Is a thermal shock sometimes necessary for a Die-Cast Aluminium IM Rotor??

[robglow]

We've recently had some of our Die-Cast Aluminium rotors pre-commissioned by another company. After doing a Motor Current Signature Analysis (MCSA) they found an imbalance in one of the rotor bars. The other rotors that passed the test, were subjected to both light and heavy loads. This rotor with the imbalance however, was only tested under light, no-load conditions. Their argument was that no further testing was required, as it is clear that the rotor was faulty.

Since the rotor cage was assembled with a molten metal die-cast, the possibility of inter-bar short circuits in the core increases. It is my opinion that this rotor needs to undergo a thermal shock (full-load current run) in order for the rotor bars to expand axially and break apart any possible short circuits between bar and core. The expansion rate of aluminium is a lot more than that of the core and would result in movement between the two parts.

My question is, can anyone advise me of any research that has been performed on this subject? Has anyone else experienced this rotor thermal before?

Rob

 

[electricpete]

I don’t know much about effects of inter-bar currents (I wish I knew more). All that I have heard about them centers around their effects on stray losses and starting performance. And it stands to reason that those effects are more pronounced in cast rotors with lower bar to iron resistance.

When you talk about “unbalance”, I don’t know what you mean. Since you mention current signature analysis, I am thinking maybe you are suggesting that variability of the bar to iron resistance can show up in current signature analysis as pole pass-frequency (2*s*FL) sidebands around line frequency?

If so, a few thoughts fwiw:
1 – Most pole pass frequency effects don’t show up well at “no-load” as I’m sure you know. Would expect them to get larger in magnitude and higher in modulating frequency as load increases.
2 – I would think from physical considerations that fabricated motors would be more susceptible to variability in bar-to-core resistance while the bar-to-core resistance of cast rotors would be more uniform. And yet variations in bar-to-core resitance has never been suggested as cause of “false positive” pole-pass frequency variations for fabricated rotor SCIM’s to my knowledge.
3 – Is it logical that inter-bar currents cause modulation of fundamental current at LF rather than slot harmonics? I vaguely thought that inter-bar currents are associated with harmonics rather than fundamental.


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

 

[waross]

Have you considered a void in the casting?

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[robglow]

When you talk about “unbalance”, I don’t know what you mean. Since you mention current signature analysis, I am thinking maybe you are suggesting that variability of the bar to iron resistance can show up in current signature analysis as pole pass-frequency (2*s*FL) sidebands around line frequency?

Yes, I am talking about pole-pass frequency showing a significant dB spike in the test result. I'm not too familiar with MCSA, but I am told this is an indication for a possible cracked rotor bar, causing higher resistance, less bar current and imbalance.

Most pole pass frequency effects don’t show up well at “no-load” as I’m sure you know. Would expect them to get larger in magnitude and higher in modulating frequency as load increases.

I would expect this too. But, the interesting thing is, when the other rotors were tested, the pole-pass frequency showed two significant spikes on the side-bands at no-load, but when the load was increased to full-load, the spikes seemed to have decreased to an even "healthier" state. This gives me some evidence that the rotor bars might have "cracked" away from a possible bar-core bond caused by the casting process. So to answer your point no.2 - I guess the overall resistance in general would be more uniform, but it also might be more susceptible to bar-core insulation breakdown, due to excessive molten temperatures.

Page 14 of this paper mentions this theory, but I can't seem to find any other publication on this.

What's your opinion?

 

[electricpete]

Page 14 of this paper mentions this theory, but I can't seem to find any other publication on this.

ok, let me post that into this thread so we can think about it:

Cast rotors are frequently heat treated after casting to reduce surface loss. For one
such process, after the rotor is turned to size, the rotors are heated to about 900 F,
then rapidly cooled by water quenching. The major effects of flame treatment:
•[1] Smear on the rotor surface is removed. This smear is caused by the cutting tool
when the rotor outside diameter is finished. The smear shorts laminations together
And cause inter-laminar currents.
[2]• The thermal shock caused by rapid water quenching of the heated rotor tends to
separate the bars and lamination, breaking up shorts between bars and laminations
which were missed by the core plating process.
•[3] The heat promotes oxidation of the bar surface. The aluminum oxide coating thus
formed on the bar adds to the insulation between the bar and lamination.
The effectiveness of flame treatment depends on the temperature, the rate of
temperature rise, and type of gas used for the flame propagation.

I see what you mean about the thermal shock part based on item 2.
Now I'll really show my ignorance...I had no idea there was a coating on the part of the core that forms the barrier for the bars. I've only seen fab rotors .... laminations are coated on the flat sides before stamping. That coating isn't going to end up in the hole that forms the cavity for the bar. I guess fab rotors must have some extra step of coating the core after it is assembled from laminations into a stack.

I would expect this too. But, the interesting thing is, when the other rotors were tested, the pole-pass frequency showed two significant spikes on the side-bands at no-load, but when the load was increased to full-load, the spikes seemed to have decreased to an even "healthier" state. This gives me some evidence that the rotor bars might have "cracked" away from a possible bar-core bond caused by the casting process

Yes, that does seem to support your theory. It's not expected behavior for a porosity defect that is static / unchanging. I wonder what would have happened if you went back down to load load again after the high-load run. Logic suggests if your theory was true it would be even lower then I think.

I don't think my input will be worth much but I'm going to ponder it some more.

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

 

[electricpete]

correction:

I wonder what would have happened if you went back down to load load again after the high-load run. Logic suggests if your theory was true it would be even lower then I think.

should have been

I wonder what would have happened if you went back down to low load again after the high-load run. Logic suggests if your theory was true it would be even lower then I think.

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

 

[electricpete]

Sorry, it's late and I have no edit button. One more correction:

I guess fab rotors must have some extra step of coating the core after it is assembled from laminations into a stack.

should have been

I guess cast rotors must have some extra step of coating the core after it is assembled from laminations into a stack.

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