It is currently Mon Sep 28, 2020 2:34 am



Post new topic Reply to topic
Author Message
ekstra   ara
 Post subject: Data Collection
PostPosted: Mon Feb 11, 2019 9:15 am 

Joined: Fri Feb 01, 2019 1:16 pm
Posts: 8
Hi all,

Pretty new here but liking the learning atmosphere. I've just started doing some arc flash studies and just wondering how critical certain data is to a study.
The system I'll do studies on is pretty much the same SLD with some minor changes (additional pump,etc.). MCC-600V has SMC/pumps attached (not shown on pic).

1. Cable lengths/resistances - how accurate should you be? I've seen some consultants who did studies for us put in a tolerance of +/- 10% and eyeball the lengths. Are cable lengths from the SMC to the pump critical? Do you need to diligently gather all the cable sizes and lengths? Or overall, given the +/- 10% assumption that it wouldn't be a big deal in the end?

2. If the pumps/motors are run by a SMC (soft starter), can you assume that there will be no fault/current contribution coming from them?

3. When having multiple pumps (say 3) in the MCC, what scenarios should be run? I've usually just seen worst case, which is to assume all pumps are running and add all the HP into one composite load. Is there any value in running a scenario when only 1 or 2 pumps are running?

4. For the ETAP gurus, I've seen different results from different ETAP versions using the exact same data(ie. breaker type, etc). Would it be attributed to updated device model/characteristics for the same device (ie. Square-D/KAL, 200A, 0.600kV) between different ETAP versions?


You do not have the required permissions to view the files attached to this post.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Collection
PostPosted: Thu Feb 14, 2019 12:13 pm 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
jvito wrote:
Hi all,

Pretty new here but liking the learning atmosphere. I've just started doing some arc flash studies and just wondering how critical certain data is to a study.
The system I'll do studies on is pretty much the same SLD with some minor changes (additional pump,etc.). MCC-600V has SMC/pumps attached (not shown on pic).

1. Cable lengths/resistances - how accurate should you be? I've seen some consultants who did studies for us put in a tolerance of +/- 10% and eyeball the lengths. Are cable lengths from the SMC to the pump critical? Do you need to diligently gather all the cable sizes and lengths? Or overall, given the +/- 10% assumption that it wouldn't be a big deal in the end?


As with everything, it depends! First off you have several impedances: the source transformer (or generator) impedance, the cable impedance, potentially the load impedance. Generally we are looking at the sum X1+X2+X3 but when one of these impedances dominates, the others don't matter. So with short cables, bus bars, etc., cable impedance is almost immaterial (transformer impedance dominates). But at longer distances (tens to hundreds of feet or more), cable impedance dominates and that's where it affects current. Now this is where it gets a little more complicated. If the operation of the OCPD is in a definite time region such as "instantaneous" tripping then lower impedances decreases incident energy so frankly a low error doesn't hurt anything. But if the overcurrent protective device is in an inverse time region (trips faster with higher currents) which is the case with all fuses and many breakers and it is using one of the "typical" curves we see in power distribution (ANSI inverse, very inverse, or extremely inverse), then as current drops time increases exponentially causing the net incident energy to actually increase. So paradoxically getting the cable size wrong on the low side is now detrimental and eliminating them altogether is definitely a bad idea. By looking at the predicted arcing current and the TCC's you can observe which part of the curve you are on and what the net effect of a measurement error will be, and act accordingly.

Quote:
2. If the pumps/motors are run by a SMC (soft starter), can you assume that there will be no fault/current contribution coming from them?


SMC is AB's product line name just like Square D uses Altistart and ABB uses PST. Soft starter is the generic term. It depends on the design. A soft start with a bypass contactor is essentially the same as a starter so you'd assume fault contribution from the motor exists. If however it is a "full time" soft starter without a bypass contactor then any fault current would have to flow inversely through the soft starter which is not likely or even if it is, too low level to matter. So summarizing the general rule is to not model the stuff on the other side of a VFD, but you do model the stuff on the other side of a soft starter, IN GENERAL. But you need to verify which particular equipment configuration you have since there are VFD's with bypass contactors, particularly around hospitals and water palnts, and there are soft starters without bypass contactors but that's not the most common design except with very small soft starters since it provides an efficiency boost, greater reliability, and a smaller enclosure for a negligible amount of additional costs.

Quote:
3. When having multiple pumps (say 3) in the MCC, what scenarios should be run? I've usually just seen worst case, which is to assume all pumps are running and add all the HP into one composite load. Is there any value in running a scenario when only 1 or 2 pumps are running?


