Fault Diagnostics of Induction Motors

Get a refund from your school. This is your final scholastic project before joining the work force. Your school was supposed to do two critical things for you; give you enough theoretical knowledge so that you can tackle real world problems and confidence in your knowledge and learning ability that you can figure out a problem on your own. It appears they have failed you in both endeavors.

Use the same trouble shooting guides you have for 3 phase but throw two phases out. Think about what you have learned about 3 phase vs single. should be easy to convert your knowledge to the single phase task at hand.

lol... thank u redfred.. i would've surely considered ur advice if it were possible.. but for now... i am but in a little trouble and hurry.... so can u please help me c through it?

thanks

Ok, here's some guidance but you will still have to do most of the work yourself. All motors are electro-mechanical devices. So I would start with generating a list of failure symptoms that you've encountered that a motor can generate. Don't worry about the really rare esoteric failures, just list the common failure modes you could expect a motor may present. After you've listed them, see if any can be grouped together. (Buzzing and grinding could both be considered a vibration.) Now generate a two dimensional cross reference table with one axis (row) being the anticipated failure modes and the other axis (column) with the listing of electrical and mechanical. In each box of the electrical column identify what electrical failure could cause each motor failure. Similarly identify what mechanical failure could cause each motor failure. Now examine your answers and possibly add a third column where a combination of electrical and mechanical failures could work together.

So when is this project due, June 1?

:).... NO SIR i've just started it... this is the 4th week and i've to submitt it in the 18... so i've quite some time in hand....

Data acquisition setup will be built to record current in the single phase induction motor. The motor will be operated under a wide range of operating conditions and baseline data will be collected. Then, faults such as broken bars, bearing faults will be physically injected in the machine and the data will be gathered under the faulty conditions. The two types of data sets, i.e., baseline and fault will be used to develop the fault diagnostic features. Finally, a diagnostic algorithm will be built for online diagnostics of any induction motor\

the final line in bold is my final objective... what i actually have to present...

the faults will be known... so will be their effects on the current spectrum... but i'm not sure whether the technique that i'm using for the diagnostics i.e motor current signature analysis works equally good for single phase induction motors also??... cuz according to my research on the internet motor current signature analysis is used for 3 phase induction motors.

:).... NO SIR i've just started it... this is the 4th week and i've to submitt it in the 18... so i've quite some time in hand....

Data acquisition setup will be built to record current in the single phase induction motor. The motor will be operated under a wide range of operating conditions and baseline data will be collected. Then, faults such as broken bars, bearing faults will be physically injected in the machine and the data will be gathered under the faulty conditions. The two types of data sets, i.e., baseline and fault will be used to develop the fault diagnostic features. Finally, a diagnostic algorithm will be built for online diagnostics of any induction motor\

the final line in bold is my final objective... what i actually have to present...

the faults will be known... so will be their effects on the current spectrum... but i'm not sure whether the technique that i'm using for the diagnostics i.e motor current signature analysis works equally good for single phase induction motors also??... cuz according to my research on the internet motor current signature analysis is used for 3 phase induction motors.

and yess.... i can assure u that i'll be doing all my work myself...:) infact i wouldnt have taken this route if my project supervisor/advisor was here.. unfortunately he has left for singapore for some sort of short course and wont be back before 3 months... so this is the main reason i'm seeking your guidance..

Ok, two things seems to missing in your set up. First, what type of single phase AC motor will you be using to test? Often the difference will be seen in the starting technique of one of these motor types; shaded pole, capacitor start (either split-capacitor or capacitor-start capacitor-run), universal motor, split-phase (centrifugal switch). Second, your measurement instrumentation must be capable of measuring the phase angle between current and voltage. Measuring just the amplitude will lead to failure in your analysis.

Hey fred!!..

The motor that has been given to me was installed on a pedestal fan. the capacitor which is there in it is a 2.5 uF +- 5%, 450 v capacitor and runs at a max of 85 deg celcius. the ratings strongly indicate that it is a motor run capacitor but then i am confused because this is the only capacitor in it... THERE IS NO STARTING CAPACITOR IN IT!!.. i mean there must have been a starting capactor with it right?

Ok, you are correct that you've yet to identify how the motor starts to spin in a particular direction. But not all electric motors require a separate starter circuitry. Did you look for a shaded pole? Since you have a capacitor somewhere in the motor windings, this tells me that your system is designed to work with a single phase power source but not your motor. If you do not have a centrifugal switch in the motor then the capacitor windings are intended to bias the direction of rotation. My point here is that strictly speaking your motor requires two voltage phases applied and not just one. If you have a centrifugal switch then this is only required during start up and momentum will do the rest. I will answer your specific question. No, there need not be a "starting" capacitor for a motor to be operated from a single phase voltage supply. You just need to know the starting mechanism. Some old motors and many with failing starter circuits require an external push in one direction for them to start. After that they will run fine, but possibly backwards.

Here's another question for you to consider. Since this motor works in a pedestal fan, when will the motor torque need to be at a maximum? Similarly, what will be the mechanical loads on the motor during certain parts of operation?

