Fractional Slot Stator Windings ? 4

[edison123]

We have a 3.3 KV, 1490 RPM (4 pole), 1000 HP Motor with 84 stator slots, which we want to convert to 990 RPM (6 pole) motor. For 6 pole, we get a fractional slot winding (84/3x6=4 2/3 slots per pole per phase). Can any one suggest the actual arrangement of coil groups around the stator periphery that will be electrically and magnetically balanced for 6 poles? For example, I claim the arrangement 555 455 544 555 455 544 is the best. But, our consultant claims the arrangement 545 545 464 545 545 464 is electrically better ? Views from any experienced ac machine designer will be appreciated.

 

[electricpete]

I am not a designer. But hey it's a free board so I'll throw in my unifnormed 2 cents for what it's worth.

I do have a reference which indicates that a balanced winding cannot be produced for this configuration.

Using terminology of EPRI5036V17
q = Q/(m*p) = N / B = a + b/B
Q=84
m = 3 phases
p = 6 poles
N and B are smallest integers
q = 84/14 = 42/9=14/3
B = 3
a - largest integer
b = remainder

Condition for balance is
d = (mNP+1)/B
where P is the smallest integer that will make d an integer.

If B/m is an integer, there will be no value of P that makes d an integer. In this case, the fractional slot winding will be unbalanced, and neither the phase voltage nor phase angles will be equal.

Sorry if I am telling you what you already know.

 

[electricpete]

typo correction:
Should have been q = 84/18 = 42/9=14/3
results are the same

 

[electricpete]

I should clarify it is a direct quote from the EPRI reference: "If B/m is an integer, there will be no value of P that makes d an integer. In this case, the fractional slot winding will be unbalanced, and neither the phase voltage nor phase angles will be equal."

NOT my own words.

 

[edison123]

electricpete,

interesting equation.

I agree that the winding we propose is not ideally balanced. My question was which of the two arrangements proposed in my posting would produce least dyssymmetry.

 

[teco]

edison123:- coil grouping
a b c a b c a b c
4 5 5 5 5 4 5 4 5 :-repeat wind 12 groups 5 + 6 groups4.
note:- unbalanced elect deg, connect series star only.
also check rotor slotts for cusp or coging.
cannot get a balanced winding for 6 pole (12.87 elect deg between slotts)config.have you looked at 8 polethis you have a balanced winding(17.14 elect deg )

kind regds

 

[jbartos]

Suggestion to edison123 (Electrical) Jun 12, 2003 ///\\We have a 3.3 KV, 1490 RPM (4 pole), 1000 HP Motor with 84 stator slots, which we want to convert to 990 RPM (6 pole) motor. For 6 pole, we get a fractional slot winding (84/3x6=4 2/3 slots per pole per phase). Can any one suggest the actual arrangement of coil groups around the stator periphery that will be electrically and magnetically balanced for 6 poles? For example, I claim the arrangement 555 455 544 555 455 544
///5+5+5+4+5+5+5+4+4+5+5+5+4+5+5+5+4+4=84
or
15+14+13+15+14+13=84
Please, what is your design basis for this configuration?\\ is the best. But, our consultant claims the arrangement 545 545 464 545 545 464
///14+14+14+14+14+14=84
This partitioning is more symmetrical.
Also, considered could be either:
545 545 545 545 545 545=84
or
464 464 464 464 464 464=84
or
545 545 545 545 545 545=84
Nowadays, the magnetic circuits are often designed by software. Has the consultant used any software for the group it is claimed better one?\\ is electrically better ? Views from any experienced ac machine designer will be appreciated.
///This may be proprietary.\\

 

[GusD]

Hi edison 123

I am curious of why would you hire a Consultant if you think you have a better answer to your problem?All of the arrangments mentioned so far would probably get your motor running.The problem starts after it runs.Can it run efficiently and up to the rated Horsepower.
I would suggest,if possible,consult the motor manufacturer. They have stator core magnectics information that are quite important for motor performance in addition to the Physical arrangement of coil groups.
If you expect to have equal motor efficiency, you have to have all of the Electrical/Magnectic information related to this motor.
We,working for a motor factory subsidiary,were always very reluctant to change motor's design whenever that change impacted on the Motor Magnectics/Harmonics.Sometimes,what appears to be a straight forward motor change,turns out pretty ugly.The Customer is very seldom aware that his
previous 1000 hp have now shrank to 950Hp.