This gets into modelling here the general theory is to ignore motors that are under 50 HP as their fault contribution (inductance) is negligible, add together the HP or all motors from 50 HP up to around 150 HP and treat it as one large motor, and model all larger motors individually. The inductance contributes fault current. So you need to model it but also take into consideration diversity...if we have say 3 pumps but only 2 can realistically run at a time, then we model 2 pumps, not 3. This is in the short circuit system analysis standards (IEC and ANSI). I suggest reading up on the requirements for the standard you are implementing for your short circuit analysis.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Collection
PostPosted: Fri Feb 15, 2019 8:39 am 

Joined: Fri Feb 01, 2019 1:16 pm
Posts: 8
Thanks for the answers! This has helped me alot.

As you can see our system is fairly basic which is good since I am a newbie to all of these. I took maybe 1-2 courses in power engineering and then took the ETAP ARC Flash course. Plus lots of reading online.

First off, is there any book or material you can recommend that has all these "rules of thumb and/or modelling rules"? Or you just get them from experience?

Quote:

Quote:
jvito wrote:
Hi all,

Pretty new here but liking the learning atmosphere. I've just started doing some arc flash studies and just wondering how critical certain data is to a study.
The system I'll do studies on is pretty much the same SLD with some minor changes (additional pump,etc.). MCC-600V has SMC/pumps attached (not shown on pic).

1. Cable lengths/resistances - how accurate should you be? I've seen some consultants who did studies for us put in a tolerance of +/- 10% and eyeball the lengths. Are cable lengths from the SMC to the pump critical? Do you need to diligently gather all the cable sizes and lengths? Or overall, given the +/- 10% assumption that it wouldn't be a big deal in the end?



As with everything, it depends! First off you have several impedances: the source transformer (or generator) impedance, the cable impedance, potentially the load impedance. Generally we are looking at the sum X1+X2+X3 but when one of these impedances dominates, the others don't matter. So with short cables, bus bars, etc., cable impedance is almost immaterial (transformer impedance dominates). But at longer distances (tens to hundreds of feet or more), cable impedance dominates and that's where it affects current. Now this is where it gets a little more complicated. If the operation of the OCPD is in a definite time region such as "instantaneous" tripping then lower impedances decreases incident energy so frankly a low error doesn't hurt anything. But if the overcurrent protective device is in an inverse time region (trips faster with higher currents) which is the case with all fuses and many breakers and it is using one of the "typical" curves we see in power distribution (ANSI inverse, very inverse, or extremely inverse), then as current drops time increases exponentially causing the net incident energy to actually increase. So paradoxically getting the cable size wrong on the low side is now detrimental and eliminating them altogether is definitely a bad idea. By looking at the predicted arcing current and the TCC's you can observe which part of the curve you are on and what the net effect of a measurement error will be, and act accordingly.


What you say makes sense with regard to cable lengths. For my case, it would probably 100 ft max from the utility power...plus other smaller lengths..so fairly negligible. I doubt total cable lengths would even add up to 200 ft.

The utility transformers dont even reach 0.5 MVA, most of the one's I've encountered are just 300 KVA.

What I did was get the name plate data from the field for the OCPD (breaker and fuses) and find them in ETAP's libraries to get the other device characteristics (impedance, TCCS). This is what the consultants we hired before did, and to me this makes sense. But my boss is hesitant and thinks we might be missing significant data by relying on ETAP libraries. What is the usual procedure for this? Can we request data (relevant to arc flash) direct from the manufacturer?

Quote:
Quote:
2. If the pumps/motors are run by a SMC (soft starter), can you assume that there will be no fault/current contribution coming from them?
SMC is AB's product line name just like Square D uses Altistart and ABB uses PST. Soft starter is the generic term. It depends on the design. A soft start with a bypass contactor is essentially the same as a starter so you'd assume fault contribution from the motor exists. If however it is a "full time" soft starter without a bypass contactor then any fault current would have to flow inversely through the soft starter which is not likely or even if it is, too low level to matter. So summarizing the general rule is to not model the stuff on the other side of a VFD, but you do model the stuff on the other side of a soft starter, IN GENERAL. But you need to verify which particular equipment configuration you have since there are VFD's with bypass contactors, particularly around hospitals and water palnts, and there are soft starters without bypass contactors but that's not the most common design except with very small soft starters since it provides an efficiency boost, greater reliability, and a smaller enclosure for a negligible amount of additional costs.


We have SMC-50, SMC Flex and some VFDs and our application is for pump stations. We previously hired consultants to do our studies for us and I doubt that they even checked if we had bypass contractor on our SMCs. When I asked them how they modeled it, is that they treated VFDs and SMCs the same: if they existed, they prevent fault contribution from the pumps back to the faulted bus (which is what is on the SLD). But even with this assumption, I've seen other reports where they lump all the motors/pumps into one lumped load and consider it 'worst case'. We usually have 2-3 pumps which are interlocked so practically speaking, only 1 can run at a time. Mind you, for most of our stations, our pumps are way below 50 HP. So either way, I doubt it would make a big difference. Maybe at the end of the day, the labels produced with pumps in the model would be the same or very close with the one were they aren't included in the model.

