Help with a formula? 2

[THooper]

Can anyone help me with figuring out a formula? Let me give you a brief background of my problem. We have been rewinding some dc traction motors for many years. We have a customer that lost eight of these motors off the same side of their car hauler. Each of the motor failures are classic runaways, the centrifugal force has caused the armatures to flare excessively taking out the field coils. The nameplate on these motors are: 25/13hp, 2100 rpm, 110vdc, series windings, We use an epoxy resin that is rated at 10,100 psi in our winding process along with a banding material that is rated at 240,000 psi. Can anyone help me figure out how fast this armature would have to be spinning over the rated 2100 rpm, to flare and mushroom?

 

[JStephen]

First off, I can't tell enough about the geometry from your post to have a real clear idea of what's happening.

Secondly, it sounds like this could be a somewhat involved analysis, and not just looking for a formula.

It would seem too, that analysis of the situation would require analysis of the dynamic forces as well as the electromagnetic forces involved.

 

[IRstuff]

Are you saying that the armatures are going so fast that they're stretching enough to take out the field coils?

Have you actually done before/after measurements of the material deformation?

When you say "same side," does that mean there are is "another side" motor where there are no failures? Are they all being run at the same rpm?

Since you are using an epoxy, isn't it more likely that the motors are being overheated, causing the epoxy to soften?

After all, 2100 rpm is hardly fast. Even relatively cheesy RC motors can run 20,000 rpm without any problems.

TTFN

 

[CESSNA1]

MOTORHEAD1:

I am just a mechancial engineer, but it seems strange that the failures are on one side, presumably on one car of one company. Anyone else reporting problems? What are the circumstances?

Is this motor connected directly to an axle? If so how can it overspeed?

I recall from my EE courses many years gao that there is one type of motor winding (I do not remember which type) that if the field collapses the armature will accelerate and self desruct, plus throw shrapnel all over the place.

I believe you need more information on the circumstances, the application, and the problem history.

Hope this helps

Dave

 

[THooper]

Thank you for all your input:
JStephen, I agree that more analysis is required and history.

IRstuff, Yes the armature coils has stretched and mushroom from excessive speed, enough to take out the field coils. We do have measurements and the coils have mushroom approximately 2.5 in. No the motors have not seen any heat or thermal degradation. As far as this type of motor holding up to what a RC motor operates at, I don't think it would be able to handle that kind of speed due to the design and how large the armature is.

CESSNA1, Yes the failures are on one side and the same car. I don't know alot about the car itself, but it has two dc motors that drive the front wheels and each motor drives the wheel independent. Each motor is also connected to a gearbox. The system has some type of logic box that is suppose to shut the motors down if an overspeed occurs. The manufacture of the car said they can not find anything wrong with the car nor the control system and therefor is questioning our rewinding process. We have rewound alot of these for other customers and this is a very uncommon problem.

We believe that forwhatever reason, it is an overspeed or runaway, and the reason I am trying and looking for help to figure out a formula, is if the materials and winding process used is done right and in good condition, with epoxy rated at 10,100 psi and banding rated at 240,000 psi it would take___?___speed over the rated 2100 rpm to cause this type of coil lifting and mushroom?
Right now this is all I have is what we are seeing on the motor evaluation, I am trying to come up with more details to help solve this probelm and wish I had more of them to share with you. I met again this morning with the manufacture and they are at a loss....

Thank's again for your input
kind regards

 

[IRstuff]

How can you get a runaway with no other damage? And why only one side? Even assuming a overspeed, there would have been damaged clutches or reports of excessive noise, etc.

Can you swap the controller outputs from one side to the other?

TTFN

 

[THooper]

Exactly, there is more to this failure then I am being told, I did find out that they did indeed find a BAD logic box, but said this would have nothing to do with a overspeed, but I ask if this box controls the motor from an overspeed, and it went bad, then could it not do what it was designed to do, which is control from an overspeed, They put a brand new motor on also and the car is running just fine....

 

[notnats]

Is it possible that the wheels on the right side are lifting off the ground and spinning? A series motor will run away, possibly before the safety circuit works.

