Topic review  IEEE 1584 Equations  Your thoughts? 
Author 
Message 


Post subject: 
Re: IEEE 1584 Equations  Your thoughts? 


I don't think the basic equations are all that bad. It seems like for instance the Wilkins "simplified" model really isn't all that simplified. The biggest problem is that with the dependent current term (sort of selfreferencing), it kind of defies certain kinds of analyses, making it difficult to develop maps of the "solution space" easily for instance. So some of the "boundary" type efforts that have been done do this. I do see some value in a "really simple" model but unfortunately most of the people asking for this really mean something on par with the 70E tables. And the really complicated part really isn't IEEE 1584 per se, it's how we get to realistic values of bolted fault current as well as opening time. Once we have these, IEEE 1584 empirical equation is pretty straight forward.
But the really big challenging part in my mind is the enclosure size table. I understand what it represents...there are at this point perhaps a half dozen "standard" test enclosure sizes that have been used and the trick is to use the enclosure most representative of the equipment actually being studied. In some cases such as a size 2U MCC bucket this is very straight forward. But as we move away into custom built equipment such as industrial control panels or even the larger MCC buckets typical of NEMA Size 4 and larger, the correspondence falls apart. This is where the current table actually does a disservice to the analyst. This is never mind the orientations and other newer tests.
My recommendation is to simply list the test enclosure dimensions and then provide one or two sentences after that suggesting equipment that is likely to be represented well by the test enclosure. That way if I am working with a "standard" enclosure such as a class E2 starter or a 2U MCC bucket I can just use the comments/recommended list. If I'm dealing with custom built equipment, I can use the raw test enclosure dimension list to attempt to assign one that is closest.
Then moving down the scale we get into the Wilkins timedomain and similar models. These are really cool from a computational point of iew because the curve fit is better (questioning the confidence interval in these results since as the number of adjustable parameters increases the model confidence interval becomes enormous) but the consequence is going to be that the model(s) are necessarily much more challenging to use. This is even more true if we can't reach agreement between the academics and the pragmatists and end up with a need to satisfy both. We already have that in the 2002 standard in that there are at least 3 models: the theoretical (Lee) model, the empirical model, and a set of equipment specific (fuse) models. Adding boundary methods or time model models simply enriches the playing field even more.
I don't think the basic equations are all that bad. It seems like for instance the Wilkins "simplified" model really isn't all that simplified. The biggest problem is that with the dependent current term (sort of selfreferencing), it kind of defies certain kinds of analyses, making it difficult to develop maps of the "solution space" easily for instance. So some of the "boundary" type efforts that have been done do this. I do see some value in a "really simple" model but unfortunately most of the people asking for this really mean something on par with the 70E tables. And the really complicated part really isn't IEEE 1584 per se, it's how we get to realistic values of bolted fault current as well as opening time. Once we have these, IEEE 1584 empirical equation is pretty straight forward.
But the really big challenging part in my mind is the enclosure size table. I understand what it represents...there are at this point perhaps a half dozen "standard" test enclosure sizes that have been used and the trick is to use the enclosure most representative of the equipment actually being studied. In some cases such as a size 2U MCC bucket this is very straight forward. But as we move away into custom built equipment such as industrial control panels or even the larger MCC buckets typical of NEMA Size 4 and larger, the correspondence falls apart. This is where the current table actually does a disservice to the analyst. This is never mind the orientations and other newer tests.
My recommendation is to simply list the test enclosure dimensions and then provide one or two sentences after that suggesting equipment that is likely to be represented well by the test enclosure. That way if I am working with a "standard" enclosure such as a class E2 starter or a 2U MCC bucket I can just use the comments/recommended list. If I'm dealing with custom built equipment, I can use the raw test enclosure dimension list to attempt to assign one that is closest.
Then moving down the scale we get into the Wilkins timedomain and similar models. These are really cool from a computational point of iew because the curve fit is better (questioning the confidence interval in these results since as the number of adjustable parameters increases the model confidence interval becomes enormous) but the consequence is going to be that the model(s) are necessarily much more challenging to use. This is even more true if we can't reach agreement between the academics and the pragmatists and end up with a need to satisfy both. We already have that in the 2002 standard in that there are at least 3 models: the theoretical (Lee) model, the empirical model, and a set of equipment specific (fuse) models. Adding boundary methods or time model models simply enriches the playing field even more.




Posted: Thu Sep 15, 2016 8:41 pm 





Post subject: 
Re: IEEE 1584 Equations  Your thoughts? 