Lastly, you did not answer my critical question. Will your instrumentation be capable of measuring the current to voltage phase angle?

Did u see the pictures i've attached to the thread in post 13??

no sir... all i've been given is this motor and Data aquisition card NI 6025 E of the national instruments... it can only measure the voltage... not even the current...

so i've planned to take the stator current, pass it through a currrent to voltage converter circuit and then give the output to the DAQ... this is the only possible way of recording the voltage...

and if we have the voltage... we can easlily convert it into current... and then analyse its frequency spectrum in MATLAB...

thats all i could think off at hand.... what do you think about it?...

and what can we do for the fact that this DAQ card cannot measure the phase difference??

Yes, I've seen the pictures.

One question I forgot to ask earlier, what academic program is this project part of? In other words is this part of a BS degree in Physics, Computer Science, Electronic Technology, Electrical Engineering, or is this a pre-college class?

You do realize that your DAQ card accepts only up to ±10V. So you will have to accurately attenuate your voltages. This is a trivial design requirement but ignoring this will damage your acquisition board.

As for measuring the current phase, this should be easy to derive from the time based current and time based voltage readings. You've mentioned the use of a current to voltage converting circuit a few times, what type of circuit topology do you intend to use?

i'm doing Bachelors in avionics engg.... with almost the same courses taught in electrical engg...

yes.. i realize it fred... the last thing i want is to upset my faculty memebers by damaging any equipment given to me.....

i intend to make a current to voltage converter using op amps on a bread board... then give the input to this circuit and its out to the Daq card's input..

input current would be constant.. so we can limit the output voltage by limiting the value of resistance.... also that opamp response in excellent upto very high frequencies...

Your principle for current measurement is a sound and a frequently used configuration for very small currents but you're going to have considerable trouble implementing this. You see all of the op-amps that will fit on your prototype board will only be able to drive much less than an ampere of current. I would expect that the inrush current in starting your fan will quickly overdrive your circuit. Now that I think about this a little more, there is a company that makes a high current op-amp like supply but buying one of these would quickly consume your whole budget for this project. A more cost effective approach will be to use a current transformer. You can buy one of these for less than $10US or your EE department might loan you one so that you don't blow up their acquisition board. There still is some very good reasons to also use a normal op-amp to condition the various signals to your DAQ so I am glad that you suggested your approach.

I doubt that you will be using a variable frequency drive to operate your fan under test. (The high current op-amp could easily be configured as a variable frequency driver.) This does bring to my mind another question that I need to know to give you proper advice. What type of power will be running your fan? Do you have 120VAC 60 Hz power with a grounded neutral or 220VAC 50 Hz split phase? If you are not precisely sure just tell me where you are and I can look this up.

One thing I forgot to mention, your DAQ board has many inputs available to it and one thing that I think would be very good to monitor will the angular velocity of the motor shaft. Have you considered this at all?

No, i didnt think about it!!.. yes it will be very good... but honestly i didnt know and still am not sure that how can i do it with my DAQ board??? can you please tell me how can i do it with this board and how will it help us doing our motor current signature analysis??

and one more thing sir, if it doesnt bother you, can i have your e mail address on either hotmail or yahoo etc so that i can talk to you directly??

yes i can do that too. But just asking.. that we're only interested in the "fluctuations" in stator current and its frequency spectrum, cant i just place a resistor in series with the stator current and then feed it to my I-V converter circuit??.. Although the amplitude of current will decrease due to this resistance but it wont effect the variations in the current! and neither would this decrease in current amplitude have any effect on its frequency spectrum !

No sir, i wont be using variable frequency drive. and the place where i live has 220V 50 hz supply. and i'm sure about it

Split phase or with a neutral?

definately not sure about that! :(

Well let's assume that you have a split phase power distribution. That will mean that each lead of your fan will have 110VAC referenced to ground. The inverting node of your current detect op-amp circuit is a virtual ground. Connecting this to either 110VAC lead of a split phase power grid will quickly release the inherent smoke of the op-amp. You will probably not damage the acquisition board but I would not risk it. If you still wish to install a shunt resistor and measure the differential voltage across the resistor, then you'll need to design a high common mode rejection amplifier that will not be damaged by the line voltages. This can be done, but since you will not be measuring DC current the current transformer will be so much easier to implement.

As for my e-mail address, I'll send you a personal message later today or Tuesday with my work e-mail address. (Monday is a holiday here.) I'd like to discuss with your professor though how much or little assistance he'll accept from me. I'd certainly hate for you to fail because I got too helpful.

Seems very convinient.... I've contacted a number of transformer suppliers. Lets see now who responds first. I'll buy it as soon as i can.

And about contacting my project advisor, i'll be glad if u would. Believe me sir you helping me wouldn't be any problem because afterall i'll be doing all the things practically myself. You would be guiding and helping me understand things better though. And if any thing, it'll just be a plus point to my project and understanding.

But then again sir, let me remind you that I have my mid-term presentation on 22nd of june in which i'll show all of my faculty what i've already done and have to convince them that i've progressed much! But what i've been doing all this time is studying the related faults in induction motors and techniques perform its diagnostics. MCSA has to be my final solution.

