Grounded Stator Winding Sychronous Generator 3

[aolalde]

In a cruise ship, 5 identical synchronous generators are connected in parallel sharing the total load.
One generator failed to ground. Preliminary testing, with Megger and feeding AC to the fault shows one grounded spot on phase 2,

Generator Name Plate Data: 14000 kVA, 6600 V, 1224.7 Amps, 60 Hz, 514.3 rpm, ODP, PF 0.80
Stator Winding: 126 slots, 126 coils lap, 42 groups of 3 coil each, 7 circuits Wye
TPC 21, coil span 1-7, six terminal leads.

Since removal of this generator from the ship is very expensive and time consuming, the owner wants to clear the ground and work the generator with the coil or coils grounded cut out of the winding (capped).

Is this possible? Assuming a single coil is grounded what is the best way to clear that ground fault?
Should it develop dangerous internal circulating currents?

Have any one solved a similar problem? Many thanks in advance for your comments.

 

[electricpete]

I think what you are describing is a cutout coil repair. The faulty coil is removed from the circuit and the associated connections are reinsulated.

It is fairly common as a short-term repair for motors. In that case you take a look at expected current unbalance and consider whether derating is required. Also I think in some cases it may cause vibration.

For generators in parallel, you have the added consideration that unbalanced voltage can create circulating currents which may require further derating.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

Sorry aolalde, I posted quickly and I didn't see that you were the original poster. Of course I know you are already very familiar with cutout coil repair and the all of the items I mentioned.

I do have some literature on this topic. I'll send you an email.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[aolalde]

Thanks Electricpete.
We are thinking of taking out the whole series of six coils on that circuit. Since this being a generator I am afraid the uneven voltage connected in parallel with the remaining six circuits will generate a large circulating current due to the low internal impedance of the winding.

Then we will think about the magnetic un-balance created when the circuit is removed. As you mentioned I have done this in motors but not in generators.

 

[petronila]

Aolalde,

Like Electricpete sayds, the first to take account when you capped a Coil is the Current umbalance in the faulty phase, for it is necessary to cut one coil in the others two phases in order to maintain the balance, you need the winding draw to analize the right point to do this(120 electrical degrees aparted)then you will have a winding balanced.Another important think is tp cut in halves each capped coil in order to avoid undesired generated voltages and current circulations.

The derrating or not will be product of wire CMA analysis.With quick Calculations 6600 V/1.732= 3810 Volts per phase. for 1224.7 Amps and 7 Wyes we have seven paralell circuits per phase and each circuit will drain: 175 Amps (22.11 Ohms per path) for 6 Coils in each Paralell Circuit each coil will have 3.685 Ohms per coil With out one coil the path resistance will be : 18.425 and the current by this path : 210 Amps(20% More).

Regards

Petronila

 

[aolalde]

Thanks Petronila.

What you are describing is the impedance of a load with seven parallel circuits, not the generator internal impedance. If the winding internal impedance of one branch should had the proposed 22.11 Ohms, the internal voltage drop when 175 amps are provided will be 3.85*175 = 3868 Volts! Almost 100% Voltage drop!
Since we have a winding with seven parallel circuits, to balance all phases, 21 coils need to be capped, one per circuit per phase. The risk is to saturate the magnetic circuit and/or overheat the field winding, since the voltage has to match the bus voltage for parallel operation with the other 4 generators.

 

[aolalde]

To get the internal synchronous impedance of a generator at least the “No load saturation” and the “Short Circuit” curves are required.

 

[edison123]

aolalde

I have bypassed coils in motors but not (so far) in generators exactly for the reasons you have stated.

Have you tried to contact EASA ?

 

[edison123]

Operation with coils cut out is discussed in IEEE Std 492 under section 7.6 - Unusual or dangerous operation conditions (!!!)

 

[aolalde]

Edison123, thank you for the recommendation, I will get IEEE Std. 492 to study that part.
To add some information; the winding has 6 rings (U1, V1, W1, U2, V2, W2); one per line and each ring has 7 evenly spaced connections to one end of two groups of three coils. The six lines are external for differential protection, three are power lines and three are to form the neutral.
The ground has been located in one group of phase W.
We have to develop a repair plan for the next port arrival of the ship and our technicians will go in the trip. The idea is to test the generator on line after it is modified by whatever agreed solution that is found. The parallel synchronization and load tests of the modified generator will be made by the time they reach another port. The implications of safety and money are obviously huge.

 

[181273]

Hello here´s Jose from Indar Maquinas Electricas(Spain) and I´d seen this problem sometimes.... I solve in the next way.. It´s pòsible to work with one ground point in the machine... if you isolate the star point, and there´s no currents, but the problem is when you got a new ground fault you will get arc between the core and the cooper and it´ll destroy your winding.

This is very dangerous. In other way to isolate (disconect) a part of the winding like in DC rotors, is not posible cause you won´t get the same voltage internal on the machine as you got in the other machines resulting in a excesive flow of reactive power...

This all i said is without knowing the electrical thing of the power plant... for me the best solution is to rewind the generator or try to isolate the winding on site, I had done it a lot of times.... no VPI but it work for a long time... as I said you I´d done a lot of time....

Maybe If you feed a transformer, is very dificult to get on primary a secondary ground, and it´s a little safe to work with one ground point if you got the star open to ground... in any time and the damage will be greater.