The formula for centrifugal force is

f = m*w^2*r

f=force in Newtons
m=mass of rotating body in kg
w=angular velocity in radians/sec
r=radius of rotation of body

If you know the mass of the winding and the effective radius of rotation, you can get some estimate of the stress on the epoxy.

Jeff

 

[THooper]

Yes, this is very possible of the wheels lifting off the ground, these cars operate in a underground mine and the road ways are very rough.. thank you for your help and formula Notnats, this will help me out tremendous!!!!!

kind regards

 

[zekeman]

Notnats,

The formula you cite is ok, namely

F=m*w^2*r

but

r= radius of the center of mass of the material above the location of stress area of interest. The maximum stress should be where the armature attaches to the hub

m= mass of everything above the area being analyzed

This is derived from:
F=Integral[w^2*x*dm] limits of x=x1 to x=x2
dm= differential mass
x= distance from reference origin

since w is constant, this simplifies to
F=w^2*Integral[x*dm] same limits
but from definition of center of mass the integral is
m*r
where m= total mass
r= center of mass

 

[notnats]

Zekeman,

Thanks for the extra maths, but as I understand the problem, the windings are being thrown out of the slots in the armature where they are mechanically anchored by an epoxy potting compound. So in that case, the stress on the epoxy is located at each wire and the epoxy is in shear between the wire and the wall of the slot. The conductors at the maximum radius will have the greatest force on them.

Motorhead1,
Are your windings potted under vacuum. This may be the difference between the mechanical strength of a new motor and a repaired motor. Vacuum potting will remove air bubbles and help eliminate weak spots.

Jef

 

[THooper]

Yes, the armatures go through a vpi process, vacuum pressure impregnation. The coils are staying put in the iron slots, but flaring and mushrooming as they exit the core iron of the armature.

 

[IRstuff]

Have you examined the windings after such an occurrence? Do they show tearing of the epoxy or do they show flow?

Have you attempted to run a test motor at higher than allowed RPM to see if the coils separate?

TTFN

 

[THooper]

The winding have expanded so far they show tearing of the banding and insulation.

We are talking about running one up past rated rpm to see when it comes apart, but have not done so as of yet.

 

[notnats]

I suppose it's obvious, but I have to ask: are your coils longer than the OEM windings and giving extra overhang out of the armature core? It seems the winding structure is loaded in bending at this point. So if there is more mass of wire hanging further out of the slot, the bending moment will increase more than linearly.

Jeff

 

[THooper]

They extend the same amount as the oem print, We also add a surge rope to the coil head during the winding process that is not oem, I did not mention previously because I was unaware of the surge rope, but this should also add to the strength. I am researching the spec's on the surge rope also.

 

[CESSNA1]

MOTORHEAD1: The coils on the drive end of the motor are cantilevered out and I am assuming they are retained to the armature somehow at the opposite end. This accounts for the "flaring" out at the drive end. The coil of wire comes out of the armature and bends around to go back in.
This is essentially a double cantilever beam with a weight(the loop back) at the end. The deflection at the end of a double cantilever beam is

d = (2*P*L^3)/3*E*I

d=deflection
P = applied load (centrifugal force on the wire)
L = length of the two wires out of the armature
E = Modulus of elasticity for copper
I = (3.14159/64)*D^4 (D=wire diamter)

The equation for centifugal force was given by someone else.

It does however, still come down to WHY and WHAT?
WHY one side
WHY the same car
WHY no other cars
WHY no other motors
WHAT is causing the centrifugal force
WHAT is different about these motors, interface
and the installation from the others
WHAT about the motor control system

How long does it take for the failure to occur? If it short time, say a day or two, it may be beneficial to instrument the drive system for voltage/current/rpm/etc. to see if the failure can be duplicated.

Try spinning up a motor to see at what RPM the failure occurs (be careful here)

Have the falied motors been check for manufactuing defects?

Is the motor cooling adequate, any sign of high temperatures, excessive current, overload, or overheating that could reduce the strength of the copper in the winding and cause them to failure prematurely.

Good Luck
Dave