Quote:
Quote:
3. When having multiple pumps (say 3) in the MCC, what scenarios should be run? I've usually just seen worst case, which is to assume all pumps are running and add all the HP into one composite load. Is there any value in running a scenario when only 1 or 2 pumps are running?
This gets into modelling here the general theory is to ignore motors that are under 50 HP as their fault contribution (inductance) is negligible, add together the HP or all motors from 50 HP up to around 150 HP and treat it as one large motor, and model all larger motors individually. The inductance contributes fault current. So you need to model it but also take into consideration diversity...if we have say 3 pumps but only 2 can realistically run at a time, then we model 2 pumps, not 3. This is in the short circuit system analysis standards (IEC and ANSI). I suggest reading up on the requirements for the standard you are implementing for your short circuit analysis.


We use ANSI since we are in North America. Again, where can I find those 'rule of thumbs'? Is it in a book somewhere? Because most of our pumps are way below 50 HP ( mostly 20 HP and below). Does pump impedance even matter? I do agree with what you say and it does make sense.

However, my boss (who is also not very experienced with Arc Flash) is very hesitant when I reference 'rule of thumbs' I find on forums or online. This is reasonable, since he would be liable if something does happen. But I do think, Arc Flash Analysis, is one big assumption/estimation, so you can't get away with 'rule of thumbs' and reasonable assumptions.

Appreciate your help!


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Collection
PostPosted: Fri Feb 15, 2019 12:02 pm 

Joined: Fri Feb 01, 2019 1:16 pm
Posts: 8
Looking into it further, it seems that we do have a bypass contractors on some on our SMCs.

However, since its below 50 HP....is it reasonable to ignore its effects?


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Collection
PostPosted: Sun Feb 17, 2019 8:13 am 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
jvito wrote:
First off, is there any book or material you can recommend that has all these "rules of thumb and/or modelling rules"? Or you just get them from experience?
Quote:


This is in the ANSI standard I believe. I don't think IEC mentions it.

Quote:
What you say makes sense with regard to cable lengths. For my case, it would probably 100 ft max from the utility power...plus other smaller lengths..so fairly negligible. I doubt total cable lengths would even add up to 200 ft.


Take your model and adjust the cable lengths with some typical values you would expect. Run it a few times and write down the incident energy. It will quickly become obvious when accuracy matters.

Quote:
What I did was get the name plate data from the field for the OCPD (breaker and fuses) and find them in ETAP's libraries to get the other device characteristics (impedance, TCCS). This is what the consultants we hired before did, and to me this makes sense. But my boss is hesitant and thinks we might be missing significant data by relying on ETAP libraries. What is the usual procedure for this? Can we request data (relevant to arc flash) direct from the manufacturer?


You can but building your own models in the software is tedious. The software companies take great pains in getting this right. I would definitely rely on the libraries when there is an available, correct solution. If you are really worried about it, you can easily get TCC's off manufacturer web sites. Look at the opening time predicted by the model based on the predicted current. Compare this to the manufacturer's published data. Note that there are tolerances to watch out for, too.

Quote:
We use ANSI since we are in North America. Again, where can I find those 'rule of thumbs'? Is it in a book somewhere? Because most of our pumps are way below 50 HP ( mostly 20 HP and below). Does pump impedance even matter? I do agree with what you say and it does make sense.


It's in the ANSI standard itself as well as IEC.

Both ANSI and IEC standards were originally developed back in the slide rule days when all calculations were done by hand without calculators. So one of the "tricks" is that we ignore some things. In the ANSI model we make the assumption that for instance transformers have no resistance (only reactance) that that cable resistance is very low, we're in a high X/R regime. So we ignore R and only concern ourselves with X, and that smaller reactances (small motors) don't matter either. This gets us quickly into a model that we can easily deal with using a slide rule and hand calculations. We'd have to do 4 times the math if we were using the full Z model. As far as accuracy goes it is usually pretty close and in most cases, does not under-estimate short circuit current. So it's enough that we can calculate short circuits easily for equipment sizing purposes. That's the whole purpose of the ANSI model...to determine short circuit current for equipment sizing. Even for coordination purposes we can use this as a worst case number and just make sure that we've coordinated everything below that point on the curves.

Now for arc flash purposes some of these interpretations are right and some aren't. Small motors still don't matter but resistance and accurate reactance matters a lot more. We get artificially low incident energy if we don't use the full Z (reactance and resistance) model so that's what the software does. ETAP does NOT use ANSI for arc flash calculations. Compare the short circuit analysis and arc flash analysis bolted fault numbers and you'll see what I mean.

