instantaneous setting for load-center-fed 460vac motor 4

[electricpete]

Instantaneous overcurrent setting for “high-efficiency” 460vac motor fed from load center breaker

For determining instantaneous overcurrent settings, I am used to applying the NEC limits (multiples of FLA) for motors fed from molded case circuit breakers in MCC’s. (For example 800 - 1300%FLA for standard motors and 1100-1700% for energy efficient motors... possibly next higher setting depending on your interpretation).
My MAIN QUESTION: Do these NEC guidelines/limits apply to motors fed from load center breakers?

The specific example motor I’m interested in:

Motor Supply breaker: Westinghouse DS-10 (load center breaker) with Amptector trip units. The tolerance band for the instantaneous setting is +/- 20%.

Motor nameplate data: 350hp, 460vac, 350A (FLA), 1200rpm, KVA code G (equates to LRC = 2109-2372 @100% voltage)
Motor efficiency is not marked on nameplate (motor purchased in early 1990’s).

Motor data sheet shows locked rotor current of 2200A (matches above range).
Motor data sheet shows efficiency of 95.7% at full load, full voltage.
When compared with recent revisions of NEMA MG-1 standard, I believe this motor would be considered an “energy efficient” motor. (Per NEMA MG-1-2005 table 12.11, 300hp 1200rpm motor with nominal energy efficiency 95.4 and min energy efficiency 94.5 can be classified as energy efficient, although the efficiency shall be marked on the nameplate under this standard and this efficiency was not... efficiency data retrieved from data sheet).

If I considered this a standard motor, the instantaneous trip would be 800% - 1300% * 341 =
2728 – 4433A.

If I considered this an energy efficient motor (as I’m inclined to), the instantaneous trip would be 1100% - 1700% * 341A = 3751 – 5797A

If I picked twice the locked rotor current, the setting would be 2*2200 = 4400.

Twice locked rotor currrent seems fairly conservative to me. We have for many years used 1.7*LRC on larger motors. We’ve been through the math before that says the true peak could be up to double using the steady state equivalent circuit and not taking any credit for expected decay of dc component before first peak.... the factor of 2 seems to account for that (without getting too in detail about assumptions of which parameter of the non-sinusoidal current the trip unit actually responds to, which can get very murky).

Personally I’m inclined to set it as high as possible, but that is an uphill battle if I don’t have good justification.

What do you think?
Is 4400A high enough?
Should the NEC guidelines be used for this motor?
Should it be considered energy efficienct?
And in that case, we should consider up to 5800 setting?


=====================================
(2B)+(2B)' ?

 

[electricpete]

Personally I'm inclined to set it as high as possible,

.... subject to review of coordination, cable protection etc.

=====================================
(2B)+(2B)' ?

 

[DRWeig]

Hi Pete,

In my experience, yes -- the NEC rules apply to load center breakers feeding motors.

As for suggested setting, I'm also inclined to go as high as possible within the bounds of coordination with upstream stuff.

Good on ya,

Goober Dave

 

[LionelHutz]

The instantaneous is there to trip during a fault. So, I would ensure it's set low enough that it will capture the exected fault current.

Upstream coordination - The level setting doesn't make any difference when the upstream device also has a instantaneous trip. I often see people "coordinating" their instantaneous trip settings and it's a complete and total waste of time. It's a level setting and not a current limit setting. Setting the trip to 3.8kA vs 5.8kA won't have any effect on the actual fault current.

Cable protection - the same comments apply. The setting could change the trip time if you set the trip level higher than the fault current, but it won't change the fault current level.

 

[jraef]

If you read the NEC section and exceptions thoroughly, you will see that the exception allowing the higher settings is valid when demonstrated that lower settings don't hold in under normal circumstances (it may not have that last part quite so clear). The implication for that follows what Lionel just said;, the setting should be as low as possible, with allowance for adjustment when what is possible may be a little higher than what is typical

Personally, I would set it at 10x and only raise it if it nuisance trips.