So please get me started already! Time is literally flying by!

i'm attaching this weekly program that i've told the faculty that i'll follow

According to this i should have had made the DAQ setup, taken baseline data of healthy and faulty motors both! So as you can see that i am seriously lagging behind and want to get things going.

To danishrizvi (only);

Just curious, have you worked out what kind of motor that is yet? What the capacitor is for? How it all works?

Hey, I'm not holding you up. You came here already four weeks late. That's not my problem. As I said in my first posting, you're school believes that you already had enough training to be able to do this project. The only one late here is yourself.

GOOD news!!. i've been permitted to buy a current transformer for my work!. i'll be off for this week end.. so i'll bring it this time...

one more thing.. my motor says that its a 220V 125W motor... doesnt it mean that the current that it will draw would already be less than 1 amp??.. about .5 A ???

Yes, your analysis is correct. Your fan will draw less than an ampere of current RMS. For your analysis though you will want to keep the peak to peak currents inside your data acquisition range.

so what rating transformer should i buy... and what should be the primary current to secondary currrent ratio??

Now this gets to the specifics of how you will design your experimental apparatus. Depending on how much and what type of signal conditioning you will be using, you could use almost any transfer ratio. One of the critical things that you need to be aware of is that there is a maximum linear current transfer range for the transformer itself. I would expect that your fan will never draw this amount of current but you should be aware that a limit here does exist.

The next critical item for you to consider will also determine which transformer to choose is the faults and expected fault currents will you be measuring. Now you may wish to have different circuit gains to maximize resolution of your acquisition board. If my memory is correct, your acquisition board has 12 bits of resolution so this may not be critical.

The glad tidings are that i've been successfull in doing all of my circuitry work. i've included the circuit diagram in this reply.

i'm using the LAbview software to analyse the voltage levels and frequency spectrum. the voltage changes are coherent to that i've been getting on the VCO. so this means that my interface of the motor and computer through the DAQ is correct.

But the problem is that i'm not getting the correct FAST FOURIER TRANSFORM (fft) waveform correctly. As the voltage signal input to the card is 50 hz, i reckon that the fft(frequency spectrum) should also contain the maximum peak at 50hz!.. but its not coming that way. What should i do about it?

You're correct that you should have 50 Hz as a dominant frequency. I don't know what might be your problem with your data acquisition system. My initial idea is that you may not have a large enough sample array size to include 50 Hz in the analysis output. But the actual root of your problem could be anything from damaged electronics to simply reading the wrong data location in your software and all steps in between. I have a few suggestions though that should help you to troubleshoot your set up. Include a sample versus time window in the display of your software for troubleshooting. Route your data stream to this display so you can see the array of data points entering the FFT analysis block. If this display does not include several cycles of your 50 Hz sine wave then you cannot get an FFT analysis that reflects 50 Hz. If I was writing your code, I would not bother to include an FFT analysis block until I was sure that I can capture an accurate time based array of data. Then I'd make sure that it is a suitable set of data to get an FFT analysis. Only once I knew the data was acceptable would I attempt to utilize an FFT module.

you say that you wouldnt have included the FFT analysis if you couldnt capture time based array... so does it mean that there are other ways and options also to do the fault diagnostics of my motor??. if so please let me know about it. i'd be glad if there are other options i could ponder about!

Of course there are other options for troubleshooting a motor failure. FFT analysis is but one of many tools a technician has for troubleshooting anything. Don't get me wrong, I'm not trying to persuade you away from ultimately using FFT analysis. My point is that an oscilloscope like display of an array of time based data points for me tells me if I'm looking at the correct data set. Since your FFT display in your software is generating a frequency based display that disagrees with what you expected, you need to identify which of the three possible scenarios are true:

1. The FFT analysis is correctly analyzing and displaying the wrong time based data array (waveform).

2. The FFT analysis is not performing a correct analysis of the correct time based data array.

3. The FFT analysis is correctly analyzing the correct data but you're not displaying it properly or you misunderstand what an FFT analysis produces.

My suggestion to use a time based display to troubleshoot your program is because the initial data capture must be a time based capture. A time based display in essence permits you to quickly observe the raw data.

danishrizvi one more tip, we don't use SMS TXT in this forum, etiquette and respect go a long way.

For general information I directed danishrizvi to this forum as I was unable to help him, I just hope the knives will be put way and some help is forthcoming.

Sure Tony.... i agree to it!!... and i will take care of it..

yes i get what u say.... but those capacitors are PERMANENT SPLIT-CAPACITOR right?... AND this one doesnt looks like it!!!...

to be honest sir.. i'm a little worried because i haven't started the project as yet... i mean yes i have done almost all the initial literature review and with your help know my pre requisites.... but still i need to start doing things practically now....

my mid semester presentation will be held in the 5th week from now...and i need to convince my faculty that i've not been sitting idle and have worked on something atleast... so i need to get things started under your advice....

so please tell me that what should i do in what order.. i think i've made my approach very clear to u that how would i proceed with things... but need u to further tell me what step to do first and what to do after that...