Quote:
However, my boss (who is also not very experienced with Arc Flash) is very hesitant when I reference 'rule of thumbs' I find on forums or online. This is reasonable, since he would be liable if something does happen. But I do think, Arc Flash Analysis, is one big assumption/estimation, so you can't get away with 'rule of thumbs' and reasonable assumptions.

Appreciate your help!


You aren't getting away with anything. All power system analysis is based on assumptions, estimations, etc. Even arc flash models are in fact an estimation based on experimental testing. To do true accurate models you'd use something like EMTP which is for transient analysis and does detailed time series analysis. To make it tractable, we use a lot of assumptions that convert a huge time series sparse matrix iterative solution model that would take hours to calculate into one with exact answers that can be crunched in seconds. All power analysis training should have covered this. But when it comes to making assumptions, estimates.,etc., those are actually defined in the standards themselves. That way we are all doing it the same way. All of the assumptions have been tested or modelled in some way, by using a far more complex numerical model or theoretical modelling and proving that the estimates work just as good as the full model (<1% error).

This should be part of the training you received. It is certainly part of the IEEE standards. It's expensive to purchase outright but you may want to look into getting access to the IEEE Color Books series that is the foundation of power system engineering as far as standards goes. Or see if you can find copies at a local engineering (university) library that you can read or borrow.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Collection
PostPosted: Fri Feb 22, 2019 1:27 pm 

Joined: Fri Feb 01, 2019 1:16 pm
Posts: 8
Quote:
SMC is AB's product line name just like Square D uses Altistart and ABB uses PST. Soft starter is the generic term. It depends on the design. A soft start with a bypass contactor is essentially the same as a starter so you'd assume fault contribution from the motor exists. If however it is a "full time" soft starter without a bypass contactor then any fault current would have to flow inversely through the soft starter which is not likely or even if it is, too low level to matter. So summarizing the general rule is to not model the stuff on the other side of a VFD, but you do model the stuff on the other side of a soft starter, IN GENERAL. But you need to verify which particular equipment configuration you have since there are VFD's with bypass contactors, particularly around hospitals and water palnts, and there are soft starters without bypass contactors but that's not the most common design except with very small soft starters since it provides an efficiency boost, greater reliability, and a smaller enclosure for a negligible amount of additional costs.


Would this (having no fault contribution) also apply to overload relays (E300) ? It seems to me that it functions similar to VFD and SMC and also has its own protection.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Collection
PostPosted: Sun Feb 24, 2019 3:26 pm 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
jvito wrote:
Would this (having no fault contribution) also apply to overload relays (E300) ? It seems to me that it functions similar to VFD and SMC and also has its own protection.


No. An overload relay is a passive device that monitors current and generates a trip signal for a contactor. Two reasons not to consider this. It does not limit fault current in any way. Furthermore, a contactor is limited to interrupting fault currents up to roughly 6-10 times it's rating. It is nowhere near the 20+ times rated current interruption capabilities of fuses and breakers.

Second overload relays are way too slow. The NEMA rated ones are rated in terms of how fast they trip at 6x FLA. So for instance a class 20 relay trips in 20 seconds at 600% of FLA. They are rated to trip in several minutes at 2x FLA. Any arc flash would be over before the overload relay gets close to tripping. But there's an exception and you're getting close to it. More advanced motor protection relays can trigger on jam or locked rotor conditions at much faster trip times. This is beneficial for larger and medium voltage motors. And they can be used in the same way as "fuse savers" tripping on low level faults instead of tripping fuses or breakers. BUT use of such things has to also consider the maximum contactor rating and to suppress tripping if the current exceeds the contactor's capabilities or the contactor is destroyed. So yes in some ways you could say an overload relay can be used in this way but it's not a typical scenario to consider and certainly not one that we'd typically rely on in protection scenarios (arcing and short circuit faults).


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Collection
PostPosted: Mon Feb 25, 2019 7:47 am 

Joined: Fri Feb 01, 2019 1:16 pm
Posts: 8
Good information!
So to be on the safe side, I'll model motors with E300s.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Data Collection
PostPosted: Mon Feb 25, 2019 4:32 pm 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
In Class E2 starters this is more common. R and E rated fuses are both slow and expensive. In a 600 V contactor though with fuse opening times in a fraction of a cycle, MCPs opening in 1-2 cycles, the typical 30-100 ms time of the contactor plus at least 1 cycle for the starter means the arcing fault will be cleared before the contactor has time to react even if you tried. E2 starters with vacuum contactors react in typically 3 cycles plus one for the relay so that’s about 50-65 ms, much faster than an R rated fuse meltibg time. Check your numbers but the E300 probably won’t matter at all. Look at the TCC for the arcing current calculated by the arc flash and short circuit models.


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