Imbalanced current on 3 phase system 1

[cjhut]

I have a 3 phase 480v 2000a main lug only switchboard being fed from a wye secondary of a 1500KVA three phase oil filled transformer. The primary voltage is 2.3kv. There are three fusible disconnects on the switchboard feeding three soft starters each driving a three phase squirrel cage motor. The switchboard is about 6ft away from the transformer secondary and is fed by parallelled conductors in wire trough conecting to the 2000amp bus.The parallel conductors all appear to be of the same length,termination,etc. The soft starts are fed from the disconnects on the switchboard through cable tray using 1 single conductor per phase.Everything is relatively new including the transformer. I am having a noticable differnce in current on all three phases on all three of these motors. Example: if I measure the voltage with a load I read phaseA-B volts 480, phase B-c volts 472, and phase A-C 475, each phase to ground around 277. So my voltage looks good. What I am having trouble understanding is, my current readings look like this A-360 B-320 C-311amps.These readings were taken from a 300 HP motor running a screw compressor that was about 85% loaded from an air demand standpoint. The FLA of this motor is 345amps, so A phase is running above that.I am seeing this same realtionship on the two other identical machine/motors. This problem has been here since
the installation, so this is nothing new. I was wondering if this might have to do with the impedence of the windings of the transformer being different? We are not using 3 single phase transformers it is one unit. It is causing me some problems because if the machines have to run at full load for an extended period of time the soft starter will trip out on phase A overload. I understand the voltage is slightly different from phase to phase, but the difference between volts/amps isn't proportional. If impedence is an issue how can I prove it?

 

[cranky108]

Did you look at the primary voltage on the transformer?

 

[waross]

I understand the voltage is slightly different from phase to phase, but the difference between volts/amps isn't proportional. If impedence is an issue how can I prove it?

You have not told us what the voltage unbalance is.
Be aware that an induction motor develops a balanced back EMF or in other words acts as an induction generator. The back EMF is balanced both as to voltage and phase relationship. Any phase error or voltage unbalance in the supply will cause an unbalanced current to flow which is disproportionate to the voltage unbalance.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[jghrist]

I was wondering if this might have to do with the impedence of the windings of the transformer being different? We are not using 3 single phase transformers it is one unit.

The winding impedances should be identical, although per phase impedances are not something that is tested. Any phase impedance difference would show up as a voltage difference, which you don't have.

Were your voltages measured at the motors or at the transformer? Are the two motors at the end of the same feeder or different feeders?

 

[cjhut]

I took some readings on the primary side today and here's what I read volts A-B 2251, volts B-C 2270, volts C-A 2280. I watched it for a while and the voltage is never more than 30 volts difference across any of the phases. SOmetimes it is as little as 10volts. I cannot guarentee the phasing on these readings to be phased with the above secondary readings due to the fact that I have no voltmeters on the switchboard or transformer, these readings were taken from the bus which feeds the primary, but not the actual phase conductors going to the primary. Would this slight diffence in voltage cause such a current unbalance?

 

[cjhut]

The 480volts/current readings were taken from the load sides of the soft starters. There is no voltage/current difference between the line and load side of the soft starts. Each motor has it's own feed coming from the switchboard-single conductors per phase in steel tray. No parallel conductors, and nothing run more than about 50 ft from the switchboard. So to answer your question jghrist, I took the readings at the motor leads.

 

[waross]

Said a different way:
Voltage unbalance in motor feeds will cause a disproportionate unbalance in motor currents.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[PHovnanian]

phaseA-B volts 480, phase B-c volts 472, and phase A-C 475, each phase to ground around 277

Not bad. The worst phase to phase voltage is about 0.9% high. But if each phase is 277 V L-N. the angles don't quite add up.

Are there any capacitors on the 480V system anywhere?

 

[cjhut]

When I gave the 277v reading to ground, I was pretty much saying that as a nominal voltage, It did vary a little from phase to phase, but just a volt or two. I was just trying to let everyone know that we have a properly grounded wye, with no floating ground.If it matters I can get some exact readings. There are no capacitors connected to this circuit. I apologize if my info is lacking some initial details, but this is something that is getting over my head. I am an electrician, not an engineer. I love troublshooting, and I am proficient at troubleshooting circuits down to a component level. But power quality isn't my cup of tea. I appreciate everyone's input, WARCROSS: do you think that this slight difference in voltage could cause my problems. Here is the real issue, if these machines run to full load (they are compressors)we run above the FLA on Aphase--past the service factor and the soft start shuts the motor down. Looking at the bigger picture, what can I do to correct this problem? If this was a smaller circuit the fix could be cheap, but because this is coming from a substation, I would imagine that if we try to correct the power quality, it could get expensive. Let me know your toughts.

