seconds, you can calculate the incident energy for scenario A, then add the incident energy for scenario B multiplied by 0.75 (0.4-0.1)/0.4 to account for A conditions for 0.1 seconds then B conditions for 0.3 seconds. This was not stated in the 2002 edition of IEEE 1584 but will likely be in the next edition.

You can't really average or add different initial conditions. This would lead to erroneous conditions. So if you have say scenario A and scenario B depending on which combination of generators are operating and you were to average them to determine either incident energy or by extension the arc flash boundary, then the worker would be underprotected in one scenario and overprotected in the other. Weighting them via probability is the same result because the hazard is not decreasing, only the probability of occurrence.

In terms of multiple scenarios such as multiple generators with different conditions where customer is running different combinations, this is what scenarios are for. In practice typically the condition with the lowest available fault current results in the highest trip time and the highest incident energy.

In terms of interpretation this is where it gets messy. A very conservative approach would be to use the incident energy from the most extreme scenario. However this may be an operating condition that only exists for a short period of time such as a startup and may not be representative of typical conditions. A typical example would be operating a main-tie-main with the tie and both mains closed, a scenario that is often disallowed via interlocks. So you kind of have four choices:
1. Use worst case for everything, however unlikely. This is applying the principle of "conservativism" but often makes the whole report irrelevant and though it may be inline with engineering ethics, it often eschews practicality and casts the profession in a bad light. There is a big difference between delivering realistic results that are disappointing and delivering unrealistic/ridiculous results. When it comes to mixtures of generator and utility feeds, this is usually not the best approach because usually the worst case (lowest available fault current) conditions are the least likely to occur in practice so the plant is unnecessarily hindered with all kinds of stupidly crazy conditions that don't apply.
2. Use one or more sets of values based on likely scenarios such as a label with a table depending on some easy to determine condition such as how many generators are operating. Workers thus have to have knowledge of current conditions to read the table. The more complex the table obviously the less useful which is why a small number of conditions is best. Accuracy is nice but if you can group them into "close enough" conditions this drastically reduces the size of the table.
3. Use the most likely/common condition with warnings or footnotes that qualify it with no further guidance (look up the value in the report) based on how frequently this may occur. This is basically the table method except the table isn't in the field, it's in the office. This works if there is only one or two typical operating conditions and everything else is a rare event.
4. If the report is for someone else, report results and let them make their own determination (with coaching). This places the decision making in the hands of the decision maker. The engineer becomes a resource.

I also have some concerns about your scenarios. 50 kA sounds more like AIC and not a realistic incident energy. Did you include cable impedances (lengths and resistances)? Skipping out on that often tends to under-report actual incident energy. Second, 9 meter arc flash boundary probably places the arc flash boundary outside of the room in many cases. Be careful that your values are even valid and meaningful. Incident energy is pure thermal radiation which is trivially stopped even by a single sheet of drywall, so using results that place the arc flash boundary beyond the confines of a room or beyond a row of equipment is irrelevant and can be replaced with a simple requirement such as not standing in the doorway.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Multiple branch study
PostPosted: Tue Jun 19, 2018 4:06 am 
User avatar

Joined: Fri Jun 08, 2012 10:53 am
Posts: 39
Thank you very much for a very detailed answer.

And yes, cables are included.
The problem we are facing is that since our software does not include multiple branches, we get a unrealistic high energy and boundary zone. We therefore have to do this manually and now I know how to deal with the different time steps.

Thanks alot.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Multiple branch study
PostPosted: Fri Apr 19, 2019 1:58 pm 

Joined: Fri Apr 19, 2019 1:57 pm
Posts: 2
Does SKM have this function?


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Multiple branch study
PostPosted: Wed Oct 30, 2019 11:39 am 

Joined: Fri Apr 19, 2019 1:57 pm
Posts: 2
I'm guessing based on the silence, the answer is no?


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Multiple branch study
PostPosted: Fri Nov 01, 2019 1:29 pm 
Plasma Level
User avatar

Joined: Wed May 07, 2008 5:00 pm
Posts: 822
Location: Rutland, VT
Why not ask SKM?

_________________
Barry Donovan, P.E.
www.workplacesafetysolutions.com


Top
 Profile Send private message  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 9 posts ] 

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