Question for experienced rewinders 5

[zlatkodo]

I would like to know how often exist in practice (on the U.S. market and in other areas outside Europe), three-phase low voltage induction motors with 96 slots and for which number of poles , power and purpose ?
Thanks in advance.
Zlatkodo

 

[electricpete]

I think that windings for 6 or 12 poles and 96 slots are asymmetric windings. Therefore, they must be unbalanced. The only question is: what we can do to reduce imbalances?

On page 102 of the above link, there is discussion of the possibility of leaving Q0 slots empty to establish symmetry, and new symmetry conditions given in equations 2.84, 2.85, 2.86

For Q = 96, 2.84 implies Q0 must be 6, 12, 18 etc
For Q = 96, 2.86 implies Q0 must be 3, 9, 15 etc
There is no Q0 that satisfies both 2.84 and 2.86 when Q=96. I conclude there is no way to create a symmetric 96 slot machine even by leaving slots empty. But that's just going by what I read at that link (and assuming I didn't make any mistakes). I guess doing a full vector diagram provides better insight as to what is the degree of imbalance, and apparently must be small if 96 is used by Reliance.

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

 

[electricpete]

apparently must be small if 96 is used by Reliance

Duh. Never mind - they use it for 8 pole, not for 6 or 12 pole.
(Should've have said no way to create symmetric 96 slot machine for 6 or 12 pole.)

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

 

[edison123]

pete - I will look at the equations given in that google book later when I have the time. But I can very much assure you that we have rewound many stator and rotor windings with 96/6 configuration and all of them are working well with balanced currents without any electrical or mechanical issues whatsoever.

Muthu

http://www.edison.co.in

 

[electricpete]

I didn't make any claims other than the balanced / symmetrical part. If the winding doesn't meet the criteria, it cannot be PERFECTLY balanced from a mathematical point of view even if eerything else is ideal and should result in some unbalance. From a real world point of view it may not be important.

Noise/vibration is also a consideration. The same reference I linked above suggests n=2 is required for fractional slot winding to avoid subharmonic noise/vibration. The 12/6 has n=3. How much a problem it is I don't know - it is somewhat mysterious to me.

We have two sets of large fractional slot machines at our site: one is 800 hp 900 rpm. The other is 3500 hp 324 rpm. During uncoupled run in the repair shop, you can stand right next to them and can't tell theyr're running. When under load, they SCREAM. And I have to put scream in capital letters because there is no other way to describe the noise. The 324 rpm 3500 hp are about a mile from the parking lot, and I can tell whether they're running when I get out of the car in the parking lot each morning. I don't know how much this is related specifically to fractional slot design features of these motors... it is just something I have always associated with fractional slot windings in general based on these two.

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

 

[esee135]

I agree with Muthu. I got to work in a large global rewind engineering company for a period, and there is ALOT going on with the slots per pole per phase ratio. I found that there is some juggling within the coil group distribution for the vector summation for phase separation, balance, what have you. Not a simple subject. I've worked on synchronous machines with s.p.p.p's of 1-33/35 with split shared stator poles between multiple circuits, and it was very balanced, fractional near 6 seems simpler to me.

I can assure you, there are some windings that look completely odd ball, but it takes effort and time to analyze it to determine whether or not it affects operation. In general if you speak with the winding engineers regarding some odd configurations, I've always had a good technical answer.

Even if it's unbalanced, you can likely still run. Which is the case ones you start bypassing stator coils.

 

[electricpete]

just so we're clear on what is being said, you guys agree it won't be PERFECTLY balanced, right?

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

 

[zlatkodo]

Hi, Electricpete,
Thanks for useful link (Reliance).
You wrote:
I guess doing a full vector diagram provides better insight as to what is the degree of imbalance.
That’s true. Right way to know what is the degree of imbalance is doing a vector diagram to see
how much the angle of mmf -vectors deviate from the ideal direction and what is the percentage difference in the size of these vectors.
If these differences are in the allowed limits (for a particular purpose, power and motor-type), then the winding can be used... Otherwise, this test should be repeated but with a different coil-order . It should be noted that the minimum imbalance is often achieved with the order in which is no repetition (repetition is not necessary). Of course, it is suitable for small power motors, because in that case we can not do parallel circuits.
Regards.
Zlatkodo

 

[edison123]

pete

Fractional slot windings are perfectly balanced windings electrically if all phases have the same no. of coils and same total no. of turns. But the magnetic balance spatially around the periphery has to be obtained by proper placing of the coil groups around the machine so that there is no uneven magnetic pull.

Your scream problem seems interesting. It is normally due to improper stator/rotor slot design. IIRM, excessive skewing of slots creates a noise. Have you talked to the OEM about that issue ?