"Dear future generations: Please accept our apologies. We were rolling drunk on petroleum."
— Kilgore Trout (via Kurt Vonnegut)
For the best use of Eng-Tips, please click here -> faq731-376

 

[electricpete]

Thanks for all the good comments. Interesting there are a variety of approaches. Some would set it as high as practical. Some set it low (10x) and adjust if trips are seen. Probably there are some situational and application differences. In some installation situations, there is plenty of time to adjust setpoint if motor trips during post-installation test.... in other cases installation and return to service are time critical. Further in some applications, spurious trip is not tolerable (it’s more than a nuisance, it’s an equipment functional failure). So knowing the random nature of the peak current, how many times do we have to start it during installation testing in order to have extremely high confidence that it won’t trip during operation (sort of a rhetorical question).

For Lionel – That’s a good point. I understand what you are saying. Attached I have roughly reproduced our existing coordination curves for discussion. For the existing setting (before motor replacement), the upper band of the 4400A motor instantaneous trip is just a hair below the lower band of the 5500A load center tie breaker instantaneous trip.

Now what if we increased our motor instantaneous setting from 4400A to 6000A. Then would now be an overlap. What you’re saying (I believe) is that this is overlap not a problem because we still have margin between 6000A instantaneous setting and 28,000A short circuit current (and motor breaker will clear motor short circuit fault before tie breaker trips). If the protection responds corrrect to normal start (no trip of either breaker) and to motor short circuit (motor trip before load center can respond), then we say protection is good (right?).

For devil’s advocate position: What if we have a fault that draws something like 5,750A. I guess that is not considered credible in the relaying world. But are we being a little presumptuous to assume we can determine minimum fault current with that much confidence. It seems to me that in random wound machine, the endwindings of different phases are adjacent at many different points in the winding which can short out just about any fraction of the winding that you choose, and if you were allowed to pick the location of the short you could generate any fault currrent between LRC and avaialble short circuit current from the supply. Is there a standard that says we don’t have to consider these type of scenarios?


=====================================
(2B)+(2B)' ?

 

[Bramb]

Hi Pete,
I think that, in your "Devil's advocate" position, you mixed two subjects....

1)

set low enough that it will capture the expected fault current

For phase faults it means to consider the minimum fault current: ph-ph fault = 0,866*I_3phmin (I_3phmin calculated with minimum source short circuit power).
For ground faults, if neutral point is solidly connected to ground, I_3phmin should be used considering arc resistance (0,2 reduction factor was suggested in some papers).

The above values are typically calculated in the nodes of the network, not inside the electrical machines (for this kind of fault my idea is expressed in point 2 of the present post).

In other words, in Pete's case the values will be referred to a fault in the cable at the motor terminals.

When you analyze the minimum phase fault current value you can discover that the minimum phase fault current is lower than the setting of instantaneous trip (which have been chosen as lower as possible to let the motor start).

What does it means?
Just that some faults will be cleared by the time delayed stage of the trip unit, not by the instantaneous stage (refer also to my signature.....).
Is it necessary to do something in order to avoid this and to clear the fault faster?
If correct protection of all the upstream network elements is guaranteed by the settings chosen, the answer is no.

2)
About internal faults in electrical machine I make reference to star connected winding (worst case).

For phase to ground faults, the closer the fault is to the star point, the lower the fault current is: it decrease from maximum value (100% if the fault is at the motor terminal) to zero (fault outside the winding, inside the connections which form the star point).

It becomes important in case of High Impedance Grounded systems.
In this case, ground fault protections are typically set maximum to 0,1 p.u. of the expected min ph-g fault current: it means that roughly the 90% of the winding is protected against ph-g faults by ground fault protection. About the 10% remaining I would say that nobody take care of it... It isn't really true as this kind of fault also causes mechanical problems leading to overcurrent phenomena which cause in the long term the electrical trip or at least vibrations which could be controlled by the process.
It also causes unbalance of the currents which causes trip of the negative sequence 46 protection of the motor if foreseen (I'm not sure, but I've also heard about molded case circuit breaker trip units reaction to unbalanced loads).

