I need a help with correcting my result.

Hi,

 

I am simulating the breathing crack effect and having a problem with the result.

The following is what I have simulated,

In order to create a thin crack, I used the debond command.

The results I got are shown below.

No Crack:

Thin Crack:

As you can see, the result is good when there is no crack. However, the result for the cracked one is weird, it almost looks like noise. I am not sure but I think it is because of the numerical dispersion. Do you have any suggestion in order to fix this problem?

Thank you.

 

Best,

Parks

7 comments

  • Official comment

    Hi Parks,

    Apologies for the delay. The set up of your model seems correct, I have tried a couple of methods to curb the numerical dispersion but I am getting the same result. Your model is only 37.5 wavelengths thick and is meshed appropriately so numerical dispersion shouldn't be an issue (as you see in the model without the crack). My suggestion would be to try modeling the crack with a line of void elements to see if the issue persists.

     

    Best Regards,

    Chloe

    Comment actions Permalink
  • Hi Parks,

    Can you share the files via your prefered cloud sharing service and we can look into this.

    Best regards,

    Oliver

    Comment actions Permalink
  • Hi Oliver,

     

    I am copying and pasting the code since I don't use a cloud.

    mem mb 310 50
    mp omp 2 *

    titl BCS
    rest no

    symb thk_out = 80.e-3
    symb thk_in = 70.e-3
    symb wid_out = 120.e-3
    symb wid_in = 25.e-3
    symb wid_tr = 5.e-3

    symb #keycord x 1 0.0 $wid_out $wid_in $wid_tr $wid_tr $wid_in $wid_out
    symb #keycord y 1 0.0 $thk_out $thk_in $thk_in $thk_out

    symb box = $wid_out / 200

    symb #keyindx i 1 7 1 $box 1
    symb #keyindx j 1 5 1 $box 1

    grid $i7 $j5

    geom
    keypnt 7 5
    skew stnd
    ldef on

    matr
    type elas
    wvsp on
    prop L1 2690 6306 3114

    site
    regn L1

    boun

    data
    hist singaus * 'singaus.txt'
    end

    func hist singaus *

    calc
    strs
    strn
    pres aprs
    disp

    plod
    pdef press func
    vctr pdir 0 -1 0
    sdef press pdir $i3 $i5 $j4 $j4

    modl
    dbnd debond1 * 0
    vctr ddir1 0 1
    elem debond1 ddir1 $i2 $i6 $j3 $j3

    pout
    hist func
    hist pres $i4 $i4 1 $j2 $j2
    hist aprs $i4 $i4 1 $j2 $j2

    prcs

    grph
    plot matr
    term

    grph
    nvew 2 2
    line on
    set imag avi 100

    proc plot save
    exec 30
    grph
    bbox user $x1 $x7 $y1 $y5
    plot aprs
    plot dgap
    imag
    end$ proc
    proc plot 100
    term

    symb #get { labl } jobname
    symb #save symb.$labl

    stop

    The user defined function I used is,

    time = 0, 1e-8, 2e-8, ..., 3.998e-5, 3.999e-5, 4e-5

    Amplitude = (59.3*10^6)*sin((2.2305*10^6)*(time-0.000008))*exp((-9.147*10^10)*((time-0.000008)^2))

    Thank you.

     

    Best,

    Parks

    Comment actions Permalink
  • Hi, I was wondering when I will get the answer.

    Comment actions Permalink
  • Dear Chloe Allison,

     

    I have tried modeling the crack with a line of void elements. It shows clear signal. However, the problem is that my goal is to see the breathing (clapping) effect of the crack. If I model the crack with a line of void elements, it does not show the breathing (clapping) effect. Is it possible to make the void element crack to breath? Thank you.

     

    Best,

    Parks

    Comment actions Permalink
  • Hi Parks,

     

    I am in talks with the developers about your model. Do you have any experimental results that you wish to compare the simulation to? 

     

    Best Regards,

    Chloe

    Comment actions Permalink
  • Dear Chloe Allison,

     

    I was expecting to receive a signal similar to the figures below.

    Or

    However, there are difference in modeling. For the above figures, the initial state of breathing crack was open, whereas the initial state of mine was closed. Even though, there seems to be too much ringing effect in my result.

     

    Best,

    Parks

    Comment actions Permalink

Please sign in to leave a comment.

Didn't find what you were looking for?

New post