One CT Phase Reading Significantly Lower 1

[mesutphen]

While reviewing the motor starting data for a 1750HP, 7.2KV motor, we noticed that B phase starting current was significantly lower (about 50%) than A and C phases. This is not something that has been known to happen in the past. The last 3 starts had the following voltages and currents:

Start 1: Ia - 958 Va - 6766 Ib - 450 Vb - 6797 Ic - 967 Vc - 6769
Start 2: Ia - 1032 Va - 6752 Ib - 583 Vb - 6777 Ic - 1086 Vc - 6760
Start 3: Ia - 1102 Va - 6748 Ib - 605 Vb - 6783 Ic - 1130 Vc - 6755

While running, all 3 phases are fairly balanced, Va - 6755 Ia - 98A, Vb - 6729 Ib - 93A, Vc - 6755 Ic 96A. We injected current from 30-130A into the CT and the indication was 27A @ 30A injected - 125A @ 132A injected, so low-range accuracy seems to be okay. I'm trying to think what would cause the B phase CT to read so much lower. The 5% unbalance while running seems to indicate the motor is OK. Is it possible that the CT is becoming saturated at higher currents? The CT is about 30 years old but I haven't heard of a transformer shifting its saturation point before...

Thank you.

 

[electricpete]

I guess if you come down to it, there is nothing that satisfies everything you told us. Seems highly doubtful that a relay problem would only show up during start and not run (unless for some reason there is a setting that controls how that inrush parameter is calculated on an individual phase basis), and also you told us relays were swapped and stayed with same phases.

So, I guess it is a need to be methodical and recheck the facts that are presented and re-examine what was ruled out, because we seem to have ruled everything out. Perhaps argues to going back for more testing if that is easily done.

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

 

[jghrist]

It is normal for transformer inrush current to be high on two phases and low on the third. Is this also the case with motors?

 

[electricpete]

Not for motors (other than random variation in dc offset which can create 2 high and one low).

For a transformer, the middle phase sees a different reluctance than the other two phases. If you look at the transformer core (3 leg or 5 leg) you can see the asymmetry. For a motor, the core is round... and all three stator phases look identical to each other, just shifted from each other in position by 2/3 of a pole pitch around the stator.

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

 

[dpc]

It's good to distinguish between motor inrush and locked rotor current during motor starting - they are not quite the same thing Inrush includes the locked rotor current PLUS the brief period of dc offset current which will vary depending on the phase angle of the voltage at the instant the contacts close to energize the winding. This can be asymmetrical in one or more phases. The locked rotor current is the symmetrical value of current after the initial inrush has decayed.

I have no idea what the OP's actual problem is, but inrush is a significant issue when setting motor short circuit protection and is quite real. Inrush is generally higher on newer high efficiency motors.

David Castor

http://www.cvoes.com

 

[electricpete]

http://www.gedigitalenergy.com/products/manuals/469/469man-dh.pdf

Page A-1 (pdf page 229/248) describes means to use 2 CT's to sense 3-phase current. There is also mention that polarity error can cause sqrt(3) error. Do you use something like this by any chance?

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

 

[electricpete]

Actually that sqrt(3) error is sqrt(3) high, can't arrange the vectors to create sqrt(3) low.

* There are 2 connections per CT channel... did you swap both of them (including the wye/neutral side connection)? If not, could perhaps have high resistance somewhere in that path on the neutral side (doesn't seem likely, but it would seem about the only thing that hasn't been ruled out).

Looking at your A and C phase currents, they might be a little high to be LRC given that voltage is below nameplate, but they also seem way too low to be peak instantaneous current (in fact peak instantaneous is always at least sqrt2 times LRC even if you don't have any dc, right?). I see in the manual they have some kind of "learned starting current" which is measured 200 msec after start, with the intention that the dc component has died about by then. Is that what you're reporting?

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

 

[mesutphen]

I believe the starting currents we are reading are peak inrush. The manual I have doesn't really describe it. It is the current in the Event History. We are definitely not reading the "learned starting current" because that's in the Learned History.

There was a portion of the B-phase wiring going to a separate building to drive a current transducer. We shorted that wiring and starting currents are now within 10A, and running currents are within 1-2A (used to be within 5-7A). Apparently our problem is CT burdening.

Considering we swapped the A and B input leads into the relay, I am guessing the transducer is between the relay output and the common ground. This is the only thing that explains to me why the problem did not move from B-indicated to A-indicated when we swapped leads.

 

[electricpete]

So, when you swapped A and B input leads, did you swap one lead per channel or 2 leads per channel?

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

 

[mesutphen]

We swapped one lead per channel. On the Multilin, CT's are on G6 (A-phase), G7 (B-phase), and G8 (C-phase), with common ground on H6, H7, H8. We swapped leads on G6 and G7, but left the leads on H6 and H7 since it should have been commonly grounded.

 

[electricpete]

The neutral wiring is not checked if you just swap one lead per channel and in theory a high resistance on the neutral could cause everything seen. I said it was a longshot, because I assume the neutral connections are something like short jumpers between H6/H7/H8, with extra lead landed on one that connects to the long lead that provide interconnection to the ct neutral. But sometimes the real world connections and daisy chains are not quite what we expect/assume. So it would be useful to do as thorough a check on the wiring as possible. Ideally you'd like to watch with your own eyes as both leads per channel are swapped, and examine closely whether any other leads are landed on these two pairs of terminals and see if they all make sense.

The known longer run of wire on B phase seems to explain everything EXCEPT the results of the lead swap on the relay input... the problem should've shown on A phase indicated (B phase actual) if that long run were the culprit, wouldn't it?

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

 

[electricpete]

Considering we swapped the A and B input leads into the relay, I am guessing the transducer is between the relay output and the common ground. This is the only thing that explains to me why the problem did not move from B-indicated to A-indicated when we swapped leads.

By common ground I think you mean wye point? I'm not clear exactly what transducer you're referring to. But if there is a difference on the neutral side, it would show up in swapping both leads per channel (vs only one lead per channel).

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

 

[electricpete]

At any rate, if you are suspecting or investigating that the neutral side has some wiring differences between A phase and B phase, then we are on the same page.

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

 

[rasevskii]

for mesutphen:

If it were really connected like the dwg that you posted everything should be ok. Now maybe time to get in and rewire exactly to that dwg, rather than endlessly swapping connections. That means unbinding cable ties and tearing out the existing possibly. There is probably some sneak circuit or unwanted ground that cannot be seen otherwise. The remote circuit on phase B that you mention is not even shown there.

On the Multilin. Is that low reading possibly a stored value?

Are there any actual ammeters in the panel that show also a low reading on phase B?

BTW, don't assume anything. Check it yourself personally.

Just some advice from field experience.

rasevskii

 

[rasevskii]

mesutphen:

On further thought you may have a bad CT on phase B. If the above checks/rewiring find no fault, then a mag curve on each of the 3 CTs may be done. Inject on the secondary (primary and secondary circuits lifted) an AC voltage and read the mA magnetizing current at about 10% steps until the CT begins to saturate (knee-point)(when a 10% V increase gives a 20% mA increase, that is the beginning of the saturation or knee-point) You will need a variac and step-up transformer for this, up to around 250...350V will be needed (or more).

On all 3 CTs the curve should be the same. If phase B CT is bad, it will show quite higher current than the others, or at least quite different.

On such a large and expensive motor, maybe ordering a complete new set of CTs would be the best answer. Otherwise how can you trust the protections?

rasevskii