The determination of the number of slots in a synchronous generator 1

[kh2]

The project is to upgrade a hydro synchronous generator 50 MVA, 24 poles, 3-phase wye, 13.8 kV L-L, lap winding, 2-turn/coil, 2-circuits, 198 slots. The upgrade shall include the rewind and to replace the stator core. The new rating shall be 60 MVA. The alternatives for the rewind are one turn coil, 2 turns, or 3 turns; one circuit, 2 circuits, or 3 circuits. The question is how many slots to have in the new stator core to accommodate one circuit verse two circuits verse three circuits. The design engineer knows the formulas to do so, but the user like me would also like to know.

Appreciate your expertise to provide the formula and the information required to calculate the number of the slots for such application.

Many thanks.

 

[starkopete]

Hi,
The number of slots will stay the same. The basic formula is: T2 = T1 x N2/N1 Where T2 is the new turns per coil, T1 is the original turns per coil, N2 is the new number of circuits and N1 is the original number of circuits.
The basic rules are: The number of circuits cannot be more than the number of poles. The number of poles must be equal to or be a multiple of the number of circuits. Thanks

 

[kh2]

Thanks for your response.

If we decided to rewind using one turn per coil (roebel bars) and 3 circuits/phase, then we will have: T2 =1, T1 =2, N1 =2, and N2 = 3.
Then we apply T2/T1 = N2/N1, the result will be 1/2 = 3/2. This is an incorrect equation, and where it shows that the number of the slots will stay the same.

Am I missing something?

Appreciate your time for providing a clarification.

 

[electricpete]

I think the original equation intended T to represent turns per coil times coils.

Let's say instead of T we use T = TPC * C
where TPC is turns per coil and C = # coils = # slots.

Then if (*) we are to keep the number of series turns per circuit the same, the condition would be:
C1 * TPC1 / N1 = C2 * TPC2 / N2

* - There may be opportunity to change the number of series turns per phase slightly if magnetic design and rotor excitation allows.

In a normal rewind C1 would be constant (keep the same number of slots). But with your new core, you can change C1 also.

Note the original winding was a fractional slot winding. The number of possible options for replacement windings seems pretty broad.



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

 

[starkopete]

electripete,
You are correct, I was ignoring the option of changing the number of slots. Thanks

 

[kh2]

Thanks again.

The existing machine slots/pole/phase =2.75 and the grouping is 3.3.3.2/3.3.3.2/3.3.3.2 repeated 6 times. Coil sides per slot is 2.

If we used C1 = 198 slots; TPC1 = 2 & TPC2 = 1; N1 =2 & [N2= 1 or 2 or 3]
Therefore C2 = [C1x(TPC1)x(N1)]/[(TCP2)x(N2)]
C2 (for one circuit and one turn) = (198 x 2 x 2)/(1 x 1) = 792 slots
C2 ( for two circuit and one turn) = (198 x 2 x 2)/(1 x 2) = 396 slots
C2 (for 3 circuits and one turn) = (198 x 2 x 2)/( 1 x 3)=264 slots

Is this correct? For example, for one circuit and one turn/coil, we need 792 slots instead of 198 slots.

Please review and advise. Thank you.

 

[electricpete]

C2 = [C1x(TPC1)x(N1)]/[(TCP2)x(N2)]
is not correct.

Starting with
C1 * TPC1 / N1 = C2 * TPC2 / N2

Solve for C2
C2 = C1 * (TPC1/TPC2) * (N2/N1)

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

 

[electricpete]

Here is a table of coils (slots) vs turns per coil and circuits using the equation above (with any luck the columns will be straight)
TPC2 N2 C2
1 1 198
1 2 396
1 3 594
2 1 99
2 2 198
2 3 297
3 1 66
3 2 132
3 3 198
I'm not saying these are all feasible, just writing the results of the equation that will keep series turns per circuit the same (i.e. keeps stator voltage the same for the same rotor voltage)


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

 

[kh2]

Thank you greatly for the correction and the formula.

This means to use one turn/coil (roebel bar) and 2 circuits will result in doubling the number of slots.