Ph-g faults not so close to the star point of the motor in solidly grounded systems can be also detected by phase overcurrent protections.

Phase to phase and three phase faults inside an asynchronous motor lead in effect to currents that are lower than the fault at the terminals, but we are always speaking about currents much higher than the motor rated current and we fall again in point 1 of the present post.

My post has grown while I was writing it.... I hope I didn't make mistakes nor expressed unclear concept...
In this case I beg your pardon in advance.

Ciao
Erminio

The difference between overload and short circuit lies in the nature of the fault, not in the value of the current.

 

[electricpete]

The system of interest is ungrounded.

The recent discussion is: Can we increase the instantaneous setting of the motor (4400 in previous attachment) above the current setting of the load center supply breaker, shown as 5500 on the attachment (I'm not sure what to call it... short time?)

The above values are typically calculated in the nodes of the network, not inside the electrical machines (for this kind of fault my idea is expressed in point 2 of the present post).

In other words, in Pete's case the values will be referred to a fault in the cable at the motor terminals.

What I am proposing in a fault inside the machine (shorted turns) which results in current at the machine terminals in the neighborhood of 5750A. In that case, with load center set at 5500A and motor instantaneous increased to 6000, we seem to lose coordination for this fault (load center breaker will trip). Do you think that is a problem, or not a fault worth considering?


=====================================
(2B)+(2B)' ?

 

[electricpete]

What I am proposing in a fault inside the machine (shorted turns)

I should clarify - not your typical shorted turns. Short between different phases of the windings (not necessarily at the line end) in the endwinding.

=====================================
(2B)+(2B)' ?

 

[DRWeig]

Hi again Pete, your analysis skills always amaze me.

Guys like me don't have that deep understanding of what goes on (or may go on) in the windings of a motor, and we're charged with designing protection at such a rapid pace that we don't have time anyway. We typically ignore the possibility of small faults that put the current level in between the locked rotor current and the instantaneous trip setting.

Here's how I was taught by my original mentor:

1. Make sure the conductors are big enough to provide acceptable voltage drop to start the motor.

2. Size instantaneous protection and its initial setting based on the tables in the NEC, it's for faults only -- to protect the conductors, starter, and motor. In your case, I'd do 11 X FLA so that I most likely wouldn't have to be called for trip-on-first-start (11X is permitted by the tables), and I would check coordination with the feeder breaker protecting the load center, adjusting as necessary). I'd make sure the instantaneous was adjustable upward from there a fair ways so that the "must start the motor" exception to the tables can be employed if needed.

3. Specify the running overload protection according to the motor service factor, design code, etc., at 1.15 or 1.25 x FLA.

A small fault that you describe, one that's in between the overload setting and the instantaneous setting, will be cleared by the overloads in short order. Even if it's slightly above the locked rotor value, it'll be OK. Although it physically is a fault, I'd call it an overload since the current is small and limited.

A solid fault will be handled by the instantaneous trip of course. Not getting into ground fault here for brevity...

I've done this often over the years, we have had hospitals with almost exactly the same situation -- 200 to 350 HP / 480/3 chilled water pumps and such.

It could very well be that your hypothetical winding fault is something that needs to be considered by the NFPA in those tables, but I bet it already has. In practice, I have not had any problems with the protection I just described. Of course, my jobs most likely never experienced the small-scale winding fault issue...

Again, them's just my thoughts and experience.

Double good on ya for the mental exercise, I needed it this week.

Goober Dave

 

[LionelHutz]

I mean a trip curve that goes straight down to 0 time when I posted instantaneous. You are showing what I would call short time trips and you do have them co-ordinated.

If both breaker trip curves went straight down to 0 time at 4.4kA and 5.5kA, then both breakers would trip on a fault of 28kA.