Muthu

http://www.edison.co.in

 

[electricpete]

"Fractional slot windings are perfectly balanced windings electrically if all phases have the same no. of coils and same total no. of turns"
I have a hard time believing that one. If we looked at voltage induced in the coils by the main flux wave, each coil has a different phasor voltage induced by virtue of different position relative to main flux wave. So looking at number of coils can't guarantee anything. If conditions are not met, then it's never going to be exact. There are lots of tricks you guys mentioned to get it close, but it can't be perfect.

The scream problem - both motors are 25 years old - 2 different OEM's, one out of business and one not particularly responsive. The fact that it occurs on 2 for 2 of our families of large fractional slot motors (and nothing like it on about 20 families of higher speed motors) makes me think it is relatively commomn among fractional slot motors. If not all fractional slot motors, then there is something to do with the design... maybe some obscure rule related to slot selection.

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

 

[rhatcher]

edison123 and wolf39 are right (stars). The 6 or 12 pole winding can be created symmetrically in a 96 slot winding. In fact, all pole combinations can be created with the right pole group sequence and the right number of circuits.

To answer the OP's question: For 96 slots; 2, 4, 8, and 16 poles results in an even group (integral slot) winding. This means that the number of coils per pole group is the same for all pole groups. This is the easiest winding to perform. In a general case, you will usually find integral slot windings on standard design motors. The manufacturer uses different stators to acheive integral slot windings for different pole combinations. For this reason, integral slot motors are more common than fractional slot motors.

If you are limited to a fixed number of stator slots, in this case 96 slots, you can still design a symmetrical winding at different speeds using an odd group (fractional slot) winding. For 96 slots; 6, 10, 12, 14, 18, 20, and 24 poles results in odd groups.

electrcipete's link gives conditions of symmetry. If we agree that there is no problem when n/m is not = 3/3 then we have agreed that for 96 slots the 10, 14, 18, and 20 pole windings would be symmetrical.

The question comes up when n/m = 3/3 such as the case for 6, 12, and 24 poles. The link suggests that this results in a non-symmetrical winding and to a certain degree, this is true. The reason is that mathematically this results in an odd grouping where the number of groups in a section is equal to the number of phases. For a 96 slot six pole, this results in a grouping of 556. When you repeat this through the phases (a,b,c) you get 556,556,556. Obviously C phase has too many coils.

The solution that is not addressed by the link is that in this case the section grouping must be varied for three sections to result in a symmetrical winding. So, the proper grouping would be 556, 565, 655. Wolf39 correctly demonstrated this when he wrote:

Phase A: 2 x (5 + 5 + 6) = 32 slots
Phase B: 2 x (5 + 6 + 5) = 32 slots
Phase C: 2 x (6 + 5 + 5) = 32 slots
Total 96 slots
Each of the 6 poles has a slot number of 16.

If you constructed this winding, it would be electrically and magnetically symmetrical. However, as suggested by electricpete and zlatkodo, I suspect that this arrangement may not result in a perfectly sinusiodal mmf wave in the air gap. But, the imperfection will be symmetrical.

 

[electricpete]

Regrouping coils is a trick I referred to (23 Dec 09 2:09) that will get you closer but not exactly perfectly symmetrical. Just putting the same number of coils in each leg does NOT guarantee a PERFECTLY symmetric winding. Because each coil does have the same voltage phasor induced... that voltage phasor also depends on position of the coil relative to the main flux wave.

I agree perfect isn't necessary.


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

 

[electricpete]

I wasn't making a careful distinction between balanced and symmetrical. I should have said would not be balanced... ie voltages induced in each leg will not be equal... and putting same number of coils in doesn't guarnatee it will be equal. Also sorry for using capital letters.

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

 

[electricpete]

http://books.google.com/books?id=fd..."fractional slot winding" symmetrical&f=false

On page 239 he defines
Ng = N1 / (p q) = A / B

Ng is slots per group, plays the role of previous q
N1 is total slots plays the role of previous Q
p is poles
q is phases, plays the role of previous m (3)
B plays the role of previous n. It is the number upon which both of the cited symmetry conditions apply.

Page 243 begins an example. 150 slots, 12 pole. It has B=6 (similar to n=6 in previous) which does not meet the symmtery. He shows at the bottom of page 239 that coils are judiciously regrouped to optimize resulting in same number of coils per phase.

But it is not a perfect balance. I assume it is a similar regrouping that you guys are talking about.

I am not a winder, and I could be wrong. But I think you guys are talking about getting close but not perfect just like the example in the link above.


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

 

[electricpete]

" He shows at the bottom of page 239 that coils are judiciously regrouped to optimize resulting in same number of coils per phase."
should have been
" He shows at the bottom of page 243 that coils are judiciously regrouped to optimize resulting in same number of coils per phase."