I thought that substituting two turns/coil with one turn/coil of a cross section equal to the two turns, will keep the number of slots unchanged. You mentioned that the number of turns/coil is for adding the voltage per turn in series in all the phase coils to be always equal to 8000 volts (1380/1.73).

Appreciate very much your comments on the above statement, and please comment on your experience with the single turn verse multi-turns, and what are the disadvantages of doubling the slots?

Our reason to choose the roebel bars is that we had stator windings failures few times that were caused by shorts between turns.

Thank you again.

 

[edison123]

2 T & 2 Ckt = 1 T & 1 Ckt. No change in slots.

And, no, you cannot double the slots just like that due to magnetic reasons.

Roebel bars have following advantages

1. No turn to turn failure

2. Better slot fill (of copper) due to elimination of turns insulations.

3. Mechanically more strong than multi-turn coils.

4. Ease of replacement in future.

5. Eliminates pole jumpers if you go for a wave winding.

Disadvantage

1. Cost due to specialized manufacture

Your case is ideal for Roebel bars.

 

[kh2]

Edison 123,

Good to hear from you. Appreciate very much your valuable recommendation.

Thank you.

 

[edison123]

kh2

You are welcome. Nice to hear from you too after a long time. Our last thread was about core loop test, wasn't it ? Must be 4-5 years back, IIRR.

 

[zlatkodo]

See attach.

Zlatkodo

 

[kh2]

Edison 123,

You are right. Last communication was in 2004. Another thread that interested you was the clearance between the coil/bar end turns. But I follow your discussions and learn a lot. Zlatkodo attachment is super.

Best regards,

kh2

 

[electricpete]

This means to use one turn/coil (roebel bar) and 2 circuits will result in doubling the number of slots [compared to original 2-turn/coil, 2circuits]

I agree you would need to double the slots if this configuration is pursued.

And, no, you cannot double the slots just like that due to magnetic reasons.

Can you elaborate? My take is that any design needs to be reviewed for constructability (can we still fit everything in the available envelope... do the teeth andcoils become too narrow) and harmonic performance. But it is not immediately obvious to me that doubling the slots must be ruled out.


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

 

[edison123]

pete

For a given core diameter, doubling the slots reduces the teeth width disproportionately leading to magnetic saturation.

Consider the original slot width and teeth width is 20 mm and the coil insulation is 5 mm total. The available copper width is 15 mm for a current of say, 100 A.

Now doubling the slots, reduces the slot width and the teeth width to 10 mm each. But the insulation requirement still stays at 5 mm. So, the available copper width is now 5 mm for half the original current of 50 A (because of two circuits).

You can see the current density in the second case has shot up leading to the necessary KVA derating, which defeats the very purpose.

If you want to maintain the same KVA rating, then the current density has to be the same => more slot width of 12.5 mm = 7.5 mm copper width => less teeth width of 7.5 mm => teeth saturation.

Remember the core ID and OD cannot be changed.

 

[electricpete]

Good points.

I hope we can agree whatever change is made needs to be evaluated. While we can't say what is feasible without reviewing everything, it also seems tough to say what is infeasible (within reasonable limits) without reviewing everything.

It may be that the slot can be made deeper (evaluate remaining back-iron depth). New insulation materials may be thinner than old materials (for all we know this is a 50 year old unit). Both of these can gain back a little space.

Also it may be possible to reduce copper based on thermal improvements:
• lower thickness of insulation (if applicable) has less thermal insulating effect.
• new core materials may be better than original.
• the temperature rating of the insulation may be higher than original.

With all that said, I have a suspicion 200 slots was somewhere near optimal. Even if doubling the number of slots were possible, it would probably push the design far away from optimal design from a price/performance perspective.

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

 

[edison123]

The same improvements you mentioned can be done with the existing slot configuration. It doesn't make technical sense (more coils, more chances of failured) or economical sense (more money to make more coils) to increase the slots. Large hydro generators, being custm built, are carefully and optimally designed taking into various technical and ecomomical aspects. One doesn't change the original design just because one can.