I voted too complex but it is also too simple. We are presenting a poster at the ESW where we believe the calculation may miss the worst case IE. We are looking at a graphical approach then calculating the IE
I voted too complex but it is also too simple. We are presenting a poster at the ESW where we believe the calculation may miss the worst case IE. We are looking at a graphical approach then calculating the IE




Posted: Mon Sep 12, 2016 8:08 am 





Post subject: 
Re: IEEE 1584 Equations  Your thoughts? 


spark wrote: It is not that they are too complex but they are complex. You can program them into Excel and provide some repeatability and reliability. I remember when these equations first came out, I could perform a calculation on my calculator three times and get three different answers. I could wash my clothes down by the river with a rock but I have a washing machine. I think for all practical purposes, a tool such as the software or other programs that have been proven to work should be used. Performing the calculations by hand for experience or become familiar with the calculation is one thing, to do it for an actual project, not practical. Same feelings here. I first did some calculations longhand before writing them into into Excel. The IEEE 1584 spreadsheet provided a better tool for reporting. It was good to go through that sequence to understand and remember how it all works.
[quote="spark"]It is not that they are too complex but they are complex. You can program them into Excel and provide some repeatability and reliability. I remember when these equations first came out, I could perform a calculation on my calculator three times and get three different answers. I could wash my clothes down by the river with a rock but I have a washing machine. I think for all practical purposes, a tool such as the software or other programs that have been proven to work should be used. Performing the calculations by hand for experience or become familiar with the calculation is one thing, to do it for an actual project, not practical.[/quote]
Same feelings here. I first did some calculations longhand before writing them into into Excel. The IEEE 1584 spreadsheet provided a better tool for reporting. It was good to go through that sequence to understand and remember how it all works.




Posted: Mon Sep 12, 2016 6:24 am 





Post subject: 
Re: IEEE 1584 Equations  Your thoughts? 


It is not that they are too complex but they are complex. You can program them into Excel and provide some repeatability and reliability. I remember when these equations first came out, I could perform a calculation on my calculator three times and get three different answers. I could wash my clothes down by the river with a rock but I have a washing machine. I think for all practical purposes, a tool such as the software or other programs that have been proven to work should be used. Performing the calculations by hand for experience or become familiar with the calculation is one thing, to do it for an actual project, not practical.
It is not that they are too complex but they are complex. You can program them into Excel and provide some repeatability and reliability. I remember when these equations first came out, I could perform a calculation on my calculator three times and get three different answers. I could wash my clothes down by the river with a rock but I have a washing machine. I think for all practical purposes, a tool such as the software or other programs that have been proven to work should be used. Performing the calculations by hand for experience or become familiar with the calculation is one thing, to do it for an actual project, not practical.




Posted: Mon Sep 12, 2016 5:35 am 





Post subject: 
Re: IEEE 1584 Equations  Your thoughts? 


I answered "seems to be OK" for a couple of reasons.
We don't really have anything else to use right now and this seems to be working i.e. I haven't heard of any series injuries when this is used.
As far as the answer "too simple/not enough detail" I question this one. We don't seem to get too specific with what we already have. For example, arc flash software has default arc gaps, working distance etc.  does anyone use anything different? Does anyone actually measure the bus gap on equipment?
The better question would be if we could include more detail, would people just use a default anyway? This is also assuming the results of the short circuit and coordination studies needed are correct. I keep hearing there is a new edition of IEEE 1584 that will be out someday, but for now, I'm sticking with what we have "seems to be OK"
I answered "seems to be OK" for a couple of reasons.
We don't really have anything else to use right now and this seems to be working i.e. I haven't heard of any series injuries when this is used.
As far as the answer "too simple/not enough detail" I question this one. We don't seem to get too specific with what we already have. For example, arc flash software has default arc gaps, working distance etc.  does anyone use anything different? Does anyone actually measure the bus gap on equipment?
The better question would be if we could include more detail, would people just use a default anyway? This is also assuming the results of the short circuit and coordination studies needed are correct. I keep hearing there is a new edition of IEEE 1584 that will be out someday, but for now, I'm sticking with what we have "seems to be OK"




Posted: Sun Sep 11, 2016 2:08 pm 



Jim Phillips (brainfiller) 


Post subject: 
IEEE 1584 Equations  Your thoughts? 


When an arc flash study is performed, most rely on software that is built around the equations from IEEE 1584. This week's question focuses on performing arc flash calculations by hand  either as an exercise or for an actual analysis.
What are your thoughts about the existing IEEE 1584 equations?
Too complex / Too much detail Too simple / Not enough detail Seems to be OK as is. I've Never performed calculations manually
When an arc flash study is performed, most rely on software that is built around the equations from IEEE 1584. This week's question focuses on performing arc flash calculations by hand  either as an exercise or for an actual analysis.
[b]What are your thoughts about the existing IEEE 1584 equations?
Too complex / Too much detail Too simple / Not enough detail Seems to be OK as is. I've Never performed calculations manually [/b]




Posted: Sat Sep 10, 2016 11:07 am 




