Debugging help Answered

Hi team

I have a model set up to inspect the microstructure of titanium and there is definitely some issues when observing the A-scan from the receiver and pulse echo response. For my first test case, I am using a homogeneous Titanium block from the material projects in OnScale. 

I can't see any issues with my model from where I am transmitting and receiving. My main concern is perhaps I am listening to the Ultrasonic waves in inside the water as I have my transmitter/receiver defined on the edge of a boundary between water and the material. The way I have this set up, I am not sure if it can be done better or not. 

I am wondering if anyone can check my model just to see if there is anything sticking out that may be wrong. In the pulse echo response from the homogeneous block, I am seeing signal noise which should not be present, so I am trying to solve this problem first so I can be confident when inspecting my microstructure.

Thanks

Dave :) 

5 comments

  • Hi David,

    Please send us your files via your preferred cloud sharing file service I can take a look for you. 

    Thanks,
    Chloe

    Comment actions Permalink
  • Thanks Chloe for looking at this, it really appreciated. 

    Couple of things to briefly show you. My full A-scan (below) shows something is off. I would expect the first echo around 3.7 microseconds to have the largest amplitude. I have no idea why this last signal is so large. There is definitely  still artefacts arising from the boundary conditions which to my knowledge is limited in the way we can deal with them.  Hopefully there is something you can notice in my script and perhaps I am not implementing things correctly.

    Originally, my water buffer at each side of the domain was only 1 mm, but I changed this to 10mm so the source is further away from the boundary and tried implementing the  following command to help absorbing boundary conditions.

    prop watr 1000. 1000. 0.
    hrgl watr visc 0.001

    Once I have this model fixed and confidence in it, this will allows be to confident when analysing my EBSD structure. This model I am sharing is just the homogenous Titanium block from OnScale project materials.

    You can access my files via the following link and the password is just  onscale

    https://strath-my.sharepoint.com/:f:/g/personal/david_harra_strath_ac_uk/Ete8RsACGGNKnThZ8dVPgSsBqHcdqjXEPaP-Ll36g-xHOQ?e=umoUgd

    Thanks

    Dave :)

     

    Comment actions Permalink
  • Hi Dave,

    Typically you want the BCs to be at least 3 wavelengths away from the source or any receive outputs so no reflections occur but this many vary depending on the frequency. I updated your model to add the buffer in y as well as x and updated all BCs to absorbing. Absorbing boundaries have a higher performance in general and what we use unless the model is unstable. Doing this results in a 3.7 us echo signal with the largest amplitude. 

    I also added code in to plot the receive A-scans in the video so you can more clearly see where the signals are coming from in the model. Here is the file.

    I hope this helps!

    Best Regards,

    Chloe

    Comment actions Permalink
  • Thank you so much Chloe, this gives a meaningful response and seeing these A-scans in live time is really helpful.

    Just one question if  I may.

    When you refer to the BCs being at least 3 wavelengths away from transmit/receive sources. Is the wavelength defined in models by, so 3 times this distance away?

    symb wavelength = $vel / $freqint 


    Thanks

    Dave :)

    Comment actions Permalink
  • Hi Dave,

    No problem at all. Yes that is the shortest wavelength in the model and what you are meshing for. See this article for more details.

    Best Regards,

    Chloe

    Comment actions Permalink

Please sign in to leave a comment.

Didn't find what you were looking for?

New post