Code covered by the BSD License  

Highlights from
Morphological Analysis for Bearing Fault detection

4.8 | 8 ratings Rate this file 31 Downloads (last 30 days) File Size: 2.34 KB File ID: #37154 Version: 1.0
image thumbnail

Morphological Analysis for Bearing Fault detection


Santhana Raj (view profile)


Applies the selected morphological operators on the signal.

| Watch this File

File Information

Y= morph_analysis(sig,fault_fr,RPM) Applies the Mathematical morphology operation on the signal "sig". To be basically used for Condition Monitoring of Rotating Equipments.
The Morphological operations done are
 Gradient 1: Beucher Gradient(Dilate - Erode)
 Gradient 2: Signal - Erode
 Gradient 3:Dilate - Signal
It outputs a FFT graph of the signal along with the gradients. It also prints on the Command Window, about the availability of Fault in the specified frequency.

Required Products Signal Processing Toolbox
MATLAB release MATLAB 7.13 (R2011b)
Tags for This File   Please login to tag files.
Please login to add a comment or rating.
Comments and Ratings (12)
16 Jul 2015 Ashish Raj Sinha  
05 Jul 2015 Santhana Raj

Santhana Raj (view profile)

The signal has been taken from CWRU bearing databased. Just search for Case Western Reserve University, Bearing data center. You would get a good collection of vibration data with bearing faults.

If you want some more data, look for NASA's prognostics site.

Comment only
05 Jul 2015 Ashish Raj Sinha

Can u give the details about value of rpm for the given signal.Please provide basic information about details of signal used in this program.Sir is there any open source link from where we can load more signals for analysis.Please help.

05 Jul 2015 Ashish Raj Sinha

From the base paper i was able to guess it as 29Hz is it correct?

22 Apr 2014 Anwesh

Anwesh (view profile)

17 Apr 2014 Santhana Raj

Santhana Raj (view profile)

Ya, You are correct.
Fs- Sampling Frequency

N- No of points in FFT, preferably of 2^n format, less than the number of values in input

fault_tym=inverse of fault frequency

X- Fs*fault_tym. Can be proven that the result will be better if the number of structural elements taken is a multiple of Fs*fault_tym. Which multiple provides better result is being compared in this code. You can chk the ieee paper for further details...

Comment only
17 Apr 2014 Anwesh

Anwesh (view profile)

The programs works well for me
can you describe the terms Fs , N i.e N= 4*2048 , X in the program.I figured out Fs to be sampling frequency

20 Feb 2014 Santhana Raj

Santhana Raj (view profile)

@Hemo Khalifa,

The program works fine for me. I cant directly provide a example/demo here. But I have already given how to call the function and the meaning of the variable in the comment portion of the file. If you want to understand the code, You can check my page,

If you are facing any specific error in executing the code, Drop me a mail with the error, I would be glad to help.

Comment only
19 Feb 2014 Hemo khalifa

it's no working with me, can you give an example or add demo

Comment only
05 Nov 2013 Mike

Mike (view profile)

23 May 2013 tk

tk (view profile)

23 Jun 2012 n chow

n chow (view profile)


Contact us