Two-ammeter check for rotor bar problem - using clamp-on ammeters vs in-line? 1

[electricpete]

I have a motor with oscillating current on in-line panel ammeter and worsening trend in apparent pole pass sidebands around 60hz in current.

In attempting to rule out process oscillation as the cause, I have verified no movement of refrigerant control valve linkages, but I'd like to get more confidence using the two-ammeter method as descirbed on page 7 here:

http://www.pump-magazine.com/pump_magazine/pump_articles/article_29/rotortesting.pdf

Page 7 talks about the two ammeter method.

Basically this method checks whether the (envelope of the) oscillation is in-phase among phases. (in-phase indicates simple load oscillation, rotating phase indicates rotor defect as bar bar passes among phases at pole pass frequency).

It is suggested to use two identical analogue ammeters.

The motor is in-service. Access to tag it out is limited. It is possible we have provisions to short ct secondary and insert extra meters in-line with motor running but we'd prefer not to do that if not necessary.
I'm wondering if we can expect reasonable results on the two ammeter test using analog clamp-on ammeters vs analogue in-line ammeters.

In absence of any advice, we will probably end up trying it to see what the results "look like". But even that non-intrusive clamp-on check will be harder for us than you might think. Would prefer not to mobilize the team for a first clamp-on check that doesn't work and then have to ask for a second check using in-line meters. So I'm still very interested to hear thoughts/predictions about what we can expect with that clampon test method.

More details on the particular case here are attached. Also discussed at another forum:

http://www.maintenance.org/topic/cast-rotor-worsening-current-signature-trend-no-change-in-vibration

Attachment doesn't seem to be sticking here... I'm not sure why.


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

 

[LionelHutz]

Clamping onto the CT secondary or clamping onto the motor leads directly?

 

[waross]

Hi Pete. About how fast is the current oscillating? If the variation is related to the slip frequency I would expect more of a "flutter" or vibration of the ammeter needle. Have you considered the possibility of displaying the currents on a two trace 'scope?

Open rotor symptoms of motor at rated load
–
Draws higher than rated amps
–
Operates slower than rated speed
•
Symptoms at any significant load (>50%)
–
Analog ammeter needle appears to vibrate

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

 

[electricpete]

We are measuring ct secondary.

Ammeter is at a frequency of about 0.3 hz (one oscillation per 3 sec) which did match pole pass frequency.
Video of the ammeter oscllations here - not too exciting

https://drive.google.com/open?id=0B3ePa0Zibh6UQ0lGZVNqeFdOXzA

Good idea about two trace scope. Motor current in range 30-60A, so output of our 30:1 CT will be in the range 1-2A.. low enough that we could use our hall effect probes (10A range I think). That would eliminate a lot of guesswork and questions. I should've thought of that, but anyway that's what I have you guys for (to keep me straight). Very helpful.

Here is summary of what we've seen for info

https://drive.google.com/open?id=0B3ePa0Zibh6UQUhyT0JydUtHN2c

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

 

[waross]

Not to rule out motor problems but I would be taking a close look at the mechanical side. Could be a sticky expansion valve. Stiction may be introducing some deadband into the valve action rather than smooth proportioning. I once had some regulator valves running colder than they should be and I got severe cycling, from proportioning to full open and back to proper proportioning several time a second.

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

 

[electricpete]

Yes, I agree. We know mechanical problems can cause sidebands. Here are our past/planned actions to rule out mechanical load variations as a cause:

1 - The first way to help rule out mechanical problems is frequency match (current oscillation frequency matches pole pass frequency). The early measurements in my attachment didn't all have great match, but I attribute that to a very indirect/inaccurate way to estimate speed (from current magnitude). The last measurement in the attachment was taken in conjunction with vibration measurement which gave us ability to determine speed accurately (from high-res vibration spectrum) and the match was good. We will continue taking monthly readings of both current and vibration at the same time over the next several months... (working toward motor replacement in the fall if all is confirmed).

2 - I have looked closely at the mechanical system for oscillations and haven't seen any. The conspicous controls on this axial flow compressor are variable inlet vanes and hot gas bypass valve... both have visible linkages which are not visibly moving while the current oscillation occurs. No oscillation of indicated pressures (although they read out on digital display.. I don't rule out it may be filtered). I think the system is complex enough that we can't truly rule it out just by inspecting the system.

3 - The two-ammeter test was suggested in the other forum and the reason I am pursuing it is to finally rule out (or in) mechanical load variation causes with higher confidence. For process induced oscillations the [envelope of the] three phase currents oscillate together. For rotor induced oscillations, the three phase current envelopes will peak at different times as the bad rotor location slowly moves in and out of poles of the rotating stator field (each stator field pole is closely associated with a single stator phase that contributes the majority of the mmf for that pole).

By the way, if I had "pdma" test equipment, they have facility to look for so-called swirl effect sidebands below 5*LF as means to distinguish between rotor cause and load/mechanical cause. But I don't have pdma or any dedicated motor test box. We just use our standard vibration analyser with current clamp-on probe connected as input. It works great and we view the spectra with same software that we use for our vibraiton. But the downside is there is nowhere enough dynamic range to separate from the noise floor those very low level swirl effect peaks below 5*LF. (I think pdma uses special analog or digital processing to help get around it somehow). Without the benefit of pdma or other special purpose motor test box, I think/hope that the two-ammeter method will be a good reliable old-school alternative.

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

 

[waross]

Thank you for the reply, Pete.
Yours
Bill

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

 

[edison123]

epete

Interesting. Never heard of the two ammeter method. Look forward to seeing what you find.

Muthu

http://www.edison.co.in

 

[electricpete]

(each stator field pole is closely associated with a single stator phase that contributes the majority of the mmf for that pole)

Wrong! My explanation at the end sounded good at the time I wrote it, but in retrospect obviously incorrect. (We can see that by comparing the number of poles and phases in a 2-pole motor) The phase most closely associated with a given rotating stator field pole varies at 60 times per second. I suspect a better explanation for this behavior (current envelope peaks at noticeably differerent times in each phase) will have to consider the reverse rotating components.


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