 

[cunin]

The info you are getting on this site is right on with your problem. The next step I would do is infered scan your system,with a motor load on it. Scan everything motors, switchgear, transformers, cable connections, enclosed and overhead. Look for hot spots that will give you a voltage drop under load.

 

[Kiribanda]

cjhut,
Just a thought.Did you use a true RMS clamp meter to measure the given currents on the load side of the soft starter?It looks to me that the output currents of soft starter contains some harmonics therby giving a higher true RMS.So if your soft starter does not have thermal type overload relays then they would definitely trip depending on the setting.Otherwise if your soft starter has electronic type overload relays which might work on 60Hz wave then it may not trip.Therefore,to eliminate this type of situation is ther anyway that you can start and run one motor by-passing its soft starter?

 

[electricpete]

The negative sequence as fraction of positive sequence can be roughly estimate from "unbalance" defined as max deviation from average over average. (under some assumptions, such as summing to zero).

For your voltages:
480, 472, 475
Average = 475.7
Max deviation from average = 4.33
Neg sequence fraction estimate =4.33/475 ~ 0.9%

For your currents:360, 320, 311
Average = 330.3
Max deviation from average = 29.7
Neg sequence fraction estimate = 29.7/330.3 ~ 9%.

So, the current unbalance (expressed as fraction of fundamental) is roughly 10* as much as voltage unbalance (expressed as fraction of fundamental).

Let 1, 2, 0 refer to positive, neg and zero seq.
V2/V1 = 9%
I2/I1~0.9%

(I2/I1)/(V2/VI) = 9%/0.9%~10.

What do we expect this ratio to be:
(I2/I1)/(V2/VI) = (I2/V2)/(V1/V2) = (1/Z2)/(1/Z1) = Z1/Z2

What do we expect the impedance ratio's to be. As a first estimate, we say negative sequence impedance is rougly the same as locked rotor impedance (there are good physics arguments to make this approximation since both conditions correspond to very high slip). What is ratio positive sequence? It is inverse of ratio locked rotor current to full load current which we know is around 7.

If we knew only the voltage unbalance and you asked me to predict the current unbalance, I would say 7 times higher. You actually measured 10 times higher. It is not such a big deviation (considering approximations in estimating unbalance and approximations in substituting locked rotor impedance for neg seq impedance). I would suspect that the current unbalance you are seeing is simply a result of the voltage unbalance you are seeing.



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

 

[electricpete]

I left out a few words:

"What is ratio positive sequence? It is inverse of ratio locked rotor current to full load current which we know is around 7."

should have been

"What is ratio positive sequence impedance to negative sequence impedance~ locked rotor impedance? It is inverse of ratio locked rotor current to full load current which we know is around 7."


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

 

[electricpete]

Sorry, one more typo (probably obvious, but I'll correct it anyway):
"What do we expect this ratio to beI2/I1)/(V2/VI) = (I2/V2)/(V1/V2) = (1/Z2)/(1/Z1) = Z1/Z2"
should have been:
"What do we expect this ratio to beI2/I1)/(V2/VI) = (I2/V2)/(I1/V1) = (1/Z2)/(1/Z1) = Z1/Z2"


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

 

[electricpete]

By the way, what I said is just more detail on what Bill (waross) said earlier: 11 Nov 11 7:58


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

 

[electricpete]

If you repeat the unbalnce calc using primary side measurements, it is 0.7%. Not too far from the secondary 0.9%. So it appears all of this is expected result of primary side unbalance.

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

 

[waross]

Is the voltage unbalance higher with no motors running?
Do the voltage unbalances and the current unbalances drop as more motors are added?

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Masbibe]

electricpete,

It will be great if you can explain from where you get this approximation:

"The negative sequence as fraction of positive sequence can be roughly estimate from "unbalance" defined as max deviation from average over average".

Thanks in advance,

Milovan Milosevic

 

[electricpete]

It will be great if you can explain from where you get this approximation:"The negative sequence as fraction of positive sequence can be roughly estimate from "unbalance" defined as max deviation from average over average".

I didn't invent that. It is used in NEMA MG-1 as the basis for derating.

Here is a link.

http://users.encs.concordia.ca/~pillay/16.pdf

Note he calls the "true unbalance" to be V+/V-, and compares it to the NEMA maximum deviation from average in the figures. The agreement is good for small unbalances and generally starts to diverge as the true unbalance gets larger.

Also attached is a spreadsheet (requires analysis tookpak installed) which compares "true unbalance" = V+/V-, maximum deviation from average, and another formulation given by GE.

For the spreadsheet, you input the three magnitudes and the spreadsheet figures out the angles under the assumption that the three vectors sum to zero (which holds for currents flowing into a motor and also for phase to phase voltages).

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

 

[Masbibe]

thanks electricpete,

it will be helpfull.



Milovan Milosevic