 

[electricpete]

Since we have a winding with seven parallel circuits, to balance all phases, 21 coils need to be capped, one per circuit per phase.

I can see where that gives you a balanced winding electrically: 7 circuit wye, now with 5 coils per leg. Each leg of a wye contains two pole-phase groups: one has two coils and the other has three coils.

Also since one half of the total groups have 2 coils/gp and the other half have 3 coils per group, you should be able to arrange the cut coils physically so that one coil is cut out of every other group around the whole machine (2/3/2/3/2/3 etc).

One interesting thing, it seems this machine has a very small coil pitch: 1-6. This is a 14-pole machine with 126 slots, so one full pole pitch would be 1-10, right? This machine I think has a coil pitch of 5/9 of full pitch {(6-1)/(10-1)}. That's the lowest I've ever heard of. Does that sound right to you?

It does seem you should be prepared for possible noise and vibration. Fractional slot machines are notoriously noisy and you have created something similar to a fractional slot machine.

Flux density as you mentioned should go up by a factor of 7/6, which can heat the core or cause exciting harmonics. Just to mention a remote possibility - maybe the system can tolerate operating at slightly reduced voltage or slightly increased frequency to reduce this overexcitation? I'm guessing that would be a tough evaluation requiring someone familiar with the ship's electrical system and electrical loads.


=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[waross]

A comment on over excitation.
Excitation varies as the load varies. Also the reactive power is dependant on excitation.
I suggest limiting excitation current to the maximum rated excitation current.
I would expect two issues.
1> The excitation should be adequate for lower levels of output power. Maximum excitation is only required at maximum output.
2> With a parallel machine if the excitation is insufficient the reactive power will drop. This will through some VARs over onto the other machines. This will probably not be a problem.

Power output. I would consider reducing the dead-rack adjustment on the governor. This setting may be more familiar by other names. It is the governor limit on the maximum fuel delivered. It is normally set up on a new machine with a load bank.
If I was working on a machine where with no information available for the governor, I would not make any changes.
If I could definitely identify the "dead-rack" adjustment, I would adjust it.
I would parallel the machine and advance the governor to pick up the amount of load that I want to set the adjustment for.
I would turn down the dead rack adjustment until the power output starts to fall.
I would increase the governor setting. (At this point, the dead-rack setting should be over riding the governaor setting.)
I would make a final adjustment of the dead-rack to the exact point that I want the power output level to be.
I would return the governor setting to the normal operating setting.
I would want to be assured that the Automatic Voltage Regulator will not damage itself or the field by trying to supply excess excitation. Other than that, I would let the other machines supply the reactive current if required.
respectfully

 

[oftenlost]

I know this is late but for my own edification, why not a 12000KVA 6Y?

 

[electricpete]

I got an email from aolalde and that is exactly what he is looking at now.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[waross]

If the removal of some coils changes the wave-form, there will be current sharing issues internal to each cycle. This may cause some unexpected heating. With 4 good generators in parallel with one bad generator, most of the heating will be in the bad generator. You can probably accept the extra heating, but I would watch the temperature as you build up the load on the repaired generator.
respectfully

 

[edison123]

oftenlost

With 14 poles, I am not sure you can get 6 circuits.

aolalde

If, as pete says, you are successful in reconfiguring for 6 circuits, would appreciate if you can let me know how.

pete

I think you have my mail id. Would appreciate if you can pass it to aolalde. Only my new domain is gmail.com

 

[oftenlost]

Thank you for the responses. My thinking was that this was a 7 circuit Y which does not have the magnitude of circulating problems seen in Delta connections. As each of the 7 circuits is delivering the line potential, each circuit becomes a parallel "pipe" for current flow. Cutting out the 6 coils in the discrepant pipe and limniting the load to that safely carried by the remaining 6 circuits would seem to be possible. Assuming a ground in Phase A, you would remove the six coils in that particular string, leaving the remaining 120 coils intact. The individual phases would still be delivering 3828 VAC per leg and there should be no reason, in my thinking to cut out more than 6.

 

[edison123]

It is not just the question of cutting out the parallel paths with corresponding output reduction. You need to make 14 poles / phase in the stator with 6 circuits. I don't see how this can be done.

May be aolalde will give the good news.

 

[aolalde]

Hi Edison.

I know it is hard to understand how you can get 14 poles with six circuits, a winder shall know that parallel circuits allowed are only whole results of the ratio P/n. For P=14 only 14, 7 & 1 are satisfactory. That is true when you are constructing a whole winding. In fact this original winding complies with that rule.

However, now a perfect balanced stator winding is being modified. For this particular case having a synchronous salient pole generator helps to easily understand why 14 poles are still there provided by the rotor excitation.

Each individual coil is a voltage cell with a certain induced value and phase angle “Volts per coil” (a vector). When the six coils in one circuit connected in series add (as vectors), the Voltage line to neutral is achieved. As far as the flux distribution in the air gap is kept, each circuit will produce the line to neutral voltage and will help to feed current to the load. In theory if I should have a way to add or remove these cells to feed my load, I could use 7,6, 5, 4 etc.

Since nothing is for free, the load current will produce an armature reaction flux that will distort the original balanced distribution around the 14 poles, because some coils have load current an some do not have any. The more coils are put out of work the higher the magnetic unbalance, that will be worst if it is left on a localized zone of the winding. That is why we have decided to remove a circuit from the other two phases at 120 and 240 geometrical grades. I will let you know the results next week.