rotor core loop test to find broken rotor bar - theory question 2

[electricpete]

I have been told by a respected motor expert that a broken rotor bar can be found during a core loop test of the rotor.

Several loops of coils looped thru center of rotor and around outside and back thru center etc. Then pass high current using a ductor to excite the core.

Then supposedly we can detect a cracked/broken rotor bar by infrared inspection or clamping onto to rotor bar extension if possible, or some similar method to look for current deviation in that bar.

Here is my question: At a given instant in time during this test, the flux is all flowing in the same direction uniformly around the core. This means that an identical voltage is induced in all rotor bars.... attempting to induce current in the same direction in all bars. Where would the return path for current flow be if current is flowing the same direction in all bars? Is it through the shaft? Or maybe this test won't work to detect broken rotor bars?

 

[SteveKW]

Someone will say I am wrong but that's never stopped me before.

You don't have a uniform flux density nor is the voltage identical in all the rotor bars. The current induced into the rotor bars is the greatest at the loop since the magnetic field decays the further from the loop you get. This current then flows (in one instant) in one direction closest to the loop and the return current path is through the rings (each end of the rotor) to the opposite bars and then back.

You could even put iron filings on a piece of paper that's on the rotor and see the filings "stand up" where the current flows through each bar. If a bar is broken you will see a gap where the filings do nothing.

 

[electricpete]

Thanks Steve. The iron filings would be a good idea.

If I look at the core from a magnetic circuit point of view: The flux needs to travel in complete loops. Since Mu of core is so much higher than air, most of flux should be in the core. If we assume that all of the flux is in the core and travels in a complete loop around the core with uniform cross section, it seems to me like the flux density should be uniform around the core. Your explanation may be the right answer, but I just can't picture it.

 

[GordS]

I would like to suggest the following in response to Pete's question which tries to also explain what Steve has noticed in reality. Consider it my best guess. Corrections are invited.

The test is actually intended to perform a rotor core loss assessment. The test current does, I think, create a uniform magnetic field within the core as Pete mentions. Areas where laminations have been damaged will show up as hot spots because the eddy currents should increase in those areas. In undamaged areas, eddy currents are limited to the lamination thickness because they see a high electrical resistance (electrical resistance being inversely proportional to cross sectional area)

The rotor bars will likely be in mechanical contact with the rotor iron at certain spots. This allows the eddy current in the laminations to continue into and spread through the rotor bar. This eddy current is perpindicular to the bar and creates a perpindicular magnetic field that lines up with the rotor diameter. This would induce poles in the filings that would cause them to stand up to satisfy the "unlike poles attract" rule. If the bar is broken however, there can be no continuation of eddy currents into the bar and therfore no magnetic field above the gap in the bar to make iron filings stand up as mentioned by Steve.

 

[electricpete]

Thanks Gord. Your comments make me realize that my simple model of flux flowing circumferentially around the core does not induce any voltage/current in the rotor bars since the flux would be flowing parallel to the plane formed by two adjacent rotor bars (flux is not cutting the area formed by the loop of two adjacent rotor bars connected at ends by ring).

But I have to admit I still can't how the eddy currents would they induce a current between the bar and the end-ring which would be heat up a loose connection (as seen by infrared)

 

[GordS]

Pete. Can your motor expert remember if a loose connection shoes up as a hot spot in this test? Or does the infra-red show the poor connection as a spot cooler than the surrounding area? The mechanism I described should show a poor connection as a cool spot.