Squirrel Cage Rotor Testing 2

[KllrWolf]

We have someone asking us about rotor testing for broken/cracked bars and joints. We use the growler test on all the rotors. They brought up an induced current loop test in the spec they sent us. I can find information on testing the stator, but the test on a rotor is slim information. About the only thing I found for a rotor was advertising or required the shaft to be hollow for the test cable to fit through. My questions are: what exactly is the test, how easy is it to perform, is it worth performing, and does it require special equipment?

Thanks

 

[Leendabests]

Hi...

We having same issues as your. But the test we have made it induced the current into the rotor to heat and then scan using infrared to detect hot spot more than other part of rotor. It is not really practical because surface of the rotor also can produce hot spot. Rotor is one of the essential part to produce electromagnetic. The best test i guess is to test the assemble and dissamble motor. Test the stator coil 1st

 

[electricpete]

For the loop test of rotor:
The shaft need not have a hole in it, could also be a spider type.
If not spider and no hollow shaft, there is no option for rotor loop test, but you could clamp onto the rotor at each end to directly apply voltage, rather than induce it.

The intent in both cases is to create current in the bars... if there is a bar/joint with high resistance connection, it should show a hot spot on thermography or flux anomaly with green paper.

In both cases, there are lots of variables that can influence the flux and current beside broken rotor bars. Not a great test imo but I like the clamp test better than the loop test.

Here is an article by EASA which is a survey of methods for checking rotor bars

http://www.pumpingmachinery.com/pump_magazine/pump_articles/article_29/rotortesting.pdf

Neither method is mentioned in there but they are mentioned in motor repair specs such as EPRI's medium voltage repair spec.

Attached are results of thermography during clamp test of an 800hp from our plant that showed rotor defect based on running current signature analsysis (45 db down). The infrared shows hot spots at same locations as visible cracks. After resoldering, the current signature analysis restults returned to normal (60db down).

Some rambling discussion of this particular case and the monitoring methods in general here:

http://maintenanceforums.com/eve/forums/a/tpc/f/7161085912/m/7071076492/p/1

=====================================
(2B)+(2B)' ?

 

[electricpete]

Also, a bad bar can show up as either hotter or colder than the rest... hotter if the increase in resistance is small, cooler if the increase in resistance is large approaching an open circuit.

Let's say I apply a voltage V from one endring to the opposite endring (or I apply a current, but changes in one path are small enough that they don't change the overall voltage much).

That voltage sees normally resistance of one bar (Rb) plus resistance of two connection resistance Rc (one for each joint from from bar to ring.
Inormal = V/Rtotal
Inormal = V/(Rb + Rc + Rc)

Power in either connection is
Pnormal = Inormal^2 * Rc
Pnormal = [V^2 /(Rb + Rc + Rc)^2] * Rc

now let's say that an abnormal connection resistance increase by a factor f (f>1) to f*Rc
Iabnormal = V/Rtotal
Iabnormal = V/(Rb + Rc + f*Rc)

Power in the abnormal connection is
Pabnormal = Iabnormal^2 * f*Rc
Pabnormal = [V^2 /(Rb + Rc + f*Rc)^2] * f* Rc

Find the ratio of the heat in abnormal connection to heat in a normal connection

Pabnormal/Pnormal = f * [(Rb + Rc + Rc) / (Rb + Rc + f*Rc)]^2

For f*Rc << Rb, then Pabnormal/Pnormal ~ f [faulty connection gets hotter]
For f*Rc >> Rb, then Pabnormal/Pnormal ~ 1/f [faulty connection gets cooler]

=====================================
(2B)+(2B)' ?

 

[electricpete]

There were lots of assumptions above (including neglecting inductance). Results are not intended to be quantitative, just an illustration of why a faulty connection can get hotter or cooler.

=====================================
(2B)+(2B)' ?

 

[petronila]

Hello KllrWolf,

This test can be performed with core tester, the core tester are intended mainly for stator test but are equipped with special clamps to perform DC Armature and squirrell cage testing in order to find hot spots like or open bars.

You can find manny manufacturers at trade like

http://www.lexseco.com/product_info.htm

Regards

Carlos

 

[KllrWolf]

Thanks everyone for the information. It has filled in the blanks, and gave me good places to look up more details.

 

[Motorwinder]

We use the Lexseco core tester, with the green magnetic paper. Works well. Sometimes we have to heat the rotor up to 400F before it shows opens. Rotor testing can be tricky.