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

 

[electricpete]

But what I want is just the sequence of 18 numbers representing number of coils per group going around the circumference. Twelve of them are 5 and six of them are 6. What is the order of the 5's and 6'x?

Here is my previous attempt:
556 565 655 556 565 655

This particular one doens't seem symmetric to me because the 6-coil B groups have another phase 6-coil group 2 groups away on both sides. While the 6 coils a and c groups have distance 4 to the next 6-coil group on one side. So if I were looking just at the sequence, I can see there is something different about B phase than A phasel That doesn't seem symmetric to me. Would you call it symmetric?

I am open to the idea there is another way to insert the 5's and 6's to create symmetry, but I sure can't see it.... so if you can provide one, that would help.

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

 

[electricpete]

I would also like to define my terms, hopefully the same way others are using the terms, but correct me if I'm wrong.

Balanced means the same voltage is induced in each phase (or circuit).

Symmetric to me means 3-phase symmetry. I can't tell A from B from C by looking at it. That symmetry also creates a pure rotating flux pattern (although it surely has spatial harmonics). I don't think we get a pure rotating flux pattern unless we have the three-phase symmetry.

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

 

[electricpete]

Another comment on definition of symmetry - although we have spatial harmonics, they all travel at the same speed and direction.. which gives the rotating flux wave (with spatial harmonics). In contrast if the winding is not symetric, then those spatial harmonics travel at different speeds and directions, rather than a single wave that moves together. Most obvious example is single phase motor with reverse rotating fundamental.

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

 

[electricpete]

My last comment about symmetry was off. We can never have all spatial harmonics travelling at same speed since the slot boundaries don't move with the wave. Sorry.

I assume symmetry is a sufficient but not necessary condition for balance.

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

 

[rhatcher]

The method used for the design is the same for all windings.

1)determine # of groups = poles * phases = 12 poles * 3 phases = 36 pole groups

2)Determine # coils per group = slots/group = A B/C = 96/36 = 2 2/3 If A B/C is a whole number, the grouping is even and you can stop with this portion of the design and skip to step 8 (design the connection, # turns, etc.) If the result is a fractional number, the result A B/C tells you this; some groups will have A coils, some will have A+1 coils, C is the number of pole groups in each section with a section being the number of poles to until the pattern repeats.

3) Determine # of repeatable patterns = poles/C = RP = 4 This is the number repeatable pattern in the winding and the maximum number of parallel circuits.

3) Determine # groups with 'A' coils = 3 * RP * (C-B) = X = 12

4) Determine # group with 'A + 1' coils = 3 * RP * B = Y = 24

5) Determine # groups per section with 'A' coils = C * X/groups = 36/36 = 1

6) Determine # group per section with 'A + 1' coils = C * Y/groups = 72/36 = 2

7) If C does not equal 3 or a (multiple), you have determined the section grouping. According to the reference you have also made a symmetrical fractional slot winding. To demonstrate what happens in this case, if you went through the steps for 10 pole 96 slot, C = 5 and at this point you would know that the section grouping is 33343 and the number of repeatable patterns in the winding is 2. You divide the section into three phases and repeat the section three times to finish one repeatable pattern. Then you repeat once more (RP=2). You get:

333,433,334,333,343/,/333,433,334,333,343 = 96 slots 10 pole // separates the two repeatable patterns

In the 12 poles 96 slot example we are looking at, C = 3, the section grouping is 233 (or 232 or 332), and the repeatable patterns is 4. The section grouping is problematic since it would result in a winding that is 233,233,233...where there is obvious unbalance. In this case you vary the section grouping for three sections to form a repeatable pattern. Specifically, instead of repeating the same section three times to get a repeatable pattern you vary the section three different ways to get a repeatable pattern. Then you repeat that pattern 4 times (RP=4). You get:

233,323,332/,/233,323,332/,/233,323,332/,/233,323,332 = 96 slots 12 poles // separates the 4 repeatable patterns.

8) The final step is to design the connection, the number of parallel circuits, and the number of turns.

Now, if you at the difference in 'symmetry' between the 10 pole/C=5 and the 12 pole/C=3, there does not appear to be a difference. It makes me wonder why the first reference that you brought said that C=3 cannot result in symmetry.

Anyway, I have some thoughts on how symmetrical this really is in the vector sense that you are thinking about and what things to consider when looking at this. However, I am out of time for now so I'll come back later.

Pete...I hope this helps and I hope I didn't make any mistakes. Let me know what you think.

 

[electricpete]

I thought that symmetry meant that the 3 phase windings look identical, just shifted from each other. I have a hard time exactly reconciling that from the definition on page 99 of the 1st link: "A winding is said to be symmetrical if when fed from symmetrical supply it creates a rotating field"

I have to admit that definition throws me a little. Is it the same as my definition? Or is it the same as my definition of balance? How do you define symmetry?

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