At my Mill we have around 5000 running motors. We have programs in place for root cause elimination, thermography, vibration, automatic diagnostics for control circuits etcera, and we are in a start up phase with ultrasound. BUT for the electrical windings in the motors we have nothing. We only have a handful unplanned shuts per year due to motor failures, the vibration guys find most of them before they fail mechanically, but when they occur it´s very expensive. Do you have any suggestions for a cost effective method to find these few failures amongst all these motors. We are in a start up phase to prioritize all equipment due to criticality. We are thinking about testing Dr. Cardosos Extended Parks Vector Apporach as a method, it looks very promising.

This message has been edited. Last edited by: STORAnders,

quote:

Dr. Cardosos Extended Parks Vector Apporach

This is the first time I have heard of this. An internet search reveals papers available for purchase that describe the method.

Is there an actual product available to provide this type of testing?

Are there available sources that discuss the test methodology?

STORAnders

The 'Extended Parks Vector Approach' is a variation of Electrical Signature Analysis (I will see if I can find the ISO TC-108 SC-5 write-up for it). However, so far as we were informed in the standards development group, it was purely theoretical and had not been implemented outside of a lab. I would like to hear if someone has any experience.

The primary issue is that it will not detect winding faults until they are already in fault - no warning on larger machines.

If you are looking for insulation degradation, in particular inter-turn (winding shorts), for field testing, then you will want to look at either a low voltage testing method (ie: PdMA or ALL-TEST Pro) or a PD-Surge tester (ie: the Baker AWA line). If you are looking at much larger machines above 6kV, you may want to consider PD monitoring (Iris Power Engineering), of which they also have newer systems that will detect problems a few weeks out as low as 2.3kV.

In any case, my recommendation would be to have the reps for each of the manufacturers come out and demonstrate their technologies for the following:

1. Ease of use - how much training is required to implement;
2. Accuracy - can the equipment identify the faults that it claims; and,
3. Portability.

Make sure during the demonstrations that your people get to operate the equipment. Test the tech support for each company, as well.

Sincerely,
Howard

Here is a portion of the statement within the ISO proposed standard:

quote:

The EPVA combines the simplicity of the former Park's Vector Approach and the detailed insight given by the
spectral analysis. Moreover, it eliminates some of the technical limitations of the conventional motor current
spectral analysis. In fact, it is well known that it is sometimes difficult to filter out the fundamental component
of the motor supply current without affecting the sideband components. In contrast, the characteristic
components in the EPVA spectrum can easily be separated since the fundamental is far apart and its
amplitude is automatically reduced by the Park's transformation. Additionally, by taking into account the
current in all three phases, the EPVA provides a more meaningful spectrum than the one obtained by using
the conventional motor current spectral analysis.

The details of the explanation primarily describe how it eliminates the problems in current signature analysis having to do with the fundamental frequency - in effect, disregarding demodulation. There are also no pass/fail limits applied to the faults that it claims to detect.

I cannot post the remainder of the explanation due to the international copyright limitations associated with the draft standard.

Howard

There are no commercial products available yet. I and one of my colleagues have participated at lectures by doctor Cardoso at Vaxjo University and the results looks promising. He showed several cases where he had found motors with faults. One case at a cement factury was specially interestising. He found a stator windning fault on a big motor that the owners where totally unaware of. On this case he also showed photos of the winding damage, I have no copys of these photos. The last time Dr. Cardoso visited Vaxjo my colleague spoke to him. Dr. Cardoso then revealed that he has developed his formula further but he has not yet published his new findings. Among other things he thinks that the latest version of his mathematic fomula can find problems in frequency converters, e.g. ageing capacitors.

quote:

Originally posted by MotorDoc:
STORAnders

The 'Extended Parks Vector Approach' is a variation of Electrical Signature Analysis (I will see if I can find the ISO TC-108 SC-5 write-up for it). However, so far as we were informed in the standards development group, it was purely theoretical and had not been implemented outside of a lab. I would like to hear if someone has any experience.

The primary issue is that it will not detect winding faults until they are already in fault - no warning on larger machines.

If you are looking for insulation degradation, in particular inter-turn (winding shorts), for field testing, then you will want to look at either a low voltage testing method (ie: PdMA or ALL-TEST Pro) or a PD-Surge tester (ie: the Baker AWA line). If you are looking at much larger machines above 6kV, you may want to consider PD monitoring (Iris Power Engineering), of which they also have newer systems that will detect problems a few weeks out as low as 2.3kV.

In any case, my recommendation would be to have the reps for each of the manufacturers come out and demonstrate their technologies for the following:

1. Ease of use - how much training is required to implement;
2. Accuracy - can the equipment identify the faults that it claims; and,
3. Portability.

Make sure during the demonstrations that your people get to operate the equipment. Test the tech support for each company, as well.

Sincerely,
Howard

A big guy from ALL-test USA visited us a few years back to present theír instruments. It all looked very promising until we served him with 5-10 motors which had different types of problems, and we knew what the problems were. We let the guy test the motors with his instruments and I can say that we were not impressed at all by the result he came up with. We didn't buy any instruments from ALL-test. Thats were we stand today with our Meggers and cheap motor testers from a Swedish manufacturer.
For our high voltage machines we have a PM- program in place in cooperation with ABB.