You show a delay time on the motor breaker. In that case, you might not have to worry about the first cycle inrush as much since there is a time delay to ride through it.

With your theoretical question, you could increase the trip level of the motor breaker to be higher than the upstream but there would be no point. One of the breakers would trip and you don't want to trip the main on a motor fault.

 

[electricpete]

Thanks Dave. I didn’t mean to get carried away trying to create a scenario. I am not really interested in answering my “what if” question as much as I am interested in understanding “standard practice” if there is such a thing.

Even if it's slightly above the locked rotor value, it'll be OK. Although it physically is a fault, I'd call it an overload since the current is small and limited.

... but my question was not really about clearing the fault to limit damage, it was about clearing the fault with the motor breaker in order to avoid tripping the load center supply (or tie) breaker.... in the example cited, overload does nothing to prevent load center breaker from tripping.

I've done this often over the years, we have had hospitals with almost exactly the same situation -- 200 to 350 HP / 480/3 chilled water pumps and such.

It could very well be that your hypothetical winding fault is something that needs to be considered by the NFPA in those tables, but I bet it already has. In practice, I have not had any problems with the protection I just described. Of course, my jobs most likely never experienced the small-scale winding fault issue...

Just to clarify what you and Lionel are saying is as follows (paraphrased): It is standard practice that we do NOT need to consider “faults” drawing an intermediate current between LRC and short circuit current at the motor terminals for coordination. i.e. in my example we can set the motor instatnaneous to 6000A even though the load center tie (acts like supply) breaker is set at 5500A. Correct?

(I can see some logic to that approach, I also remember hearing it various places before.... just double checking that it is a valid and acceptable approach).


=====================================
(2B)+(2B)' ?

 

[electricpete]

Lionel

Sorry, I skipped to the latest response and overlooked your earlier one. Your input is very valuable to me on this (I’m glad you got me started looking at coordination from the standpoint of times rather than currents).

You are showing what I would call short time trips and you do have them co-ordinated.

If both breaker trip curves went straight down to 0 time at 4.4kA and 5.5kA, then both breakers would trip on a fault of 28kA.

You show a delay time on the motor breaker. In that case, you might not have to worry about the first cycle inrush as much since there is a time delay to ride through it.

Sorry, my previous slide had somewhat of an error with regard to motor breaker. The correct existing settings are as shown in slide 1 and 2 of this post. Curve 3 is tie/supply breaker in red and Curve 4 is motor breaker in blue. The tie/supply breaker definitely has a built in delay for coordination, which causes the curve to turn horizontal (constant time delay) for currents above the setpoint. When I look at motor curve closer, I see the line representing right half of the tolerance band for motor breaker turns horizontal, but the line representing the left half of the toelrance band goes straight down. So I think the motor breaker is a true instantaneous with no intentional delay, and the right half of the curve is intended to show the max time to clear the fault (useful for coordination).

Slide 3 is a possible change (increase motor instantaneous to 6000A) to the settings to for installation of new high efficiency motor, in order to reduce concerns of spurious trip during start (either during post-installation testing or subsequent operation).

Do you think this slide 3 represents reasonable coordination between motor breaker and tie breaker ? (it is not coordinated for fault of approx 5575, but that type of fault above starting current but far below short circuit current is not credible, right?)

With your theoretical question, you could increase the trip level of the motor breaker to be higher than the upstream but there would be no point. One of the breakers would trip and you don't want to trip the main on a motor fault.

The reason for the change would be to prevent spurious trip during motor start.


=====================================
(2B)+(2B)' ?

 

[electricpete]

ignore the part of slide 3 which shows the time overcurrrent curve shifting to right. I intended of course only to shift the instantanous to the right.

=====================================
(2B)+(2B)' ?

 

[electricpete]

Attached corrects the previously-mentioned error, as well as colors of labels in slide 1

=====================================
(2B)+(2B)' ?

 

[DRWeig]

Hi Pete,

Yes, you got it right -- but what confuses me is that you have a load center breaker followed by a motor (instantaneous-only?) breaker. Isn't there nothing in between them but some cable? You should only need one.

