Unbalanced Currents during Generator short circuit test

[edison123]

During a short circuit test on a brand new 100 MW hydro generator, at each excitation level, the phase currents were found to be different (7098 / 7042 / 7056 A; 6790 / 6734 / 6748 A down to 602 / 560 / 574 A). The ammeters used for this test are calibrated and checked for the same reading by connecting them in series and passing a known current. ( I even switched around the ammeters and still found the above pattern of the differing phase currents). Also, the CT’s were brand new from reputed manufacturer and had been calibrated prior to SC test with primary injection. What could be the reason for this consistent different phase currents under short circuit ? Does the fact that the stator winding is fractional slot winding has anything do with it and if so, how ? Has anybody else seen the same phenomenon ?

 

[jwerthman]

It's not unusual to see some unbalance in the currents during this type of test. It is frequently due to unbalances in the shorting connections. At high current levels, the impedance of the shorting connections is a significant portion of the total circuit impedance and it is difficult to keep the connections balanced. It is common to apply two straps - one between T1-T2 and a second between T2-T3. This is not a balanced connection and will result in unbalanced currents. Differences in the impedance of bolted joint connections can also cause the same.

 

[jbartos]

Suggestion: What was a manufacturing tolerance specified to the generator? -0%,+0% of nameplate parameters?

 

[edison123]

Thanks jwerthman. A single heavy short was directly across all the 3 terminals of leads (neutrals are already shorted in similar fashion. Each phase lead is bolted to the single short with 6 nos. M20 bolts. So, that would rule out any impedance drop across the shorts on both ends of the winding. Could uneven airgaps play a part though in this case the airgap variation was within 3% ?

jb, could you pls elaborate a little on your question ?

 

[buzzp]

I would think this unbalance could be from the impedance differences in the windings. No expert on this type of testing but just a thought.

 

[rbulsara]

I thought same as buzzp but did not want to be the first respondent as I am not an expert in large generators.

The percentage difference is barely 0.8% at higher end which is probably within winding impedance tolerance.

 

[RAMConsult]

I have to go with jwerthman, The first place I would look is the connection of the three phases to the short circuit point. This is really a low voltage, high current connection with tremendous physical forces and heating which can affect the contact resistance.

It is still unclear to me how the short circuit is formed. Are the phases held on one bolt or are they attached to a piece of busbar? If bolted together with a single bolt my guess is phase B is in the middle. If busbar then my guess is that the leads are attached in a linear manner, again phase B is in the middle. Try rearranging the order of your connections to see if the imbalance shifts from phase B.

The ideal short would involve a triangular piece of copper with each phase in a corner, held in place by a bolt, with a belleville washer and copper washer pair on both sides of the connection, all with identical torque on each nut. The goals are to create a symmetrical path and to ensure that the contact resistance is consistent between phases.

You are wise to investigate the source of the apparent imbalance since a machine of this size should be as symmetric as possible. It may be necessary to do an internal inspection where you may find that a coil has shifted or that an interwinding connection needs to be resoldered.

 

[jbartos]

Suggestion: Manufacturing tolerances to the generator Xd" are specified in a range -10% to +15%, for example.
Visit and check with:

http://www.dieselgenerators.com/marinep.htm

etc.

 

[jbartos]

Suggestion: Tolerances of meters may be about 0.5% for voltage and current each. Quality inspection may be with 10% accuracy for Xd", etc.

 

[electricpete]

I'm not familiar with the test.

It seems on the edge and tough to call. It seems like you have 0.8% deviation max-min and approx 0.5% max deviation from average. That is about as tight as any measurement can be expected that I am familiar with: for example winding resistance, motor current balance etc. But then again winding resistance is a low signal measurement which may be susceptible to some othe errors. Motor current balance definitely influenced by voltage.

In fractional slot winding as you know not all pole phase groups produce equal voltage but all phases do.

I'm curious how closely your ct primary injection tests agreed and whether you are operating at higher current in the test than what the ct's were measured at.

 

[edison123]

RAMConsult,

All the three phase leads and neutral leads of the generator are in side by side and in a single plane. A single bar, bolted (6 nos. M 20 bolts for each phase) to all the three phase leads, shorts all the three leads on the phase. The generator is brand new and the stator bars are mechanically braced in a very rigid manner.


pete,

Open circuit characteristic (OCC) and short circuit characteristic (SCC) tests are routine tests done on new and rewound generators to check parameters like synchronous impedance, short circuit ratio etc.

In this case, all the CT’s are 7000 / 5 A. Though the unbalance in current is not worrisome, I was wondering what would lead to such a consistent unbalance currents. Since the working flux during short circuit is below normal, the question of saturation and related harmonics do not come into the scene.


jb,

As I said, prior to this test, all the three ammeters were connected in series and they all showed exactly the same value for a particular current. Also, the meters were interchanged and the unbalance problem did not travel with the meters.

 

[ScottyUK]

Edison,

I agree with RAMConsult's theory about busbar jointing and path lengths. Have you tried testing the joint resistances? You won't need much additional resistance to account for the error you are observing.

I've not found a reasonably portable accurate instrument to measure such low resistances with, say 1% accuracy. We use a 12V 100A DC power supply through a resistor to deliver current through the joint and measure the voltdrop using a laboratory bench meter with averaging and 100nV resolution (Keithley type 2000), because our 200A T&R Ductor doesn't have the resolution required. Even then we have to reverse the voltage connections to eliminate the thermal EMF's from the probes.

 

[edison123]

Scotty,

When compared to stator winding resistance (which in this case was about 5 milliohms per phase as per design and on acutal measurements), how much effect can the shorting connnection bar to stator leads joint resistance really have ? As you say, this joint resistance is so low that you are unable to measure with a 200 A ductor, with which I can measure up to one micro-ohm (not that I have actually measured one micro-ohm in my real life).

I was leaning towards non-uniform air gaps around the machine leading unbalanced flux and resultant differing voltags and currents in each phase. What say you ?

 

[ScottyUK]

Hi Edison,

I'm certainly not going to discount your theory - someone from a generator design background might wish to comment on the likelihood.

We've measured joint resistances typically about 6.0uOhm on our 170MVA sets, but occasionally we find one that is as high as 20uOhm. I think this sort of difference in joint resistance would account for the 1% or so mismatch in your high current measurements if the stator resistance is 5mOhm.

Did you check the joints at the neutral point as well as the line ends?

The outer phases are spanned by two 'long' paths - one at the line ends and one at the neutral - between them, while the centre phase has a short path to either of the outer phases. This might contribute to the mismatch too, although if I was betting a round of drinks on the cause I would back joint resistance being the cause.

 

[edison123]

Hi Scotty,

Since the shorts at the both line and neutral ends were tightened with 6 nos. M 20 bolts per phase, it didn't occur to me check the contact resistance on these shorts. Now, it is too late for the machine is running.

Kumar

 

[buzzp]

Not being a generator expert, based on all the posts, I am still sticking with my original suggestion that this is just impedance differences in the windings. One test should of been to measure the resistance with a micro-ohmmeter measure the simple resistance. These numbers would likely reflect differences that you are seeing on your short circuit test.