I can't say much about the operator of the ATPro instrument. I've had pretty good luck with it, as well as the PdMA and Baker AWA, for off-line testing. Guess they should use someone with some experience

However, did you evaluate any of the MCSA or ESA instruments from any of the manufacturers?

I am still interested in hearing who has had luck with the Parks Vector approach. Is there an instrument developed with it, or is the professor the only one who knows how to do it?

Howard

quote:

Originally posted by MotorDoc:
I can't say much about the operator of the ATPro instrument. I've had pretty good luck with it, as well as the PdMA and Baker AWA, for off-line testing. Guess they should use someone with some experience

However, did you evaluate any of the MCSA or ESA instruments from any of the manufacturers?

I am still interested in hearing who has had luck with the Parks Vector approach. Is there an instrument developed with it, or is the professor the only one who knows how to do it?

Howard

The ALL-test guy were supposed to be one of their most experienced technicians.

We only evaluated a couple of MCA-instruments. We didn't look at any on-line instruments.
On our high voltage motors ABB uses a ESA-approach when they test them, but we are talking of about 20 motors or so. I am searching for a cost effective method to test the other 4980 running motors, and to take in consideration that we have very few failures per year.

Dr. Cardoso says that there are no commercial products at the moment. We have plans to feed his formula into an ABB Argus measuring system and to try to evaluate it both on faulty motors that come in to our work shop, and on a pulper that are known to break motors. I can't tell when this evaluation will be finished because the engineer that I planned to do the evaluation are leaving us very soon.

This message has been edited. Last edited by: STORAnders,

The comment was made tongue-in-cheek. When I left ALL-TEST Pro as their general manager on January 1, 2005, the next person along had only a month or so experience. So, they were right, I am sure. Regardless of ease of use, it does take some time to get experience with any of the technologies.

But I am not here to take a position on ATPro. If they wish to defend their product, then they can make comments in this forum. My statement was simple: I have had good experiences with the ATPro, PdMA, Baker, Areva, and a few others that will be on the market soon.

As there are a number of concepts and technologies that have entered the market only to die a slow, or fast, death. Simply put, it sounds like you will be putting the technology in place. I am sure I will, as will others on this forum, be interested in the results, findings, and learning curve. It may be the next best thing - it just hasn't been applied outside of a controlled environment. It will also be interesting to understand its limitations, in addition to the capabilities.

Howard

When searching for more information on the Extended Park’s Vector and the Dr. Cardosos method of analysis I found an excellent paper on the net on that topic. The beauty of it was, that it was for free. You can check on the paper at: http://www.aedie.org/9CHLIE-paper-send/339-melero.pdf
The authors examined a 5.5 kW motor with 1 artificially shorted turn, which represented 0.85% of all turns in that phase. A very little number. Yet they were able to see the fault quite clearly by seeing the negative sequence both in the voltage and the current. They mention the Park’s Vector, but I don’t know if they actually used it.
In order to see the progressive deterioration of the winding, they came up with the idea of the “ageing cycles”. The ageing cycle was a startup, run for 10 minutes and then shut down (Hmm, is it a typical loading of an induction motor? I do not think so). The idea was this: After each ageing cycle the insulation will get worst, and they could see the deteriorating process on negative sequence current and negative sequence voltage. They went through 16 cycles and no deterioration was found.
They also found that the current through the shorted turns increased from 3.2 Amps to 40 Amps, by the factor of 13. In other words the losses in the shorted turns increased 13^2 = 169 times. The 10 minutes was probably a good choice, because if they tried to run for 20 minutes, they would end with a hole in the winding and the paper would be in jeopardy.
It just confirms the experience. If you want to find a short in the winding, no matter how fast you are, you have very little time left (minutes, seconds).
Also: If you find a short, the motor has to be rewound. On the other hand, if you do not find that short, the motor has to be rewound.
Did anybody found anything interesting on Extended Park’s Vector?
jank

quote:

electrical windings

Have a look at http://www.psdinsight.se
We can help you in a number of ways. Please contact us for more information.
\Ã…ke

It would be very interesting to hear from people on other industries with many motors running. Do you have a systematic approach to prevent your motors creating unplanned shutdowns. If so, can you explain the process.

Vendor.

Hi Anders,

We can now offer a new ESA based and on-line technology combination including both vibration and ESA possibility from one of our new major suppliers. Special is that it primarily focus on permanently installed continous monitoring but it could be used portable with the limitations that the sensors would most likely need to be installed permanently. There are solutions to make that wired to a connector in the cubicle. This technology was developed for nuclear RCP units and will now be available for in the frist step asynch AC motors expanding to VFD and generators. System/software is now undergoing testing and will be available after that finished. We also have our own flux based technology that aim at being included in normal vibration route activity that can be used as a tool to indicate what motors have a worsened status and need further analysis or replacement. You are welcome to inquire further. Olov