Not sure about the model numbers you quoted, but normally I would size the load center breaker per NEC 430, then only have a disconnect switch at the starter (or at the variable speed drive, etc...). No need to have another instantaneous trip device at the motor controller if it's already protected by one upstream. You can often get a magnetic-only motor breaker to fit in a load center, but we normally (in my shop) use a standard thermal-magnetic since the thermal part of the curve is slow and the motor controller's overloads are set well below it.

That way you don't have two breakers with similar (or identical) trip settings. The only coordination you have is from the load center breaker backward to whatever device feeds that load center.

Wish I had a one-line to share for more clarity, but I'm hanging in a most wonderful Korean coffee shop instead of at my desk.

Let me know if I'm missing something -- the old brain is tired today.

Good on ya again,

Goober Dave

 

[electricpete]

Yes, you got it right –

Thanks for confirming that.

but what confuses me is that you have a load center breaker followed by a motor (instantaneous-only?) breaker. Isn't there nothing in between them but some cable? You should only need one.

It is part of a double-ended load center (all buswork between these breakers). This portion of the load center can be supplied directly from transformer via supply breaker (2) or from the other half of load center via tie breaker (3). Additionally, we can feed the other half of the load center with both 2 and 3 closed... which is why the settings of 2 are lower than 3 (to avoid tripping this half of load center due to fault in other half in this cross-tied configuration). There are other motor loads besides this 300hp motor fed from this portion of the load center... not shown on this diagram.

Not sure about the model numbers you quoted, but normally I would size the load center breaker per NEC 430, then only have a disconnect switch at the starter (or at the variable speed drive, etc...). No need to have another instantaneous trip device at the motor controller if it's already protected by one upstream. You can often get a magnetic-only motor breaker to fit in a load center, but we normally (in my shop) use a standard thermal-magnetic since the thermal part of the curve is slow and the motor controller's overloads are set well below it.

There is no starter or controller. Just a breaker which closed or opened to switch the motor on and off.... and includes the both the time-overcurrent protection (relied upon for overload protection) and the instantaneous protection. The motors do not cycle on/off frequently.


=====================================
(2B)+(2B)' ?

 

[Bramb]

Hi Pete,

What I am proposing in a fault inside the machine (shorted turns) which results in current at the machine terminals in the neighborhood of 5750A. In that case, with load center set at 5500A and motor instantaneous increased to 6000, we seem to lose coordination for this fault (load center breaker will trip). Do you think that is a problem, or not a fault worth considering?

I think it is problem to loose coordination between incomer or tie and motor feeder, as the portion of the load center feeds more than one user (motor).

I consider the neighborhood of 5750 A a possible fault value, so I would increase current setting of both the incomer and the tie breaker high set (or short time) in order to obtain full coordination.

This as per my experience in European industrial plants protection relay coordination. Also seen in some guideline of a main Oil & Gas Corporation.

Ciao
Erminio

The difference between overload and short circuit lies in the nature of the fault, not in the value of the current.

 

[LionelHutz]

I would not overlap the curves and lose co-ordination in that region unless there was no other way around it. And there is almost always another way.

 

[electricpete]

Thanks for clarifying Lionel, I had misunderstood your comment.

The goal is to increase motor instantaneous setting to avoid possible spurious trip. If the tie breaker currrent setting is increased to coordinate based on current, then the supply breaker must be increased and also settings further upstream must be reviewed. Might be possible, but unknown at this point.

Regarding the proposed change shown in slide 3 of my previous attachment (3 Nov 11 15:41): where motor instantaneous is increased to 6000 while leaving tie/supply at 5500 (with time delay that will coordinate for expected fault currents which are well above 6000)...

Dave agrees with this proposed change.
Lionel and Erminio do not agree with this proposed change.

Any further comments... or anyone else wants to weigh in on this proposed change?

=====================================
(2B)+(2B)' ?