machine bearing fault detection

[JChec]

Hi,

I've just started researching bearing analysis to monitor machine conditions. My intial presumption was to just use FFT and look at the power spectrum generated from using accelerometers. Another technique is envelope detection. Does anyone have any good input and why envelope signals should be used over my intitail thought? 

Thanks

[Norbert_B]

This is a "classic" usecase for sound and vibration.

hope this helps,

Norbert 

[Jacco K]

I too am working on bearing fault detection. It seemed real easy at first .

---

JChec wrote:

Hi,

I've just started researching bearing analysis to monitor machine conditions. My intial presumption was to just use FFT and look at the power spectrum generated from using accelerometers. Another technique is envelope detection. Does anyone have any good input and why envelope signals should be used over my intitail thought? 

Thanks

---

The spectrum usually contains huge peaks at the axle rotation speed and it's harmonics, whereas the faults seem relatively minor. Apparently, after enveloping, these faults show up much better on the FFT. 

But I have to be fair here: I don't have the sound and vibration toolkit, so I haven't tried the enveloping and demodulation thingy yet.

I'm currently redesigning my program because I did some false assumptions, and coded my program not flexible enough. Good thing it's a school project, or it would've been an expensive project. Unfortunately that also means I'm not getting paid... 

Message Edited by Jacco K on 01-16-2009 04:15 AM

[Jacco K]

Something to keep in mind, is that there are always unwanted vibrations. You'll have to find a way to remove or ignore them.

Properly conditioning the signal before doing an FFT is probably the trickiest thing in bearing fault detection. 

Will your solution be used to monitor one particular machine, or will it be a more "general" application that's to be used by maintenance personnel? 

[JChec]

Hi Jacco

Thanks for your input, I'll try get a data from accelerometers on a few different motors and i'll get the envelope signal first before the FFT and see how that goes. Wanting to do analysis on a specific motor but will also soon need to develop a general purpose system too. I'll post my findings when I get some data.

Thanks,

Jevon

[Anil_Reddy]

Hi,

you may want to check   

http://www.spectraquest.com/tech/index.html

Anil

[JChec]

Hi Anil

Thanks, that is a really useful site!

[arsi986]

Hi dear all,

I am arslan and trying to develop software using labveiw for diagnosis of ball bearing inner, outer and ball faults... i am using cepstrum technique...my problem is i have developed some what software and now i wanted to check my software through data files....As i dont have any experimental setup, so can any one provide me good noise free data of healthy, outer race, inner race and ball defect...the data should be in time domain and of single accerlometer....

It would be great help for me,

Regards,

Arslan

[arsi986]

Hi dear all,

I am arslan and trying to develop software using labveiw for diagnosis of ball bearing inner, outer and ball faults... i am using cepstrum technique...my problem is i have developed some what software and now i wanted to check my software through data files....As i dont have any experimental setup, so can any one provide me good noise free data of healthy, outer race, inner race and ball defect...the data should be in time domain and of single accerlometer....

It would be great help for me,

Regards,

Arslan

[altenbach]

---

arsi986 wrote:

I am arslan ...

---

Duplicate post. Contined here.

[altenbach]

I am arslan ...