GusD

 

[edison123]

teco
Thanks for your input. I have decided on 455 545 554 ..repeat, which is closer to what you have suggested.

gusD
don't we all live and perish by consultants? it was management (?) decision to counter check an engineer's decision. As usual, the consultant was proved wrong. In our case, we are changing the motor design to suit the new speed requirement on the load side. naturally, the new output at 1000 RPM will be 700 HP.

electricpete
thanks again for your memorable equation

 

[jbartos]

Suggestion: Is your design based on prerogatives or on any engineering design basis or empiricism?

 

[edison123]

My design is based on 1. Pole balance (All pole-phase bands have same no. of coils). 2. Magnetic balance (Opposite pole bands have same no. of coils). These two conditions are required to get electrical as well as magnetic balance.

 

[jbartos]

Suggestion: Please, would you numerically elaborate on how those pole-phase bands have the same no. of coils.

 

[edison123]

If you add up the no. of coils in each pole-phase band in my above posting (4+5+5) (5+4+5) (5+5+4)... repeat, you get the same no. of coils(14)in each pole-phase band. This satisfies the electrical balance.

 

[jbartos]

Suggestion to previous posting: Thank you for the clarification that appears to be somewhat different from your original posting.

 

[electricpete]

Well, since the subject is opened up....
I have to say I am absolutely positive that I don't understand the requirements for a balanced fractional slot winding. But I'm interested to learn from anyone out there who knows more about it.

Here is one thing that strikes me about
(4+5+5) (5+4+5) (5+5+4) (4+5+5) (5+4+5) (5+5+4)

The grouping of three together is somewhat articial. It does highlight that there are same number of coils in phase A, B, C. However 60 degrees later in time I think the same winding could be described as :

..4) (+5+5+5) (+4+5 +5) (+5+4 +4) (+5+5+5) (4+5+5) (+5+4..

The groups range from 13 to 15. I think it is a reflection as discussed above there is no perfectly symmetrical solution. Maybe there is more to it.

 

[electricpete]

Appendix 1 of Electric Machinery, Volume 2, by Liwschitz/Garak has a detailed discussion of this subject. Far above my head. If it can be described in 3 or 4 simple rules I'd be interested to understand, but it looks more complicated to me.

 

[jbartos]

Suggestion: The winding symmetry is leading to temperature distribution symmetry. Visit

http://www.magsoft-flux.com/Products/software/motorcad/motorcad.html

for the Motor-CAD software.

 

[electricpete]

I have already qualified I don't know the subject. I read a little bit and trying to understand.

I think the 3 induced phase voltages could be calculated for each of the proposed configurations. The goal is to generate induced phase voltages for three phases which are as close as possible to equal magnitude with phase angle 0, 120 degrees, 240 degrees.

If all coils were connected in series, the voltage for a given phase is the sum of each of the included induced coil voltages.

For 84 slots, 6 poles, we could represent the induced voltage for the coil whose top leg is in kth slot as a phasor:
Vk = exp(6*Pi/84)*(-1)^t for k=1, 2, 3, ...84.

Va = Sum(Vk) for all k that belong to phase A.
In the case of proposed solution:
(4+5+5) (5+4+5) (5+5+4) (4+5+5) (5+4+5) (5+5+4)
for phase A we would use k in the set of 1,2,3,4 (first 4), 15, 16, 17, 18,19, 20 (batch of 5), etc.

for phase B we use 5,6,7,8,9 (batch of 5), etc

t is an odd or even number depending on the two possible coil polarities.

 

[electricpete]

typo correction... should add I (sqrt(-1) as follows:

Vk = exp(I*6*Pi/84)*(-1)^t for k=1, 2, 3, ...84.

 

[jbartos]

Sugestion: Visit

http://www.ansoft.com/device/motor/3phase10.pdf

a motor design software for more info